Abstract
Rationale
Incidence rates for pulmonary hypertension using diagnostic data in patients with cardiopulmonary disease are not known.
Objectives
To determine incidence rates of, risk factors for, and mortality hazard associated with pulmonary hypertension among patients referred for transthoracic echocardiography.
Methods
A retrospective cohort study was conducted using data from the U.S. Department of Veterans Affairs (VA) (1999–2020) and Vanderbilt University Medical Center (1994–2020). Pulmonary hypertension was defined as pulmonary artery systolic pressure >35 mm Hg, with prevalent cases excluded. Heart failure and chronic obstructive pulmonary disease were the primary exposures of interest. The primary outcome was incident pulmonary hypertension. Secondarily, we examined mortality rate after incident diagnosis.
Results
We identified 245,067 VA patients (94% men, 20% Black) and 117,526 Vanderbilt patients (46% men, 11% Black) without pulmonary hypertension, of whom 38,882 VA patients and 8,061 Vanderbilt patients developed pulmonary hypertension. Only 18–19% of patients with echocardiography-based pulmonary hypertension also had diagnostic codes. The hazard of pulmonary hypertension was fourfold higher in patients with heart failure and chronic obstructive pulmonary disease compared with patients without either. Mortality rates increased from pulmonary artery systolic pressure of 35–45 mm Hg and then plateaued. Independent risk factors for incident pulmonary hypertension included older age, male sex, Black race, and cardiometabolic comorbidities.
Conclusions
Pulmonary hypertension incidence rates estimated by diagnostic data are higher than code-based rates. Heart failure and chronic obstructive pulmonary disease strongly associate with incident pulmonary hypertension. Pulmonary artery systolic pressure >45 mm Hg at diagnosis is associated with high mortality. New pulmonary hypertension on echocardiography is an important prognostic sign.
Keywords: echocardiography, pulmonary hypertension, epidemiology
Pulmonary hypertension is associated with substantial morbidity and mortality and is most commonly due to heart failure (HF) or chronic obstructive pulmonary disease (COPD) (1, 2). The prevalence of pulmonary hypertension in HF is substantial, ranging from 40–75% in HF with reduced ejection fraction (HFrEF) and 36–83% in HF with preserved ejection fraction (HFpEF) (3). Pulmonary hypertension prevalence in COPD is less well described, varying with severity of underlying disease and present in up to 90% of patients with Global Initiative for Chronic Obstructive Lung Disease stage IV disease (4, 5) Despite the notable prevalence of pulmonary hypertension and its strong association with increased mortality, it is systemically underrecognized (6, 7). In contrast to virtually all common cardiopulmonary conditions, incident rates (IRs) for pulmonary hypertension are not well established. Right heart catheterization (RHC) is the gold standard for pulmonary hypertension diagnosis, but it is invasive and not routinely performed in all patients with HF or COPD. As a result, estimates of pulmonary hypertension incidence in the RHC referral population skew toward sicker, more complex patients. In contrast, transthoracic echocardiography (TTE) use for the estimation of pulmonary pressures has earned widespread acceptance as a noninvasive tool for identifying pulmonary hypertension and is routinely used for evaluating symptoms suggestive of pulmonary hypertension (8). Moreover, TTE-estimated pulmonary arterial systolic pressure (PASP) effectively identifies patients at risk of poor outcomes, with PASP as low as 33 mm Hg carrying an increased risk for mortality (9).
Only two prior studies have examined pulmonary hypertension IRs (10, 11). The first relied on International Classification of Diseases (ICD) coding to establish pulmonary hypertension diagnosis (10), with findings limited by the unclear accuracy of ICD codes for identifying pulmonary hypertension (12, 13). The second used TTE for diagnosis (11) but excluded left heart disease, the most common cause of pulmonary hypertension. To address the gaps in our understanding of pulmonary hypertension epidemiology, we examined IRs among individuals referred for TTE in a national healthcare system and a large, sex-balanced tertiary care center. We hypothesized that IRs of pulmonary hypertension would be substantially higher than previously reported, particularly among patients with HF and/or COPD, and that incident pulmonary hypertension would be associated with increased hazard of mortality.
