a comparison of albumin and saline for fluid resuscitation ... › 2020 › 03 › ...(116 [12%] vs...

For personal use. Only reproduce with permission from The Lancet Publishing Group.

ARTICLES

THE LANCET • Vol 357 • May 5, 2001 1385

Summary

Background The beneficial effects of �-blockers on long-term outcome after acute myocardial infarction were shownbefore the introduction of thrombolysis and angiotensin-converting-enzyme (ACE) inhibitors. Generally, the patientsrecruited to these trials were at low risk: few had heartfailure, and none had measurements of left-ventricularfunction taken. We investigated the long-term efficacy ofcarvedilol on morbidity and mortality in patients with left-ventricular dysfunction after acute myocardial infarctiontreated according to current evidence-based practice.

Methods In a multicentre, randomised, placebo-controlledtrial, 1959 patients with a proven acute myocardialinfarction and a left-ventricular ejection fraction of �40%were randomly assigned 6·25 mg carvedilol (n=975) orplacebo (n=984). Study medication was progressivelyincreased to a maximum of 25 mg twice daily during thenext 4–6 weeks, and patients were followed up until therequisite number of primary endpoints had occurred. Theprimary endpoint was all-cause mortality or hospitaladmission for cardiovascular problems. Analysis was byintention to treat.

Findings Although there was no difference between thecarvedilol and placebo groups in the number of patients withthe primary endpoint (340 [35%] vs 367 [37%], hazard ratio0·92 [95% CI 0·80–1·07]), all-cause mortality alone waslower in the carvedilol group than in the placebo group (116 [12%] vs 151 [15%], 0·77 [0·60–0·98], p=0·03).Cardiovascular mortality, non-fatal myocardial infarctions,and all-cause mortality or non-fatal myocardial infarctionwere also lower on carvedilol than on placebo.

Interpretation In patients treated long-term after an acutemyocardial infarction complicated by left-ventricular systolicdysfunction, carvedilol reduced the frequency of all-causeand cardiovascular mortality, and recurrent, non-fatalmyocardial infarctions. These beneficial effects areadditional to those of evidence-based treatments for acutemyocardial infarction including ACE inhibitors.

Lancet 2001; 357: 1385–90

*Members listed at end of paper

Correspondence to: Prof H J Dargie, Department of Cardiology,Western Infirmary, Glasgow G11 2NT, UK(e-mail: [email protected])

IntroductionThe efficacy of �-blockers in reducing majorcoronary events and improving short-term andlong-term outcome has established their beneficialrole in the management of acute myocardial infarction.1–3

However, the randomised trials whose results showedthese effects were done before thrombolysis or primaryangioplasty were used for reperfusion, and beforethe introduction of angiotensin-converting-enzyme(ACE) inhibitors. �-blockers have now been shownto reduce mortality and morbidity substantially,and improve left-ventricular function in patients withchronic heart failure when given together with ACEinhibitors.4–7

Since coronary heart disease is a major cause ofheart failure, attention has focused once more on the useof �-blockers in patients with acute myocardialinfarction. Although there have been many randomised,placebo-controlled trials of � blockade in acutemyocardial infarction, none has studied patients withconfirmed left-ventricular systolic dysfunction whomight also have had clinical evidence of heart failureduring the index hospital admission. Post-hoc subgroupanalyses of previous trials, however, have suggested asimilar mortality benefit in patients with heart failure.8–10

Conversely, several trials of ACE inhibitors haveconclusively shown substantial improvement in mortalityand morbidity in this group of patients.11–13

Registries from Europe and the USA indicate thatthe use of �-blockers in eligible patients post myocardialinfarction is substantially lower than would be expectedfrom the convincingly positive results of the oldertrials.14,15 One explanation for this finding could be theabsence of contemporary data from trials in the post-thrombolytic era, specifically in patients who havesubstantial left-ventricular dysfunction, who might alsohave heart failure, and in whom ACE inhibitors will havebeen prescribed. We designed the Carvedilol Post-Infarct Survival Control in LV Dysfunction(CAPRICORN) study to test the hypothesis that theaddition of carvedilol to standard modern managementof acute myocardial infarction in patients with left-ventricular dysfunction with or without heart failurewould improve outcome in terms of mortality andmorbidity.

Patients and methodsPatientsThe CAPRICORN study, whose design and protocolhave been published elsewhere,16 was a multicentre,double-blind, randomised controlled trial of carvedilolversus placebo involving 17 countries and 163 centresworldwide. Eligible patients were aged 18 years or olderwith a stable, definite myocardial infarction occurring

Effect of carvedilol on outcome after myocardial infarction inpatients with left-ventricular dysfunction: the CAPRICORNrandomised trial

The CAPRICORN Investigators*

Articles


3–21 days before randomisation. Other inclusion criteriawere: left-ventricular ejection fraction of 40% orless by two-dimensional echocardiography or byradionuclide or contrast ventriculography, or wall-motion-score index of 1·3 or less; and receipt ofconcurrent treatment with ACE inhibitors for at least 48 h and stable dose for more than 24 h unless there wasproven intolerance of ACE inhibitors. We includedpatients who had heart failure appropriately treatedwith diuretics and ACE inhibitors during the acutephase, but excluded those who continued to requireintravenous diuretics or inotropes, or who haduncontrolled heart failure. Unstable angina, hypotension(systolic blood pressure <90 mm Hg), uncontrolledhypertension, bradycardia (heart rate <60 beats permin), and unstable insulin-dependent diabetesmellitus were further reasons for exclusion. Patientswith a continuing indication for �-blockers for anyclinical indication other than heart failure wereexcluded, as were those requiring ongoing therapywith inhaled �2-agonists or steroids.

Informed consent was obtained from all patients, andethics committees from all participating countries gavetheir approval.

MethodsPatients were randomly assigned carvedilol or identical-looking placebo by use of permuted blocks withstratification by centre. Study medication was uptitratedto the highest tolerated dose for each patient, to amaximum of 25 mg twice daily. The initial dose of 6·250 mg, if tolerated, was continued on a twicedaily basis. If it was not tolerated, the same dosewas readministered or reduced by half. If that dosewas not tolerated, the patient received no studymedication, but was followed up anyway. Aftersuccessful initial dosing, the patient returned as anoutpatient every 3–10 days for assessment of tolerabilityand further uptitration. In the absence of adverseevents or evidence of clinical heart failure, andif the heart rate was greater than 50 beats per min andthe systolic blood pressure greater than 80 mm Hg, thedosew as increased to the next level. The patientremained in the outpatient department for 2 h to ensureno side effects ensued.

During the maintenance period, patients werereviewed every 3 months during the first year, and every4 months thereafter. Investigators were encouraged toreview the dose of study medication at each visit and toensure that doses of other drugs, especially ACEinhibitors, were adjusted accordingly to ensure optimumdose levels. At specified visits, an electrocardiogram wasdone, New York Heart Association class ascertained,and venous blood taken for routine biochemistry andhaematological analysis.

The maintenance phase continued until 633 validatedprimary endpoints had occurred, whereupon down-titration began. All patients had a minimum of 3months’ follow up. Study medication was withdrawnin a stepwise manner over 1–2 weeks, decreasing onedose level at a time every 3–4 days. Subsequentuse of open-label �-blockade was at the discretionof the investigator.

The original primary endpoint was all-cause mortality,but, during a masked analysis, the data and safetymonitoring board noted that overall mortalitywas lower than had been predicted and that thestudy could not be completed with the sample sizeand power originally planned. The steering committee

therefore decided to adopt co-primary endpointsof all-cause mortality (the original primary endpoint),together with all-cause mortality or cardiovascularhospital admissions (the first prespecified secondaryendpoint). The other secondary endpoints weresudden death and hospital admission for heartfailure. Other outcomes assessed were recurrentnon-fatal myocardial infarction, and all-causemortality or recurrent non-fatal myocardialinfarction.

The trial was overseen by a steering committee, whichmet monthly by teleconference and at face-to-facemeetings at least twice a year. An endpoints committeewas responsible for masked adjudication of allprespecified endpoints, which were described in detail ina manual of operating procedures agreed by the steeringcommittee.

Statistical analysisThe target sample size was revised on the basis ofthe new co-primary endpoint of all-cause mortality orcardiovascular hospital admission. For 90% power,and assuming a hazard ratio of 0·77, we calculatedthat recruitment of a minimum of 1850 patientsrandomised on a one/one basis with 633 deathsor cardiovascular hospital admissions would berequired. In view of the advice from the data andsafety monitoring board concerning the mortalityrate, we decided to divide the �=0·05 adopted forthe previous primary endpoint of all-causemortality alone into 0·005 for all-cause mortality and0·045 for all-cause mortality or cardiovascular hospitaladmissions.

All analyses were by intention to treat. The basis of allprincipal analyses was time to first event, and results were assessed by the log-rank test and quantified byhazard ratios and 95% CIs, calculated with Cox’sproportional hazards model.

ResultsWe recruited 1959 patients, of whom 975 were assignedcarvedilol and 984 placebo, and who were followed upfor a mean of 1·3 years (figure 1). The trial continuedto its planned conclusion when 633 primary endpointshad been validated. In total, 707 such eventswere judged by the endpoints committee to haveoccurred.

Baseline characteristics, which were similar betweenthe two groups, are shown in table 1. The mean left-

ARTICLES

1386 THE LANCET • Vol 357 • May 5, 2001

1959 patients randomised

984 assigned placebo

151 died174 withdrew permanently

116 died192 withdrew permanently

975 assigned carvedilol

659 patients on treatment at end of study

667 patients on treatment at end of study

Figure 1: Trial profile


ventricular ejection fraction was 32·8%, and intravenousdiuretics and nitrates were required in about a third andthree-quarters of patients, respectively, as treatment forthe acute event. Reperfusion therapy, mainly bythrombolysis but also by primary angioplasty, wasapplied in 46% of all patients. The site of the indexmyocardial infarction was anterior in 57%, and therewas a history of previous myocardial infarction or anginain 30% and 56% of all patients. Hypertension, diabetes,or hyperlipidaemia were present in 54, 22, and 33%,respectively.

Of the 940 patients who entered the maintenancephase in the carvedilol group, 692 (74%) reached themaximum attainable dose of 25 mg twice daily, and 103(11%) and 65 (7%), respectively, reached 12·5 mg and6·25 mg twice daily. Excluding deaths, carvedilol andplacebo were withdrawn permanently in 192 (20%) and174 (18%) of patients, respectively.

Fewer patients in the carvedilol group than in theplacebo group died (table 2, figure 2); however,there was no significant difference between the twogroups in the co-primary endpoint of death orcardiovascular hospital admission (table 2, figure 3),or in the secondary endpoints of sudden deathand admission to hospital because of heart failure(table 2). Fewer patients on carvedilol than on placebodied from cardiovascular causes, including heartfailure, or had a non-fatal myocardial infarction (table 2, figure 4).

DiscussionThe results of this study show substantial benefitfrom carvedilol with respect to major coronary events.The 23% relative reduction in mortality is identicalto that reported in a meta-analysis of 22 long-term,randomised, controlled trials of �-blockers in acutemyocardial infarction.3 However, in CAPRICORN,the all-cause mortality rate on placebo was 15%compared with 12% on carvedilol after an average follow-up of 1·3 years, whereas in the previous trials,the average mortality was 10% on placebo and 8%on �-blockers. Although these benefits cannot becompared exactly because of variations in lengthof follow-up in the trials included in the meta-analysis,the higher mortality on placebo in the CAPRICORNstudy emphasises that these patients were atparticularly high risk. The reduction in all-causemortality was additional to the effects of ACEinhibitors and reperfusion therapy, which wereprescribed in 98% and 46% of patients,respectively.

The reduction in all-cause mortality was unexpected,since concern about insufficient power to detect asignificant difference in all-cause mortality hadpersuaded the steering committee to change the primaryendpoint from all-cause mortality to all-cause mortalityor cardiovascular hospital admissions. Althoughnominally significant for the outcome of all-causemortality alone, the p value of 0·03 does not meet thehigher level of significance specified when the primaryendpoint was adopted. Nevertheless, death is the mostimportant outcome, it was the original primary endpoint,and, in practical terms, the observed 23% reduction inall-cause mortality represents a clinically importantoutcome.

Despite these benefits in terms of major coronaryevents, the new co-primary endpoint of all-causemortality or cardiovascular hospital admission was only8% lower for carvedilol than for placebo. The apparentinconsistency between the results for these two endpointswas not caused by an excess of cardiovascular hospitaladmissions in the carvedilol group. The numbers ofcardiovascular hospital admissions for any reason otherthan myocardial infarction and heart failure were aboutequal in the two treatment groups. However, many ofthese events preceded episodes of myocardial infarction,heart failure, and death, and hence masked the benefit on

ARTICLES


Carvedilol group Placebo group (n=975) (n=984)

DemographicsMean (range) age (years) 63 (29–88) 63 (25–90)Sex

Men 716 (73%) 724 (74%)Women 259 (27%) 260 (26%)

Smoking historyCurrent 326 (33%) 319 (32%)Previous 264 (27%) 243 (25%)Never 383 (39%) 418 (43%)

Medical historyPrevious MI 299 (31%) 290 (29%)Previous angina 559 (57%) 531 (54%)Previous hypertension 541 (55%) 514 (52%)Previous diabetes 207 (21%) 230 (23%)Other vascular disease 168 (17%) 159 (16%)Previous revascularisation 118 (12%) 107 (11%)Hyperlipidaemia 315 (32%) 322 (33%)

Infarct characteristicsMean (SD) LVEF (%) 32·9 (6·4) 32·7 (6·4)Mean (SD) SBP (mm Hg) 121·6 (17·3) 120·7 (16·1)Mean (SD) DBP (mm Hg) 73·7 (10·3) 73·4 (10·0)Mean (SD) heart rate (beats/min) 77·3 (11·4) 77·2 (11·3)Site of MI

Anterior 572 (59%) 536 (54%)Inferior 205 (21%) 205 (21%)Other 198 (20%) 243 (25%)

Treatment for index myocardial infarctionNitrates 715 (73%) 717 (73%)Intravenous �-blockers 112 (11%) 100 (10%)Intravenous heparin 617 (63%) 635 (65%)Subcutaneous heparin 460 (47%) 481 (49%)Intravenous diuretics 338 (35%) 320 (33%)Thrombolysis/primary angioplasty 442 (45%) 465 (47%)

Medications at time of randomisationACE inhibitor 953 (98%) 955 (97%)Aspirin 838 (86%) 847 (86%)

MI=myocardial infarction; LVEF=left-ventricular ejection fraction; SBP=systolic bloodpressure; DBP=diastolic blood pressure; ACE=angiotensin-converting enzyme.

Table 1: Baseline characteristics

Prop

ortio

n ev

ent-f

ree

Years

0 0·5 1·0 1·5 2·0 2·5

1·0

0·9

0·8

0·7

0·6

0·5

0·4

0·3

0·2

0·1

0

Carvedilol

Placebo

p=0·031

Numbers at riskCarvedilolPlacebo

975984

856861

648638

364358

117123

168

Figure 2: Kaplan-Meier estimates of all-cause mortality


these outcomes in a time to first event analysis. In fact,289 patients in the placebo group were admitted tohospital with cardiovascular problems compared with275 in the carvedilol group. Overall, 14% fewer patientswere admitted to hospital for heart failurein the carvedilol group than in the placebo group,and fewer died from heart failure.

We cannot directly compare CAPRICORN with earlier�-blocker trials with regard to endpoints, such as hospitaladmissions for heart failure, that were not assessed inthese trials. Nevertheless, in the Norwegian study oftimolol and the BHAT trial of propranolol—the twolargest landmark trials of �-blockers in myocardialinfarction—there were more reports of heart failure as anadverse event in the groups treated with �-blockers thanin the placebo groups. In both these trials, there was arapid escalation (over1–2 days) to the maximum dose ofstudy medication. By comparison, dose titration inCAPRICORN was more gradual.

In CAPRICORN, all-cause mortality or non-fatalrecurrent myocardial infarction—the most commonlyadopted primary endpoint in contemporaryclinical trials of acute coronary disease—was 29%lower with carvedilol than with placebo (p=0·002).The reduction in deaths and recurrent myocardialinfarction occurred during acute and chronic phases,as it has done in trials of �-blockers in acutemyocardial infarction and chronic heart failure. Theresults of our trial show that these improvements in

outcome apply equally to patients with objective evidenceof clinically significant left-ventricular dysfunction.

The acute phase of myocardial infarction is anintrinsically unstable period, especially in those with left-ventricular dysfunction. In our study, the meanleft-ventricular ejection fraction was substantiallyimpaired (32·8%), and treatment with intravenousdiuretics was required in a third of patients duringthe acute event. In the trials of �-blockers in heart failure,evidence of stability was required, not only in clinicalstatus but also in terms of drug therapy for heart failure.By contrast, in addition to being started on increasingdoses of carvedilol, almost all patients in CAPRICORNwere treated with an ACE inhibitor at the same time.

Recruitment to CAPRICORN was slow in somecountries where it was widely perceived that thecase for �-blockers in all patients with myocardialinfarction was proven and that the previous specificcontraindication of heart failure did not apply afterthe results of the trials of �-blockers in heart failure.The results of our study show that the approachto high-risk patients with left-ventricular dysfunctionafter a myocardial infarction should resemble thatof the studies of �-blockers in heart failure rather thanthe previous trials of �-blockers in myocardial infarction.

In light of these results, especially when taken togetherwith those of the previous trials of �-blockers for heart

ARTICLES


Carvedilol group (n=975) Placebo group (n=984) Hazard ratio (95% CI) p

Primary endpointsAll-cause mortality 116 (12%) 151 (15%) 0·77 (0·60–0·98) 0·031All-cause mortality or cardiovascular-cause 340 (35%) 367 (37%) 0·92 (0·80–1·07) 0·296hospital admission

Secondary endpointsSudden death 51 (5%) 69 (7%) 0·74 (0·51–1·06) 0.098Hospital admission for heart failure 118 (12%) 138 (14%) 0·86 (0·67–1·09) 0·215

Other endpointsCardiovascular-cause mortality 104 (11%) 139 (14%) 0·75 (0·58–0·96) 0·024Death due to heart failure 18 (2%) 30 (3%) 0·60 (0·33–1·07) 0·083Non-fatal myocardial infarction 34 (3%) 57 (6%) 0·59 (0·39–0·90) 0·014All-cause mortality or non-fatal myocardial 139 (14%) 192 (20%) 0·71 (0·57–0·89) 0·002infarction

Table 2: Primary, secondary, and other endpoints

Prop

ortio

n ev

ent-f

ree

Years

0 0·5 1·0 1·5 2·0 2·5

1·0

0·9

0·8

0·7

0·6

0·5

0·4

0·3

0·2

0·1

0

CarvedilolPlacebo


975984

693701

266246

481473

7578

94

p=0·30

Figure 3: Kaplan-Meier estimates of all-cause mortality orcardiovascular hospital admission

Prop

ortio

n ev

ent-f

ree

Years

0 0·5 1·0 1·5 2·0 2·5

1·0

0·9

0·8

0·7

0·6

0·5

0·4

0·3

0·2

0·1

0

Carvedilol

Placebo


975984

842834

634615

355338

108112

158

p=0·002

Figure 4: Kaplan-Meier estimates of all-cause mortality or non-fatal myocardial infarction


attack and heart failure, future studies of �-blockers inacute myocardial infarction should probably not be done.Although these results might be generalisable to other �-blockers, the CAPRICORN trial has provided the basisof a clinical template for future clinical practice in high-risk patients with left-ventricular dysfunction after amyocardial infarction. CAPRICORN reaffirms thatpatients with left-ventricular dysfunction after myocardialinfarction remain at high risk despite the benefitsafforded by modern care. In a meta-analysis of the threelarge trials of ACE inhibitors in patients with left-ventricular dysfunction or heart failure after an acutemyocardial infarction—SAVE, AIRE, and TRACE—theabsolute reduction in risk was 2·3%. The number ofpatients who need to be treated for 1 year with an ACEinhibitor to save one life is, therefore, 43.17 This findinghighlights the benefits of ACE inhibitors in patients withheart failure or left-ventricular dysfunction after an acutemyocardial infarction. The absolute reduction inmortality in CAPRICORN at 1 year was also 2·3%,resulting in an identical number needed to treat for 1 year. However, this benefit is additional to those ofACE inhibitors alone.

Our results indicate that treatment with carvedilolprovides a mechanism to further reduce the high rate ofmortality and other major coronary events in patientswith left-ventricular dysfunction after acute myocardialinfarction.

The CAPRICORN InvestigatorsWriting committee—Henry Dargie, Wilson Colucci, Ian Ford, Jose-Luis Lopez Sendon, Willem Remme, Norman Sharpe, Antje Blank, Terry L Holcslaw.Steering committee—Henry J Dargie (UK),Wilson S Colucci (USA), Jose Luis Lopez-Sendon (Spain), Willem Remme (Netherlands), Norman Sharpe (New Zealand), Ian Ford (UK; non-voting ).Endpoint committee—John McMurray (UK), Eugene Connolly (assistant to committee, UK), Lars Kober (Denmark), Jonathan Sackner-Bernstein (USA),Jordi Soler-Soler (Spain), Faiez Zannad (France).Data and safety monitoring board—Desmond Julian (UK), Barry Massie (USA), Simon Thompson (UK), Lars Wilhelmson(Sweden), Ian Ford (non-voting secretary, UK).Independent statisticians—Ian Ford, Jan Love, Michele Robertson, Adele Trainer (Robertson Centre, University of Glasgow, UK)Co-ordination—Antje Blank, Christoph Staiger, Deborah Hallett,Michaela Jahn (Roche Pharmaceuticals), Terry L Holcslaw, HelenWhitelegg, Ameet Nathwani (Glaxo SmithKline).Australia—J Amerena, T J Campbell, B Corliss, S Coverdale, D Cross, PJ Currie, S D’Arcy, M Dooris, L Fitzpatrick, M A Fitzpatrick, D Gallagher, P Garrahy, J Gibcus, A Haner, A Hopkins, J D Horowitz, B Jackson, C Juergens, J Kesby, S Leslie, D Leung, S MacKinley, E Newland, D Ninio, J G O’Driscoll, D Owensby, J Paull, L Roberts, M Rowe, G Rudge, A Russell, P Sage, A Smith-Bridges, W F Walsh. Belgium—P J C F Block, L De Wolf, A Depaepe, C Gassavelis, E Lecoq, M Massoz, P H F Materne, M Schoofs, M Van Camp.Canada—E V Auersperg, R B H Myers.France—J L Bonnet, C Cassat, Z Chati, C Darodes, S Dennetiere, L Fabrice, BA Fethi, JP Guillot, M Hubert, JY Ketelers, K Khalife, F Leroy, H Mann, P Pruvost, F Thoin, JL Weber, M Zupan.Germany—J B Bockenforde, B Darius, H Darius, G Dörr, F Freytag, P Hanrath, M Knap, H Neubauer, CA Nienaber, V Regitz-Zagrosek, JU Röhnisch, H Schmidt, CM Schönborn, A Thilo, O J Titlbach, F Weber.Hungary—L Barna, E Bod, A Borthaiser, G Buday, M Csanády, I Czuriga, I Edes, E Farkas, C Fejer, M Foos, P Goroek, T Habron, G Hagi, L Hegyi, A Ivanov, E Kalo, N Kanakaridisz, C Kiraly,G Kovacs, P Kovacs, V Kruzslicz, J Lippai, L Lorinczy, M Mezöfi, B Mezey, A Mohacsi, E Rabai, A Savkozy, P Solte’sz, E Somogyi, M Szabo, L Szeder, S Timar, K Toth, K Zamolyi, A Zsoldos.Ireland—J Barton.Israel—E G Abinader, L Arcavi, A B Ari, A Asman, A Benari, M Bergman,A Caspi, G Cotter, I Davidovich, A Dniella, N Friedberg, Z Gassan, S Harel, Y Hasin, A Keren, F Khoury, Y Kishon, T Jakrasovsky, S Lanciano, L Lerner, G Levine, A Loukitch, W Markewicz, A T Marmor, S Masarwa-Shibi, K Mikhailova, Y Milovanov, T Nakrasovsky, D Peres, N Pearlstein, E Pinchasov, M Popov, B Rabinowitz, L H Reisin, T Rivka,T Rosenfeld, A Salah, D Schepps, A Shotan, L Sigal, R Soloman,

Z Vered, N Vogel, D Wechsler, O Winrovsky, V Witzling, S Yanovsky, G Zaid, R Zimlichman. Italy–F Anglano, L Argenziano, G Binaghi, M G Cagliarelli, L Carvita, A Conti, S Damiani, M Fabrizio, C Giuseppe, R Mantovani, A Maresta, L Pirazzini, A Raviele, M Sanguinetti, M Sanguinetti, M L Sica, L Spinolo, B Trimarco.Lithuania–R Babarskiené, G Barauskiené, R Dzikaité, D Jarasuniene, R Kaualiauskiene, A Kirkutis, A Laucevicius, A Simaitis, A Statkeviciené,P Zabiela, C Zanella.Luxembourg–C Delagardelle, K Marianne.Netherlands–S C Baldew, J V Beek, G M Jochemsen, M Lageweg, M M Peters, E C M Schavemaker, H H Tan, P R M van Dykman, R A M van Langeveld.New Zealand–U Balzat, H Brannigan, C Cocks, S Coverdale, A Culpan, M Denton, R Doughty, T Grey, H Hart, J Hinge, L Howitt, H Ikram, C J S Low, M Lund, H F McAlister, M McLelland, S Muncaster,DS Scott, N Sharpe, I Straznicky, B van Gruting, D Penney, H Walsh,H D White, S Wright. Russia–O S Akimova, G P Arutyunov, Y B Belousov, T K Chernyavskaya,I V Demidova, M G Glezer, A G Gorbunov, I G Gordeev, E V Kokorin, I E Mikhailova, V S Moyseev, N B Perepech, I Persianov-Dubrov, S G Porokhnenko, A V Rozanov, E I Rubashkina, G E G Saveljeva, D Scott, D P Sementsov, V Y Sibirsky, K E Sobolev, I N Sokolova, O V Solowiew, S N Tereshtchenko, A A Upnitsky, A A Vershinin, N A Volov, A V Zorin.Spain–J M Aguirre Salcedo, C Almería, J Alonso, A Alonso Garcia, N Alvarenga, L Andraca, I Antorrena, F F Avilés, J E Barthe, A Bautista, J Bermejo, A Bethencourt, R Bilbao, X Bosch Genover, M T Castelló, A Castro Beiras, A Conde, M Crespo, E Cruz, J M Cruz-Fernández, E De Teresa Galvan, C Expósito, F Fernandez Aviles, F FernandezVazquez, R Fuentes, E Galve Basilio, E Galve, M C García, G Gómez, I Gómez, E González, J M Gonzalez, A Grau, M Jimenez, E A López de Sa, V López Garcia, J Lupón, A Mallol, E Méndez, J M Nicolás, J Palet, J Palomo, M J Paniagua, M G de la Peña, F Ridocci, E Rodriguez, R Rubio, L Ruiz, J Toquero, V Valle Tudela, J Velasco Rami,F Worner Diz, JL Zamorano Gomez.UK–A A J Adgey, J Akhter, J Anderson, M N A l Khafaji, J S Birkhead, S Campbell, B S P Chin, S M Cobbe, A J Cowley, H Dargie, M A de Belder, M A Devine, F G Dunn, D S Fluck, J Gardener, S Gent,R A Greenbaum, R Haynes, D Hogg, R S Hornung, S Humayun, J B Irving, M Jamieson, A Joy, M D Joy, W Keeble, M Lal, P S Lewis, F Lie, G Y H Lip, Y McKay, S Maynard, M J Metcalfe, M W Millar-Craig, E Morris, R J Northcote, H Nouriel, C P O'Regan, M P Pye, K Ranjadayalan, S G Ray, S Saltissi, L Shipley, J A Singer, R Singh, I B Squire, A D Timmis, C M Travill, M P S Varma, D Willis, A Wright, H Young.US–J Batty, L Brookfield, R Chernick, C Dahl, D Ferry, R Gilmore, M Givertz, S Goldman, S Hall, R Heuser, D Karalis, K Konzen, R Krishnan, R Lewin, F McGrew, B Molk, A Mooss, J Morrison, F Navetta, I Niazi, A Niederman, W O'Neill, R Patel, G Phillipedes.

References1 Norwegian Multicentre Group. Timolol induced reduction

in mortality and reinfarction in patients surviving acutemyocardial infarction. N Engl J Med 1981; 304: 801–07.

2 �-Blocker Heart Attack Trial Research Group. A randomised trial ofpropranolol in patients with acute myocardial infarction, I: mortalityresults. JAMA 1982; 247: 1707–14.

3 Yusuf S, Peto R, Lewis J, et al. Beta blockade during and aftermyocardial infarction: an overview of the randomized trials. ProgCardiovasc Dis 1985; 27: 335–71.

4 Packer M, Bristow M, Cohn J, et al. The effect of carvedilol onmorbidity and mortality in patients with chronic heart failure. N EnglJ Med 1996; 334: 1349–55.

5 CIBIS-II Investigators. The Cardiac Insufficiency Bisoprolol Study II(CIBIS-II): a randomised trial. Lancet 1999; 353: 9–13.

6 The MERIT-HF Study Group. Effect of metoprolol CR/XL inchronic heart failure: Metoprolol CR/XL Randomised InterventionTrial in Congestive Heart Failure (MERIT-HF). Lancet 1999; 353:2001–07.

7 The Australia-New Zealand Carvedilol Study Heart Failure ResearchCollaborative Group. Randomised,placebo-controlled trial of carvedilol in patients with congestive heartfailure due to ischaemic heart disease. Lancet 1997; 349: 375–80.

8 Spargias KS, Hall AS, Greenwood DC, et al. � Blocker treatmentand other prognostic variables in patients with clinical evidence ofheart failure after acute myocardial infarction: evidence from theAIRE study. Heart 1999; 81: 25–32.

9 Freemantle N, Cleland J, Young P, Mason J, Harrison J.Beta-blockade after myocardial infarction: systematic review andmeta regression analysis. BMJ 1999; 318: 1730–38.

10 Vantrimpont P, Rouleau JC, Wun CC, for the SAVEinvestigators. Additive beneficial effects of beta-blockers toangiotensin-converting enzyme inhibitors in the Survival and

ARTICLES



Ventricular Enlargement (SAVE) study. J Am Coll Cardiol 1997; 29:229–36.

11 Pfeffer MA, Braunwald E, Moye LA, et al, on behalf theSAVE investigators. Effect of captopril on mortalityand morbididty in patients with left ventricular dysfunctionafter myocardial infarction. Results of the Survival andVentricular Enlargement Trial. N Engl J Med 1992;327: 669–77.

12 The Acute Infarction Ramipril Efficacy (AIRE) Study Investigators.Effect of ramipril on mortality and morbidity of survivors of acutemyocardial infarction with clinical evidence of heart failure. Lancet1993; 342: 821–28.

13 Kober L, Torp-Pedersen C, Carlsen JE, et al. A clinical trial of theangiotensin-converting-enzyme inhibitor trandolapril in patients withleft ventricular dysfunction after myocardial infarction. N Engl J Med1995; 333: 1670–76.

14 Krumholz HM, Radford MS, Wang Y, et al. National use andeffectiveness of beta blockers for the treatment of elderly patientsafter acute myocardial infarction. JAMA 1998; 280: 623–29.

15 Gottlieb SS, McCarter RJ, Vogel RA. Effect of beta blockade onmortality among high risk and low risk patients after myocardialinfarction. N Engl J Med 1998; 339: 489–97.

16 Dargie HJ, on behalf of the CAPRICORN Steering Committee.Design and methodology of the CAPRICORN trial: a randomiseddouble blind placebo controlled study of the impact of carvedilol onmorbidity and mortality in patients with left ventricular dysfunctionafter myocardial infarction. Eur J Heart Failure 2000; 2: 325–32.

17 Flather MD, Yusuf S, Kober L, Pfeffer M, Hall A, Murray G.Long-term ACE-inhibitor therapy in patients with heartfailure or left-ventricular dysfunction: a systematic overviewof data from individual patients. Lancet 2000; 355: 1575–81.

ARTICLES


n engl j med

350;22

www.nejm.org may

27, 2004

The

new england journal

of

medicine

2247

original article

A Comparison of Albumin and Saline for Fluid Resuscitation in the Intensive Care Unit

The SAFE Study Investigators*

The Saline versus Albumin Fluid Evalua-tion (SAFE) Study is a collaboration of theAustralian and New Zealand Intensive CareSociety Clinical Trials Group, the AustralianRed Cross Blood Service, and the GeorgeInstitute for International Health. The writ-ing committee (Simon Finfer, M.B., B.S.,Rinaldo Bellomo, M.B., B.S., M.D., NeilBoyce, M.B., B.S., Ph.D., Julie French, R.N.,John Myburgh, M.B., B.Ch., Ph.D., and Ro-byn Norton, Ph.D., M.P.H.) takes respon-sibility for the content of this article. Addressreprint requests to Dr. Finfer at ANZICSCTG, Level 3, 10 Ievers St., Carlton, VIC3053, Australia, or at [email protected].

*The Saline versus Albumin Fluid Evalua-tion (SAFE) Study investigators are list-ed in the Appendix.

N Engl J Med 2004;350:2247-56.

Copyright © 2004 Massachusetts Medical Society.

background

It remains uncertain whether the choice of resuscitation fluid for patients in intensivecare units (ICUs) affects survival. We conducted a multicenter, randomized, double-blindtrial to compare the effect of fluid resuscitation with albumin or saline on mortality in aheterogeneous population of patients in the ICU.

methods

We randomly assigned patients who had been admitted to the ICU to receive either4 percent albumin or normal saline for intravascular-fluid resuscitation during the next28 days. The primary outcome measure was death from any cause during the 28-dayperiod after randomization.

results

Of the 6997 patients who underwent randomization, 3497 were assigned to receive al-bumin and 3500 to receive saline; the two groups had similar baseline characteristics.There were 726 deaths in the albumin group, as compared with 729 deaths in the salinegroup (relative risk of death, 0.99; 95 percent confidence interval, 0.91 to 1.09; P=0.87).The proportion of patients with new single-organ and multiple-organ failure was sim-ilar in the two groups (P=0.85). There were no significant differences between thegroups in the mean (±SD) numbers of days spent in the ICU (6.5±6.6 in the albumingroup and 6.2±6.2 in the saline group, P=0.44), days spent in the hospital (15.3±9.6and 15.6±9.6, respectively; P=0.30), days of mechanical ventilation (4.5±6.1 and4.3±5.7, respectively; P=0.74), or days of renal-replacement therapy (0.5±2.3 and0.4±2.0, respectively; P=0.41).

conclusions

In patients in the ICU, use of either 4 percent albumin or normal saline for fluid resus-citation results in similar outcomes at 28 days.

abstract

The New England Journal of Medicine Downloaded from nejm.org at HOSPITAL ITALIANO on February 27, 2020. For personal use only. No other uses without permission.

Copyright © 2004 Massachusetts Medical Society. All rights reserved.

n engl j med

350;22

www.nejm.org may

27

,

2004

The

new england journal

of

medicine

2248

he administration of intrave-

nous fluids to maintain or increase intra-vascular volume is a common intervention

in the intensive care unit (ICU), but there is uncer-tainty whether the choice of fluid significantly influ-ences patients’ outcomes.

1-7

In particular, no ade-quately powered randomized, controlled trials haveexamined the effect of fluid choice on the survivalof patients in the ICU. In the absence of such trials,a number of meta-analyses have examined how thechoice of crystalloid or colloid solution and of al-bumin-containing or albumin-free fluid affects sur-vival in critically ill patients and in patients who areless severely ill.

1-3,7

A meta-analysis published bythe Cochrane Injuries Group Albumin Reviewers in-cluded 24 studies involving a total of 1419 patientsand suggested that the administration of albumin-containing fluids resulted in a 6 percent increase inthe absolute risk of death when compared with theadministration of crystalloid solutions.

1

However,a subsequent meta-analysis of 55 trials involving atotal of 3504 patients examined the effect of resus-citation with albumin-containing fluid on the riskof death in a general population of patients and didnot find a significant increase in the risk of death.

3

The conflicting results of such meta-analyseshave left many clinicians unsure about the effect ofalbumin-containing fluids on survival in critically illpatients. To address this uncertainty, we conductedthe Saline versus Albumin Fluid Evaluation (SAFE)Study in 16 ICUs in Australia and New Zealand. Wetested the hypothesis that when 4 percent albuminis compared with 0.9 percent sodium chloride (nor-mal saline) for intravascular-fluid resuscitation inpatients in the ICU, there is no difference in the28-day rate of death from any cause.

study design and treatment protocol

Patients 18 years of age or older who had been ad-mitted to the closed, multidisciplinary ICUs of 16academic tertiary hospitals in Australia and NewZealand between November 2001 and June 2003were assessed for eligibility for the study. Eligiblepatients were those whom the treating clinicianjudged to require fluid administration to maintainor increase intravascular volume, with this decisionsupported by the fulfillment of at least one objectivecriterion. Patients admitted to the ICU after cardiacsurgery, after liver transplantation, or for the treat-ment of burns were excluded. Details of the inclu-

sion and exclusion criteria are given in Table S1 ofthe Supplementary Appendix (available with the fulltext of this article at www.nejm.org). A detailed de-scription of the study design has been publishedelsewhere.

8

The study protocol was approved by the ethicscommittees of the University of Sydney and of eachparticipating institution. Written informed consentwas to be obtained from all competent patients; incases in which prior consent could not be obtainedfrom the patient because of critical illness or the useof sedative or anesthetic drugs, consent could be de-layed, and a provision for delayed consent was ap-plied. In such cases, the patient or his or her surro-gate decision maker was informed of the study assoon as practicable, and consent was sought to con-tinue the study procedures and to access the partic-ipant’s medical records for study-related data. Thepatients or their legal surrogates were informed oftheir right to request that the study procedures bediscontinued and their right to refuse the study-related use of their medical records.

Eligible patients were randomly assigned to re-ceive either 4 percent albumin (Albumex, CSL) ornormal saline, with the random assignments strat-ified according to institution and according towhether there was a diagnosis of trauma on admis-sion to the ICU. Randomization was carried outcentrally with the use of a minimization algorithm,and the service was accessed on the Internet througha secure Web site. Study fluids were supplied inidentical 500-ml bottles, and blinding was ensuredthrough the use of specially designed maskingcartons and specially designed and manufacturedadministration sets.

8

The effectiveness of the blind-ing was confirmed in a formal study before the trialwas initiated. The treating clinicians determined theamount and rate of fluid administration accordingto each patient’s clinical status and response to treat-ment. The allocated study treatment was to be usedfor all fluid resuscitation in the ICU until death ordischarge or until 28 days after randomization. Theadministration of intravenous fluids outside the ICUwas not controlled.

In addition to the study fluid, patients receivedmaintenance fluids, specific replacement fluids, en-teral or parenteral nutrition, and blood products atthe discretion of the treating clinicians. The moni-toring of central venous pressure, pulmonary-arterycatheterization, and all other aspects of patient carewere performed at the discretion of the treating cli-nicians.

t

methods



n engl j med

350;22

www.nejm.org may

27, 2004

albumin vs. saline for fluid resuscitation in the icu

2249

baseline assessment and follow-up data collection

Data collected at baseline included the Acute Phys-iology and Chronic Health Evaluation II score,

9

aswell as information pertaining to diagnostic crite-ria for severe sepsis

10

(Table S2 of the Supplemen-tary Appendix) and for the acute respiratory distresssyndrome.

11

Patients were identified as having trau-matic brain injury at baseline if they had a history oftrauma, a Glasgow Coma Score

12

while not sedatedof less than 14, and an abnormality consistent withtraumatic brain injury on a computed tomographicscan of the head. The cardiovascular, respiratory,renal, hematologic, and hepatic components ofthe Sequential Organ-Failure Assessment (SOFA)score,

13

as described in Table S3 of the Supplemen-tary Appendix, were recorded at the time of random-ization, daily for the next seven days, and then everythird day until discharge from the ICU or until day28. After randomization, the heart rate, centralvenous pressure, mean arterial blood pressure, vol-ume of study fluid administered, volume of non-study fluid and blood products administered, netfluid balance (calculated as the total fluid inputminus the total fluid output), use of mechanicalventilation, and use of renal-replacement therapy(intermittent or continuous hemodialysis, hemo-filtration, or hemodiafiltration) were recorded dai-ly until discharge from the ICU or death or untilday 28.

outcome measures

The primary outcome measure was death from anycause within 28 days after randomization. Second-ary outcome measures were the survival time duringthe first 28 days, the proportion of patients who hadone, two, three, four, or five new organ failures (de-fined as a documented change in the cardiovascular,respiratory, renal, hematologic, or hepatic compo-nent of the SOFA score from 0, 1, or 2 at baseline to3 or 4 during the ICU stay, where higher scores in-dicate increasingly severe organ dysfunction), theduration of mechanical ventilation, the duration ofrenal-replacement therapy, and the duration of theICU and hospital stay.

Death from any cause within 28 days after ran-domization was also examined in six predefinedsubgroups according to the presence or absence oftrauma, the presence or absence of severe sepsis,and the presence or absence of the acute respiratorydistress syndrome at baseline.

study and data management

Two preplanned interim analyses were performedby an independent statistician after recruitment ofthe first 2333 patients (33 percent of the plannedtotal) and the first 4666 patients (67 percent), andthe results were reviewed by the independent data-monitoring committee.

The George Institute for International Health atthe University of Sydney performed the data man-agement, the site management, and the data analy-sis, independently of the funding agencies. Themanuscript was prepared by the writing committeeand was revised by the study investigators, who ap-proved the final manuscript.

statistical analysis

The trial was designed to enroll 7000 patients, there-by providing a power of 90 percent to detect a 3 per-cent difference in absolute mortality rates betweenthe two groups from an estimated baseline mortal-ity rate of 15 percent. The data were exported fromthe study database and analyzed with the use ofSPSS software (version 11.5). All analyses were per-formed on an intention-to-treat basis. Where datawere missing, we report the number of available ob-servations, and we make no assumptions about themissing data.

Proportions were compared by means of thechi-square test or Fisher’s exact test, and continu-ous variables were compared by means of unpairedt-tests. The results of comparisons of event rates inthe two groups are presented as relative risks with95 percent confidence intervals. Survival times werecompared by means of the log-rank test and are pre-sented as Kaplan–Meier curves without adjustmentfor baseline covariates. Heterogeneity of treatmenteffects among subgroups was assessed with the useof the test for a common relative risk.

14

study patients

Seven thousand random assignments to a studytreatment were made (3499 to the albumin groupand 3501 to the saline group). Three patients mis-takenly underwent randomization twice within 28days; they were followed for 28 days beginning atthe time of the first randomization, and for purpos-es of data analysis were considered part of the groupto which they were first assigned. Thus, the studypopulation comprised 6997 patients, 3497 of whom

results



n engl j med

350;22

www.nejm.org may

27

,

2004

The

new england journal

of

medicine

2250

were assigned to receive albumin and 3500 of whomwere assigned to receive saline. The majority of thepatients (6628 [94.7 percent] — 3312 of those inthe albumin group [94.7 percent] and 3316 of thosein the saline group [94.7 percent]) were enrolledwith the use of the provision for delayed consent.Delayed consent was obtained from the patient in2713 cases (38.8 percent) and from a surrogate de-cision maker (a relative, a legally recognized surro-gate, or an institutional ethics committee) in the re-maining cases. Prior consent was obtained from thepatient in 45 cases (0.6 percent) and from a surro-gate decision maker in 335 cases (4.8 percent).

At baseline, the only statistically significant dif-ference between the two groups was a higher mean(±SD) central venous pressure in the albumin group(9.0±4.7 mm Hg, vs. 8.6±4.6 mm Hg in the salinegroup; P=0.03). The baseline characteristics of the6997 patients are summarized in Table 1, and theirprogress through the study is summarized in Fig-ure S1 of the Supplementary Appendix.

Study fluid was administered to all but 197 pa-tients (2.8 percent), including 90 in the albumingroup and 107 in the saline group. Resuscitationfluids in addition to the allocated study fluid wereadministered to 309 patients in the albumin group(8.8 percent) and 375 in the saline group (10.7 per-cent). The most common reason for the adminis-tration of nonstudy resuscitation fluid was error (in189 patients in the albumin group [5.4 percent] and190 in the saline group [5.4 percent]). Clinicians’preference for a specific nonstudy resuscitation flu-id was the reason for its administration in 68 pa-tients in the albumin group (1.9 percent) and 103in the saline group (2.9 percent). At the completionof the trial, information on vital status 28 days afterrandomization was unavailable for 67 patients (1.0percent), including 26 in the albumin group and 41in the saline group. In 56 of these 67 cases, vital sta-tus was missing because the patient or his or her le-gal surrogate had withheld or withdrawn consent.

fluids administered and treatment effects

On each of the first three study days, the patientswho had been randomly assigned to receive albuminreceived significantly less study fluid than did thoseassigned to saline, resulting in a significantly great-er net positive fluid balance in the saline group oneach of those days (Table 2). The ratios of the volumeof albumin to the volume of saline administeredduring the first four days were as follows: 1:1.3 onday 1, 1:1.6 on day 2, 1:1.3 on day 3, and 1:1.2 on

* Plus–minus values are means ±SD. Percentages were calculated according to the number of patients for whom data were available: for sex, 3497 in the albu-min group and 3500 in the saline group; for severe sepsis, 3339 in the albumin group and 3338 in the saline group; and for all the other variables, 3428 in the albumin group and 3423 in the saline group. Because of rounding, not all per-centages total 100. ICU denotes intensive care unit, and APACHE II Acute Physiology and Chronic Health Evaluation II.

† Higher scores on APACHE II indicate more severe illness.‡P=0.03 for the comparison with the value in the albumin group (without cor-

rection for multiple-hypothesis testing).§ Organ failure was defined as a Sequential Organ-Failure Assessment score

13

of 3 or 4 for any individual organ system.

