insights into the contemporary epidemiology and outpatient management of congestive heart failure

8
nosis, and more than 40% of patients with CHF require readmission within 3 to 6 months of hospital discharge. 7,8 Although significant advances have been made in determining both the pathophysiology and therapy for CHF, there has been surprisingly little change in mortality rates over the past 4 decades. 5 Although this may be caused by recent demographic shifts and higher proportions of older and sicker patients, 9,10 a second possibility may be that proven efficacious therapies for CHF, particularly angiotensin converting enzyme (ACE) inhibitors, have not been maximally implemented. Innovative specialized clinics providing intensive out- patient ambulatory care for patients with CHF have been established to enhance the appropriate use of proven efficacious therapies to bring about the desired benefits. At the University of Alberta Hospital, a multidisciplinary Heart Function Clinic was established in 1989, with a mandate to monitor patients with CHF closely, to apply maximally proven efficacious therapy, and to provide counseling, education, and intensive follow-up. This study was undertaken to evaluate our experience in this clinic and to investigate the epidemiology, management, and prognosis of CHF in the 1990s, with particular atten- Congestive heart failure (CHF) is a major clinical and public health problem with an increasing incidence and prevalence and currently affects 1% to 2% of the North American adult population. 1,2 It is associated with poor functional capacity, decreased quality of life, and increased morbidity and mortality risk. 3-8 Mortality rates in excess of 40% within 2 years of initial diagnosis 4,5 are commonly reported. Patients with CHF require frequent hospitalization; 19% of the patients in the SOLVD Reg- istry 6 were hospitalized within 1 year of the initial diag- From The a Division of General Internal Medicine and c Division of Cardiology, Uni- versity of Alberta, and the b Division of General Internal Medicine, University of Ottawa. Drs McAlister and Teo are supported by the Alberta Heritage Foundation for Med- ical Research; Dr McAlister is supported by the Medical Research Council of Canada. Submitted May 7, 1998; accepted August 7, 1998. Reprint requests: Koon K. Teo, MD, Division of Cardiology, 2C2 Walter Mackenzie Centre, University of Alberta Hospital, Edmonton, Alberta, Canada T6G 2B7. E-mail: [email protected] Copyright © 1999 by Mosby, Inc. 0002-8703/99/$8.00 + 0 4/1/96664 Insights into the contemporary epidemiology and outpatient management of congestive heart failure Finlay A. McAlister, MD, FRCPC, a,b Koon K. Teo, MB, PhD, FRCPC, FACC, c Muba Taher, BMSc, c Terrence J. Montague, MD, FRCPC, FACC, c Dennis Humen, MD, FRCPC, FACC, c Lawrence Cheung, MD, a Mercedeh Kiaii, MD a Rita Yim, MHSA, c and Paul W. Armstrong, MD, FRCPC, FACC c Edmonton, Alberta, and Ottawa, Ontario, Canada Objectives To evaluate the epidemiology, prognosis, and patterns of practice in patients with chronic congestive heart failure (CHF) treated and followed at a specialized clinic. Methods Prospective cohort study of consecutive patients referred to and followed up in a specialized heart failure clinic between September 1989 and March 1996. Results Of the 628 patients referred, 566 were confirmed to have CHF. Mean duration of follow-up was 518 ± 490 days (range 1 to 2192 days). Vital status was available for 99.3% of patients. Mean age at enrollment was 66 years, 68% were men, 67% had an ischemic cause of heart disease, and 78% had systolic dysfunction. Patients with preserved systolic function were older, more often female, had higher mean systolic blood pressures, and a lower prevalence of ischemic heart disease, ventricular arrhythmias, or impaired renal function when compared with those with systolic dysfunction (all P .001). Although there was a significant negative trend in survival with decreasing ejection fraction (P = .03), the survival experience of those with CHF and preserved systolic function did not significantly differ from those with systolic failure (P = .25). Multiple logistic regression analysis showed increased mortality risk was associated with increasing age, New York Heart Association class IV, ischemic cause of disease, elevated serum creatinine level, use of diuretics, and systolic dysfunc- tion, whereas use of β-blockers was associated with reduced risk. Conclusions Our data suggest that a specialized outpatient clinic can improve practice patterns in patients with CHF. The high mortality risk in CHF with preserved systolic function suggests the need to find efficacious (and effective) therapies for this condition. (Am Heart J 1999;138:87-94.) See related Editorial on page 5.

Upload: finlay-a-mcalister

Post on 31-Oct-2016

212 views

Category:

Documents


0 download

TRANSCRIPT

nosis, and more than 40% of patients with CHF requirereadmission within 3 to 6 months of hospitaldischarge.7,8 Although significant advances have beenmade in determining both the pathophysiology andtherapy for CHF, there has been surprisingly littlechange in mortality rates over the past 4 decades.5

Although this may be caused by recent demographicshifts and higher proportions of older and sickerpatients,9,10 a second possibility may be that provenefficacious therapies for CHF, particularly angiotensinconverting enzyme (ACE) inhibitors, have not beenmaximally implemented.