Some of the results of these studies have been previously reported in preprint form (https://doi.org/10.1101/2024.10.08.24315117).
Methods
The Institutional Review Boards of the U.S. Department of Veterans Affairs (VA) Tennessee Valley Healthcare System, West Haven VA Medical Center, and Vanderbilt University Medical Center approved the study. The VA Birth Cohort allows data analysis through a waiver of the requirement to obtain consent.
We derived the exploratory cohort from participants in the VA Birth Cohort (14) and evaluated patients who underwent TTE between September 30, 1999, and October 1, 2020. For the validation cohort we extracted data from Vanderbilt’s deidentified electronic health record (EHR). We evaluated patients who underwent TTE between January 1, 1994, and August 31, 2020. Baseline was defined as the date of the first TTE with a reported PASP value. We excluded patients with pulmonary hypertension on baseline TTE. In the VA cohort, we excluded patients who underwent baseline TTE after December 31, 2016, to coincide with the end of available Medicare data, which supply several baseline variables.
Exposures, Covariates, and Outcomes
We selected HF and COPD as our primary exposures, defined by the presence of at least two outpatient or one inpatient ICD, Ninth Revision (ICD-9) or ICD, 10th Revision (ICD-10) code, as in prior studies (1, 15) The diagnostic codes used to determine the presence of COPD and HF are detailed in Table E1 in the data supplement. Our definition of COPD yielded a specificity greater than 90% compared with gold-standard diagnosis in large healthcare databases (16, 17). As a sensitivity analysis, we also assessed an EHR definition of COPD optimized for accuracy (incorporating ICD codes with age > 50 yr, history of smoking, and inhaler use) (17).
We collected data on demographics (age, sex, and race or ethnicity), body mass index, common clinical conditions or surrogates for disease associated with pulmonary hypertension (atrial fibrillation, scleroderma, hypertension, diabetes, hepatitis C infection, renal function, and Fibrosis-4 [FIB-4] score), substance use (smoking use, cocaine abuse, and alcohol abuse), and laboratory findings that may affect PASP estimation on TTE (hemoglobin) for cohort description and inclusion in regression models. These data were ascertained using a combination of clinical, laboratory, and/or ICD-9 or ICD-10 code data collected closest to the date of baseline TTE (up to 180 d afterward), as previously described (18–24).
We used a validated natural language processing tool to extract measurements from TTE reports (25, 26). When absent, we calculated the PASP using the simplified Bernoulli equation using the reported tricuspid regurgitant velocity. We assumed a right atrial pressure of 5 mm Hg when missing (27, 28). We excluded PASP estimates outside a physiological range (<8 or >159 mm Hg) on the basis of clinical experience. We defined HFpEF as left ventricular ejection fraction (LVEF) ≥ 50% and HFrEF as LVEF < 50%, on the basis of LVEF cutoffs frequently used in epidemiologic cohorts and clinical trials (29). Covariate definitions were identical between cohorts, except for smoking and cirrhosis data, which were unavailable in the Vanderbilt cohort.
The primary outcome was incident pulmonary hypertension, defined as PASP > 35 mm Hg. The rationale for this selected cutoff is detailed in the data supplement. We completed additional sensitivity analyses defining pulmonary hypertension as PASP > 40 mm Hg. Follow up time was defined by the time from baseline TTE to follow up TTE with pulmonary hypertension, with censoring for death or end of study (September 30, 2020). In the primary analysis, we conservatively assumed that patients with only one measured PASP value never developed pulmonary hypertension. This assumption categorizes patients without a detectable tricuspid regurgitant jet as never developing pulmonary hypertension, while in fact a missing tricuspid regurgitant jet does not rule out pulmonary hypertension (30). A sensitivity analysis required at least one additional TTE examination after baseline to define pulmonary hypertension incidence. To compare between TTE-based and ICD-based IRs, we calculated pulmonary hypertension IRs using ICD-9 and ICD-10 diagnostic codes using the same time frame as the TTE-based analyses.
