Statin trials and goals of cholesterol-loweringtherapy after AMITerje R. Pedersen, MD Oslo, Norway

Several randomized clinical trials using statins in the prevention of coronary heart disease (CHD) have demonstrated bene-fit, both in terms of retardation of the progression of signs of coronary atherosclerosis and in reduced morbidity and mortalityrates. Three of these trials have examined the long-term effect of statins in patients with previous myocardial infarction. TheScandinavian Simvastatin Survival Study (4S) showed that a mean reduction of low-density-lipoprotein (LDL) cholesterol by35% reduced coronary mortality rates by 42% and total mortality rates by 30%. In the Cholesterol and Recurrent Eventstrial, a 28% reduction in LDL-cholesterol was associated with a reduction in major coronary events of 24%. In the Long TermIntervention with Pravastatin in Ischemic Disease study, the 25% LDL-cholesterol reduction produced a 24% reduction incoronary disease mortality rates and 22% reduction in death from all causes. All event reductions were highly statisticallysignificant. Other trials using statins in patients without signs of CHD have yielded similar risk reductions. Post hoc analysis ofthe results of the trials have produced diverging indications as to what is the optimal goal of cholesterol lowering. Analysis ofthe 4S indicates that aggressive treatment aiming at LDL-cholesterol levels lower than the current recommendations of expertpanels in the United States and in Europe may yield additional benefit. This strategy finds some support in epidemiologicalstudies and in a study with angiographic end points. Analysis of two trials using pravastatin contradict this and conclude thatthere is little or no additional benefit of reducing LDL-cholesterol below 125 mg/dL (3.2 mmol/L). Future studies need toaddress this question prospectively. (Am Heart J 1999;138:S177-S182.)

Three large-scale, randomized, controlled clinical tri-als have recently been published, demonstrating thelong-term benefit of statin therapy after acute myocar-dial infarction (Fig 1). These trial results have put anend to a discussion over several decades on the useful-ness of cholesterol lowering to reduce the risk inpatients with coronary heart disease (CHD). Before thepublication of these trials, expert panels in Europe andthe United States had published recommendations tolower cholesterol in patients at high risk of morbidityand mortality of CHD.1,2 These recommendations werebased on a strong body of evidence from epidemiologi-cal studies but slim evidence from randomized clinicaltrials that cholesterol lowering might yield clinical bene-fit. Now the discussion among scientists and cliniciansis focusing on how to implement the evidence from thetrials in daily clinical practice. In particular, there is stilluncertainty about how aggressive lipid lowering shouldbe. The evidence from the statin trials is not clear at thispoint. In the United States, the Expert Panel has longago recommended that the goal of lipid interventionshould be a level of low-density-lipoprotein (LDL) cho-lesterol of less than 100 mg/dL (2.6 mmol/L).3 This rec-ommendation was later endorsed by the American

Heart Association and American College of Cardiology.4

Only a small proportion of patients included in the 3statin trials in CHD patients reached this goal, and thetrial results are conflicting for LDL-levels below 130mg/dL. In Europe, recommendations of the SecondJoint Task Force of European and other Societies oncoronary prevention are aiming at an LDL-cholesterollevel of <3.0 mmol/L (115 mg/dL).5 This goal is appar-ently less ambitious than in the United States, however,cholesterol levels are generally higher in Europe. Fortotal cholesterol the European target was set at 5.0mmol/L (193 mg/dL).

The Scandinavian SimvastatinSurvival Study (4S)

The first study with a statin and with clinical end pointswas the Scandinavian Simvastatin Survival Study (4S),which randomly allocated 4444 patients with CHD, age35 to 70 years, to treatment with placebo or simvastatin20 mg.6 The simvastatin dosage was titrated up to 40 mgin 37% of the patients who did not reach the target of aserum total cholesterol of 5.2 mmol/L or less during thefirst 18 weeks. After 1 year, 77% of all patients in the sim-vastatin group had reached the target for the study. Onaverage, total cholesterol was reduced by 25% in the sim-vastatin group, with a 35% reduction in LDL-cholesteroland 8% increase in high-density lipoprotein (HDL)-choles-terol, whereas serum triglycerides were reduced 10%from baseline. In the placebo group, serum triglyceridesrose by 7%, whereas each of the other aforementioned

From the University of Oslo, Cardiology Department, Aker Hospital.Reprint requests: Terje R. Pedersen, MD, University of Oslo, Cardiology Depart-ment, Aker Hospital, N 0514 Oslo, Norway.Copyright © 1999 by Mosby, Inc.0002-8703/99/$8.00 + 0 4/0/98555

lipids and lipoproteins rose by 1%. These changes wereassociated with a 42% relative risk reduction of CHDmortality rate (P < .0001), which explained a 30% reduc-tion in all cause-mortality rate (P = .0003).