Innovative specialized clinics providing intensive out-patient ambulatory care for patients with CHF have beenestablished to enhance the appropriate use of provenefficacious therapies to bring about the desired benefits.At the University of Alberta Hospital, a multidisciplinaryHeart Function Clinic was established in 1989, with amandate to monitor patients with CHF closely, to applymaximally proven efficacious therapy, and to providecounseling, education, and intensive follow-up. Thisstudy was undertaken to evaluate our experience in thisclinic and to investigate the epidemiology, management,and prognosis of CHF in the 1990s, with particular atten-

Congestive heart failure (CHF) is a major clinical andpublic health problem with an increasing incidence andprevalence and currently affects 1% to 2% of the NorthAmerican adult population.1,2 It is associated with poorfunctional capacity, decreased quality of life, andincreased morbidity and mortality risk.3-8 Mortality ratesin excess of 40% within 2 years of initial diagnosis4,5 arecommonly reported. Patients with CHF require frequenthospitalization; 19% of the patients in the SOLVD Reg-istry6 were hospitalized within 1 year of the initial diag-

From The aDivision of General Internal Medicine and cDivision of Cardiology, Uni-versity of Alberta, and the bDivision of General Internal Medicine, University ofOttawa.Drs McAlister and Teo are supported by the Alberta Heritage Foundation for Med-ical Research; Dr McAlister is supported by the Medical Research Council ofCanada.Submitted May 7, 1998; accepted August 7, 1998.Reprint requests: Koon K. Teo, MD, Division of Cardiology, 2C2 Walter MackenzieCentre, University of Alberta Hospital, Edmonton, Alberta, Canada T6G 2B7.E-mail: [email protected] © 1999 by Mosby, Inc.0002-8703/99/$8.00 + 0 4/1/96664

Insights into the contemporary epidemiology andoutpatient management of congestive heart failureFinlay A. McAlister, MD, FRCPC,a,b Koon K. Teo, MB, PhD, FRCPC, FACC,c Muba Taher, BMSc,c Terrence J.Montague, MD, FRCPC, FACC,c Dennis Humen, MD, FRCPC, FACC,c Lawrence Cheung, MD,a Mercedeh Kiaii, MDa

Rita Yim, MHSA,c and Paul W. Armstrong, MD, FRCPC, FACCc Edmonton, Alberta, and Ottawa, Ontario, Canada

Objectives To evaluate the epidemiology, prognosis, and patterns of practice in patients with chronic congestive heartfailure (CHF) treated and followed at a specialized clinic.

Methods Prospective cohort study of consecutive patients referred to and followed up in a specialized heart failureclinic between September 1989 and March 1996.

Results Of the 628 patients referred, 566 were confirmed to have CHF. Mean duration of follow-up was 518 ± 490days (range 1 to 2192 days). Vital status was available for 99.3% of patients. Mean age at enrollment was 66 years, 68%were men, 67% had an ischemic cause of heart disease, and 78% had systolic dysfunction. Patients with preserved systolicfunction were older, more often female, had higher mean systolic blood pressures, and a lower prevalence of ischemic heartdisease, ventricular arrhythmias, or impaired renal function when compared with those with systolic dysfunction (all P ≤.001). Although there was a significant negative trend in survival with decreasing ejection fraction (P = .03), the survivalexperience of those with CHF and preserved systolic function did not significantly differ from those with systolic failure (P =.25). Multiple logistic regression analysis showed increased mortality risk was associated with increasing age, New YorkHeart Association class IV, ischemic cause of disease, elevated serum creatinine level, use of diuretics, and systolic dysfunc-tion, whereas use of β-blockers was associated with reduced risk.

Conclusions Our data suggest that a specialized outpatient clinic can improve practice patterns in patients with CHF.The high mortality risk in CHF with preserved systolic function suggests the need to find efficacious (and effective) therapiesfor this condition. (Am Heart J 1999;138:87-94.)

See related Editorial on page 5.

tion to comparisons between patients with left ventricu-lar systolic dysfunction and preserved left ventricularfunction.

MethodsPatients

Between September 1989 and December 1995, 628 adultpatients were referred by family physicians, internists, and car-diologists to the University of Alberta Hospital Heart FunctionClinic for assessment or management of chronic CHF. Thisclinic is managed by a team of physicians, specialized nursepractitioners, pharmacists, dieticians, and social workers incooperation with primary care physicians. Diagnosis of CHFwas made by the Heart Function Clinic physicians on the basisof the clinical criteria published by the Framingham HeartStudy.11 For the purpose of this study, 27 patients who werediagnosed by the clinic physicians with a non-CHF conditionand 35 patients with incomplete information in the database ororiginal charts were excluded from further analysis. The finalstudy sample was composed of 566 patients. Management ofeach patient was at the discretion of the attending clinic physi-cian and, although the maximal application of ACE inhibitortherapy was stressed, no explicit protocols or critical pathwayswere used during the period of this study.