The secondary outcome was all-cause mortality after incident pulmonary hypertension. In the VA cohort, participants were followed from the date of incident pulmonary hypertension through death or end of study. In the Vanderbilt cohort, the mortality analysis was truncated at the end of 2017, after which mortality data are unavailable. Patients who did not develop pulmonary hypertension served as a comparator group and were followed from the latest TTE examination until end of study. Further methodologic details on cohort derivation, exposures, covariates, and outcome analysis are available in the data supplement.
Statistical Analysis
Baseline characteristics are presented as frequencies and percentages for categorical variables and as median (interquartile range [IQR]) for continuous variables. We used Poisson regression to estimate IRs per 1,000 person-years. Cox proportional-hazards regression models were used to estimate the hazard ratios (HRs) attributable to clinical characteristics. For the mortality outcome, PASP was modeled using restricted cubic splines with three knots in a Cox model. The proportional-hazards assumption was tested using Schoenfeld residuals. We repeated the analysis using time-updated variables of HF, COPD, and estimated glomerular filtration rate. Kaplan-Meier curves were used to assess the differences in survival among subjects stratified by a four-level HF and COPD status variable and by PASP category with log-rank testing. All tests with two-sided P values <0.05 were considered statistically significant. All missing data were imputed using multiple imputation by chained equations (31). Cox survival models were fitted in each imputed dataset and then combined to obtain pooled HRs and standard errors according to Rubin’s rules (32). All analyses were performed using R version 4.0.2 (www.r-project.org).
Results
VA and Vanderbilt Cohorts
We identified 245,067 VA patients without pulmonary hypertension on baseline TTE. Patients were predominantly male (93.7%; n = 15,430 women), with a median age of 59.9 years (IQR, 55.1–64.1 yr). Black and Hispanic patients accounted for 19.7% and 5.2% of the cohort, respectively. Cardiac, metabolic, and pulmonary comorbidities were common at baseline (Table 1), particularly HF (18.2%) and COPD (28.1%). At the end of follow up, the prevalence of HF and COPD had increased to 36% and 40%, respectively. The median baseline PASP was 28 mm Hg (IQR, 23–32 mm Hg). A total of 87,007 patients had at least one repeat PASP estimate.
Table 1.
Clinical characteristics of the U.S. Department of Veterans Affairs cohort
| Overall* (n = 389,331) | Prevalent Pulmonary Hypertension (n = 137,939) | Final Analytic Cohort† (n = 245,067) | |
|---|---|---|---|
| Age, yr | 60.1 (55.4–64.2) | 60.4 (55.8–64.3) | 59.9 (55.1–64.1) |
| Male sex | 367,653 (94.4) | 132,116 (95.8) | 229,637 (93.7) |
| BMI, kg/m2 | 29.6 (25.8–34.2) | 29.8 (25.7–34.8) | 29.5 (25.9–33.9) |
| Race/ethnicity | |||
| Other | 12,443 (3.2) | 4,203 (3.0) | 7,986 (3.3) |
| Black | 80,364 (20.6) | 31,062 (22.5) | 48,396 (19.7) |
| Hispanic | 18,776 (4.8) | 5,525 (4.0) | 12,770 (5.2) |
| White | 259,145 (66.6) | 89,691 (65.0) | 165,165 (67.4) |
