Sex differences in mortality and hospitalization in heart failure with preserved and mid-range ejection fraction: a systematic review and meta-analysis of cohort studies

Introduction The influence of sex on the prognosis of heart failure with preserved or intermediate ejection fraction (HFpEF and HFmrEF) remains uncertain. This study aimed to investigate whether sex differences impact the prognosis of patients diagnosed with HFpEF and HFmrEF. Methods A comprehensive search across three databases (PubMed, the Cochrane Library, and Embase) was conducted to identify sex-related prognostic cohort studies focusing on HFpEF and HFmrEF. Risk estimates were synthesized using the random effects model. The analysis included 14 cohorts comprising 41,508 HFpEF patients (44.65% males) and 10,692 HFmrEF patients (61.79% males). Results Among HFpEF patients, men exhibited significantly higher rates of all-cause mortality (13 studies; hazard ratio (HR): 1.24, 95% confidence interval (CI): 1.15 to 1.33)) and cardiovascular disease mortality (5 studies; HR: 1.22, 95% CI: 1.14 to 1.31) compared to women. However, no significant difference was observed in HF admissions. For HFmrEF patients, men displayed notably higher all-cause mortality (HR: 1.21, 95% CI: 1.12 to 1.31) but no significant differences in cardiovascular mortality or HF admissions. Discussion These findings suggest that male patients diagnosed with HFpEF and HFmrEF may face a more unfavorable prognosis in terms of all-cause mortality. Variations were noted in cardiovascular mortality and HF admissions, indicating potential complexities in sex-related prognostic factors within these heart failure categories. In summary, male patients with HFpEF and HFmrEF may have a more unfavorable prognosis.


Introduction:
The influence of sex on the prognosis of heart failure with preserved or intermediate ejection fraction (HFpEF and HFmrEF) remains uncertain.This study aimed to investigate whether sex differences impact the prognosis of patients diagnosed with HFpEF and HFmrEF.Methods: A comprehensive search across three databases (PubMed, the Cochrane Library, and Embase) was conducted to identify sex-related prognostic cohort studies focusing on HFpEF and HFmrEF.Risk estimates were synthesized using the random effects model.The analysis included 14 cohorts comprising 41,508 HFpEF patients (44.65% males) and 10,692 HFmrEF patients (61.79% males).Results: Among HFpEF patients, men exhibited significantly higher rates of allcause mortality (13 studies; hazard ratio (HR): 1.24, 95% confidence interval (CI): 1.15 to 1.33)) and cardiovascular disease mortality (5 studies; HR: 1.22, 95% CI: 1.14 to 1.31) compared to women.However, no significant difference was observed in HF admissions.For HFmrEF patients, men displayed notably higher all-cause mortality (HR: 1.21, 95% CI: 1.12 to 1.31) but no significant differences in cardiovascular mortality or HF admissions.Discussion: These findings suggest that male patients diagnosed with HFpEF and HFmrEF may face a more unfavorable prognosis in terms of all-cause mortality.Variations were noted in cardiovascular mortality and HF admissions, indicating potential complexities in sex-related prognostic factors within these heart failure categories.In summary, male patients with HFpEF and HFmrEF may have a more unfavorable prognosis.

Introduction
There are approximately 64 million people in the world with heart failure (HF), and more than half of them are women (1)(2)(3).With economic development and the acceleration of population aging, the incidence of global HF is still increasing (3,4).The latest HF guidelines classify HF by ejection fraction (5): HF with reduced ejection fraction (HFrEF; EF = <40%), HF with midrange EF (HFmrEF; EF: 41%-49%), and HF with preserved EF (HFpEF; EF ≥ 50%) (6,7).More than half of HF cases are HFpEF and HFmrEF, with an increasing trend in recent years (2).Epidemiological studies provide evidence that sex influences the outcomes of HFrEF patients, particularly with regard to men, who exhibit higher all-cause mortality rates compared to women (5, 8, 9).However, the impact of sex differences on the prognosis of patients with HFpEF or HFmrEF is a topic that lacks clarity in the existing literature.Therefore, further research is needed to comprehensively understand the relationship between sex differences and prognosis in both HFpEF and HFmrEF.We reviewed the literatures to elucidate whether sex differences influence the prognosis of patients with HFpEF or HFmrEF.

