Author + information
- Received February 23, 2002
- Revision received April 3, 2002
- Accepted April 17, 2002
- Published online July 17, 2002.
- Daniel L Dries, MD, MPH*,* (, )
- Mark H Strong, MD*,
- Richard S Cooper, MD† and
- Mark H Drazner, MD, MSc*
- ↵*Reprint requests and correspondence:
Dr. Daniel L. Dries, Heart Failure Research Group, Division of Cardiology, Department of Internal Medicine, Room H8.116, 5323 Harry Hines Boulevard, Dallas, Texas 75390-9034, USA.
Objectives This study was undertaken to determine whether enalapril had comparable efficacy in black and white patients with asymptomatic left ventricular dysfunction (ALVD) in preventing the development of symptomatic heart failure (HF).
Background Recent studies have suggested that black patients with HF due to systolic dysfunction may derive less benefit than white patients with HF when treated with the same medication.
Methods This is a post hoc analysis of the 4,054 black and white participants of the Studies of Left Ventricular Dysfunction Prevention Trial.
Results Randomization to enalapril was associated with a comparable reduction in the relative risk of the development of symptomatic HF in black (relative risk [RR] 0.67, 95% confidence interval [CI] 0.49, 0.92, p = 0.01) and white patients (RR 0.61, 95% CI 0.53, 0.70, p < 0.001). Treatment with enalapril was also associated with a comparable reduction in the risk of the development of HF requiring medical therapy and the composite end point of death or development of HF in black and white patients. Black as compared with white patients with ALVD were at increased risk of the development of symptomatic HF (RR 1.81, 95% CI 1.51, 2.17, p < 0.001) despite adjustment for available measures of disease severity.
Conclusions Despite the increased absolute risk in black patients compared with white patients for the progression of ALVD, enalapril was equally efficacious in reducing the risk of progression of ALVD in these two ethnic groups.
Although the possibility has been raised that black patients derive less benefit than white patients from some medications used in the treatment of heart failure (HF) due to systolic dysfunction (1,2), further studies are needed to test this hypothesis, given its important public health implications. One recent report from the Studies of Left Ventricular Dysfunction (SOLVD) databases demonstrated that white but not black participants randomized to enalapril had a significant reduction in the risk of hospitalization for HF (1). We designed the present post hoc analysis to determine whether enalapril delayed the progression of asymptomatic left ventricular dysfunction (ALVD) to symptomatic HF in both black and white subjects participating in the SOLVD Prevention Trial. The SOLVD Prevention Trial had end points designed to represent progression from ALVD to symptomatic HF (3). These end points were unique to the SOLVD Prevention Trial and had not been assessed in the previous report (1), which had combined participants from the SOLVD Treatment and Prevention trials. If there is a differential response to enalapril therapy between these two ethnic groups, we hypothesized that enalapril should also be less efficacious in preventing the progression from ALVD to symptomatic HF in black than in white subjects.
The SOLVD Prevention Trial was a randomized, double-blind, placebo-controlled trial. A total of 4,228 patients with ALVD were randomly assigned to receive either enalapril or placebo at one of 83 hospitals linked to 23 centers in the U.S., Canada, and Belgium. Patients known to have heart disease who had an ejection fraction 0.35 or less and who were not receiving diuretics, digoxin, or vasodilators for the treatment of HF were eligible for the SOLVD Prevention Trial. Patients were allowed to receive diuretic therapy for hypertension, digoxin for current or past atrial fibrillation, or nitrates for angina. The rationale, design, and methods of the SOLVD Trials have been described (3,4).
Defining the etiology of ALVD
For each enrolled participant, the SOLVD site investigator determined the most likely etiology of HF and indicated it on the baseline form. Choices included: “ischemic,” “nonischemic,” and “other” as determined by the individual SOLVD site investigators after reviewing all available information. Subjects did not routinely undergo cardiac catheterization or non-invasive testing to make this determination. For the present analysis, we refined the definition of ischemic etiology to minimize misclassification. As done previously (5–7), the present analysis defined a participant as having an ischemic etiology of heart failure if anyof the following was present: 1) classification as ischemic etiology by original SOLVD investigator; 2) history of prior myocardial infarction; or 3) history of coronary artery bypass graft operation.
Definition of race
Self-identified racial/ethnic classification was reported on the baseline SOLVD forms. The choices that were available on the form included: black, white, Hispanic, American Indian, Asian, and other.