| HF | 102,105 (26.2) | 56,370 (40.9) | 44,637 (18.2) |
| HFrEF (LVEF < 50%) | 61,122 (15.7) | 35,754 (25.9) | 25,043 (10.2) |
| HFpEF (LVEF ≥ 50%) | 40,817 (10.5) | 20,494 (14.9) | 19,594 (8.0) |
| COPD | 125,722 (32.3) | 55,165 (40.0) | 68,877 (28.1) |
| Atrial fibrillation | 65,288 (16.8) | 30,281 (22.0) | 34,085 (13.9) |
| Alcohol abuse | 103,881 (26.7) | 39,616 (28.7) | 62,668 (25.6) |
| Cocaine abuse | 32,541 (8.4) | 12,267 (8.9) | 19,880 (8.1) |
| Scleroderma | 11,570 (3.0) | 4,436 (3.2) | 6,942 (2.8) |
| Smoking | |||
| Current | 147,385 (37.9) | 54,490 (39.5) | 77,130 (31.5) |
| Former | 117,090 (30.1) | 42,116 (30.5) | 57,897 (23.6) |
| Never | 113,694 (29.2) | 36,911 (26.8) | 56,503 (23.1) |
| Hypertension | |||
| None | 10,817 (2.8) | 2,797 (2.0) | 7,757 (3.2) |
| Controlled | 287,625 (73.9) | 99,357 (72.0) | 183,639 (74.9) |
| Uncontrolled | 77,558 (19.9) | 30,803 (22.3) | 45,609 (18.6) |
| Diabetes mellitus | 170,748 (43.9) | 67,588 (49.0) | 100,935 (41.2) |
| Hepatitis C infection | 47,028 (12.1) | 17,802 (12.9) | 28,558 (11.7) |
| FIB-4 score >3.25 | 32,359 (8.3) | 14,125 (10.2) | 13,980 (5.7) |
| eGFR, ml/min/1.73 m2 | 82.0 (66.0–98.0) | 80.0 (61.0–97.0) | 85.0 (71.0–101.0) |
| Hgb, g/dl | 14.0 (12.5–15.1) | 13.5 (11.7–14.8) | 14.20 (12.9–15.2) |
Definition of abbreviations: BMI = body mass index; COPD = chronic obstructive pulmonary disease; eGFR = estimated glomerular filtration rate; FIB-4 = Fibrosis-4; HF = heart failure; HFpEF = heart failure with preserved ejection fraction; HFrEF = heart failure with reduced ejection fraction; Hgb = hemoglobin; LVEF = left ventricular ejection fraction.
Data are expressed as median (interquartile range) or as n (%).
All with baseline pulmonary arterial systolic pressure values before December 2016.
Missing values in the final analytic cohort: race, 10,858 (4.4%); smoking status, 53,714 (21.9%); hypertension, 8,201 (3.3%); eGFR, 2,394 (1%); BMI, 3,332 (1.4%); FIB-4 score, 8,122 (3.3%); hemoglobin, 3,522 (1.4%).
The median time between baseline TTE and incident pulmonary hypertension was 3.4 years (IQR, 1.5–6.4 yr). Table E2 compares clinical characteristics stratified by the presence of a second TTE examination after the baseline study, and Table E3 compares clinical characteristics stratified by time between TTE examinations. There was a higher prevalence of HF (25.5% vs. 14.2%) and diabetes (46.6% vs. 38.2%) among individuals with follow up TTE compared with those without. Patients with shorter intervals between TTE examinations were generally older, with a higher prevalence of HF, COPD, and liver fibrosis (by FIB-4 score). The proportion of patients who developed incident pulmonary hypertension was highest in patients with follow up TTE within two years (47.6%) compared with longer time intervals (42.3–42.9%).
We identified 77,548 Vanderbilt patients without pulmonary hypertension on baseline TTE. Similar to the VA cohort, HF (21.4%) and COPD (12.3%) were common in the Vanderbilt cohort, as were other comorbidities (Table 2). Echocardiographic characteristics of both cohorts are displayed in Table E4. The number of PASP measurements was similarly distributed between the VA and Vanderbilt cohorts (see Figure E1).
Table 2.