All patients were followed up until March 31, 1996, withdemographic, clinical, and laboratory data collected in aprospective fashion at baseline and each subsequent clinicvisit and entered into a computerized database. Vital status asof March 31, 1996, was available from the clinic charts or anadministrative database (Alberta Health Registries, Vital Statis-tics) for 99.2% of the study sample (562 of 566 patients).

Study variablesObjective assessments (by echocardiography, radionuclide,

or contrast ventriculography) of left ventricular ejection frac-tion (LVEF) were obtained in all patients within 3 months oftheir initial presentation to the clinic. Although the precisionof these methods may differ, data from the SOLVD trialsrevealed that each method gave comparable prognostic infor-mation for mortality.12 As we used LVEF to categorize patientsrather than as an end point and as a reflection of actual prac-tice, it was appropriate to accept measurements by any one ofthese methods. Patients were defined as having CHF causedby systolic dysfunction if their LVEF was less than 45%. Ifpatients had an LVEF ≥45% or if the tests indicated that thepredominant problem was in ventricular diastolic relaxation,they were classified as having “preserved systolic left ventricu-lar function.” This term was used because not all patients ful-filled the strict diagnostic criteria required for diastolic dys-

American Heart JournalJuly 1999McAlister et al88

PreservedOverall Systolic dysfunction systolic function

(n = 566)* (n = 441) (n = 121)

Age (mean ± SD) 66 ± 14 65 ± 14 69 ± 14†

Male 68 72 50†

CauseIschemia 67 74 43†

Hypertension 9 4 28†

Other 24 22 29NYHA class

I 12 12 12II 35 36 31III 38 38 41IV 15 15 16

Concomitant conditionsDiabetes mellitus 22 20 21Systemic hypertension 28 24 40†

Prior myocardial infarction 51 56 31†

Ventricular arrhythmia 18 21 6†

Atrial arrhythmia 29 28 36Left ventricular hypertrophy on electrocardiogram 23 23 25Left bundle branch block 37 40 28

Clinical featuresSystolic blood pressure (mm Hg, mean ± SD) 123 ± 24 120 ± 22 132 ± 28†

Diastolic blood pressure (mm Hg, mean ± SD) 75 ± 12 75 ± 11 76 ± 14Heart rate (mean ± SD) 80 ± 15 80 ± 15 78 ± 16Peripheral edema present 29 28 33Serum sodium (mmol/L) 139 ± 4 139 ± 4 139 ± 3Serum potassium (mmol/L) 4.4 ± 0.5 4.4 ± 0.5 4.3 ± 0.5Serum creatinine (µmol/L) 128 ± 70 133 ± 77 113 ± 35†

Serum hemoglobin (g/L) 123 ± 37 127 ± 14 119 ± 32

*Includes 4 patients with undetermined dysfunction type.†P = .001 for comparison between patients with systolic dysfunction and preserved systolic function.

Table I. Percentage distribution of demographic and clinical variables at baseline in 566 Heart Function Clinic patients

function. The causes of CHF were classified as ischemia (his-tory of myocardial infarction, coronary revascularization, doc-umented stenosis >75% of at least 1 coronary artery, or focalareas of akinetic myocardium detected on echocardiographyor contrast ventriculography), hypertension (history of hyper-tension before development of CHF and no evidence ofmyocardial ischemia), or other (including idiopathic dilatedcardiomyopathy, valvular heart disease of any cause, history ofexposure to cardiotoxic drugs or excess alcohol consumptionwith objective evidence of dilated cardiomyopathy and noischemic cause). Ventricular arrhythmias were defined as ahistory of cardiac arrest, documented ventricular tachycardiaor fibrillation, or frequent premature ventricular complexes(>5/min) detected by 24-hour Holter monitor or electrocardio-gram. Atrial arrhythmias were defined as atrial fibrillation,multifocal atrial tachycardia, or atrial flutter. The diagnosis ofleft ventricular hypertrophy was determined by electrocardio-gram on the basis of the Romhilt-Estes criteria. Causes ofdeath were determined from death certificates or review ofmedical records and were classified as cardiac (including sud-den death, progressive pump failure, or acute myocardialinfarction) or noncardiac.

The study protocol was reviewed and approved by theResearch Ethics Board of the University of Alberta.

Data collection/assuranceThe main study variables collected were patient demo-

graphics, cause and pathophysiologic characteristics of CHF,New York Heart Association (NYHA) functional classification,presence or absence of concomitant conditions (diabetes mel-litus, hypertension, prior myocardial infarction, ventricular oratrial arrhythmias, left ventricular hypertrophy by electrocar-

diogram, left bundle branch block), clinical features (systolicand diastolic blood pressure, heart rate, presence of periph-eral edema), results of laboratory investigations (serumsodium, potassium, creatinine, hemoglobin levels), use of car-diovascular medications, and death. These variables wereprospectively collected and entered into a database. The studydata were retrospectively extracted from this database by acomputer-based data acquisition tool and a priori defined cri-teria. In those cases in which the database was incomplete,the original charts were reviewed to extract the relevant data.Quality assurance, collation, and analysis of the data weredone at the Epidemiology Coordinating Research Centre, Divi-sion of Cardiology, University of Alberta.

