Author + information
- Received November 29, 2018
- Revision received January 11, 2019
- Accepted January 13, 2019
- Published online February 25, 2019.
- aDepartment of Internal Medicine, University of Colorado School of Medicine, Aurora, Colorado
- bDivision of Cardiology, Sarver Heart Center, University of Arizona College of Medicine, Tucson, Arizona
- cDivision of Cardiovascular Medicine, Vanderbilt University Medical Center, Vanderbilt Heart and Vascular Institute, Nashville, Tennessee
- dDivision of Cardiology, University of Colorado School of Medicine, Aurora, Colorado
- ↵∗Address for correspondence:
Dr. David Kao, Division of Cardiology, University of Colorado School of Medicine, 12700 East 19th Avenue, Campus Box B-139, Aurora, Colorado 80045.
Objectives This study sought to investigate sex differences in outcomes and responses to spironolactone in patients with heart failure with preserved ejection fraction (HFpEF).
Background HFpEF affects women more frequently than men. Sex differences in responses to effects of mineralocorticoid antagonists have not been reported.
Methods This was an exploratory, post hoc, non-pre-specified analysis of the TOPCAT (Aldosterone Antagonist Therapy for Adults With Heart Failure and Preserved Systolic Function) trial. Subjects with symptomatic HF and a left ventricular ejection fraction ≥45% were randomized to spironolactone or placebo therapy. Subjects enrolled from the Americas were analyzed. The primary outcome was a composite of cardiovascular (CV) death, cardiac arrest, or HF hospitalization. Secondary outcomes included all-cause mortality, CV, and non-CV mortality and CV, HF, and non-CV hospitalization. Sex differences in outcomes and treatment effects were determined using time-to-event analysis.
Results In total, 882 of 1,767 subjects (49.9%) were women. Women were older with fewer comorbidities but worse patient-reported outcomes. There were no sex differences in outcomes in the placebo arm or in response to spironolactone for the primary outcome or its components. Spironolactone therapy was associated with reduced all-cause mortality in women (hazard ratio: 0.66; p = 0.01) but not in men (pinteraction = 0.02).
Conclusions In TOPCAT, women and men presented with different clinical profiles and similar clinical outcomes. The interaction between spironolactone and sex in TOPCAT overall and in the present analysis was nonsignificant for the primary outcome, but there was a reduction in all-cause mortality associated with spironolactone therapy in women, with a significant interaction between sex and treatment arm. Prospective evaluation is needed to determine whether spironolactone therapy may be effective for treatment of HFpEF in women. (Aldosterone Antagonist Therapy for Adults With Heart Failure and Preserved Systolic Function [TOPCAT]; NCT00094302)
Heart failure with preserved ejection fraction (HFpEF) has been consistently observed to affect more women than men, but the underlying reasons for this sex difference are incompletely understood. Despite multiple large clinical trials, no medical therapies have proven beneficial in HFpEF, but potential sex differences in treatment responses have not been thoroughly explored. Recently, 6 clinically different HFpEF phenotypes were identified that differed significantly by sex, rates of adverse outcomes, and nominally significant differences in treatment responses, suggesting that sex may be 1 of multiple factors important in defining HFpEF phenotypes (1). Additional data suggest possible sex differences in the effects of aldosterone on cardiac remodeling and clinical response to mineralocorticoid antagonists (MRAs) (2–4), raising the question of whether a sex-specific benefit of MRAs may exist. Sex differences in response to treatment with MRAs among patients with HFpEF have not been reported.
The TOPCAT (Aldosterone Antagonist Therapy for Adults With Heart Failure and Preserved Systolic Function; NCT00094302) trial (5) presented the first opportunity to explore sex-specific treatment responses to MRAs in HFpEF in a large, randomized placebo-controlled trial. There were no significant interactions in TOPCAT between treatment arm and sex with respect to the primary outcome in pre-specified subgroup analysis of the full TOPCAT cohort (pinteraction = 0.99). However, there has been controversy regarding the apparent benefit of spironolactone in subjects enrolled from the Americas but not from Russia or Georgia (6,7), and an analysis of possible sex differences within the Americas cohort alone has not yet been reported. The goal of the present study was to describe sex differences in characteristics, outcomes, and treatment responses in subjects enrolled in TOPCAT from the Americas. The authors hypothesized that women would have lower rates of adverse outcomes and a larger benefit from spironolactone therapy than men.