Table 1. Baseline Characteristics of the Patients.*

Characteristic Albumin Group Saline Group

Age — yr 58.6±19.1 58.5±18.7

Female sex — no. (%) 1424 (40.7) 1376 (39.3)

Reason for admission to ICU — no. (%)SurgicalMedical

1473 (43.0)1955 (57.0)

1465 (42.8)1958 (57.2)

Source of admission to ICU — no. (%)

Emergency department 948 (27.7) 977 (28.5)

Hospital floor 614 (17.9) 573 (16.7)

Another ICU 63 (1.8) 66 (1.9)

Another hospital 323 (9.4) 341 (10.0)

Operating room (emergency surgery) 801 (23.4) 780 (22.8)

Operating room (elective surgery) 662 (19.3) 678 (19.8)

Same ICU (readmission) 17 (0.5) 8 (0.2)

Predefined subgroups — no. (%)TraumaSevere sepsisAcute respiratory distress syndrome

597 (17.4)603 (18.1)

61 (1.8)

590 (17.2)615 (18.4)66 (1.9)

APACHE II score† 18.7±7.9 19.0±8.0

Physiological variables

Heart rate — beats/min 91.4±23.5 92.3±23.5

Mean arterial pressure — mm Hg 77.8±16.4 78.2±16.3

Central venous pressure — mm Hg 9.0±4.7 8.6±4.6‡

Urine output — ml/hr 89.7±132.4 95.0±161.4

Serum albumin — g/liter 27.4±7.8 27.7±7.9

Organ failure— no. (%)§

No failure 1962 (57.2) 1885 (55.1)

1 organ 1075 (31.4) 1148 (33.5)

2 organs 335 (9.8) 329 (9.6)

3 organs 50 (1.5) 57 (1.7)

4 organs 5 (0.1) 4 (0.1)

5 organs 1 (<0.1) 0

Mechanical ventilation — no. (%) 2186 (63.8) 2217 (64.8)

Renal-replacement therapy — no. (%) 45 (1.3) 41 (1.2)

Albumin in previous 72 hr — no. (%) 127 (3.7) 135 (3.9)



n engl j med

350;22

www.nejm.org may

27, 2004


2251

day 4. The overall ratio of the volume of albumin tothe volume of saline administered during the firstfour days was approximately 1:1.4. Patients in thetwo groups received similar volumes of other fluidsduring the first four days, except on days 1 and 2,when the patients in the albumin group received agreater volume of packed red cells than did those inthe saline group; on average, during the first fourdays, patients assigned to receive albumin received71.0 ml more packed red cells than those assigned

to receive saline. On day 2, patients in the salinegroup received a greater volume of nonstudy fluidsthan did those in the albumin group (Table 2). Afterday 4, there were no differences between the twogroups in the volume of study fluids administered.There were no significant differences between thegroups in the mean arterial pressure measured atthe end of each of the first four days of the study. Thepatients assigned to receive albumin had a lowerheart rate at the end of the first day than those as-

* Plus–minus values are means ±SD.

† P values are for the comparison between the two means for each variable at each time point.

Table 2. Fluids Administered and Physiological Effects of Treatment.*

Variable Albumin Group Saline Group P Value†

No. of Patients Value No. of Patients Value

Study fluid (ml)Day 1Day 2Day 3Day 4

3410305922101686

1183.9±973.6602.7±892.7268.0±554.5192.3±427.0

3418306822021664

1565.3±1536.1954.0±1484.4348.3±753.5228.6±642.6

<0.001<0.001

0.030.57

Nonstudy fluid (ml)Day 1Day 2Day 3Day 4

3392305121991680

1459.4±1183.22615.9±1372.52618.5±1346.52691.5±1228.7

3405305721911656

1505.6±1254.32707.3±1435.72660.9±1319.32707.7±1255.4

0.300.0090.150.36

Packed red cells (ml)Day 1Day 2Day 3Day 4

3411306622171692

97.8±360.7106.5±321.459.8±225.543.6±167.5

3415307422101668

71.7±296.861.1±235.249.5±190.846.0±189.0

<0.001<0.001

0.300.77

Net positive fluid balance (ml)Day 1Day 2Day 3Day 4

3363304421901671

1543.6±1619.71015.3±1826.9422.1±1633.3137.2±1491.0

3382305221821649

1990.5±2061.71505.1±2215.9553.0±1732.3155.7±1650.6

<0.001<0.001

0.0070.70

Mean arterial pressure (mm Hg)Day 1Day 2Day 3Day 4

3406306822151688

81.4±14.484.4±15.187.2±15.388.3±15.9

3408307522091666

80.9±14.584.2±15.786.9±16.188.4±16.3

0.140.490.620.87

Heart rate (beats/min)Day 1Day 3Day 3Day 4

3398307122161691

88.0±20.288.5±19.588.8±19.189.5±18.9

3406307522131668

89.7±20.889.5±19.289.7±18.889.9±18.5

<0.0010.060.100.52

Central venous pressure (mm Hg)Day 1Day 2Day 3Day 4

2204209515311221

11.2±4.811.6±4.911.4±4.811.1±4.8

2270213515891230

10.0±4.510.4±4.310.7±4.410.5±4.4

<0.001<0.001<0.001<0.001

Serum albumin (g/liter)Day 1Day 2Day 3Day 4

2081270819211498

28.7±7.030.8±6.430.0±6.429.0±6.2

2061270319051478

24.7±6.524.5±5.923.6±5.623.1±5.5

<0.001<0.001<0.001<0.001



n engl j med

350;22

www.nejm.org may

27

,

2004

The

new england journal

of

medicine

2252

signed to receive saline. Central venous pressurewas significantly higher in the albumin group thanin the saline group at all time points during the firstfour days, and the serum albumin concentration washigher in the albumin group throughout the studyperiod (Table 2).

outcomes

Within 28 days after randomization, 726 of 3473 pa-tients in the albumin group (20.9 percent) and 729of 3460 patients in the saline group (21.1 percent)had died. For the albumin group as compared withthe saline group, the absolute difference in mortal-ity was –0.2 percent (95 percent confidence inter-val, –2.1 to +1.8 percent). The relative risk of deathamong patients assigned to receive albumin as

compared with those assigned to receive saline was0.99 (95 percent confidence interval, 0.91 to 1.09;P=0.87). At 28 days, 111 patients in the albumingroup (3.2 percent) and 87 patients in the salinegroup (2.5 percent) remained in the ICU (relativerisk, 1.27; P=0.09); 793 (22.8 percent) and 848(24.5 percent), respectively, remained in the hospi-tal (relative risk, 0.93; 95 percent confidence inter-val, 0.86 to 1.01; P=0.10) (Table 3). There was nosignificant difference in survival times between thetwo groups (Fig. 1).

The number of patients who had new single-organ or multiple-organ failure, assessed accordingto their SOFA scores, was similar in the two groups(P=0.85 by Fisher’s exact test) (Table 3). During the28-day study period the mean length of stay in the

* Plus–minus values are means ±SD. CI denotes confidence interval, and ICU intensive care unit.† The data include the numbers of patients in the ICU or the length of stay in the ICU.‡ Data were available for 2649 patients in the albumin group and 2673 patients in the saline group. New organ failure was defined as a Sequential

Organ-Failure Assessment score

13

of 0, 1, or 2 in any individual organ system at baseline, followed by an increase in the score to 3 or 4 in the same system.

§ The P value pertains to the comparison between the albumin and saline groups in the numbers of patients who had no new organ failure or

new failure of one, two, three, four, or five organs.

Table 3. Primary and Secondary Outcomes.*

Outcome Albumin Group Saline GroupRelative Risk

(95% CI)Absolute Difference

(95% CI) P Value

Status at 28 days — no./total no. (%)

Dead 726/3473 (20.9) 729/3460 (21.1) 0.99 (0.91 to 1.09) 0.87

Alive in ICU 111/3473 (3.2) 87/3460 (2.5) 1.27 (0.96 to 1.68) 0.09

Alive in hospital† 793/3473 (22.8) 848/3460 (24.5) 0.93 (0.86 to 1.01) 0.10

Length of stay in ICU — days 6.5±6.6 6.2±6.2 0.24 (–0.06 to 0.54) 0.44

Length of stay in hospital — days† 15.3±9.6 15.6±9.6 –0.24 (–0.70 to 0.21) 0.30

Duration of mechanical ventilation — days

4.5±6.1 4.3±5.7 0.19 (–0.08 to 0.47) 0.74

Duration of renal-replacement therapy — days

0.48±2.28 0.39±2.0 0.09 (–0.0 to 0.19) 0.41

New organ failure — no. (%)‡ 0.85§

No failure 1397 (52.7) 1424 (53.3)

1 organ 795 (30.0) 796 (29.8)

2 organs 369 (13.9) 361 (13.5)

3 organs 68 (2.6) 75 (2.8)

4 organs 18 (0.7) 17 (0.6)

5 organs 2 (0.1) 0

Death within 28 days according to sub-group — no./total no. (%)

Patients with trauma 81/596 (13.6) 59/590 (10.0) 1.36 (0.99 to 1.86) 0.06

Patients with severe sepsis 185/603 (30.7) 217/615 (35.3) 0.87 (0.74 to 1.02) 0.09

Patients with acute respiratory dis-tress syndrome

24/61 (39.3) 28/66 (42.4) 0.93 (0.61 to 1.41) 0.72



n engl j med

350;22

www.nejm.org may

27, 2004


2253

ICU was 6.5±6.6 days in the albumin group and6.2±6.2 days in the saline group (P=0.44). The meanlength of stay in the hospital was 15.3±9.6 days and15.6±9.6 days, respectively (P=0.30). The numbersof days of mechanical ventilation and days of renal-replacement therapy were similar in the two groups(Table 3).

subgroup analyses

During the 28-day study period, the relative risk ofdeath among patients with trauma in the albumingroup as compared with such patients in the salinegroup was 1.36; the corresponding relative risk ofdeath among patients without trauma was 0.96(P=0.04 by the test for a common relative risk). Thisdifference in the relative risk of death was due to thegreater number of patients with trauma and an as-sociated brain injury who died after random assign-ment to albumin as opposed to saline: 59 of 241such patients in the albumin group died (24.5 per-cent), as compared with 38 of 251 such patients inthe saline group (15.1 percent) (relative risk, 1.62;95 percent confidence interval, 1.12 to 2.34; P=0.009). Among patients who had trauma withoutbrain injury, there was no difference between thegroups in terms of mortality: 22 such patients inthe albumin group (6.2 percent) and 21 in the sa-line group (6.2 percent) died (relative risk, 1.00; 95percent confidence interval, 0.56 to 1.79; P=1.00).Among all the patients who had trauma (596 in thealbumin group and 590 in the saline group), therewere 81 (13.6 percent) deaths in the albumin groupand 59 (10.0 percent) in the saline group (relativerisk, 1.36; 95 percent confidence interval, 0.99 to1.86; P=0.06) (Fig. 2 and Table 3).

In a subgroup analysis of patients with severesepsis, the relative risk of death during the 28-daystudy period among those randomly assigned to re-ceive albumin as opposed to saline was 0.87, ascompared with a corresponding relative risk of 1.05among patients without severe sepsis (P=0.06 bythe test for a common relative risk). Of the 603 pa-tients with severe sepsis who had been assigned toreceive albumin, 185 (30.7 percent) died, and of the615 patients with severe sepsis who had been as-signed to receive saline, 217 (35.3 percent) died (rel-ative risk, 0.87; 95 percent confidence interval, 0.74to 1.02; P=0.09) (Table 3). In a subgroup analysis ofpatients with the acute respiratory distress syn-drome, the relative risk of death among those as-signed to receive albumin as opposed to saline was0.93; the corresponding relative risk among patients

without this syndrome was 1.00 (P=0.74 by the testfor a common relative risk).

In this randomized trial, we found that the use of4 percent albumin or normal saline for intravascu-lar volume resuscitation in a heterogeneous popu-lation of patients in the ICU resulted in equivalentrates of death from any cause during the 28-daystudy period. Requirements for mechanical venti-lation and renal-replacement therapy, time spent inthe ICU and in the hospital during the 28-day studyperiod, and the time until death (among the patientswho died) were also equivalent. The proportion ofpatients in the two groups in whom new single-organ or multiple-organ failure developed were sim-ilar. Our findings do not support the results of theCochrane Injuries Group Albumin Reviewers’ meta-analysis, which suggested that the use of albuminwas associated with an increased mortality rateamong critically ill patients.

1

Our study was conducted as a double-blind, ran-domized trial. Albumin and saline are not consid-ered equipotent intravascular volume expanders, buttheir relative potencies have not previously beenexamined in an adequately powered, blinded trial.In our study, patients who were resuscitated with al-bumin received less fluid than those who were re-suscitated with saline. During the first four days,

discussion

Figure 1. Kaplan–Meier Estimates of the Probability of Survival.

P=0.96 for the comparison between patients assigned to receive albumin and those assigned to receive saline.

Prob

abili

ty o

f Sur

viva

l

0.8

0.9

0.7

0.6

0.00 84 12 16 20 24 28

Days

AlbuminSaline

1.0



n engl j med

350;22

www.nejm.org may

27

,

2004

The

new england journal

of

medicine

2254

the ratio of albumin administered to saline admin-istered was approximately 1:1.4. However, therewas no significant difference in mean arterial pres-sure between the groups, and the differences incentral venous pressure and heart rate were small.Thus, we believe that the patients in the two groupswere resuscitated to similar and acceptable endpoints.

Our study was also a large, pragmatic study in apopulation of patients subject to a large number ofconcurrent interventions. We did not collect infor-mation on all concurrent interventions performed inthe ICU. However, randomization was stratified ac-cording to participating institution, so that each in-stitution treated equal numbers of patients assignedto saline or to albumin. As a result, we do not believethat an imbalance in concurrent interventions couldhave influenced the results.

Patients who were assigned to albumin receiveda significantly greater volume of packed red cellsduring the first two days of the study. The reasonsfor this difference remain speculative but may in-clude greater hemodilution with albumin than withsaline or increased blood loss with albumin due totransient alterations in coagulation. In a study oftransfusion requirements in critically ill patientsconducted by Hébert and colleagues, a liberal trans-fusion policy resulted in the administration of an

excess of 3.0 units of packed red cells. This was as-sociated with an increase in in-hospital mortality of5.9 percentage points.

15

During the first four daysof our study, the excess volume of fluid transfusedin the albumin group averaged 71.0 ml per patient(less than one quarter of a unit). Accordingly, we donot believe this small excess in transfused volumeinfluenced the results.

Given that our study had insufficient power todetect small but important differences in mortalityamong the predefined subgroups, the results pro-vide only limited evidence that the treatment effectsvaried among these subgroups. The finding that pa-tients with trauma might benefit more from resus-citation with saline than patients without traumaappears to be consistent with the results of a meta-analysis by Choi et al., who suggested that colloidresuscitation was associated with increased mortal-ity in patients with trauma.

2

In our study, however,the increased relative risk of death among patientswith trauma as compared with those without trau-ma resulted from a small excess number of deathsamong patients who had trauma with brain injury,whereas the meta-analysis by Choi et al. did not in-clude studies in patients with brain injury.

2

In our study, the difference in mortality between

the albumin and saline groups among patients withtrauma involving brain injury should be interpreted

Figure 2. Relative Risk of Death from Any Cause among All the Patients and among the Patients in the Six Predefined Subgroups.

The size of each symbol indicates the relative number of events in the given group. The horizontal bars represent the confidence intervals (CI). ARDS denotes the acute respiratory distress syndrome.

1.0 2.0

SalineBetter

AlbuminBetter

Severe sepsis

No

Overall

Trauma

Yes

Yes

No

ARDS

Yes

No

185/603

726/3473

81/596

641/2831

518/2734

24/61

697/3365

217/615

729/3460 0.99 (0.91–1.09)

1.36 (0.99–1.86)

0.96 (0.88–1.06)

0.87 (0.74–1.02)

1.05 (0.94–1.17)

0.93 (0.61–1.41)

0.5

1.00 (0.91–1.09)

59/590

666/2830

492/2720

28/66

697/3354

Relative Risk (95% CI)PatientsAlbuminGroup

SalineGroup

no. of deaths/total no.



n engl j med

350;22

www.nejm.org may

27, 2004


2255

with caution. Patients with traumatic brain injuryconstituted only 7 percent of the study population,and the excess number of deaths in the albumingroup was only 21. In large studies, such subgroupdifferences frequently occur by chance.

16

In addi-tion, the rate of death from any cause over a 28-dayperiod is not considered the most appropriate out-come measurement with which to assess treatmenteffects in patients with brain injury. Assessment ofmortality and functional neurologic status at leastsix months after injury is recommended.

17

In con-trast with our findings in patients with trauma, thecomparison of the relative risk of death among pa-tients with severe sepsis and those without severesepsis provides limited evidence of a treatment effectthat favors albumin in patients with severe sepsis.It should be noted that such differences betweensubgroups frequently occur by chance and that onlyspecifically designed and appropriately poweredstudies can determine whether any such treatmenteffects are real.

In conclusion, our study provides evidence thatalbumin and saline should be considered clinicallyequivalent treatments for intravascular volume re-

suscitation in a heterogeneous population of pa-tients in the ICU. Whether either albumin or salineconfers benefit in more highly selected populationsof critically ill patients requires further study. Ac-cording to the current state of knowledge, factorsthat may influence the choice of resuscitation fluidfor a critically ill patient include the individual clini-cian’s preference, the tolerability of the treatment,its safety, and its cost.

Supported by the Auckland District Health Board, MiddlemoreHospital, and the Health Research Council of New Zealand — all inNew Zealand; and by the Australian Commonwealth Department ofHealth and Aged Care, CSL (Melbourne, Victoria), the AustralianNational Health and Medical Research Council, the Health Depart-ment of Western Australia, the New South Wales Health Depart-ment, the Northern Territory Health Services, the QueenslandHealth Services Department, Royal Hobart Hospital (Tasmania), theSouth Australian Department of Human Services, and the VictorianDepartment of Human Services — all in Australia.

Dr. A. Davies and Dr. D. Stephens report owning shares in CSL.We are indebted to the nursing and medical staff of the intensive

care units and the blood banks of the participating institutions; tothe staff of the Australian Red Cross Blood Service and the NewZealand Blood Service, whose enthusiasm and hard work made theSAFE Study possible; to the staff of Tuta Healthcare, who manufac-tured and supplied the fluid-administration sets; and to the manag-ers and staff of CSL, who manufactured, blinded, packed, and sup-plied the saline and albumin used in the study.

appendix

The Saline versus Albumin Fluid Evaluation (SAFE) Study investigators are as follows:

Writing Committee

— S. Finfer (Chair), R. Bellomo, N.Boyce, J. French, J. Myburgh, and R. Norton.

Management Committee

— R. Norton (Chair), J. French (Senior Project Manager), R. Bellomo, S.Finfer, J. Myburgh, G. Doig, M. Hayek, and S. O’Donnell.

Steering Committee

— S. Finfer (Chair), A. Bell, R. Bellomo, N. Boyce, D. Blythe, J.Cade, M. Chapman, L. Cole, J. Cooper, A. Davies, C. French, J. French, C. Joyce, C. McArthur, S. MacMahon, J. Myburgh, B. Neal, R. Norton,J. Presneill, P. Saul, I. Seppelt, D. Stephens, A. Turner, A. Williams, and C. Woolfe.

External Safety and Data Monitoring Committee

— R. Peto(Chair), P. Sandercock, C. Sprung, and D. Young.

Statistical Analysis

(George Institute for International Health, University of Sydney, Sydney,N.S.W., Australia) — S.K. Lo, S. Sivarajasingham, L. Francis, M. Woodward.

Site investigators

(all in Australia unless otherwise specified) —Alfred Hospital, Melbourne: J. Charlton, J. Cooper, A. Davies, C. Harry, L. Higgins, K. Moulden, and S. Vallance. Auckland Hospital, Auck-land, New Zealand: J. Chadderton, L. Newby, and C. McArthur. Austin and Repatriation Medical Centre, Melbourne: S. Bates, R. Bellomo,D. Goldsmith, and A. Voss. Australian Red Cross Blood Service, Melbourne: N. Boyce. Fremantle Hospital, Fremantle: D. Blythe and A. Pal-ermo. George Institute for International Health, University of Sydney, Sydney: L. Francis, J. French, M. Hayek, K. Jayne, S. MacMahon, M.Merai, B. Neal, R. Norton, S. Pandey, S. O’Donnell, M. Schmidt, S. Sivarajasingham, and M. Woodward. John Hunter Hospital, Newcastle:R. Carroll, B. McFadyen, and P. Saul. Middlemore Hospital, Auckland, New Zealand: J. Clarke, J. Powell, A. Williams, and J. Tai. NepeanHospital, Penrith: L. Cole, I. Hynesova, I. Seppelt, and L. Weisbrodt. Princess Alexandra Hospital, Brisbane, Queensland: L. Bradley, C.Joyce, T. Kelly, A. Limpus, and R. Moore. Royal Adelaide Hospital, Adelaide: M. Chapman, S. Creed, S. Kaplan, J. Rivett. Royal Darwin Hos-pital, Darwin: D. Stephens and J. Thomas. Royal Hobart Hospital, Hobart: A. Bell, K. Marsden, and A. Turner. Royal Melbourne Hospital,Melbourne: C. Boyce, J. Cade, B. Howe, J. Presneill, and M. Robertson. Royal North Shore Hospital, Sydney: G. Doig, S. Finfer, A. O’Con-nor, J. Potter, and N. Ramakrishnan. Royal Prince Alfred Hospital, Sydney: C. Powell, D. Rajbhandari, and C. Woolfe. St. George Hospital,Sydney: K. Girling, M. Hodgetts, A. Jovanovska, and J. Myburgh. Western Hospital, Melbourne: C. French and L. Little.

references

1.

Cochrane Injuries Group Albumin Re-viewers. Human albumin administrationin critically ill patients: systematic review ofrandomised controlled trials. BMJ 1998;317:235-40.

2.

Choi PT, Yip G, Quinonez LG, Cook DJ.Crystalloids vs. colloids in fluid resuscita-tion: a systematic review. Crit Care Med1999;27:200-10.

3.

Wilkes MM, Navickis RJ. Patient sur-

vival after human albumin administration:a meta-analysis of randomized, controlledtrials. Ann Intern Med 2001;135:149-64.

4.

Cook D, Guyatt G. Colloid use for fluidresuscitation: evidence and spin. Ann InternMed 2001;135:205-8.

5.

Whether albumin therapy improves orworsens survival of critically ill patients isnot known. Ann Intern Med 2001;135:S-25.

6.

Boldt J. New light on intravascular vol-

ume replacement regimens: what did welearn from the past three years? AnesthAnalg 2003;97:1595-604.

7.

Bunn F, Alderson P, Hawkins V. Colloidsolutions for fluid resuscitation. CochraneDatabase Syst Rev 2003;1:CD001319.

8.

ANZICS Clinical Trials Group, Insti-tute for International Health SAFE StudyInvestigators. The Saline vs. Albumin FluidEvaluation (SAFE) Study (ISRCTN76588266):



n engl j med

350;22

www.nejm.org may

27

,

2004

2256


design and conduct of a multi-centre, blind-ed randomised controlled trial of intravenousfluid resuscitation in critically ill patients.London: BMJ Publishing, 2004. (AccessedMay 3, 2004, at http://bmj.com/cgi/content/full/326/7389/559/DC1.)

9.

Knaus WA, Draper EA, Wagner DP, Zim-merman JE. APACHE II: a severity of diseaseclassification system. Crit Care Med 1985;13:818-29.

10.

Bone RC, Balk RA, Cerra FB, et al. Defini-tions for sepsis and organ failure and guide-lines for the use of innovative therapies insepsis. Chest 1992;101:1644-55.

11.

Bernard GR, Artigas A, Brigham KL, et al.

The American-European Consensus Confer-ence on ARDS: definitions, mechanisms, rel-evant outcomes, and clinical trial coordina-tion. Am J Respir Crit Care Med 1994;149:818-24.

12.

Teasdale G, Jennett B. Assessment ofcoma and impaired consciousness: a practi-cal scale. Lancet 1974;2:81-4.

13.

Vincent JL, de Mendonca A, Cantraine F,et al. Use of the SOFA score to assess the in-cidence of organ dysfunction/failure in in-tensive care units: results of a multicenter,prospective study. Crit Care Med 1998;26:1793-800.

14.

Woodward M. Epidemiology: study de-

sign and data analysis. Boca Raton, Fla.:Chapman & Hall/CRC Press, 1999.

15.

Hébert PC, Wells G, Blajchman MA, etal. A multicenter, randomized, controlledclinical trial of transfusion requirements incritical care. N Engl J Med 1999;340:409-17.[Erratum, 1999;340:1056.]

16.

Collins R, MacMahon S. Reliable as-sessment of the effects of treatment on mor-tality and major morbidity. I. Clinical trials.Lancet 2001;357:373-80.

17.

Narayan RK, Michel ME, Ansell B, et al.Clinical trials in head injury. J Neurotrauma2002;19:503-57.


posting presentations at medical meetings on the internet

Posting an audio recording of an oral presentation at a medical meeting on the Internet, with selected slides from the presentation, will not be considered prior publication. This will allow students and physicians who are unable to attend the meeting to hear the presentation and view the slides. If there are any questions about this policy, authors should feel free to call the

Journal’

s Editorial Offices.



JAMA-EXPRESSORIGINAL CONTRIBUTION

Major Outcomes in High-RiskHypertensive Patients Randomized toAngiotensin-Converting Enzyme Inhibitoror Calcium Channel Blocker vs DiureticThe Antihypertensive and Lipid-Lowering Treatmentto Prevent Heart Attack Trial (ALLHAT)The ALLHAT Officers andCoordinators for the ALLHATCollaborative Research Group

TREATMENT AND COMPLICA-tions among the 50 to 60 mil-lion people in the United Stateswith hypertension are esti-

mated to cost $37 billion annually, withantihypertensive drug costs alone ac-counting for an estimated $15.5 bil-lion per year.1 Antihypertensive drugtherapy substantially reduces the riskof hypertension-related morbidity andmortality.2-6 However, the optimalchoice for initial pharmacotherapy ofhypertension is uncertain.7

Earlier clinical trials documented thebenefit of lowering blood pressure (BP)using primarily thiazide diuretics or�-blockers.2,3,8 After these studies, sev-eral newer classes of antihypertensiveagents (ie, angiotensin-converting en-zyme [ACE] inhibitors, calcium chan-nel blockers [CCBs], �-adrenergicblockers, and more recently angioten-sin-receptor blockers) became avail-able. Over the past decade, major pla-cebo-controlled trials have documentedthat ACE inhibitors and CCBs reducecardiovascular events in individualswith hypertension.9-11 However, theirrelative value compared with older, lessexpensive agents remains unclear.There has been considerable uncer-tainty regarding effects of some classesof antihypertensive drugs on risk of

Author Affiliations: ALLHAT Authors, Their Finan-cial Disclosures, and Group Members are listed at theend of this article.Corresponding Authors and Reprints: Jackson T.Wright, Jr, MD, PhD, Case Western Reserve Univer-sity, General Clinical Research Center, Suite 7311,

Horvitz Tower, 11000 Euclid Ave, Cleveland, OH44106-5041 (e-mail: [email protected]); Barry R.Davis, MD, PhD, University of Texas-Houston HealthScience Center, School of Public Health, 1200 Her-man Pressler St, Suite E801, Houston, TX 77030(e-mail: [email protected]).

Context Antihypertensive therapy is well established to reduce hypertension-related morbidity and mortality, but the optimal first-step therapy is unknown.

Objective To determine whether treatment with a calcium channel blocker or anangiotensin-converting enzyme inhibitor lowers the incidence of coronary heart dis-ease (CHD) or other cardiovascular disease (CVD) events vs treatment with a diuretic.

Design The Antihypertensive and Lipid-Lowering Treatment to Prevent Heart At-tack Trial (ALLHAT), a randomized, double-blind, active-controlled clinical trial con-ducted from February 1994 through March 2002.

Setting and Participants A total of 33357 participants aged 55 years or older withhypertension and at least 1 other CHD risk factor from 623 North American centers.

Interventions Participants were randomly assigned to receive chlorthalidone, 12.5to 25 mg/d (n=15255); amlodipine, 2.5 to 10 mg/d (n=9048); or lisinopril, 10 to 40mg/d (n=9054) for planned follow-up of approximately 4 to 8 years.

Main Outcome Measures The primary outcome was combined fatal CHD or non-fatal myocardial infarction, analyzed by intent-to-treat. Secondary outcomes were all-cause mortality, stroke, combined CHD (primary outcome, coronary revascularization,or angina with hospitalization), and combined CVD (combined CHD, stroke, treated an-gina without hospitalization, heart failure [HF], and peripheral arterial disease).

Results Mean follow-up was 4.9 years. The primary outcome occurred in 2956 par-ticipants, with no difference between treatments. Compared with chlorthalidone (6-year rate, 11.5%), the relative risks (RRs) were 0.98 (95% CI, 0.90-1.07) for amlo-dipine (6-year rate, 11.3%) and 0.99 (95% CI, 0.91-1.08) for lisinopril (6-year rate,11.4%). Likewise, all-cause mortality did not differ between groups. Five-year sys-tolic blood pressures were significantly higher in the amlodipine (0.8 mm Hg, P=.03)and lisinopril (2 mm Hg, P�.001) groups compared with chlorthalidone, and 5-yeardiastolic blood pressure was significantly lower with amlodipine (0.8 mm Hg, P�.001).For amlodipine vs chlorthalidone, secondary outcomes were similar except for a higher6-year rate of HF with amlodipine (10.2% vs 7.7%; RR, 1.38; 95% CI, 1.25-1.52).For lisinopril vs chlorthalidone, lisinopril had higher 6-year rates of combined CVD (33.3%vs 30.9%; RR, 1.10; 95% CI, 1.05-1.16); stroke (6.3% vs 5.6%; RR, 1.15; 95% CI,1.02-1.30); and HF (8.7% vs 7.7%; RR, 1.19; 95% CI, 1.07-1.31).

Conclusion Thiazide-type diuretics are superior in preventing 1 or more major formsof CVD and are less expensive. They should be preferred for first-step antihyperten-sive therapy.JAMA. 2002;288:2981-2997 www.jama.com

See also pp 2998 and 3039.

©2002 American Medical Association. All rights reserved. (Reprinted) JAMA, December 18, 2002—Vol 288, No. 23 2981

Downloaded From: https://jamanetwork.com/ by a Hospital Italiano de Buenos Aires User on 02/27/2020

coronary heart disease (CHD).6,12-16 Therelative benefit of various agents inhigh-risk hypertensive subgroups suchas older patients, black patients, and pa-tients with diabetes also needed to beestablished.17

The Antihypertensive and Lipid-Lowering Treatment to Prevent HeartAttack Trial (ALLHAT), a random-ized, double-blind, multicenter clini-cal trial sponsored by the NationalHeart, Lung, and Blood Institute, wasdesigned to determine whether the oc-currence of fatal CHD or nonfatal myo-cardial infarction is lower for high-risk patients with hypertension treatedwith a CCB (represented by amlo-dipine), an ACE inhibitor (repre-sented by lisinopril), or an �-blocker(represented by doxazosin), each com-pared with diuretic treatment (repre-sented by chlorthalidone).18 Chlortha-lidone was found to be superior todoxazosin and was previously re-ported after early termination of thedoxazosin arm of the trial.19,20 Second-ary outcomes included all-cause mor-tality, stroke, and other cardiovascu-lar disease (CVD) events. A lipid-lowering subtrial was designed todetermine whether lowering choles-terol with 3-hydroxy-3-methylglu-taryl coenzyme A reductase inhibitor(pravastatin) compared with usual carereduced all-cause mortality in a mod-erately hypercholesterolemic subset ofALLHAT participants.18,21 To evaluatedifferences in CVD effects of the vari-ous first-step drugs, ALLHAT was de-signed with a large sample size (9000-15000 participants/intervention arm)and long follow-up (4-8 years). Thisstudy presents results of the amlo-dipine and lisinopril vs chlorthali-done comparisons on major CVD out-comes.

METHODSStudy Design

The rationale and design of ALLHAThave been presented elsewhere.18 Par-ticipants were men and women aged 55years or older who had stage 1 or stage2 hypertension with at least 1 addi-tional risk factor for CHD events.18,22

The risk factors included previous (�6months) myocardial infarction orstroke, left ventricular hypertrophydemonstrated by electrocardiographyor echocardiography, history of type 2diabetes, current cigarette smoking,high-density lipoprotein cholesterol ofless than 35 mg/dL (�0.91 mmol/L),or documentation of other atheroscle-rotic CVD. Individuals with a historyof hospitalized or treated symptom-atic heart failure (HF) and/or known leftventricular ejection fraction of less than35% were excluded.

Unless the drug regimen had to betapered for safety reasons, individualscontinued any prior antihypertensivemedications until they received ran-domized study drug, at which pointthey stopped taking all previous medi-cations. Treatment with the study drugwas initiated the day after randomiza-tion. By telephone, participants wererandomly assigned to chlorthalidone,amlodipine, or lisinopril in a ratio of 1.7:1:1. The concealed randomizationscheme was generated by computer,implemented at the clinical trials cen-ter, stratified by center and blocked inrandom block sizes of 5 or 9 to main-tain balance. Participants (n=33357)were recruited at 623 centers in theUnited States, Canada, Puerto Rico, andthe US Virgin Islands between Febru-ary 1994 and January 1998. (The origi-nal reported number of 625 siteschanged because 2 sites and their pa-tients with poor documentation of in-formed consent were excluded.20) Allparticipants gave written informed con-sent, and all centers obtained institu-tional review board approval. Fol-low-up visits were at 1 month; 3, 6, 9,and 12 months; and every 4 monthsthereafter. The range of possible fol-low-up was 3 years 8 months to 8 years1 month. The closeout phase began onOctober 1, 2001, and ended on March31, 2002.

TreatmentTrained observers using standardizedtechniques measured BPs during thetrial.20 Visit BP was the average of 2seated measurements. Goal BP in each

randomized group was less than 140/90mm Hg achieved by titrating the as-signed study drug (step 1) and addingopen-label agents (step 2 or 3) when nec-essary. The choice of step 2 drugs (aten-olol, clonidine, or reserpine) was at thephysician’s discretion. Nonpharmaco-logic approaches to treatment of hyper-tension were recommended accordingto national guidelines.4,23 Step 1 drugswere encapsulated and identical in ap-pearance so that the identity of eachagent was double-masked at each dos-age level. Dosages were 12.5, 12.5 (shamtitration), and 25 mg/d for chlorthali-done; 2.5, 5, and 10 mg/d for amlo-dipine; and 10, 20, and 40 mg/d for li-sinopril. Doses of study-supplied open-label step 2 drugs were 25 to 100 mg/dof atenolol; 0.05 to 0.2 mg/d of reser-pine; or 0.1 to 0.3 mg twice a day ofclonidine; step 3 was 25 to 100 mg twicea day of hydralazine. Other drugs, in-cluding low doses of open-label step 1drug classes, were permitted if clini-cally indicated.18,20

OutcomesThe primary outcome was fatal CHD ornonfatal myocardial infarction com-bined.18 Four major prespecified sec-ondary outcomes were all-causemortality, fatal and nonfatal stroke,combined CHD (the primary out-come, coronary revascularization, hos-pitalized angina), and combined CVD(combined CHD, stroke, other treatedangina, HF [fatal, hospitalized, ortreated nonhospitalized], and periph-eral arterial disease). Coronary revas-cularization included coronary arterybypass graft, percutaneous angio-plasty, insertion of stents, and ather-ectomy. Individual components of thecombined outcomes were prespeci-fied and examined, as were other sec-ondary outcomes including cancer,incident electrocardiographic left ven-tricular hypertrophy, end-stage renaldisease (ESRD) (dialysis, renal trans-plant, or death), and slope of thereciprocal of longitudinal serum cre-atinine measurements. Change in es-timated glomerular filtration rate24,25

was examined post hoc.

ANTIHYPERTENSIVE TREATMENT IN ALLHAT

2982 JAMA, December 18, 2002—Vol 288, No. 23 (Reprinted) ©2002 American Medical Association. All rights reserved.


Study outcomes were assessed at fol-low-up visits and reported to the clini-cal trials center.18 Hospitalized out-comes were primarily based on clinicinvestigator reports, and copies of deathcertificates and hospital discharge sum-maries were requested. Among all com-bined CVD events that resulted indeaths, hospitalizations, or both, theproportion with documentation (ie, adeath certificate or a hospital dis-charge summary) was 99% in all 3 treat-ment groups. In addition, searches foroutcomes were accomplished throughthe Center for Medicare and MedicaidServices, the Department of VeteransAffairs, the National Death Index, andthe Social Security Administration da-tabases. A death was ascertained byclinic report or by match with the afore-mentioned databases plus a confirma-tory death certificate. A death pend-ing confirmation is one found usingdatabases but for which a confirma-tory death certificate has not yet beenobtained. Medical reviewers from theclinical trials center verified the phy-sician-assigned diagnoses of out-comes using death certificates and hos-pital discharge summaries. Moredetailed information was collected ona random (10%) subset of CHD andstroke events to validate the proce-dure of using physician diagnoses.18

When a large excess of HF became evi-dent in the doxazosin arm, a 1-timesample of HF hospitalizations was re-viewed by the ALLHAT Endpoints Sub-committee. Agreement rates betweenthe subcommittee and clinic investiga-tors were 90% (155/172) for the pri-mary outcome, 85% (33/39) for HF hos-pitalizations,26 and 84% (129/153) forstroke, and were similar in all treat-ment groups.

Two major safety outcomes, angio-edema and hospitalization for gastro-intestinal bleeding, were prespecified.Occurrence of gastrointestinal bleed-ing was ascertained from Center forMedicare and Medicaid Services andDepartment of Veterans Affairs hospi-talization databases, representing 74%of ALLHAT participants (persons �65years, Department of Veterans Affairs

participants, or both).27 Angioedemawas ascertained using a solicited eventquestion on a serious adverse eventform.

Statistical MethodsTo maximize statistical power, 1.7 timesas many participants were assigned tothe diuretic group as to each of the other3 groups.18 Given the achieved samplesize and expected event rate, treat-ment crossovers, and losses to follow-up, ALLHAT had 83% power to detecta 16% reduction in risk of the primaryoutcome between chlorthalidone andeach other group at a 2-sided �=.0178(z =2.37) to account for the 3 originalcomparisons.28 Data were analyzedaccording to participants’ randomizedtreatment assignments regardless oftheir subsequent medications (intent-to-treat analysis). Cumulative eventrates were calculated using the Kaplan-Meier method. Although rates are pre-sented only through 6 years, both thelog-rank test and Cox proportional haz-ards regression model incorporated theparticipant’s entire trial experience toevaluate differences between cumula-tive event curves and to obtain 2-sidedP values. Only the Cox proportionalhazard regression results are pre-sented, because P values were essen-tially identical. Hazard ratios (relativerisks [RRs]) and 95% confidence inter-vals (CIs) were obtained from the Coxproportional hazards regressionmodel.29 For consistency with �=.0178,95% CIs may be converted to 98.2% lim-its by multiplying the upper limit anddividing the lower limit by RR(0.41/Z),where Z is the value of the test statisticfor the RR estimate. The Cox propor-tional hazards regression modelassumption was examined by using log-log plots and testing a treatment � time(time-dependent) interaction term; if itwas violated, the RR estimate from a2-by-2 table was used.29 Heterogeneityof effects in prespecified subgroups, (1)men and women, (2) participants lessthan 65 and 65 years or older, (3) blackand nonblack participants, and (4) dia-betic and nondiabetic participants, andthe post hoc subgroups presence or

absence of CHD at baseline, was exam-ined by testing for treatment-covariateinteraction with the Cox proportionalhazards regression model by usingP�.05. SAS version 8.0 (SAS Institute,Cary, NC) and STATA version 7 (StataCorp, College Station, Tex) were usedfor statistical analyses.

A National Heart, Lung, and BloodInstitute–appointed data and safetymonitoring board met at least annu-ally to review the accumulating data andto monitor for safety and efficacy. TheLan-DeMets version of the O’Brien-Fleming group sequential boundarieswas used to assess treatment group dif-ferences, and conditional power wasused to assess futility.30,31

RESULTSPatient Characteristics

TABLE 1 presents baseline characteris-tics for the 33357 participants in thechlorthalidone, amlodipine, and lisin-opril treatment groups. The mean agewas 67 years; 47% were women, 35%were black, 19% were Hispanic, and36% were diabetic. There were nearlyidentical distributions of baseline fac-tors in the 3 treatment groups.22

Visit and Medication AdherenceFIGURE 1 shows the number of partici-pants randomized and followed up to thetime of closeout. In all 3 treatmentgroups, the mean (SD) length of fol-low-up was 4.9 years (1.4 years), and99% of expected person-years were ob-served. The maximum range of fol-low-up was 8.0, 7.9, and 8.1 years in thechlorthalidone, amlodipine, and lisin-opril groups, respectively. At trial close-out, 419 (2.7%) of the chlorthalidonegroup, 258 (2.8%) of the amlodipinegroup, and 276 (3.0%) of the lisinoprilgroup had unknown vital status. Amongparticipants with unknown vital sta-tus, the distributions of most baselinefactors were similar among the 3 treat-ment groups, but participants assignedto lisinopril were less likely to be blackand more likely to be women, have un-treated hypertension, evidence of CHDor atherosclerotic CVD, and a lowermean serum glucose.




Visit adherence decreased over timefrom about 92% at 1 year to 84% to 87%at 5 years in all 3 treatment groups(TABLE 2). Among participants in thechlorthalidone group who were con-

tacted in the clinic or by telephonewithin 12 months of annual sched-uled visits, 87.1% were taking chlortha-lidone or another diuretic at 1 year, de-creasing to 80.5% at 5 years; 67.5%

(n=4387) were taking a diuretic with-out a CCB or an ACE inhibitor; and13.2% were taking a diuretic with a CCB(5.8% [n=399]) or an ACE inhibitor(9.3% [n=641]). Only 9.0% were tak-ing either a CCB (5.8% [n=399]) or anACE inhibitor (5.6% [n=385]) with-out a diuretic at 5 years.

Among participants in the amlo-dipine group, 87.6% were taking amlo-dipineoranotherCCBat1year,decreas-ing to 80.4% at 5 years; and 63.8%(n=2502)were takingaCCBalonewith-out a diuretic. Another 16.6% were tak-ing a CCB with a diuretic, and only 6.9%were taking a diuretic without a CCB.Among participants in the lisinoprilgroup, 82.4% were taking lisinopril oranotherACEinhibitor at1year,decreas-ing to 72.6% at 5 years; 56.9% (n=2143)were takinganACEinhibitoralonewith-out a diuretic; and 15.7% were taking anACE inhibitor with a diuretic at 5 years.About 8.5% were taking a diuretic with-out an ACE inhibitor.

The most common reasons for nottaking step 1 medication at 5 years in thechlorthalidone, amlodipine, and lisin-opril groups were unspecified refusals(41.4% [n=775], 40.5% [n=443], and37.9% [n=552], respectively) and symp-tomatic adverse effects (15.0% [n=282],16.4% [n=180], and 18.1% [n=264],respectively). Elevated BP (4.5% [n=84],3.5% [n=38], and 9.0% [n=131]) orother adverse effects such as abnormallaboratory values (3.8% [n=71], 1.6%[n=17], and 2.3% [n=34]) were otherreasons given for discontinuation of step1 medications. Among participants withavailable medication information at 1year, 26.7%, 25.9%, and 32.6% of thoseassigned to chlorthalidone, amlo-dipine, and lisinopril, respectively, weretaking a step 2 or step 3 drug. At 5 years,the corresponding percentages were40.7%, 39.5%, and 43.0%, respec-tively. Usage patterns of specific step 2drugs were similar among groups.Participants could be taking more than1 step-up drug. At 1 year, 40.0%(n=4645), 44.0% (n=3017), and 43.8%(n=2764) of participants assigned tochlorthalidone, amlodipine, and lisin-opril, respectively, still taking their

Table 1. Baseline Characteristics of the ALLHAT Participants*

Characteristic

No. of Participants (%)

Chlorthalidone(n = 15 255)

Amlodipine(n = 9048)

Lisinopril(n = 9054)

Age, mean (SD), y 66.9 (7.7) 66.9 (7.7) 66.9 (7.7)

Age range, y55-64 6471 (42.4) 3844 (42.5) 3869 (42.7)

�65 8784 (57.6) 5204 (57.5) 5185 (57.3)

EthnicityWhite, non-Hispanic 7202 (47.2) 4305 (47.6) 4262 (47.1)

Black, non-Hispanic 4871 (31.9) 2911 (32.2) 2920 (32.3)

White Hispanic 1912 (12.5) 1108 (12.2) 1136 (12.5)

Black Hispanic 498 (3.3) 302 (3.3) 290 (3.2)

Other 772 (5.1) 422 (4.7) 446 (4.9)

Women 7171 (47.0) 4280 (47.3) 4187 (46.2)

Education, mean (SD), y 11.0 (4.0) 11.0 (3.9) 11.0 (4.1)

Receiving antihypertensivetreatment

13 754 (90.2) 8171 (90.3) 8164 (90.2)

Blood pressure, mean (SD), mm Hg 146 (16)/84 (10) 146 (16)/84 (10) 146 (16)/84 (10)

Treated at baseline 145 (16)/83 (10) 145 (16)/83 (10) 145 (16)/84 (10)

Untreated at baseline 156 (12)/89 (9) 157 (12)/90 (9) 156 (12)/89 (9)

Eligibility risk factors†Cigarette smoker 3342 (21.9) 1980 (21.9) 1981 (21.9)

Atherosclerotic CVD‡ 7900 (51.8) 4614 (51.0) 4684 (51.7)

History of MI or stroke 3581 (23.5) 2098 (23.2) 2058 (22.7)

History of coronaryrevascularization

1986 (13.0) 1106 (12.2) 1218 (13.5)

Other atherosclerotic CVD 3604 (23.6) 2145 (23.7) 2152 (23.8)

Major ST depression orT-wave inversion

1572 (10.4) 908 (10.1) 940 (10.5)

Type 2 diabetes 5528 (36.2) 3323 (36.7) 3212 (35.5)

HDL-C �35 mg/dL 1798 (11.8) 1018 (11.3) 1061 (11.7)

LVH by electrocardiogram 2467 (16.2) 1533 (16.9) 1474 (16.3)

LVH by echocardiogram 695 (4.6) 411 (4.6) 402 (4.5)

History of CHD at baseline§ 3943 (26.0) 2202 (24.5) 2270 (25.3)

Body mass index, mean (SD) 29.7 (6.2) 29.8 (6.3) 29.8 (6.2)

Current medication useAspirin 5426 (35.6) 3268 (36.1) 3258 (36.0)

Estrogen supplementation(women only)

1273 (17.8) 752 (17.6) 727 (17.4)

Lipid trial participants 3755 (24.6) 2240 (24.8) 2167 (23.9)

*ALLHAT indicates Antihypertensive and Lipid-Lowering Treatment to Prevent Heart Attack Trial; CVD, cardiovasculardisease; MI, myocardial infarction; HDL-C, high-density lipoprotein cholesterol; LVH, left ventricular hypertrophy; andCHD, coronary heart disease. Body mass index was calculated as weight in kilograms divided by the square of heightin meters. To convert HDL-C to mmol/L, multiply by 0.0259.