Data analysisIn addition to the description of overall practice patterns

and survival data, 2 principal null hypothesis were tested: (1)there are no differences in treatment received betweenpatients with systolic dysfunction and preserved systolic func-tion, and (2) there is no relation between demographic or ther-apeutic variables and death among patients attending the HeartFunction Clinic. The baseline characteristics of the systolic dys-function and preserved systolic function groups were com-pared by the chi-square test and, because multiple compar-isons were planned, the Bonferroni correction was used andstatistical significance was assumed at P < .01. The firsthypothesis was tested by chi-square analyses for each thera-peutic agent; the Bonferroni correction was again used toderive statistical significance at P < .002. To adjust for baselineclinical risk factors and presenting characteristics, the secondhypothesis was tested by multiple logistic regression analysiswith the forward stepwise technique and entering all the pre-

American Heart JournalVolume 138, Number 1, Part 1 McAlister et al 89

Systolic PreservedOverall dysfunction systolic function

(n = 1681 visits)* (n = 1325 visits) (n = 352 visits)

ACE inhibitors 83 86 71†

Enalapril 50 53 40Captopril 8 8 7Lisinopril 23 24 21Other 2 1 3

DiureticsLoop 81 82 74†

Metolazone 15 15 13Thiazide 8 6 14†

Potassium-sparing 5 5 6Aspirin 59 62 48†

Digoxin 58 60 54Nitrates 47 48 41Potassium supplements 41 40 44Warfarin 37 37 34β-Blocker 26 24 33†

Calcium channel blocker 16 14 25†

Magnesium supplements 10 11 5Amiodarone 6 8 2†

Other antiarrhythmics 5 5 4

*Includes 4 visits by patients with undetermined dysfunction type.†P < .001 for comparison between patients with systolic dysfunction and preserved systolic function.

Table II. Frequency (percentage) of medication prescriptions for 566 patients with CHF at 1681 outpatient visits

specified factors with P < .25 on univariate analysis. Age wasrun as a continuous variable in the analysis and it was decideda priori to repeat the analysis dichotomizing age into “youngerthan 70 years” or “70 years or older” if age was significant inthe first model. Survival curves were generated with theKaplan-Meier method and compared by the Mantel log-ranktest. All analyses were performed with SPSS and SAS software(SAS Institute). For the multiple logistic regression and survivalanalyses, statistical significance was accepted at P < .05.

ResultsMean duration of follow-up of the 566 patients was

518 ± 490 days (median 373 days; range 1 to 2192days). Baseline assessments of LVEF were obtained in562 (99%) patients by echocardiography (49%),radionuclide ventriculography (35%), or contrast ven-triculography (15%). The majority of patients (65%) hadmoderate to severe systolic dysfunction (LVEF <40%),but 22% patients had ejection fractions ≥45% and thesepatients were categorized as having CHF associatedwith preserved systolic function.

Demographic and clinical variablesDistribution of baseline demographic and clinical

variables in the study population is shown in Table I.Ischemia and hypertension were the most commoncauses for CHF in these patients, whereas idiopathiccardiomyopathy (13%), valvular disease (7%), and alco-holic cardiomyopathy (2%) were much less common.Compared with those with systolic dysfunction,patients with preserved systolic function were older

at presentation, more often female, had a higherprevalence of prior hypertension, but a lower preva-lence of ischemic heart disease as the primary cause(all P = .001). This group included a few patients whohad CHF marked by acute decompensation associatedwith transient myocardial ischemia, but there were nopatients with hypertrophic or restrictive cardiomyopa-thy. There were no significant differences betweenthe groups with respect to NYHA class at presenta-tion, but the patients with preserved systolic functionhad higher systolic blood pressures and a lower inci-dence of ventricular arrhythmia and renal dysfunction(all P = .001).

Medication useThe prescribing patterns differed for the 2 groups of

patients (Table II): there was less use of ACE inhibitors,diuretics, aspirin, and amiodarone and more use of β-blockers and calcium channel blockers in patients withpreserved systolic function versus systolic dysfunction.Overall 83% of our patients received an ACE inhibitor.The most commonly prescribed (and their mean dailydose) were enalapril (10.7 mg/day), lisinopril (10.3mg/day), and captopril (62.1 mg/day). Of patients pre-scribed digoxin, only 44% had atrial arrhythmia; 42% ofpatients with preserved systolic function and 59% ofpatients with systolic dysfunction who receiveddigoxin were in sinus rhythm. Anticoagulation withwarfarin for atrial arrhythmia was received by 61% ofthe 44 patients with preserved systolic function and

American Heart JournalJuly 1999McAlister et al90

Figure 1

Survival analysis for 566 patients with CHF, by NYHA class at presentation.

