Sex Differences in Left Ventricular Function and Cardiac Mechanics

Women have higher left ventricular ejection fraction (LVEF) than men, yet clinical cutpoints for heart failure classification and treatment criteria are sex agnostic. ¹ LVEF is a widely used but imperfect global measure of systolic function. By contrast, speckle‐based measures of myocardial deformation have greater sensitivity for characterizing subclinical abnormalities in cardiac function. Although women are also known to have greater (more negative) myocardial strain than men. Whether sex differences in myocardial deformation mirror those observed with LVEF remains an unanswered question. ¹ We sought to examine sex differences in the association of LVEF and cardiac strain among community‐based individuals free of cardiovascular disease with preserved LVEF.

All data and materials have been made publicly available at the National Institutes of Health database of genotypes and phenotypes and can be accessed at https://www.ncbi.nlm.nih.gov/gap/. We included FHS (Framingham Heart Study) participants with LVEF≥50% and available echocardiographic measures. ² We excluded those with prevalent heart failure or myocardial infarction (N=218) and missing covariates (N=165). All participants provided informed consent and the Boston University Medical Center Institutional Review Board approved study protocols.

The linearity assumption of the relationship between LVEF and myocardial strain measures was confirmed using restricted cubic splines. We examined the association of self‐reported sex with myocardial strain ¹ (primary: global longitudinal strain [GLS], secondary: global circumferential strain [GCS]) using multivariable‐adjusted linear regression models adjusted for age, LVEF, heart rate, body mass index, diabetes, systolic blood pressure, hypertension treatment, and smoking. We tested whether sex modified the relation between LVEF and strain using sex*LVEF interaction terms. Finally, we compared differences in the relationship between LVEF and strain between men and women with high, moderate, and low cardiometabolic risk burden (based on adapted American Heart Association Life's Simple 7 score ³ ). A 2‐sided P value of <0.05 was statistically significant (R v.4.2.2).

Among 6107 individuals (55% women), mean age was similar in men and women (50±15 versus 51±16 years). Compared with men, women had lower body mass index (26.6±5.7 versus 28.1±4.5 kg/m²) and fewer comorbidities (hypertension: 23% versus 26%, diabetes: 4% versus 5%). LVEF was higher in women (66±5% versus 64±4%) and myocardial strain was greater in women versus men (GLS: −21±3% versus −19±3%, GCS: −31±5% versus −29±5%).

In sex‐pooled analyses, GLS was lower (more negative) by 0.7% for every 5% greater LVEF (β_adj−0.661 per 5% increase in LVEF, SE 0.040, P<0.0001). When examining sex differences, women had greater GLS compared with men, without statistically significant effect modification by sex across the LVEF range (Figure A). Specifically, GLS was 1.8% greater in women versus men (β_adj−1.783, SE 0.074, P<0.0001). Average GLS in participants with an LVEF of 50% was −16.5% in men versus −18.3% in women, with 33% of men versus 16% of women with LVEF between 50% and 55% meeting “abnormal GLS” criteria (clinical cutpoint of −18% ⁴ ). Female sex was similarly associated with greater GCS across the LVEF range again without significant effect modification by sex (β−1.202, SE 0.125, P < 0.0001, Figure B).

Figure . Comparison of the association of LVEF and myocardial deformation measures in women vs men.

The linear relationships between LVEF with myocardial deformation in men (blue) vs women (red) are displayed. For both sexes, lower LVEF was associated with worse cardiac mechanics including GLS and GCS. Women displayed greater (more negative) GLS (A) and GCS (B) compared with men across the LVEF range without effect modification. In (C), increasing risk factor burden was associated with worse (more positive) GLS across LVEF range (higher risk factor burden displayed in bolder curves, lower risk factor burden curves displayed in less bold curves). Women at all risk burden levels demonstrated greater GLS levels across the LVEF range compared with men. Dotted line in (A) and (C) display the clinical cutpoint for abnormal GLS (defined as > −18%). Risk factor burden was assessed using a composite point system adapted from the American Heart Association Life's Simple 7 with 2 points assigned for ideal health metrics excluding diet (high: >10 points, moderate: 7–9 points, low: <6 points). GLS indicates global longitudinal strain and LVEF, left ventricular ejection fraction.

Given known sex differences in cardiac remodeling in the setting of cardiometabolic disease, we compared LVEF with strain measures in men versus women stratified by cardiometabolic risk burden. Increasing risk factor burden was associated with worse GLS at any given LVEF (high versus low risk burden: β 0.634, SE 0.194, P=0.001). Women at all risk burden levels demonstrated greater GLS levels versus men (Figure C). Cardiometabolic risk burden did not modify the association of LVEF with GCS.

In a community‐based sample, we observed that sex differences in myocardial strain measures mirrored differences in LVEF, such that GLS was on average 2% greater across the range of preserved LVEF in women versus men. Although greater cardiometabolic risk burden was associated with worse GLS in both men and women, women had greater GLS versus men across the range of cardiometabolic risk burden. Our findings highlight 2 points. First, the persistence of greater GLS in women after stratification by cardiometabolic risk burden argues that our findings are not explained by greater sensitivity to risk exposures in women but potentially reflect inherent sex differences in cardiac mechanics, although individual cardiometabolic risk factors were not tested. Second, our findings may suggest that sex‐specific response to heart failure therapy (eg, greater benefit of angiotensin receptor neprilysin inhibitors at higher LVEF among women versus men) is not explained by worse cardiac mechanics in women but further examination of sex differences in cardiac mechanics among individuals with heart failure is warranted. ⁵ Our study has limitations. First, observed effect sizes were modest and clinical significance of findings requires further study. Second, we assessed cardiometabolic risk using an adapted Life's Simple 7 score as diet and sleep histories were not routinely ascertained in FHS. Finally, residual confounding cannot be excluded in observational studies. Taken together, our findings underscore the need to consider sex‐specific normative values for assessment of both LVEF and cardiac mechanics and to improve our understanding of distinct patterns of cardiac remodeling that occur in response to risk exposures and targeted therapies in men versus women.

Source of Funding

The Framingham Heart Study is supported by grants from the National Institutes of Health N01‐HC25195 and HHSN2682015000011. Dr Lau is supported by grants from the National Institutes of Health K23‐HL159243 and the American Heart Association 853 922. Dr Benjamin is supported by grants from National Institutes of Health R01‐HL092577 and the American Heart Association AF AHA‐18SFRN34110082. Dr Cheng is supported by grants R01‐HL131532, R01‐HL151828, and U54‐AG065141. Dr Ho is supported by grants from the National Institutes of Health R01‐HL160003, R01‐HL168889, and K24‐HL153669.

Disclosures

Dr Lau reports previous advisory board service for Astellas Pharma. The remaining authors have no disclosures to report.