†For trial eligibility, participants had to have at least 1 other risk factor in addition to hypertension. Thus, the indicatedrisk factors are not mutually exclusive or exhaustive and may not represent prevalence.

‡History of MI or stroke, history of coronary revascularization, major ST segment depression or T-wave inversion onany electrocardiogram in the past 2 years, other atherosclerotic CVD (history of angina pectoris; history of intermit-tent claudication, gangrene, or ischemic ulcers; history of transient ischemic attack; coronary, peripheral vascular, orcarotid stenosis �50% documented by angiography or Doppler studies; ischemic heart disease documented byreversible or fixed ischemia on stress thalium or dipyridamole thalium, ST depression �1 mm for �1 minute on ex-ercise testing or Holter monitoring; reversible wall motion abnormality on stress echocardiogram; ankle-arm index�0.9; abdominal aortic aneurysm detected by ultrasonography, computed tomography scan, or radiograph; carotidor femoral bruits).

§P = .03 for comparison of groups.




blinded medication were receiving themaximal study dose. At 5 years, the per-centages were 56.9% (n=2629), 65.7%(n=1856), and 60.3% (n=1391), re-spectively.

Intermediate OutcomesGiven the large sample size in ALLHAT,almost all differences in follow-up BPand biochemical measurements werestatistically significant (TABLE 3 andTABLE 4). Mean seated BP at random-ization was about 146/84 mm Hg in all3 groups, with 90% of participants re-porting current antihypertensive drugtreatment (Table 1). Follow-up BPs inall 3 groups are shown in Table 3 andFIGURE 2.

Mean total serum cholesterol levelsat baseline and 4 years follow-up areshown in Table 4. At 4 years, about 35%to 36% of participants in all 3 groups

reported taking lipid-lowering drugs,largely statins, some as a result of par-ticipation in the ALLHAT lipid trial.Mean serum potassium levels at base-line and follow-up are also shown;about 8% of the chlorthalidone groupwere receiving potassium supplemen-tation at 5 years compared with 4% inthe amlodipine group and 2% in the li-sinopril group. Among individuals clas-sified as nondiabetic at baseline, withbaseline fasting serum glucose less than126 mg/dL (7.0 mmol/L), incidence ofdiabetes (fasting serum glucose, �126mg/dL [7.0 mmol/L]) at 4 years was11.6%, 9.8%, and 8.1%, respectively.

Mean estimated glomerular filtra-tion rate at baseline was about 78 mL/min per 1.73 m2 in all groups. At 4years, it was 70.0, 75.1, and 70.7 mL/min per 1.73 m2 in the chlorthali-done, amlodipine, and lisinopril groups,

respectively. The slopes of the recip-rocal of serum creatinine over time werevirtually identical in the chlorthali-done and lisinopril groups (–0.018 and–0.019 dL/mg per year), whereas thedecline in the amlodipine slope (–0.012dL/mg per year) was less than that ofthe chlorthalidone slope (P�.001).

Primary and Secondary OutcomesAmlodipine vs Chlorthalidone. No sig-nificant difference was observed be-tween amlodipine and chlorthalidonefor the primary outcome (RR, 0.98; 95%CI, 0.90-1.07) or for the secondary out-comes of all-cause mortality, com-bined CHD, stroke, combined CVD,angina, coronary revascularization,peripheral arterial disease, cancer,or ESRD (TABLE 5, FIGURE 3, andFIGURE 4). The amlodipine group hada 38% higher risk of HF (P�.001) with

Figure 1. Randomization and Follow-up of Participants in the Antihypertensive and Lipid-Lowering Treatment to Prevent Heart Attack Trial

7479 Known Alive1256 Confirmed Dead

55 Deaths Pending Confirmation200 Lost to Follow-up58 Refused Follow-up

Status at Study Closeout7412 Known Alive1314 Confirmed Dead

52 Deaths Pending Confirmation218 Lost to Follow-up58 Refused Follow-up

Status at Study Closeout769 Confirmed Dead361 Vital Status Unknown

Status as of February 15, 200012 530 Known Alive

2203 Confirmed Dead103 Deaths Pending Confirmation339 Lost to Follow-up80 Refused Follow-up

Status at Study Closeout

Year 113 854 Completed Visit

2235 Discontinued Study Drug

6210 Completed Visit1873 Discontinued Study Drug

775 Unspecified Refusal282 Symptomatic Adverse Event84 Blood Pressure Elevation91 Blood Pressure Too Low73 Morbid Event71 Other Adverse Effects

125 Other Nonmedical Reasons638 Other

Year 5

Year 18215 Completed Visit1357 Discontinued Study Drug


443 Unspecified Refusal180 Symptomatic Adverse Event38 Blood Pressure Elevation51 Blood Pressure Too Low45 Morbid Event17 Other Adverse Effects83 Other Nonmedical Reasons

387 Other

Year 5



552 Unspecified Refusal264 Symptomatic Adverse Event131 Blood Pressure Elevation76 Blood Pressure Too Low49 Morbid Event34 Other Adverse Effects94 Other Nonmedical Reasons

484 Other

Year 5


Year 5NA

42 418 Patients Randomized

15 255 Assigned to ReceiveChlorthalidone

9048 Assigned to ReceiveAmlodipine

9054 Assigned to ReceiveLisinopril

9061 Assigned to ReceiveDoxazosin

15 255 Included in Analysis 9048 Included in Analysis 9054 Included in Analysis

NA indicates not applicable. Eligibility data were not collected for nonrandomized screenees. All randomized participants were included in the analyses. A patient mayhave more than 1 reason for discontinuing study drug; therefore, numbers do not sum to total. On January 24, 2000, the National Heart, Lung, and Blood Institutedecided to discontinue the doxazosin group and report results.18 Study closeout for chlorthalidone, amlodipine, and lisinopril groups was from October 1, 2001, throughMarch 31, 2002. Collection of last events for the doxazosin group had a closeout interval from October 15, 1999, through February 15, 2000, which captures moreinformation than that reported previously.24




a 6-year absolute risk difference of 2.5%and a 35% higher risk of hospitalized/fatal HF (P�.001). The treatment ef-fects for all outcomes were consistentacross the predefined subgroups(FIGURE 5) and by absence or pres-ence of CHD at baseline. Cause-specific mortality rates (except forunintentional injuries/suicides/homicides in amlodipine comparedwith chlorthalidone, not a prespeci-fied hypothesis) were similar for the 2groups (TABLE 6).

Lisinopril vs Chlorthalidone. No sig-nificant difference was observed be-tween lisinopril and chlorthalidone forthe primary outcome (RR, 0.99; 95% CI,

0.91-1.08) or for the secondary out-comes of all-cause mortality, combinedCHD, peripheral arterial disease, can-cer, or ESRD (Table 5, Figures 3 and 4).Cause-specific mortality rates were alsosimilar in the 2 groups (Table 6). The li-sinopril group had a 15% higher risk forstroke (P=.02) and a 10% higher risk ofcombined CVD (P�.001), with a 6-yearabsolute risk difference for combinedCVD of 2.4%. Included in this analysiswas a 19% higher risk of HF (P�.001),a 10% higher risk of hospitalized/fatal HF(P=.11), an 11% higher risk of hospital-ized/treated angina (P=.01), and a 10%higher risk of coronary revasculariza-tion (P=.05). The treatment effects for

all outcomes were consistent across sub-groups by sex, diabetic status (FIGURE 6),and baseline CHD status. For com-bined CHD, there was a significantdifferential effect by age (P=.01 for in-teraction) with RRs (lisinopril vschlorthalidone) of 0.94 for those lessthan 65 years vs 1.11 in those 65 yearsor older. However, when age was mod-eled as a continuous variable, there wasno significant interaction. For stroke andcombined CVD, there was a significantdifferential effect by race (P=.01 andP=.04 for interaction, respectively). TheRRs (lisinopril vs chlorthalidone) were1.40 (95% CI, 1.17-1.68) and 1.00 (95%CI, 0.85-1.17) for stroke and 1.19 (95%

Table 2. Visits Expected and Completed and Antihypertensive Medication Use at Annual Visits

Years, No. (%)

1 2 3 4 5

ChlorthalidoneExpected visits 15 067 (98.8) 14 711 (96.4) 14 272 (93.6) 12 380 (81.2) 7243 (47.5)

Completed visits 13 854 (91.9) 12 988 (88.3) 12 335 (86.4) 10 618 (85.8) 6210 (85.7)

Receiving blinded study drug 11 618 (83.9) 10 367 (79.8) 9372 (76.0) 8149 (72.9) 4623 (71.2)

Receiving blinded study drug or same class 12 063 (87.1) 11 001 (84.7) 10 202 (82.7) 9034 (80.8) 5247 (80.5)

Full crossovers* 707 (5.1) 865 (6.7) 944 (7.7) 921 (8.2) 583 (9.0)

Partial crossovers† 469 (3.4) 770 (5.9) 1054 (8.5) 1223 (10.9) 860 (13.2)

Receiving step 2 or 3‡ 3703 (26.7) 4185 (32.2) 4395 (35.6) 4244 (38.0) 2642 (40.7)

Other antihypertensive medication 594 (4.3) 586 (4.5) 618 (5.0) 609 (5.5) 320 (4.9)

No. of antihypertensive medications, mean (SD) 1.4 (0.7) 1.5 (0.8) 1.6 (0.9) 1.7 (1.0) 1.8 (1.0)

AmlodipineExpected visits 8937 (98.8) 8733 (96.5) 8510 (94.0) 7411 (81.9) 4343 (48.0)

Completed visits 8215 (91.9) 7672 (87.9) 7355 (86.4) 6341 (85.6) 3769 (86.8)

Receiving blinded study drug 6858 (83.5) 6106 (79.6) 5630 (76.6) 4886 (73.3) 2826 (72.1)

Receiving blinded study drug or same class 7192 (87.6) 6532 (85.2) 6116 (83.2) 5367 (80.5) 3151 (80.4)

Full crossovers* 232 (2.8) 342 (4.5) 369 (5.0) 401 (6.0) 270 (6.9)





LisinoprilExpected visits 8942 (98.8) 8725 (96.4) 8458 (93.4) 7356 (81.2) 4315 (47.7)

Completed visits 8158 (91.2) 7574 (86.8) 7185 (84.9) 6142 (83.5) 3605 (83.5)

Receiving blinded study drug 6316 (77.4) 5418 (71.5) 4897 (68.2) 4155 (64.4) 2307 (61.2)

Receiving blinded study drug or same class 6721 (82.4) 5944 (78.4) 5536 (77.1) 4824 (74.8) 2736 (72.6)

Full crossovers* 285 (3.5) 387 (5.1) 430 (6.0) 455 (7.0) 320 (8.5)





*Full crossovers: (1) assigned to chlorthalidone, not on step 1, no open-label diuretic, but on open-label calcium channel blocker (CCB) or angiotensin-converting enzyme (ACE)inhibitor; (2) assigned to lisinopril, not on step 1, no open-label ACE inhibitor, but on open-label diuretic; (3) assigned to amlodipine, not on step 1, no open-label CCB, but onopen-label diuretic.

†Partial crossovers: (1) assigned to chlorthalidone, on step 1 or open-label diuretic and on open-label CCB or ACE inhibitor; (2) assigned to lisinopril, on step 1 or open-label ACEinhibitor and on open-label diuretic; (3) assigned to amlodipine, on step 1 or open-label CCB and on open-label diuretic.

‡Step 2: atenolol, clonidine, or reserpine; step 3: hydralazine.




CI, 1.09-1.30) and 1.06 (95% CI, 1.00-1.13) for combined CVD in blacks andnonblacks, respectively.

The mean follow-up systolic BP forall participants was 2 mm Hg higher inthe lisinopril group than the chlortha-lidone group, 4 mm Hg higher inblacks, and 3 mm Hg higher in those65 years or older. Adjustment for fol-low-up BP as time-dependent covari-ates in a Cox proportional hazards re-gression model slightly reduced the RRsfor stroke (1.15 to 1.12) and HF (1.20to 1.17) overall and in the black sub-group (stroke, 1.40 to 1.35; and HF,1.32 to 1.26), but the results re-mained statistically significant.

Primary Safety OutcomesSix-year rates of hospitalization for gas-trointestinal bleeding, available only inMedicare and Department of VeteransAffairs participants, occurred in 8.8%,8.0%, and 9.6% participants in thechlorthalidone, amlodipine, and lisin-opril treatment groups, respectively,with no significant differences (Table5). Angioedema occurred in 8 of 15255(0.1%), 3 of 9048 (�0.1%), and 38 of9054 (0.4%) persons in the chlortha-lidone, amlodipine, and lisinopril treat-ment groups, respectively. Significantdifferences were seen for the lisinoprilvs chlorthalidone comparison overall(P�.001), in blacks (2 of 5369 [�0.1%]for chlorthalidone, 23 of 3210 [0.7%]for lisinopril; P�.001), and in non-blacks (6 of 9886 [0.1%] for chlortha-lidone, 15 of 5844 for lisinopril [0.3%];P=.002). The only death from angio-edema was in the lisinopril group.

COMMENTNeither amlodipine (representing CCBs,particularly dihydropyridine [DHP]–CCBs) nor lisinopril (representing ACEinhibitors) was superior to chlorthali-done (representing thiazide-type diuret-ics) in preventing major coronary eventsor in increasing survival. Chlorthali-done was superior to amlodipine (byabout 25%) in preventing HF, overall,and for hospitalized or fatal cases, al-though it did not differ from amlo-dipine in overall CVD prevention.

Chlorthalidone was superior to lisino-pril in lowering BP and in preventing ag-gregate cardiovascular events, princi-pally stroke, HF, angina, and coronaryrevascularization. ALLHAT previouslyreported that chlorthalidone was supe-

rior to doxazosin (representing �-block-ers) in reducing BP and preventing car-diovascular events, particularly HF.19,20

It is not surprising that no signifi-cant differences in CHD and stroke rateswere found between chlorthalidone and

Table 3. Number of Participants, Mean Blood Pressure, Achieved Blood Pressure Goal, andBlood Pressure Difference at Baseline and Annual Visits

Chlorthalidone Amlodipine Lisinopril

P Value

Amlodipine vsChlorthalidone

Lisinopril vsChlorthalidone

No. of Participants (%)

Baseline 15 255 (100) 9048 (100) 9054 (100)

1 Year 12 862 (84.3) 7609 (84.1) 7521 (83.1)

2 Years 11 740 (77.0) 6883 (76.1) 6700 (74.0)

3 Years 10 698 (70.1) 6381 (70.5) 6076 (67.1)

4 Years 9379 (61.5) 5637 (62.3) 5325 (58.8)

5 Years 5301 (34.7) 3195 (35.3) 2963 (32.7)

Systolic Blood Pressure, Mean (SD), mm Hg

Baseline 146.2 (15.7) 146.2 (15.7) 146.4 (15.7) .98 .39

1 Year 136.9 (15.8) 138.5 (14.9) 140.0 (18.5) �.001 �.001

2 Years 135.9 (15.9) 137.1 (15.0) 138.4 (17.9) �.001 �.001

3 Years 134.8 (15.4) 135.6 (15.2) 136.7 (17.3) .001 �.001

4 Years 133.9 (15.7) 134.8 (15.0) 135.5 (17.2) .002 �.001

5 Years 133.9 (15.2) 134.7 (14.9) 135.9 (17.9) .03 �.001

Diastolic Blood Pressure, Mean (SD), mm Hg

Baseline 84.0 (10.1) 83.9 (10.2) 84.1 (10.0) .52 .49

1 Year 79.3 (9.9) 78.7 (9.5) 79.9 (10.5) �.001 �.001

2 Years 78.3 (9.6) 77.7 (9.6) 78.6 (10.3) �.001 .03

3 Years 77.2 (9.5) 76.4 (9.6) 77.3 (10.3) �.001 .42

4 Years 76.5 (9.7) 75.7 (9.6) 76.6 (10.4) �.001 .48

5 Years 75.4 (9.8) 74.6 (9.9) 75.4 (10.7) �.001 .94

Achieved Blood Pressure Goal of �140/90 mm Hg, No. (%)

Baseline 4149 (27.2) 2497 (27.6) 2381 (26.3) .56 .12

1 Year 7434 (57.8) 4200 (55.2) 3806 (50.6) �.001 �.001

2 Years 7161 (61.0) 3951 (57.4) 3625 (54.1) �.001 �.001

3 Years 6836 (63.9) 4046 (63.4) 3597 (59.2) .54 �.001

4 Years 6293 (67.1) 3709 (65.8) 3360 (63.1) .15 �.001

5 Years 3615 (68.2) 2118 (66.3) 1813 (61.2) .09 �.001

Systolic Blood Pressure, � mm Hg*

Baseline 0 0.2

1 Year 1.6 3.1

2 Years 1.2 2.5

3 Years 0.8 1.9

4 Years 0.9 1.6

5 Years 0.8 2.0

Diastolic Blood Pressure, � mm Hg*

Baseline −0.1 0.1

1 Year −0.6 0.6

2 Years −0.6 0.3

3 Years −0.8 0.1

4 Years −0.8 0.1

5 Years −0.8 0

*Compared with chlorthalidone group.




amlodipine-based therapy in ALLHAT.In the Systolic Hypertension in theElderly Program and the Systolic Hy-pertension in Europe trial, in which athiazide-like diuretic (chlorthali-done) or a DHP-CCB was comparedwith a placebo, major CHD events werereduced by 27% and 30%, and strokeby 37% and 42%, respectively.8,9 Moredirect evidence comes from 2 large ac-tive-controlled trials that comparedDHP-CCB and traditional first-stepdrugs. The Swedish Trial in Old Pa-tients with Hypertension-2 and the In-ternational Nifedipine GITS (long-acting gastrointestinal formulation)Study: Intervention as a Goal in Hy-pertension Treatment (INSIGHT),found no significant differences for ma-jor CHD or stroke rates between thetreatment groups.32,33 Some of these in-dividual trials have had limited powerto evaluate differences between treat-ments.34 In meta-analyses of 5 positive-controlled trials, which included bothDHP-CCB and non–DHP-CCB trials,there were trends that favored CCB-based therapy for stroke and tradi-tional treatment for CHD, with no dif-ference for all-cause mortality.13,14

However, ALLHAT observed approxi-mately the same number of strokes andnearly twice as many CHD events as all5 trials combined, which suggests thatthe aggregate of the evidence would in-dicate no difference between CCB-based treatment and diuretic-basedtreatment for these outcomes.

The amlodipine vs chlorthalidonefindings for HF reinforce previous trialresults. In the diuretic-based Systolic Hy-pertension in the Elderly Program, ac-tive therapy reduced HF occurrence by49% compared with placebo (P�.001),although in the DHP-CCB–based Sys-tolic Hypertension in Europe trial, it wasreduced by 29% (not statistically sig-nificant).9,35 In the INSIGHT trial, HFwas approximately twice as frequent inthe CCB vs the diuretic arm.33 The pre-viously cited meta-analyses reported ahigher rate of HF with CCB-based treat-ment than traditional regimens, with nodifference in RR for DHPs comparedwith non–DHPs.13,14

Table 4. Biochemical Changes by Treatment Group*


P Value



Cholesterol, mg/dL

No. of participants (%)Baseline 14 483 (94.9) 8586 (94.9) 8573 (94.7)

2 Years 10 206 (66.9) 6025 (66.6) 5739 (63.4)

4 Years 8495 (55.7) 5025 (55.5) 4711 (52.0)

Mean (SD)Baseline 216.1 (43.8) 216.5 (44.1) 215.6 (42.4) .47 .38

2 Years 205.3 (42.1) 202.5 (42.2) 202.0 (42.8) �.001 �.001

4 Years 197.2 (42.1) 195.6 (41.0) 195.0 (40.6) .009 �.001

�240 mg/dL, No. (%)Baseline 3838 (26.5) 2284 (26.6) 2178 (25.4) .89 .06

2 Years 1898 (18.6) 1018 (16.9) 976 (17.0) .005 .03

4 Years 1223 (14.4) 673 (13.4) 603 (12.8) .13 .005

Potassium, mEq/L


2 Years 9877 (64.7) 5794 (64.0) 5516 (60.9)

4 Years 8315 (54.5) 4919 (54.4) 4616 (51.0)

Mean (SD)Baseline 4.3 (0.7) 4.3 (0.7) 4.4 (0.7) .59 .001

2 Years 4.0 (0.7) 4.3 (0.7) 4.5 (0.7) �.001 �.001

4 Years 4.1 (0.7) 4.4 (0.7) 4.5 (0.7) �.001 �.001

�3.5 mEq/L, No. (%)Baseline 493 (3.4) 292 (3.4) 223 (2.6) .99 .001

2 Years 1254 (12.7) 151 (2.6) 83 (1.5) �.001 �.001

4 Years 707 (8.5) 93 (1.9) 37 (0.8) �.001 �.001

Fasting Glucose, mg/dL


2 Years 5980 (39.2) 3506 (38.7) 3333 (36.8)

4 Years 4972 (32.6) 2954 (32.6) 2731 (30.2)

Mean (SD)Baseline 123.5 (58.3) 123.1 (57.0) 122.9 (56.1) .71 .54

2 Years 127.6 (59.2) 122.4 (54.2) 120.8 (54.0) �.001 �.001

4 Years 126.3 (55.6) 123.7 (52.0) 121.5 (51.3) .20 .002

�126 mg/dL, No. (%)Baseline 3258 (28.9) 1941 (29.2) 1985 (29.4) .68 .55

2 Years 1967 (32.9) 1048 (29.9) 947 (28.4) �.001 �.001

4 Years 1626 (32.7) 901 (30.5) 784 (28.7) .11 �.001

Fasting Glucose Among Nondiabetics With Baseline Fasting Glucose �126 mg/dL

No. of participants (%)Baseline 6766 (100) 3954 (100) 4096 (100)

2 Years 3074 (45.4) 1787 (45.2) 1737 (42.4)

4 Years 2606 (40.3) 1567 (39.6) 1464 (35.7)

Mean (SD)Baseline 93.1 (11.7) 93.0 (11.4) 93.3 (11.8) .52 .45

2 Years 102.2 (27.1) 99.0 (22.5) 97.4 (20.0) �.001 �.001

4 Years 104.4 (28.5) 103.1 (27.7) 100.5 (19.5) .11 �.001

�126 mg/dL, No. (%)2 Years 295 (9.6) 132 (7.4) 101 (5.8) .006 �.001

4 Years 302 (11.6) 154 (9.8) 119 (8.1) .04 �.001(continued)




A body of literature based on obser-vational studies and secondary CHD pre-vention trials of short-acting CCBs hassuggested that CCBs, especially DHP-CCBs, may increase the risk of cancer,gastrointestinal bleeding, and all-causemortality.14,36,37 The results of ALLHATdo not support these findings. In fact,the mortality from noncardiovascularcauses was significantly lower in theCCB group (Table 6).

There were no significant differ-ences in the incidence of ESRD be-tween chlorthalidone and amlodipine,consistent with findings from theINSIGHT trial.33 Comparison of the re-ciprocal serum creatinine slopes sug-gested a slower decline in kidney func-tion in the amlodipine group. However,this finding requires cautious interpre-tation because studies assessing glomer-ular filtration rate more directly haveshown a hemodynamically mediatedacute increase in glomerular filtrationrate followed by a more rapid rate of de-cline with chronic therapy using amlo-dipine and other CCBs.38-40

Comparison of the lisinopril andchlorthalidone groups revealed betterdrug tolerance and BP control withchlorthalidone. Angioedema, a rare butpotentially serious adverse effect of ACEinhibitor use, occurred 4 times morefrequently in participants randomizedto lisinopril than in those randomizedto chlorthalidone. Cholesterol levels,the prevalence of hypokalemia (se-rum potassium �3.5 mEq/L), and newdiabetes (fasting glucose �126 mg/dL[�7.0 mmol/L]) were higher in thechlorthalidone than the other groupsfollowing 2 and 4 years of follow-up.Overall, these metabolic differences didnot translate into more cardiovascularevents or into higher all-cause mortal-ity in the chlorthalidone group com-pared with the other 2 groups.

The ALLHAT findings for some ma-jor outcomes are consistent with pre-dictions from placebo-controlled tri-als involving ACE inhibitors anddiuretics. Specifically, for ACE inhibi-tor and diuretic trials, respectively, thereductions in CHD rates were 20% and18%, and for all-cause mortality, 16%

and 10%.13 The 10% greater rate of com-bined CVD in the lisinopril than in thechlorthalidone group was due to in-creased occurrences of stroke, HF, an-gina, and coronary revascularization.Results for some of these outcomes mayseem surprising, because of reports ofbeneficial effects of ACE inhibitors onsurrogate markers of atherosclerosis andreductions in vascular and renal eventsin individuals with HF, diabetes, kid-ney disease, and cerebrovascular dis-ease in placebo-controlled trials.41-43

However, the finding in ALLHAT thatHF incidence was lower in the di-uretic vs the ACE inhibitor group is alsoconsistent with previous reports. In theSystolic Hypertension in the ElderlyProgram trial (chlorthalidone vs pla-cebo), there was a 49% decrease in the

development of HF, whereas in theStudies of Left Ventricular Dysfunc-tion Prevention (enalapril vs placebo)and Heart Outcomes Prevention Evalu-ation trials (ramipril vs placebo), therewere only 20% and 23% reductions,respectively.8,10,44 In published meta-analyses of placebo-controlled trials, thereductions in rates for stroke with ACEinhibitor and diuretics were 30% and34%, translating into nearly equiva-lent results.3,13 The 15% relative in-crease in stroke incidence for lisino-pril compared with chlorthalidonetreatment in ALLHAT must be consid-ered in the context of heterogeneity ofthe results by race. The Swedish Trialin Old Patients with Hypertension-2trial, which compared ACE inhibitorswith conventional treatment (diuret-

Figure 2. Mean Systolic and Diastolic Blood Pressure by Year During Follow-up

150

135

140

145

1300 1 3 4 5 62

Follow-up, y

Mean Systolic Blood Pressure

mm

Hg

90

75

80

85

700 1 3 4 5 62

Follow-up, y

Mean Diastolic Blood Pressure

mm

Hg

AmlodipineLisinopril

Chlorthalidone

Number measured at baseline through 5 years is given in Table 3; numbers at 6 years for chlorthalidone, am-lodipine, and lisinopril are 2721, 1656, and 1551, respectively.

Table 4. Biochemical Changes by Treatment Group* (cont)


P Value



Estimated Glomerular Filtration Rate, mL/min per 1.73 m2†


2 Years 9877 (64.7) 5794 (64.0) 5516 (60.9)

4 Years 8316 (54.5) 4924 (54.4) 4621 (51.0)

Mean (SD)Baseline 77.6 (19.7) 78.0 (19.7) 77.7 (19.9) .08 .57

2 Years 73.3 (19.9) 78.0 (20.5) 74.0 (20.0) �.001 .002

4 Years 70.0 (19.7) 75.1 (20.7) 70.7 (20.1) �.001 .03

*To convert serum cholesterol to mmol/L, multiply by 0.0259; fasting glucose to mmol/L, multiply by 0.0555.†Simplified 4-variable Modification of Diet in Renal Disease Study formula.24,25




ics and/or �-blockers), showed no sig-nificant differences in CHD, stroke, HF,or all-cause mortality.32 Although thesefindings are somewhat different fromthe experience in ALLHAT, consider-ation needs to be given to respectiveconfidence limits, population differ-ences (especially race), and study de-signs (open vs double-blind).

No substantial differences in inci-dence of ESRD, glomerular filtration rate,or reciprocal creatinine slopeswerenotedfor the lisinopril vs chlorthalidone com-parisons. The ALLHAT study popula-tion was selected for high CVD risk andhad a baseline mean creatinine of only1.0 mg/dL (88.4 µmol/L). More de-tailed analyses of high renal risk sub-groups (ie, diabetic, renal-impaired, andblack patients) will be the subject of sub-sequent reports.

Analyses of RRs for stroke and HF ad-justed for follow-up BP suggest that the2-mm Hg systolic BP difference over-

Figure 3. Cumulative Event Rates for the Primary Outcome (Fatal Coronary Heart Disease orNonfatal Myocardial Infarction) by Treatment Group

20

4

8

12

16

0

No. at Risk

1 3 4 5 6 72

15 255Chlorthalidone 14 477 13 102 11 362 6340 2956 20913 8209048Amlodipine 8576 7843 6824 3870 1878 21582189054Lisinopril 8535 7711 6662 3832 1770 1958123

Time to Event, y

Cum

ulat

ive

Eve

nt R

ate,

%

Chlorthalidone


No significant difference was observed for amlodipine (relative risk [RR], 0.98; 95% confidence interval [CI],0.90-1.07; P = .65) or lisinopril (RR, 0.99; 95% CI, 0.91-1.08; P = .81) vs chlorthalidone with a mean fol-low-up of 4.9 years.

Table 5. Clinical Outcomes by Antihypertensive Treatment Group*

Chlorthalidone Amlodipine Lisinopril Amlodipine vsChlorthalidone Lisinopril vs Chlorthalidone

No. ofTotal

Events

6-Year Rateper 100

Persons (SE)

No. ofTotal

Events

6-Year Rateper 100

Persons (SE)

No. ofTotal

Events

6-Year Rateper 100

Persons (SE)RR

(95% CI)Z

ScoreP

ValueRR

(95% CI)Z

ScoreP

Value

Primary outcomeCHD† 1362 11.5 (0.3) 798 11.3 (0.4) 796 11.4 (0.4) 0.98 (0.90-1.07) −0.46 .65 0.99 (0.91-1.08) −0.24 .81

Secondary outcomesAll-cause mortality 2203 17.3 (0.4) 1256 16.8 (0.5) 1314 17.2 (0.5) 0.96 (0.89-1.02) −1.27 .20 1.00 (0.94-1.08) 0.12 .90Combined CHD‡ 2451 19.9 (0.4) 1466 19.9 (0.5) 1505 20.8 (0.5) 1.00 (0.94-1.07) 0.04 .97 1.05 (0.98-1.11) 1.35 .18Stroke 675 5.6 (0.2) 377 5.4 (0.3) 457 6.3 (0.3) 0.93 (0.82-1.06) −1.09 .28 1.15 (1.02-1.30) 2.31 .02Combined CVD‡ 3941 30.9 (0.5) 2432 32.0 (0.6) 2514 33.3 (0.6) 1.04 (0.99-1.09) 1.55 .12 1.10 (1.05-1.16) 3.78 �.001End-stage renal

disease193 1.8 (0.1) 129 2.1 (0.2) 126 2.0 (0.2) 1.12 (0.89-1.40) 0.98 .33 1.11 (0.88-1.38) 0.87 .38

Cancer 1170 9.7 (0.3) 707 10.0 (0.4) 703 9.9 (0.4) 1.01 (0.92-1.11) 0.30 .77 1.02 (0.93-1.12) 0.42 .67Hospitalized for

gastrointestinalbleeding§

817 8.8 (0.3) 449 8.0 (0.4) 526 9.6 (0.4) 0.92 (0.82-1.03) −1.44 .15 1.11 (0.99-1.24) 1.82 .07

Components of secondaryoutcomes

Heart failure 870 7.7 (0.3) 706 10.2 (0.4) 612 8.7 (0.4) 1.38 (1.25-1.52) 6.29 �.001 1.19 (1.07-1.31)� 3.33 �.001Hospitalized/fatal

heart failure724 6.5 (0.3) 578 8.4 (0.4) 471 6.9 (0.4) 1.35 (1.21-1.50) 5.37 �.001 1.10 (0.98-1.23)� 1.59 .11

Angina (hospitalizedor treated)

1567 12.1 (0.3) 950 12.6 (0.4) 1019 13.6 (0.4) 1.02 (0.94-1.10) 0.42 .67 1.11 (1.03-1.20) 2.59 .01

Angina(hospitalized)

1078 8.6 (0.3) 630 8.4 (0.4) 693 9.6 (0.4) 0.98 (0.89-1.08) −0.41 .68 1.09 (0.99-1.20) 1.85 .06

Coronaryrevascularizations

1113 9.2 (0.3) 725 10.0 (0.4) 718 10.2 (0.4) 1.09 (1.00-1.20) 1.88 .06 1.10 (1.00-1.21) 1.95 .05

Peripheral arterialdisease(hospitalized ortreated)

510 4.1 (0.2) 265 3.7 (0.2) 311 4.7 (0.4) 0.87 (0.75-1.01) −1.86 .06 1.04 (0.90-1.19) 0.48 .63

*RR indicates relative risk; CI, confidence interval; CHD, coronary heart disease; and CVD, cardiovascular disease. CHD includes nonfatal myocardial infarction (MI) and fatal CHD;end-stage renal disease: kidney disease death, kidney transplant, or start of chronic renal dialysis; and heart failure: fatal, nonfatal hospitalized, or treated.

†Nonfatal MIs comprise 64% to 66% of the primary outcome.‡Combined CHD indicates CHD death, nonfatal MI, coronary revascularization procedures, and hospitalized angina. Combined CVD indicates CHD death, nonfatal MI, stroke, coro-

nary revascularization procedures, hospitalized or treated angina, treated or hospitalized heart failure, and peripheral arterial disease (hospitalized or outpatient revascularization).§Denominators are 11 361 chlorthalidone, 6757 amlodipine, and 6665 lisinopril.�Proportional hazards assumption violated; data are RRs from a 2 � 2 table.




all (4 mm Hg in black patients) be-tween the lisinopril and chlorthali-done groups only partially accounts for

the observed CVD event difference.However, such analyses are limited bythe infrequency and imprecision of BP

measurements for individual partici-pants and regression dilution, whichunderestimates CVD risk associated

Figure 4. Cumulative Event Rates for All-Cause Mortality, Stroke, Combined Coronary Heart Disease, Combined Cardiovascular Disease,Heart Failure, and Hospitalized Plus Fatal Heart Failure by Treatment Group

15

3

6

9

12

0

No. at RiskChlorthalidone

Amlodipine15 255

90489054Lisinopril

1

14 52885358496

3

13 22478017689

4

11 51167856698

5

636937753789

6

301617801837

7

384210313

2

13 89881858096

Time to Event, y

Heart Failure

Cum

ulat

ive

Eve

nt R

ate,

%

15

3

6

9

12

0

15 25590489054

1

14 55685798538

3

13 31278847780

4

11 60668666768

5

644338333895

6

313018211894

7

396216322

2

13 96282518169

Time to Event, y

Hospitalized Plus Fatal Heart Failure

Cum

ulat

ive

Eve

nt R

ate,

%



50

10

20

30

40

0 1 3 4 5 6 72Time to Event, y

Combined Cardiovascular DiseaseC

umul

ativ

e C

ombi

ned

Eve

nt R

ate,

%

30

10

5

15

20

25

0



1 3 4 5 6 72Time to Event, y

All-Cause Mortality

Cum

ulat

ive

Eve

nt R

ate,

%

Chlorthalidone


15 25590489054

14 51586178543

13 30979497784

11 57069376765

638538453891

321718131828

567506949

13 93482718172

10

2

4

6

8


Stroke

Cum

ulat

ive

Eve

nt R

ate,

%

15 25590489054

14 21184288347

12 41574227264

10 58763576207

580435123461

269816721622

187142170

13 32079407789

30

10

20


Combined Coronary Heart Disease

Cum

ulat

ive

Com

bine

d E

vent

Rat

e, %

15 25590489054

14 93388478853

14 07783918318

12 48074427382

718543124304

352321012121

428217144

14 56486548612

15 25590489054

13 75281187962

11 51768376631

964357245560

516730493011

236214111375

288153139

12 59474517259




with BP differences based on single-visit (or even visit-averaged) measure-ments.45 Such modeling is also unable

to account for differences among indi-viduals due to other unmeasured orpoorly represented risk factors; thus,

participants who lower their BP by agiven amount with one drug may notbe comparable to those who lower their

Figure 5. Relative Risks and 95% Confidence Intervals (CIs) for Amlodipine/Chlorthalidone Comparisons in Prespecified Subgroups

Total

Age <65 y

Age ≥65 y

Men

Women

Black

Nonblack

Diabetic

Nondiabetic

Total

Age <65 y

Age ≥65 y

Men

Women

Black

Nonblack

Diabetic

Nondiabetic

Nonfatal Myocardial InfarctionPlus Coronary Heart Disease Death

Relative Risk(95% CI)

0.98 (0.90-1.07)

0.99 (0.85-1.16)

0.97 (0.88-1.08)

0.98 (0.87-1.09)

0.99 (0.85-1.15)

1.01 (0.86-1.18)

0.97 (0.87-1.08)

0.99 (0.87-1.13)

0.97 (0.86-1.09)

FavorsAmlodipine

FavorsChlorthalidone



1.00 (0.94-1.07)

0.94 (0.84-1.05)

1.04 (0.96-1.12)

0.99 (0.92-1.08)

1.02 (0.91-1.13)

1.03 (0.91-1.17)

0.99 (0.92-1.07)

1.04 (0.94-1.14)

0.97 (0.89-1.06)

0.5 21

0.5 21

Relative Risk

FavorsAmlodipine




All-Cause Mortality

FavorsAmlodipine


0.96 (0.89-1.02)

0.96 (0.83-1.10)

0.96 (0.88-1.03)

0.95 (0.87-1.04)

0.96 (0.86-1.07)

0.97 (0.87-1.09)

0.94 (0.87-1.03)

0.96 (0.87-1.07)

0.95 (0.87-1.04)

Combined Cardiovascular Disease

0.5 21

0.5 21

Relative Risk

FavorsAmlodipine


1.04 (0.99-1.09)

1.03 (0.94-1.12)

1.05 (0.99-1.12)

1.04 (0.98-1.11)

1.04 (0.96-1.13)

1.06 (0.96-1.16)

1.04 (0.97-1.10)

1.06 (0.98-1.15)

1.02 (0.96-1.09)



Stroke

FavorsAmlodipine


0.93 (0.82-1.06)

0.93 (0.73-1.19)

0.93 (0.81-1.08)

1.00 (0.85-1.18)

0.84 (0.69-1.03)

0.93 (0.76-1.14)

0.93 (0.79-1.10)

0.90 (0.75-1.08)

0.96 (0.81-1.14)

Heart Failure

0.5 21

0.5 21

Relative Risk

FavorsAmlodipine


1.38 (1.25-1.52)

1.51 (1.25-1.82)

1.33 (1.18-1.49)

1.41 (1.24-1.61)

1.33 (1.14-1.55)

1.47 (1.24-1.74)

1.33 (1.18-1.51)

1.42 (1.23-1.64)

1.33 (1.16-1.52)

Scales are shown in natural logarithm.

Table 6. Causes of Death by Antihypertensive Treatment Group*

No. (6-Year Rate per 100 Persons) P Value






Total deaths 2187 (17.1) 1237 (16.5) 1303 (17.0) .12 .90

Cardiovascular 992 (8.0) 592 (8.4) 609 (8.4) .98 .53

Myocardial infarction 298 (2.4) 168 (2.3) 157 (2.2) .56 .22

Definite CHD 118 (1.1) 74 (1.2) 78 (1.1) .73 .47

Possible CHD 123 (1.1) 69 (1.1) 93 (1.4) .68 .08

Stroke 163 (1.4) 91 (1.4) 116 (1.6) .62 .14

Heart failure 116 (1.1) 79 (1.3) 68 (1.1) .36 .92

Other CVD 174 (1.4) 111 (1.7) 97 (1.5) .58 .62

Noncardiovascular 1058 (8.9) 559 (7.8) 606 (8.3) .02 .47

Cancer 513 (4.3) 280 (3.7) 297 (4.0) .23 .72

Kidney disease 36 (0.4) 23 (0.5) 28 (0.5) .80 .29

Unintentional injury/suicide/homicide 65 (0.6) 18 (0.3) 27 (0.4) .004 .12

Other non−CVD 444 (3.9) 238 (3.6) 254 (3.7) .18 .62

Unknown 137 (1.2) 86 (1.2) 88 (1.3) .72 .58

*CHD indicates coronary heart disease; CVD, cardiovascular disease.




BP by the same magnitude with an-other drug.

Using an external standard of pooledresults of long-term hypertension treat-ment trials and observational studies(10-12 mm Hg systolic BP differenceassociated with 38% stroke reduction),a 2- to 3-mm Hg difference in BP mightaccount for a 6% to 12% difference instroke rates.45,46 This is consistent withthe observed 15% difference for strokeoverall but not with the difference seenin black patients (13%-16% expected,40% observed). For the HF outcome,trial results in isolated systolic hyper-tension suggest that a 3-mm Hg highersystolic BP could explain a 10% to 20%increase in risk.8,47 The forgoing ignoresthe absence of a diastolic BP differencein ALLHAT; however, the relationshipof diastolic pressure and CVD events inelderlypersonswhooftenhave increasedpulse pressure is not entirely clear.48

The primary and secondary out-come results for the amlodipine vschlorthalidone group comparisons wereconsistent for all subgroups of partici-pants: older and younger, men andwomen, black and nonblack, diabeticand nondiabetic. For the lisinopril vschlorthalidone comparisons, resultswere generally consistent by age, sex,and diabetic status. Thus, for the im-portant diabetic population, lisinoprilappeared to have no special advantage(and amlodipine no particular detri-mental effect) for most CVD and renaloutcomes when compared withchlorthalidone. In fact, chlorthali-done was superior to lisinopril for sev-eral CVD outcomes and superior to am-lodipine for HF in both diabetic andnondiabetic participants. The consis-tency of the ALLHAT findings acrossmultiple patient subgroups providesconfidence in the ability to generalize

the findings to most patients with hy-pertension.

In the lisinopril vs chlorthalidonecomparisons, there were 2 outcomeswith significant interactions. Thegreater differences observed in black vsnonblack patients for combined CVDand stroke, along with a similar trendfor HF and lesser BP lowering with li-sinopril, are in accord with the mul-tiple reports of poorer BP response withACE inhibitor in black patients.49-51

They are also consistent with reportsof lesser effects of ACE inhibitors in sec-ondary prevention of HF in this popu-lation,52,53 although these findings havebeen recently questioned.54 The differ-ential responses for disease outcomesparallel the lesser response in the blacksubgroup for BP, although the differ-ences in outcomes are not substan-tially reduced by statistically adjust-ing for systolic BP.

Figure 6. Relative Risks and 95% Confidence Intervals (CIs) for Lisinopril/Chlorthalidone Comparisons in Prespecified Subgroups

Nonfatal Myocardial InfarctionPlus Coronary Heart Disease Death

Total

Men

Black



0.99 (0.91-1.08)

0.95 (0.81-1.12)

1.01 (0.91-1.12)

0.94 (0.85-1.05)

1.06 (0.92-1.23)

1.10 (0.94-1.28)

0.94 (0.85-1.05)

1.00 (0.87-1.14)

0.99 (0.88-1.11)

FavorsLisinopril



Total 1.05 (0.98-1.11)

0.94 (0.84-1.05)

1.11 (1.03-1.20)

Men 1.03 (0.95-1.12)

1.05 (0.95-1.17)

Black 1.15 (1.02-1.30)

1.01 (0.93-1.09)

1.03 (0.93-1.14)

Age <65 y

Age ≥65 y

Women

Nonblack

Diabetic

Nondiabetic

Age <65 y

Age ≥65 y

Women

Nonblack

Diabetic

Nondiabetic 1.06 (0.97-1.15)

0.5 21

0.5 21

Relative Risk

FavorsLisinopril



All-Cause Mortality


1.00 (0.94-1.08)

0.93 (0.81-1.08)

1.03 (0.95-1.12)

0.99 (0.91-1.08)

1.02 (0.91-1.13)

1.06 (0.95-1.18)

0.97 (0.89-1.06)

1.02 (0.91-1.13)

1.00 (0.91-1.09)

FavorsLisinopril


Combined Cardiovascular Disease

1.10 (1.05-1.16)

1.05 (0.97-1.15)

1.13 (1.06-1.20)

1.08 (1.02-1.15)

1.12 (1.03-1.21)

1.19 (1.09-1.30)

1.06 (1.00-1.13)

1.08 (1.00-1.17)

1.12 (1.05-1.19)

0.5 21

0.5 21

Relative Risk

FavorsLisinopril




Stroke

FavorsLisinopril


1.15 (1.02-1.30)

1.21 (0.97-1.52)

1.13 (0.98-1.30)

1.10 (0.94-1.29)

1.22 (1.01-1.46)

1.40 (1.17-1.68)

1.00 (0.85-1.17)

1.07 (0.90-1.28)

1.23 (1.05-1.44)

Heart Failure

0.5 21

0.5 21

Relative Risk

FavorsLisinopril


1.20 (1.09-1.34)

1.23 (1.01-1.50)

1.20 (1.06-1.35)

1.19 (1.03-1.36)

1.23 (1.05-1.43)

1.32 (1.11-1.58)

1.15 (1.01-1.30)

1.22 (1.05-1.42)

1.20 (1.04-1.38)

Scales are shown in natural logarithm.




Although subordinate to safety andefficacy, the cost of drugs and medicalcare for the individual and society is afactor that should be considered in theselection of antihypertensives. One ofthe stated objectives of ALLHAT was toanswer the question, “Are newer typesof antihypertensive agents, which arecurrently more costly, as good or bet-ter than diuretics in reducing CHD in-cidence and progression?”18 Consider-ation of drug cost could have a majorimpact on the nation’s health care ex-penditures. Based on previous data thatshowed that diuretic use declined from56% to 27% of antihypertensive pre-scriptions between 1982 and 1992, thehealth care system would have saved$3.1 billion in estimated cost of anti-hypertensive drugs had the pattern ofprescriptions for treatment of hyper-tension remained at the 1982 level.55

Further economic analyses based on theresults of ALLHAT are under way.

The strengths of ALLHAT include itsrandomized double-blind design, sta-tistical power to detect clinically mean-ingful differences in CVD outcomes ofinterest, diverse population with ad-equate representation from subgroupsof special interest in the treatment ofhypertension, and varied practice-based settings. In addition, the agentsthat were directly compared represent3 of the most commonly used newerclasses of antihypertensives vs the beststudied of the older classes.

