Gender Differences in Risk Factors Associated With Pulmonary Artery Systolic Pressure, Heart Failure, and Mortality in Blacks: Jackson Heart Study.

Background Pulmonary hypertension is prevalent in black individuals, especially women. Elevated pulmonary artery systolic pressure (PASP) is associated with significant morbidity and mortality. Methods and Results We developed linear and proportional hazards models to examine potential gender-related differences in risk factors for elevated PASP (estimated by transthoracic echocardiography) and PASP-associated clinical outcomes (incident heart failure admissions and mortality) in JHS (Jackson Heart Study) participants. JHS is a prospective observational cohort study of heart disease in blacks from the Jackson, Mississippi, metropolitan area. The study cohort included participants with measurable transtricuspid gradients (n=3286) at the time of first/baseline examination, 2000-2004. The median age (interquartile range) of patients at baseline was 57.8 years (18.6 years) with 67.5% being women. The median PASP at baseline was higher in women (men: 26 mm Hg [interquartile range 8], women: 27 mm Hg [interquartile range 9]. In multivariate linear regression analyses with PASP, significant gender interactions were noted for age, chronic lung disease, pulse pressure, and obstructive spirometry. In exploratory analyses stratified by gender, body mass index, and obstructive and restrictive spirometry patterns were associated with PASP in women, and chronic lung disease was associated with PASP in men. Age and pulse pressure had stronger associations with PASP in women compared with men. There was a significant interaction between gender and PASP for heart failure admissions but not mortality. Conclusions Specific cardiopulmonary risk factors are associated with elevated PASP in women and men. Women with elevated PASP have a higher risk of incident heart failure admissions. Future research is needed to understand associated gender-specific mechanisms that can help identify targeted prevention and management strategies for patients with elevated PASP.

Clinical Perspective

What Is New?

Specific cardiopulmonary risk factors are associated with elevated pulmonary artery systolic pressure (PASP) in black women and men.

Body mass index and obstructive and restrictive spirometry patterns are associated with elevated PASP in women and chronic lung disease is associated with PASP in men.

Women with elevated PASP have a higher risk of incident heart failure admissions, compared with men.

What Are the Clinical Implications?

Elevated PASP is associated with gender‐specific risk factors in blacks, highlighting the need for guidelines and resource mobilization to enable early screening, increased surveillance, and target risk factor modification.

Introduction

Pulmonary hypertension (PH) is an important chronic illness in the black community,1, 2 with higher prevalence in women.1 PH is a known complication of left heart disease, diseases of the lung, metabolic diseases, and thromboembolic disease,3 and is associated with significant morbidity and mortality in black women.4, 5, 6, 7 Surveillance studies in the United States report a steady increase in PH‐related deaths and hospitalizations particularly in women and blacks.6 Despite these relevant clinical and epidemiological observations, reasons for high prevalence of PH, specifically nongroup 1 PH in black women, are poorly understood. Many comorbid risk factors for PH, including systemic hypertension, obesity, and diabetes mellitus, are prevalent in blacks8, 9, 10, 11 Whether these clinical comorbidities may disproportionately affect the risk for elevated pulmonary artery (PA) systolic pressure (PASP) and associated clinical outcomes in women compared with men is unknown.

Echocardiography with measurement of tricuspid regurgitant jet for estimation of PASP allows for further studies of factors associated with elevated PASP in large epidemiologic cohorts.1, 2 We sought to examine whether there are gender‐related differences in risk factors and clinical outcomes associated with elevated PASP in a large community‐based cohort of black patients. We hypothesized that the relationship between PASP and associated clinical characteristics and outcomes, namely future heart failure (HF) admissions and mortality, would be significantly different between black men and women.

Methods

Study Design and Population

We conducted post hoc analyses using data from JHS (Jackson Heart Study). The conduct of the JHS was approved by the University of Mississippi Medical Center Institutional Review Board. The participants gave written informed consent to participate in the research study. Requests to access the data set from qualified researchers trained in human subject confidentiality protocols may be sent to JHS at https://www.jacksonheartstudy.org/Research/Study-Data. JHS is a longitudinal, population‐based cohort study of cardiovascular disease that recruited noninstitutionalized black adult participants (N=5306) residing in the Jackson, Mississippi, metropolitan area.12 The conduct of the study was approved by the institutional review board of the University of Mississippi Medical Center. Participants answered predefined questionnaires, underwent venipuncture, echocardiography, and spirometry at the first examination between 2000 and 2004. The study cohort included participants who had echocardiography data available (n=5076), of whom those with a measureable tricuspid regurgitant jet velocity on echocardiography (allowing for estimation of the PASP, as described in the Outcome section) (n=3286) were included for final analysis (Figure 1). Additional details of the study population are included in Data S1.