57% of the 122 patients with systolic dysfunction whohad this rhythm.

OutcomesDuring follow-up, 148 patients died, 122 (82%) from

cardiac causes and 26 (18%) from noncardiac causes.Survival analysis (Figure 1) revealed 1-year survival ratesof 95% for NYHA class I, 93% for class II, 83% for classIII, and 70% for class IV patients. At 2-year the survivalrates were 87% for NYHA class I, 83% for class II, 69%for class III, and 52% for class IV. At 3-year they were84%, 77%, 60%, and 34%, respectively. Although therewas a significant negative trend (P = .03) in survival asthe LVEF decreased, the survival curves (Figure 2) inpatients with systolic dysfunction or preserved systolicfunction were not significantly different (P = .25): the1-, 2-, and 3-year survival rates (systolic dysfunction vspreserved systolic function) were 83% versus 88%, 69%versus 77%, and 62% versus 66%, respectively. More-over, in those patients with an ischemic cause, the sur-vival curves of the 2 groups were almost identical (P =.95), with 1-, 2-, and 3-year survival rates (systolic dys-function versus preserved systolic function) of 82% ver-sus 83%, 66% versus 66%, and 59% versus 60%, respec-tively. Also, the survival experience of the 2 groups wasnot significantly different (P = .38) in those patientswith nonischemic causes: 87% versus 92% at 1 year,77% versus 85% at 2 years, and 70% versus 70% at 3years (systolic dysfunction vs preserved systolic func-

tion). Finally, although those patients with an ischemiccause appeared to have an increased mortality riskcompared with those with a nonischemic cause, thedifference did not reach statistical significance in eitherthe systolic dysfunction group (P = .12) or the groupwith preserved systolic function (P = .11).

Multiple logistic regression analysis (Table III)revealed age (odds ratio [OR] for patients older than 70years 1.7, P = .03), NYHA functional status at presenta-tion (OR 2.0, P = .01), elevated serum creatinine >130µmol/L (OR 2.5, P = .05), and ischemic cause (OR 2.0,P = .01) to be associated with increased mortality risk.Interesting but nonsignificant associations were notedwith systolic dysfunction (OR 1.6, P = .1, increasedmortality risk), presence of left ventricular hypertrophy(OR 0.7, P = .1, decreased risk) or atrial arrhythmias(OR 0.7, P = .16, decreased risk). Regarding medica-tions, β-blockers (OR 0.5, P = .006) were associatedwith a reduced mortality risk, whereas metolazone (OR2.1, P = .01), thiazides (OR 3.5, P = .003), and loopdiuretics (OR 1.9, P = .03) were associated withincreased mortality risk. ACE inhibitors were associatedwith a trend toward a reduced risk (OR 0.9, P = .74),although this was not statistically significant, probablybecause of the small number of patients who were notprescribed an ACE inhibitor thus limiting the contrastin event rates for statistical comparison. A second mul-tiple logistic regression analysis limited to the 121patients with preserved systolic function revealed that

American Heart JournalVolume 138, Number 1, Part 1 McAlister et al 91

Figure 2

Survival analysis for 562 patients with CHF, by type of dysfunction.

age alone was significantly associated with mortalityrisk (in particular, those patients older than 70 yearshad a markedly increased risk of death [OR 7.7, P =.003]). None of the medications was significantly asso-ciated with a survival benefit (or a mortality excess). Afurther multiple logistic regression analysis, carried outafter substituting ejection fractions with the 2 broadcategories of systolic dysfunction and preserved systolicfunction, did not show significance for either group.

DiscussionEffectiveness of a specialized heart failure clinic

Of the findings from this study, perhaps most impor-tant is that a specialized outpatient clinic devoted tothe management of CHF can optimize the use ofproven efficacious therapy in this condition. The find-ing that 83% of patients received ACE inhibitors is amarked improvement from 32% to 54% reported inrecent North American practice audits13-15 and clearlydemonstrates the effectiveness of a specialized clinic inclosing the care gap in the use of ACE inhibitors. Thesurvival data in our study compare favorably to datafrom other comprehensive heart failure managementclinics16,17 and clinical trials with similar clinical pro-files, such as the SOLVD trials18,19 and DIG study20 inwhich participants also received regular and careful fol-low-up, and suggest that the enhanced application ofproven efficacious therapy can reduce mortality rate.