The patients who needed titration to 40 mg simvas-tatin daily differed from the rest, both in their base-line lipoprotein levels and in the percent reductionachieved with 20 mg.7 Patients needing titration to40 mg had higher baseline total cholesterol thanthose remaining on 20 mg7 (mean 7.1 vs 6.5 mmol/L),higher LDL-cholesterol (mean 5.1 vs 4.7 mmol/L), butalso higher levels of HDL cholesterol (mean 1.25 vs1.14 mmol/L), whereas the triglyceride level was thesame. The percent change in serum lipids andlipoproteins achieved with 20 mg simvastatin in thetitrated group versus those taking 20 mg throughoutwas for total cholesterol –18.6% versus 30.7%, forLDL-cholesterol –26.4% versus 41.4%, and for triglyc-erides –6.5% versus –18.2%, respectively. For HDL-cholesterol the change was +8.9% and +7.7%, respec-tively. After 12 months, the mean changes frombaseline were similar in the group titrated to 40 mgand those taking 20 mg. Thus the titrated group wascharacterized by higher cholesterol levels on baselinebut also a much less response in percent reduction ofLDL-cholesterol and triglycerides. Doubling the sim-vastatin dose was necessary in such patients toachieve the same reduction in lipids and lipoproteinsas the remaining patients.

At 1 year of simvastatin therapy, half the patients had

LDL-cholesterol levels <3.0 mmol/L and a quarter hadlevels <2.6 mmol/L.

There was no differential relative risk reduction ofmajor coronary events with simvastatin in relation toquartiles of baseline serum LDL-cholesterol or otherlipids and lipoproteins.8 In the lowest LDL-cholesterolquartile (3.0 to 4.4 mmol/L) the relative risk of a majorcoronary event was 0.65 (95% confidence interval [CI]0.50 to 0.85) and in the highest quartile (5.4 to 6.8) therelative risk was 0.64 (95% CI 0.51 to 0.81).

The relation between the lipid and lipoprotein levelsachieved after 1 year of treatment in the placebo andsimvastatin groups, and the subsequent risk of a majorcoronary event has been analyzed for all patients,including the 21% who qualified for the study on thebasis of a diagnosis of angina pectoris only and no pre-vious myocardial infarction.7 This method has theadvantage that prognostic importance of a variable atone defined time point can be assessed. It has the disad-vantage that a single measurement only decreases theprecision of the estimate of the true serum level, andtherefore the prognostic estimate tends to be less pre-cise. The alternative method is to take into account alllipid and lipoprotein measurements over the entireduration of the study for the correlation with outcome.Such time-dependent methods, although possiblyimproving the precision of the estimate of the averageserum levels, has the disadvantage of increasing thecomplexity of the analysis and may involve the biasrelated to “using the future to predict the future”.9

American Heart JournalAugust 1999PedersenS178

Figure 1

Rate of CHD death and nonfatal myocardial infarction over the first 5 years of the trials 4S, LIPID, and CARE.In 4S and LIPID, end points comprised silent myocardial infarction, as evidenced by new Q waves in electrocardio-graphic recordings.

The results of the analysis of the relation of 1-year LDL-cholesterol levels and major coronary events is illustratedin Fig 2. There was a near linear relation of the LDL-cho-lesterol level reached at 1 year and the subsequent risk.The range of LDL-cholesterol levels that could beassessed was 2.2 to 6.4 mmol/L. The change and levelsreached of LDL-cholesterol explained most of the effectof treatment. The observed change in triglycerides didnot have an independent predictive relation of riskreduction. It should be noted though, that patients withbaseline serum triglyceride levels >2.5 mmol/L (220mg/dL) were excluded from participation. In patientswith graver forms of hypertriglyceridemia, reductions inserum triglycerides may yield clinical benefit.

The Cholesterol and Recurrent EventsStudy

Patients included in the Cholesterol and RecurrentEvents (CARE) study all had had a myocardial infarc-tion between 3 and 20 months before randomizationand were aged between 21 and 75 years (mean 59years).10 The trial comprised 4159 patients from theUnited States and Canada. They had total cholesterollevels <240 mg/dL (6.2 mmol/L) with LDL-cholesterolin the range of 115 to 174 mg/dL (3.0 to 4.5 mmol/L)and a serum triglyceride level <350 mg/dL (4.0mmol/L). The placebo-controlled study treatment wasa fixed dose of 40 mg pravastatin daily. Pravastatintherapy produced mean reductions in total, LDL-cho-lesterol, and triglycerides of 20%, 28%, and 14%,

respectively, and increased HDL-cholesterol 5%. Onaverage, approximately 60% of the patients achievedan LDL-cholesterol level <100 mg/dL (2.6 mmol/L).These changes in lipids and lipoproteins were associ-ated with a relative risk reduction in the principalend point, fatal CHD or confirmed myocardial infarc-tion of 24% in comparison with the placebo group (P= .003). All-cause mortality rates and death from CHDwere not reduced significantly in comparison withplacebo, but the changes were in the direction favor-ing pravastatin therapy.