Figure 1

A schematic depiction of our study cohort. JHS indicates Jackson Heart Study; TR, tricuspid regurgitant.

The characteristics of 2020 excluded participants with no tricuspid regurgitant jet measurements were comparable with the included study participants and are detailed in Table S1.

Exposure

The main exposure was participant gender, self‐identified by participants as male or female in the predefined questionnaire.

Outcome

The main outcome for linear regression analyses was PASP, estimated by echocardiography in units of millimeters of mercury. PASP was calculated by addition of 5 mm Hg (to account for the typical right atrial pressure [RAP]) to the transtricuspid gradient, derived by the modified Bernoulli equation.2, 13

For Cox proportional hazards analyses, the main outcome was all‐cause mortality, with time to death calculated from the time of the index echocardiographic examination; the mortality cutoff date was December 31, 2012. We also conducted a Cox proportional hazards analysis in which outcome was the time to incident HF requiring hospital admission, after adjudication based on available data on history, physical examination, laboratory analysis, and medication use as per procedures used in the ARIC (Atherosclerosis Risk in Communities) study. Event adjudication began on January 1, 2005, and HF admission data were available for a median of 8 years (35–2921 days). More details regarding adjudication procedures in JHS have been outlined elsewhere.14 Participants with a self‐reported HF hospitalization history before January 1, 2005, (n=112) were excluded for analysis of incident HF admissions.14

Clinical Covariates

Data on clinical covariates were collected at the time of the first/baseline examination, when echocardiographic measurements were also performed. Covariates for linear regression analyses with PASP included age, body mass index (BMI), coronary heart disease (CHD), diabetes mellitus, hypertension, brachial pulse pressure measurement, presence of severe mitral or aortic valvular heart disease on baseline echocardiography, history of chronic lung disease, and pattern on baseline spirometry measurement (normal, obstructive, restrictive). For Cox proportional hazards model for mortality, covariates included age, sex, BMI, American Heart Association (AHA) physical activity category, smoking status, total cholesterol measured at first examination, diabetes mellitus, history of HF, history of CHD, presence of hypertension, history of stroke, and estimated glomerular filtration rate. For Cox proportional hazards model of incident HF hospitalization, covariates included age, sex, diabetes mellitus, CHD, systolic blood pressure, BMI, heart rate on ECG at time of first examination, use of antihypertensive agents, and smoking status. Definitions for the clinical covariates used in this study are outlined in Data S2.

Echocardiographic Parameters

Details regarding echocardiographic data and procedures are outlined in Data S3.

Statistical Analysis

Baseline characteristics between men and women were compared using the Wilcoxon–Mann–Whitney test for continuous variables and χ2 analysis for categorical variables. Continuous variables were described as median with interquartile range. Linear regression analyses were conducted to assess for gender interactions for clinical covariates associated with PASP to assess possible effect modification by gender on associations of covariates with PASP. The PASP models were adjusted for age, BMI (kg/m2), brachial pulse pressure (mm Hg), hypertension, diabetes mellitus, CHD, severe mitral/aortic valvular heart disease, history of chronic lung disease, and spirometry profile (normal, obstructive, and restrictive)––a model adapted from Choudhary et al1 Interaction terms with gender were developed for each of the covariates outlined above and added individually to the models. Exploratory analyses stratified by gender were performed if variables with significant multiplicative interaction testing were present.

Cox proportional hazards modeling was used for analyses of mortality and incident HF hospitalization. The hazard ratio (HR) for all‐cause mortality associated with PASP was determined with an adjusted model for mortality adapted from Gu et al,15 adjusting for age, sex, BMI, physical activity, smoking status, high cholesterol, diabetes mellitus, history of HF, history of CHD, hypertension, estimated glomerular filtration rate, and history of stroke. To assess for effect modification of gender on PASP association with these outcomes, a gender×PASP interaction term was then added to the model. Next, the association of PASP with decompensated incident HF events requiring hospital admission was assessed. Cox proportional hazards modeling was used to determine the HR for HF events associated with PASP in an adjusted model of HF (ARIC model) from Agarwal et al,16 adjusting for age, sex, CHD, diabetes mellitus, systolic blood pressure, blood pressure medication use, heart rate, smoking status, and BMI. Participants who died before an HF event were censored. Similar to the mortality models, a gender×PASP interaction term was then added to the model.