Defining the development of HF
Information on clinical status, the development of HF, use of medications other than those prescribed as part of the study, adherence to the study regimen, and side effects was systematically recorded at each follow-up visit. For patients who died, were hospitalized, or developed HF, the cause of death, the primary reason for hospitalization, and the development of HF were ascertained and classified using standardized forms by each center’s principal investigator, who was unaware of the patient’s treatment assignment.
Four overlapping definitions of HF, of increasing severity, were used by the SOLVD Prevention Trial investigators: 1) development of HF, identified by the study physician on the basis of symptoms, signs, or need for changes in therapy; 2) HF requiring the addition of a diuretic, digoxin, or a vasodilator to the patients regimen (in the case of patients already receiving these drugs at baseline, the additional drug had to be prescribed for this indication); 3) hospitalization for HF; and 4) death due to progressive HF (“pump-failure”). Patients classified as reaching the end point of developing HF requiring therapy were also classified as developing HF. Other end points included the composite of death or development of HF and the composite of death or HF hospitalization.
Of 4,228 subjects in the SOLVD Prevention Trial, we excluded subjects if New York Heart Association (NYHA) class was not I or II or if race was not white or black, leaving a study population of 4,054. Dichotomous variables included race (black or white), gender, etiology of left ventricular (LV) dysfunction (ischemic or nonischemic), history of diabetes or hypertension, baseline use of diuretics or beta-blockers, NYHA (I vs. II), serum sodium (<135 mEq/l or not). Continuous variables included LV ejection fraction, age, heart rate, and serum creatinine. Continuous variables were compared between groups by the Student ttest assuming unequal variance where appropriate, and dichotomous variables were compared with the Fisher exact test.
Univariate Cox proportional hazards models determined the association of enalapril therapy with the time to development of end points in black and in white patients. Although patients were randomized to enalapril or placebo, there were some baseline differences in black patients randomized to enalapril and those randomized to placebo. These baseline differences were adjusted for in multivariate Cox proportional hazards models.
Analyses were conducted to examine the independent association of race with the progression from ALVD to HF. Bivariate screens identified 17 variables associated with black race and with the development of HF (systolic blood pressure (BP), heart rate, gender, etiology of LV dysfunction, serum creatinine, digoxin, hypertension, randomization to enalapril, anti-arrhythmic use, age, hyponatremia, etiology of LV dysfunction, diabetes mellitus, and diuretic use). An automated stepwise selection process was used to select nine of these potential confounders. Adding the remaining eight potential confounders back to the model (systolic BP, anti-arrhythmic use, beta-blocker use, atrial fibrillation, diabetes, history of hypertension, digoxin use, and diuretic use) did not change the relative risk associated with black race; thus, these eight covariates were left out of the final model. Kaplan-Meier curves, stratified by randomization to enalapril therapy, were constructed in black participants and in white participants for three end points—the development of HF, the development of HF requiring therapy, and death or the development of HF. The log-rank statistic was used to compare the event-free survival in subjects randomized to enalapril or placebo in both blacks and whites for each of these end points. Formal statistical tests for evidence of interaction between randomization assignment (enalapril or placebo) and race (black vs. white participants) were conducted to test for evidence of heterogeneity of treatment effect with regard to the association of enalapril use and the risk for progression of ALVD in the two groups. The interaction analysis was conducted using a simple unadjusted model and a fully adjusted model using the same covariates as in multivariate analysis. A p value ≤0.05 for the interaction term was considered significant for evidence of heterogeneity of treatment effect for enalapril in the two groups. All statistical analyses were conducted using Statistical Analysis Software (SAS), version 8.2 (SAS, Inc., Cary, North Carolina).
The baseline characteristics of black compared with white subjects are presented in Table 1, with each group further stratified according to randomization to placebo or enalapril. In white patients, there were no significant differences in baseline characteristics between subjects randomized to enalapril or placebo. In black patients, those randomized to enalapril were more likely to be women, have a history of hypertension, and had a higher systolic BP. In comparing white subjects with black subjects, white patients had a lower proportion of women and history of diabetes mellitus or hypertension but had a higher proportion with an ischemic etiology of systolic dysfunction and baseline beta-blocker use.