The CARE trial included patients with lower levels ofcholesterol than is typical for CHD patients in Europe.In a quartile analysis, the beneficial effect of pravastatinon the primary study endpoint was confined to thethree upper quartile of baseline LDL-cholesterol10 andapparently no clinical benefit was observed in patientswith LDL-cholesterol in the range of 115 to 125 mg/dL(3.0 to 3.2 mmol/L). The number of patients with endpoints in this quartile was 93 and 89 in the placebo andpravastatin groups, respectively, and the 95% CI for therelative risk of 1.03 was 0.77 to 1.38. A potential true23% risk reduction was thus not reliably excluded.

The relation between levels of LDL-cholesterolreached during the trial and the risk of a CHD eventwas different compared with the 4S results.11 In thisanalysis, a time-dependent model was used with updat-ing of the lipid concentrations for each patient through-out the trial. The results were presented in deciles ofLDL-cholesterol concentrations reached in the pravas-tatin group that were related to the relative risk of an

American Heart JournalVolume 138, Number 2, Part 2 Pedersen S179

Figure 2

4S results showing relation between LDL-cholesterol levels at year 1 and subsequent risk of major coronary eventsin the placebo and simvastatin groups with 95% CIs. Reproduced from Pedersen TR, Olsson AG, Færgeman O,et al. Lipoprotein changes and reduction in the incidence of major coronary heart disease events in the ScandinavianSimvastatin Survival Study (4S). Circulation 1998;97:1453-60.

expanded endpoint consisting of CHD mortality,nonfatal myocardial infarction, coronary artery bypasssurgery, or percutaneous transluminal coronary angio-plasty (Fig 3). The analysis comprising 430 patientswith end point indicates that reduction in LDL-choles-terol below 125 mg/dL (3.2 mmol/L) does not yield fur-ther reduction in event rate.

The Long-Term Intervention WithPravastatin in Ischemic DiseaseStudy

A second study using 40 mg pravastatin comparedwith placebo was the Long-term Intervention withPravastatin in Ischemic Disease (LIPID) study, comprising9014 patients with CHD, 64% qualifying for the study onthe basis of a previous myocardial infarction.12 In thisstudy, baseline total cholesterol levels ranged from 4.0 to7.0 mmol/L (155 to 271 mg/dL) and triglycerides were<5.0 mmol/L (443 mg/dL). The average change in plasmalipid levels in the pravastatin group over the first 5 yearsof the 6-year study included a reduction in total choles-terol of 18%; LDL-cholesterol fell by 25% and triglyc-erides by 11% whereas HDL-cholesterol rose by 5% morethan in the placebo group. Toward the end of the study,

24% of the patients assigned to placebo had started open-label therapy with cholesterol-lowering drugs.

The primary end point of the study, death from CHD,was reduced by 24% (P < .0001) and all-cause mortalityby 22% (P < .0001).

In the LIPID study no analysis of the relation betweenlipid changes and risk reduction has been published tothis date. In the first report on the main results,12 a sub-group analysis of the relation of baseline LDL-choles-terol and the effects on major coronary events showeda no-significant 16% risk reduction in patients with LDL-cholesterol <135 mg/dL (3.5 mmol/L). This subcategorycomprised 29% of all patients and 27% of all major coro-nary events. In an expanded analysis of the relationshipbetween CHD events and baseline cholesterol pre-sented at the Scientific Sessions of the American Col-lege of Cardiology in 1998, similar reductions in riskwas observed in all quintiles, with a 23.7% reduction inthe lowest quintile (overall reduction was 24% for thiscombined end point).