An unsettled issue in the management of CHF is theappropriate dose of ACE inhibitors. Patients in thisstudy, and also observed in other settings,21,22 receivedlower doses of ACE inhibitors than were used in themajor CHF clinical trials.18,19,23,24 Although higherdoses of ACE inhibitors appear hemodynamically morebeneficial,25-29 patients attending heart failure clinics,particularly the sicker patients, are often flail, hypoten-sive, and cannot tolerate higher doses of ACEinhibitors. The recently completed ATLAS trial,designed to test whether a high dose of lisinopril (32.5mg to 35 mg daily) was more beneficial than a low dose

(2.5 mg to 5 mg daily), suggests some benefit in reduc-ing the combined secondary end point of death andhospitalization but leaves unanswered the effect of theprimary end point of all-cause mortality, which did notdiffer between the 2 doses.28

Systolic dysfunction versus preserved systolic functionPatients with CHF whose left ventricular systolic func-

tion may be normal or mildly impaired or may showclear evidence of diastolic dysfunction form a distinctgroup of patients of interest in this and other stud-ies.14,29-31 The proportion of patients with preservedsystolic function in our study sample (22%) is lowerthan that reported in previous studies of hospitalizedpatients14,29 but is consistent with previously publishedreports of outpatients with chronic CHF.30,31 Althoughthis cohort tended to be older and with a higher propor-tion of women and patients with hypertension, therewere no features on history or physical examination thatconsistently and accurately differentiated betweenpatients with preserved systolic function or systolic dys-function and emphasizes the need for objective mea-surement of LVEF in CHF.14,29,32 Reported annual mor-tality rates for patients with CHF caused by diastolicdysfunction and in those with preserved systolic func-tion varies from 1.3% to 17.5%, giving rise to the generalimpression that CHF with diastolic dysfunction or pre-served systolic function carries a favorable progno-sis.29,31,33,34 However, the mortality rates in our patientswith preserved systolic function did not differ fromthose seen in the patients with systolic dysfunction,regardless of the presence or absence of ischemic heartdisease. Although this may be because of chance, lack ofstatistical power, or a selection bias in that only sickerpatients are referred to the Heart Function Clinic, ourdata would still suggest that the mortality risk from dias-tolic dysfunction or CHF with preserved systolic func-tion can be substantially worse than previously thoughtand highlights a continuing investigative priority to findthe optimal treatment strategies for this condition.

Predictors of outcomesAdvancing age,6,30,35 poor functional capacity,30,36

reduced LVEF,6,30,36-38 ischemic cause,39 and elevatedserum creatinine level30 have all been identified in pre-vious studies as predictors of mortality in CHF. Ourstudy found that, contrary to other reports, low serumsodium,30,38,40 ventricular arrhythmias,36,37,41 and dia-betes mellitus6,35 were not negative prognostic factorsin CHF. Other studies37,42-44 have also been unable toconfirm the purported negative implications of thesefactors on multivariate analysis. It is possible that thesefactors may simply reflect the severity of illness and arenot independent prognostic indicators after adjustmentfor other factors. This inconsistency may be caused bythe heterogeneity in patient populations and treatment

American Heart JournalJuly 1999McAlister et al92

Variable Odds ratio P value

Increased riskAge >70 y 1.7 .03NYHA class IV 2.0 .01Serum creatinine >130 µmol/L 2.5 .05Ischemic cause 2.0 .01Metolazone 2.1 .01Thiazide 3.5 .003Loop diuretics 1.9 .03

Decreased riskβ-Blockers 0.5 .006

Table III. Results of multiple logistic regression analysis ofbaseline variables predicting risks of subsequent death for 566patients with CHF

regimens among studies. In particular, the use of ACEinhibitors varies markedly and, because ACE inhibitorsexert a modulating effect on the neurohumoral system,it seems reasonable to hypothesize that some prognos-tic indicators such as low serum sodium or ventricularectopic beats may be less relevant in patients takingACE inhibitors.

LimitationsThe major limitation of our study is the likelihood of

selection bias in the patient population. Although ourstudy sample represents a consecutive series of patientsreferred to a Heart Function Clinic and prospectivelyfollowed up, it is not a random sample of community-dwelling individuals with CHF. However, because allcases of CHF were confirmed by a cardiology serviceusing standard criteria and objective testing of LVEF,our study does provide insights into the epidemiologyof chronic CHF unbiased by problems with classifica-tion that can occur in population-based studies. Othersources of information on the severity of CHF in thecommunity are baseline data on clinical trial enrollees,such as those from the CONSENSUS, SOLVD, and DIGtrials.18-20,23 These are likely to be affected by selectionbias, whether by protocol or because of physician andpatient choice. Thus, although our study sample canprovide some insights into the prognosis of patientswith CHF of varying severity associated with systolicdysfunction or preserved systolic function, only large,population-based epidemiologic studies can adequatelydefine the current status of CHF.

ConclusionDespite the great strides that have been made in the

past 2 decades in identifying efficacious therapies forCHF, a number of challenges remain for researchersand clinicians. The foremost is the translation of thisproven efficacious therapy into effective practice. Ourdata support the contention that a specialized outpa-tient clinic can improve practice patterns, as shown bythe high rate of ACE inhibitors use, and possibly out-comes, in patients with CHF. Also, the demonstrationof high mortality risk in patients with preserved systolicleft ventricular function emphasizes the need to findefficacious (and effective) therapies for this condition.