If there was evidence of a significant gender interaction with PASP in the Cox proportional hazards models, exploratory analyses stratified by gender were then performed. Competing risk analysis for incident HF was also performed with all‐cause mortality as a competing event. Finally, to assess for potential differences in incidence of HF hospitalization by level of PASP in women and men, the incident rate of HF was plotted against groupings of PASP in men and women.

In the regression analyses with PASP as the outcome, we included all participants (N=3286). However, to ensure that incident HF episodes were identified, patients who self‐reported HF history (n=112) in their questionnaire were excluded from proportional hazards analyses that had incident HF as the outcome.

Missing data for clinical covariates were handled using multiple imputation, as outlined in Data S4. Distributions of missing covariates are listed in Table S2.

All analysis was performed using Stata/SE, version 15.1 software (StataCorp LP). A 2‐sided P value of <0.05 was considered significant.

Results

Baseline Characteristics

Table 1 shows the baseline clinical characteristics of the cohort by gender. The median (interquartile range) age of the study population was 57.8 years (18.6 years), with 67.5% of participants being women. The overall prevalence of obesity was 51.1%, hypertension was 56.1%, and diabetes mellitus was 19.4%. Women were more likely to be obese, exhibit poor to intermediate levels of physical activity per AHA categorization, and have hypertension and normal spirometry profiles. However, men were more likely to be smokers, to have a reduced left ventricular ejection fraction (EF), and to have obstructive or restrictive spirometry profiles. Men also had higher brachial pulse pressures than women. Median PASP was higher in women (men: 26 mm Hg [quartile 1–quartile 3: 22–30], women: 27 mm Hg [quartile 1–quartile 3: 23–32], P<0.05). The overall prevalence of PH (PASP ≥40 mm Hg) was 5.6%, with a higher prevalence in women (men: 4.04%, women: 6.35%; P<0.05).

Table 1

Baseline Characteristics by Gender

Characteristic	Total	Women, No. (%)	Men, No. (%)	P Value
Total	3286 (100)	2222 (67.6)	1064 (32.4)
Age, y
<45	666 (20.27)	445 (20.03)	221 (20.77)	0.560
45 to <55	784 (23.86)	517 (23.27)	267 (25.09)
55 to <65	950 (28.91)	651 (29.30)	299 (28.10)
≥65	886 (26.96)	609 (27.41)	277 (26.03)
AHA BMI categories
Obese	1674 (51.05)	1279 (57.69)	395 (37.19)	<0.001
Overweight	1106 (33.73)	646 (29.14)	460 (43.31)
Normal	499 (15.22)	292 (13.17)	207 (19.49)
Hypertension	1842 (56.06)	1305 (58.73)	537 (50.47)	<0.001
Diabetes mellitus	630 (19.40)	444 (20.25)	186 (17.65)	0.079
Severe left‐sided valvular heart disease	7 (0.22)	4 (0.18)	3 (0.29)	0.544
History of CHD	334 (10.69)	214 (10.08)	120 (11.98)	0.110
Reduced EF (EF <40%)	31 (0.95)	13 (0.6)	18 (1.69)
Mid‐range EF (EF 40–55%)	191 (5.84)	101 (4.59)	90 (8.48)
Preserved EF (EF >55%)	3045 (93.2)	2092 (94.83)	953 (89.82)	<0.001
Lung disease history	229 (6.98)	165 (7.44)	64 (6.03)	0.137
Smoker	639 (19.62)	347 (15.75)	292 (27.70)	<0.001
AHA physical activity categorization
Poor health	1634 (49.82)	1115 (50.27)	519 (48.87)	<0.001
Intermediate health	1039 (31.68)	742 (33.45)	297 (27.97)
Ideal health	607 (18.51)	361 (16.28)	246 (23.16)
Stroke history	146 (4.44)	89 (4.01)	57 (5.36)	0.08
BP medication use	1573 (51.9)	1161 (56.2)	412 (42.74)	<0.001
History of HF	112 (4.19)	72 (3.89)	40 (4.85)	0.25
Spirometry profile
Normal	2209 (71.05)	1530 (72.93)	679 (67.16)	0.004
Obstructive	278 (8.94)	172 (8.20)	106 (10.48)
Restrictive	622 (20.01)	396 (18.88)	226 (22.35)

Data are presented as number (percentage) or median (interquartile range). AHA indicates American Heart Association; BMI, body mass index; BP, blood pressure; CHD, coronary heart disease; EF, ejection fraction; HF, heart failure.