Other studies

Several other studies with cholesterol-lowering thera-pies have examined the relation of the reduction

American Heart JournalAugust 1999PedersenS180

Figure 3

CARE study results showing LDL-cholesterol concentration during follow-up in the pravastatin group in decileswith 95% CI and the relative risk of the expanded end point CHD death, nonfatal myocardial infarction,coronary artery bypass grafting, or percutaneous transluminal coronary angioplasty. Reproduced from SacksFM, Moyé LA, Davis BR, et al. Relationship between plasma LDL concentrations during treatment with pravastatinand recurrent coronary events in the Cholesterol and Recurrent Events Trial. Circulation 1998;97:1446-52.

achieved in lipid levels and the reduction in risk. ThePost Coronary Artery Bypass Graft Trial examined theadditional benefit of aggressive lowering of LDL-choles-terol with lovastatin and, if needed, cholestyramine, incomparison with a moderate treatment.13 In the aggres-sively treated group, mean LDL-cholesterol ranged from93 to 97 mg/dL (2.4 to 2.5 mmol/L) during the study,compared with 132 to 136 mg/dL (3.4 to 3.5 mmol/L)in the moderately treated group. In the aggressivelytreated group the 676 patients had significantly lessprogression of angiographically determined coronarygraft stenosis and occlusions and fewer new lesionsthan the moderately treated group of 675 patients.

It is interesting to note that regression analyses simi-lar to those carried out in the 4S and the CARE trial havebeen performed in the Lipid Research Clinics Primary Pre-vention Trial (LRC-CPPT)14 and in the Helsinki HeartStudy.15 The results were quite similar to those in 4S. Withthe use of various statistical techniques in 4S, 1% reduc-tion in LDL-cholesterol was associated with 1.5% to 2.3%reduction in risk. In the LRC-CPPT, 1% LDL-cholesterolreduction was associated with 1.9% reduction in risk (95%CI –3.0 to –0.8), and in the Helsinki Heart Study the corre-sponding risk reduction was 2.3% (95% CI –4.4 to –0.1).

In the West of Scotland Coronary Prevention Study(WOSCOPS) with the use of 40 mg pravastatin daily inhypercholesterolemic men, no further benefit of ther-apy was demonstrated beyond a 24% reduction in LDL-cholesterol.16 In this study, the number of end pointswithin each quintile of pravastatin-treated patients wassmaller than in the secondary prevention studies withstatins, and consequently the statistical power to detectdifferential effects within subgroups was small.

In a comprehensive meta-analysis of cholesterol low-ering including results of several statin trials in primaryand secondary prevention of CHD, Gould et al17 founda linear relation between percent reduction in choles-terol and reduction in the odds ratio of both total andCHD mortality, without any sign of a threshold effect.

A large number of epidemiological studies fromseveral countries have shown a clear curve-linear(exponential) relation of cholesterol levels and the riskof CHD.18 This appears to be the case even in regionswhere cholesterol levels are much lower than in theWestern societies. Chen et al19 found no sign of athreshold below which the relative risk of CHD mortal-ity leveled out, all the way down to a total cholesterolof <3.5 mmol/L. In some rural regions of the PeoplesRepublic of China where mean non–HDL-cholesterollevels were as low as 1 mmol/L (39 mg/dL), deathsfrom CHD comprised less than 1% of all deaths.20

Discussion

Patients with myocardial infarction continue to haveworse prognosis than those without, despite combina-

tion therapies that include statins, β-blockers, aspirin,or angiotensin-converting enzyme inhibitors. Whethermore aggressive lipid-lowering therapy than used in theclinical trials to date will improve prognosis remainsuncertain. There is compelling evidence from epidemi-ological trials that CHD is rare when cholesterol levelsare extremely low. However, coronary atherosclerosistoday is regarded as a multifactorial disease character-ized by inflammatory changes of the vascular wall.Although oxidized LDL-cholesterol is probably the maintrigger of such endothelial damage, the process mayonly be slowed down and not entirely prevented orhealed with statin therapy after having progressed overseveral decades.

The post hoc analyses performed within the statintrials cannot deliver reliable and final proof of theoptimal levels or percent reductions of cholesterolfor CHD patients. What is needed now are random-ized clinical trials in CHD patients, comparing a muchmore aggressive treatment strategy with those cur-rently applied by most clinicians. Three such trialshave been started in 1998 to 1999. The SEARCHstudy carried out in the United Kingdom will exam-ine the potential benefit of 80 mg simvastatin dailycompared with 20 mg. The TNT study carried out inseveral countries world wide is comparing atorvas-tatin 80 mg with 10 mg. In Scandinavia, the IDEALstudy will compare atorvastatin 80 mg daily with aregimen similar to that used in 4S: 20 mg simvastatin,titrated to 40 mg daily in patients not reaching thegoal of a total cholesterol of 5.0 mmol/L. These trials,comprising approximately 30,000 patients with CHDwill hopefully provide an answer to the importantquestion whether “lower is better” within the next 5years. Until the results of these trials become avail-able, clinicians are best advised to adhere to the rec-ommendations of the regional expert panels and sci-entific societies.