Clinical characteristics of the Vanderbilt cohort
| Overall (n = 117,526) | Prevalent Pulmonary Hypertension (n = 36,586) | Final Analytic Cohort* (n = 77,548) | |
|---|---|---|---|
| Age, yr | 61.8 (48.5–72.0) | 66.9 (55.6–76.4) | 59.6 (46.0–69.9) |
| Male sex | 54,847 (46.7) | 17,488 (47.8) | 35,775 (46.1) |
| BMI, kg/m2 | 28.2 (24.3–33.1) | 28.4 (24.4–33.8) | 28.1 (24.2–32.9) |
| Race/ethnicity | |||
| Other | 8,262 (7.0) | 2,503 (6.8) | 4,783 (6.2) |
| Black | 14,178 (12.1) | 5,251 (14.4) | 8,634 (11.1) |
| Hispanic | 1,721 (1.5) | 403 (1.1) | 1,251 (1.6) |
| White | 93,365 (79.4) | 28,429 (77.7) | 62,880 (81.1) |
| HF | 34,810 (29.6) | 17,907 (48.9) | 16,603 (21.4) |
| HFrEF (LVEF < 50%) | 14,389 (12.2) | 8,050 (22.0) | 6,235 (8.0) |
| HFpEF (LVEF ≥ 50%) | 19,473 (16.6) | 9,300 (25.4) | 9,984 (12.9) |
| COPD | 18,344 (15.6) | 8,624 (23.6) | 9,549 (12.3) |
| Atrial fibrillation | 25,440 (21.6) | 11,850 (32.4) | 13,407 (17.3) |
| Alcohol abuse | 5,280 (4.5) | 1,809 (4.9) | 3,399 (4.4) |
| Cocaine abuse | 1,212 (1.0) | 407 (1.1) | 793 (1.0) |
| Scleroderma | 5,211 (4.4) | 1,852 (5.1) | 3,301 (4.3) |
| Smoking | |||
| Current | 7,654 (6.5) | 2,171 (5.9) | 5,219 (6.7) |
| Former | 17,773 (15.1) | 5,644 (15.4) | 11,759 (15.2) |
| Never | 34,755 (29.6) | 8,546 (23.4) | 25,163 (32.4) |
| Hypertension | |||
| None | 9,949 (8.5) | 1,020 (2.8) | 7,926 (10.2) |
| Controlled | 75,173 (64.0) | 23,460 (64.1) | 50,336 (64.9) |
| Uncontrolled | 21,677 (18.4) | 7,780 (21.3) | 13,418 (17.3) |
| Diabetes mellitus | 16,050 (13.7) | 6,197 (16.9) | 9,798 (12.6) |
| Hepatitis C infection | 3,365 (2.9) | 1,106 (3.0) | 2,234 (2.9) |
| FIB-4 score >3.25 | 12,817 (10.9) | 5,758 (15.7) | 6,918 (8.9) |
| eGFR, ml/min/1.73 m2 | 77.0 (57.0–97.0) | 67.0 (45.0–88.5) | 81.0 (63.0–100.0) |
| Hgb, g/dl | 12.50 (10.60–13.90) | 11.60 (9.80–13.30) | 12.80 (11.10–14.10) |
Definition of abbreviations: BMI = body mass index; COPD = chronic obstructive pulmonary disease; eGFR = estimated glomerular filtration rate; FIB-4 = Fibrosis-4; HF = heart failure; HFpEF = heart failure with preserved ejection fraction; HFrEF = heart failure with reduced ejection fraction; Hgb = hemoglobin; LVEF = left ventricular ejection fraction.
Data are expressed as median (interquartile range) or as n (%).
Missing values in the final analytic cohort: gender, 6 (<0.01%); smoking, 35,407 (45.7%); hypertension, 5,868 (7.6%); FIB-4 score, 28,617 (36.9%); eGFR, 12,673 (16.3%); BMI, 9,658 (12.5%); LVEF, 384 (0.5%); hemoglobin, 14,779 (19.1%).
Pulmonary Hypertension Incidence Rates
Incident pulmonary hypertension developed in 38,882 (15.9%) VA patients over a median follow up period of 6.6 years (IQR, 4.1–10.2 yr) and 8,061 (10.4%) Vanderbilt patients over a median follow up period of 2.7 years (IQR, 0.8–6.2 yr). The median change in PASP in the VA cohort was 14.0 mm Hg (IQR, 8.4–21.7 mm Hg) and in the Vanderbilt cohort was 13.1 mm Hg (IQR, 8.3–19.9 mm Hg). A total of 39,741 VA patients (16.2%) and 9,845 Vanderbilt patients (12.7%) had increases of 5 mm Hg or more between TTE examinations. PASP was 32–34 mm Hg on baseline TTE in 13.1% of the VA cohort and 8.2% of the Vanderbilt cohort. Patients with 32–34 mm Hg on their baseline TTE accounted for 17% of incident pulmonary hypertension cases in the VA cohort and 12.6% of cases in the Vanderbilt cohort.