We thank Ms R. Gutierrez and Ms L. Hill at the Uni-versity of Alberta Hospital Heart Function Clinic forexpert assistance, Ms J. Sieben for maintaining thecomputer database, and Mr T. Taher for piloting anearly version of the data collection form.

References1. Yusuf S, Thom T, Abbot RD. Changes in hypertension and in conges-

tive heart failure mortality. Hypertension 1989;13(Suppl I):I74-9.2. Garg R, Packer M, Pitt B, Yusuf S. Heart failure in the 1990s: evolu-

tion of a major public health problem in cardiovascular medicine. JAm Coll Cardiol 1993;22(Suppl A):3A-5A.

3. Eriksson H, Svardsudd K, Larsson B, et al. Quality of life in earlyheart failure: the study of men born in 1913. Scand J Prim HealthCare 1988;6:161-7.

4. Rodeheffer RJ, Jacobsen SJ, Gersch BJ, et al. The incidence andprevalence of congestive heart failure in Rochester, Minnesota.Mayo Clin Proc 1993;68:1143-50.

5. Ho KKL, Anderson KM, Kannel WB, Grossman W, Levy D. Survivalafter the onset of congestive heart failure in Framingham HeartStudy subjects. Circulation 1993;88:107-15.

6. Bourassa MG, Gurne O, Bangdiwala S, et al. for the SOLVD Inves-tigators. Natural history and patterns of current practice in heart fail-ure. J Am Coll Cardiol 1993;22(Suppl A):14A-19A.

7. Krumholz HM, Parent EM, Tu N, Vaccarino V, Wang Y, Radford MJ,et al. Readmission after hospitalization for congestive heart failureamong Medicare beneficiaries. Arch Intern Med 1997;157:99-104.

8. Rich MW, Beckham V, Wittenberg C, Leven CL, Freedland KE, Car-ney RM. A multidisciplinary intervention to prevent the readmissionof elderly patients with congestive heart failure. N Engl J Med1995;333:1190-5.

9. Bonneux L, Barendregt JJ, Meeter K, Bonsel GJ, van der Maas PJ.Estimating clinical morbidity due to ischemic heart disease and con-gestive heart failure: the future rise of heart failure. Am J Pub Health1994;84:20-8.

10. Brophy JM. Epidemiology of congestive heart failure: Canadiandata from 1970 to 1989. Can J Cardiol 1992;8:495-8.

11. McKee PA, Castelli WP, McNamara PM, Kannel WB. The naturalhistory of congestive heart failure: the Framingham study. N Engl JMed 1971;285:1441-6.

12. Yusuf S, Held P, Teo KK. Selection of patients for randomized con-trolled trials: implications of wide or narrow eligibility criteria. StatMed 1990;9:73-86.

13. The Clinical Quality Improvement Network (CQIN) Investigators.Mortality risk and patterns of practice in 4606 acute care patientswith congestive heart failure. The relative importance of age, sex,and medical therapy. Arch Intern Med 1996;156:1669-73.

14. McDermott MM, Feinglass J, Sy J, Gheorghiade M. Hospitalizedcongestive heart failure patients with preserved versus abnormal leftventricular systolic function: clinical characteristics and drug ther-apy. Am J Med 1995;99:629-35.

15. Young JB, Weiner DH, Yusuf S, et al. Patterns of medication use inpatients with heart failure: a report from the registry of studies of leftventricular dysfunction (SOLVD). South Med J 1995;88:514-23.

16. Fonarow GC, Stevenson LW, Walden JA, et al. Impact of a com-prehensive heart failure management program on hospital readmis-sion and functional status of patients with advanced heart failure. JAm Coll Cardiol 1997;30:725-32.

17. Stevenson WG, Stevenson LW, Middlekauff HR, et al. Improvingsurvival for patients with advanced heart failure: a study of 737consecutive patients. J Am Coll Cardiol 1995;26:1417-23.

18. The SOLVD Investigators. Effect of enalapril on survival in patientswith reduced left ventricular ejection fractions and congestive heartfailure. N Engl J Med 1991;325:293-302.

19. The SOLVD Investigators. Effect of enalapril on mortality and thedevelopment of heart failure in asymptomatic patients with reducedleft ventricular ejection fractions. N Engl J Med 1992;327:685-91.

20. The DIG Investigators. The effect of digoxin on mortality and mor-bidity in patients with heart failure. The Digitalis InvestigationGroup. N Engl J Med 1997;336:525-33.

21. Missouris CG, MacGregor GA. The use of angiotensin-converting

American Heart JournalVolume 138, Number 1, Part 1 McAlister et al 93

enzyme inhibitors in the treatment of heart failure in hospital prac-tice. Postgrad Med J 1997;73:409-11.