Some limitations are worth noting.After ALLHAT was designed, neweragents have been or may soon be re-leased (eg, angiotensin-receptor block-ers, selective aldosterone antago-nists), which were not evaluated.Although clinical centers were blindedto the regimen and urged to achieve rec-ommended BP goals, equivalent BP re-duction was not fully achieved in thetreatment groups. Furthermore, be-cause diuretics, ACE inhibitors, CCBs,and �-blockers were evaluated in thetrial, the agents available for step-up ledto a somewhat artificial regimen (useof sympatholytics rather than diuret-ics and CCBs) of step-up drugs in theACE inhibitor group. This may have

contributed to the higher BPs in theACE inhibitor group, especially in theblack subgroup. However, mean fol-low-up BPs were well below 140/90mm Hg in all treatment groups. Al-though ALLHAT did not compare a�-blocker to chlorthalidone, previoustrials have suggested equivalence45 oreven inferiority3 for major CVD events.

The ALLHAT results apply directlyto chlorthalidone, amlodipine, and li-sinopril. Combined with evidence fromother trials, we infer that the findingsalso broadly apply to the drug classes(or subclass in the case of the dihy-dropyridine CCBs) that the study drugsrepresent. The evidence base for selec-tion of antihypertensive agents has beenmarkedly strengthened by the addi-tion of ALLHAT.

In conclusion, the results of ALLHATindicate that thiazide-type diureticsshould be considered first for pharma-cologic therapy in patients with hyper-tension. They are unsurpassed in low-ering BP, reducing clinical events, andtolerability, and they are less costly. Forpatients who cannot take a diuretic(which should be an unusual circum-stance), first-step therapy with CCBsand ACE inhibitors could be consid-ered with due regard for their higherrisk of 1 or more major manifestationsof CVD. Since a large proportion of par-ticipants required more than 1 drug tocontrol their BP, it is reasonable to in-fer that a diuretic be included in all mul-tidrug regimens, if possible. Althoughdiuretics already play a key role in mostantihypertensive treatment recommen-dations, the findings of ALLHAT shouldbe carefully evaluated by those respon-sible for clinical guidelines and bewidely applied in patient care.

ALLHAT Authors/Officers and Coordinators: Curt D.Furberg, MD, PhD; Jackson T. Wright, Jr, MD, PhD;Barry R. Davis, MD, PhD; Jeffrey A. Cutler, MD, MPH;Michael Alderman, MD; Henry Black, MD; WilliamCushman, MD; Richard Grimm, MD, PhD; L. JulianHaywood, MD; Frans Leenen, MD; Suzanne Oparil,MD; Jeffrey Probstfield, MD; Paul Whelton, MD, MSc;Chuke Nwachuku, MA, MPH; David Gordon, MD,PhD; Michael Proschan, PhD; Paula Einhorn, MD, MS;Charles E. Ford, PhD; Linda B. Piller, MD, MPH; J. KayDunn, PhD; David Goff, MD, PhD; Sara Pressel, MS;Judy Bettencourt, MPH; Barbara deLeon, BA; Lara M.Simpson, MS; Joe Blanton, MS; Therese Geraci, MSN,RN, CS; Sandra M. Walsh, RN; Christine Nelson, RN,

BSN; Mahboob Rahman, MD; Anne Juratovac, RN;Robert Pospisil, RN; Lillian Carroll, RN; Sheila Sulli-van, BA; Jeanne Russo, BSN; Gail Barone, RN; RudyChristian, MPH; Sharon Feldman, MPH; Tracy Lu-cente, MPH; David Calhoun, MD; Kim Jenkins, MPH;Peggy McDowell, RN; Janice Johnson, BS; Connie Kin-gry, RN, BSN; Juan Alzate, MD; Karen L. Margolis, MD;Leslie Ann Holland-Klemme, BA; Brenda Jaeger; Jef-frey Williamson, MD, MHS; Gail Louis, RN; PamelaRagusa, RN, BSN; Angela Williard, RN, BSN; R. L. SueFerguson, RN; Joanna Tanner; John Eckfeldt, MD, PhD;Richard Crow, MD; John Pelosi, RPh, MS.Financial Disclosures: The following listed authorshave served as consultants for, received personal com-pensation from, were grant recipients of, or own stockin the following companies: Furberg: Merck, Pfizer,Pharmacia & Upjohn, Takeda, Wyeth-Ayerst; Wright:Aventis, Bayer, Bristol-Myers Squibb, Forrest Labs,King/Monarch, Merck, Novartis, Pfizer; Davis: Ab-bott, Bristol-Myers Squibb, Forrest Labs, Merck, Pfizer,Pharmacia & Upjohn, GlaxoSmithKline; Alderman: Bris-tol-Myers Squibb, Merck, Novartis, Pfizer, Glaxo-SmithKline; Black: Abbott, AstraZeneca, Bristol-Myers Squibb, Novartis, Pfizer, Pharmacia & Upjohn,GlaxoSmithKline, Solvay; Cushman: AstraZeneca, Bris-tol-Myers Squibb, Forrest Labs, Merck, Pfizer, Phar-macia & Upjohn, Sankyo, Searle, Solvay, Takeda;Grimm: AstraZeneca, Merck, Novartis, Pfizer, Roche,Solvay; Haywood: Pharmacia & Upjohn; Leenen:AstraZeneca, Bayer, Bristol-Myers Squibb, Merck,Nu-Pharm, Pfizer, Pharmacia & Upjohn; Oparil: Ab-bott, AstraZeneca, Aventis, Bayer, Bristol-Myers Squibb,DuPont, Forrest Labs, King/Monarch, Merck, Novar-tis, Parke-Davis, Pfizer, Pharmacia & Upjohn, Roche,Sankyo, Schering-Plough, Searle, GlaxoSmithKline,Texas Biotechnology; Probstfield: AstraZeneca, King/Monarch, Pfizer; Whelton: Merck, Novartis, Pfizer,Pharmacia & Upjohn; Ford: Bristol-Myers Squibb; Rah-man: Abbott, Novartis, Pfizer, Searle; Lucente: Glaxo-SmithKline; Calhoun: AstraZeneca, Aventis, Merck,Novartis, GlaxoSmithKline; Solvay; McDowell: Am-gen, King/Monarch, Merck, Pfizer; Alzate: Pfizer; Tan-ner: SGP; Eckfeldt: Johnson & Johnson.Author Contributions: Dr Davis had full access to allthe data in the study and takes responsibility for theintegrity of the data and the accuracy of the data analy-ses in this article and its companion article on page2998.Study concept and design: Furberg, Wright, Davis,Cutler, Alderman, Black, Cushman, Grimm, Oparil,Whelton, Proschan, Ford, Piller, Goff, Lucente,Margolis, Williamson, Ragusa.Acquisition of data: Wright, Davis, Alderman, Black,Cushman, Grimm, Haywood, Leenen, Oparil,Probstfield, Whelton, Einhorn, Ford, Piller, Pressel,deLeon, Simpson, Blanton, Geraci, Walsh, Nelson,Rahman, Juratovac, Pospisil, Carroll, Sullivan, Russo,Christian, Feldman, Lucente, Calhoun, Jenkins,McDowell, Johnson, Kingry, Alzate, Margolis, Holland,Jaeger, Williamson, Louis, Ragusa, Williard, Fergu-son, Tanner, Eckfeldt, Crow, Pelosi.Analysis and interpretation of data: Furberg, Wright,Davis, Cutler, Black, Cushman, Grimm, Haywood,Leenen, Oparil, Probstfield, Whelton, Nwachuku,Gordon, Proschan, Einhorn, Ford, Piller, Dunn, Goff,Pressel, Bettencourt, Simpson, Rahman, Barone,Williamson.Drafting of the manuscript: Furberg, Wright, Davis,Cutler, Alderman, Black, Cushman, Grimm, Haywood,Leenan, Oparil, Probstfield, Whelton, Nwachuku,Gordon, Proschan, Einhorn, Ford, Piller, Dunn, Goff,Pressel, Bettencourt, Simpson, Rahman, Kingry,Margolis, Williamson.Critical revision of the manuscript for important in-tellectual content: Furberg, Wright, Davis, Cutler,Alderman, Black, Grimm, Haywood, Leenen, Oparil,Probstfield, Whelton, Nwachuku, Gordon, Proschan,Einhorn, Ford, Piller, Dunn, Goff, Pressel, Bettencourt,




deLeon, Simpson, Geraci, Walsh, Rahman, Pospisil,Carroll, Sullivan, Russo, Barone, Christian, Feldman,Lucente, Calhoun, Jenkins, McDowell, Johnson, Kingry,Alzate, Margolis, Williamson, Louis, Williard, Ferguson,Tanner, Pelosi.Statistical expertise: Davis, Whelton, Proschan, Ford,Dunn, Pressel.Obtained funding: Davis, Cutler, Black, Einhorn, Ford,Goff, Sullivan.Administrative, technical, or material support: Furberg,Wright, Davis, Cutler, Alderman, Black, Cushman,Grimm, Haywood, Oparil, Probstfield, Whelton,Nwachuku, Gordon, Einhorn, Ford, Piller, Pressel,Bettencourt, deLeon, Simpson, Blanton, Geraci, Walsh,Nelson, Rahman, Juratovac, Pospisil, Carroll, Russo,Barone, Christian, Feldman, Lucente, Jenkins,McDowell, Johnson, Kingry, Alzate, Margolis, Holland,Jaeger, Louis, Williard, Ferguson, Tanner, Eckfeldt,Pelosi.Study supervision: Furberg, Wright, Davis, Cutler,Black, Cushman, Grimm, Haywood, Leenen, Oparil,Probstfield, Ford, Pressel, Lucente, Alzate, Holland,Jaeger, Eckfeldt.Funding/Support: This study was supported by con-tract NO1-HC-35130 with the National Heart, Lung,and Blood Institute (NHLBI). ALLHAT investigators re-ceived contributions of study medications supplied byPfizer (amlodipine and doxazosin), AstraZeneca (aten-olol and lisinopril), and Bristol-Myers Squibb (prava-statin), and financial support provided by Pfizer.Role of the Sponsor: The NHLBI sponsored the studyand was involved in all aspects other than direct op-erations of the study centers. This included collec-tion, analysis, and interpretation of the data in addi-tion to the decision to submit the manuscript forpublication.Dedication: Special recognition is due to 3 ALLHATleaders who died in recent years after making verysignificant contributions to initiating the trial and over-seeing most of its course: Richard Carleton, MD, Chair-man of the Data and Safety Monitoring Board (1994-2000); H. Mitchell Perry, Jr, MD, member of theSteering Committee and Deputy Physician Coordina-tor for Region 1 (1994-2001); and Peter Frommer, MD,National Heart, Lung, and Blood Institute DeputyDirector Emeritus, advisor to the Project Officers, andliaison to participating pharmaceutical companies(1993-2002).Members of the ALLHAT Group: Steering Commit-tee: Furberg, Wright, Davis, Cutler, Alderman, Black,Cushman, Grimm, Haywood, Leenen, Oparil, Probst-field, Whelton; NHLBI Project Office: Cutler, Nwa-chuku, Gordon, Proschan, Einhorn; ALLHAT ClinicalTrials Center: Davis, Ford, Piller, Dunn, Pressel, Betten-court, deLeon, Simpson, Blanton; ALLHAT Regions:Veterans Administration, Memphis, Tenn: Cush-man, Geraci, Walsh, Nelson; Cleveland, Ohio: Wright,Rahman, Juratovac, Pospisil, Suhan; Bronx, NY: Al-derman, Carroll, Russo, Sullivan; Chicago, Ill: Black,Barone, Christian, Feldman, Lucente; Birmingham, Ala:Oparil, Calhoun, Jenkins, McDowell; Seattle, Wash:Probstfield; Alzate, Johnson, Kingry; Minneapolis,Minn: Grimm, Margolis, Holland, Jaeger; New Or-leans, La (formerly located in Baltimore, Md): Whelton,Williamson, Louis, Ragusa, Williard, Adler; Ottawa, On-tario, Canada: Leenen, Ferguson, Tanner; ALLHATCentral Laboratory: J. Eckfeldt, J. Bucksa, M. Now-icki; ALLHAT Drug Distribution Center: J. Pelosi;ALLHAT Electrocardiogram Reading Center: R. Crow,S. Thomas; ALLHAT Data and Safety MonitoringBoard: R. Califf, W. Applegate, J. Buring, E. Cooper,K. Ferdinand, M. Fisher, R. Gifford, S. Sheps.Investigators and Coordinators Participating in the An-tihypertensive and Lipid Trials, United States: Ala-bama: L. Ada, D. Alexander, L. Black, C. Davis, W.Davis, S. Farooqui, H. Fritz, T. Kessler, S. Ledbetter,L. Means, J. Patterson, N. Qureshi, L. Redcross, R.Reeves, T. Tucker, N. Wettermark, A. Williams, W. Yar-

brough; Arizona: I. Cohen, W. Dachman, N. Estrada,J. Felicetta, D. Fowler, R. Fowler, S. Goldman, C. Lui,S. Morris, D. Morrison, J. Nelson, J. Ohm, D. Paull,G. Pulliam, D. Roberts, I. Ruiz, H. Thai; Arkansas: J.Acklin, M. Azhar, F. Berry, D. Burns, W. Carter, M.Dixon, S. Eldridge, A. Fendley, H. Fendley, M. Flow-ers, S. Goss, M. Guyer, G. Harris, M. Hawkins, D. Hop-son, P. Kern, R. King, M. Lynch, E. Maples, R.McCafferty, M. McGehee, J. Miller, D. Neil, M.Oakum, N. Paslidis, K. Riordan, G. Robbins, D. Sim-mons, C. Vilayvanh, S. Whitmer; California: C. Al-varez, D. Anderson, M. Ariani, S. Barrett, J. Boggess,B. Brackeen, A. Bui, P. Callaham, M. Calong, J. Ca-macho, J. Cavendish, G. Chao, D. Cheung, B. Chris-tianson, W. Dempsey, G. Dennish, V. DeQuattro, R.Dharawat, D. Dizmang, N. Doherty, M. Donnell, S.Edmondson, D. Falcone, S. Franklin, J. Frazee, G. Friv-old, S. Ghattas, D. Goldfarb-Waysman, T. Haskett, L.Haywood, N. Horton, Y. Huang, K. Hui, N. Jacob, K.Jolley, B. Jurado, A. Karns, R. Karns, K. Karunaratne,A. Katchem, L. Katchem, J. Khoo, E. Kiger, L. Klein-man, J. Kozlowski, D. Kramer, E. Lee, D. Li, C. Liba-nati, P. Linz, D. Lyle, T. Maekawa, M. Mahig, J. Mal-lery, D. Martins, B. Massie, R. Mikelionis, S. Myers, J.Neutel, N. Nguyen, U. Okoronkwo, K. Owens, T. Pan,R. Petersen, A. Schultz, H. Schultz, E. Schwartz, J.Schwartz, P. Schwartz, C. Scott, Z. Song, J. Taylor, D.Townsend, S. Turitzin, D. Ujiiye, A. Usman, D. VanOstaeyen, R. Wadlington, C. Wan, L. Wang, H. Ward,L. Wieland, P. Williams-Brown, N. Wong, R. Wright;Colorado: K. Castleman, M. Chase, R. Hildenbrand,P. Lowe, P. Mehler, S. Mroz, R. Simpson, R. Tello; Con-necticut: J. Bernene, L. Ciarcia, A. Grover, J. Judge,A. Lachman, J. Lawson, N. Medina, E. Nestler, R.Schwartz, B. Sicignano, S. Solinsky; Washington, DC:J. Golden, E. Lewis, D. Mateski, P. Narayan, A. No-targiacomo, D. Ordor, V. Papademetriou, O. Ran-dall, T. Retta, J. Theobalds, S. Xu; Delaware: D. Crane,J. Lenhard; Florida: K. Anderson, S. Beery, G. Bhaskar,B. Booker, K. Broderick, E. Capili-Rosenkranz, J. Cio-con, G. Cohn, T. Connelly, V. Dallas, G. Duren, J. Durr,J. Evans, S. Feld, R. Feldman, L. Fischer, S. Fisher, M.Formoso, S. Fulford, M. Galler, J. Hildner, K. Hol-man, A. Jackson, C. Jackson, G. Khan, M. Khan, S.Kronen, J. Lehmann, A. Littles, R. Lopez, N. Madhany,L. McCarty, K. Mullinax, M. Murray, J. Navas, A.Peguero-Rivera, R. Preston, N. Rolbiecki, J. Rolle, L.Rosenfield, O. Saavedra, A. Schlau, M. Stein, J. Stokes,S. Strickland, U. Tran, B. Videau, J. Webster, T. Web-ster, A. Weinstein, T. Westfall, D. Williams, M. Yo-ham; Georgia: D. Anderson, R. Anderson, J. Barzilay,S. Boyce, P. Brackett, P. Bradley, W. Brown, R. Carter,S. Carter, D. Castro, L. Duty, H. Ellison, A. Francis, L.Goodman, D. Harrelson, T. Hartney, J. Heldreth, J. He-neisen, A. Hicks, L. Hornsby, J. Hudson, S. Hurst, L.Iskhakova; S. James, S. James, Y. Jones, K. Kersey, W.Kitchens, N. London, M. Loraditch, G. Lowe, R. Mad-dox, R. Malcolm, D. Mathis, C. Mayers, M. McDaniel,N. McPhail, A. Mikhail, H. Muecke, R. Noel, W. North,N. Parikh, D. Parish, G. Peters, P. Poulos, M. Ram, W.Rawlings, R. Remler, C. Rice, M. Salles, D. Sauers, A.Scheetz, C. Scott, L. Stevenson, J. Sumner, M. Swee-ney, E. Taylor, K. Upadhya, T. Vu, M. Walsh, K. Wil-liams, H. Yager; Illinois: M. Arron, C. Bareis, J. Bar-nett, G. Barone, C. Bermele, T. Bertucci, J. Cheng, J.Cruz, T. Denecke-Dattalo, S. Durfee, E. Edwards, L.Fahrner, D. Farley, T. Flegel, M. Friedman, C. Gaca,J. Gilden, S. Goldman, J. Graumlich, A. Hoffman, K.Hunt, C. Johnson, P. Kellums, A. Lasala, N. Lasala, V.Lauderdale, M. Lesko, F. Lopez, M. Mansuri, S. Man-suri, M. Martin, L. Moody, L. Morowczyneski, S. Mou-ritzen, N. Novotny, A. Ovalle, P. Pedersen, N. Perlman,P. Porcelli, B. Ragona, R. Sadiq, P. Sands, C. Sim-mons, K. Stevens, G. Sussman, D. Vicencio, A. Villa-fria, R. Villafria, R. Watkins; Indiana: J. Addo, J.Beliles, V. Dave, D. Fausset, J. Fox, D. Fryman, J. Hall,J. Koehler, L. Leavy, P. Linden, E. Long, H. Ma-cabalitaw, T. Nguyen, B. Peterson, J. Pratt, D.

Rosanwo, D. Ross, H. Shah, V. Shah, T. Smith, M. So-bol, B. Viellieu-Fischer, J. Wachs, B. Weinberg; Iowa:V. Butler, A. Durbin, R. Glynn, B. Hargens, W. Law-ton, M. Roberts, J. Roepke, R. Schneider, G. Stanley;Idaho: M. Baker, R. Force, T. Gillespie, S. Hillman, K.Krell, M. Macdonald; Kansas: D. Courtney, B. Craw-ford, D. DeVore, J. Moppin, N. Premsingh, K. Reuben-Hallock, R. Schanker, D. Wilson; Kentucky: R. Berk-ley, M. DeMuro, L. Kazmierzak, A. Rayner, C. Tyler,E. Wells, S. Winters; Louisiana: E. Aguilar, L. Bass, V.Batuman, B. Beard, L. Borrouso, M. Campbell, C.Chubb, P. Connor, C. Conravey, D. Doucet, M. Doucet,J. Dunnick, D. Eldridge, T. Eldridge, P. Galvan, A. Gupta,J. Hollman, D. Hull, B. Jackson, T. Jones, A. Klenk, P.Lakshmiprasad, B. Mahl, J. Paranilam, E. Reisin, H.Rothschild, J. Sampson, B. Samuels, J. Schmitt, A. Smith,V. Valentino, C. Verrett, P. Willhoit; Maine: B. Blake,T. Lebrun, C. Walworth, R. Weiss; Maryland: J. Bur-ton, W. Carr, P. Chance, S. Childs, C. Compton, J.Cook, V. Coombs, J. Daniels, P. Death, L. Essandoh,Y. Ferguson, D. Fraley, M. Freedman, M. Gary, F.Gloth, S. Gottlieb, M. Gregory, S. Hairston, P. Hall,B. Hamilton, J. Hamilton, D. Harrison, D. James, B. Ker-zner, A. Lancaster, H. Lutz, J. Marks, J. Martin, J. Mer-sey, L. Nelson, E. Obah, S. Ong, J. Palacios, S. Park,M. Partlow, M. Posner, H. Rachocka, M. Rubin, M.Rubinstein, M. Rykiel, C. Smith, B. Socha, K. Thomp-son, K. Walker, J. Webber, K. Williams; Massachu-setts: L. Bradshaw, A. Chakraborty, F. DiMario, J.Ingelfinger, J. Pincus, A. Sobrado; Michigan: L. Bey-Knight, D. Carson, A. Cavanaugh, M. Chertok, K.Church, H. Colfer, I. Diaz, B. Dobbs, G. Edelson, J.Fabello-Gamiao, S. Gappy, J. Grove, D. Johnson, M.Johnson, C. Jones, E. Jones, T. Kelly, N. Kerin, B.Letzring, M. Oleszkowicz, A. Raffee, K. Rasikas, C.Shaw, M. Siddique, B. VanOver, M. Zervos; Minne-sota: D. Berman, V. Canzanello, J. Curtis, V. Erick-son, W. Goodall, J. Graves, K. Guthrie, J. Haight, S.Hassing, J. Heegard, J. Holtzman, D. Jespersen, L. Klein,C. Kubajak, L. Nylund, P. Spilseth; Missouri: B. Apple-ton, R. Baird, S. Carmody, C. Carter, F. Charles, T. Finni-gan, S. Giddings, K. Gorman, M. Gregory, L. Johnson,S. Joseph, L. Kennington, R. Kevorkian, J. LaSalle, B.Nolfo, J. Nunnelee, A. Orf, D. Palmer, H. Perry, A.Quick, B. Rogers, B. Rosemergey, C. Scott, S. Sharma,V. Shortino, D. Smith, K. Smith, C. Stanford, C. Tu-dor, T. Wiegmann; Mississippi: C. Adair, S. Arm-strong, C. Brown, N. Brown, R. Brown, S. Burke, L.Burrell, L. Clark, S. Cooks, W. Crowell, D. Ellis, D. Gra-ham, V. Green, R. Hall, S. Hamler, D. Haymon, A. Hin-ton, M. Holman, A. James, P. Karim, K. Kirchner, A.Knotts, A. Lott, W. McArthur, F. McCune, B. Miller,H. Morrow, R. Murphy, R. Myers, S. Myers, A. Phil-lips, M. Puckett, E. Rankin, O. Ransome-Kuti, M. Red-dix, R. Rigsby, E. Searcy, D. Smith, A. Spann, Y. Tan-ner, E. Taylor-McCune, J. Tramuta, H. Wheeler, M.Wofford; Montana: L. Bigwood-Pecarina, S. English,H. Knapp, L. Sokoloski; Nebraska: M. Berry, E. But-kus, S. Byers, D. Colan, R. Dobesh, N. Hilleman, R.Hranac, P. Klein, T. McKnight, S. Mohiuddin, A. Mooss,R. Moyer, P. Myers, L. Rasmussen, J. Schafersman;Nevada: J. Chinn, R. Collins, E. Samols; New Jersey:S. Akgun, A. Bastian, L. Bordone, N. Cosgrove, A.Costa, A. Cuyjet, S. Daniels, L. DeEugenio, L.DeEugenio, R. Denniston, L. Duh, M. Farber, M. Far-ber, S. Ferguson, K. Ferranti, G. Flanagan, J. Garo-falo, H. Hassman, J. Hassman, H. Jacobs, J. Kostis, A.Kudryk, M. Kutza, R. Liang, G. McArthur, B. McGann,R. Miller, E. Moser, F. Nash, P. Niblack, E. Ogunme-fun, M. Raghuwanshi, S. Sastrasinh, T. Seely, J. Stan-ley, S. Suarez, A. Vaughn, R. Wong-Liang, J. Young,S. Yuchnovitz, M. Zolnowski; New Mexico: D. Graves,M. Groves, E. Iwan, J. Shipley; New York: N. Almelda,S. Anderson, J. Andres, N. Ankomah, E. Anteola, C.Assadi, M. Assadi, S. Atlas, J. Baruth, D. Barz, J. Begley,T. Bharathan, A. Bova, D. Brautigam, C. Brown, S.Canaan, M. Candelas, P. Caraballo, J. Chapman, L.Clark, K. Desai, D. Dowie, C. Dwyer, A. Farag, C.




Flanders, P. Foster, L. Gage, A. Gartung, S. Gedan, P.Gehring, J. Gorkin, D. Graber, H. Guber, P. Gugli-uzza, J. Halbach, A. Henriquez, M. Henriquez, D. Hoff-man, J. Holland, C. Hopkins, C. Hull, E. Ilamathi, K.Johnston, M. Karim, L. Katz, K. Kellick, S. Kerlen, M.Krishnamurthy, D. Lainoff, R. Levin, V. Littauer, J. Lohr,M. Lorenz, C. Lynott, J. Maddi, L. Marquart, K. Mar-tin, M. Maw, R. Mendelson, S. Monrad, A. Musta-pha, A. Nafziger, M. Neary, J. Ngheim, A. Niarchos,M. Noor, M. Omoh, J. Pickard, M. Pier, V. Pogue, C.Reddy, J. Ringstad, T. Rocco, C. Rosendorff, H. San-defur, A. Sass, R. Schifeling, D. Scott, P. Scriber, K.Sharma, C. Shmukler, D. Shrivastava, M. Siegelheim,G. Smith, B. Snyder, C. Spiller, M. Srivastava, S. Steven-son, A. Stewart, B. Sumner, M. Sweeney, K. Thomas,L. Thomas, L. Trawlick, N. Velez, J. Vento, H. Vis-waswariah, M. Yevdayeva, D. Zimmerman; NorthCarolina: T. Barringer, V. Bland, M. Burke-Ziglar, K.Caldwell, R. Caldwell, F. Celestino, G. Cole, M. Dar-row, B. Dunn, S. Fox, J. Holbrook, K. Jacobs, J. Li-sane, L. Loggans, A. Lowdermilk, R. Merrill, P. Miller,C. Perkins, L. Rodebaugh, V. Schlau, R. Smith, J. Spruill,J. Summerson; North Dakota: N. Chelliah, E. Garten,K. Hagen, S. Jafri, D. Vold, B. Westacott; Ohio: L.Barnes-Lark, C. Blanck, K. Casterline, D. Chen, K.Cowens, M. Cubick, D. Davidson, P. Dockery, J. Finoc-chio, T. Gundrum, T. Hentenaar, D. Hulisz, D. Hull,K. Keaton, G. Kikano, K. Klyn, L. Lazaron, D. Lukie,S. Medwid, L. Miller, R. Murden, H. Neff, E. Ospelt,M. Patel, E. Pelecanos, E. Pfister, L. Sadler, M. Sak-layen, A. Salomon, A. Schmidt, S. Stein, D. Subich, D.Thiel, L. Thompson, R. Toltzis, J. Tucker, D. Vidt, G.Wise, D. Wray; Oklahoma: D. Abott, J. Cook-Greenwood, M. Jelley, R. Kipperman, J. Leverett, C.Manion, S. Mears, B. Parker, R. Ringrose, L. Scholl, J.Schoshke, F. Shelton, M. Stephens, U. Thadani, K. Wal-ters; Oregon: M. Dissanayake, S. Falley, H. Harris, S.MacKenzie, F. McBarron, S. Murray; Pennsylvania: G.Abbott, C. Baessler, M. Benioff, A. Bowens, J. Burke,L. Carradine, K. Devine, M. Duzy, G. Dy, J. Fontaine,D. Fox, W. Gilhool, J. Grasso, T. Ham, S. Heaney, J.Hefner, D. Herr, L. Hollywood, L. Jones, M. Kauff-man, E. Kemler, S. Koduri, N. Kopyt, S. Kutalek, M.MacIntyre, R. Martsolf, A. McLeod, A. Miller, A. Min-nock, Y. Mishriki, D. Nace, L. Nagy, R. Olasin, C. Os-chwald, N. Potts, R. Reinhard, R. Reinhard, N. Rob-erts, B. Rogers, D. Sant Ram, F. Sessoms, M. Shore,S. Shore, D. Singley, J. Spencer, D. Spigner, B. Springer,W. Swagler, P. Tanzer, S. Walker, N. Walls, D. Whyte,S. Worley, G. Ziady; Puerto Rico: A. Agosto, J. Aguilera-Montalvo, H. Algarin-Sanchez, J. Alvarado, I. An-dino, J. Aponte Pagan, M. Arce, J. Benabe, J. Cangi-ano, L. Catoni, J. Cianchini, J. Claudio, M. Collazo, P.Colon, Y. Cruz-Lugo, J. DaMore, E. Edwards Volquez,A. Feliberti-Irizarri, P. Felix-Ramos, J. Fernandez-Quintero, M. Geo, M. Gomez, R. Gomez Adrover, L.Gonzalez-Bermudez, M. Guerrero, E. Guzman, J. He-redia, C. Irizarry, A. Leon, T. Lugardo, G. Martinez,R. Martinez, M. Melendez, M. Natal, M. Padilla, W.Pagan, Z. Perez, J. Pimentel, M. Pimentel Lebron, A.Ramos, M. Rios, C. Rivera, E. Rivera, J. Rivera San-tiago, E. Rodriquez, D. Romero, R. Ruiz, C. Sanchez,J. Sanchez, M. Sosa-Padilla, I. Sotomayor-Gonzalez,J. Tavarez, I. Toro-Grajales, B. Torres, N. Vazquez, S.Vazquez, M. Vega, Z. Vidal Oviedo, V. Zapata, I. Zayas-Toro; Rhode Island: C. Alteri, J. Galli, A. Hordes, L.Laflamme, K. MacLean, L. Marquis, R. Ruggieri, S.Sharma; South Carolina: J. Basile, L. Clarke, I. Coley,D. Devlin, S. Eggleston, G. Goforth, D. Ham, A. Hamp-ton, P. Hill, K. Jones, R. Jones, P. Jumper, A. Kitch-ens, C. Lieberman, J. McAlpine, J. Moloo, A. Saenz,D. Sheek, A. Smith-Salley, P. Snape, J. Sterrett, C. Stone,M. Strossner, C. Sullivan, T. Vear, D. Weathers, M.Weeks, J. Williams, M. Williams; South Dakota: C.Ageton, M. Brown, L. Dale, L. Duncan, S. Eckrich, P.Kearns, B. Lankhorst, K. McDougall, V. Schuster, J. We-genke, J. Woehl, E. Zawada; Tennessee: D. Ander-son, C. Bounds, J. Caldwell, W. Cannon, R. Cassidy,

W. Cushman, C. DeJesus, L. Dilworth, S. Duffy, B.Hamilton, T. Harrell, K. Harris, M. Herr, J. Jones, L.Jones, H. Marker, J. Miller, S. Miller, F. Putman, A.Reaves, V. Rhule, H. Ross-Clunis, S. Satterfield, G.Siami, R. Smith, A. Smuckler, C. Snorton, T. Stern, D.Venugopal; Texas: A. Abbas, H. Adrogue, A. Ama-dor, L. Arango, C. Arroyo, V. Battles, M. Beard, J. Bea-sley, R. Bhalla, G. Chauca, P. Damico, S. Davison, P.Dlabal, N. Duronio, C. East, F. Eelani, C. Farmerie, E.Fowler, O. Gambini, E. Griego, G. Habib, S. Hanna,D. Harden, T. Harrington, C. Herrera, T. Hicks, B.Hiltscher, D. Hyman, I. Lalani, A. Levine, S. Lu, I. Mar-tinez, Y. Martinez, N. Mata, R. Motaparthi, B. Norch,M. Ottosen, V. Pavlik, L. Pearce, J. Periman, M. Pick-ard, N. Pokala, A. Ray, D. Richard, K. Rogers, M.Ruggles, L. Seals, D. Shafer, T. Shamsi, D. Sherwood-Berner, E. Soltero, A. Sy, J. Tomlinson, C. Vallbona,D. Verrett, R. Victor, W. Vongpatanasin, R. Young;Utah: R. Callihan, G. Henderson, J. O’Donnell, C. Slot,J. Swauger, C. Westenfelder, C. Williams; Vermont:B. Armstrong, B. Buckley, P. Courchesne, P. Cush-man, F. Gallant, T. Howard, J. Osborne, R. Primeau,T. Tanner; Virgin Islands: K. Bryan-Christian, C. Chris-tian, M. Morris; Virginia: D. Bryan, D. Connito, K.Damico, L. Gendron, E. Goudreau, M. Juarez, R. Lemly,L. Macklin, K. McCall, J. Moore, D. Panebianco, D.Paulson, A. Pemberton, R. Renzi, D. Rice, J. Schmitt,S. Speese, J. Sperling, L. Thompson, G. Vetrovec, A.Williams, D. Williams, B. Zambrana; Washington: J.Anderson, K. Capoccia, G. Deger, A. Ellsworth, A. Mick-etti, W. Neighbor, S. Yarnall; West Virginia: H. Black-wood, S. Grubb; Wisconsin: P. Ackell, A. Arnold, S.Blumenthal, P. Bodmer, R. Dart, D. David, D. Duffy,L. Egbujiobi, M. Faignant, A. Friedman, B. Friedman,C. Koeppl, M. Lintereur, J. Morledge, D. Neu, M.Noble, M. Rassier, G. Shove, M. Stevens, R. Wergin,L. Wollet, B. Yug, C. Zyniecki; Investigators and Co-ordinators, Canada: New Brunswick: C. Baer, J.LeBlanc, R. Withers, J. Yang; Newfoundland: J. Col-lingwood, P. Crocker, F. Jardine, S. Newman, G. Ride-out, B. Sussex; Ontario: J. Baker, D. Bishop, C. Brose,D. Carswell, L. Charles, D. Coates, E. Coletta, M. Court-land, S. Crocker, R. Dhaliwal, T. Doey, D. Guy, D. Har-terre, G. Harterre, C. Henry, D. Henry, D. Hutton, I.Janzen, H. Kafka, W. Kendrick, N. Kumar, R. Lan, F.Leenen, R. Lovell, B. McAuley, B. Melbourne, S. Mel-bourne, H. Morwood, S. Munro, S. Nawaz, T.O’Callahan, S. Prasad, P. Richardson, R. Rose, C. San-derson-Guy, N. Schmidt, D. Spink, P. Spink, A. Stajfer,R. Tee, K. Usher, M. Wahby, R. Wahby, D. Wattam,L. Wells, M. Wiebe, K. Zarnke, P. Zuliani; Prince Ed-ward Island: D. Cameron.Investigators and Coordinators Participating in the An-tihypertensive Trial Only, United States: California:P. Bailey-Walton, N. Bednarski, M. Chen, S. Fochler,S. Gross, T. Harper, G. Hilliard, B. Holmes, E. Jacob-son, P. Kirkland, N. Lepor, K. Moorehead, E. Port-noy, S. Rieux, N. Rodriguez, D. Schneidman, F. Yuen;Delaware: J. Holleger, T. Tonwe; Florida: U. Ander-son, B. Austin, L. Bianco, F. Griffith, J. Jaffe, E. Kil-leavy, A. Kwon, C. Lewis, M. Manoucheri, L. Nitzberg,G. Ramos, P. Seabrooks, K. Sheikh, H. St John, T. StJohn, F. Zafar; Georgia: P. Douglass, R. Rhoades, R.Williams, A. Woodburn; Illinois: A. Chavarria, L. Cha-varria, M. Davidson, S. Ifft, J. Mathien, B. Smith, D.Steinmuller, M. Steinmuller; Indiana: A. Artis, J. Carter,M. Hutchinson, D. Smith; Kansas: P. Bowen, J. Cham-bers, J. Fullard, L. Terry, S. Waldren; Louisiana: P.Daigle, J. Diggs, P. Lakshmiprasad, A. Leitz, B. Rich-ardson; Maryland: E. Brightwell, J. Chandler, G. Den-ton, M. Kelemen, D. Tesch; Massachusetts: M. Cas-sidy, T. Sbarra; Michigan: R. Gudipati, C. Janners, S.Janners, M. Keshishian, W. Packard, B. Sheridan; Min-nesota: L. Loes, K. Margolis; Missouri: S. Brennac, C.Crosdale, K. Gage, T. McKeel, T. McKeel; New Hamp-shire: J. Aliseo, M. Jacobs; New York: C. Anderson,S. Athanail, D. Castaldo, R. Castaldo, D. Clark, D.Copley, B. Dobrzynski, D. Dobrzynski, R. Farron, B.

Hoffman, J. McLaughlin, K. Ong, T. Peoples, M. Price,I. Salom, S. Sears, R. Sutton, A. Zugibe, F. Zugibe; Ohio:L. Ballone, G. Barnett, D. Bradford, W. Feeman, C. Grif-fin, S. Moore, A. Narraway, G. Novak, G. Schroeder,J. Wiggins; Oklahoma: V. Christy, Y. Ong; Pennsyl-vania: A. Friedman, C. Matelan, M. Reyes, F. Sessoms,S. Silver, D. Watson; Puerto Rico: C. LaSalle-Ruiz; Ten-nessee: L. Hays, M. Houston; Texas: L. Alexander, D.Corral, B. Montgomery, J. Pappas, R. Rocha; VirginIslands: D. Galiber, S. Healy; Investigators and Coor-dinators, Canada: Nova Scotia: T. Machel, J. Mo-rash; Ontario: J. Cha, D. Dejewski, D. Jones, L. Jones,B. Lubelsky, R. Luton, A. Maczko, J. Otis.Acknowledgment: The ALLHAT Collaborative Re-search Group extends sincere appreciation to the42418 randomized participants without whom the trialcould not have been done. Thanks are also extendedto officers and coordinators of the research group whoparticipated in previous years: Steering Committee:Charles Francis, MD, John LaRosa, MD; NHLBI ProjectOffice: Gerald Payne, MD, Terry Manolio, MD, MS,Debra Egan, MS, MPH; ALLHAT Clinical Trials Cen-ter: C. Morton Hawkins, ScD, Cheryl Jones, ScD, Chris-tine Lusk, MPH, Barbara Kimmel, MS, MS, HeatherParks-Huitron, MHE, CHES, Melanie Gross, AdrianaBabiak-Vazquez, MPH, Gaston Benavides, PatrickCourtney, MA; ALLHAT Regions: Bronx, NY: Kim Bren-nan, Crystal Howard, MA; Chicago, Ill: Margaret Ga-zollo, RD, Julie Hynes, MS, RD, Charisse O’Neill, RN,BS; Birmingham, Ala: Cora E. Lewis, MD, MSPH; Se-attle, Wash: Kim Damon, Rebecca Letterer, RN, BSN,Susan Ross, RN, BSN; Minneapolis, Minn: Mukul Gan-guli, MVSc, PhD, Holly Jensen, Salma Koessel, MD,MPH, Carla Yunis, MD, MPH; ALLHAT Drug Distri-bution Center: Mary Mease, RPh, MPH; ALLHAT Elec-trocardiogram Reading Center: Carmen Christian-son, Bernadette Gloeb, MLS, Marsha McDonald.

REFERENCES

1. American Heart Association. 2002 Heart and StrokeStatistical Update. Dallas, Tex: American Heart As-sociation; 2001.2. Cutler JA, MacMahon SW, Furberg CD. Con-trolled clinical trials of drug treatment for hyperten-sion: a review. Hypertension. 1989;13:I36-I44.3. Psaty BM, Smith NL, Siscovick DS, et al. Health out-comes associated with antihypertensive therapies usedas first-line agents: a systematic review and meta-analysis. JAMA. 1997;277:739-745.4. The sixth report of the Joint National Committeeon prevention, detection, evaluation, and treatmentof high blood pressure. Arch Intern Med. 1997;157:2413-2446.5. Chalmers J, Zanchetti A. The 1996 report of a WorldHealth Organization expert committee on hyperten-sion control. J Hypertens. 1996;14:929-933.6. Collins R, Peto R, Godwin J, MacMahon S. Bloodpressure and coronary heart disease. Lancet. 1990;336:370-371.7. World Health Organization-International Societyof Hypertension Blood Pressure Lowering TreatmentTrialists’ Collaboration. Protocol for prospective col-laborative overviews of major randomized trials ofblood pressure lowering treatments. J Hypertens. 1998;16:127-137.8. SHEP Cooperative Research Group. Prevention ofstroke by antihypertensive drug treatment in older per-sons with isolated systolic hypertension: final resultsof the Systolic Hypertension in the Elderly Program(SHEP). JAMA. 1991;265:3255-3264.9. Staessen JA, Fagard R, Thijs L, et al. Randomiseddouble-blind comparison of placebo and active treat-ment for older patients with isolated systolic hyper-tension: the Systolic Hypertension in Europe(Syst-Eur) Trial Investigators. Lancet. 1997;350:757-764.10. Yusuf S, Sleight P, Pogue J, et al. Effects of an




angiotensin-converting-enzyme inhibitor, ramipril, oncardiovascular events in high-risk patients: the HeartOutcomes Prevention Evaluation Study Investiga-tors. N Engl J Med. 2000;342:145-153.11. PROGRESS Collaborative Group. Randomised trialof a perindopril-based blood-pressure-lowering regi-men among 6,105 individuals with previous stroke ortransient ischaemic attack. Lancet. 2001;358:1033-1041.12. Grimm RH Jr. Antihypertensive therapy: taking lip-ids into consideration. Am Heart J. 1991;122:910-918.13. Neal B, MacMahon S, Chapman N. Effects of ACEinhibitors, calcium antagonists, and other blood-pressure-lowering drugs: results of prospectively de-signed overviews of randomised trials: Blood Pres-sure Lowering Treatment Trialists’ Collaboration.Lancet. 2000;356:1955-1964.14. Pahor M, Psaty BM, Alderman MH, et al. Healthoutcomes associated with calcium antagonists com-pared with other first-line antihypertensive thera-pies: a meta-analysis of randomised controlled trials.Lancet. 2000;356:1949-1954.15. Poulter NR. Treatment of hypertension: a clini-cal epidemiologist’s view. J Cardiovasc Pharmacol.1991;18(suppl 2):S35-S38.16. Neaton JD, Grimm RH Jr, Prineas RJ, et al. Treat-ment of mild hypertension study: final results: Treat-ment of Mild Hypertension Study Research Group.JAMA. 1993;270:713-724.17. Materson BJ, Reda DJ, Cushman WC, et al. Single-drug therapy for hypertension in men: a comparisonof six antihypertensive agents with placebo: the De-partment of Veterans Affairs Cooperative Study Groupon Antihypertensive Agents. N Engl J Med. 1993;328:914-921.18. Davis BR, Cutler JA, Gordon DJ, et al. Rationaleand design for the Antihypertensive and Lipid-Lowering Treatment to Prevent Heart Attack Trial(ALLHAT). Am J Hypertens. 1996;9:342-360.19. The ALLHAT Officers and Coordinators for the ALL-HAT Collaborative Research Group. Major cardiovas-cular events in hypertensive patients randomized to dox-azosin vs chlorthalidone: the Antihypertensive and Lipid-Lowering Treatment to Prevent Heart Attack Trial(ALLHAT). JAMA. 2000;283:1967-1975.20. Davis BR, Cutler JA, Furberg CD, et al. Relation-ship of antihypertensive treatment regimens and changein blood pressure to risk for heart failure in hyperten-sive patients randomly assigned to doxazosin or chlortha-lidone: further analyses from the Antihypertensive andLipid-Lowering Treatment to Prevent Heart Attack Trial.Ann Intern Med. 2002;137:313-320.21. The ALLHAT Officers and Coordinators for theALLHAT Collaborative Research Group. Major out-comes in moderately hypercholesterolemic, hyper-tensive patients randomized to pravastatin vs usualcare: the Antihypertensive and Lipid-Lowering Treat-ment to Prevent Heart Attack Trial. JAMA. 2002;288:2998-3007.22. Grimm RH Jr, Margolis KL, Papademetriou V, etal. Baseline characteristics of participants in the Anti-hypertensive and Lipid-Lowering Treatment to Pre-vent Heart Attack Trial (ALLHAT). Hypertension. 2001;37:19-27.23. The fifth report of the Joint National Committee onDetection, Evaluation, and Treatment of High Blood Pres-sure ( JNC V). Arch Intern Med. 1993;153:154-183.