This paper was prepared using TOPCAT clinical data obtained from the National Heart, Lung, and Blood Institute’s Biological Specimen and Data Repository Information Coordinating Center (BioLINCC, Calverton, Maryland). TOPCAT conformed to the principles outlined in the Declaration of Helsinki and was approved by institutional review boards at all sites (8). The present authors’ analysis was approved by the Colorado Multiple Institution Review Board and by BioLINCC.
The design of TOPCAT has been reported previously (5). Briefly, 3,445 patients with a left ventricular ejection fraction (LVEF) of ≥45% and ≥50 years of age and a history of nonadjudicated hospitalization for heart failure (HF) in the previous 12 months, a B-type natriuretic peptide level of ≥100 pg/ml or an N-terminal pro–B-type natriuretic peptide level of ≥360 pg/ml were randomized in a double-blind fashion to receive either spironolactone or placebo therapy. The mean follow-up was 3.3 years. The primary outcome was a composite of cardiovascular (CV) mortality, aborted cardiac arrest, or HF hospitalization. Secondary outcomes for the present analysis included all-cause mortality, CV mortality, and non-CV mortality and hospitalization for CV, HF, and non-CV. Because of previously described concerns about the veracity of HF diagnosis and poor treatment compliance in subjects from Russia and Georgia (6,7,9), analysis was limited to the 1,767 patients enrolled from the Americas, in accordance with multiple secondary analyses recently published by the TOPCAT investigators (10–14).
Data were stratified according to sex and treatment arm. Baseline characteristics in women and men were compared using the chi-square test and Mann-Whitney U test for categorical and continuous variables, respectively. To account for the possibility of differential treatment effects in men and women, the presence of sex differences in outcomes was based on comparisons between men and women within the placebo arm. Significance of changes in serum potassium, serum creatinine, and systolic blood pressure (SBP) from baseline to 4 and then 12 months was determined by using the paired Wilcoxon signed-rank test. Differences in changes of serum potassium, serum creatinine, and SBP from baseline between treatment groups were compared using the Mann-Whitney U test. Univariate and multivariate associations between sex and outcomes were determined using Cox proportional hazards models. Effects of spironolactone therapy versus placebo on primary and secondary outcomes were analyzed by sex, and interaction terms between sex and treatment arm were calculated. Multivariate associations were adjusted for all patient characteristics that differed significantly between women and men in frequency or magnitude (Table 1). The proportional hazards assumption was tested for all covariates and outcomes by testing the correlation of scaled Schoenfeld residuals with time. Where a covariate showed a significant correlation with time (p < 0.05), a coefficient for the interaction between the covariate and time included in multivariate and interaction analyses. A p value <0.05 was considered significant throughout.
Baseline characteristics of women and men are summarized in Tables 1, 2, and 3⇓⇓. Of the 1,767 subjects, 882 (49.9%) were women. All baseline demographics and comorbidities in women were significantly different from those in men (Table 1). Generally, women were older and had fewer comorbidities than men, including coronary artery disease, tobacco use, atrial fibrillation, chronic obstructive pulmonary disease, and diabetes mellitus. Women had significantly higher LVEF, blood pressure, and body mass indexes but lower estimated glomerular filtration rate and serum hemoglobin. Compared with men, women were assessed at a higher New York Heart Association functional class and had lower Kansas City Cardiomyopathy Questionnaire scores and higher Patient Health Questionnaire 9th edition scores (Table 2). Use of antihypertensive medications overall was high in both men and women (99%). There were no significant differences between men and women in use of angiotensin-converting enzyme (ACE) inhibitors, angiotensin receptor blockers (ARB), or diuretics (Table 3). Women were significantly more likely to be taking calcium channel blockers, whereas men were more likely to be taking beta-blockers or other antihypertensive medications.
Trends in serum potassium, serum creatinine, and SBP are illustrated in Figure 1, and summary statistics are provided in Online Table 1. Serum potassium and creatinine concentrations increased and SBP decreased significantly at 4 and 12 months in the spironolactone arms for both men and women, with most changes occurring between baseline and 4 months. Compared to men, women demonstrated a significantly greater increase in creatinine at 4 months (pinteraction = 0.03) and in potassium at 12 months (pinteraction = 0.04).