Protocol registration and search strategy 2.2 Selection criteria and study selection
The criteria included in this study were as follows: (1) participant type: patients (age > 18 years) who were diagnosed with HFpEF or HFmrEF; (2) exposure and comparator: men vs. women; (3) outcomes: mortality [all-cause and cardiovascular diseases (CVDs)], HF admission, and all-cause admission; and (4) types of studies: observational cohort studies or post hoc analyses of clinical trials.
Studies with the following criteria were excluded: (1) no studies reported multivariate adjusted results.(2) articles with insufficient data (reviews, editorials, preclinical studies, practice guidelines, comments); and (3) studies with irrelevant purposes of this meta-analysis.
We imported all the literature preliminarily retrieved into management software (Endnote X9.2 software, Thomson Reuters, New York, NY).Then, we manually and automatically removed the duplicate literature and eliminated the remaining literature by reading the title and abstract.Finally, after preliminary screening, the complete literature that may meet the requirements was obtained.If there were any inconsistencies in the retrieval process, we resolved them through discussion (X.L.) to reach a consensus.

Data collection and quality assessment
We collected the following information by the predefined requirements for inclusion: study characteristics (first author's name, year of publication, region, origin of patients, type of design, and mean follow-up time), patient characteristics (sample size, age, sex, HF phenotype, and definition), and outcomes (adjusted hazard ratios (HRs), the corresponding 95% confidence interval (CI), and adjustments).Study quality was determined using the Newcastle-Ottawa Quality Scale (NOS) (11).

Statistical analysis
To elucidate the relationship between sex differences and prognosis in HFpEF and HFmrEF patients, we pooled the adjusted HRs with 95% CIs and used the inverse-variance method.We assessed the heterogeneity across the included articles using Cochrane's Q test (P < 0.1 marks significant).The inconsistency was assessed by the I 2 test (30%-50%: low, 50%-75%: moderate, >75%: high) (12).We used a random effects model due to potential heterogeneity within observational studies.
Subgroup analysis would be performed when the number of studies used for outcomes is greater than 10.Subgroup analysis was stratified according to the following factors: study design, sample size, region, mean follow-up time, and adjustment.According to the guidelines, when the number of studies included was more than 10, publication bias was evaluated by funnel plots, Egger's test, and Begg's test (13).Graphic abstracts and mechanisms were created in the Biorender web-based tool.We used sensitivity analysis by omitting each study or excluding studies with HFpEF with a definition of ejection fraction not less than 50% to evaluate the robustness.Data analysis was processed by Stata software (Version 16.0, Stata Corp LP, College Station, Texas, USA).P < 0.05 indicated a significant difference, and all results were tested bilaterally.

Quality of evidence
We assessed the quality or certainty of each outcome using the Grading of Recommendations Assessment, Development and Evaluation (GRADE) (14, 15).The quality of evidence for each result was evaluated by two authors, who provided evidence profile tables from the GRADEpro GDT (Guideline Development Tool).

Study features and study quality
The main information of the 14 qualified studies is presented in Table 1.Of the 14 included cohort studies published between 2006 and 2021, 7 were prospective cohort studies, and 7 were retrospective cohort studies.

Subgroup and sensitivity analyses and publication bias
Subgroup analyses for all-cause death were performed according to study design, sample size, region, mean follow-up time, and adjustment for confounders (Table 2).In addition to the adjustment of the subgroup, the differences between the other subgroups were similar (P > 0.05).A stronger association was shown in the group adjusted for New York Heart Association (NYHA) and estimated glomerular filtration rate (eGFR)/glomerular filtration rate (GFR) (P < 0.1).
Egger's test (P = 0.632), Begg's test (P = 0.583) and funnel plots did not show statistically significant bias in potential publication.Sensitivity analyses confirmed the robustness performed by omitting each study or excluding studies with HFpEF with a definition of ejection fraction not less than 50% (Supplementary Figures S1-S2).

Quality of evidence assessment
Evidence was graded according to GRADE.The studies included in this meta-analysis were all reasonable, rigorous, and high-quality cohort studies.Finally, from the six included outcomes, the GRADE assessment showed moderate certainty for all-cause mortality, CV mortality, and HF admission in patients with HFpEF and HFmrEF (Supplementary Tables S6-S7).