Unadjusted incidence in each group according to randomization assignment
The unadjusted incidence of end points that represent progression of ALVD for black and white participants stratified by randomization to enalapril or placebo is shown in Table 2. The black participants demonstrated an increased incidence of each end point defining disease progression as compared with whites. There was a significant reduction in incidence rates for each end point in the patients randomized to enalapril therapy in both black and white subjects. The unadjusted event curves (Fig. 1) demonstrate large differences in the absolute risk between the respective placebo groups in black and white subjects for the development of HF, development of HF requiring therapy, and the composite of death or development of HF. However, the relative risk reduction associated with randomization to enalapril for these end points was significant and of similar magnitude in both the black and white participants.
Univariate analyses (Table 2)
Randomization to enalapril was associated with a significant reduction in the risk for development of symptoms of HF, symptoms requiring the addition of a vasodilator or digoxin, and the combined end point of death or development of HF in both black and white subjects. Randomization to enalapril was associated with a significant reduction in the risk of first hospitalization for HF or the combined end point (death or first HF hospitalization) in the white subjects but not in the black subjects. However, the 95% confidence intervals of the relative risk associated with randomization to enalapril overlapped between the black and white subjects, and formal tests for heterogeneity of treatment effect (interaction) between the groups were not significant for these end points (Table 3).
Multivariate analyses (Table 3)
We performed multivariate analyses to adjust for the differences in baseline characteristics of black patients randomized to placebo or enalapril. In addition, formal statistical testing for interaction between randomization and race was conducted to test for heterogeneity of treatment effect using unadjusted and adjusted modeling. The adjusted analyses demonstrated little difference compared with the treatment effects demonstrated in unadjusted analyses. In addition, there was no significant statistical evidence for heterogeneity of treatment effect with enalapril for any of the end points.
Increased progression of ALVD in black subjects
Multivariate analysis was performed to determine the independent risk associated with black race for the progression of ALVD (Table 4). In these analyses, black subjects remained at increased risk compared with the white subjects for the progression of ALVD to each of the end points despite adjustment for other markers of disease severity.
Despite recent concerns that angiotensin-converting enzyme (ACE) inhibitors may be less efficacious in black patients with HF (1), the present study demonstrates that enalapril significantly reduced the risk of development of HF in both blacks and whites with asymptomatic LV systolic dysfunction. Enalapril resulted in a comparable reduction in the risk of the development of symptoms of HF, the need to start medications for symptoms of HF, and the combined end point of death or the development of symptoms of HF in both blacks and whites with ALVD. These data demonstrate that there is no evidence for heterogeneity of treatment effect in black and white patients with regard to the ability of ACE inhibitors to reduce the risk for progression from ALVD to symptomatic HF. The consistency of results in black and white subjects strengthens the argument that ACE-inhibitor therapy should continue to be used in black patients with HF (8,9).
Comparison with previous studies
Previous retrospective studies that have addressed whether ACE inhibitors are efficacious in black patients with HF are inconclusive. Data from the Vasodilator–Heart Failure Trial investigators can be interpreted to show either that ACE inhibitors are not as efficacious in blacks as in whites or that the combination of isosorbide and hydralazine is particularly beneficial in black patients (10,11). A recent report that combined the SOLVD Prevention and Treatment Trials and used a matched analysis of black and white subjects showed that enalapril use was associated with a reduction in the risk for HF hospitalization in white but not black subjects, and there was statistical evidence for heterogeneity of treatment effect in this analysis (1). In the present study, the effect of enalapril in reducing the risk of either first hospitalization for HF or the composite end point (death or first hospitalization for HF) was not nominally significant in the black subjects but was significant in the white subjects, although formal evidence of statistical interaction between treatment effect and ethnicity was not present. In light of the previous study by Exner et al. (1), we are unable to determine whether these findings represent a lack of statistical power to detect heterogeneity of treatment effects in these two ethnic groups or whether they represent lack of statistical power to detect efficacy of enalapril for these end points in the small number of black participants. The latter hypothesis is supported both by the fact that the point estimates of relative risk associated with enalapril use in black subjects for these end points were less than one, and of comparable magnitude to those associated with enalapril use in the white subjects, and because enalapril was equally efficacious in reducing the risk for the development of symptomatic HF in both groups.