Rates of primary and secondary outcomes as well as documented causes of CV deaths stratified by sex for the placebo arm are summarized in Table 4 and for the spironolactone arm in Online Table 2. Kaplan-Meier curves for primary and secondary outcomes stratified by sex are shown in Online Figures 1a to 1d. In the placebo arm, rates of all events were numerically lower in women than in men, but there were no statistically significant differences. In the spironolactone arm, event rates were also numerically lower in women than in men, of which the differences in all-cause mortality and non-CV mortality were statistically significant. There were no significant differences in rates of specific CV causes of death (pump failure versus sudden death versus other) between men and women in either the placebo (chi-square p = 0.43) or spironolactone arms (chi-square p = 0.16).
Sex-specific multivariate hazard ratios (HR) in the placebo and spironolactone arms for all outcomes adjusted for age, race, LVEF, atrial fibrillation, coronary artery disease, chronic obstructive pulmonary disease, hypertension, diabetes mellitus, body mass index, heart rate, SBP, estimated glomerular filtration rate, and New York Heart Association functional class are summarized in Online Figure 2. There were no significant differences between men and women in the HRs of any outcomes in the placebo arm, whereas in the spironolactone arm, women had a significantly lower multivariate risk of the primary outcome, CV mortality, all-cause mortality, and non-CV mortality.
Kaplan-Meier survival estimates and event rates for the primary and secondary outcomes stratified by sex and treatment for subjects from the Americas are shown in Figures 2A and 2B. Multivariate HRs for outcomes comparing spironolactone versus placebo are summarized in Figure 3, and comprehensive results of univariate and multivariate analyses, including interaction terms, are found in the Online Table 3. The rates of the primary outcome and HF hospitalization between either male or female patients receiving spironolactone and placebo, were not significantly different, although there was a trend toward less CV mortality associated with spironolactone in women (9.0% vs. 13.2%, respectively; p = 0.051). Women receiving spironolactone had lower all-cause mortality than those receiving placebo (15.8% vs. 22.3%, respectively; p = 0.015). There was no significant reduction in any outcomes between treatment arms in men. In multivariate analysis, the association between spironolactone treatment and the primary outcome in women (HR: 0.81; 95% confidence interval [CI]: 0.63 to 1.05; p = 0.12) was similar in men (HR: 0.85; 95% CI: 0.67 to 1.08; p = 0.18; pinteraction = 0.84). There was a reduced likelihood of CV mortality associated with spironolactone in women (HR: 0.63; 95% CI: 0.42 to 0.95; p = 0.03 respectively) but not in men (p = 0.37), but the interaction term was nonsignificant (pinteraction = 0.31). Reductions in HF hospitalization with spironolactone were nonsignificant in both women and men.
There was a significant reduction in all-cause mortality associated with spironolactone in women (HR: 0.66; 95% CI: 0.48 to 0.90; p = 0.01) but not in men (p = 0.68), with a significant sex-treatment interaction (pinteraction = 0.024). To estimate the impact of death from unknown causes on the associations between all-cause mortality, CV mortality, and spironolactone treatment, analysis was repeated where deaths of unknown causes were combined with CV deaths. The multivariate HR for mortality was significant in women (HR: 0.57; 95% CI: 0.39 to 0.83; p = 0.003) but not in men (p = 0.80), with a significant interaction (pinteraction = 0.049). There were no significant associations between spironolactone and either of the sexes in non-CV mortality, CV hospitalization, or non-CV hospitalization (women: p = 0.43; men: p = 0.63).
In this secondary analysis of TOPCAT, several sex differences in HFpEF patient characteristics and evidence of a possible sex-specific treatment response to spironolactone were found. Women generally had fewer comorbidities but had worse patient-reported outcomes than men. There were no sex differences in rates of primary or secondary outcomes in patients receiving placebo and no significant interactions between sex and treatment arm with respect to the primary outcome or its components. Treatment with spironolactone in women was associated with a lower multivariate risk of all-cause mortality with a statistically significant interaction between sex and treatment arm, suggesting a possible sex-specific benefit of spironolactone in women. This finding is hypothesis-generating only and will require substantial validation prior to clinical application. However, given the paucity of effective therapies in patients with HFpEF, a prospective study of the effect of spironolactone in women with HFpEF is warranted.
Patient characteristics and outcomes
Sex differences in patient characteristics and outcomes were described previously in the 3 following HFpEF clinical trials: the I-PRESERVE (Irbesartan in Heart Failure with Preserved Ejection Fraction; NCT02973035) study (15), the CHARM-Preserved (Candesartan in Heart failure: Assessment of Reduction in Mortality and Morbidity; NCT00634712) study (16); and the DIG-PEF (Digitalis Investigation Group-Preserved Ejection Fraction) study (17,18). Directionality of sex differences in baseline characteristics among our analysis and those of I-PRESERVE and DIG-PEF are summarized in Table 5. Differences in baseline characteristics observed in the present study were concordant in all 3 studies, except for diabetes mellitus, which was less common in women in TOPCAT but the same between men and women in I-PRESERVE, and more common in women in DIG-PEF.