24. Levey AS, Bosch JP, Lewis JB, et al. A more ac-curate method to estimate glomerular filtration ratefrom serum creatinine: a new prediction equation:Modification of Diet in Renal Disease Study Group.Ann Intern Med. 1999;130:461-470.25. Levey AS, Greene T, Kusek JW, et al. A simpli-fied equation to predict glomerular filtration rate fromserum creatinine. J Am Soc Nephrol. 2000;11:155A.26. Piller LB, Davis BR, Cutler JA, et al. Validation ofheart failure events in ALLHAT participants assignedto doxazosin. Curr Control Trials Cardiovasc Med.2002;3:10.27. Public Health Service-Health Care Financing Ad-ministration. International Classification of Dis-eases, Ninth Revision (ICD-9). 6th ed. Bethesda, Md:US Dept of Health and Human Services; 2001. DHHSPublication No. (PHS) 80-1260.28. Dunnett CW. A multiple comparisons procedurefor comparing several treatments with a control. J AmStat Assoc. 1955;50:1096-1121.29. Klein JP, Moeschberger ML. Survival Analysis:Techniques for Censored and Truncated Regression.New York, NY: Springer-Verlag; 1997.30. Davis BR, Hardy RJ. Upper bounds for type I andII error rates in conditional power calculations. Com-mun Stat. 1990;19:3571-3584.31. Lan KKG, DeMets DL. Discrete sequential bound-aries for clinical trials. Biometrika. 1983;70:659-663.32. Hansson L, Lindholm LH, Ekbom T, et al. Ran-domised trial of old and new antihypertensive drugsin elderly patients: cardiovascular mortality and mor-bidity: the Swedish Trial in Old Patients with Hyper-tension-2 study. Lancet. 1999;354:1751-1756.33. Brown MJ, Palmer CR, Castaigne A, et al. Mor-bidity and mortality in patients randomised to double-blind treatment with a long-acting calcium-channelblocker or diuretic in the International Nifedipine GITSstudy: Intervention as a Goal in Hypertension Treat-ment (INSIGHT). Lancet. 2000;356:366-372.34. MacMahon S, Neal B. Differences between blood-pressure-lowering drugs. Lancet. 2000;356:352-353.35. Kostis JB, Davis BR, Cutler J, et al. Prevention of heartfailure by antihypertensive drug treatment in older per-sons with isolated systolic hypertension: SHEP Coop-erative Research Group. JAMA. 1997;278:212-216.36. Ad Hoc Subcommittee of the Liaison Committeeof the World Health Organisation and the Interna-tional Society of Hypertension. Effects of calcium an-tagonists on the risks of coronary heart disease, can-cer and bleeding. J Hypertens. 1997;15:105-115.37. Furberg CD, Psaty BM, Meyer JV. Nifedipine: dose-related increase in mortality in patients with coronaryheart disease. Circulation. 1995;92:1326-1331.38. Agodoa LY, Appel L, Bakris GL, et al. Effect oframipril vs amlodipine on renal outcomes in hyper-tensive nephrosclerosis: a randomized controlled trial.JAMA. 2001;285:2719-2728.39. Hall WD, Kusek JW, Kirk KA, et al. Short-term ef-fects of blood pressure control and antihypertensivedrug regimen on glomerular filtration rate: the African-American Study of Kidney Disease and HypertensionPilot Study. Am J Kidney Dis. 1997;29:720-728.40. ter Wee PM, De Micheli AG, Epstein M. Effectsof calcium antagonists on renal hemodynamics andprogression of nondiabetic chronic renal disease. ArchIntern Med. 1994;154:1185-1202.41. Lonn EM, Yusuf S, Jha P, et al. Emerging role of

angiotensin-converting enzyme inhibitors in cardiacand vascular protection. Circulation. 1994;90:2056-2069.42. Weir MR, Dzau VJ. The renin-angiotensin-aldosterone system: a specific target for hyperten-sion management. Am J Hypertens. 1999;12:205S-213S.43. Jafar TH, Schmid CH, Landa M, et al. Angiotensin-converting enzyme inhibitors and progression of non-diabetic renal disease: a meta-analysis of patient-level data. Ann Intern Med. 2001;135:73-87.44. Effect of enalapril on mortality and the develop-ment of heart failure in asymptomatic patients withreduced left ventricular ejection fractions: the SOLVDInvestigators. N Engl J Med. 1992;327:685-691.45. Collins R, Peto R, MacMahon S, et al. Blood pres-sure, stroke, and coronary heart disease: Part 2, short-term reductions in blood pressure: overview of ran-domised drug trials in their epidemiological context.Lancet. 1990;335:827-838.46. MacMahon S, Peto R, Cutler J, et al. Blood pres-sure, stroke, and coronary heart disease: Part 1, pro-longed differences in blood pressure: prospective ob-servational studies corrected for the regression dilutionbias. Lancet. 1990;335:765-774.47. Staessen JA, Byttebier G, Buntinx F, et al. Anti-hypertensive treatment based on conventional or am-bulatory blood pressure measurement: a randomizedcontrolled trial: Ambulatory Blood Pressure Monitor-ing and Treatment of Hypertension Investigators.JAMA. 1997;278:1065-1072.48. Franklin SS. Cardiovascular risks related to in-creased diastolic, systolic and pulse pressure: an epi-demiologist’s point of view. Pathol Biol (Paris). 1999;47:594-603.49. Saunders E, Weir MR, Kong BW, et al. A com-parison of the efficacy and safety of a beta-blocker, acalcium channel blocker, and a converting enzyme in-hibitor in hypertensive blacks. Arch Intern Med. 1990;150:1707-1713.50. Rahman M, Douglas JG, Wright JT. Pathophysi-ology and treatment implications of hypertension inthe African-American population. Endocrinol MetabClin North Am. 1997;26:125-144.51. Cushman WC, Reda DJ, Perry HM, et al. Re-gional and racial differences in response to antihyper-tensive medication use in a randomized controlled trialof men with hypertension in the United States. ArchIntern Med. 2000;160:825-831.52. Exner DV, Dries DL, Domanski MJ, Cohn JN. Lesserresponse to angiotensin-converting-enzyme inhibi-tor therapy in black as compared with white patientswith left ventricular dysfunction. N Engl J Med. 2001;344:1351-1357.53. Carson P, Ziesche S, Johnson G, Cohn JN. Racialdifferences in response to therapy for heart failure:analysis of the vasodilator-heart failure trials: Vaso-dilator-Heart Failure Trial Study Group. J Card Fail.1999;5:178-187.54. Dries DL, Strong MH, Cooper RS, Drazner MH.Efficacy of angiotensin-converting enzyme inhibitionin reducing progression from asymptomatic left ven-tricular dysfunction to symptomatic heart failure in blackand white patients. J Am Coll Cardiol. 2002;40:311-317.55. Manolio TA, Cutler JA, Furberg CD, et al. Trendsin pharmacologic management of hypertension in theUnited States. Arch Intern Med. 1995;155:829-837.




CORRECTIONS

Incorrect Byline: In the Original Contribution entitled “Myocardial Perfusion Imaging for Evaluation and Triage of Patients With Suspected Acute Cardiac Ischemia: ARandomized Controlled Trial” published in the December 4, 2002, issue of THE JOURNAL (2002;288:2693-2700), the order of authors in the byline was incorrect.Jonathan Handler, MD, should have been listed between John L. Griffith, PhD, and Gary V. Heller, MD, PhD.

Incorrect Data in Table: In the Original Contribution entitled “Major Outcomes in High-Risk Hypertensive Patients Randomized to Angiotensin-Converting EnzymeInhibitor or Calcium Channel Blocker vs Diuretic: The Antihypertensive and Lipid-Lowering Treatment to Prevent Heart Attack Trial (ALLHAT)” published in the De-cember 18, 2002, issue of THE JOURNAL (2002;288:2981-2997), there were incorrect data in TABLE 6. This table replaces the one on page 2992. The new data do notaffect the results or conclusions of the original article.

Table 6. Causes of Death by Antihypertensive Treatment Group*

No. (6-Year Rate per 100 Persons) P Value






Total deaths 2203 (17.3) 1256 (16.8) 1314 (17.2) .20 .90

Cardiovascular 996 (8.0) 603 (8.5) 618 (8.5) .76 .39

Myocardial infarction 296 (2.4) 169 (2.3) 157 (2.2) .66 .25

Definite CHD 118 (1.1) 72 (1.2) 77 (1.0) .88 .52

Possible CHD 128 (1.1) 71 (1.1) 95 (1.4) .62 .10

Stroke 162 (1.4) 92 (1.4) 121 (1.7) .71 .06

Heart failure 114 (1.0) 83 (1.4) 68 (1.1) .17 .98

Other CVD 178 (1.4) 116 (1.7) 100 (1.5) .46 .66

Noncardiovascular 1067 (8.9) 571 (8.0) 616 (8.6) .04 .57

Cancer 515 (4.3) 285 (3.8) 302 (4.1) .31 .86

Kidney disease 36 (0.4) 24 (0.5) 27 (0.5) .68 .37

Unintentional injury/suicide/homicide 66 (0.6) 19 (0.4) 28 (0.4) .005 .14

Other non−CVD 450 (4.0) 243 (3.7) 259 (3.9) .21 .68

Unknown 140 (1.2) 82 (1.2) 80 (1.1) .89 .78

*CHD indicates coronary heart disease; CVD, cardiovascular disease.

CME ANNOUNCEMENTOnline CME to Begin in Mid-2003

In mid-2003, online CME will be available for JAMA/Archives journalsand will offer many enhancements:

• Article-specific questions• Hypertext links from questions to the relevant content• Online CME questionnaire• Printable CME certificates and ability to access total CME credits

We apologize for the interruption in CME and hope that you willenjoy the improved online features that will be available in mid-2003.

LETTERS

178 JAMA, January 8, 2003—Vol 289, No. 2 (Reprinted) ©2003 American Medical Association. All rights reserved.


in digestion, absorption, and metabolism further increase de-mand and decrease utilization of critical nutrients. When re-stricted diets are necessary, alternative methods of deliveringessential nutrients should be considered.

Jonelle E. Wright, [email protected] W. Reynolds Department of Geriatric MedicineGarth J. Willis, MHSUniversity of Oklahoma College of MedicineOklahoma CityMarilyn S. Edwards, PhD, RDDepartment of Internal MedicineUniversity of Texas Medical SchoolHouston

Funding/Support: This research was funded by National VA Merit Award #E2117,the Retirement Research Foundation, and the National VA Center for Healthy Ag-ing with Disabilities. It was supported by grants M01 RR02719 and MO1 RR-14467 from the National Institutes of Health, National Center for Research Re-sources, General Clinical Research Center.Acknowledgment: Biostatistical assistance was provided by Donald E. Parker, PhD,and Christie E. Burgin, PhD, of Applied Research Consultants and the Department

of Biostatistics and Epidemiology, University of Oklahoma Health Sciences Cen-ter, Oklahoma City. We also thank Ji H. Park, BS, for her help with organizing thedata and constructing the Figure.

1. Sullivan DH. The role of nutrition in increasing morbidity and mortality. ClinGeriatr Med. 1995;11:661-674.2. McGee M, Jensen GL. Nutrition in the elderly. J Clin Gastroenterol. 2000;30:372-380.3. Panel on Micronutrients, Subcommittees on Upper Reference Levels of Nutri-ents and of Interpretation and Use of Dietary Reference Intakes, Institute of Medi-cine. Dietary Reference Intakes for Vitamin A, Vitamin K, Arsenic, Boron, Chro-mium, Copper, Iodine, Iron, Manganese, Molybdenum, Nickel, Silicon, Vanadium,and Zinc. Washington, DC: National Academies Press; 2002.4. Levine M, Dhariwal KR, Welch RW, Wang Y, Park JB. Determination of opti-mal vitamin C requirements in humans. Am J Clin Nutr. 1995;62:1347S-1356S.

CORRECTION

Investigator Omitted: In the Original Contribution entitled “Major Outcomes inHigh-Risk Hypertensive Patients Randomized to Angiotensin-Converting EnzymeInhibitor or Calcium Channel Blocker vs Diuretic: the Antihypertensive and Lipid-Lowering Treatment to Prevent Heart Attack Trial (ALLHAT)” published in the De-cember 18, 2002, issue of THE JOURNAL (2002;288:2981-2997), Pasquale F. Nes-tico, MD, was inadvertently omitted from the list of ALLHAT investigators. Hisname should appear on page 2996 under “Pennsylvania.”

LETTERS

2196 JAMA, May 12, 2004—Vol 291, No. 18 (Reprinted) ©2004 American Medical Association. All rights reserved.


n engl j med

350;15

www.nejm.org april

8, 2004

1495

The

new england

journal

of

medicine

established in 1812

april

8

,

2004

vol. 350 no. 15

Intensive versus Moderate Lipid Lowering with Statins after Acute Coronary Syndromes

Christopher P. Cannon, M.D., Eugene Braunwald, M.D., Carolyn H. McCabe, B.S., Daniel J. Rader, M.D., Jean L. Rouleau, M.D., Rene Belder, M.D., Steven V. Joyal, M.D., Karen A. Hill, B.A., Marc A. Pfeffer, M.D., Ph.D., and Allan M. Skene, Ph.D., for the Pravastatin or Atorvastatin Evaluation and Infection Therapy–Thrombolysis

in Myocardial Infarction 22 Investigators*

abstract

From the Thrombolysis in Myocardial In-farction (TIMI) Study Group, Cardiovas-cular Division, Department of Medicine,Brigham and Women’s Hospital and Har-vard Medical School, Boston (C.P.C., E.B.,C.H.M., M.A.P.); the University of Penn-sylvania, Philadelphia (D.J.R.); the Univer-sity of Montreal, Montreal (J.L.R.); Bristol-Myers Squibb, Princeton, N.J. (R.B., S.V.J.);and the Nottingham Clinical ResearchGroup, Nottingham, United Kingdom(K.A.H., A.M.S.). Address reprint requeststo Dr. Cannon at the TIMI Study Group,Cardiovascular Division, Brigham andWomen’s Hospital, 75 Francis St., Boston,MA 02115, or at [email protected].

*The investigators and research coordina-tors who participated in the Pravastatinor Atorvastatin Evaluation and InfectionTherapy–Thrombolysis in Myocardial In-farction 22 (PROVE IT–TIMI 22) studyare listed in the Appendix.

This article was published at www.nejm.orgon March 8, 2004.

N Engl J Med 2004;350:1495-504.


background

Lipid-lowering therapy with statins reduces the risk of cardiovascular events, but theoptimal level of low-density lipoprotein (LDL) cholesterol is unclear.

methods

We enrolled 4162 patients who had been hospitalized for an acute coronary syndromewithin the preceding 10 days and compared 40 mg of pravastatin daily (standard therapy)with 80 mg of atorvastatin daily (intensive therapy). The primary end point was a com-posite of death from any cause, myocardial infarction, documented unstable angina re-quiring rehospitalization, revascularization (performed at least 30 days after randomiza-tion), and stroke. The study was designed to establish the noninferiority of pravastatinas compared with atorvastatin with respect to the time to an end-point event. Follow-uplasted 18 to 36 months (mean, 24).

results

The median LDL cholesterol level achieved during treatment was 95 mg per deciliter(2.46 mmol per liter) in the standard-dose pravastatin group and 62 mg per deciliter(1.60 mmol per liter) in the high-dose atorvastatin group (P<0.001). Kaplan–Meier es-timates of the rates of the primary end point at two years were 26.3 percent in the prav-astatin group and 22.4 percent in the atorvastatin group, reflecting a 16 percent reduc-tion in the hazard ratio in favor of atorvastatin (P=0.005; 95 percent confidence interval,5 to 26 percent). The study did not meet the prespecified criterion for equivalence butdid identify the superiority of the more intensive regimen.

conclusions

Among patients who have recently had an acute coronary syndrome, an intensive lipid-lowering statin regimen provides greater protection against death or major cardiovas-cular events than does a standard regimen. These findings indicate that such patientsbenefit from early and continued lowering of LDL cholesterol to levels substantially be-low current target levels.



n engl j med

350;15

www.nejm.org april

8

,

2004

The

new england journal

of

medicine

1496

everal large, randomized, con-

trolled trials have documented that cholester-ol-lowering therapy with 3-hydroxy-3-meth-

ylglutaryl coenzyme A reductase inhibitors (statins)reduces the risk of death or cardiovascular eventsacross a wide range of cholesterol levels whether ornot patients have a history of coronary artery dis-ease.

1-7

The doses of statins used in these trials re-duced low-density lipoprotein (LDL) cholesterollevels by 25 to 35 percent, and current guidelinesrecommend a target LDL cholesterol level of lessthan 100 mg per deciliter (2.59 mmol per liter) forpatients with established coronary artery disease ordiabetes.

8,9

It is not clear whether lowering lipid lev-els further would increase the clinical benefit. Ac-cordingly, the Pravastatin or Atorvastatin Evaluationand Infection Therapy–Thrombolysis in MyocardialInfarction 22 (PROVE IT–TIMI 22) trial was de-signed to compare the standard degree of LDL cho-lesterol lowering to approximately 100 mg per deci-liter with the use of 40 mg of pravastatin daily

2,3

withmore intensive LDL cholesterol lowering to approx-imately 70 mg per deciliter (1.81 mmol per liter)with the use of 80 mg of atorvastatin daily

10

as amean of preventing death or major cardiovascularevents in patients with an acute coronary syndrome.

patient population

Between November 15, 2000, and December 22,2001, 4162 patients were enrolled at 349 sites ineight countries (see the Appendix). The protocolwas approved by the relevant institutional reviewboards, and written informed consent was obtainedfrom all patients. As described previously,

11

menand women who were at least 18 years old were el-igible for inclusion if they had been hospitalizedfor an acute coronary syndrome — either acute my-ocardial infarction (with or without electrocardio-graphic evidence of ST-segment elevation) or high-risk unstable angina — in the preceding 10 days.Patients had to be in stable condition and were tobe enrolled after a percutaneous revascularizationprocedure if one was planned. Finally, patients hadto have a total cholesterol level of 240 mg per deci-liter (6.21 mmol per liter) or less, measured at thelocal hospital within the first 24 hours after the onsetof the acute coronary syndrome or up to six monthsearlier if no sample had been obtained during thefirst 24 hours. Patients who were receiving long-term lipid-lowering therapy at the time of their in-

dex acute coronary syndrome had to have a totalcholesterol level of 200 mg per deciliter (5.18 mmolper liter ) or less at the time of screening in the localhospital.

Patients were ineligible for the study if they hada coexisting condition that shortened expected sur-vival to less than two years, were receiving therapywith any statin at a dose of 80 mg per day at thetime of their index event or lipid-lowering therapywith fibric acid derivatives or niacin that could notbe discontinued before randomization, had receiveddrugs that are strong inhibitors of cytochrome P-4503A4 within the month before randomization or werelikely to require such treatment during the studyperiod (because atorvastatin is metabolized by thispathway), had undergone percutaneous coronaryintervention within the previous six months (otherthan for the qualifying event) or coronary-artery by-pass surgery within the previous two months or werescheduled to undergo bypass surgery in response tothe index event, had factors that might prolong theQT interval, had obstructive hepatobiliary disease orother serious hepatic disease, had an unexplainedelevation in the creatine kinase level that was morethan three times the upper limit of normal and thatwas not related to myocardial infarction, or had acreatinine level of more than 2.0 mg per deciliter(176.8 µmol per liter).

study protocol

The protocol specified that patients were to receivestandard medical and interventional treatment foracute coronary syndromes, including aspirin at adose of 75 to 325 mg daily, with or without clopid-ogrel or warfarin. Patients were not permitted to betreated with any lipid-modifying therapy other thanthe study drug. Eligible patients were randomly as-signed in a 1:1 ratio to receive 40 mg of pravastatinor 80 mg of atorvastatin daily in a double-blind, dou-ble-dummy fashion. In addition, patients were alsorandomly assigned to receive with the use of a two-by-two factorial design a 10-day course of gatiflox-acin or placebo every month during the trial. The re-sults of the antibiotic component of the trial are notreported here.

Patients were seen for follow-up visits and re-ceived dietary counseling

8

at 30 days, at 4 months,and every 4 months thereafter until their final visit inAugust or September 2003. Patients who discontin-ued the study drug during the trial were followed bymeans of telephone calls. Blood samples were ob-tained at randomization, at 30 days, at 4, 8, 12, and

s

methods



n engl j med

350;15

www.nejm.org april

8, 2004

intensive vs. moderate lipid lowering with statins

1497

16 months, and at the final visit for the measure-ment of lipids and other components that were partof the safety assessment. Measurements were madeat the core laboratories listed in the Appendix. LDLcholesterol levels were monitored, and the protocolspecified that the dose of pravastatin was to increaseto 80 mg in a blinded fashion if the LDL cholesterollevel exceeded 125 mg per deciliter (3.23 mmol perliter) on two consecutive visits and the patient hadbeen taking study medication and had returned forthe required study visits. The dose of either studydrug could be halved in the event of abnormal liver-function results, elevations in creatine kinase levels,or myalgias.

Patients were followed for 18 to 36 months, withan average follow-up of 24 months. The trial con-tinued until 925 events had been reported to the co-ordinating center, after which time all patients wererequested to return for a final study visit. Eight pa-tients (0.2 percent) were lost to follow-up.

end points

The primary efficacy outcome measure was the timefrom randomization until the first occurrence of acomponent of the primary end point: death fromany cause, myocardial infarction, documented un-stable angina requiring rehospitalization, revascu-larization with either percutaneous coronary inter-vention or coronary-artery bypass grafting (if theseprocedures were performed at least 30 days afterrandomization), and stroke. Myocardial infarctionwas defined by the presence of symptoms suggestiveof ischemia or infarction, with either electrocardio-graphic evidence (new Q waves in two or more leads)or cardiac-marker evidence of infarction, accordingto the standard TIMI and American College of Car-diology definition.

12,13

Unstable angina was de-fined as ischemic discomfort at rest for at least 10minutes prompting rehospitalization, combinedwith one of the following: ST-segment or T-wavechanges, cardiac-marker elevations that were abovethe upper limit of normal but did not meet the cri-teria for myocardial infarction, or a second episodeof ischemic chest discomfort lasting more than10 minutes and that was distinct from the episodethat had prompted hospitalization. Secondary endpoints were the risk of death from coronary heartdisease, nonfatal myocardial infarction, or revascu-larization (if it was performed at least 30 days afterrandomization), the risk of death from coronaryheart disease or nonfatal myocardial infarction, andthe risk of the individual components of the primaryend point.

statistical analysis

Although the trial was designed as a time-to-eventstudy, the definition of noninferiority was arrived atthrough a consideration of two-year event rates. Forthe comparison of pravastatin with atorvastatin, wedefined the prespecified boundary for noninferior-ity as an upper limit of the one-sided 95 percent con-fidence interval of the relative risk at two years of lessthan 1.17 (corresponding to a hazard ratio through-out follow-up of 1.198). Assuming a two-year eventrate of 22 percent in the atorvastatin group and thatthe two treatments had equivalent efficacy, we de-termined that enrollment of 2000 patients per groupwould give the study a statistical power of 87 per-cent and that this power would be preserved if fol-low-up continued until 925 end-point events hadoccurred.

14

A central randomization system wasused that involved a permuted-block design inwhich assignment was stratified according to cen-ter. Three interim assessments of efficacy and safetywere carried out by the data and safety monitoringboard. Rules for stopping the study early in the eventthat the superiority of either treatment was estab-lished were not prespecified.

All efficacy analyses are based on the intention-to-treat principle. Estimates of the hazard ratios andassociated 95 percent confidence intervals compar-ing pravastatin with atorvastatin were obtained withthe use of the Cox proportional-hazards model, withrandomized treatment as the covariate and stratifi-cation according to the receipt of gatifloxacin orplacebo. (Using the two-by-two factorial design,we conducted a preliminary test for interaction andfound none. For the primary end point, the interac-tion P value was 0.90 and the hazard ratios compar-ing pravastatin with atorvastatin were almost iden-tical for the gatifloxacin and placebo groups.) Whenit was determined that noninferiority was not dem-onstrated, the subsequent assessment of superioritywas carried out with the use of two-sided confidenceintervals. The investigators designed the trial andhad free and complete access to the data. Data co-ordination was performed by the Nottingham Clin-ical Research Group (see the Appendix). Investi-gators at TIMI, the sponsor, and members of theNottingham Clinical Research Group performeddata analysis jointly.

The two groups of patients were well matched withregard to base-line characteristics, with the excep-tion of a history of peripheral arterial disease, which

results



n engl j med

350;15

www.nejm.org april

8

,

2004

The

new england journal

of

medicine

1498

was more common in the pravastatin group thanthe atorvastatin group (P=0.03) (Table 1). Their av-erage age was 58 years, and 22 percent were women.Before their index event, 18 percent of patients hadhad a myocardial infarction, 11 percent had previ-ously undergone coronary-artery bypass surgery,and 18 percent had diabetes mellitus. The indexevent was high-risk unstable angina in approxi-mately one third of the patients, myocardial infarc-

tion without electrocardiographic evidence of ST-segment elevation in approximately one third, andmyocardial infarction with ST-segment elevationin one third. Sixty-nine percent of patients under-went percutaneous coronary intervention for thetreatment of their index acute coronary syndromebefore randomization. One quarter of the patientswere taking statin drugs at the time of the indexevent. Concomitant medications were administered

* Plus–minus values are means ±SD. None of the differences between groups were significant with the exception of a his-tory of peripheral arterial disease (P=0.03). Two patients did not have information regarding the electrocardiographic type of acute coronary syndrome, and one patient had missing information regarding prior statin use. MI denotes myo-cardial infarction, LDL low-density lipoprotein, and HDL high-density lipoprotein. To convert values for cholesterol to

millimoles per liter, multiply by 0.02586. To convert values for triglycerides to millimoles per liter, multiply by 0.01129.

Table 1. Base-Line Characteristics of the Patients.*

Characteristic40 mg of Pravastatin

(N=2063)80 mg of Atorvastatin

(N=2099)

Age — yr 58.3±11.3 58.1±11.2

Male sex — no. (%) 1617 (78.4) 1634 (77.8)

White race — no. (%) 1865 (90.4) 1911 (91.0)

Diabetes mellitus — no. (%) 361 (17.5) 373 (17.8)

Hypertension — no. (%) 1014 (49.2) 1077 (51.3)

Current smoker — no. (%) 766 (37.1) 763 (36.4)

Prior MI — no. (%) 395 (19.1) 374 (17.8)

Percutaneous coronary intervention — no. (%)Before index eventFor treatment of index event

320 (15.5)1426 (69.1)

322 (15.3)1442 (68.7)

Coronary bypass surgery before index event — no. (%) 221 (10.7) 233 (11.1)

Peripheral arterial disease — no. (%) 136 (6.6) 105 (5.0)

Prior statin therapy — no. (%) 514 (24.9) 535 (25.5)

Index event — no. (%)Unstable anginaMI without ST-segment elevationMI with ST-segment elevation

614 (29.8)757 (36.7)690 (33.4)

604 (28.8)747 (35.6)748 (35.6)

Lipid valuesTotal cholesterol

No. of patientsMedian — mg/dlInterquartile range — mg/dl

1981180

158–202

2014181

160–205

LDL cholesterolNo. of patientsMedian — mg/dlInterquartile range — mg/dl

1973106

87–127

2003106

89–128

HDL cholesterolNo. of patientsMedian — mg/dlInterquartile range — mg/dl

198139

33–46

201438

32–46

TriglyceridesNo. of patientsMedian — mg/dlInterquartile range — mg/dl

1984154

115–207

2016158

119–214



n engl j med

350;15

www.nejm.org april

8, 2004


1499

to patients during the treatment period as follows:aspirin to 93 percent, warfarin to 8 percent, clo-pidogrel or ticlodipine to 72 percent initially and20 percent at one year, beta-blockers to 85 per-cent, angiotensin-converting–enzyme inhibitors to69 percent, and angiotensin-receptor blockers to14 percent.

At the time of randomization, a median of sevendays after the onset of the index event, the medianLDL cholesterol levels were 106 mg per deciliter(2.74 mmol per liter) before treatment in each group(Fig. 1). The LDL cholesterol levels achieved duringfollow-up were 95 mg per deciliter (2.46 mmol perliter; interquartile range, 79 to 113 mg per deciliter[2.04 to 2.92 mmol per liter]) in the pravastatingroup and 62 mg per deciliter (1.60 mmol per liter;interquartile range, 50 to 79 mg per deciliter [1.29to 2.04 mmol per liter]) in the atorvastatin group(P<0.001). Among 2985 patients (75 percent) whohad not previously received statin therapy, the me-dian LDL cholesterol levels had fallen by 22 percentat 30 days in the pravastatin group and by 51 per-cent in the atorvastatin group (P<0.001). As antici-pated, among the 990 patients who had previouslyreceived statin therapy (25 percent), LDL cholesterollevels were essentially unchanged from base line(during statin therapy) in the pravastatin group,whereas they fell by an additional 32 percent in theatorvastatin group (P<0.001). Median high-densitylipoprotein cholesterol levels rose during follow-upby 8.1 percent in the pravastatin group and 6.5 per-cent in the atorvastatin group (P<0.001). MedianC-reactive protein levels fell from 12.3 mg per literat base line in each group to 2.1 mg per liter in thepravastatin group and 1.3 mg per liter in the ator-vastatin group (P<0.001).

primary end point

For all randomized patients, the Kaplan–Meierevent rates of the primary end point at two yearswere 26.3 percent in the standard-dose pravastatingroup and 22.4 percent in the high-dose atorva-statin group, representing a 16 percent reduction inthe hazard ratio favoring atorvastatin (P=0.005; 95percent confidence interval, 5 to 26 percent) (Fig. 2);this difference did not meet the criteria for equiva-lence. The benefit of high-dose atorvastatin as com-pared with standard-dose pravastatin emerged asearly as 30 days and was consistent over time (Fig.3). The risk of the secondary end point of death dueto coronary heart disease, myocardial infarction, orrevascularization was similarly reduced by 14 per-

cent in the atorvastatin group (P=0.029), with atwo-year event rate of 19.7 percent, as comparedwith 22.3 percent in the pravastatin group. The riskof death, myocardial infarction, or urgent revascu-larization was reduced by 25 percent in the atorva-statin group (P<0.001).

Among the individual components of the pri-mary end point, there was a consistent pattern ofbenefit favoring high-dose atorvastatin over stan-dard-dose pravastatin, which included a significant14 percent reduction in the need for revasculariza-tion (P=0.04), a 29 percent reduction in the risk ofrecurrent unstable angina (P=0.02), and nonsignif-icant reductions in the rates of death from any cause(28 percent, P=0.07) and of death or myocardial in-farction (18 percent, P=0.06) (Fig. 4). Stroke was in-frequent, but the rates did not differ significantly be-tween the groups.

The benefit of high-dose atorvastatin was con-sistent across the prespecified subgroups, includingmen and women, patients with unstable angina andthose with myocardial infarction, and those withand those without diabetes mellitus (Fig. 5). Thebenefit appeared to be greater among patients witha base-line LDL cholesterol level of at least 125 mgper deciliter, a prespecified subgroup, with a 34 per-cent reduction in the hazard ratio, as compared witha 7 percent reduction among patients with a base-line LDL cholesterol below 125 mg per deciliter(P for interaction=0.02).

Figure 1. Median Low-Density Lipoprotein (LDL) Cholesterol Levels during the Study.

To convert values for LDL cholesterol to millimoles per liter, multiply by 0.02586.

120

100

LDL

Cho

lest

erol

(mg/

dl)

80

60

40

20

0Base line 30 Days 4 Mo 8 Mo 16 Mo Final

40 mg of pravastatin

80 mg of atorvastatin

Time of Visit

No. of PatientsPravastatinAtorvastatin

1883 1910

1445 1461

1647 1645

1761 1758

1844 1856

1973 2003



n engl j med

350;15

www.nejm.org april

8

,

2004

The

new england journal

of

medicine

1500

tolerability and safety

The rates of discontinuation of treatment becauseof an adverse event or the patient’s preference or forother reasons were 21.4 percent in the pravastatingroup and 22.8 percent in the atorvastatin group atone year (P=0.30) and 33.0 percent and 30.4 per-cent, respectively, at two years (P=0.11). Duringtreatment, the dose of pravastatin was increased to80 mg in 8 percent of patients, and the dose washalved among 1.4 percent of the patients in the prav-astatin group and 1.9 percent of those in the atorva-statin group (P=0.20), owing to side effects or liver-function abnormalities. The percentages of patientswho had elevations in alanine aminotransferase lev-els that were more than three times the upper limitof normal were 1.1 percent in the pravastatin groupand 3.3 percent in the atorvastatin group (P<0.001).The study medication was discontinued by the in-vestigators because of a report of myalgias or mus-cle aches or elevations in creatine kinase levels in2.7 percent of pravastatin-treated patients, as com-pared with 3.3 percent of atorvastatin-treated pa-tients (P=0.23). There were no cases of rhabdomy-olysis in either group.

In this comparison of two statin regimens of differ-ent lipid-lowering intensities for the prevention of

discussion

cardiovascular events, intensive therapy with high-dose atorvastatin resulted in a median LDL choles-terol level of 62 mg per deciliter, as compared witha level of 95 mg per deciliter for standard-dose prav-astatin. Among patients who had recently been hos-pitalized for an acute coronary syndrome, the moreintensive regimen resulted in a lower risk of deathfrom any cause or major cardiac events than did amore moderate degree of lipid lowering with the useof a standard dose of a statin. Although prior place-bo-controlled studies have shown that a standard-dose statin is beneficial,

1-7

we demonstrated thatmore intensive lipid lowering significantly increasedthis clinical benefit.

Although the exact mechanism of the benefitcannot be established solely on the basis of our re-sults, the extent of the benefit afforded by the 80-mgdose of atorvastatin is in keeping with what wouldbe expected on the basis of the greater degree of lip-id lowering produced by this regimen. In the HeartProtection study, statin treatment resulted in an LDLcholesterol level that was 40 mg per deciliter (1.03mmol per liter) lower than the value in the placebogroup and that was accompanied by a 25 percent re-duction in cardiovascular events. In our study, theLDL cholesterol level was 33 mg per deciliter (0.85mmol per liter) lower in the atorvastatin group thanin the pravastatin group. This difference shouldtranslate into a 20 percent reduction in clinicalevents, which is very similar to the 16 percent re-duction we observed and suggests that much of thebenefit is attributable to the difference in the degreeof LDL cholesterol lowering. However, we cannotexclude the possibility that the difference in clinicaloutcomes may be due in part to non–lipid-relatedpleiotropic effects, which may differ between thetwo statins we used.

15

Future trials involving differ-ent doses of a single statin should help address thispossibility.

Intensive therapy with high-dose atorvastatinhad a consistent beneficial effect on cardiac events,including a significant 29 percent reduction in therisk of recurrent unstable angina and a 14 percentreduction in the need for revascularization. The re-duction in the rate of death from any cause was ofborderline significance (28 percent, P=0.07), sug-gesting that more aggressive lipid lowering is im-portant not only to reduce the risk of recurrent is-chemia, but possibly also to decrease the risk offatal events.

The reduction in clinical events with the moreintensive lipid-lowering therapy was apparent as

Figure 2. Kaplan–Meier Estimates of the Incidence of the Primary End Point of Death from Any Cause or a Major Cardiovascular Event.

Intensive lipid lowering with the 80-mg dose of atorvastatin, as compared with moderate lipid lowering with the 40-mg dose of pravastatin, reduced the hazard ratio for death or a major cardiovascular event by 16 percent.

30

25

Dea

th o

r M

ajor

Car

diov

ascu

lar

Even

t (%

) 20

15

10

5

00 3 6 9 12 15 18 21 24 27 30

40 mg of pravastatin80 mg of atorvastatin

P=0.005

Months of Follow-up

No. at RiskPravastatinAtorvastatin

138 133

2063 2099

1688 1736

1536 1591

1423 1485

810 842



n engl j med

350;15

www.nejm.org april

8, 2004


1501

early as 30 days after the start of therapy. This rapidtime frame is similar to that reported with statintreatment in the placebo-controlled MyocardialIschemia Reduction with Aggressive CholesterolLowering (MIRACL) trial

16

and in prior observation-al studies.

17,18

We studied patients who had beenhospitalized for an acute coronary syndrome and

enrolled them immediately after their condition hadstabilized. Three quarters of the patients were treat-ed with an early invasive strategy, and the majoritywere treated with multiple medications for second-ary prevention, including antiplatelet therapy, beta-blockers, and angiotensin-converting–enzyme in-hibitors (and a statin as part of the trial design).

Figure 3. Hazard Ratio for the the Primary End Point of Death from Any Cause or a Major Cardiovascular Event at 30, 90, and 180 Days and at the End of Follow-up in the High-Dose Atorvastatin Group, as Compared with the Standard-Dose Pravastatin Group.

Event rates are Kaplan–Meier estimates censored at the time points indicated with the use of the average duration of fol-low-up (two years). CI denotes confidence interval.

Censoring Time Hazard Ratio (95% CI) Risk ReductionEvent Rates

0.50 0.75 1.00 1.25 1.50

High-DoseAtorvastatin Better

Standard-DosePravastatin Better

Atorvastatin Pravastatin

30 Days

90 Days

180 Days

End of follow-up

percent

17

18

14

16

1.9

6.3

12.2

22.4

2.2

7.7

14.1

26.3

Figure 4. Estimates of the Hazard Ratio for the Secondary End Points and the Individual Components of the Primary End Point in the High-Dose Atorvastatin Group, as Compared with the Standard-Dose Pravastatin Group.

CI denotes confidence interval, CHD coronary heart disease, and MI myocardial infarction. Revascularization was per-formed at least 30 days after randomization.

End Point2-Yr Event Rates

0.50 1.00 1.50



Atorvastatin Pravastatin

Death from any cause

Death from CHD

Death from other causes

MI

Death or MI

Death from CHD or MI

Stroke

Revascularization

Unstable angina requiringhospitalization

MI, revascularization, or deathfrom CHD

percent

28

30

27

13

2.2

1.1

1.2

6.6

3.2

1.4

1.8

7.4

18

16

14

14

8.3

7.2

16.3

19.7

10.0

8.3

18.8

22.3

29

¡9

3.8

1.0

5.1

1.0

Hazard Ratio (95% CI) Risk Reduction



n engl j med

350;15

www.nejm.org april

8

,

2004

The

new england journal

of

medicine

1502

Nonetheless, early and continued separation of theevent curves was observed in the more intensivelipid-lowering group. This early reduction in eventrates in patients with acute coronary syndromescontrasts with the lag of approximately one to twoyears in prior studies of statins in patients withchronic atherosclerosis.

1-6

These data suggest thatthis population of patients with acute coronary syn-

dromes, who have a culprit lesion and frequentlymultiple additional vulnerable plaques as well,

19,20

can derive particular benefit from early and intensivelipid lowering with statins. The current guidelinesof the American College of Cardiology and Ameri-can Heart Association recommend instituting lipid-lowering therapy at the time of hospital dischargein patients with acute coronary syndromes, on the

Figure 5. Two-Year Event Rates and Estimates of the Hazard Ratio for the Primary End Point in the High-Dose Atorva-statin Group, as Compared with the Standard-Dose Pravastatin Group, According to Base-Line Characteristics.

A test for interaction was significant only for a base-line low-density lipoprotein (LDL) value of at least 125 mg per deci-liter, as compared with a value of less than 125 mg per deciliter (P=0.02). LDL cholesterol was measured at base line in a total of 3976 patients, and high-density lipoprotein (HDL) cholesterol was measured in 3995. Two patients did not have information regarding the electrocardiographic type of acute coronary syndrome, and one patient had missing in-formation regarding prior statin use. MI denotes myocardial infarction.

Base-Line Characteristic No. of Patients (%) Pravastatin

Sex

Male

Female

Age

<65 yr

≥65 yr

Prior smoking

Diabetes

No

No

Yes

Yes

Prior statin therapy

Index event

No

MI without ST-segment elevation

Unstable angina

LDL cholesterol

MI with ST-segment elevation

<125 mg/dl

≥125 mg/dl

≥40 mg/dl

HDL cholesterol

<40 mg/dl

Yes

percent

3251 (78)

911 (22)

23.0

20.3

26.2

27.0

2932 (70)

1230 (30)

734 (18)

3428 (82)

28.1

20.1

28.8

29.5

25.0

34.6

21.0 24.6

3077 (74)

1049 (25)

3112 (75)

1218 (29)

1504 (36)

1438 (35)

1091 (27)

2885 (73)

1776 (44)

2219 (56)

22.8 26.5

1085 (26) 21.3 25.9

27.5 28.9

20.6 25.5

26.5 31.4

19.0 24.1

22.6 24.2

20.1 28.2

23.5 25.6

21.7 26.7

23.1 26.0

AtorvastatinHazard Ratio

0.50 1.00 1.50



2-Yr Event Rates



n engl j med

350;15

www.nejm.org april

8, 2004


1503

theory that it will improve patients’ compliance withthe use of statins for long-term secondary preven-tion.

21

Our data are evidence that such therapy willalso provide protection against early recurrent car-diovascular events.

After the early separation of the clinical-eventcurves, we also observed a continued benefit of ator-vastatin therapy throughout the follow-up period oftwo and one-half years (Fig. 2). It cannot be deter-mined from this study whether this longer-termbenefit was due to ongoing intensive lipid-loweringtherapy or was the result of an early benefit in stabi-lizing vulnerable plaques with the early and inten-sive treatment after the acute event. It also cannot bedetermined whether other differences between thetwo statins used explain the observed clinical bene-fit. Nonetheless, our findings suggest that patientswith acute coronary syndromes who receive earlyand intensive lipid-lowering therapy continue to de-rive benefit in the chronic phase of atherosclerosiswhen high-dose statin therapy is maintained.

Our finding of a continued benefit of intensivelipid-lowering therapy during the follow-up phaseis consistent with studies showing that such an ap-proach results in a slower rate of progression of ath-erosclerosis in patients with stable coronary arterydisease

22

or in those who undergo coronary-arterybypass grafting,

23

as well as in greater reductions incarotid intimal–medial thickening.

24-26

Althoughour study documented a benefit for up to an aver-age of two years of follow-up, several large, ongoingtrials involving patients with stable atherosclerosiswill determine the five-year outcomes of intensiveas compared with moderate lipid lowering.

27

It is important to note that our safety and effica-cy results were obtained in a carefully selected andmonitored study population (for example, we ex-cluded patients who were concomitantly receivingstrong inhibitors of cytochrome P-450 3A4, be-cause this is integral to the route of metabolism ofatorvastatin). Although both drugs were generallywell tolerated, there were significantly more liver-related side effects with high-dose atorvastatin thanwith standard-dose pravastatin. Patients in clinicalpractice generally have more coexisting conditionsthan did our patients, and they may not tolerate ahigh-dose statin regimen as well as our patients did.Thus, clinicians must take these factors into accountwhen applying the results of our trial in clinicalpractice.

The National Cholesterol Education Programand European guidelines currently recommend thatthe goal of treatment in patients with establishedcoronary artery disease should be an LDL choles-terol level of less than 100 mg per deciliter.

8,9

Al-though our data provide support for the use of thisapproach, given the substantially lower LDL choles-terol levels achieved in the group given 80 mg ofatorvastatin daily (median, 62 mg per deciliter), ourresults suggest that after an acute coronary syn-drome, the target LDL cholesterol level may be low-er than that recommended in the current guidelines.

Supported by Bristol-Myers Squibb and Sankyo.Drs. Rader and Pfeffer report having received consulting fees and

lecture fees from Bristol-Myers Squibb. Drs. Joyal and Belder areemployees of Bristol-Myers Squibb and have equity in the company.Dr. Rouleau reports having received consulting fees and lecture feesfrom Novartis.

appendix

The following investigators and research coordinators participated in the study (the complete list of investigators and coordinators is avail-able at www.timi.org):

Operations Committee

— E. Braunwald (Study Chair), C. Cannon (Principal Investigator), T. Grayston, B. Muhlestein,D. Rader, J. Rouleau, and members of the Data Coordinating Center and Sponsor are indicated with an asterisk;

Steering Committee

— Mem-bers of the Operations Committee and R. Byington, A. Castaigne, H. Darius, G. DeFerrari, B. Gersh, D. Gilbert, S. Grundy, G. Jackson, R.Knopp, I. Meredith, E. Ofili, M. Pfeffer, F. Sacks, P. Shah, S. Smith, A. Tonkin, J. Velasco;

TIMI Study Group

— C. McCabe (Project Director),*S. McHale (Project Manager);

Sponsor

(Bristol-Myers Squibb

,

Princeton, N.J., and Wallingford, Conn.) — R. Belder,* J. Breen,* G. Cucinotta,S. Joyal,* C.-S. Lin, K. Natarajan,* S. Nichols;

Data Coordinating Center

(Nottingham Clinical Research Group, Nottingham, United King-dom) — A. Skene,* K. Hill;

Clinical Events Committee —

M. Pfeffer (Chairman), R. Guertin-Mercier; J. Potzka, R. Messing;

Physician Reviewers

— E. Ascher, P. Finn, R. Giugliano, J. Kirdar, D. Lee, A. Mirza, T. Rocco;

Biomarker Core Laboratory

(Brigham and Women’s Hospital, Boston)— D. Morrow, P. Ridker, E. Danielson, G. Borkowski;

Chlamydia/Serology Core Laboratory

(Johns Hopkins University, Baltimore) — T. Quinn;C. Gaydos, B. Wood;

Electrocardiographic Core Laboratory

(eResearch Technology, Philadelphia) — J. Morganroth;

Special Lipid Core Laboratory

(University of Pennsylvania, Philadelphia) — D. Rader, M. Wolfe;

Chemistry Core Laboratory

(LabCorp, Raritan, N.J.);

Data and Safety MonitoringBoard —

A. Gotto (Chair), J. Bartlett, D. DeMets, J. Banas, T. Pearson;

Clinical Centers Enrolling the Most Patients (in order of enrollment)

— Hunts-ville Hospital, Huntsville, Ala.: W. Haught and K. Griffin; Fremantle Hospital, Fremantle, Western Australia: R. Hendriks and D. Greenwell;Detar Hospital, Victoria, Tex.: H. Chandna and D. Holly; St. Francis Hospital, Tulsa, Okla.: J. Cassidy and N. Ritchie; Advanced Health In-stitute, Galax, Va.: J. Puma and E. Jones; Michigan Heart, Ypsilanti: J. Bengtson and C. Carulli; North Mississippi Medical Center, Tupelo: B.Bertolet and M. Jones; Wilford Hall Medical Center, Lackland Air Force Base, Lackland, Tex.: R. Krasuski and U. Ward; Queen ElizabethHospital, Woodville, Saskatchewan, Canada: J. Horowitz and R. Prideaux; Moses H. Cone Hospital, Greensboro, N.C.: T. Kelly and K.Cochran; Deaconess Medical Center, Spokane Wash.: D. Hollenbaugh and J. Mansfield; Louisiana Cardiology Associates, Baton Rouge: J.McLachlan and A. Yoches; New Mexico Heart Institute, Albuquerque: R. Orchard and S. Justice; River Cities Cardiology, Jeffersonville, Ind.:D. Denny and B. Vanvactor; Laval Hospital, Quebec, Canada: P. Bogaty and L. Boyer; Altru Health System Research Center, Grand Forks,



n engl j med

350;15

www.nejm.org april

8

,

2004

1504


N.D.: A. Ahmed and D. Vold; Iowa Heart Center, Des Moines: W. Wickmeyer and N. Coffman; MetroHealth Medical Center, Cleveland: R.Finkelhor and M. Dettmer; Scarborough Cardiology Research, Scarborough, Ont., Canada: A. Ricci, and B. Bozek; Moses Cone Hospital,Greensboro, N.C.: T. Stuckey and S. Milks; Lake Forest Hospital, Bannockburn, Ill.: J. Alexander and K. Anderson; Centre Hospitalier Uni-versitaire, Fleurimont, Que., Canada: S. LePage and L. Larrivee; Butterworth Hospital, Grand Rapids, Mich.: R. McNamara and B. Van Over;Centre Hospitalier Regional de Lanaudière, Quebec, Canada: S. Kouz and M. Roy; Craigavaon Area Hospital, Portadown, County Armagh,United Kingdom: A. Moriarty and D. McEneaney.

references

1.