Demographic and Clinical Characteristics Associated With PASP and Interaction of PASP With Gender

After adjusting for relevant clinical characteristics in linear multivariate regression analyses, significant gender interactions with PASP were seen with age, BMI, diagnosis of hypertension, pulse pressure, history of chronic lung disease, and obstructive spirometry pattern. The results of interaction testing for each of the covariates are outlined in Table 2.

Table 2

Adjusted β Coefficients for Association of Gender Interaction Terms With PASP

Variable	Adjusted β Coefficient for Interaction Term (95% CI)a	P Value for Interaction Term
Age×gender	−0.06 (−0.09 to −0.02)	0.005
Male	NA	NA
BMI (kg/m2)×gender	−0.09 (−0.17 to −0.01)	0.029
Diabetes mellitus×gender	−0.21 (−1.04 to +1.46	0.745
Hypertension×gender	−1.21 (−2.18 to −0.24)	0.015
Pulse pressure (mm Hg)×gender	−0.04 (−0.08 to −0.01)	0.014
Severe left‐sided valve disease×gender	9.09 (−0.84 to +19.0)	0.073
CHD×gender	1.23 (−0.33 to +2.80)	0.122
Normal spirometry	1.00 (Ref)
Obstructive spirometry×gender	−1.98 (−3.67 to −0.30)	0.021
Restrictive spirometry×gender	−0.45 (−1.68 to +0.78)	0.47
Chronic lung disease×gender	2.79 (0.82–4.76)	0.005

Each individual regression model was adjusted for age, gender, body mass index (BMI), coronary heart disease (CHD), diabetes mellitus, hypertension, pulse pressure, severe mitral or aortic valvular disease, history of chronic lung disease, and spirometry category, as well as the designated gender interaction term. PASP indicates pulmonary artery systolic pressure.

Stratified Analysis in Men and Women

In exploratory analyses stratified by gender (Table 3), higher BMI and obstructive and restrictive spirometry patterns were associated with PASP in women, while no evidence of an association of these factors with PASP was found in men. Also, age and pulse pressure, while associated with PASP in both men and women, demonstrated a stronger association with PASP in women compared with men (adjusted β coefficient [SE] for age 0.18 [0.01] in women versus 0.14 [0.02] in men; adjusted β coefficient [SE] for pulse pressure 0.07 [0.01] in women versus 0.05 [0.02] in men). A history of chronic lung disease was associated with PASP in men, but no evidence of an association of chronic lung disease with PASP was found in women. In a supplementary analysis of spirometry values (forced expiratory volume in the first second of expiration [FEV1], forced vital capacity), there was a significant gender interaction with FEV1 and forced vital capacity (Table S3). In stratified analyses, FEV1 was significantly associated with PASP in both men and women, with a stronger association in women (adjusted β coefficient [SE] for FEV1 in women −2.63 [0.58] versus men −1.23 [0.62]) (Table S4). However, there was no evidence of an association of PASP with forced vital capacity in men or women (Table S4).

Table 3

Association of Clinical Characteristics With PASP Stratified by Gender

Variable	Men		Women
	Adjusted β Coefficient (95% CI)a	P Value	Adjusted β Coefficient (95% CI)a	P Value
Age, y	0.14 (0.10–0.17)	<0.001	0.18 (0.15–0.21)	<0.001
BMI, kg/m2	0.06 (−0.01 to 0.13)	0.092	0.17 (0.13–0.21)	<0.001
Hypertension	−0.23 (−0.48 to 1.70)	0.616	0.48 (−0.18 to 1.14)	0.157
Pulse pressure, mm Hg	0.05 (0.02–0.08)	0.004	0.07 (0.05–0.09)	<0.001
Obstructive spirometry	0.48 (−0.89 to 1.85)	0.494	2.72 (1.65–3.79)	<0.001
Restrictive spirometry	0.49 (−0.56 to 1.54)	0.360	1.05 (0.31–1.79)	0.006
Chronic lung disease	2.94 (1.27–4.62)	0.001	0.13 (−0.92 to 1.18)	0.813

Adjusted for age, body mass index (BMI), pulse pressure, hypertension, diabetes mellitus, coronary heart disease, severe valvular disease, history of chronic lung disease, spirometry categories. PASP indicates pulmonary artery systolic pressure.