References1. Expert Panel of the National Cholesterol Education Program

(NECP). Report of the National Cholesterol Education ProgramExpert Panel on detection, evaluation, and treatment of high bloodcholesterol in adults. Arch Intern Med 1988;148:36-9.

2. Pyörälä K, De Backer G, Graham I, et al. Prevention of coronaryheart disease in clinical practice: recommendations of the TaskForce of the European Society of Cardiology, European Atheroscle-rosis Society and European Society of Hypertension. Eur Heart J1994;15:1300-31.

3. Expert Panel on Detection Evaluation, and Treatment of High BloodCholesterol in Adults. Summary of the second report of the nationalcholesterol education program (NCEP) Expert Panel on Detection,Evaluation, and Treatment of High Blood Cholesterol in Adults(Adult Treatment Panel II). JAMA 1993;269:3015-23.

4. Smith SC, Blair SN, Criqui MH, et al. Preventing heart attack anddeath in patients with coronary disease. Circulation 1995;92:2-4.

5. Wood D, De Backer G, Faergeman O, et al. Prevention of coro-

American Heart JournalVolume 138, Number 2, Part 2 Pedersen S181

nary heart disease in clinical practice: recommendations of the Sec-ond Joint Task Force of European and other societies on coronaryprevention. Eur Heart J 1998;19:1434-503.

6. Scandinavian Simvastatin Survival Study Group. Randomized trialof cholesterol lowering in 4444 patients with coronary heart dis-ease: the Scandinavian Simvastatin Survival Study (4S). Lancet1994;344:1383-9.

7. Pedersen TR, Olson AG, Faergeman O, et al. Lipoprotein changesand reduction in the incidence of major coronary heart diseaseevents in the Scandinavian Simvastatin Survival Study (4S). Circula-tion 1998;97:1453-60.

8. Scandinavian Simvastatin Survival Study Group. Baseline serumcholesterol and treatment effect in the Scandinavian Simvastatin Sur-vival Study (4S). Lancet 1995;345:1274-5.

9. Wolfe RA, Strawderman RL. Logical and statistical fallacies in theuse of Cox regression models. Am J Kidney Dis 1996:27:124-9.

10. Sacks FM, Pfeffer MA, Moye LA, et al. The effect of pravastatin oncoronary events after myocardial infarction in patients with averagecholesterol levels. N Engl J Med 1996;335;1001-9.

11. Sacks FM. Moyé LA, Davis BR, et al. Pfeffer treatment with pravas-tatin and recurrent plasma LDL concentrations during treatment withpravastatin and recurrent coronary events in the Cholesterol andRecurrent Events trial. Circulation 1998;97:1446-52.

12. The Long-term Intervention with Pravastatin in Ischemic Disease (LIPID)study group. Prevention of cardiovascular events and death withpravastatin in patients with coronary heart disease and a broadrange of initial cholesterol levels. N Engl J Med 1998;339:1349-57.

13. The Post Coronary Artery Bypass Graft Trail Investigators. The effectof aggressive lowering of low-density lipoprotein cholesterol levelsand low-dose anticoagulation on obstructive changes in saphenous-vein coronary-artery bypass grafts. N Engl J Med 1997;336:153-62.

14. Lipid Research Clinics Program. The Lipid Research Clinics Coro-nary Primary Prevention Trial Results, II: the relationship of reductionin incidence of coronary heart disease to cholesterol lowering.JAMA 1984;251:365-74.

15. Manninen V, Elo MO, Frick MH, et al. Lipid alterations and declinein the incidence of coronary heart disease in the Helsiniki HeartStudy. JAMA 1988;260:641-51.

16. West of Scotland Coronary Prevention Study Group. Influence ofpravastatin and plasma lipids on clinical events in the West of Scot-land Coronary Prevention Study (WOSCOPS). Circulation1998;97:1440-5.

17. Gould AL, Rossouw ER, Santanello NC, et al. Cholesterol reductionyields clinical benefit: impact of statin trials. Circulation 1998;97:946-52.

18. Law MR, Wald NJ, Thompson SG. By how much and how quicklydoes reduction in serum cholesterol concentration lower risk ofischaemic heart disease? BMJ 1994;308:367-72.

19. Chen Z, Peto R, Collins R, et al. Serum cholesterol concentrationand coronary heart disease in population with low cholesterol con-centrations. BMJ 1991;303:276-82.

20. Chen JS, Campbell TC, Li JY, et al. Diet lifestyle and mortality inChina: a study of the characteristics of 65 Chinese counties.Oxford: Oxford University Press. 1990.

American Heart JournalAugust 1999PedersenS182