The IR of pulmonary hypertension in the VA cohort was 22.0 per 1,000 person-years. HF, COPD, or both were present at baseline in 51% of patients who developed incident pulmonary hypertension. IRs were lower for COPD alone than HF alone and lowest among those with neither HF nor COPD (Table 3 and see Figure E2). The highest IRs were observed among those with both HF and COPD (60.3 vs. 15.7 per 1,000 person-years). Among patients with HF at baseline, pulmonary hypertension IRs were higher in those with HFrEF compared with HFpEF (62.7 vs. 43.9 per 1,000 person-years). We observed similar but consistently higher pulmonary hypertension IRs in the Vanderbilt validation cohort compared with the VA cohort (Table 3). Among the 76,007 VA patients and 21,653 Vanderbilt patients with at least one subsequent PASP measurement, IRs were twofold to threefold higher (see Table E5). Using a stringent EHR definition of COPD did not change IR estimates in the VA cohort (see Table E6).
Table 3.
Pulmonary hypertension IRs in the VA and Vanderbilt cohorts, stratified by HF/COPD status and within HF by LVEF
| VA Pulmonary Hypertension IRs |
Vanderbilt Pulmonary Hypertension IRs |
|||||
|---|---|---|---|---|---|---|
| n | Incident Pulmonary Hypertension | Rate/1,000 PY | n | Incident Pulmonary Hypertension | Rate/1,000 PY | |
| Total | 245,067 | 38,882 | 22.0 (21.8–22.3) | 77,548 | 8,061 | 25.5 (24.9–26.0) |
| HF−/COPD− | 152,456 | 19,111 | 15.7 (15.5–16.0) | 55,258 | 4,199 | 17.2 (16.7–17.7) |
| HF−/COPD+ | 47,974 | 7,128 | 22.6 (22.1–23.1) | 5,687 | 555 | 31.4 (28.9–34.1) |
| HF+/COPD− | 23,734 | 6,856 | 49.8 (48.6–51.0) | 12,741 | 2,524 | 56.4 (54.2–58.6) |
| HF+/COPD+ | 20,903 | 5,787 | 60.3 (58.8–61.9) | 3,862 | 783 | 79.8 (74.3–85.5) |
| HF+/LVEF < 50% | 25,043 | 7,966 | 62.7 (61.3–64.1) | 6,235 | 1,478 | 78.2 (74.3–82.2) |
| HF+/LVEF ≥ 50% | 18,594 | 4,677 | 43.9 (42.7–45.2) | 9,984 | 1,751 | 51.0 (48.6–53.4) |
Definition of abbreviations: COPD = chronic obstructive pulmonary disease; HF = heart failure; IR = incident rate; LVEF = left ventricular ejection fraction; PY = patient-years; VA = U.S. Department of Veterans Affairs.
Pulmonary hypertension IRs did not vary over time in the VA cohort (see Table E7) when stratified by five-year intervals beginning in 2000. Pulmonary hypertension IRs were approximately 30–40% lower when pulmonary hypertension was defined as PASP > 40 mm Hg (see Table E8) in both cohorts. Among patients with TTE-based incident pulmonary hypertension, only 19% of the VA cohort and 18% of the Vanderbilt cohort also had ICD codes for pulmonary hypertension (see Table E9). When defined by ICD coding, the pulmonary hypertension IRs were lower in both the VA (7.5 vs. 22.0 per 1,000 person-years) and Vanderbilt (5.9 vs. 25.5 per 1,000 person-years) cohorts (see Table E10).