Relationship of PASP With Mortality and Interaction With Gender

The median (range) follow‐up time for the mortality analysis was 10.1 years (0–12.3 years). During the follow‐up period, 363 deaths occurred, including 216 deaths in women and 147 deaths in men. After adjustment for potential confounders, PASP was significantly associated with mortality (adjusted HR per 1 mm Hg of PASP, 1.03; CI, 1.01–1.04). No evidence of a gender interaction with PASP and mortality was seen in the adjusted analysis when an interaction term for PASP and gender was included (P interaction >0.05).

Relationship of PASP With Incident HF Admissions and Interaction With Gender

The median (range) follow‐up time for incident HF analysis was 8.0 years (0–8.0 years).

During the follow‐up period, 141 patients were admitted with decompensated HF, including 104 women and 37 men.

For HF events, there was evidence of a significant interaction between gender and PASP (P interaction <0.05). In analyses stratified by gender, elevated PASP increased the hazards of incident HF in women, while no evidence of a relationship was found between PASP and hazards of HF in men (adjusted HR per 1 mm Hg of PASP, 1.05; 95% CI, 1.02–1.07 in women versus HR per 1 mm Hg of PASP, 0.99; 95% CI, 0.95–1.05 in men) (Table 4). No evidence of a gender interaction was observed in competing risk analysis for incident HF, when death was treated as a competing risk.

Table 4

Gender‐Stratified Cox Proportional Hazards Analysis of Association of PASP With Incident Decompensated HF

	Men		Women
	Adjusted HR (95% CI)a	P Value	Adjusted HR (95% CI)a	P Value
PASP, mm Hg	0.99 (0.95–1.05)	0.87	1.05 (1.02–1.07)	<0.001

Patients who self‐reported heart failure (HF) history (n=112) in their questionnaire were excluded from analyses that had incident HF as the outcome.

Adjusted for age, gender, diabetes mellitus, coronary heart disease, systolic blood pressure, body mass index, heart rate, use of antihypertensive agents, and smoking status. HR indicates hazard ratio; PASP, pulmonary artery systolic pressure.

Differences in incidence rates of HF in men and women by groupings of PASP is displayed in Figure 2, showing an increase in the incident rate of HF with increasing PASP grouping in women but not in men.

Figure 2

Incidence rates of heart failure in men and women according to pulmonary artery systolic pressure (PASP) groupings. Vertical lines represent 95% CIs of incidence rates of heart failure.

Discussion

To the best of our knowledge, our study is the first to examine gender‐specific differences in pathogenesis and prognosis of elevated PASP in a vulnerable black cohort. While female predominance has been observed in group 1 PH,7, 17 epidemiological data from population‐based studies and nongroup 1 PH is limited. We report that there was evidence of a significant gender interaction with a number of clinical risk factors for elevated PASP in blacks. Gender‐stratified exploratory analyses found that black women with specific risk factors, namely higher BMI and obstructive or restrictive spirometry profile, are likely to have higher mean PASP compared with men with similar risk factor profiles. Furthermore, age and brachial pulse pressure were more strongly associated with PASP in women than in men in these exploratory analyses. Finally, elevated PASP was noted to be related to subsequent incident HF admissions in women, but there was no evidence of such a relationship in men, while there was no evidence of a gender‐interaction with PASP with regards to overall mortality. The relationship between PASP and incident HF events in women was independent of important clinical comorbidities, suggesting that elevated PASP can serve as an important marker to identify women who are at risk for subsequent HF admission. Given the known higher prevalence of PH and associated poor prognosis in black women, our study findings can potentially help identify relevant patients to be targeted for the study of prevention and management strategies.