Major findings
In total, the study included 14 studies involving 52,200 patients with HFpEF (41,508) and HFmrEF (10,692).For HFpEF patients, men were significantly more likely than women to die from all causes and CVDs, but HF admission was not associated with sex differences; for HFmrEF patients, men were significantly more likely than women to die from all causes, but CV mortality and HF admission were not associated with sex differences.We systematically evaluated whether there are prognostic differences between men and women in patients with HFpEF and HFmrEF.
The sex difference in the prognosis of HFpEF and HPmrHF remains controversial.Previous studies reported a similar crude d mortality rate between sexes in patients with HF (27).Stolfo et al. showed that women had lower all-cause mortality (HR: 0.81, 95% CI: 0.76 to 0.87) and CV mortality (HR: 0.82, 95% CI: 0.76 to 0.89) than men among HFpEF patients, but HF admission did not decrease significantly (HR: 0.98, 95% CI: 0.91 to 1.05) (9).However, some studies have found no significant difference in the prognosis of HF by sex.For example, Blumer et al. showed that there was no prominent increase in all-cause mortality among male HFpEF patients.(HR: 1.12, 95% CI: 0.79 to 1.58) (16).Wang et al. showed that the prognosis was similar between men and women in HFpEF patients, including all-cause mortality (HR: 0.619, 95% CI: 0.240 to 1.593), CV mortality (HR: 0.690, 95% CI: 0.249 to 1.915) and HF admission (HR: 0.812, 95% CI: 0.312 to 2.114) (25).For patients with HFpEF, our results show that men were at greater risk for allcause and CV death, while HF admission was similar to that in women.In general, statistical power was generally reduced when there were fewer studies included or insufficient follow-up.Therefore, the preliminary conclusion that sex has a prominent effect on the prognosis of HFpEF needs to be established by more large sample size and prospective studies.
For HFmrEF patients, our results showed that men were at greater risk for all-cause death, while CV death and HF admission were similar to those in women.Insufficient studies may have resulted in a nonsignificant increase in CV mortality (3 studies), and more prospective studies are needed to demonstrate the association between sex differences and CV mortality in HFmrEF.
Our results showed no statistically prominent differences in the outcomes of death from any causes or CVDs and HF admission Forest plot of the association between sex differences and prognosis in patients with HFpEF.(A) Association between sex differences and all-cause mortality in patients with HFpEF.(B) Association between sex differences and CV mortality in patients with HFpEF.(C) Association between sex differences and HF admission in patients with HFpEF.In the forest plot, the diamond indicates the pooled estimate.Gray boxes are relative to study size, and the black vertical lines indicate 95% CIs around the effect size estimate.HFpEF, heart failure with preserved ejection fraction; CV, cardiovascular.between HFpEF and HFmrEF (all P > 0.1).Additionally, a large IPD meta-analysis consistently showed women had a lower ageadjusted all-cause mortality in either patient with HFpEF or HFrEF (interaction p value for EF group × sex = 0.72) (27), which reinforced our observation of better outcomes for women with HF compared with males, regardless of EF.The etiology of HF is an important confounding factor.Studies have shown that men are more likely to suffer from ischemic heart disease (IHD) (28,29).Our results showed that men had a higher all-cause mortality than women even after adjusting for IHD, and there was no difference between groups stratified by IHD (P = 0.966).These results suggested that IHD has no effect on death from any cause in HFpEF patients.
Diabetes is another vital potential mediator.Martínez's findings suggested that diabetes did not affect mortality for any cause (HR: 1.41, 95% CI: 1.35 to 1.47).It also found that among diabetic HFpEF patients, the HRs of men and women who died from any cause were not significantly different.However, among nondiabetic patients with HFpEF, men were more likely to die from any cause (27).The results from another study also suggested that sex did not influence mortality in HFpEF patients with diabetes but not in nondiabetic patients (18).Our subgroup analysis suggested a stronger relationship between men and all-cause death in subgroups with adjustment for diabetes mellitus (30) (P = 0.046).Subgroups stratified by adjustment for eGFR, NYHA, and heart rate had statistically prominent differences but not among subgroups adjusted for age, atrial fibrillation (AF), prior HF hospitalization, coronary heart disease (CHD), and obesity.Overall, these results suggested that the sex difference in prognosis in HFpEF could be partly explained by the kidney, diabetes, and severity of HF rather than IHD, AF, hypertension, age, and obesity.