Event rates for all secondary outcomes in our analysis are summarized in Table 6. In the present analysis, there were no significant sex differences in any clinical endpoints in subjects receiving placebo. Event rates for all mortality outcomes and CV hospitalization in the present study were similar to those in CHARM-Preserved, although formal statistical tests were not reported (16). Multivariate risks of all outcomes were lower in I-PRESERVE (15), whereas there was a lower multivariate risk of all-cause mortality in women in DIG-PEF but no significant differences in HF or CV hospitalization (17,18). Larger studies that included observational data such as the MAGGIC (Meta-Analysis Global Group in Chronic Heart Failure) found reduced mortality in women compared to that in men with HFpEF, using a large individual patient data meta-analysis (19). Taken together, these study results reveal mixed evidence of sex differences in both mortality and hospitalization outcomes in patients with HFpEF, with somewhat more evidence of reduced mortality in women.
Treatment response to spironolactone
There was no significant interaction between sex and the treatment arm in the primary TOPCAT analysis with respect to the primary outcome (pinteraction = 0.99). However, subjects from the Americas and Russia and Georgia regions were analyzed together for all pre-specified subgroup analyses in TOPCAT, including the interaction between treatment and sex (5). In our analysis of subjects from the Americas only, there was no significant sex/treatment interaction effect for the primary outcome or its component CV mortality and HF hospitalization rates. We did observe a possible sex difference in treatment response to spironolactone with respect to all-cause mortality (HR: 0.66; 95% CI: 0.48 to 0.90; p = 0.01; pinteraction = 0.02), although there were no significant sex-specific treatment effects in either CV or non-CV mortality alone. This discrepancy may be due to lack of statistical power as well as the inclusion of death of unknown cause in the all-cause mortality outcome. Although the present findings raise the possibility of a sex-specific response to spironolactone with respect to all-cause mortality, these authors’ results are not conclusive, and existing supporting evidence is mixed.
Sex differences in the biologic activity of spironolactone in the kidney and the myocardium could promote sex-specific treatment response to spironolactone. Increases in serum potassium and serum creatinine and decreases in SBP associated with spironolactone in both men and women were observed as expected (6), with greater increases observed in women in creatinine at 4 months (pinteraction = 0.03) and in potassium at 12 months (pinteraction = 0.04). Previously, a greater decline in renal function associated with eplerenone in women than in men was also observed in the EPHESUS (Eplerenone Post-Acute Myocardial Infarction Heart Failure Efficacy and Survival Study) study (3). Given that there were no differences between men and women in baseline use of ACE inhibitors and ARBs in either the present study or those in EPHESUS, these findings could reflect greater renal activity of spironolactone in women.
Large studies of ACE inhibitors and ARBs in HFpEF have failed to show a sex difference in reduction of adverse clinical outcomes, and sex differences in HFpEF outcomes with MRAs have not been reported (16,20,21). MRA studies in HFrEF, such as the Randomized Aldactone Evaluation Study (22) or the Eplerenone in Mild Patients Hospitalization and Survival Study in Heart Failure (23), have not shown evidence of sex-specific treatment effects, whereas MRAs were associated with improved outcomes in both studies. There was an apparent reduction in all-cause mortality associated with eplerenone in women but not in men with acute myocardial infarction complicated by LV dysfunction (EF <40%) in EPHESUS, although as in the present analysis, the interaction between sex and treatment arm was not significant (24). Animal studies have shown a greater impact of MRAs on cardiac remodeling in females versus males including normalization of myocardial gene expression profiles post-myocardial infarction (4,25,26). It is feasible that MRAs may have greater impact on cardiac remodeling in women than in men in HFpEF due to greater myocardial sensitivity to aldosterone in women, but significant additional clinical and translational investigation is required to demonstrate sex differences in biologic effects of spironolactone sufficient to improve clinical outcomes such as mortality.