22. Chin MH, Wang JC, Zhang JX, Lang RM. Utilization and dosing ofangiotensin-converting enzyme inhibitors for heart failure. Effect ofphysician specialty and patient characteristics. J Gen Intern Med1997;12:563-6.

23. The CONSENSUS Trial Study Group. Effects of enalapril on mortal-ity in severe congestive heart failure: results of the CooperativeNorth Scandinavian Enalapril Survival Study (CONSENSUS). NEngl J Med 1987;316:1429-35.

24. Cohn JN, Johnson G, Ziesche S, et al. A comparison of enalaprilwith isosorbide-dinitrate in the treatment of chronic congestive heartfailure. N Engl J Med 1991;325:303-10.

25. Riegger GAJ. Effects of quinapril on exercise tolerance in patientswith mild to moderate heart failure. Eur Heart J 1991;12:705-11.

26. Pacher R, Globits S, Bergler-Klein J, et al. Clinical and neurohumoralresponse of patients with severe congestive heart failure treated withtwo different captopril dosages. Eur Heart J 1993;14:273-8.

27. Vagelos R, Nejedly M, Wilson K, Yee YG, Fowler M. Comparison oflow versus high dose enalapril therapy for patients with severe con-gestive heart failure [abstract]. J Am Coll Cardiol 1991;17:275A.

28. The Assessment of Treatment with Lisinopril and Survival (ATLAS)Investigators. Preliminary result presented at the 1998 Annual Sci-entific Meeting, American College of cardiology.

29. Vasan RS, Benjamin EJ, Levy D. Prevalence, clinical features, andprognosis of diastolic heart failure: an epidemiologic perspective. JAm Coll Cardiol 1995;26:1565-74.

30. Madsen BK, Hansen JF, Stokholm KH, Brons J, Husum D, MortensenLS. Chronic congestive heart failure: description and survival of 190consecutive patients with a diagnosis of chronic congestive heartfailure based on clinical signs and symptoms. Eur Heart J 1994;15:303-10.

31. Cohn JN, Johnson G. Heart failure with normal ejection fraction: theV-HeFT study. Circulation 1990;81(Suppl III):III 48-53.

32. Ghali JK, Kadakia S, Cooper RS, Liao Y. Bedside diagnosis of pre-served versus impaired left ventricular systolic function in heart fail-ure. Am J Cardiol 1991;67:1002-6.

33. Setaro JF, Soufer R, Remetz MS, et al. Long-term outcome in patientswith congestive heart failure and intact systolic left ventricular per-formance. Am J Cardiol 1992;69:1212-6.

34. Gaasch WH. Diagnosis and treatment of heart failure based onleft ventricular systolic or diastolic dysfunction. JAMA 1994;271:1276-80.

35. Andersson B, Waagstein F. Spectrum and outcome of congestiveheart failure in a hospitalized population. Am Heart J 1993;126:632-40.

36. Matoba M, Matsui S, Hirakawa T, et al. Long-term prognosis ofpatients with congestive heart failure. Jpn Circ J 1990;54:57-61.

37. Rouleau J, Shenasa M, de Champlain J, Nadeau R. Predictors ofsurvival and sudden death in patients with stable severe congestiveheart failure due to ischemic and nonischemic causes: A prospectivelong term study of 200 patients. Can J Cardiol 1990;6:453-60.

38. Parameshwar J, Keegan J, Sparrow J, Sutton GC, Poole-Wilson PA.Predictors of prognosis in severe chronic heart failure. Am Heart J1992;123:421-6.

39. Bart BA, Shaw LK, McCants CB, Fortin DF, Lee KL, O’Connor CM.Clinical determinants of mortality in patients with angiographicallyproven ischemic and non-ischemic cardiomyopathy [abstract]. Cir-culation 1996;94(Suppl I):I-193.

40. Saxon LA, Stevenson WG, Middlekauff HR, et al. Predicting deathfrom progressive heart failure secondary to ischemic or idiopathicdilated cardiomyopathy. Am J Cardiol 1993;72:62-5.

41. Gradman AH, Deedwania PC. Predictors of mortality in patientswith heart failure. Cardiol Clin 1994;12:25-35.

42. Dargie HJ, Cleland JGF, Leckie VJ, Inglis CG, East BW, Ford I. Rela-tion of arrhythmias and electrolyte abnormalities to survival in patientswith severe chronic heart failure. Circulation 1987;75:98-107.

43. Anguita M, Arizon JM, Bueno G, et al. Clinical and hemodynamicpredictors of survival in patients age <65 years with severe conges-tive heart failure secondary to ischemic or nonischemic dilated car-diomyopathy. Am J Cardiol 1993;72:413-7.

44. Cohn J, Levine TB, Olivari MT, et al. Plasma norepinephrine as aguide to prognosis in patients with chronic congestive heart failure.N Engl J Med 1984;311:819-23.

American Heart JournalJuly 1999McAlister et al94