Gender Differences in Risk Factors Associated With PASP

We report that obese women are more likely to have elevated PASP compared with men with similar risk factors. Our observation is particularly relevant given the high prevalence of obesity in women participants in JHS (57% in our cohort), and in the overall US population with higher age‐adjusted prevalence of obesity in women (40.4%) compared with men (35.0%).18 Obesity is a major global health and economic challenge19 with significant associated morbidity and mortality.20 Obesity is a well‐known risk factor for global microvascular dysfunction including in the pulmonary circulation and has been linked with PH through common physiologic syndromes such as obstructive sleep apnea (OSA) and obesity hypoventilation syndrome (OHS) causing group 3 PH and cardiomyopathy of obesity causing group 2 PH. Although previous studies have reported strong correlations between obesity and elevated PASP,21, 22 data on gender‐based differences are limited. In the background of high obesity burden in women, our study results can also potentially be attributed to significantly higher leptin/adiponectin ratio in obese women compared with men,23 and subsequent effects on leptin‐mediated proinflammatory pathways,24, 25 structural modeling, and endothelial dysfunction.26, 27 Women with OSA are reported to have higher waist‐hip ratio and elevated BMI,28, 29 causing differential hormonal regulation of ventilatory responses,30 also likely contributing to PH.

We also observed that obstructive and restrictive spirometry patterns appear to be associated with elevated PA pressures in women but not in men, while a history of chronic lung disease was associated with higher PASP in men but not women. Furthermore, decreased FEV1 values were more strongly associated with elevated PASP in women than men. These gender‐related differences have not been extensively studied previously but are significant given the higher prevalence of obstructive airway disorders in women31 and associated increased rates of hospitalizations32 and mortality.33 Some potential mechanisms that can explain this association include exaggerated vascular inflammation related to increased secretion of interleukin 4 and interleukin 13 in women,34 role of estrogen in hypoxia‐related vascular remodeling,35 and estrogen receptor–mediated expression of CYP1A1 that promotes susceptibility to oxidant damage.36 Restrictive spirometry patterns have also been shown to be associated with increased arterial stiffness in men and women.37 The exact mechanisms mediating the association of PASP and restrictive spirometry especially in women remains unclear but can be linked to similar hormone‐related activation of inflammatory pathways.38, 39 The association of a prior diagnosis of chronic lung disease being associated with higher PASP values in men but not women may relate to differences in presentation and the likelihood of accurate diagnosis of lung disease in men versus women.40 However, these intriguing gender differences in the association of lung disease with PASP require confirmation and further study in diverse populations.

In our study, elevated PASP in women is more strongly influenced by age compared with men. PH is being increasingly diagnosed in the elderly and associated with poor prognosis41; however, studies on differences in gender distribution are sparse. Redfield et al42 evaluated age‐ and gender‐related interaction effects on vascular and cardiac parameters and reported that advancing age in women was associated with an increase in vascular and ventricular systolic/diastolic stiffness even in the absence of cardiovascular disease. Previous studies have reported other vascular indices such as central large artery stiffness, pulse pressures, and systemic pressures to be increased in older women compared with men.43, 44 These effects on cardiac and vascular indices likely contribute to the development of PH through a decrease in PA compliance,45, 46 increased pulmonary vascular resistance,46 associated ventricular diastolic dysfunction, and left atrial enlargement.47, 48 Animal studies have attributed this age‐dependent gender‐related vascular heterogeneity to the effect of menopause‐related hormonal changes on vascular remodeling49 and differential modulation of the renin‐angiotensin aldosterone system and sympathetic systems.50, 51

Gender Differences With PASP and Associated Clinical Outcomes

Finally, we found that in blacks, elevated PASP was a risk factor for poorer prognosis, including HF admissions and mortality. Our study specifically suggests that women with elevated PASP have a higher risk of incident/new decompensated HF admissions requiring hospitalization. The result was consistent after adjustment for relevant risk factors, suggesting that pulmonary pressures could be an independent predictor for decompensated HF admissions in women. Many studies have previously reported increased risk of HF admissions (mainly HF with preserved EF phenotype) in women compared with men.52, 53 Although elevated PASP has been identified as a possible risk factor for HF admissions in multiple studies,54, 55 including an analysis of the JHS cohort,2 to the best of our knowledge, this is the first study to report elevated PASP as a distinct gender‐specific risk factor for HF admissions in black women. Mechanisms underlying this observation could be related to aforementioned direct hormonal influences on increased vascular stiffness and risk of diastolic dysfunction in women that manifests as clinical HF. HF is a major public health problem56, 57 with significant morbidity and mortality burden; hence, future research in understanding the exact mechanisms outlining the role and mechanisms of PASP in HF risk is needed.

