Sex Differences in Right Ventricular Systolic Function and All-Cause Mortality in Tricuspid Regurgitation

There are important sex differences in the incidence, management, and outcome of valvular heart disease (VHD) [1]. However, most research has so far focused on left-sided VHD, primarily aortic stenosis [2], and left ventricular (LV) remodeling. The impact of sex differences on tricuspid valve (TV) disease has received little attention. The Framingham Heart Study showed a prevalence of more than mild tricuspid regurgitation (TR) in 18.4% of women and 14.8% of men. On multivariable analysis, the principal determinants were shown to be age, body mass index, and female sex [3]. Other studies have similarly demonstrated a consistent relationship with TR and sex [3‒5]. Although mild TR is generally considered to be associated with a benign prognosis, moderate and severe TR is associated with right ventricular (RV) dysfunction and comparable mortality rates if left untreated [6]. With evidence also suggesting that women represent an independent risk factor for poor perioperative outcomes [7, 8], it is surprising that there continues to remain a lack of scientific data in the literature about sex differences in TR, and whether aside from prevalence, sex influences RV performance, pulmonary hypertension, and clinical outcomes. Hence, the principal aim of the current study was to explore whether sex differences in patients with significant TR influence RV function and pulmonary hypertension. The secondary aim was to determine whether sex influences all-cause mortality in patients with significant TR.

A total of 12,791 consecutive transthoracic echocardiograms, performed between January 2010 and December 2010 at Guy’s and St Thomas’ Hospital, were reviewed to identify patients aged >18 years with significant TR (moderate or severe). Patients with prior TV replacement or repair or congenital heart disease involving the TV were excluded. The definition of hypertension was based upon history of hypertension, use of antihypertensive medications, or clinic blood pressure (BP) >140/90 mm Hg. Hypercholesterolemia was defined as the use of lipid-lowering drugs. The time to death or valve intervention was recorded for all patients.

Transthoracic echocardiography was performed with Vingmed systems 5, 7, or 9 (GE Medical, Milwaukee, WI, USA) or an “Epiq 7” (Philips Medical Systems, Bothell, WA, USA) ultrasound machines. LV ejection fraction (LVEF) was measured mainly by the biplane Simpson’s rule [9] or by visual estimates when image quality was poor. RV function and dimensions were measured according to the guidelines [9]. RV S’ was measured in an apical 4-chamber view [10]. The peak trans-tricuspid pressure gradient was estimated from the velocity of tricuspid regurgitant jet (V_max) by applying the simplified Bernoulli equation. The systolic pulmonary artery pressure (SPAP) was defined as a sum of peak trans-tricuspid pressure gradient and estimated RA pressure based on the inspiratory changes in the inferior vena cava dimension [11]. Pulmonary hypertension was defined as SPAP ≥40 mm Hg [10, 11]. The tricuspid annular plane systolic excursion (TAPSE) was measured from a zoomed M-mode on the lateral tricuspid annulus using an apical 4-chamber view, and divided by SPAP to calculate the TAPSE/SPAP index [12]. RV wall tension (RV WT) was derived from the RV basal-to-apex length at end-diastole multiplied by RV systolic pressure. Increased RV WT was defined as SPAP × RV length >3,300 mm Hg × mm [13]. RV systolic dysfunction was defined by either TAPSE <16 mm or RV S’ <10 cm/s [12]. The severity of TR was assessed semiquantitatively based on the vena contracta width and density of the continuous wave signal as previously reported [12]. Mitral and aortic valve disease was graded according to guidelines [14, 15].

All-cause deaths and the frequencies of valvular interventions were recorded during follow-up by reviewing the electronic patient record or a death certificate with February 17, 2019, as the censoring date.

SPSS version 26.0 (IBM corporation, Armonk, NY, USA) was used for statistical analyses. Comparison between women and men was performed by Student’s t test and χ². Kaplan-Meier curves were used to calculate event-free survival rates between women and men, and the differences were tested by Log-Rank test. The predictors of all-cause mortality in women and men were identified in univariate and multivariate Cox proportional Hazard models. All variables with p < 0.1 in univariate Cox regression analyses or clinically relevant, or being known prognosticators (TAPSE, LVEF), were included in the multivariate Cox regression models. Because of collinearity, SPAP, pulmonary hypertension, or TAPSE/SPAP index were not entered in the same model with RV WT. A p value <0.05 was considered to be statistically significant.