Scandinavian Simvastatin Survival StudyGroup. Randomised trial of cholesterol low-ering in 4 444 patients with coronary heartdisease: the Scandinavian Simvastatin Sur-vival Study (4S). Lancet 1994;344:1383-9.

2.

Sacks RM, Pfeffer MA, Moye LA, et al.The effect of pravastatin on coronary eventsafter myocardial infarction in patients withaverage cholesterol levels. N Engl J Med1996;335:1001-9.

3.

The Long-Term Intervention with Prava-statin in Ischaemic Disease (LIPID) StudyGroup. Prevention of cardiovascular eventsand death with pravastatin in patients withcoronary heart disease and a broad range ofinitial cholesterol levels. N Engl J Med 1998;339:1349-57.

4.

Heart Protection Study CollaborativeGroup. MRC/BHF Heart Protection Study ofcholesterol lowering with simvastatin in20 536 high-risk individuals: a randomisedplacebo-controlled trial. Lancet 2002;360:7-22.

5.

Shepard J, Cobbe SM, Ford I, et al. Pre-vention of coronary heart disease with prav-astatin in men with hypercholesterolemia.N Engl J Med 1995;333:1301-7.

6.

Downs JR, Clearfield M, Weis S, et al.Primary prevention of acute coronary eventswith lovastatin in men and women with aver-age cholesterol levels: results of AFCAPS/TexCAPS: Air Force/Texas Coronary Athero-sclerosis Prevention Study. JAMA 1998;279:1615-22.

7.

Shepherd J, Blauw GJ, Murphy MB, et al.Pravastatin in elderly individuals at risk ofvascular disease (PROSPER): a randomisedcontrolled trial. Lancet 2002;360:1623-30.

8.

Expert Panel on Detection, Evaluation,and Treatment of High Blood Cholesterol inAdults. Executive summary of the ThirdReport of the National Cholesterol Educa-tion Program (NCEP) Expert Panel on Detec-tion, Evaluation, and Treatment of HighBlood Cholesterol in Adults (Adult Treat-ment Panel III). JAMA 2001;285:2486-97.

9.

De Backer G, Ambrosioni E, Borch-Johnsen K, et al. European guidelines on car-diovascular disease and prevention in clinicalpractice. Atherosclerosis 2003;171:145-55.

10.

Davidson MH, Nawrocki JW, Weiss SR,et al. Effectiveness of atorvastatin for reduc-

ing low-density lipoprotein cholesterol toNational Cholesterol Education Programtreatment goals. Am J Cardiol 1997;80:347-8.

11.

Cannon CP, McCabe CH, Belder R,Breen J, Braunwald E. Design of the Prava-statin or Atorvastatin Evaluation and Infec-tion Therapy (PROVE IT)-TIMI 22 trial. Am JCardiol 2002;89:860-1.

12.

Cannon CP, McCabe CH, Wilcox RG, etal. Oral glycoprotein IIb/IIIa inhibition withorbofiban in patients with unstable coro-nary syndromes (OPUS-TIMI 16) trial. Cir-culation 2000;102:149-56.

13.

Cannon CP, Battler A, Brindis RG, et al.American College of Cardiology key dataelements and definitions for measuring theclinical management and outcomes ofpatients with acute coronary syndromes: areport of the American College of Cardiol-ogy Task Force on Clinical Data Standards(Acute Coronary Syndromes Writing Com-mittee). J Am Coll Cardiol 2001;38:2114-30.

14.

Lachin JM, Foulkes MA. Evaluation ofsample size and power for analyses of sur-vival with allowance for nonuniform patiententry, losses to follow-up, noncompliance,and stratification. Biometrics 1986;42:507-19.

15.

Bonetti PO, Lerman LO, Napoli C, Ler-man A. Statin effects beyond lipid lowering— are they clinically relevant? Eur Heart J2003;24:225-48.

16.

Schwartz GG, Olsson AG, EzekowitzMD, et al. Effects of atorvastatin on earlyrecurrent ischemic events in acute coronarysyndromes: the MIRACL study: a random-ized controlled trial. JAMA 2001;285:1711-8.

17.

Stenestrand U, Wallentin L. Early statintreatment following acute myocardial infarc-tion and 1-year survival. JAMA 2001;285:430-6.

18.

Aronow HD, Topol EJ, Roe MT, et al.Effect of lipid-lowering therapy on early mor-tality after acute coronary syndromes: anobservational study. Lancet 2001;357:1063-8.

19.

Asakura M, Ueda Y, Yamaguchi O, et al.Extensive development of vulnerable plaquesas a pan-coronary process in patients withmyocardial infarction: an angioscopic study.J Am Coll Cardiol 2001;37:1284-8.

20.

Goldstein JA, Demetriou D, Grines CL,Pica M, Shoukfeh M, O’Neill WW. Multiplecomplex coronary plaques in patients withacute myocardial infarction. N Engl J Med2000;343:915-22.

21.

Braunwald E, Antman EM, Beasley JW,et al. ACC/AHA guideline update for themanagement of patients with unstableangina and non-ST-segment elevation myo-cardial infarction — 2002: summary article:a report of the American College of Cardiol-ogy/American Heart Association Task Forceon Practice Guidelines (Committee on theManagement of Patients with UnstableAngina). Circulation 2002;106:1893-900.

22.

Nissen SE, Tuzcu EM, Schoenhagen P,et al. Effect of intensive compared with mod-erate lipid-lowering therapy on progressionof coronary atherosclerosis. JAMA 2004;291:1071-80.

23.

Campeau L, Hunninghake DB, Knat-terud GL, et al. Aggressive cholesterol low-ering delays saphenous vein graft athero-sclerosis in women, the elderly, and patientswith associated risk factors: NHLBI postcoronary artery bypass graft clinical trial.Circulation 1999;99:3241-7.

24.

Taylor AJ, Kent SM, Flaherty PJ, CoyleLC, Markwood TT, Vernalis MN. ARBITER:Arterial Biology for the Investigation of theTreatment Effects of Reducing Cholesterol:a randomized trial comparing the effects ofatorvastatin and pravastatin on carotidintima medial thickness. Circulation 2002;106:2055-60.

25.

Kent SM, Coyle LC, Flaherty PJ, Mark-wood TT, Taylor AJ. Marked low-density lipo-protein cholesterol reduction below currentNational Cholesterol Education Program tar-gets provides the greatest reduction in carotidatherosclerosis. Clin Cardiol 2004;27:17-21.

26.

Smilde TJ, van Wissen S, WollersheimH, Trip MD, Kastelein JJ, Stalenhoef AF.Effect of aggressive versus conventional lipidlowering on atherosclerosis progression infamilial hypercholesterolaemia (ASAP): aprospective, randomised, double-blind trial.Lancet 2001;357:577-81.

27.

LaRosa JC. New and emerging datafrom clinical trials of statins. Curr Athero-scler Rep 2004;6:12-9.




Volume 341 Number 10

·

709

The New England

Journal

of

Medicine

© Copyr ight, 1999, by the Massachusetts Medical Society

VOLUME 341

S

EPTEMBER

2, 1999

NUMBER 10

THE EFFECT OF SPIRONOLACTONE ON MORBIDITY AND MORTALITYIN PATIENTS WITH SEVERE HEART FAILURE

B

ERTRAM

P

ITT

, M.D., F

AIEZ

Z

ANNAD

, M.D., W

ILLEM

J. R

EMME

, M.D., R

OBERT

C

ODY

, M.D., A

LAIN

C

ASTAIGNE

, M.D., A

LFONSO

P

EREZ

, M.D., J

OLIE

P

ALENSKY

, M.S.,

AND

J

ANET

W

ITTES

, P

H

.D.,

FOR

THE

R

ANDOMIZED

A

LDACTONE

E

VALUATION

S

TUDY

I

NVESTIGATORS

*

A

BSTRACT

Background and Methods

Aldosterone is importantin the pathophysiology of heart failure. In a double-blind study, we enrolled 1663 patients who had se-vere heart failure and a left ventricular ejection frac-tion of no more than 35 percent and who were beingtreated with an angiotensin-converting–enzyme inhib-itor, a loop diuretic, and in most cases digoxin. A totalof 822 patients were randomly assigned to receive 25mg of spironolactone daily, and 841 to receive place-bo. The primary end point was death from all causes.

Results

The trial was discontinued early, after amean follow-up period of 24 months, because an in-terim analysis determined that spironolactone wasefficacious. There were 386 deaths in the placebogroup (46 percent) and 284 in the spironolactonegroup (35 percent; relative risk of death, 0.70; 95 per-cent confidence interval, 0.60 to 0.82; P<0.001). This30 percent reduction in the risk of death among pa-tients in the spironolactone group was attributed toa lower risk of both death from progressive heartfailure and sudden death from cardiac causes. Thefrequency of hospitalization for worsening heartfailure was 35 percent lower in the spironolactonegroup than in the placebo group (relative risk of hos-pitalization, 0.65; 95 percent confidence interval, 0.54to 0.77; P<0.001). In addition, patients who receivedspironolactone had a significant improvement in thesymptoms of heart failure, as assessed on the basisof the New York Heart Association functional class(P<0.001). Gynecomastia or breast pain was reportedin 10 percent of men who were treated with spirono-lactone, as compared with 1 percent of men in theplacebo group (P<0.001). The incidence of serioushyperkalemia was minimal in both groups of patients.

Conclusions

Blockade of aldosterone receptorsby spironolactone, in addition to standard therapy,substantially reduces the risk of both morbidity anddeath among patients with severe heart failure.(N Engl J Med 1999:341:709-17.)

©1999, Massachusetts Medical Society.

From the Department of Internal Medicine, Division of Cardiology,University of Michigan, Ann Arbor (B.P., R.C.); the Centre d’Investiga-tion, Clinique de Nancy, Nancy, France (F.Z.); STICARES, CardiovascularResearch Foundation, Rotterdam, the Netherlands (W.J.R.); the Service deCardiologie, Hôpital Henri Mondor, Creteil, France (A.C.); Global Medi-cal Operations, Searle, Skokie, Ill. (A.P.); and the Statistics Collaborative,Washington, D.C. (J.P., J.W.). Address reprint requests to Dr. Pitt at theDivision of Cardiology, University of Michigan Medical Center, 3910Taubman, 1500 E. Medical Center Dr., Ann Arbor, MI 48109-0366.

*Other investigators are listed in the Appendix.

LDOSTERONE has an important role inthe pathophysiology of heart failure.

1-4

Al-dosterone promotes the retention of sodi-um, the loss of magnesium and potassium,

sympathetic activation, parasympathetic inhibition,myocardial and vascular fibrosis, baroreceptor dys-function, and vascular damage and impairs arterialcompliance.

4-8

Many physicians have assumed thatinhibition of the renin–angiotensin–aldosterone sys-tem by an angiotensin-converting–enzyme (ACE)inhibitor will suppress the formation of aldosterone.In addition, treatment with an aldosterone-receptorblocker in conjunction with an ACE inhibitor hasbeen considered relatively contraindicated becauseof the potential for serious hyperkalemia.

9,10

Consequently, aldosterone-receptor blockers are

used infrequently in patients with heart failure.

11,12

There is increasing evidence to suggest, however,that ACE inhibitors only transiently suppress theproduction of aldosterone.

7,13-16

Furthermore, treat-ment with the aldosterone-receptor blocker spirono-lactone at a daily dose of 12.5 to 25 mg in conjunc-tion with standard doses of an ACE inhibitor, a loopdiuretic, and in most cases digoxin is pharmacolog-ically effective and well tolerated, decreases atrial na-triuretic peptide concentrations, and does not leadto serious hyperkalemia (defined as a serum potas-

A



710

·

September 2, 1999

The New England Journal of Medicine

sium concentration of at least 6.0 mmol per liter).

17

On the basis of this information, we designed theRandomized Aldactone Evaluation Study (RALES)to test the hypothesis that daily treatment with25 mg of spironolactone would significantly reducethe risk of death from all causes among patientswho had severe heart failure as a result of systolicleft ventricular dysfunction and who were receivingstandard therapy, including an ACE inhibitor, if tol-erated.

METHODS

Patients

Patients were eligible for enrollment if they had had New YorkHeart Association (NYHA) class IV heart failure within the sixmonths before enrollment and were in NYHA class III or IV atthe time of enrollment, had been given a diagnosis of heart failureat least six weeks before enrollment, were being treated with anACE inhibitor (if tolerated) and a loop diuretic, and had a leftventricular ejection fraction of no more than 35 percent withinthe six months before enrollment (with no clinically significantintercurrent event). Treatment with digitalis and vasodilators wasallowed, but potassium-sparing diuretics were not permitted.Oral potassium supplements were not recommended unless hy-pokalemia (defined as a serum potassium concentration of lessthan 3.5 mmol per liter) developed.

Patients were excluded from the study if they had primary op-erable valvular heart disease (other than mitral or tricuspid regur-gitation with clinical symptoms due to left ventricular systolicheart failure), congenital heart disease, unstable angina, primaryhepatic failure, active cancer, or any life-threatening disease (otherthan heart failure). Patients who had undergone heart transplan-tation or were awaiting the procedure were also ineligible. Othercriteria for exclusion were a serum creatinine concentration ofmore than 2.5 mg per deciliter (221 µmol per liter) and a serumpotassium concentration of more than 5.0 mmol per liter. The in-stitutional review boards or ethics committees of all participatinginstitutions approved the protocol, and all patients gave writteninformed consent.

Procedures

After the initial evaluation, patients were randomly assigned ina double-blind fashion to receive either 25 mg of spironolactone(Aldactone, Searle, Skokie, Ill.) once daily or a matching placebo.After eight weeks of treatment, the dose could be increased to 50mg once daily if the patient showed signs or symptoms of pro-gression of heart failure without evidence of hyperkalemia. If hy-perkalemia developed at any time, the dose could be decreased to25 mg every other day; however, the investigator was encouragedfirst to adjust the doses of concomitant medications. Follow-upevaluations and laboratory measurements, including measure-ments of serum potassium, were conducted every 4 weeks for thefirst 12 weeks, then every 3 months for up to 1 year and every6 months thereafter until the end of the study. Additional clinicallaboratory tests were also performed at weeks 1 and 5. Serum po-tassium was also measured at week 9 in patients for whom thedose was increased to 50 mg. Study medication could be with-held in the event of serious hyperkalemia, a serum creatinine con-centration of more than 4.0 mg per deciliter (354 µmol per liter),intercurrent illness, or any condition in which such a course wasdeemed medically necessary to protect the patient’s best interests.However, all patients remained in the study so that we could trackhospitalizations and deaths.

An independent data and safety monitoring board periodicallyreviewed the results in a blinded fashion. Event committees whosemembers were unaware of the patients’ treatment assignments as-sessed the causes of death and reasons for hospitalization.

End Points

The primary end point of the study was death from any cause.Secondary end points included death from cardiac causes, hospi-talization for cardiac causes, the combined incidence of deathfrom cardiac causes or hospitalization for cardiac causes, and achange in the NYHA class. The effect of spironolactone was alsoassessed with the use of six prerandomization variables: left ven-tricular ejection fraction, the cause of heart failure, the serum cre-atinine concentration, age, the use of ACE inhibitors, and the useof digitalis.


The analysis of death from all causes (the primary end point)included all patients, according to the intention-to-treat principle.Kaplan–Meier

18

methods were used to construct cumulative sur-vival curves for the two groups. The primary comparison be-tween the two groups was based on a log-rank test.

19

Cox pro-portional-hazards regression models

20

were developed to explorethe effects of base-line variables on the estimated effect of spirono-lactone. Formal assessment of efficacy used a group sequentialmonitoring plan with a Lan–DeMets

21

stopping boundary andan O’Brien–Fleming

22

spending function.The sample size was calculated on the basis of the following

assumptions: the mortality rate in the placebo group would be38 percent, the risk of death would be 17 percent lower in thespironolactone group than in the placebo group, and approxi-mately 5 percent of the patients in the spironolactone groupwould discontinue treatment during each year of the study.

23

Thepower of the study to detect a difference between treatmentgroups was set at 90 percent (with a two-tailed

a

level of 0.05).At each of its meetings, the data and safety monitoring board

evaluated the available data for evidence of efficacy and safety andcalculated the cumulative type I error with respect to efficacy. Intwo large studies of patients with heart failure,

24,25

the distributionsof the time to death were nonexponential; therefore, the compu-tations for group sequential monitoring of mortality from all caus-es were based on life-table calculations of event rates. The criticalz value required to establish that treatment with spironolactonewas efficacious was 2.02, corresponding to a P value of 0.043.

RESULTS

Randomization was begun on March 24, 1995;recruitment was completed on December 31, 1996,with follow-up scheduled to continue through De-cember 31, 1999. However, at the fifth planned in-terim analysis, the observed effect of spironolactoneon the risk of death from all causes exceeded theprespecified critical z value. Hence, the trial wasstopped on August 24, 1998, after a mean follow-upof 24 months, on the recommendation of the dataand safety monitoring board. The analysis includesall events through midnight on August 24, 1998.

A total of 1663 patients from 195 centers in 15countries underwent randomization: 841 were as-signed to receive placebo and 822 were assigned toreceive spironolactone. As shown in Table 1, the twogroups had similar characteristics at base line. Sevenpatients (three in the placebo group and four in thespironolactone group) who had a history of NYHAclass IV heart failure were in NYHA class II at thetime of randomization. During the study, 414 patients(200 in the placebo group and 214 in the spirono-lactone group) discontinued treatment because of alack of response, because of adverse events, or for



THE EFFECT OF SPIRONOLACTONE ON MORBIDITY AND MORTALITY IN PATIENTS WITH SEVERE HEART FAILURE


·

711

administrative reasons. Treatment was stopped in anadditional 19 patients (11 in the placebo group and8 in the spironolactone group) because of the needfor heart transplantation; 2 patients, both of whomwere in the placebo group, died after heart trans-plantation. Patients who discontinued treatment werefollowed by means of regularly scheduled telephonecalls to determine their vital status. After 24 monthsof follow-up, the mean daily dose of study medica-tion for the patients who continued to receive treat-ment was 31 mg in the placebo group and 26 mg inthe spironolactone group.

Survival

There were 386 deaths in the placebo group (46percent) and 284 deaths in the spironolactone group(35 percent), representing a 30 percent reduction inthe risk of death (relative risk of death among thepatients in the spironolactone group, 0.70 by a Cox

proportional-hazards model; 95 percent confidenceinterval, 0.60 to 0.82; P<0.001) (Fig. 1 and Table2). A total of 314 deaths in the placebo group (37percent) and 226 deaths in the spironolactone group(27 percent) were attributed to cardiac causes, rep-resenting a 31 percent reduction in the risk of deathfrom cardiac causes (relative risk, 0.69; 95 percentconfidence interval, 0.58 to 0.82; P<0.001). The re-duction in the risk of death among the patients in thespironolactone group was attributed to significantlylower risks of both death from progressive heart fail-ure and sudden death from cardiac causes (Table 2).

The reduction in the risk of death among patientsin the spironolactone group was similar in analysesof all six prespecified subgroups as well as in retro-spective analyses performed according to sex, NYHAclass, base-line serum potassium concentration, useof potassium supplements, and use of beta-blockers(Fig. 2). The estimated beneficial effect was similaracross geographic regions.

Death from Cardiac Causes and Hospitalization for Cardiac Causes

During the trial, 336 patients in the placebo groupand 260 patients in the spironolactone group werehospitalized at least once for cardiac reasons (Table2). In total, there were 753 hospitalizations for car-diac causes in the placebo group and 515 in thespironolactone group, representing a 30 percent re-duction in the risk of hospitalization for cardiac caus-es among patients in the spironolactone group (rel-ative risk, 0.70; 95 percent confidence interval, 0.59to 0.82; P<0.001) (Table 2). Analysis of the com-bined end point of death from cardiac causes or hos-pitalization for cardiac causes revealed a 32 percentreduction in the risk of this end point among patientsin the spironolactone group as compared with thosein the placebo group (relative risk, 0.68; 95 percentconfidence interval, 0.59 to 0.78; P<0.001) (Table 3).

Changes in NYHA Class

Three categories were used to assess changes in thesymptoms of heart failure: improvement, no change,and worsening or death. The condition of patientswho were in NYHA class III at base line was consid-ered to have improved if they were in NYHA class Ior II at the end of the study and considered to haveworsened if they were in NYHA class IV (or haddied). The condition of patients who were in NYHAclass IV at base line was considered to have improvedif they were in NYHA class I, II, or III at the endof the study; other patients in NYHA class IV atbase line either had no change at the end of thestudy or died. In the placebo group, the conditionof 33 percent of the patients improved; it did notchange in 18 percent, and it worsened in 48 percent.In the spironolactone group, the condition of 41percent of the patients improved; it did not change

*Plus–minus values are means ±SD. ACE denotes angiotensin-convert-ing enzyme.

†The ejection fraction could be measured by contrast ventriculography,gated radionuclide ventriculography, or echocardiography.

‡The cause of heart failure was determined on the basis of a patient’shistory, angiographic evidence, or both. Data on the cause of heart failurewere not available for two patients in the placebo group.

T

ABLE

1.

B

ASE

-L

INE

C

HARACTERISTICS

OF

THE

P

ATIENTS

.*

C

HARACTERISTIC

P

LACEBO

G

ROUP

(N=841)

S

PIRONOLACTONE

G

ROUP

(N=822)

Age — yr 65±12 65±12

White race — % 86 87

Sex — no. (%)MaleFemale

614 (73)227 (27)

603 (73)219 (27)

Blood pressure — mm HgSystolicDiastolic

122±2075±11

123±2175±12

Heart rate — beats/min 81±15 81±14

New York Heart Association class — no. (%)

IIIIIIV

3 (0.4)581 (69)257 (31)

4 (0.5)592 (72)226 (27)

Left ventricular ejection fraction — %† 25.2±6.8 25.6±6.7

Cause of heart failure — no. (%)‡IschemicNonischemic

453 (54)386 (46)

454 (55)368 (45)

Medications — %Loop diureticsACE inhibitorsDigitalisAspirinPotassium supplementsBeta-blockers

1009472372710

1009575362911

Mean dose of ACE inhibitors — mg/day

CaptoprilEnalaprilLisinopril

62.116.513.1

63.413.515.5



712

·

September 2, 1999


in 21 percent, and it worsened in 38 percent. Thedifference between groups was significant (P<0.001by the Wilcoxon test).

Safety

There were no significant differences between thetwo groups in serum sodium concentration, bloodpressure, or heart rate during the study. The medi-an creatinine and potassium concentrations did notchange in the placebo group during the first year offollow-up, the period for which the data were mostcomplete. During the same period, however, themedian creatinine concentration in the spironolac-tone group increased by approximately 0.05 to 0.10mg per deciliter (4 to 9 µmol per liter) and the me-dian potassium concentration increased by 0.30 mmolper liter. The differences between the two groupswere significant (P<0.001) but were not clinically im-portant.

Table 4 lists the adverse reactions in the two groups.Serious hyperkalemia occurred in 10 patients in the

placebo group (1 percent) and 14 patients in thespironolactone group (2 percent, P=0.42). Gyneco-mastia or breast pain was reported by 10 percent ofthe men in the spironolactone group and 1 percentof the men in the placebo group (P<0.001), causingmore patients in the spironolactone group than in theplacebo group to discontinue treatment (10 vs. 1,P=0.006).

DISCUSSION

We found that treatment with spironolactone re-duced the risk of death from all causes, death fromcardiac causes, hospitalization for cardiac causes, andthe combined end point of death from cardiac caus-es or hospitalization for cardiac causes among pa-tients who had severe heart failure as a result of leftventricular systolic dysfunction and who were receiv-ing standard therapy including an ACE inhibitor.Spironolactone also improved the symptoms of heartfailure, as measured by changes in the NYHA func-tional class. The reductions in the risk of death and

Figure 1.

Kaplan–Meier Analysis of the Probability of Survival among Patients in the Placebo Group and Patientsin the Spironolactone Group.The risk of death was 30 percent lower among patients in the spironolactone group than among patients in theplacebo group (P<0.001).

0.00

1.00

0 36

0.45

0.50

0.55

0.60

0.65

0.70

0.75

0.80

0.85

0.90

0.95

3 6 9 12 15 18 21 24 27 30 33

Months

Placebo

Spironolactone

NO. AT RISK

PlaceboISpironolactone

841I822

775I766

723I739

678I698

628I669

592I639

565I608

483I526

379I419

280I316

179I193

92I122

36I43

Pro

bab

ility

of

Su

rviv

al





·

713

hospitalization were observed after 2 to 3 months oftreatment and persisted throughout the study (meanfollow-up, 24 months). The results were consistentamong subgroups. Serious hyperkalemia requiringthe discontinuation of treatment was uncommon,occurring in one patient in the placebo group andthree in the spironolactone group.

The patients in our study were at higher risk thanthose in studies of the effects of bisoprolol,

26

digox-in,

27

amlodipine,

28

or carvedilol

29

on heart failure re-sulting from systolic left ventricular dysfunction andtreated with standard therapy, including an ACE in-hibitor, but they were at lower risk than patients in astudy of the effects of enalapril.

25

The reduction inthe risk of death with spironolactone treatment wasdue to significant decreases in the risk of both deathfrom progressive heart failure and sudden death fromcardiac causes. These results are consistent with thecurrent understanding of the effect of aldosterone inpatients with heart failure.

30-32

Aldosterone was originally thought to be impor-

tant in the pathophysiology of heart failure only be-cause of its ability to increase sodium retention andpotassium loss. However, in the past several years,research has shown that aldosterone also causes my-ocardial and vascular fibrosis,

33,34

direct vascular dam-

age,

8

and baroreceptor dysfunction

6

and prevents theuptake of norepinephrine by myocardium.

4,32

Thereduction in the risk of death in our study does notappear to be due entirely to an effect of spironolac-tone on sodium retention or potassium loss; instead,it is likely that spironolactone is also cardioprotec-tive. In our previous dose-finding study,

17

a dose of25 mg of spironolactone daily had no apparent di-uretic effect — that is, there was no change in totalbody weight, the sodium-retention score, or urinarysodium excretion. In the present study, spironolac-tone (mean dose, 26 mg daily) did not have a clini-cally significant hemodynamic effect. Although wecannot rule out the possibility that spironolactonehad some effect on sodium excretion in the presentstudy, this effect would most likely be minor, ascompared with the effect of the high doses of loopdiuretics used. Also, although there was a significantincrease from base line in serum potassium concen-trations in the patients in the spironolactone group,this change was not clinically important.

The 35 percent reduction in the risk of hospital-ization for worsening heart failure may be attribut-able to the ability of spironolactone to reduce myo-cardial and vascular fibrosis. Although the exact causeof the reduction in the risk of death in our study re-

*CI denotes confidence interval.

†This category includes death due to worsening heart failure (defined as increasing symptoms orsigns requiring an increase in treatment).

‡This category includes witnessed death from cardiac causes heralded by abrupt loss of conscious-ness within one hour after the onset of symptoms in a patient in whom death was unexpected.

§Some patients were hospitalized for more than one cardiac cause.

T

ABLE

2.

R

ELATIVE

R

ISKS

OF

D

EATH

AND

H

OSPITALIZATION

.

V

ARIABLE

P

LACEBO

G

ROUP

(N=841)

S

PIRONOLACTONE

G

ROUP

(N=822)R

ELATIVE

R

ISK

(95% CI)* P V

ALUE

no. of patients

Cause of death

Cardiac causesProgression of heart failure†Sudden death‡Myocardial infarction

31418911015

2261278217

0.69 (0.58–0.82)0.64 (0.51–0.80)0.71 (0.54–0.95)

<0.001<0.001

0.02

Other cardiovascular causes 13 12Stroke 11 8Noncardiovascular causes 41 29Unknown 7 9Total 386 284 0.70 (0.60–0.82) <0.001

no. of patients/no. of events

Reason for hospitalization

Cardiac causes§Worsening heart failureAnginaVentricular arrhythmiasMyocardial infarction

336/753300/66335/4424/3114/15

260/515215/41343/6623/2510/11

0.70 (0.59–0.82)0.65 (0.54–0.77)

<0.001<0.001

Other cardiovascular causes 112/163 117/169Stroke 20/24 14/15Noncardiovascular causes 232/377 223/361



714

·

September 2, 1999


mains speculative, we postulate that an aldosterone-receptor blocker can prevent progressive heart fail-ure by averting sodium retention and myocardial fi-brosis and prevent sudden death from cardiac causesby averting potassium loss and by increasing the my-ocardial uptake of norepinephrine. Spironolactonemay prevent myocardial fibrosis by blocking theeffects of aldosterone on the formation of colla-gen,

5,35,36

which in turn could play a part in reducingthe risk of sudden death from cardiac causes, sincemyocardial fibrosis could predispose patients to vari-ations in ventricular-conduction times and, hence,to reentry ventricular arrhythmias.

32,35-37

Few patients (11 percent) in the spironolactone

Figure 2.

Relative Risks of Death from All Causes and According to Demographic and Clinical Characteristics.The horizontal lines indicate 95 percent confidence intervals. LV denotes left ventricular, ACE angiotensin-con-verting enzyme, and NYHA New York Heart Association. To convert values for creatinine to micromoles per liter,multiply by 88.4.

Death from all causes

Median ageI<67 yrI»67 yr

LV ejection fractionI<26%I»26%

Cause of heart failureINonischemicI

Ischemic

Median creatinineI<1.2 mg/dlI»1.2 mg/dl

Use of digitalisINoI

Yes

Use of ACE inhibitorINoI

Yes

SexIFemaleI

Male

Median potassiumI<4.2 mmol/literI»4.2 mmol/liter

NYHA classIIIIIIV

Beta-blocker useINoI

Yes

Use of potassium supplementsINoI

Yes

0.2 1.40.3 0.4 0.5 0.6 0.7 0.8 0.9 1.0 1.1 1.2 1.3

Spironolactone Better Placebo Better

*Each analysis represents the time to the first occurrence of an event.For patients with both events, the analysis includes only the first event. CIdenotes confidence interval.

T

ABLE

3.

R

ELATIVE RISKS OF THE COMBINED END POINTS OF DEATH OR HOSPITALIZATION IN THE

SPIRONOLACTONE GROUP.*

END POINT

RELATIVE RISK

(95% CI)P

VALUE

Death from cardiac causes orhospitalization for cardiac causes

0.68 (0.59–0.78) <0.001

Death from any cause or hospitalizationfor any reason

0.77 (0.68–0.86) <0.001

Death from any cause or hospitalization for cardiac causes

0.68 (0.60–0.77) <0.001




Volume 341 Number 10 · 715

group were receiving a beta-blocker at base line, andthe reduction in the risk of death did not differ sig-nificantly between those who were treated with abeta-blocker and those who were not so treated.Since our patients were at higher risk than patientswho were evaluated in recent studies of beta-block-ers in heart failure,26,29 studies are needed to exam-ine both the tolerability and the effectiveness ofbeta-blockers in such a high-risk population as wellas the effects of the concomitant use of an aldoste-rone-receptor blocker and a beta-blocker.

Our finding that an aldosterone-receptor blockerreduced the risk of both morbidity and death amongpatients who were receiving an ACE inhibitor em-phasizes the point that standard doses of an ACEinhibitor do not effectively suppress the productionof aldosterone.7,14 Although higher doses of ACE in-hibitors may be more effective than lower doses inreducing the risk of morbidity and death among pa-

tients with heart failure,38 there is no evidence thathigher doses suppress aldosterone production moreeffectively in the long term. ACE inhibitors cannottotally suppress the production of aldosterone, be-cause other factors in addition to angiotensin II (e.g.,serum potassium) are important in the productionof aldosterone and may override the effects of angi-otensin II.39-41 Since aldosterone remains in the cir-culation, only the presence of an aldosterone-recep-tor blocker will completely suppress the effects ofthis hormone.

The fact that spironolactone significantly reducedthe risk of both morbidity and death among thehigh-risk patients in our study with only a very lowincidence of serious hyperkalemia can be attributedto our previous efforts in determining an effectiveand safe dose of spironolactone when used in con-junction with an ACE inhibitor.17 We found thatspironolactone at a dose of 12.5 to 25 mg daily waspharmacologically effective in blocking the aldoste-rone receptors and decreasing atrial natriuretic pep-tide concentrations and that serious hyperkalemiaoccurred most frequently with daily doses of 50 mgor greater.17 In the present study, therefore, spirono-lactone therapy was initiated at a daily dose of 25mg, and physicians were given the option of reduc-ing the dose to 25 mg every other day if serum po-tassium concentrations started to rise to a hyperka-lemic level or of increasing the dose to 50 mg dailyafter eight weeks in patients who had symptoms orsigns of worsening heart failure but no evidenceof hyperkalemia. It should be emphasized, however,that a serum creatinine concentration of more than2.5 mg per deciliter and a serum potassium concen-tration of more than 5.0 mmol per liter were exclu-sion criteria. In addition, the long-term use of agentsknown to interact with spironolactone, increase therisk of hyperkalemia, or do both was not allowed.Although potassium supplements were used by 29percent of the patients in the spironolactone group,the benefit of spironolactone in these patients wassimilar to that in patients who did not use potassiumsupplements.

Overall, spironolactone therapy was tolerated well:8 percent of the patients in the spironolactone groupdiscontinued treatment because of adverse events, ascompared with 5 percent of the patients in the pla-cebo group. This difference was due in part to a sig-nificant incidence of gynecomastia or breast painamong men in the spironolactone group (P<0.001).The rate of discontinuation of treatment because ofthis event was higher in the spironolactone group thanin the placebo group (2 percent vs. 0.2 percent,P=0.006). Gynecomastia has previously been ob-served in patients who were treated with spironolac-tone.42,43 Specifically, gynecomastia has been report-ed to occur in 6.9 percent of men who received dailydoses of spironolactone of 50 mg or less for hyper-

*P=0.17 for the comparison with the placebo group.

†There were 614 men in the placebo group and 603 in the spironolac-tone group.

‡P<0.001 for the comparison with the placebo group.

§P=0.006 for the comparison with the placebo group.

TABLE 4. ADVERSE EVENTS.

ADVERSE EVENT

PLACEBO

GROUP

(N=841)

SPIRONOLACTONE

GROUP

(N=822)

no. of patients (%)

One or more events 667 (79) 674 (82)*

Discontinuation because of adverse event

40 (5) 62 (8)

Cardiovascular disordersAnginaHeart failure

251 (30)83 (10)80 (10)

248 (30)103 (13)52 (6)

Respiratory tract disordersCoughDyspneaPneumoniaPulmonary edemaPleural effusion

285 (34)117 (14)39 (5)25 (3)7 (0.8)

11 (1)

262 (32)103 (13)34 (4)17 (2)5 (0.6)3 (0.4)

Metabolic and nutritional disordersHyperuricemia

215 (26)25 (3)

269 (33)16 (2)

Neoplasm 10 (1) 13 (2)

Urinary system disorders 89 (11) 99 (12)

Disorders of skin and appendages 72 (9) 73 (9)

Musculoskeletal disorders 118 (14) 101 (12)

Nervous system disorders 173 (21) 185 (23)

Psychiatric disorders 126 (15) 122 (15)

Gastrointestinal disorders 241 (29) 236 (29)

Endocrine disordersGynecomastia in men†Breast pain in men†Gynecomastia or breast pain

in men†

26 (3)8 (1)1 (0.1)9 (1)

84 (10)55 (9)‡10 (2)§61 (10)‡

Edema 21 (2) 18 (2)

Serious hyperkalemia 10 (1) 14 (2)



716 · September 2, 1999


tension.43 The use of a selective aldosterone-receptorantagonist such as eplerenone, which has a lower af-finity for androgen and progesterone receptors thandoes spironolactone,44 may minimize the risk of gy-necomastia. The risk of gynecomastia should not,however, be an argument against the use of spirono-lactone in men with severe heart failure, since spirono-lactone reduces the risk of both morbidity and death.The effectiveness and risks of treatment with spirono-lactone in patients at lower risk than those in ourstudy, such as those with less severe heart failure,will require further prospective study.

Our finding that an aldosterone-receptor antago-nist, when used in conjunction with an ACE inhib-itor, reduces the risk of both death from progressiveheart failure and sudden death from cardiac causescontributes to our understanding of the pathophys-iology of heart failure and has implications for thetreatment of patients with other conditions in whichACE inhibitors are beneficial, such as patients withhypertension and those who have had a myocardialinfarction.

Supported by a grant from Searle, Skokie, Ill.Preliminary data were presented at the American Heart Association

meeting, Dallas, November 8–11, 1998.

We are indebted to Lorraine R. Baer, Pharm.D., for editorial con-tributions and assistance in the preparation of the manuscript.

APPENDIX

In addition to the authors, the following persons participated in thestudy: Data and Safety Monitoring Board — D. Julian (chair), J.-P. Boissel,C. Furberg, H. Kulbertus, S. Pocock; Primary End-Point Committee — J.Blumenfeld, J.A. Ramires; Nonfatal Hospitalization End-Point Committee— S. Sasayama (chair), C. Brilla, D. Duprez, R. Muñoz; Medical Monitors— D. Asner, B. Roniker; Investigators: Belgium — P. Block, G. Boxho,J.-M. Chaudron, V. Conraads, J. Creplet, P. De Salle, F. Deman, D. Du-prez, O. Gurnee, G. Heyndrickx, S. Janssens, G. Jouret, C. Mortier, L.Pierard, P. Timmermans, J.L. Vandenbossche, W. Van Mieghem, J. Van-welden, J. Vincke; Brazil — F.M. Albanesi Filho, F.A. de Almeida, J.C.A.Ayoub, E.T. Barbosa, M. Batlouni, L.C. Bodanese, R.M. Carrasco, A.C. deC. Carvalho, I. Castro, O.R. Coelho, D. Dauar, C. Drumond Neto, G.S.Feitosa, R.A. Franken, P.C.B.V. Jardim, C. Mady, M.F. de C. Maranhão,J.A. Marin Neto, L.F. de Miranda, J.C. Nicolau, W. Oigman, W.A. de Ol-iveira Júnior, W.C. Pereira Filho, J.A.F. Ramires, J.J.F. Rapozo Filho, S.Rassi, J.M. Ribeiro, J.P. Ribeiro, P.R.F. Rossi, J.F.K. Saraiva, A.S. Sbissa,M.A.D. da Silva, J.E. de Sigueira, J. Souza Filho; Canada — I.M. Arnold,D. Beanlands, C. Koilppillai, S. Lepage, A. Morris, M. White; France — F.Albert, G. Amat, F. Apffel, M.C. Aumont, S. Baleynaud, P. Battistella, J.Beaune, L. Bonnefoy, A. Bonneau, J. Bonnet, M. Bory, J.P. Bousser, J.A.Boutarin, B. Charbonnier, A. Cohen, A. Cohen-Solal, M.T. Courbet-Andrejak, A. Cribier, F. Delahaye, C. D’Ivernois, J.P. Doazan, V. Dormagen,H. Douard, F. Dravet, A. Dutoit, J.M. Fayard, M. Ferriere, C. Fournier, Y.Frances, F. Funck, M. Galinier, L.F. Garnier, P. Gibelin, P. Gosse, B. Grivet,L. Guize, B. d’Hautefeuille, A. Heraudeau, J.F. Huret, L. Janin-Manificat,G. Jarry, Y. Jobic, E. Jullien, J.C. Kahn, K. Khalife, A. Koenig, F. Latour,C. Leclercq, F. Leclercq, L. Ledain, H. Le Marec, S. Levy, J.M. Mallion,P. Maribas, G. Mialet, P.L. Michel, J.P. Millet, B. Moquet, J.P. Normand,T. Olive, P. Poncelet, J. Ponsonnaille, J. Puel, A. Rifai, P. Sans, J.P. Simon,M. Toussaint, A. Verdun, B. Veyre, S. Werquin; Germany — C. Brilla, G.Riegger, M. Zehender; Japan — Y. Aizawa, M. Hori, H. Inoue, H.Kasanuki, A. Kitabatake, M. Matsuzaki, S. Ogawa, M. Omata, S. Sasayama,A. Takeshita, Y. Yazaki, M. Yokoyama; Mexico — L. Avila, F.J. Guerrero, H.Gutiérrez-Leonard, J.L. Leyva-Garza; the Netherlands — P.J.L.M. Bernink,H. Fintelman, J.A. Kragten, J.B.L. ten Kate, D.J.A. Lok, A.R. RamdatMisier, G.P. Molhoek, G.M.G. Paulussen, L.H.J. van Kempen, D.J. vanVeldhuisen, L.G.P.M. van Zeijl, A.J.A.M. Withagen; New Zealand — H.Ikram; South Africa — J.D. Marx, D.P. Naidoo; Spain — J.M. Aguirre, S.

Alcasena, M. Artaza, J. Azpitarte, J.R. Berrazueta, A. Castro-Beiras, P.Conthe, A. Cortina, J.M. Cruz-Fernández, J. Farré, I. Ferreira, M. García-Moll, V. López-García-de-Aranda, J.L. López-Sendón, R. Muñoz, F. Na-varro, J. Palomo, J.M. Ribera-Casado, J.L. Rodríguez-Lambert, J. Soler-Soler, E. de-Teresa, J.A. de-Velasco; Switzerland — P. Delafontaine, O.M.Hess, L. Kappenberger, G. Noll, W. Rutishauser, J. Sztajzel; United King-dom — A.J.S. Coats, T.S. Callaghan, A.D. Struthers; United States — G.W.Dec, P. Deedwania, J. Nicklas, K.T. Weber; Venezuela — N. Lopez, S. Waich.

REFERENCES

1. Dzau VJ, Colucci WS, Hollenberg NK, Williams GH. Relation of the renin-angiotensin-aldosterone system to clinical state in congestive heart failure. Circulation 1981;63:645-51.2. Swedberg K, Eneroth P, Kjekshus J, Wilhelmsen L. Hormones regulat-ing cardiovascular function in patients with severe congestive heart failure and their relation to mortality. Circulation 1990;82:1730-6.3. Weber KT, Villarreal D. Aldosterone and antialdosterone therapy in congestive heart failure. Am J Cardiol 1993;71:Suppl:3A-11A.4. Barr CS, Lang CC, Hanson J, Arnott M, Kennedy N, Struthers AD. Effects of adding spironolactone to an angiotensin-converting enzyme in-hibitor in chronic congestive heart failure secondary to coronary artery dis-ease. Am J Cardiol 1995;76:1259-65.5. MacFadyen RJ, Barr CS, Struthers AD. Aldosterone blockade reduces vascular collagen turnover, improves heart rate variability and reduces early morning rise in heart rate in heart failure patients. Cardiovasc Res 1997;35:30-4.6. Wang W. Chronic administration of aldosterone depresses baroreceptor reflex function in the dog. Hypertension 1994;24:571-5.7. Duprez DA, De Buyzere ML, Rietzschel ER, et al. Inverse relationship between aldosterone and large artery compliance in chronically treated heart failure patients. Eur Heart J 1998;19:1371-6.8. Rocha R, Chander PN, Khanna K, Zuckerman A, Stier CT Jr. Miner-alocorticoid blockade reduces vascular injury in stroke-prone hypertensive rats. Hypertension 1998;31:451-8.9. Capoten tablets: captopril tablets. In: Physicians’ desk reference. 52nd ed. Montvale, N.J.: Medical Economics, 1998:784-7.10. Vasotec tablets: enalapril maleate. In: Physicians’ desk reference. 52nd ed. Montvale, N.J.: Medical Economics, 1998:1771-4.11. Cleland JG, Swedberg K, Poole-Wilson PA. Successes and failures of current treatment of heart failure. Lancet 1998;352:Suppl 1:SI19-SI28.12. Cohn JN. The management of chronic heart failure. N Engl J Med 1996;335:490-8.13. Borghi C, Boschi S, Ambrosioni E, Melandri G, Branzi A, Magnani B. Evidence of a partial escape of renin-angiotensin-aldosterone blockade in patients with acute myocardial infarction treated with ACE inhibitors. J Clin Pharmacol 1993;33:40-5.14. Cleland JGF, Dargie HJ, Hodsman GP, et al. Captopril in heart failure: a double blind controlled trial. Br Heart J 1984;52:530-5.15. Biollaz J, Brunner HR, Gavras I, Waeber B, Gavras H. Antihyperten-sive therapy with MK 421: angiotensin II-renin relationships to evaluate ef-ficacy of converting enzyme blockade. J Cardiovasc Pharmacol 1982;4:966-72.16. Staessen J, Lijnen P, Fagard R, Verschueren LJ, Amery A. Rise in plas-ma concentration of aldosterone during long-term angiotensin II suppres-sion. J Endocrinol 1981;91:457-65.17. The RALES Investigators. Effectiveness of spironolactone added to an angiotensin-converting enzyme inhibitor and a loop diuretic for severe chronic congestive heart failure (the Randomized Aldactone Evaluation Study [RALES]). Am J Cardiol 1996;78:902-7.18. Kaplan EL, Meier P. Nonparametric estimation from incomplete ob-servations. J Am Stat Assoc 1958;53:457-81.19. Mantel N. Evaluation of survival data and two new rank order statistics arising in its consideration. Cancer Chemother Rep 1966;50:163-70.20. Cox DR. Regression models and life-tables. J R Stat Soc [B] 1972;34:187-202.21. Lan KKG, DeMets DL. Discrete sequential boundaries for clinical tri-als. Biometrika 1983;70:659-63.22. O’Brien PC, Fleming TR. A multiple testing procedure for clinical tri-als. Biometrics 1979;35:549-56.23. Lakatos E. Sample sizes based on the log-rank statistic in complex clin-ical trials. Biometrics 1988;44:229-41. [Erratum, Biometrics 1988;44:923.]24. The SOLVD Investigators. Effect of enalapril on survival in patients with reduced left ventricular ejection fractions and congestive heart failure. N Engl J Med 1991;325:293-302.25. The CONSENSUS Trial Study Group. Effects of enalapril on mortal-ity in severe congestive heart failure: results of the Cooperative North Scandinavian Enalapril Survival Study (CONSENSUS). N Engl J Med 1987;316:1429-35.