Elevated PASP in this cohort of black patients was associated with decreased survival with no evidence of gender interaction. Our observation is consistent with surveillance studies that have reported highest mortality rates in blacks with PH compared with other ethnicities.7 Multiple studies have established black race as an independent risk factor for mortality in patients with PH.58, 59 These observations have significant public health and clinical implications given the known background of under‐recognition and delayed diagnoses of PH7 in this at‐risk population. Our results further emphasize the need for guidelines and resource mobilization to enable early screening and increased surveillance and target risk factor modification in this vulnerable population. While interaction analysis in our study did not suggest gender‐based differences in mortality risk, multiple epidemiological studies have reported trends of higher mortality rates in women with PH,4, 7, 60 but the mechanisms driving differential mortality risk remain unclear. A few earlier studies had observed that women with connective tissue disease–related PAH, namely systemic sclerosis, had higher risk of death, even after accounting for demographic and hemodynamic characteristics.4, 61 The risk factor profile and distribution of PH groups of the women participants in our study is likely different from these studies, which could explain our equivocal results.

Limitations

Our study has some limitations. Our study design is observational and hence these results cannot be used to determine causality. Residual confounding may also be present despite attempts to account for known confounding variables. A few of the clinical covariates were self‐reported through questionnaires and subject to recall bias. Our main outcome, PASP, has been estimated by echocardiography and is likely prone to measurement bias. There were no inferior vena cava measurements, so we used an assumed right atrial pressure of 5 mm Hg as in a previous population‐based cohort study.13 This may have led to underestimation of PA pressure in participants with more elevated right atrial pressures. Presence of obesity, which was prevalent in this cohort, can also make echocardiography measurements technically difficult.62 Although an invasive right heart catheterization–derived PA pressure measurement is considered the gold standard, this is not practical to perform in population‐based studies. There may be a lack of precision in measured PASP that may have slightly affected the associations identified in the study. However, many cohort studies have demonstrated good sensitivity, specificity, and reliability of echocardiographic pulmonary hemodynamic measurements in comparison with right heart catheterization measurements.63, 64 We do not have detailed phenotyping data to adjudicate for PH classification (World Health Organization groups 1–5); however, ongoing studies such as PVDOMICS (Redefining Pulmonary Hypertension Through Pulmonary Vascular Disease Phenomics) that are performing detailed phenotypic assessment of patients may be able to explore these in the future.65

While the study enrollment started between 2000 and 2004, adjudication of HF events started in 2005. Thus, we may be underestimating the association between PASP and HF hospitalization assuming that the HF event rates remained constant throughout the study period. We are unable to comment on temporal trends in relation to HF events or mortality, since we are using PASP measurements from the baseline visit only. We do not have data regarding EF at the time of admission for decompensated HF and therefore cannot differentiate between the nature of HF (HF with preserved EF, HF with reduced EF, or HF with midrange EF) leading to hospitalization. Finally, given the multiple tests performed to assess gender interaction with covariates in the models, type 1 errors may have been made in concluding significant gender interactions were present.

Conclusions

Distinct cardiopulmonary risk factors are associated with elevated PASP in black women and men. This knowledge may assist in screening and early diagnosis of PH in at‐risk members of the black population. Women with elevated PASP appear predisposed to a higher risk of incident HF admissions compared with men. These findings warrant future research to identify specific genetic and hormonal mechanisms to enable targeted prevention and management of PH.

Sources of Funding

JHS is supported and conducted in collaboration with Jackson State University (HHSN268201800013I), Tougaloo College (HHSN268201800014I), the Mississippi State Department of Health (HHSN268201800015I/HHSN26800001), and the University of Mississippi Medical Center (HHSN268201800010I, HHSN268201800011I, and HHSN268201800012I) contracts from the NHLBI and the National Institute for Minority Health and Health Disparities. G.C. is supported by a Department of Veterans Affairs MERIT award I01CX001892, NIH/NHLBI R01HL128661, and NIH/NHLBI R01HL148727.

Disclosures

None.

Data S1. Study design and population.

Data S2. Clinical covariates.

Data S3. Echocardiographic parameters.

Data S4. Statistical analysis.

Table S1. Comparison of Baseline Characteristics in Study Participants vs Excluded Participants

Table S2. Distribution of Missing Covariates in Analytic Sample

Table S3. Adjusted β Coefficients for PASP and Interaction Analysis With Gender

Table S4. Association of Clinical Characteristics With PASP Stratified by Gender

Click here for additional data file.