Of the 209 patients included in the present study, 117 (56%) were women and 92 (44%) were men. The mean (±SD) ages of women and men were 72.6 ± 13 and 70.8 ± 15.8 years, respectively (p = 0.366). There was no difference in body mass index, clinic BP, cardiovascular morbidities, and TR severity grade between women and men (Tables 1, 2). TR was primary in 2.6% and secondary in 97.4% in women, and primary in 9.8% and secondary in 90.2% in men (Table 2). The raw (unindexed) RV basal diameter was smaller and RV end-diastolic length (RVED-L) shorter in women than men (Table 2). When adjusted for body height, there was no significant difference in RVED-L between men and women, while there was a trend for shorter BSA-adjusted RVED-L in women than men (Table 2).

The mean RV WT was 3,170 ± 1,220 mm Hg × mm in women and 3,817 ± 1,499 mm Hg × mm in men (p = 0.002) (Fig. 1). The individual measures of RV systolic function (TAPSE, SPAP, and RV S’) did not differ between the sexes, but overall RV systolic dysfunction and LVEF <50% were more common in men. There was a strong inverse correlation between TAPSE/SPAP index and RV WT in both women (R² = −0.66, p < 0.001) and men (R² = −0.69, p < 0.001).

The rates of tricuspid and aortic valve replacement did not differ between the sexes, but women were more likely to undergo a mitral valve replacement compared to men (Table 1). The frequencies of moderate or severe mitral stenosis were also higher in women.

During a mean (±SD) follow-up of 69.4 + 33.4 months, a total of 123 deaths occurred: 68 (32.5%) in women and 55 (26%) in men (p = 0.808). On Kaplan-Meier analysis, no difference in mortality rates was found between women and men (Log-Rank p = 0.528) (Fig. 2). The univariable predictors of all-cause mortality in women and men are presented in Tables 3 and 4, respectively. In multivariable Cox regression analysis in women, higher age and increased RV WT were associated with a higher hazard ratio (HR) of all-cause mortality, and mitral valve replacement with a lower HR of all-cause mortality independent of coronary artery disease, TAPSE and LVEF <50% (Table 3). In the same multivariate model (Table 3), when increased RV WT (>3,300 mm Hg × mm) was replaced by RV WT as a continuous variable, 1SD higher RV WT (1,499 mm Hg × mm) was associated with a 2.2-fold increased HR of all-cause mortality (HR 2.18; 95% CI: 1.50–3.18, p < 0.001).

In a separate multivariable Cox regression analysis for men, higher age and increased RV WT were similarly associated with a higher HR of all-cause mortality independent of diastolic BP, coronary artery disease, atrial fibrillation, overall pacemaker or implantable cardioverter defibrillator (ICD) device, and TAPSE (Table 4). When TAPSE was replaced by TAPSE/SPAP index, increased RV WT remained a powerful predictor of all-cause mortality (HR 3.41; 95% CI: 1.35–8.62, p = 0.009) independent of age (HR 1.08; 95% CI: 1.04–1.12, p < 0.001), TAPSE/SPAP index (HR 0.36; 95% CI: 0.03–4.41, p = 0.424) and other nonsignificant covariables in the model (diastolic BP, coronary artery disease, atrial fibrillation, and overall pacemaker or ICD during follow-up) (data not shown). In the primary multivariate model (Table 4), when increased RV WT (>3,300 mm Hg × mm) was replaced by RV WT as a continuous variable, the association of 1SD higher RV WT (1,499 mm Hg × mm) with all-cause mortality attenuated (HR 1.32; 95% CI: 0.93–1.88, p = 0.119).

The principal finding of the current study is that women more commonly have preserved RV and LV systolic function and lower RV WT than men, despite comparable SPAP. The secondary finding is that both sexes have similar mortality rates. Age and increased RV WT are independent predictors of all-cause mortality in both sexes, while in women mitral valve replacement was associated with a lower HR of all-cause mortality. Although the rates of tricuspid and aortic valve replacement did not differ between the sexes, women had higher frequencies of moderate or severe mitral stenosis and were more likely to undergo a mitral valve replacement.

Over the past few years, there has been increasing focus on sex differences in VHD. Studies to date have largely investigated differences in patients with left-sided VHD, while the right heart has gained little attention. This is surprising in light of data showing an increased prevalence of TR in women [3‒5]. Some studies have indicated that female sex is associated with higher risk during combined valve and coronary bypass surgery [7], while other echocardiographic follow-up studies of TR patients have indicated male sex as a predictor of mortality [16]. Apart from some historical studies [3] or recent reviews [1, 17], there are very few studies exploring sex differences in TR [18‒20]. In a small study of 92 patients (60% men) with symptomatic severe TR undergoing isolated TV surgery, there were no significant sex differences in postoperative outcome after TV surgery, with 5-year survival 70.2 ± 11.4% for men and 76.3 ± 6.8% for women (p = 0.3) [19]. However, female patients had higher cardiovascular disease risk, were 11 years older, had more previous cardiac surgery, and had more cardiac arrhythmia than men. Similarly, in a cohort of 756 patients with severe aortic regurgitation, significant TR (≥2 + TR grade) was independently associated with a high mortality in both sexes [20]. However, female patients with ≥2 + TR had worse prognosis compared with those without TR, and the same held true for men.

