DISCUSSION

We confirmed the findings in V-HeFT I of the efficacy and survival advantage of a combination of isosorbide dinitrate plus hydralazine added to standard background therapy in patients with moderate-to-severe heart failure. The trial population was limited to patients who identified themselves as black, a subgroup that had had a particularly favorable response to this therapy in retrospective analyses.6,38

The 43 percent reduction in the mortality rate in the group given isosorbide dinitrate plus hydralazine occurred among patients who were well treated with a background regimen of neurohormonal-inhibitor drugs. Thus, the reduction in mortality is consistent with the existence of an alternative mechanism controlling the progression of heart failure. A similar reduction in mortality was previously reported among blacks with NYHA class II, III, or IV heart failure in V-HeFT I who were receiving only diuretics and digitalis rather than neurohormonal-inhibiting agents.6,38 These results suggest that a fixed-dose combination of isosorbide dinitrate plus hydralazine may benefit patients in a manner that is independent of background therapy.

Growing evidence supports the concept that nitric oxide protects against myocardial and vascular remodeling.11,22-27,45-47 The combination of isosorbide dinitrate and hydralazine may serve as a nitric oxide donor, with hydralazine conferring protection against the degradation of nitric oxide induced by oxidative stress.48-52 Thus, data from A-HeFT support, but do not prove, the existence of a protective role of nitric oxide even in the presence of neurohormonal blockade.

We used two strategies to demonstrate the efficacy of isosorbide dinitrate plus hydralazine as therapy for heart failure in a moderate-sized cohort. First, we designed the trial to include only patients self-identified as black, since previous data suggested that such patients had increased responsiveness to this therapy.6,38 Second, we used a composite score, composed of weighted values for death from any cause, a first adjudicated hospitalization for heart failure, and change in the quality of life, as the primary end point. We hypothesized that including in the primary end point data relevant to all important outcomes of chronic disease would enhance the statistical power of the study to detect efficacy in a moderate-sized cohort.53 The significant benefit of isosorbide dinitrate plus hydralazine demonstrated by the differences in the composite score (–0.1±1.9, as compared with –0.5±2.0 in the placebo group; P=0.01) provides support for the use of such an end point in future trials.

Nitric oxide regulates cardiovascular processes including myocardial hypertrophy, remodeling, and substrate use, as well as vascular function, inflammation, and thrombosis.11,22-27,45-47 Substantial evidence exists that endothelial dysfunction and impaired bioavailability of nitric oxide occur in both ischemic and nonischemic models of heart failure and contribute to the pathophysiology of congestive heart failure.12-19,21 Studies of murine models of heart failure — mice deficient in endothelial nitric oxide synthase and mice overexpressing nitric oxide synthase — have demonstrated the role of nitric oxide in preserving left ventricular performance, inhibiting myocardial remodeling, and improving survival.23,24,27,46 Nitric oxide has also been shown to regulate the use of myocardial substrate47 and mitochondrial respiration.25 In clinical models of heart failure, basal cardiac release of nitric oxide is decreased,54 sensitivity to inhibition of nitric oxide synthase is increased,55 nitric oxide–mediated processes are impaired,12,13,18 and oxidative stress is increased.14,15,17,19,20

Our choice of a study cohort of patients identifying themselves as black is based on observations of differences in prevalence, risk profiles, causation, disease severity, outcomes, and response to therapy between black patients and white patients with heart failure.28-30 Retrospective analyses according to race in V-HeFT I, V-HeFT II, and the Studies of Left Ventricular Dysfunction have shown significant differences between blacks and whites in the response to pharmacotherapy for heart failure.30,38,39 In a novel experimental system, Kalinowski et al.37 found that, as compared with endothelial cells from healthy white women, endothelial cells from healthy black women had diminished bioavailability of nitric oxide as a result of increased oxidative stress. Data suggesting that endothelial function and bioavailability of nitric oxide may be less robust in blacks than in whites32-37 lend credibility to the hypothesis that the balance of mechanisms leading to the progression of heart failure may vary with geographic origin.

The combination of isosorbide dinitrate and hydralazine was used as a vasodilator in V-HeFT I because of its balanced effect on arteriolar dilation and venodilation. Once the physiologic role of nitric oxide was elucidated, it became apparent that isosorbide dinitrate exerts its vasodilatory effect by donating nitric oxide or forming related compounds.48,56 Concomitant administration of hydralazine prolongs the vasodilatory effects of isosorbide dinitrate in experimental and clinical models.19,49-52 Munzel et al.51 have shown that hydralazine is an effective antioxidant, inhibiting the generation of nitric oxide–inactivating reactive oxygen species by inhibiting vascular NADH and NADPH oxidases. Similarly, other antioxidants have been shown to reduce the mitochondrial production of reactive oxygen species.57 Thus, hydralazine, by reducing oxidative stress, may enhance the effects of nitric oxide derived from nitric oxide donors as well as from endogenous sources. This study, however, does not establish that these mechanisms explain the clinical benefit of isosorbide dinitrate plus hydralazine in heart failure.

Our trial represents a departure from the recent approach to the design of cardiovascular trials. Rather than studying a large heterogeneous population, we examined a specific population in whom efficacy was more likely to be established. A heterogeneous population may have substantial variations in genetic and environmental factors that influence disease progression and the response to therapy. Since subgroups of the trial cohort are rarely large enough for statistical analyses of the heterogeneity of an effect, the standard thinking is that the overall response to therapy should be accepted as a mandate for the use of that therapy in all subgroups in the trial. Recent insights into mechanistic variability, however, lend credence to the concept that the average effects in heterogeneous populations may obscure therapeutic efficacy in some subgroups and the lack of such efficacy in others.