Volume 341 Number 10 · 717

26. The CIBIS-II Investigators and Committees. The Cardiac Insufficiency Bisoprolol Study II (CIBIS-II): a randomised trial. Lancet 1999;353:9-13.27. The Digitalis Investigation Group. The effect of digoxin on mortality and morbidity in patients with heart failure. N Engl J Med 1997;336:525-33.28. Packer M, O’Connor CM, Ghali JK, et al. Effect of amlodipine on morbidity and mortality in severe chronic heart failure. N Engl J Med 1996;335:1107-14.29. Packer M, Bristow MR, Cohn JN, et al. The effect of carvedilol on morbidity and mortality in patients with chronic heart failure. N Engl J Med 1996;334:1349-55.30. Zannad F. Angiotensin-converting enzyme inhibitor and spironolac-tone combination therapy: new objectives in congestive heart failure treat-ment. Am J Cardiol 1993;71:Suppl:34A-39A.31. Pitt B. “Escape” of aldosterone production in patients with left ven-tricular dysfunction treated with an angiotensin converting enzyme inhib-itor: implications for therapy. Cardiovasc Drugs Ther 1995;9:145-9.32. Struthers AD. Aldosterone escape during angiotensin-converting en-zyme inhibitor therapy in chronic heart failure. J Card Fail 1996;2:47-54.33. Weber KT, Brilla CG. Pathological hypertrophy and cardiac intersti-tium: fibrosis and renin-angiotensin-aldosterone system. Circulation 1991;83:1849-65.34. Brilla CG, Pick R, Tan LB, Janicki JS, Weber KT. Remodeling of the rat right and left ventricles in experimental hypertension. Circ Res 1990;67:1355-64.35. Klug D, Robert V, Swynghedauw B. Role of mechanical and hormonal factors in cardiac remodeling and the biologic limits of myocardial adapta-tion. Am J Cardiol 1993;71:Suppl:46A-54A.

36. Brilla CG, Matsubara LS, Weber KT. Anti-aldosterone treatment and the prevention of myocardial fibrosis in primary and secondary hyperaldos-teronism. J Mol Cell Cardiol 1993;25:563-75.37. Barr CS, Naas A, Freeman M, Lang CC, Struthers AD. QT dispersion and sudden unexpected death in chronic heart failure. Lancet 1994;343:327-9.38. Hobbs RE. Results of the ATLAS study — high or low doses of ACE inhibitors for heart failure? Cleve Clin J Med 1998;65:539-42.39. Hajnoczky G, Varnai P, Hollo Z, et al. Thapsigargin-induced increase in cytoplasmic Ca2+ concentration and aldosterone production in rat adre-nal glomerulosa cells: interaction with potassium and angiotensin-II. En-docrinology 1991;128:2639-44.40. Szalay KS, Beck M, Toth M, de Chatel R. Interactions between ouabain, atrial natriuretic peptide, angiotensin-II and potassium: effects on rat zona glomerulosa aldosterone production. Life Sci 1998;62:1845-52.41. Okubo S, Niimura F, Nishimura H, et al. Angiotensin-independent mechanism for aldosterone synthesis during chronic extracellular fluid vol-ume depletion. J Clin Invest 1997;99:855-60.42. Aldactone: spironolactone. In: Physicians’ desk reference. 53rd ed. Montvale, N.J.: Medical Economics, 1999:2928-9.43. Jeunemaitre X, Chatellier G, Kreft-Jais C, et al. Efficacy and tolerance of spironolactone in essential hypertension. Am J Cardiol 1987;60:820-5.44. de Gasparo M, Joss U, Ramjoué HP, et al. Three new epoxy-spirolac-tone derivatives: characterization in vivo and in vitro. J Pharmacol Exp Ther 1987;240:650-6.



n engl j med 365;10 nejm.org september 8, 2011 883

The new england journal of medicineestablished in 1812 september 8, 2011 vol. 365 no. 10

Rivaroxaban versus Warfarin in Nonvalvular Atrial FibrillationManesh R. Patel, M.D., Kenneth W. Mahaffey, M.D., Jyotsna Garg, M.S., Guohua Pan, Ph.D., Daniel E. Singer, M.D.,

Werner Hacke, M.D., Ph.D., Günter Breithardt, M.D., Jonathan L. Halperin, M.D., Graeme J. Hankey, M.D., Jonathan P. Piccini, M.D., Richard C. Becker, M.D., Christopher C. Nessel, M.D., John F. Paolini, M.D., Ph.D.,

Scott D. Berkowitz, M.D., Keith A.A. Fox, M.B., Ch.B., Robert M. Califf, M.D., and the ROCKET AF Steering Committee, for the ROCKET AF Investigators*

A bs tr ac t

From the Duke Clinical Research Institute (M.R.P., K.W.M., J.G., J.P.P., R.C.B.) and Duke Translational Medicine Institute (R.M.C.), Duke University Medical Center, Durham, NC; Johnson & Johnson Phar-maceutical Research and Development, Raritan (G.P., C.C.N.), and Bayer HealthCare Pharmaceuticals, Montville (J.F.P., S.D.B.) — both in New Jersey; Massachusetts Gen-eral Hospital and Harvard Medical School — both in Boston (D.E.S.); Ruprecht-Karls-University, Heidelberg (W.H.), and Hospi-tal of the University of Münster, Münster (G.B.) — both in Germany; the Cardiovas-cular Institute, Mount Sinai Medical Cen-ter, New York (J.L.H.); Royal Perth Hospi-tal, Perth, WA, Australia (G.J.H.); and the University of Edinburgh and Royal Infir-mary of Edinburgh — both in Edinburgh (K.A.A.F.). Address reprint requests to Dr. Patel at Duke Clinical Research Institute, Duke University Medical Center, Rm. 0311 Terrace Level, 2400 Pratt St., Durham, NC 27705, or at [email protected].

* A complete listing of the steering com-mittee members and trial investigators in the Rivaroxaban Once Daily Oral Di-rect Factor Xa Inhibition Compared with Vitamin K Antagonism for Preven-tion of Stroke and Embolism Trial in Atrial Fibrillation (ROCKET AF) is pro-vided in the Supplementary Appendix, available at NEJM.org.

This article (10.1056/NEJMoa1009638) was published on August 10, 2011, at NEJM.org.

N Engl J Med 2011;365:883-91.Copyright © 2011 Massachusetts Medical Society.

Background

The use of warfarin reduces the rate of ischemic stroke in patients with atrial fibril-lation but requires frequent monitoring and dose adjustment. Rivaroxaban, an oral factor Xa inhibitor, may provide more consistent and predictable anticoagulation than warfarin.

Methods

In a double-blind trial, we randomly assigned 14,264 patients with nonvalvular atrial fibrillation who were at increased risk for stroke to receive either rivaroxaban (at a daily dose of 20 mg) or dose-adjusted warfarin. The per-protocol, as-treated primary analysis was designed to determine whether rivaroxaban was noninferior to warfa-rin for the primary end point of stroke or systemic embolism.

Results

In the primary analysis, the primary end point occurred in 188 patients in the riva-roxaban group (1.7% per year) and in 241 in the warfarin group (2.2% per year) (hazard ratio in the rivaroxaban group, 0.79; 95% confidence interval [CI], 0.66 to 0.96; P<0.001 for noninferiority). In the intention-to-treat analysis, the primary end point occurred in 269 patients in the rivaroxaban group (2.1% per year) and in 306 patients in the warfarin group (2.4% per year) (hazard ratio, 0.88; 95% CI, 0.74 to 1.03; P<0.001 for noninferiority; P = 0.12 for superiority). Major and nonmajor clinically rel-evant bleeding occurred in 1475 patients in the rivaroxaban group (14.9% per year) and in 1449 in the warfarin group (14.5% per year) (hazard ratio, 1.03; 95% CI, 0.96 to 1.11; P = 0.44), with significant reductions in intracranial hemorrhage (0.5% vs. 0.7%, P = 0.02) and fatal bleeding (0.2% vs. 0.5%, P = 0.003) in the rivaroxaban group.

Conclusions

In patients with atrial fibrillation, rivaroxaban was noninferior to warfarin for the prevention of stroke or systemic embolism. There was no significant between-group difference in the risk of major bleeding, although intracranial and fatal bleeding occurred less frequently in the rivaroxaban group. (Funded by Johnson & Johnson and Bayer; ROCKET AF ClinicalTrials.gov number, NCT00403767.)



T h e n e w e ngl a nd j o u r na l o f m e dic i n e

n engl j med 365;10 nejm.org september 8, 2011884

A trial fibrillation is associated with an increase in the risk of ischemic stroke by a factor of four to five1 and ac-

counts for up to 15% of strokes in persons of all ages and 30% in persons over the age of 80 years.2 The use of vitamin K antagonists is highly effec-tive for stroke prevention in patients with nonval-vular atrial fibrillation and is recommended for persons at increased risk.3-5 However, food and drug interactions necessitate frequent coagulation monitoring and dose adjustments, requirements that make it difficult for many patients to use such drugs in clinical practice.6-8

Rivaroxaban is a direct factor Xa inhibitor that may provide more consistent and predictable anti-coagulation than warfarin.9,10 It has been reported to prevent venous thromboembolism more effec-tively than enoxaparin in patients undergoing orthopedic surgery11,12 and was noninferior to enoxaparin followed by warfarin in a study involv-ing patients with established venous thrombosis.13 This trial was designed to compare once-daily oral rivaroxaban with dose-adjusted warfarin for the prevention of stroke and systemic embolism in patients with nonvalvular atrial fibrillation who were at moderate-to-high risk for stroke.14

Me thods

Study Design and Oversight

The Rivaroxaban Once Daily Oral Direct Factor Xa Inhibition Compared with Vitamin K Antago-nism for Prevention of Stroke and Embolism Trial in Atrial Fibrillation (ROCKET AF) was a multi-center, randomized, double-blind, double-dummy, event-driven trial that was conducted at 1178 par-ticipating sites in 45 countries.14 The study was supported by Johnson & Johnson Pharmaceutical Research and Development and Bayer HealthCare. The Duke Clinical Research Institute coordinated the trial, managed the database, and performed the primary analyses independently of the spon-sors. Pertinent national regulatory authorities and ethics committees at participating centers ap-proved the protocol, which is available with the full text of this article at NEJM.org. The members of an international executive committee designed the trial, were responsible for overseeing the study’s conduct, retained the ability to independently ana-lyze and present the data, made the decision to submit the manuscript for publication, and take responsibility for the accuracy and completeness

of the data and all analyses. The first academic author wrote the initial draft of the manuscript.

Study Participants

We recruited patients with nonvalvular atrial fibril-lation, as documented on electrocardiography, who were at moderate-to-high risk for stroke. Elevated risk was indicated by a history of stroke, transient ischemic attack, or systemic embolism or at least two of the following risk factors: heart failure or a left ventricular ejection fraction of 35% or less, hypertension, an age of 75 years or more, or the presence of diabetes mellitus (i.e., a CHADS2 score of 2 or more, on a scale ranging from 1 to 6, with higher scores indicating a greater risk of stroke). According to the protocol, the proportion of pa-tients who had not had a previous ischemic stroke, transient ischemic attack, or systemic embolism and who had no more than two risk factors was limited to 10% of the cohort for each region; the remainder of patients were required to have had either previous thromboembolism or three or more risk factors. Complete inclusion and exclu-sion criteria are provided in the Supplementary Ap-pendix, available at NEJM.org. All patients provided written informed consent.

Study Treatment

Patients were randomly assigned to receive fixed-dose rivaroxaban (20 mg daily or 15 mg daily in patients with a creatinine clearance of 30 to 49 ml per minute) or adjusted-dose warfarin (target in-ternational normalized ratio [INR], 2.0 to 3.0). Patients in each group also received a placebo tab-let in order to maintain blinding. Randomization was performed with the use of a central 24-hour, computerized, automated voice-response system. A point-of-care device was used to generate en-crypted values that were sent to an independent study monitor, who provided sites with either real INR values (for patients in the warfarin group in order to adjust the dose) or sham values (for pa-tients in the rivaroxaban group receiving placebo warfarin) during the course of the trial. Sham INR results were generated by means of a validated algorithm reflecting the distribution of values in warfarin-treated patients with characteristics sim-ilar to those in the study population.15

It was intended that patients would continue to take the assigned therapy throughout the course of the trial, unless discontinuation was considered to be clinically indicated. Follow-up



Rivaroxaban vs. Warfarin in Atrial Fibrillation


procedures and restrictions on concomitant med-i cations are summarized in the Supplementary Appendix.

Outcomes

The primary efficacy end point was the compos-ite of stroke (ischemic or hemorrhagic) and sys-temic embolism. Brain imaging was recommend-ed to distinguish hemorrhagic from ischemic stroke. In the presence of atherosclerotic periph-eral arterial disease, the diagnosis of embolism required angiographic demonstration of abrupt arterial occlusion.

Secondary efficacy end points included a com-posite of stroke, systemic embolism, or death from cardiovascular causes; a composite of stroke, sys-temic embolism, death from cardiovascular causes, or myocardial infarction; and individual compo-nents of the composite end points. The principal safety end point was a composite of major and nonmajor clinically relevant bleeding events. Bleed-ing events involving the central nervous system that met the definition of stroke were adjudicated as hemorrhagic strokes and included in both the primary efficacy and safety end points. Other overt bleeding episodes that did not meet the criteria for major or clinically relevant nonmajor bleeding were classified as minor episodes.

An independent clinical end-point committee applied protocol definitions to adjudicate all sus-pected cases of stroke, systemic embolism, myo-cardial infarction, death, and bleeding events that contributed to the prespecified end points. De-tailed definitions of the end-point events are pro-vided in the Supplementary Appendix.


The primary hypothesis was that rivaroxaban would be noninferior to warfarin for the preven-tion of stroke or systemic embolism. The primary analysis was prespecified to be performed in the per-protocol population, which included all pa-tients who received at least one dose of a study drug, did not have a major protocol violation, and were followed for events while receiving a study drug or within 2 days after discontinuation (group A in Fig. 1 in the Supplementary Appendix).16-19

For the primary analysis, we determined that a minimum of 363 events would provide a power of 95% to calculate a noninferiority margin of 1.46 with a one-sided alpha level of 0.025. How-ever, 405 events were selected as the prespecified

target to ensure a robust statistical result. On the basis of a projected event rate of 2.3% per 100 patient-years in the warfarin group and a projected 14% rate of annual attrition, it was estimated that approximately 14,000 patients would need to be randomly assigned to a study group.

If noninferiority was achieved in the primary analysis, a closed testing procedure was to be con-ducted for superiority in the safety population during treatment, which included patients who received at least one dose of a study drug and were followed for events, regardless of adherence to the protocol, while they were receiving the assigned study drug or within 2 days after discontinuation (group B in Fig. 1 in the Supplementary Appen-dix). Key secondary efficacy end points were also tested for superiority in the as-treated safety popu-lation.20 Testing for noninferiority and superior-ity was also performed in the intention-to-treat population, which included all patients who un-derwent randomization and were followed for events during treatment or after premature dis-continuation (group C in Fig. 1 in the Supplemen-tary Appendix).

In addition, we performed post hoc analyses of events in the intention-to-treat population and events occurring during the end-of-study transi-tion to open-label treatment with conventional anticoagulant agents. In the warfarin group, we used the method of Rosendaal et al.21 to calculate the overall time that INR values fell within the therapeutic range. Comparative analyses of treat-ment efficacy were performed according to quar-tiles of time that INR values fell within the thera-peutic range at the participating clinical sites.

Event rates per 100 patient-years are presented as proportions of patients per year. Hazard ratios, confidence intervals, and P values were calculated with the use of Cox proportional-hazards mod-els with treatment as the only covariate. Testing for noninferiority was based on a one-sided sig-nificance level of 0.025; testing for superiority was based on a two-sided significance level of 0.05.

R esult s

Recruitment and Follow-up

From December 18, 2006, through June 17, 2009, a total of 14,264 patients underwent randomiza-tion (Fig. 1 in the Supplementary Appendix). The study was terminated on May 28, 2010. The pro-portions of patients who permanently stopped





their assigned therapy before an end-point event and before the termination date were 23.7% in the rivaroxaban group and 22.2% in the warfarin group. The median duration of treatment expo-sure was 590 days; the median follow-up period was 707 days. Only 32 patients were lost to follow-up. Because of violations in Good Clinical Prac-tice guidelines at one site that made the data un-reliable, 93 patients (50 in the rivaroxaban group and 43 in the warfarin group) were excluded from all efficacy analyses before unblinding. An addi-tional issue with data quality was raised at an-other trial site, but this issue was resolved with-out the exclusion of the patients from the analysis (for details, see the Supplementary Appendix).

Patient Characteristics and Treatments

Key clinical characteristics of the patients who underwent randomization are shown in Table 1. The median age was 73 years (a quarter of the patients were 78 years of age or older), and 39.7% of the patients were women. The patients had sub-stantial rates of coexisting illnesses: 90.5% had hypertension, 62.5% had heart failure, and 40.0% had diabetes; 54.8% of the patients had had a previous stroke, systemic embolism, or transient ischemic attack. The mean and median CHADS2 scores were 3.5 and 3.0, respectively. Data on med-ication use at baseline are provided in Table 1 in the Supplementary Appendix. Previous use of vi-tamin K antagonists was reported by 62.4% of pa-tients. At some time during the study, 34.9% of patients in the rivaroxaban group and 36.2% of those in the warfarin group took aspirin concur-rently with the assigned study drug. Among pa-tients in the warfarin group, INR values were within the therapeutic range (2.0 to 3.0) a mean of 55% of the time (median, 58%; interquartile range, 43 to 71).

Primary Outcome

In the per-protocol population (the patients in-cluded in the primary efficacy analysis), stroke or systemic embolism occurred in 188 patients in the rivaroxaban group (1.7% per year) and in 241 patients in the warfarin group (2.2% per year) (hazard ratio in the rivaroxaban group, 0.79; 95% confidence interval [CI], 0.66 to 0.96; P<0.001 for noninferiority) (Table 2 and Fig. 1A). In the as-treated safety population, primary events oc-curred in 189 patients in the rivaroxaban group (1.7% per year) and in 243 patients in the warfarin

Table 1. Characteristics of the Intention-to-Treat Population at Baseline.

CharacteristicRivaroxaban(N = 7131)

Warfarin(N = 7133)

Age — yr

Median 73 73

Interquartile range 65–78 65–78

Female sex — no. (%) 2831 (39.7) 2832 (39.7)

Body-mass index*

Median 28.3 28.1

Interquartile range 25.2 –32.1 25.1–31.8

Blood pressure — mm Hg

Systolic

Median 130 130


Diastolic

Median 80 80


Type of atrial fibrillation — no. (%)

Persistent 5786 (81.1) 5762 (80.8)

Paroxysmal 1245 (17.5) 1269 (17.8)

Newly diagnosed or new onset 100 (1.4) 102 (1.4)

Previous medication use — no. (%)

Aspirin 2586 (36.3) 2619 (36.7)

Vitamin K antagonist 4443 (62.3) 4461 (62.5)

CHADS2 risk of stroke†

Mean score (±SD) 3.48±0.94 3.46±0.95

Score — no. (%)

2 925 (13.0) 934 (13.1)

3 3058 (42.9) 3158 (44.3)

4 2092 (29.3) 1999 (28.0)

5 932 (13.1) 881 (12.4)

6‡ 123 (1.7) 159 (2.2)

Coexisting condition — no. (%)

Previous stroke, systemic em-bolism, or transient ischemic attack

3916 (54.9) 3895 (54.6)

Congestive heart failure 4467 (62.6) 4441 (62.3)

Hypertension 6436 (90.3) 6474 (90.8)

Diabetes mellitus 2878 (40.4) 2817 (39.5)

Previous myocardial infarction‡ 1182 (16.6) 1286 (18.0)

Peripheral vascular disease 401 (5.6) 438 (6.1)

Chronic obstructive pulmonary disease

754 (10.6) 743 (10.4)

Creatinine clearance — ml/min§

Median 67 67


* The body-mass index is the weight in kilograms divided by the square of the height in meters.

† The CHADS2 score for the risk of stroke ranges from 1 to 6, with higher scores indicating an increased risk. Three patients (one in the rivaroxaban group and two in the warfarin group) had a CHADS2 score of 1.

‡ P<0.05 for the between-group comparison.§ Creatinine clearance was calculated with the use of the Cockcroft–Gault formula.





group (2.2% per year) (hazard ratio, 0.79; 95% CI, 0.65 to 0.95; P = 0.01 for superiority). Among all randomized patients in the intention-to-treat anal-ysis, primary events occurred in 269 patients in the rivaroxaban group (2.1% per year) and in 306 pa-tients in the warfarin group (2.4% per year) (haz-ard ratio, 0.88; 95% CI, 0.74 to 1.03; P<0.001 for noninferiority; P = 0.12 for superiority) (Fig. 1B).

During treatment in the intention-to-treat pop-ulation, patients in the rivaroxaban group had a lower rate of stroke or systemic embolism (188 events, 1.7% per year) than those in the warfarin group (240 events, 2.2% per year) (P = 0.02) (Ta-ble 2 and Fig. 2). Among patients who stopped taking the assigned study drug before the end of the study, during a median of 117 days of follow-up after discontinuation, primary events occurred in 81 patients in the rivaroxaban group (4.7% per year) and in 66 patients in the warfarin group (4.3% per year) (P = 0.58). (Details regarding the time to events in patients who completed the study and were switched to standard medical therapy are provided in Fig. 2 in the Supplementary Appendix.)

Bleeding Outcomes

Major and clinically relevant nonmajor bleeding occurred in 1475 patients in the rivaroxaban group and in 1449 patients in the warfarin group (14.9% and 14.5% per year, respectively; hazard ratio in the rivaroxaban group, 1.03; 95% CI, 0.96 to 1.11; P = 0.44) (Table 3). Rates of major bleeding were

similar in the rivaroxaban and warfarin groups (3.6% and 3.4%, respectively; P = 0.58). Decreases in hemoglobin levels of 2 g per deciliter or more and transfusions were more common among pa-tients in the rivaroxaban group, whereas fatal bleeding and bleeding at critical anatomical sites were less frequent. Rates of intracranial hemor-rhage were significantly lower in the rivaroxaban group than in the warfarin group (0.5% vs. 0.7% per year; hazard ratio, 0.67; 95% CI, 0.47 to 0.93; P = 0.02). Major bleeding from a gastrointestinal site was more common in the rivaroxaban group, with 224 bleeding events (3.2%), as compared with 154 events in the warfarin group (2.2%, P<0.001) (Table 2 in the Supplementary Appendix). (Data on nonhemorrhagic adverse events are provided in Table 3 in the Supplementary Appendix.)

Secondary Efficacy Outcomes

The rates of secondary efficacy outcomes in the as-treated safety population are presented in Ta-ble 4 in the Supplementary Appendix. During treatment, myocardial infarction occurred in 101 patients in the rivaroxaban group and in 126 pa-tients in the warfarin group (0.9% and 1.1% per year, respectively; hazard ratio in the rivaroxaban group, 0.81; 95% CI, 0.63 to 1.06; P = 0.12). In the same analysis population, there were 208 deaths in the rivaroxaban group and 250 deaths in the warfarin group (1.9% and 2.2% per year, respec-tively; hazard ratio, 0.85; 95% CI, 0.70 to 1.02;

Table 2. Primary End Point of Stroke or Systemic Embolism.*

Study Population Rivaroxaban WarfarinHazard Ratio

(95% CI)† P Value

No. of Patients

No. of Events

Event Rate

No. of Patients

No. of Events

Event Rate Noninferiority Superiority

no./100 patient-yr

no./100 patient-yr

Per-protocol, as-treated population‡

6958 188 1.7 7004 241 2.2 0.79 (0.66–0.96) <0.001

Safety, as-treated population 7061 189 1.7 7082 243 2.2 0.79 (0.65–0.95) 0.02

Intention-to-treat population§ 7081 269 2.1 7090 306 2.4 0.88 (0.75–1.03) <0.001 0.12

During treatment 188 1.7 240 2.2 0.79 (0.66–0.96) 0.02

After discontinuation 81 4.7 66 4.3 1.10 (0.79–1.52) 0.58

* The median follow-up period was 590 days for the per-protocol, as-treated population during treatment; 590 days for the safety, as-treated population during treatment; and 707 days for the intention-to-treat population.

† Hazard ratios are for the rivaroxaban group as compared with the warfarin group.‡ The primary analysis was performed in the as-treated, per-protocol population during treatment.§ Follow-up in the intention-to-treat population continued until notification of study termination.





P = 0.07). In addition, in the intention-to-treat analy-sis throughout the trial, there were 582 deaths in the rivaroxaban group and 632 deaths in the war-farin group (4.5% and 4.9% per year, respectively; hazard ratio, 0.92; 95% CI, 0.82 to 1.03; P = 0.15).

Selected Subgroup Analyses

The effect of rivaroxaban, as compared with war-farin, in both efficacy and safety analyses was con-sistent across all prespecified subgroups (Fig. 3, 4, and 5 in the Supplementary Appendix). Fur-

thermore, the effect of rivaroxaban did not differ across quartiles of the duration of time that INR values were within the therapeutic range accord-ing to study center (P = 0.74 for interaction) (Ta-ble 5 in the Supplementary Appendix). Within the highest quartile according to center, the haz-ard ratio with rivaroxaban versus warfarin was 0.74 (95% CI, 0.49 to 1.12).

Discussion

In this randomized trial, we compared rivaroxa-ban with warfarin for the prevention of stroke or systemic embolism among patients with nonval-vular atrial fibrillation who were at moderate-to-high risk for stroke. In both the primary analy-sis, which included patients in the per-protocol population, and in the intention-to-treat analysis, we found that rivaroxaban was noninferior to war-farin. In the primary safety analysis, there was no significant difference between rivaroxaban and warfarin with respect to rates of major or nonma-jor clinically relevant bleeding.

As prespecified in the statistical-analysis plan, we analyzed the trial data in a variety of ways be-cause we anticipated that some patients would discontinue the study treatment and we wished to evaluate both noninferiority and superiority. Al-though an intention-to-treat analysis is the stan-dard method for assessing superiority in a ran-domized trial, noninferiority is best established when patients are actually taking the randomized treatment.16-19 Thus, the primary analysis was per-formed in the per-protocol population during re-ceipt of the randomly assigned therapy. In the intention-to-treat population, we found no signifi-cant between-group difference in a conventional superiority analysis. In contrast, in the analyses of patients receiving at least one dose of a study drug who were followed for events during treat-ment, we found that rivaroxaban was superior to warfarin. The difference between these results reflects the fact that among patients who discon-tinued therapy before the conclusion of the trial, no significant difference in outcomes would have been anticipated, and none was seen.

The most worrisome complication of antico-agulation is bleeding. Rates of major and nonma-jor clinically relevant bleeding, the main measure of treatment safety, were similar in the rivaroxa-ban and warfarin groups. Bleeding that proved fatal or involved a critical anatomical site occurred less frequently in the rivaroxaban group, mainly

Cum

ulat

ive

Even

t Rat

e (%

)100

80

90

70

60

40

30

10

50

20

00 120 240 360 480 600 720 840

6

4

3

1

5

2

00 120 240 360 480 600 720 840

Days since Randomization

B Events in Intention-to-Treat Population

A Events in Per-Protocol Population

No. at RiskRivaroxabanWarfarin

69587004

62116327

57865911

54685542

44064461

24722539

34073478

14961538

Warfarin

Rivaroxaban

Cum

ulat

ive

Even

t Rat

e (%

)

100

80

90

70

60

40

30

10

50

20

00 120 240 360 480 600 720 840



70817090

68796871

66836656

64706440

52645225

29512944

41054087

17851783

6

4

3

1

5

2

00 120 240 360 480 600 720 840

Warfarin

Rivaroxaban

Figure 1. Cumulative Rates of the Primary End Point (Stroke or Systemic Embolism) in the Per-Protocol Population and in the Intention-to-Treat Population.





because of lower rates of hemorrhagic stroke and other intracranial bleeding. In contrast, bleeding from gastrointestinal sites, including upper, lower, and rectal sites, occurred more frequently in the rivaroxaban group, as did bleeding that led to a drop in the hemoglobin level or bleeding that re-quired transfusion. Even though patients in our trial were at increased risk for bleeding events, rates of major bleeding were similar to those in other recent studies involving patients with atrial fibrillation.4,15,22,23

Among patients in our study who survived and did not reach the primary end point, the rate of premature, permanent cessation of randomized treatment (14.3% in year 1) was slightly higher than in other studies (average, 11%).15,23 This may have been a consequence of the trial’s double-blind design or the inclusion of patients with more co-existing illnesses. Among patients who perma-nently discontinued their assigned treatment be-fore the end of the study, only about half were treated thereafter with a vitamin K antagonist. This observation suggests that for at least some of the patients who participated in the trial, the risks of open-label therapy with currently available anticoagulants were ultimately judged to outweigh the risk of stroke or systemic embolism. Event rates were similar at 30 days and 1 year after with-drawal, suggesting that the mechanism of events did not involve hypercoagulability early after with-drawal of rivaroxaban. Events occurring at the end of the study were probably related to increased difficulty in achieving the transition from blinded trial therapy to the open-label use of a vitamin K antagonist when the patient had previously been assigned to the rivaroxaban group, since presum-ably many patients who had previously been as-signed to the warfarin group would have already had a therapeutic INR.

Among patients in the warfarin group, the pro-portion of time in which the intensity of anti-coagulation was in the therapeutic range (mean, 55%), which was calculated from all INR values during the study and for 7 days after warfarin interruptions, was lower than in previous studies of other new anticoagulants in patients with atrial fibrillation (range, 64 to 68%). Among these trials, the only study of blinded treatment was limited to North American sites, which may have facilitated trial compliance.15 Most earlier trials of warfarin included fewer high-risk patients,3 and no previous studies addressed patient populations with overall levels of coexisting illnesses and geographic diver-

sity that were similar to those of the patients in our study.24 Significant variations in the duration of time in the therapeutic range may reflect re-gional differences and differential skill in manag-ing warfarin.25 In a recent analysis of anticoagu-lation management involving more than 120,000 patients in the Veterans Affairs health care system, the mean proportion of time in the therapeutic range was 58%, with significant variation across sites.24 The efficacy of rivaroxaban, as compared with warfarin, was as favorable in centers with the best INR control as in those with poorer control.

Cum

ulat

ive

Even

t Rat

e (%

)

100

80

90

70

60

40

30

10

50

20

00 120 240 360 480 600 720 840

10

6

4

8

2

00 120 240 360 480 600 720 840


B Events after Discontinuation

A Events during Treatment


70817090

63096397

58745976

55435602

43944432

23722408

33543401

13921407

Warfarin

Rivaroxaban

Cum

ulat

ive

Even

t Rat

e (%

)

100

80

90

70

60

40

30

10

50

20

00 120 240 360 480 600 720 840

Days since Drug Discontinuation


20881962

12701193

986880

775681

543470

211196

364326

10196

10

6

4

8

2

00 120 240 360 480 600 720 840

Warfarin

Rivaroxaban

Figure 2. Cumulative Rates of the Primary End Point during Treatment and after Discontinuation in the Intention-to-Treat Population.





In conclusion, in this trial comparing a once-daily, fixed dose of rivaroxaban with adjusted-dose warfarin in patients with nonvalvular atrial fibrillation who were at moderate-to-high risk for stroke, rivaroxaban was noninferior to warfarin in the prevention of subsequent stroke or sys-temic embolism. There were no significant dif-ferences in rates of major and clinically relevant nonmajor bleeding between the two study groups, although intracranial and fatal bleeding occurred less frequently in the rivaroxaban group.

Supported by Johnson & Johnson Pharmaceutical Research and Development and Bayer HealthCare.

Dr. Patel reports receiving consulting fees from Ortho McNeil Janssen and Bayer HealthCare and serving on an advisory board for Genzyme; Dr. Mahaffey, receiving consulting fees from Adolor, Alexion, Amgen, Argolyn Bioscience, AstraZeneca, Bayer, Bristol-Myers Squibb, Daiichi Sankyo, Eli Lilly, Elsevier, Forest Labs, Genentech, GlaxoSmithKline, Guidant, Ikarja, Johnson & Johnson, Merck, Novartis, Pfizer, Proctor & Gamble, Sanofi Aventis, Schering Plough, Scios, and WebMD, grant support from Abbott Vascular, Amgen, Amylin, AstraZeneca, Bayer, Boehringer Ingelheim, Bristol-Myers Squibb, CardioKinetix, Cierra, Cordis, Edwards LifeSciences, Eli Lilly, Genentech, GlaxoSmithKline, Guidant, Innocoll Pharma-ceuticals, Johnson & Johnson, KCI Medical, Luitpold Pharmaceuti-cal, Medtronic, Merck, Momenta Pharmaceuticals, Novartis, Portola Pharmaceutical, Pozen, Regado Biotechnologies, Sanofi Aventis, Schering Plough, and the Medicines Company, and lecture fees from Johnson & Johnson; Dr. Pan, being an employee of Johnson & Johnson; Dr. Singer, receiving consulting fees from Johnson & Johnson, Bayer HealthCare, Boehringer Ingelheim, Daiichi Sankyo,

Merck, Sanofi Aventis, Medtronic, and St. Jude Medical, grant sup-port from Daiichi Sankyo, and lecture fees from Bristol-Myers Squibb and Pfizer; Dr. Hacke, serving as a board member for Syg-nis Pharma Germany and receiving consulting fees, grant support, lecture fees, and payments for the development of educational presentations from Boehringer Ingelheim, consulting fees from Photothera USA and Codman USA, and lecture fees from Bayer and Photothera USA; Dr. Breithardt, receiving consulting fees from Bayer HealthCare, Sanofi Aventis, Bristol-Myers Squibb, Boehringer Ingelheim, Boston Scientific, and Otsuka Pharmaceu-tical, grant support from Sanofi Aventis, St. Jude, and Meda Phar-ma, and lecture fees from Sanofi Aventis, Boehringer Ingelheim, Bayer HealthCare, and Boston Scientific; Dr. Halperin, receiving consulting fees from Astellas Pharma, Boehringer Ingelheim, Bristol-Myers Squibb, Daiichi Sankyo, Pfizer, and Sanofi Aventis; Dr. Hankey, receiving consulting fees from Sanofi Aventis, Schering Plough, and Boehringer Ingelheim and lecture fees from Sanofi Aventis and Pfizer; Dr. Becker, receiving grant support from Johnson & Johnson, Bayer, Regado Biosciences, and Bristol-Myers Squibb, consulting fees from Boehringer Ingelheim, and lecture fees from Merck and AstraZeneca; Dr. Nessel, being an employee of and having an equity interest in Johnson & Johnson; Drs. Paolini and Berkowitz, being employees of Bayer HealthCare, and Dr. Paolini, having an equity interest in Bayer HealthCare; Dr. Fox, receiving grant support and lecture fees from Eli Lilly and lecture fees from Sanofi Aventis and AstraZeneca; Dr. Califf, receiving con-sulting fees from Kowa, Nile, Orexigen, Sanofi Aventis, Novartis, and Xoma and grant support from Novartis, Merck, and Amilyn/Lilly and having an equity interest in Nitrox. No other potential conflict of interest relevant to this article was reported. Disclosure forms provided by the authors are available with the full text of this article at NEJM.org.

We thank Lisa G. Berdan, Karen Hannan, and Kimberly Schwabe for their operational oversight of the trial and Elizabeth Cook for providing editorial support in the preparation of the manuscript.

Table 3. Rates of Bleeding Events.*

VariableRivaroxaban(N = 7111)

Warfarin(N = 7125)

Hazard Ratio(95% CI)† P Value‡

Events Event Rate Events Event Rate

no. (%)no./100

patient-yr no. (%)no./100

patient-yr

Principal safety end point: major and nonmajor clinically relevant bleeding§

1475 (20.7) 14.9 1449 (20.3) 14.5 1.03 (0.96–1.11) 0.44

Major bleeding

Any 395 (5.6) 3.6 386 (5.4) 3.4 1.04 (0.90–1.20) 0.58

Decrease in hemoglobin ≥2 g/dl 305 (4.3) 2.8 254 (3.6) 2.3 1.22 (1.03–1.44) 0.02

Transfusion 183 (2.6) 1.6 149 (2.1) 1.3 1.25 (1.01–1.55) 0.04

Critical bleeding¶ 91 (1.3) 0.8 133 (1.9) 1.2 0.69 (0.53–0.91) 0.007

Fatal bleeding 27 (0.4) 0.2 55 (0.8) 0.5 0.50 (0.31–0.79) 0.003

Intracranial hemorrhage 55 (0.8) 0.5 84 (1.2) 0.7 0.67 (0.47–0.93) 0.02

Nonmajor clinically relevant bleeding 1185 (16.7) 11.8 1151 (16.2) 11.4 1.04 (0.96–1.13) 0.35

* All analyses of rates of bleeding are based on the first event in the safety population during treatment.† Hazard ratios are for the rivaroxaban group as compared with the warfarin group and were calculated with the use of Cox proportional-hazards

models with the study group as a covariate.‡ Two-sided P values are for superiority in the rivaroxaban group as compared with the warfarin group.§ Minimal bleeding events were not included in the principal safety end point.¶ Bleeding events were considered to be critical if they occurred in intracranial, intraspinal, intraocular, pericardial, intraarticular, intramuscular

(with compartment syndrome), or retroperitoneal sites.





1. Wolf PA, Abbott RD, Kannel WB. Atrial fibrillation as an independent risk factor for stroke: the Framingham Study. Stroke 1991;22:983-8.2. Idem. Atrial fibrillation: a major con-tributor to stroke in the elderly: the Fram-ingham Study. Arch Intern Med 1987;147: 1561-4.3. Hart RG, Benavente O, McBride R, Pearce LA. Antithrombotic therapy to pre-vent stroke in patients with atrial fibrilla-tion: a meta-analysis. Ann Intern Med 1999;131:492-501.4. Hart RG, Pearce LA, Aguilar MI. Meta-analysis: antithrombotic therapy to pre-vent stroke in patients who have nonval-vular atrial fibrillation. Ann Intern Med 2007;146:857-67.5. Singer DE, Albers GW, Dalen JE, Go AS, Halperin JL, Manning WJ. Antithrom-botic therapy in atrial fibrillation: the Seventh ACCP Conference on Antithrom-botic and Thrombolytic Therapy. Chest 2004;126:Suppl:429S-456S.6. Albers GW, Yim JM, Belew KM, et al. Status of antithrombotic therapy for pa-tients with atrial fibrillation in university hospitals. Arch Intern Med 1996;156:2311-6.7. Go AS, Hylek EM, Borowsky LH, Phil-lips KA, Selby JV, Singer DE. Warfarin use among ambulatory patients with nonval-vular atrial fibrillation: the Anticoagula-tion and Risk Factors in Atrial Fibrillation (ATRIA) study. Ann Intern Med 1999; 131:927-34.8. Piccini JP, Hernandez AF, Zhao X, et al. Quality of care for atrial fibrillation among patients hospitalized for heart failure. J Am Coll Cardiol 2009;54:1280-9.9. Kubitza D, Becka M, Wensing G, Voith B, Zuehlsdorf M. Safety, pharmaco-dynamics, and pharmacokinetics of BAY 59-7939 — an oral, direct Factor Xa in-hibitor — after multiple dosing in healthy

male subjects. Eur J Clin Pharmacol 2005; 61:873-80.10. Kubitza D, Becka M, Roth A, Mueck W. Dose-escalation study of the pharma-cokinetics and pharmacodynamics of riv-a roxaban in healthy elderly subjects. Curr Med Res Opin 2008;24:2757-65.11. Eriksson BI, Borris LC, Friedman RJ, et al. Rivaroxaban versus enoxaparin for thromboprophylaxis after hip arthroplas-ty. N Engl J Med 2008;358:2765-75.12. Lassen MR, Ageno W, Borris LC, et al. Rivaroxaban versus enoxaparin for thromboprophylaxis after total knee ar-throplasty. N Engl J Med 2008;358:2776-86.13. The EINSTEIN Investigators. Oral riva-roxaban for symptomatic venous throm-boembolism. N Engl J Med 2010;363:2499-510.14. ROCKET AF Study Investigators. Riv-aroxaban — once daily, oral, direct factor Xa inhibition compared with vitamin K antagonism for prevention of stroke and Embolism Trial in Atrial Fibrillation: ratio-nale and design of the ROCKET AF study. Am Heart J 2010;159(3):340.e1-347.e1.15. Albers GW, Diener HC, Frison L, et al. Ximelagatran vs warfarin for stroke pre-vention in patients with nonvalvular atrial fibrillation: a randomized trial. JAMA 2005;293:690-8.16. Matilde Sanchez M, Chen X. Choosing the analysis population in non-inferiority studies: per protocol or intent-to-treat. Stat Med 2006;25:1169-81.17. Brittain E, Lin D. A comparison of intent-to-treat and per-protocol results in antibiotic non-inferiority trials. Stat Med 2005;24:1-10.18. Draft: Guidance for Industry: non-inferiority clinical trials. Silver Spring, MD: Food and Drug Administration, 2010. (http://www.fda.gov/Drugs/GuidanceComplianceRegulatoryInformation/ Guidances/default.htm.)

19. Piaggio G, Elbourne DR, Altman DG, Pocock SJ, Evans SJ. Reporting of nonin-feriority and equivalence randomized tri-als: an extension of the CONSORT state-ment. JAMA 2006;295:1152-60. [Erratum, JAMA 2006;296:1842.]20. Soulakova JN. Comparison of several testing strategies for combination drug efficacy trials based on the closure prin-ciple. Stat Med 2009;28:260-73.21. Rosendaal FR, Cannegieter SC, van der Meer FJ, Briët E. A method to deter-mine the optimal intensity of oral antico-agulant therapy. Thromb Haemost 1993; 69:236-9.22. Connolly S, Pogue J, Hart R, et al. Clopidogrel plus aspirin versus oral anti-coagulation for atrial fibrillation in the Atrial fibrillation Clopidogrel Trial with Irbesartan for prevention of Vascular Events (ACTIVE W): a randomised controlled trial. Lancet 2006;367:1903-12.23. Connolly SJ, Ezekowitz MD, Yusuf S, et al. Dabigatran versus warfarin in pa-tients with atrial fibrillation. N Engl J Med 2009;361:1139-51. [Erratum, N Engl J Med 2010;363:1877.]24. Rose AJ, Hylek EM, Ozonoff A, Ash AS, Reisman JI, Berlowitz DR. Risk- adjusted percent time in therapeutic range as a quality indicator for outpatient oral anticoagulation: results of the Veter-ans Affairs Study to Improve Anticoagu-lation (VARIA). Circ Cardiovasc Qual Out-comes 2011;4:22-9.25. Wallentin L, Yusuf S, Ezekowitz MD, et al. Efficacy and safety of dabigatran compared with warfarin at different levels of international normalised ratio control for stroke prevention in atrial fibrillation: an analysis of the RE-LY trial. Lancet 2010;376:975-83.Copyright © 2011 Massachusetts Medical Society.

References

receive immediate notification when an article is published online first

To be notified by e-mail when Journal articles are published Online First, sign up at NEJM.org.



494

·

N Engl J Med, Vol. 345, No. 7

·

August 16, 2001

·

www.nejm.org


EFFECTS OF CLOPIDOGREL IN ADDITION TO ASPIRIN IN PATIENTS WITH ACUTE CORONARY SYNDROMES WITHOUT ST-SEGMENT ELEVATION

T

HE

C

LOPIDOGREL

IN

U

NSTABLE

A

NGINA

TO

P

REVENT

R

ECURRENT

E

VENTS

T

RIAL

I

NVESTIGATORS

*

A

BSTRACT

Background

Despite current treatments, patientswho have acute coronary syndromes without ST-seg-ment elevation have high rates of major vascularevents. We evaluated the efficacy and safety of theantiplatelet agent clopidogrel when given with aspi-rin in such patients.

Methods

We randomly assigned 12,562 patientswho had presented within 24 hours after the onset ofsymptoms to receive clopidogrel (300 mg immediately,followed by 75 mg once daily) (6259 patients) or pla-cebo (6303 patients) in addition to aspirin for 3 to 12months.

Results

The first primary outcome — a compositeof death from cardiovascular causes, nonfatal myocar-dial infarction, or stroke — occurred in 9.3 percent ofthe patients in the clopidogrel group and 11.4 percentof the patients in the placebo group (relative risk withclopidogrel as compared with placebo, 0.80; 95 per-cent confidence interval, 0.72 to 0.90; P<0.001). Thesecond primary outcome — the first primary outcomeor refractory ischemia — occurred in 16.5 percent ofthe patients in the clopidogrel group and 18.8 percentof the patients in the placebo group (relative risk, 0.86,P<0.001). The percentages of patients with in-hospitalrefractory or severe ischemia, heart failure, and revas-cularization procedures were also significantly lowerwith clopidogrel. There were significantly more pa-tients with major bleeding in the clopidogrel groupthan in the placebo group (3.7 percent vs. 2.7 percent;relative risk, 1.38; P=0.001), but there were not signif-icantly more patients with episodes of life-threateningbleeding (2.1 percent vs. 1.8 percent, P=0.13) or hem-orrhagic strokes.

Conclusions

The antiplatelet agent clopidogrel hasbeneficial effects in patients with acute coronary syn-dromes without ST-segment elevation. However, therisk of major bleeding is increased among patientstreated with clopidogrel. (N Engl J Med 2001;345:494-502.)


The Manuscript Writing Committee (Salim Yusuf, D.Phil., F.R.C.P.C.,Feng Zhao, M.Sc., Shamir R. Mehta, M.D., F.R.C.P.C., Susan Chrolavicius,B.Sc., Gianni Tognoni, M.D., and Keith K. Fox, M.D., F.R.C.P.) assumesresponsibility for the overall content of the manuscript. Address reprint re-quests to Dr. Yusuf at the Canadian Cardiovascular Collaboration ProjectOffice, Population Health Research Institute, McMaster University, Hamil-ton General Hospital, 237 Barton St. E., Hamilton, ON L8L 2X2, Canada,or at [email protected].

*The Clopidogrel in Unstable Angina to Prevent Recurrent Events (CURE)trial investigators are listed in the Appendix.

HROMBOSIS caused by a ruptured oreroded atherosclerotic plaque is the usualunderlying mechanism of acute coronarysyndromes.

1

Aspirin and heparin reduce therisk of death from cardiovascular causes, new myocar-dial infarction, and recurrent ischemia,

2,3

but there isstill a substantial risk of such events in both the shortterm and the long term. Intravenous glycoproteinIIb/IIIa receptor blockers have been shown to reducethe incidence of early events, mainly among patients

T

who are treated according to an invasive strategy,

4,5

butlong-term oral therapy with glycoprotein IIb/IIIareceptor blockers is not beneficial and may even in-crease mortality.

6

Similarly, continuing treatment withlow-molecular-weight heparin beyond one week hasnot been shown to be effective.

7

Although the long-term use of oral anticoagulants may be useful, noconvincing evidence of their benefit is yet available.