This is a post-hoc, exploratory subgroup analysis that stratified the TOPCAT subjects enrolled from the Americas according to sex, and all findings are hypothesis-generating only. An important limitation is the subgroup-within-subgroup analysis plan stratifying the population first by region, then sex, then treatment arm. As has been discussed elsewhere, many subjects enrolled from Russia and Georgia may not have had HF or taken the study medication as reliably (6,7,27), raising questions of the validity of combining cohorts in the Americas with those in Russia and Georgia for an analysis of HFpEF outcomes. Therefore, only subjects from the Americas were enrolled, despite the potential methodologic ramifications, which included an increased likelihood of both type I and II errors (10–14). The reduced sample sizes are relatively underpowered to detect outcomes of interest particularly with respect to formal treatment-subgroup interaction statistics. Increasing numbers of subgroups also increase the risk of false positive associations due to random chance. These issues reinforce the importance of conducting an appropriately powered, prospective trial examining the role of spironolactone prior to recommending use of spironolactone for management of HFpEF in women.
Women with HFpEF enrolled in TOPCAT were older and had higher LVEF and fewer comorbidities but worse health status than men. There were no significant sex differences in clinical endpoints. The authors observed a significant reduction in all-cause mortality associated with spironolactone in women but not in men. Prospective evaluation is warranted to determine whether spironolactone should be recommended for treatment of HFpEF in women.
COMPETENCY IN MEDICAL KNOWLEDGE: Heart failure with preserved ejection fraction represent roughly half of all heart failure cases and predominantly affects women. There are currently no proven therapies that clearly improve long-term survival or reduce heart failure hospitalizations.
TRANSLATIONAL OUTLOOK: This analysis of subjects with HFpEF enrolled from the Americas in the TOPCAT trial shows that women and men present with distinct clinical profiles, although prognosis is poor for both. It also shows a possible improvement in all-cause mortality outcome associated with spironolactone therapy in women but not in men. Previous studies have also suggested that spironolactone could have a greater protective effect against adverse cardiac remodeling in women than in men. If future prospective studies confirm the present observations, spironolactone therapy may be a new treatment option for HFpEF in women.
The authors thank TOPCAT investigators for conducting this trial and making these data available. The authors also wish to thank Karen and Steven Leaffer for their generous support of this research.
Supported by American Heart Association award 17IG3366030, U.S. National Institutes of Health/National Heart, Lung, and Blood Institute grants 1K08HL125725 and HH SN268200425207C, and Jacqueline’s Research Fund, Center for Women’s Health Research, University of Colorado School of Medicine. Dr. Sweitzer has received grants from Novartis and Merck Sharpe & Dohme; and is a consultant for Myokardia. Dr. Lindenfeld is a consultant for Novartis, Abbott, Edwards Lifesciences, Boston Scientific, Relypsa, VWave, CVRx, Boehringer Ingelheim, and ImpulseDynamics. All other authors have reported that they have no relationships relevant to the contents of this paper to disclose. John Teerlink, MD, served as Guest Editor for this paper.
- Abbreviations and Acronyms
- angiotensin-converting enzyme
- angiotensin receptor blocker
- heart failure with preserved ejection fraction
- heart failure with reduced ejection fraction
- left ventricular ejection fraction
- mineralocorticoid antagonist
- renin-angiotensin-aldosterone system
- systolic blood pressure
- Received November 29, 2018.
- Revision received January 11, 2019.
- Accepted January 13, 2019.
- 2019 American College of Cardiology Foundation
- Rossignol P.,
- Cleland J.G.F.,
- Bhandari S.,
- et al.
- Pfeffer M.A.,
- Claggett B.,
- Assmann S.F.,
- et al.
- de Denus S.,
- O’Meara E.,
- Desai A.S.,
- et al.
- Rickham P.P.
- Bristow M.R.,
- Enciso J.S.,
- Gersh B.J.,
- et al.
- Hegde S.M.,
- Claggett B.,
- Shah A.M.,
- et al.
- Selvaraj S.,
- Claggett B.,
- Shah S.J.,
- et al.
- Vaduganathan M.,
- Claggett B.L.,
- Chatterjee N.A.,
- et al.
- Cikes M.,
- Claggett B.,
- Shah A.M.,
- et al.
- Sanders N.A.,
- Supiano M.A.,
- Lewis E.F.,
- et al.
- Lam C.S.P.,
- Carson P.E.,
- Anand I.S.,
- et al.
- O’Meara E.,
- Clayton T.,
- McEntegart M.B.,
- et al.
- Bristow M.R.,
- Enciso J.S.,
- Gersh B.J.,
- et al.
- Lam C.S.P.,
- Brutsaert D.L.