It has been demonstrated that black patients with symptomatic HF are at increased risk for death and HF progression compared with whites despite adjustments for differences in HF severity, comorbidities, and socioeconomic factors (5). The present analysis extends these findings by demonstrating that the absolute risk for progression from ALVD to symptomatic HF is substantially greater in blacks than in whites (Fig. 1). Despite the comparable relative reduction in risk associated with enalapril in the two groups, the differences in the baseline magnitude of risk was such that blacks randomized to enalapril remained at higher risk than whites randomized to placebo for the development of HF. These differences between black and white subjects in the risk of progression of ALVD persisted after adjusting for potential confounders including ejection fraction, NYHA class, serum sodium, and etiology of LV dysfunction. We interpret these findings to suggest either that residual confounding persists or that there are differences in the natural history of ALVD in blacks compared with whites. Clearly, statistical adjustment based upon a single, unstandardized measurement of the variables collected in SOLVD could result in residual confounding. On the other hand, if the latter hypothesis is true, one potential explanation may be ethnic differences in the etiology of LV systolic dysfunction (11).
There are several important limitations of this retrospective study. The absolute number of black patients in the study was small. The composite end point of death or development of HF was predominantly driven by the development of HF, whereas the end point of development of HF requiring therapy was a subset of the end point of the development of HF. Thus, the efficacy of enalapril in black subjects demonstrated in this study was predominantly based on this one end point (development of HF). In comparing the risk for progression to symptomatic HF between black and white subjects, we cannot exclude the possibility of residual confounding by variables not adjusted for in the multivariate models.
The validity of race as a construct in medical research has been questioned (12–14). However, the possibility of differential responses to pharmacologic therapy between ethnic groups has been suggested by recent data showing that black patients with HF derive less benefit than white patients from enalapril (1)and some (2)but not all (15)beta-blockers. We believe the present study demonstrates that ACE inhibitors are efficacious in black patients with systolic dysfunction. These findings are concordant with the recent African American Study of Kidney Disease and Hypertension that demonstrated a benefit of ACE-inhibitor therapy in black patients with renal disease (16). Our study adds support to the recommendation (9)that therapies demonstrated to improve survival in large randomized trials continue to be prescribed to all patients with HF irrespective of their ethnicity.
Potential racial variation in drug response will become an increasingly important question in clinical medicine. Racial categories combine social and biological effects in complex ways, however, and there are no adequate methods for eliminating this confounding (13,14). Genomic research continues to demonstrate the limited variation across population groups: virtually all human genetic diversity, including that found at the ACE gene, is represented in African-origin populations (17–19), and they therefore should be unlikely to display phenotypic responses that diverge sharply from those found in all other populations. In a sample of populations, Wilson et al. (20)recently compared standard ethnic labels against a clustering algorithm based directly on genetic markers to predict variation in enzymes that modify drug responses. Genetic markers of population origin improved prediction over ethnic labels in 3 of 4 enzymes studied, although neither system was very precise. These data demonstrate that race can be a very weak proxy for drug response (14).
Black patients with ALVD are at increased risk as compared with white patients for the development of HF. Enalapril appears to be equally efficacious in black and white patients in reducing the risk of progression from ALVD to symptomatic HF.
☆ Drs. Dries, Cooper and Drazner received support from the Donald W. Reynolds Cardiovascular Clinical Research Center, Dallas, Texas. Dr. Dries received support from the NIH (HL04455) and AHA, Texas affiliate (0160002Y). Dr. Drazner was the recipient of a Doris Duke Clinical Scientist Development Award from the Doris Duke Charitable Foundation, New York, New York.
- angiotensin-converting enzyme
- asymptomatic left ventricular dysfunction
- blood pressure
- heart failure
- left ventricular
- New York Heart Association
- Studies of Left Ventricular Dysfunction
- Received February 23, 2002.
- Revision received April 3, 2002.
- Accepted April 17, 2002.
- American College of Cardiology Foundation
- Dries D.L.,
- Sweitzer N.K.,
- Drazner M.H.
- Kaufman J.S.,
- Cooper R.S.
- Gabriel S.B.,
- Schaffner S.F.,
- Nguyen H.,
- Moore J.M.,
- Roy J.,
- Blumensteil B.,
- Higgins J.,
- DeFellice M.,
- Lochner A.,
- Faggart M.,
- Ward R.,
- Cooper R.S.,
- Adeyemo A.A.,
- Forimi C.N.,
- Lander E.S.,
- Daly M.J.,
- Altshuler D.