Comparison with previous studies
The prior meta-analysis conducted by Manuel et al. showed that being male is independently associated with an increased risk of all-cause mortality in patients with both HFrEF and HFpEF (27).Furthermore, our study has revealed a link between male sex and cardiovascular mortality in HFpEF.
HFpEF is increasingly recognized as a syndrome with diverse phenotypes and various comorbidities.Notably, cardiac-related deaths account for only 27% of all-cause mortality in HFpEF patients, as opposed to 65% in HFrEF.This finding suggests that the disparity in mortality between sexs in HFpEF can be partly attributed to cardiac factors.Additionally, there has been limited exploration of sex-related differences and outcomes in HFmrEF.Our research has demonstrated that women tend to have better survival rates among patients with HFmrEF, underscoring the persistence of sex-related variations in prognosis regardless of ejection fraction.

Potential mechanism
The underlying mechanism of sex differences related to the prognosis of HF is unclear.In general, women with typical HFpEF have more complications (31), with hypertension and diabetes being the main cardiovascular risk factors associated with HFpEF (29).Men are more likely to suffer from HFrEF and HFmrEF (31), and ischemic cardiomyopathy is more common as a cause of HF (28,29).
The reason why women have a higher survival rate than men may be that they have better heart function and less ischemic cardiomyopathy (28,32).Studies have shown that estrogen, the main sex hormone in women, plays a crucial role in heart health.In addition to protecting the heart from cardiomyocyte hypertrophy and apoptosis, myocardial infarction size, and arrhythmia, estrogen reduces ischemic-reperfusion injury (IRI) (33)(34)(35)(36)(37).In addition, estrogen can regulate some risk factors for CHD, such as hypertension and hyperlipidemia, by reducing the vasoconstrictor endothelin and increasing the activity of lipoprotein lipase to prevent CHD and HF (38-41).

Clinical implications
HF treatment is aimed at reducing symptoms, improving survival, enhancing physical activity, and making patients live better (42).Treatment of patients with systolic dysfunction aims to reduce elevated filling pressures, decrease neurohormonal levels, and increase cardiac output.In patients with diastolic dysfunction, the main purpose of treatment is to improve ventricular relaxation and filling and reduce preload (30).However, there is no model for classifying treatment by sex.Our comprehensive study revealed that male patients with HFpEF and HFmrEF have a worse prognosis.Consistent with prior studies, women generally exhibited better prognoses than men, irrespective of EF.Consequently, further research is essential to better understand the observed sex difference in prognosis in patients with HFpEF and HFmrEF and how both pathophysiology and treatments contribute to this.

Study limitations
The present systematic reviews and meta-analyses have several limitations.First, half of the retrospective cohorts were included in the study.However, the subgroup analysis of prospective and retrospective studies was consistent, showing the robustness of the present study.Second, the EF of patients with HFpEF across the included studies was not uniform, and the EF range of some patients with HFpEF overlapped the EF range of patients with HFmrEF (Table 1).However, sensitivity analysis of all studies with an ejection fraction of no less than 50% still showed that our results were stable and reliable (Supplementary Figure S1B).There was variability in the HFpEF definition (with a cutoff of 40%, 45%, or 50%), resulting in some studies including HFmrEF patients as HFpEF, which is inconsistent with the latest HF Universal definition.These cutoffs might have over/underestimated the current findings.This constitutes one of the significant limitations of the present study and may limit its generalizability.Third, the number of included studies was limited, and more studies were included to prove the reliability of the conclusions.The other limitation is an inherent restriction of observational studies and the potential for some confounders not adjusted for-this should be included in the limitations.Last, patients included in these studies might have been categorized into HF phenotypes based on only one single measure of EF.In addition to the variability of study definitions, the variability in clinical assessment might also contribute to patients' misclassifications.Despite these limitations, it is important to consider sex differences in clinical settings, and our study provides valuable information for the design and analysis of clinical trials and animal studies related to HFmrEF and HFpEF, which are two types of HF with limited treatment options.

TABLE 2
Subgroup analysis of the impact of gender differences on all-cause mortality in patients with hFpEF.