In the present study of 209 patients with moderate or severe TR, we found no difference in age and cardiovascular disease burden between women and men. The mortality rates were high in both sexes with a 5-year survival of 66 ± 5% for women and 61 ± 5% for men. Although tricuspid and aortic valve interventions were comparable in women and men, mitral valve interventions were performed more frequently in women, likely due to a higher prevalence of mitral stenosis in women.

In our study, RV and LV systolic function were more commonly preserved in women than men. This difference was not explained by the TR severity grade, or the presence of coronary artery disease or aortic or mitral valve disease, which were similar in both sexes. This is consistent with the observation that heart failure with reduced LVEF is more common in men, while in the Framingham Heart Study, heart failure with preserved LVEF (≥50%) was twice as common in women compared to men [21]. In aortic stenosis, men have more often LV systolic dysfunction and dilatation than women [2]. However, there is very little evidence regarding sex differences in LV and RV remodeling and their impact on outcome in TR patients. In a previous large study of moderate and severe TR, Dietz et al. [22] showed that patients with a distinct phenotype of dilated RV with RV systolic dysfunction were more frequently men with coronary artery disease, impaired renal function, and systolic LV dysfunction. Patients presenting with systolic RV dysfunction/remodeling showed poorer outcome compared with patients with normal RV size and systolic function. However, although women in our study were less likely to have RV and LV dysfunction, there was no difference in all-cause mortality compared to men. This may be because the survival benefit of preserved systolic RV and LV function in women were offset by an equally high burden of complex comorbidities including ischemic heart disease, diabetes, atrial fibrillation, left-sided VHD as well as advanced age and a pulse pressure of ≥60 mm Hg, which is a surrogate marker of increased arterial stiffness in both women and men.

Both sexes shared age and increased RV WT as independent predictors of all-cause mortality. Recently, we presented data from this cohort of patients showing that increased RV WT (>3,300 mm Hg × mm) was not only common but associated with a 2.6-fold higher risk of all-cause mortality [13]. Male sex was the only independent determinant of increased RV WT. Consistent with this finding we now show that RV WT was significantly lower in women than men (3,170 ± 1,220 mm Hg × mm vs. 3,817 ± 1,499, p = 0.002) (Table 2). This finding may in part explain the observation of a stronger association between increased RV WT and all-cause mortality in men compared with women. A possible explanation for this finding may be preserved RV long-axis function and geometry and thereby lower RV WT in women than men despite a comparable SPAP. The later finding may be explained by lesser RV remodeling reflected by RVED-L.

There are some limitations to our study. First, although data were prospectively collected and entered in a clinical database at a heart valve clinic, the current analysis was retrospective and observational, which carries the risk of selection bias. The ideal method would be a propensity-matching. However, due to relatively low number of patients, a propensity-matching would result in further lower equal pairs for comparison between men and women. Second, the number of patients in the subgroups of men and women was low and therefore the results should be cautiously interpreted. Larger prospective studies are warranted to explore sex differences in RV function and association with all-cause mortality in patients with significant TR. Third, the data on functional class, heart failure-related hospitalizations, and medications were not available. Fourth, the data on diastolic dysfunction, which is more common in elderly women, were not collected. Finally, primary TR was more common in men than women; however, the information on the etiologies of primary TR was not available.

In patients with moderate or severe TR, age and cardiovascular disease burden were comparable in both women and men, while women more frequently had preserved RV and LV systolic function and lower RV WT. There was no difference in all-cause mortality between women and men, with and age and increased RV WT remaining predictors of all-cause mortality in both sexes.

Samples used in this study were obtained as part of routine medical care. Written informed consent from participants was not required in accordance with local guidelines. The study was approved by the local Institutional Review Board and NHS Health Research Authority (Protocol ID 260396/2019) and conducted according to the Good Clinical Practice guidelines and the principle outlined in the Declaration of Helsinki for all human or animal experimental investigations.

There are no conflicts of interest for any authors.

This paper was not funded.

Sahrai Saeed and Ronak Rajani: Conceptualization and design, data collection, analysis and interpretation, drafting of the manuscript, editing and final approval. Jenna Smith, Karine Grigoryan, Dario Freitas, Øyvind Bleie, and John B. Chambers: interpretation of data, writing-review and editing, and final approval.

All data generated or analyzed during this study are included in this article. Further enquiries can be directed to the corresponding author.