Sex differences in clinical phenotypes and prognostic associations in acute heart failure

Background Women are under-represented in heart failure trials, and the risk markers used in acute heart failure (AHF) derive largely from male-predominant cohorts. In this study, we compared the metabolic, inflammatory, and renal profiles of men and women hospitalized for AHF and tested whether the prognostic value of these markers differs by sex. Methods We analyzed 530 consecutive patients from a prospective single-center AHF registry. Postdischarge mortality was modeled by Cox regression in the 503 patients discharged alive; each of eight covariates was tested for a sex interaction, and 19 metabolic, inflammatory, and renal markers were screened for effect modification by sex. Results The total number of male patients was 284 (53.6%). Women were older (82 vs. 79 years) than men and more often had preserved ejection fraction (65.9% vs. 35.9%), frailty (74.0% vs. 59.9%), and chronic kidney disease; they had a higher total cholesterol level (139 vs. 126 mg/dL), a lower troponin level, a higher E/e′ ratio within preserved ejection fraction, and a lower ferritin level (42 vs. 71 ng/mL; below 100 ng/mL in 76.8% vs. 63.7%). Men more often had an ischemic etiology (45.8% vs. 15.0%), although coronary anatomy was more often unassessed in women (all p < 0.05). In-hospital (5.3% vs. 4.9%) and postdischarge mortality rates (23.6% vs. 26.7%) and renal outcomes did not differ; after full adjustment, male sex was associated with higher postdischarge mortality [hazard ratio (HR) 1.51, 95% confidence interval (CI) 1.03–2.22], but this depended on the inclusion of the left ventricular ejection fraction (LVEF) category. No covariate showed significant effect modification by sex (all p for interaction ≥0.06), with similar optimism-corrected C-indices (men 0.683, women 0.687). Among the 19 markers, only uric acid showed an interaction (women HR 1.15 per mg/dL, 95% CI 1.00–1.33; men 0.91, 0.80–1.03; p = 0.008), not surviving multiplicity correction. Conclusions Men and women hospitalized for AHF are clinically and biochemically distinct populations. Unadjusted mortality rates were similar in both sexes; an adjusted association with male sex was sensitive to the inclusion of the LVEF category and therefore should be interpreted cautiously. No robust sex-related effect modification was found; given the limited number of events, this represents the absence of evidence rather than evidence of similar prognostic associations between the sexes. Low ferritin levels, frequent in women, warrants assessment with complete iron indices; the uric acid interaction is hypothesis-generating.