8

Therefore, there is a need to reduce further the riskof ischemic events in a broad spectrum of patients bothwhen they first present with acute coronary syndromesand in the long term.

The thienopyridine derivatives, ticlopidine and clo-pidogrel, are antiplatelet agents that inhibit the plate-let aggregation induced by adenosine diphosphate,thereby reducing ischemic events.

9

Combining one ofthese drugs with aspirin, which blocks the thrombox-ane-mediated pathway, may have an additive effect. Inpatients who are undergoing percutaneous translumi-nal coronary angioplasty (PTCA) with stenting, short-term aspirin treatment plus a thienopyridine derivativeresults in a substantially lower rate of myocardial in-farction than does either aspirin alone or warfarin.

10

However, the role of long-term combined therapy withaspirin and an antiplatelet agent in a broader group ofpatients at high risk for cardiovascular events is un-known. We therefore designed the Clopidogrel in Un-stable Angina to Prevent Recurrent Events (CURE)trial to compare the efficacy and safety of the early andlong-term use of clopidogrel plus aspirin with thoseof aspirin alone in patients with acute coronary syn-dromes and no ST-segment elevation.

METHODS

Study Design

We undertook a randomized, double-blind, placebo-controlledtrial comparing clopidogrel with placebo in patients who presentedwith acute coronary syndromes without ST-segment elevation. Thedesign and rationale of the study have been reported previously.

9

Study Patients

Patients were eligible for the study if they had been hospital-ized within 24 hours after the onset of symptoms and did nothave ST-segment elevation. Initially, patients older than 60 years



CLOPIDOGREL IN ADDITION TO ASPIRIN FOR ACUTE CORONARY SYNDROMES WITHOUT ST-SEGMENT ELEVATION


·

August 16, 2001

·

www.nejm.org

·

495

of age with no new electrocardiographic changes but with a historyof coronary artery disease were included. However, after a review ofthe overall rates of events among the first 3000 patients, the steeringcommittee recommended that we enroll only patients who had ei-ther electrocardiographic changes or an elevation in the serum levelof cardiac enzymes or markers at entry. We excluded patients withcontraindications to antithrombotic or antiplatelet therapy, thosewho were at high risk for bleeding or severe heart failure, those whowere taking oral anticoagulants, and those who had undergone cor-onary revascularization in the previous three months or had receivedintravenous glycoprotein IIb/IIIa receptor inhibitors in the previ-ous three days.

After we had obtained written informed consent, patients wererandomly assigned to either the clopidogrel group or the placebogroup by a central, 24-hour, computerized randomization service.Permuted-block randomization, stratified according to clinical cen-ter, was used. A loading dose of clopidogrel (300 mg orally) ormatching placebo was administered immediately, followed by clo-pidogrel (75 mg per day) or matching placebo for 3 to 12 months(mean duration of treatment, 9 months). Aspirin (recommendeddose, 75 to 325 mg daily) was started or continued simultaneouslywith the study drug. Follow-up assessments occurred at discharge,at one and three months, and then every three months until theend of the study.

Study Organization

Patients were recruited between December 1998 and Septem-ber 2000 at 482 centers in 28 countries. The ethics review boardat each institution approved the study. The study was organized andcoordinated and all the data were managed and analyzed by theCanadian Cardiovascular Collaboration Project Office, McMasterUniversity, Hamilton, Ontario. A steering committee consisting ofnational coordinators oversaw the study. The data were periodicallyreviewed by an independent data and safety monitoring board.

Outcomes

The first primary outcome was the composite of death fromcardiovascular causes, nonfatal myocardial infarction, or stroke, andthe second primary outcome was the composite of the first primaryoutcome or refractory ischemia. The secondary outcomes were se-vere ischemia, heart failure, and the need for revascularization. Thesafety-related outcomes were bleeding complications, which werecategorized as life-threatening, major (requiring the transfusion of2 or more units of blood), or minor. All primary outcomes andlife-threatening and major bleeding complications were adjudicatedby persons who were unaware of the patients’ treatment-groupassignments.

Definitions

Death from cardiovascular causes was defined as any death forwhich there was no clearly documented nonvascular cause. Myocar-dial infarction was defined by the presence of at least two of thefollowing: ischemic chest pain; the elevation of the serum levelsof cardiac markers or enzymes (troponin, creatine kinase, creatinekinase MB isoenzyme, or other cardiac enzymes) to at least twicethe upper limit of the normal reference range or three times the up-per limit of normal within 48 hours after percutaneous coronaryintervention (or to a level 20 percent higher than the previous val-ue if the level had already been elevated because of an early myo-cardial infarction); and electrocardiographic changes compatiblewith infarction.

9

Stroke was defined as a new focal neurologic def-icit of vascular origin lasting more than 24 hours. Stroke was fur-ther classified as the result of intracranial hemorrhage, ischemia (ifa computed tomographic or magnetic resonance imaging scan wasavailable), or uncertain cause.

Refractory ischemia in the hospital was defined as recurrent chestpain lasting more than five minutes with new ischemic electrocar-diographic changes while the patient was receiving optimal med-ical therapy (two antianginal agents, one of which was intravenousnitrate unless such therapy was contraindicated) and leading to ad-

ditional interventions (such as thrombolytic therapy, cardiac cath-eterization, the insertion of an intraaortic balloon pump, coronaryrevascularization, or transfer to a referral hospital for an invasive pro-cedure) by midnight of the next calendar day. Refractory ischemiaafter discharge was defined by rehospitalization lasting at least 24hours for unstable angina, with ischemic electrocardiographicchanges. Severe ischemia (in the hospital) was defined as ischemiathat was similar to in-hospital refractory ischemia but for which nourgent intervention was performed. Recurrent angina (in the hos-pital) was defined similarly, but electrocardiographic changes werenot required.

Major bleeding episodes were defined as substantially disablingbleeding, intraocular bleeding leading to the loss of vision, or bleed-ing necessitating the transfusion of at least 2 units of blood. Majorbleeding was classified as life-threatening if the bleeding episodewas fatal or led to a reduction in the hemoglobin level of at least5 g per deciliter or to substantial hypotension requiring the use ofintravenous inotropic agents, if it necessitated a surgical interven-tion, if it was a symptomatic intracranial hemorrhage, or if it neces-sitated the transfusion of 4 or more units of blood. Minor bleedingepisodes included other hemorrhages that led to the interruptionof the study medication.


The study was initially designed to include 9000 patients, withan expected rate of events in the placebo group of 12 to 14 percent.However, because the rate of events appeared to be lower than hadoriginally been expected, the size of the study was increased. As-suming a rate of 10 percent in the placebo group for the first pri-mary outcome and a two-sided alpha level of 0.045, a study with12,500 patients would have 90 percent power to detect a 16.9 per-cent reduction in risk. For the second primary outcome, assuminga 14 percent rate of events in the placebo group and a two-sidedalpha level of 0.01, the study had 90 percent power to detect areduction of 16.4 percent in risk. Partitioning the alpha maintainsan overall level of 0.05, after adjustment for the overlap betweenthe two sets of outcomes. All analyses were based on the inten-tion-to-treat principle and used either the log-rank statistic or thechi-square test. Subgroup analyses were conducted with the useof tests for interactions in the Cox regression model.

The data and safety monitoring board monitored the incidenceof the primary outcome to determine the benefit of clopidogrel,using a modified Haybittle–Peto boundary of 4 SD in the firsthalf of the study and 3 SD in the second half of the study. Theboundary had to be exceeded at two or more consecutive timepoints, at least three months apart, for the board to consider ter-minating the study early. There were two formal interim assess-ments performed at the times when approximately one third andtwo thirds of the expected events had occurred. Despite the factthat the preset boundary indicating efficacy had been crossed bythe time of the second interim analysis, the board recommendedthat the trial continue until its planned end, in order to definemore clearly whether the risks of major bleeding episodes couldoffset the benefits of therapy.

All unrefuted events that occurred up to the end of the sched-uled follow-up period on December 6, 2000, are included in theanalyses. Vital status was ascertained for 12,549 of the 12,562 pa-tients who underwent randomization (99.9 percent), with 6 patientsin the clopidogrel group and 7 in the placebo group lost to fol-low-up.

RESULTS

The base-line characteristics of the patients areshown in Table 1.

Primary Outcomes

The first primary outcome — death from cardiovas-cular causes, nonfatal myocardial infarction, or stroke— occurred in 582 of the 6259 patients in the clo-



496

·


·

August 16, 2001

·

www.nejm.org


pidogrel group (9.3 percent) as compared with 719of the 6303 patients in the placebo group (11.4 per-cent); relative risk, 0.80; 95 percent confidence inter-val, 0.72 to 0.90; P<0.001) (Fig. 1 and 2 and Table2). The rate of the second primary outcome — deathfrom cardiovascular causes, nonfatal myocardial infarc-tion, stroke, or refractory ischemia — was also higherin the placebo group (1187 patients [18.8 percent])than in the clopidogrel group (1035 patients [16.5percent]; relative risk, 0.86; 95 percent confidenceinterval, 0.79 to 0.94; P<0.001). The rate of eachcomponent of these composite outcomes also tendedto be lower in the clopidogrel group. However, theclearest difference was observed in the rates of myo-cardial infarction (Table 2). With respect to refractoryischemia, the difference was observed primarily in firstevents that occurred during the initial hospitalization(85 in the clopidogrel group as compared with 126in the placebo group; relative risk, 0.68; 95 percentconfidence interval, 0.52 to 0.90; P=0.007), with lit-tle difference in the rate of rehospitalization for unsta-ble angina.

Other In-Hospital Outcomes

Significantly fewer patients in the clopidogrel groupthan in the placebo group had severe ischemia (176patients [2.8 percent] vs. 237 patients [3.8 percent];relative risk, 0.74; 95 percent confidence interval, 0.61to 0.90; P=0.003) or recurrent angina (1307 [20.9percent] vs. 1442 [22.9 percent]; relative risk, 0.91;95 percent confidence interval, 0.85 to 0.98; P=0.01)(Fig. 3). Slightly fewer patients in the clopidogrelgroup underwent coronary revascularization duringthe study (36.0 percent vs. 36.9 percent), but the dif-ference was accounted for entirely by a difference inthe rate of revascularization during the initial periodof hospitalization (20.8 percent in the clopidogrelgroup vs. 22.7 percent in the placebo group, P=0.03).Radiologic evidence of heart failure was found in few-er patients in the clopidogrel group (229 [3.7 per-cent], vs. 280 [4.4 percent] in the placebo group; rel-ative risk, 0.82; 95 percent confidence interval, 0.69to 0.98; P=0.03).

Temporal Trends

The rate of the first primary outcome was lowerin the clopidogrel group both within the first 30 daysafter randomization (relative risk, 0.79; 95 percentconfidence interval, 0.67 to 0.92) and between 30days and the end of the study (relative risk, 0.82; 95percent confidence interval, 0.70 to 0.95) (Fig. 1 and2). Further analysis indicated that the benefit of clo-pidogrel was apparent within a few hours after ran-domization, with the rate of death from cardiovascularcauses, nonfatal myocardial infarction, stroke, or refrac-tory or severe ischemia significantly lower in the clo-pidogrel group by 24 hours after randomization (1.4percent in the clopidogrel group vs. 2.1 percent in

the placebo group; relative risk, 0.66; 95 percent con-fidence interval, 0.51 to 0.86).

Subgroup Analyses

The consistency of the results in a number of keysubgroups is documented in Figure 4. The benefitswere also consistent among subgroups receiving dif-ferent doses of aspirin and among those receiving ornot receiving lipid-lowering drugs, beta-blockers, hep-arin, or angiotensin-converting–enzyme inhibitors at

*Plus–minus values are means ±SD. CABG denotes coronary-artery by-pass grafting, PTCA percutaneous transluminal coronary angioplasty, LMWlow molecular weight, and ACE angiotensin-converting enzyme.

†An associated myocardial infarction was defined as a myocardial infarc-tion associated with the episode of pain that occurred before randomization.

‡Data on the particular type of abnormality were missing for one patientin the placebo group.

T

ABLE

1.

B

ASE

-L

INE

D

EMOGRAPHIC

C

HARACTERISTICS

, M

EDICAL

H

ISTORY

, E

LECTROCARDIOGRAPHIC

C

HANGES

,

AND

D

RUG

T

HERAPY

.*

C

HARACTERISTIC

C

LOPIDOGREL

G

ROUP

(N=6259)

P

LACEBO

G

ROUP

(N=6303)

Age — yr 64.2±11.3 64.2±11.3

Female sex — no. (%) 2420 (38.7) 2416 (38.3)

Time from onset of pain to ran-domization — hr

14.2±7.2 14.1±7.1

Heart rate — beats/min 73.2±14.8 73.0±14.6

Systolic blood pressure — mm Hg 134.4±22.5 134.1±22.0

Diagnosis at study entry — no. (%)Unstable anginaSuspected myocardial infarction

4690 (74.9)1569 (25.1)

4724 (74.9)1579 (25.1)

Associated myocardial infarction — no. (%)†

1624 (25.9) 1659 (26.3)

Medical history — no. (%)Myocardial infarctionCABG or PTCAStrokeHeart failureHypertensionDiabetesCurrent or former smoker

2029 (32.4)1107 (17.7)274 (4.4)462 (7.4)

3750 (59.9)1405 (22.4)3790 (60.6)

2015 (32.0)1139 (18.1)232 (3.7)492 (7.8)

3642 (57.8)1435 (22.8)3841 (60.9)

Electrocardiographic abnormality— no. (%)‡

AnyST segment

Depression »1 mmElevation «1 mmTransient elevation >2 mm

5863 (93.7)

2642 (42.2)203 (3.2)38 (0.6)

5921 (93.9)

2646 (42.0)199 (3.2)37 (0.6)

T-wave inversionMajor (»2 mm)Other (<2 mm)

Other

1589 (25.4)721 (11.5)670 (10.7)

1635 (25.9)713 (11.3)690 (10.9)

Medications at time of randomiza-tion — no. (%)

AspirinHeparin or LMW heparinACE inhibitorBeta-blockerCalcium-channel blockerLipid-lowering agentIntravenous nitrate

4168 (66.6)4522 (72.3)2347 (37.5)3678 (58.8)1784 (28.5)1599 (25.6)2836 (45.3)

4134 (65.6)4605 (73.1)2309 (36.6)3690 (58.5)1771 (28.1)1586 (25.2)2906 (46.1)





·

August 16, 2001

·

www.nejm.org

·

497

the time of randomization. There was a tendency to-ward a greater benefit among patients who had pre-viously undergone revascularization (relative risk ofthe first primary outcome, 0.56; 95 percent confidenceinterval, 0.43 to 0.72) than among those who hadnot (relative risk, 0.88; 95 percent confidence interval,0.78 to 0.99; P for interaction=0.002). However,these results should be interpreted cautiously, giventhe large numbers of subgroup analyses that wereperformed. Furthermore, consistent benefits were ob-served irrespective of whether patients underwent re-vascularization procedures after randomization.

Safety

Major bleeding was significantly more common inthe clopidogrel group (3.7 percent in the clopidogrelgroup as compared with 2.7 percent in the placebogroup; relative risk, 1.38; 95 percent confidence in-terval, 1.13 to 1.67; P=0.001) (Table 3). There were135 patients with life-threatening bleeding episodesin the clopidogrel group (2.2 percent) as comparedwith 112 in the placebo group (1.8 percent; relativerisk, 1.21; 95 percent confidence interval, 0.95 to1.56). There was no excess rate of fatal bleeding, bleed-ing requiring surgical intervention, or hemorrhagicstroke. The excess major bleeding episodes were gas-trointestinal hemorrhages and bleeding at the sites of

Figure 1.

Cumulative Hazard Rates for the First Primary Outcome (Death from Cardiovascular Causes,Nonfatal Myocardial Infarction, or Stroke) during the 12 Months of the Study.The results demonstrate the sustained effect of clopidogrel.

0.00

0.14

0 12

0.12

0.10

0.08

0.06

0.04

0.02

3 6 9

Months of Follow-up

P<0.001

NO. AT RISKF

PlaceboFClopidogrel

F

6303F6259

F

5780F5866

F

4664F4779

F

3600F3644

F

2388F2418

Clopidogrel

Placebo

Cu

mu

lati

ve H

azar

d R

ate

Figure 2.

Cumulative Hazard Rates for the First Primary Outcome(Death from Cardiovascular Causes, Nonfatal Myocardial Infarc-tion, or Stroke) during the First 30 Days after Randomization.The results demonstrate the early effect of clopidogrel.

0.00

0.06

0 30

0.05

0.04

0.03

0.02

0.01

10 20

P=0.003

Placebo

Clopidogrel

Days of Follow-up

NO. AT RISKF

PlaceboFClopidogrel

F

6303F6259

F

6108F6103

F

5998F6035

F

5957F5984

Cu

mu

lati

ve H

azar

d R

ate



498

·


·

August 16, 2001

·

www.nejm.org


*The number of patients who died from cardiovascular causes or had a nonfatal myocardial infarc-tion was 539 (8.6 percent) in the clopidogrel group and 660 (10.5 percent) in the placebo group(P<0.001; relative risk, 0.81; 95 percent confidence interval, 0.72 to 0.91). The corresponding num-bers at 30 days were 241 (3.9 percent) and 305 (4.8 percent) (relative risk, 0.79; 95 percent confi-dence interval, 0.67 to 0.94; P=0.007). CI denotes confidence interval.

†Some patients had both a Q-wave and a non–Q-wave myocardial infarction.

‡Only the first ischemic event was counted for each patient.

T

ABLE

2.

I

NCIDENCE

OF

THE

M

AIN

S

TUDY

O

UTCOMES

.*

O

UTCOME

C

LOPIDOGREL

G

ROUP

(N=6259)

P

LACEBO

G

ROUP

(N=6303)

R

ELATIVE

R

ISK

(95% CI) P V

ALUE

no. (%)

First primary outcome: nonfatal myo-cardial infarction, stroke, ordeath from cardiovascular causes

582 (9.3) 719 (11.4) 0.80 (0.72–0.90) <0.001

Second primary outcome: first primary outcome or refractory ischemia

Death from cardiovascular causesMyocardial infarction†

Q-waveNon–Q-wave

StrokeRefractory ischemia‡

During initial hospitalizationAfter discharge

1035 (16.5)

318 (5.1)324 (5.2)116 (1.9)216 (3.5)75 (1.2)

544 (8.7)85 (1.4)

459 (7.6)

1187 (18.8)

345 (5.5)419 (6.7)193 (3.1)242 (3.8)87 (1.4)

587 (9.3)126 (2.0)461 (7.6)

0.86 (0.79–0.94)

0.93 (0.79–1.08)0.77 (0.67–0.89)0.60 (0.48–0.76)0.89 (0.74–1.07)0.86 (0.63–1.18)0.93 (0.82–1.04)0.68 (0.52–0.90)0.99 (0.87–1.13)

<0.001

Death from noncardiovascular causes 41 (0.7) 45 (0.7) 0.91 (0.60–1.39)

arterial punctures. The number of patients who re-quired the transfusion of 2 or more units of blood washigher in the clopidogrel group (177 [2.8 percent])than in the placebo group (137 [2.2 percent], P=0.02). The rate of major bleeding episodes was higherearly (within 30 days after randomization: 2.0 percentvs. 1.5 percent; relative risk, 1.31; 95 percent confi-dence interval, 1.01 to 1.70) and also late (more than30 days after randomization: 1.7 percent vs. 1.1 per-cent; relative risk, 1.48; 95 percent confidence interval,1.10 to 1.99). Overall, there was no significant excessof major bleeding episodes after coronary-artery by-pass grafting (CABG) (1.3 percent vs. 1.1 percent; rel-ative risk, 1.26; 95 percent confidence interval, 0.93 to1.71). However, in most patients scheduled for CABGsurgery, the study medication was discontinued be-fore the procedure (median time before the procedure,five days). In the 910 patients in whom the study med-ication was discontinued more than five days beforethe procedure (five days being the duration of the ef-fect of clopidogrel), there was no apparent excess ofmajor bleeding within seven days after surgery (4.4percent of the patients in the clopidogrel group vs.5.3 percent of those in the placebo group). In the 912patients who stopped taking the medications withinfive days before CABG surgery, the rate of majorbleeding was 9.6 percent in the clopidogrel group and6.3 percent in the placebo group (relative risk, 1.53;

P=0.06). Overall, the risk of minor bleeding wassignificantly higher in the clopidogrel group than inthe placebo group (322 [5.1 percent] vs. 153 [2.4percent]; P<0.001). The numbers of patients withthrombocytopenia (28 in the placebo group and 26in the clopidogrel group) or neutropenia (5 and 8,respectively) were similar.

Adherence to Study Medication and Aspirin

A total of 46.2 percent of the patients in the clopid-ogrel group discontinued the study medication tem-porarily (for more than five days), as compared with45.4 percent in the placebo group. The most commonreason for the temporary discontinuation of the studymedication was the need for revascularization or an-other surgical procedure; 84 percent of the patientswith such a need discontinued the medication beforethe procedure. A total of 21.1 percent of the patientsin the clopidogrel group discontinued the study med-ication permanently, as compared with 18.8 percentin the placebo group. A total of 99 percent of the pa-tients in both groups were taking aspirin while theywere in the hospital, 96 percent were taking it at threemonths, and 94 percent at the final visit. The use ofall other medications (other than thrombolytic ther-apy and glycoprotein IIb/IIIa receptor inhibitors)was similar in the clopidogrel group and the placebogroup.





·

August 16, 2001

·

www.nejm.org

·

499

Thrombolytic Therapy and Glycoprotein IIb/IIIaReceptor Inhibitors

A total of 71 patients in the clopidogrel group (1.1percent) and 126 patients in the placebo group (2.0percent) received thrombolytic therapy (relative risk,0.57; 95 percent confidence interval, 0.43 to 0.76;P<0.001); 369 patients in the clopidogrel group (5.9percent) and 454 in the placebo group (7.2 percent)received a glycoprotein IIb/IIIa receptor inhibitor(relative risk, 0.82; 95 percent confidence interval,0.72 to 0.93; P=0.003).

DISCUSSION

Our study demonstrates the benefit of adding clo-pidogrel to the regimen of treatment for patients withacute coronary syndromes without ST-segment eleva-tion who are receiving aspirin and other medications.Treatment with clopidogrel reduced the risk of myo-cardial infarction and recurrent ischemia, with a trendtoward lower rates of stroke and death from cardiovas-cular causes. Fewer patients in the clopidogrel groupreceived a thrombolytic agent or an intravenous gly-

coprotein IIb/IIIa receptor inhibitor. The benefits weobserved were in addition to those of aspirin, whichwas recommended for all patients, indicating thatblocking the adenosine diphosphate–receptor path-way with clopidogrel leads to further benefit.

Our study primarily included centers in which therewas no routine policy of early use of invasive proce-dures, since such a policy would have led to a high rateof immediate discontinuation of the study medicationand the use of an open-label thienopyridine derivative.Once a patient had been randomly assigned to a treat-ment group, there were no restrictions on the use ofany therapy or intervention. In particular, if the clini-cian believed that angiography and revascularizationwere needed or that a thienopyridine derivative wasindicated, the study medication could be stopped oropen-label clopidogrel or ticlopidine could be used.In fact, 5491 patients (43.7 percent) underwent an-giography, 2072 patients (16.5 percent) underwentCABG, and 2658 patients (21.2 percent) underwentPTCA. In 85.8 percent of the patients who under-went PTCA and 84.9 percent of those who underwent

Figure 3.

Proportions of Patients Who Had Events Other Than Those Included in the First Primary Outcome while TheyWere in the Hospital.The numbers and percentages of patients in each group with the specified outcome are given above the bars. RR de-notes relative risk.

0

25

RefractoryŁischemia

Other severeŁischemia

Other recurrentŁangina

RevascularizationŁprocedure

HeartŁfailure

PlaceboFClopidogrel

126F(2.0%)

85F(1.4%)

RR=0.68FP=0.007

237F(3.8%)

176F(2.8%)

RR=0.74FP=0.003

1442F(22.9%)

1307F(20.9%)

RR=0.91FP=0.01

1431F(22.7%)

1302F(20.8%)

RR=0.92FP=0.03

280F(4.4%)

229F(3.7%)

RR=0.82FP=0.026

20

15

10

5

Other In-Hospital Outcomes

Pat

ien

ts w

ith

Ou

tco

me

(%)



500

·


·

August 16, 2001

·

www.nejm.org


CABG, the use of the study medication was tempo-rarily interrupted for more than five days, and thevast majority of the patients who underwent PTCAreceived a thienopyridine-type antiplatelet agent forabout two to four weeks. In the patients who under-went CABG, the study medication was restarted aftera median of 11 days. Although these interruptions oftherapy with the study medication would tend to re-sult in an underestimate of the difference between theclopidogrel group and the placebo group, they alsopermit us to make useful estimates of the benefits andrisks of clopidogrel when it is used routinely and overthe long term, as compared with a strategy of more

selective and short-term use among those undergoingimplantation of a coronary stent.

Clopidogrel prevented a range of ischemic coronaryevents — among them, myocardial infarction and se-vere and refractory ischemia. Clopidogrel was associ-ated with a trend toward fewer ischemic strokes, andthere was no increase in the rate of hemorrhagic strokethat would offset these benefits. There was a significantreduction in the incidence of heart failure with clo-pidogrel that was of about the same magnitude as thereduction in the incidence of ischemic events, suggest-ing that the reduction of ischemia can prevent heartfailure. The benefits of clopidogrel were observed in

Figure 4.

The Rates and Relative Risks of the First Primary Outcome (Death from Cardiovascular Causes, Nonfatal Myocardial In-farction, or Stroke) in Various Subgroups.The data show the consistency of the benefit of clopidogrel. The dotted line represents the average treatment effect. The size ofeach box is proportional to the number of patients in the individual analysis. An associated myocardial infarction (MI) was definedas a myocardial infarction associated with the episode of pain that occurred before randomization. Because of missing data, sixpatients could not be classified in a risk category. CI denotes confidence interval.

0.4 1.20.6 0.8

Relative Risk (95% CI)

1.0

Clopidogrel better Placebo better

OverallF

Associated MIFNo associated MIF

Male sexFFemale sexF

«65 yr oldF>65 yr oldF

ST-segment deviationFNo ST-segment deviationF

Enzymes elevated at entryFEnzymes not elevated at entryF

DiabetesFNo diabetesF

Low riskFIntermediate riskFHigh riskF

History of revascularizationFNo history of revascularizationF

Revascularization after randomizationFNo revascularization after randomization

125620FF

3283F9279F

F7726F4836F

F6354F6208F

F6275F6287F

F3176F9386F

F2840F9722F

F4187F4185F4184F

F2246F

103160FF

4577F7985

11.4FF

13.7F10.6F

F11.9F10.7F

F07.6F15.3F

F14.3F08.6F

F13.0F10.9F

F16.7F09.9F

F06.7F09.4F18.0F

F14.4F10.7F

F13.9F10.0

CharacteristicNo. ofŁ

PatientsPercentage ofŁ

Patients with Event

09.3FF

11.3F08.6F

F09.1F09.5F

F05.4F13.3F

F11.5F07.0F

F10.7F08.8F

F14.2F07.9F

F05.1F06.5F16.3F

F08.4F09.5F

F11.5F08.1F

FF

ClopidogrelPlacebo





·

August 16, 2001

·

www.nejm.org

·

501

a range of patients, including both patients who wereundergoing revascularization procedures and thosewho were not. The benefits were also observed inthose at low, medium, and high risk of cardiovascularevents and those who were receiving various proventherapies such as aspirin, lipid-lowering drugs, an-giotensin-converting–enzyme inhibitors, and beta-blockers.

13-16

The benefits of clopidogrel were appar-ent as early as the first 24 hours after randomization,indicating that the oral loading dose was rapidly ef-fective. Thereafter, the differences between the twogroups were maintained until the end of the study.

Clopidogrel increased the risk of minor and majorbleeding episodes. For every 1000 patients treatedwith clopidogrel, 6 will require a transfusion. However,there was no excess in bleeding that caused strokes,required surgical intervention or inotropic agents, orcaused permanent disability. Furthermore, the excessrisk of bleeding we observed is similar to that observedwith aspirin alone, as compared with a control, in pre-vious studies and lower than that observed in mosttrials of the short-term intravenous use or the pro-longed oral use of glycoprotein IIb/IIIa receptor in-hibitors.

4,16,17

The risk of bleeding may have been part-ly mitigated by the temporary discontinuation of thestudy medication before surgery. Treatment with clo-

pidogrel was not associated with an excess rate of anyother type of adverse event that necessitated the dis-continuation of the study drug; this finding indicatesthat the combination of clopidogrel and aspirin is aswell tolerated as aspirin alone.

In summary, clopidogrel significantly reduces therisk of the composite outcome of death from cardio-vascular causes, nonfatal myocardial infarction, orstroke, as well as a range of related ischemic events.The use of the drug, in addition to aspirin, is asso-ciated with an increased risk of bleeding.

Supported by Sanofi-Synthelabo and Bristol-Myers Squibb. Dr. Yusuf isthe recipient of a Senior Scientist award from the Canadian Institutes ofHealth Research and holds an endowed chair from the Heart and StrokeFoundation of Ontario. Dr. Mehta is a research fellow of the Heart andStroke Foundation of Canada.

Drs. Yusuf, Mehta, Tognoni, and Fox have received honorariums for ed-ucational activities or have served as consultants to Bristol-Myers Squibband Sanofi-Synthelabo.

We are indebted to Judy Lindeman for secretarial assistance.

APPENDIX

The following persons participated in the CURE trial:

Steering Com-mittee:

S. Yusuf (chair and principal investigator), K.A.A. Fox (cochair),G. Tognoni (cochair), S.R. Mehta (project officer), S. Chrolavicius (studycoordinator), S. Anand, A. Avezum, N. Awan, M. Bertrand, A. Budaj, L.

*The number of patients with bleeding that met the criteria for major bleeding established by theThrombolysis in Myocardial Infarction trial

11

was 68 in the clopidogrel group and 73 in the placebogroup (relative risk, 0.94; 95 percent confidence interval, 0.68 to 1.30; P=0.70). The number withbleeding that met the criteria for life-threatening or severe bleeding established by the Global Utilizationof Streptokinase and Tissue Plasminogen Activator for Occluded Coronary Arteries trial

12

was 78 in theclopidogrel group and 70 in the placebo group (relative risk, 1.12; 95 percent confidence interval, 0.81to 1.55; P=0.48). Some patients had more than one bleeding episode. CI denotes confidence interval.

TABLE 3. BLEEDING COMPLICATIONS.*

VARIABLE

CLOPIDOGREL GROUP

(N=6259)

PLACEBO GROUP

(N=6303)RELATIVE RISK

(95% CI) P VALUE

no. (%)

Major bleedingNecessitating transfusion of »2 units

of bloodLife-threatening

FatalCausing 5 g/dl drop in hemoglobin

levelRequiring surgical interventionCausing hemorrhagic strokeRequiring inotropic agentsNecessitating transfusion of »4 units

of bloodNon–life-threatening

231 (3.7)177 (2.8)

135 (2.2)11 (0.2)58 (0.9)

45 (0.7)7 (0.1)

34 (0.5)74 (1.2)

96 (1.5)

169 (2.7)137 (2.2)

112 (1.8)15 (0.2)57 (0.9)

43 (0.7)5 (0.1)

34 (0.5)60 (1.0)

57 (0.9)

1.38 (1.13–1.67)1.30 (1.04–1.62)

1.21 (0.95–1.56)

1.70 (1.22–2.35)

0.0010.02

0.13

0.002

Site of major bleedingGastrointestinalRetroperitonealUrinary (hematuria)Arterial puncture siteSurgical site

83 (1.3)8 (0.1)4 (0.1)

36 (0.6)56 (0.9)

47 (0.7)5 (0.1)5 (0.1)

22 (0.3)53 (0.8)

Minor bleeding 322 (5.1) 153 (2.4) 2.12 (1.75–2.56) <0.001

Total with bleeding complications 533 (8.5) 317 (5.0) 1.69 (1.48–1.94) <0.001



502 · N Engl J Med, Vol. 345, No. 7 · August 16, 2001 · www.nejm.org


Ceremuzynski, J. Col, P.J. Commerford, R. Diaz, M. Flather, M.-G. Fran-zosi, B. Gersh, W. Grossman, D.A. Halon, D. Hunt, C. Joyner (chair ofthe Events Adjudication Committee), N. Karatzas, M. Keltai, S. Kopecky,B.S. Lewis, A. Maggioni, K. Malmberg, T. Moccetti, J. Morais, M. Nat-arajan, E. Paolasso, R. Peters, L. Piegas, A. Pipilis, J. Pogue, M. Ramos-Corrales, H.-J. Rupprecht, E. Sitkei, V. Valentin, J. Varigos, P. Widimsky,T. Wittlinger, M. Blumenthal, J. Bouthier, C. Gaudin, T. Hess, N. Khurmi,M. Sotty; Canadian Cardiovascular Collaboration (CCC) Project Of-fice: I. Copland, B. Cracknell, C. Demers, J. Eikelboom, B. Morris, J. Pas-quale, V. Yacyshyn, F. Zhao; Data Safety and Monitoring Board: G.Wyse (chair), J. Cairns, R. Hart, J. Hirsh, M. Gent, T. Ryan, J. Wittes; In-vestigators who recruited more than 10 patients (the numbers of pa-tients appear in parentheses): Argentina (512) — R. Ahuad Guerrero, V.Arias, F.H. Bello, J.O. Bono, L. Bujan, A. Caccavo, A.A. Fernandez, J.J.Fuselli, A.J. Gambarte, E.G. Hasbani, E. Marzetti, G. Mon, R. Nordaby,G. Quijano, A. Salvati, E. San Martin, P. Schygiel, A. Sosa Liprandi, H.Torre, E. Tuero; Australia (742) — J. Amerena, J.H.N. Bett, J. Botha, A.Buncle, D. Careless, A. Ewart, A. Fagan, D. Fitzpatrick, P. Garrahy, K.Gunawardane, A. Hamer, A. Hill, N. Jepson, G. Lane, H. LeGood, G. Nel-son, M. Sallaberger, G. Tulloch, D. Owensby, J. Padley, D. Rees, K. Rob-erts, D. Rosen, J. Sampson, B. Singh, C. Singh, R. Taylor, A. Thomson,W. Walsh; Austria (5); Belgium (38) — H. De Raedt, J. Rankin; Brazil(598) — J.A. Abrantes, D.C. de Albuquerque, C. Blacher, L.C. Bodanese,A.C. Carvalho, M. Coutinho, O. Dutra, J.P. Esteves, P. Leães, N.S. Lima,N. Lobo, J.A. Marin Neto, R.L. Marino, J.C. Nicolau, R.C. Pedrosa, I.Pereira Filho, W. Rabelo, A. Rabelo Jr., R.F. Ramos, R.M. Rocha, E.C.Schramm, C.V. Serrano, Jr., V.P. Souza, A. Timerman, R. Vaz, S.S. Xavier;Canada (1761) — P. Auger, K. Barbin, I. Bata, L. Bergeron, R.K. Bharga-va, P. Bogaty, P. Bolduc, T. Boyne, L. Boyer, B. Bozek, T. Calmusky, Y.K.Chan, D. Cole, M. D’Astous, M. David, T. Davies, L. Desjardins, F.C.Dohaney, M. Doody, F. Dumont, C. Fortin, A. Fung, G. Galdreault, P.B.Gervais, J.P. Giannoccaro, G. Gosselin, D. Grandmont, A. Grover, M.Gupta, H. Hink, J.G. Hiscock, D. Hutton, J.W.H. Hynd, A. Iless, C. Kent,A. Kitching, S. Kouz, K. Kwok, M. LaForest, H. Lee, C. Lefkowitz, B.Lubelsky, P. Ma, J.-F. Marquis, A. McCallum, R. Major, B. May, M. Mer-cier, M. Montigny, A. Morris, S. Nawaz, J. Norris, S. Pallie, A. Panju, P.Parekh, C. Patterson, Y. Pesant, C. Pilon, A.R.J. Rajakumar, T. Rebane, J.Ricci, M. Ruel, R. Schuld, R. Starra, M. Thornley, A.S. Weeks, L.H. Wink-ler, G. Wisenberg, I. Witt, K. Woo, E. Yu, R. Zadra, D. Zaniol; CzechRepublic (165) — P. Bocek, M. Branny, V. Cepelak, P. Gregor, L. Groch,P. Jansky, P. Svitil, A. Vaclavicek; Denmark (40) — S. Husted, B. Ziegler;Finland (161) — A. Karhu, J. Mustonen, M. Nieminen, K. Nissinen, K.Peuhkurinen, P. Tuomainen, A. Ylitalo; France (360) — M. Adam, G. Amat,K. Asward, G. Bessede, P. Dambrine, E. Decoulx, B. D’Hautefeuille, J.Dujardin, R. Fouche, P. Fournier, Y. Haftel, J.Y. Ketelers, R. Lallemant, M.Lang, F. Leroy, M. Richard; Germany (778) — W. Dippold, J. Gronholz,J. Harenberg, C. Hilpert, T. Horacek, A. Moritz, H. Neuss, K. Nitsche,H.R. Ochs, M. Plumer-Schmidt, B. Pollock, G. Post, M. Sauer, A.Schmidt, H. Schmitt, H.-O. Schulze, P. Schuster, E.-R. von Leitner; Greece(144) — S. Christakos, K. Karidis, K. Kifnidis, K. Papadopoulos, D. Sym-conidis, A. Tirologos; Hungary (599) — B. Garai, T. Gesztesi, K. Heltai,B. Herczeg, A. Janosi, A. Kadar, E. Kalo, P. Karpati, B. Kiss, A. Nagy, M.Okros, F. Poor, L. Regos, J. Sebo, E. Sitkei, P. Soltesz, J. Stefan, F. Szaboki,S. Timar, Z. Varallyay, K. Veres; Ireland (37) — K. Daly; Israel (694) —G. Cotter, A. Marmor, D. Nazzal, M. Omary, L. Reisin, T. Rosenfeld, M.Shargorodsky, S. Shasha, A. Smulovitz, Z. Vered, R. Zimlichman; Italy(538) — G. Beria, D. Bernardi, E. Bonfanti, V. Ceci, R. De Caterina, S.De Servi, M. DelPinto, A. Di Chiara, G. Di Pasquale, L. Filippucci, M.Galli, E. Gardinale, M.T. Landoni, A. Martoni, F. Mauri, N. Maurea, P.Meneghetti, M. Mennuni, A. Murrone, A. Pani, E. Paolini, P. Passarelli, G.Pettinati, B. Pontillo, L. Rossi, R. Rossi, M. Santini, G. Sabino, P. Tricoci,F.M. Turazza; Mexico (240) — A. Cruz Diaz, A. Garcia-Sosa, I. HernandezSantamaria, J. Leiva Pons, J.N. Medecigo, F.J. Rangez Rojo, C. Jorges-Sanchez Diaz, J. Trujillo Castro, E. Villegas Morales; the Netherlands (499)— S. Baldew, D.C.G. Basart, N. Bijsterfeld, P. Bronzwaer, A. Chardon, M.den Hartoog, H. Heijmen, J. Heijmeriks, G. Jochemsen, H. Klomps, P.Landsaat, H. Michels, A. Moons, M. Pieterse, A. Sadée, G. van Beek, M.Van Hessen, J. Verheul, E. Viergever; New Zealand (89) — M. Audeau, M.Hills, H. Ikram, M. Johnstone, A. Smart; Norway (61) — J. Erikssen, T.Morstel, J. Winther Haerem; Poland (2050) — P. Achremczyk, A. Bara-nowska, P. Burduk, D. Cichon, A. Czepiel, H. Danielewicz, P. Danielewicz,W. Dworzanski, J. Gessek, A. Gorecki, J. Gorny, K. Janik, A. Jedrzejowski,G. Kania, T. Kawka-Urbanek, J. Klaudel, D. Kopcik, A. Kozlowski, W. Kra-sowski, A. Ladach, M. Laniec, J. Maciejewicz, P. Maciejewski, Z. Majcher,S. Malinowski, T. Marczyk, P. Miekus, T. Myczka, M. Ogorek, M. Piepior-ka, K. Religa, Z. Reszka, M. Rozwodowska, W. Smielak-Korombel, D.Smiglak, J. Spyra, S. Stec, D. Susol, M. Szpajer, P. Szymanski, G. Szumczyk-Muszytowska, M. Tomzynski, M. Ujda, W. Wasek, T. Waszyrowski, P.Wojewoda, A. Wnorowska, P. Wrobel, J. Zadrozny, A. Zebrowski, Z. Zie-linski; Portugal (129) — J.M. Bastos, P. Cardoso, M. Carrageta, C. Catari-

no, R. Ferreira, M. Veloso Gomes; South Africa (308) — A. Aboo, L.Bobak, B. Brown, S. Cassim, S. Harripersadh, P. Manga, F. Maritz, J. Marx,D. Myburgh, H. Nell, A. Okreglicki, H. Theron; Spain (330) — N. AlonsoOrcajo, F. Colomina, I. Echanove, J. Garcia Guerrero, E. Gomez-Martinez,I. Lekuona, L. Miralles, C. Pasaval, A. Rovira, J. San Jose, A. RodriguezLlorian, V. Valentin; Sweden (260) — N. Abdon, B. Bartholdson, E. Basilier,I. Johansson, B. Kristensson, K. Larsson, P. Lindell, B. Moller, G. Ras-manis, A. Stjerna, K. Tolhagen; Switzerland (225) — B. Caduff, G. Noseda,A. Rossi, D. Schmidt, J. Zerega; United States (462) — K. Carr, R. Detrano,R. Geer, G. Grewal, P. Hermany, N. Lakkis, W.R. Lewis, T.R. Pacheco, G.Papuchis, J. Puma, S. Runckey, U. Thadani, J. Whitaker; United Kingdom(737) — Y.J.A. Adgey, D. Blackman, M. Brack, A. Bridges, A. Cohen, P.Currie, S. Duffy, I. Findlay, J. Goodfellow, K. Gray, D. Hogg, D. Holdright,D. Hughes, N. Irvine, K. Jennings, P. Keeling, P. Ludman, T. Mathew, P.McIntyre, R. Oliver, N. Palmer, E. Rodrigues, L. Smyth, D. Sprigings, J.Stephens, A. Timmis, R. Vincent.

REFERENCES

1. Yeghiazarians Y, Braunstein JB, Askari A, Stone PH. Unstable angina pectoris. N Engl J Med 2000;342:101-14.2. Antiplatelet Trialists’ Collaboration. Collaborative overview of random-ised trials of antiplatelet therapy. I. Prevention of death, myocardial infarc-tion, and stroke by prolonged antiplatelet therapy in various categories of patients. BMJ 1994;308:81-106. [Erratum, BMJ 1994;308:1540.]3. Low-molecular-weight heparin during instability in coronary artery dis-ease. Lancet 1996;347:561-8.4. The Platelet Receptor Inhibition in Ischemic Syndrome Management in Patients Limited by Unstable Signs and Symptoms (PRISM-PLUS) Inves-tigators. Inhibition of the platelet glycoprotein IIb/IIIa receptor with ti-rofiban in unstable angina and non–Q-wave acute myocardial infarction. N Engl J Med 1998;338:1488-97.5. The GUSTO IV-ACS Investigators. Effect of glycoprotein IIb/IIIa re-ceptor blocker abciximab on outcome in patients with acute coronary syn-dromes without early coronary revascularization: the GUSTO IV-ACS ran-domized trial. Lancet 2001;357:1915-24.6. Chew DP, Bhatt D, Sapp S, Topol EJ. Increased mortality with oral platelet glycoprotein IIb/IIIa antagonists: a meta-analysis of phase III mul-ticenter randomized trials. Circulation 2001;103:201-6.7. FRagmin and Fast Revascularisation during InStability in Coronary artery disease (FRISC II) Investigators. Long-term low-molecular-mass heparin in unstable coronary-artery disease: FRISC II prospective randomised multi-centre study. Lancet 1999;354:701-7. [Erratum, Lancet 1999;354:1478.]8. The Organization to Assess Strategies for Ischemic Syndromes (OASIS) Investigators. Effects of long-term, moderate-intensity oral anticoagulation in addition to aspirin in unstable angina. J Am Coll Cardiol 2001;37:475-84.9. CAPRIE Steering Committee. A randomised, blinded, trial of clopido-grel versus aspirin in patients at risk of ischaemic events (CAPRIE). Lancet 1996;348:1329-39.10. CURE Study Investigators. The Clopidogrel in Unstable angina to prevent Recurrent Events (CURE) trial programme: rationale, design and baseline characteristics including a meta-analysis of the effects of thieno-pyridines in vascular disease. Eur Heart J 2000;21:2033-41.11. Rao AK, Pratt C, Berke A, et al. Thrombolysis in Myocardial Infarc-tion (TIMI) Trial — phase I: hemorrhagic manifestations and changes in plasma fibrinogen and the fibrinolytic system in patients treated with re-combinant tissue plasminogen activator and streptokinase. J Am Coll Car-diol 1988;11:1-11.12. The PURSUIT Trial Investigators. Inhibition of platelet glycoprotein IIb/IIIa with eptifibatide in patients with acute coronary syndromes. N Engl J Med 1998;339:436-43.13. The Heart Outcomes Prevention Evaluation Study Investigators. Ef-fects of an angiotensin-converting–enzyme inhibitor, ramipril, on cardio-vascular events in high-risk patients. N Engl J Med 2000;342:145-53. [Erratum, N Engl J Med 2000;342:748.]14. Yusuf S, Wittes J, Friedman L. Overview of results of randomized clin-ical trials in heart disease. II. Unstable angina, heart failure, primary preven-tion with aspirin, and risk factor modification. JAMA 1988;260:2259-63.15. Tonkin AM, Colquhoun D, Emberson J, et al. Effects of pravastatin in 3260 patients with unstable angina: results from the LIPID study. Lan-cet 2000;356:1871-5.16. Anti-thrombotic Trialists Collaboration. Prevention of death, myocar-dial infarction and stroke by antiplatelet therapy: collaborative meta-analysis of 266 trials involving 200,000 patients at high risk of occlusive vascular disease. BMJ (in press).17. The SYMPHONY Investigators. Comparison of sibrafiban with aspirin for prevention of cardiovascular events after acute coronary syndromes: a randomised trial. Lancet 2000;355:337-45.




a comparison of albumin and saline for fluid resuscitation ... › 2020 › 03 › ...(116 [12%] vs...

Documents