Clinical Characteristic Associations
Within the VA cohort, there was an increased hazard of incident pulmonary hypertension with COPD alone (HR, 1.30 [95% confidence interval (CI), 1.26–1.33]), HF alone (HR, 2.38 [95% CI, 2.31–2.44]), and concurrent HF and COPD (HR, 2.54 [95% CI, 2.46–2.62]) in comparison with individuals without HF or COPD. The hazard for incident pulmonary hypertension associated with HF and COPD increased substantially after time updating these comorbidities (Table 4). Among individuals with HF, HFrEF was associated with greater risk of incident pulmonary hypertension compared with HFpEF (HR, 1.44 [95% CI, 1.39–1.49]). The hazard of pulmonary hypertension increased with age, hypertension, diabetes, and atrial fibrillation. We found higher risk of pulmonary hypertension among men (HR, 1.40 [95% CI, 1.33–1.47]; P < 0.001) and non-Hispanic Black patients (HR, 1.04 [95% CI, 1.01–1.07]; P < 0.001). The covariates associated with incident pulmonary hypertension in the Vanderbilt cohort were similar to those in the VA cohort (Table 4). The predictors of incident pulmonary hypertension were nearly identical between the HF and COPD populations, including the degree of hazard attributable to hypertension, diabetes, and atrial fibrillation (see Table E11). Smoking history and FIB-4 score were excluded from the Vanderbilt predictor models because of a high degree of missingness. Excluding these from the VA models did not change the significance of any covariates. The factors associated with incident pulmonary hypertension were unchanged on sensitivity analysis defining pulmonary hypertension as PASP > 40 mm Hg or defining COPD stringently (see Tables E12 and E13).
Table 4.
Predictors of incident pulmonary hypertension in the VA and Vanderbilt cohorts
| Characteristic | VA Cohort |
Vanderbilt Cohort |
||
|---|---|---|---|---|
| Hazard Ratio (95% CI) |
Hazard Ratio (95% CI) |
|||
| Baseline | Time Updated* | Baseline | Time Updated* | |
| HF−/COPD− | 1 | 1 | 1 | 1 |
| HF−/COPD+ | 1.30 (1.26–1.33) | 1.52 (1.47–1.57) | 1.38 (1.26–1.51) | 1.57 (1.43–1.73) |
| HF+/COPD− | 2.38 (2.31–2.44) | 3.25 (3.14–3.35) | 2.27 (2.15–2.39) | 2.93 (2.77–3.09) |
| HF+/COPD+ | 2.54 (2.46–2.62) | 4.07 (3.95–4.20) | 2.60 (2.40–2.82) | 3.63 (3.37–3.91) |
| Age, yr (per SD, 6.4) | 1.12 (1.10–1.13) | 1.07 (1.06–1.08) | 1.29 (1.26–1.33) | 1.18 (1.15–1.21) |
| Male sex | 1.40 (1.33–1.47) | 1.32 (1.25–1.39) | 1.14 (1.09–1.20) | 1.11 (1.06–1.17) |
| BMI, kg/m2 (per SD, 6.4) | 1.01 (1.00–1.02) | 0.98 (0.97–0.99) | 0.99 (0.94–1.05) | 0.99 (0.94–1.05) |
| Race/ethnicity (vs. non-Hispanic White) | ||||
| Other | 0.95 (0.90–1.01) | 0.95 (0.90–1.00) | 0.76 (0.67–0.88) | 0.73 (0.63–0.85) |
| Non-Hispanic Black | 1.04 (1.01–1.07) | 1.06 (1.03–1.09) | 1.18 (1.10–1.26) | 1.15 (1.08–1.24) |
| Hispanic | 0.92 (0.88–0.97) | 0.95 (0.90–1.00) | 0.82 (0.66–1.02) | 0.82 (0.65–1.03) |
| Hypertension (vs. no hypertension) | ||||
| Controlled hypertension | 2.33 (2.07–2.62) | 2.04 (1.79–2.32) | 1.49 (1.31–1.70) | 1.32 (1.15–1.50) |
| Uncontrolled hypertension | 2.72 (2.41–3.06) | 2.25 (1.98–2.56) | 1.60 (1.38–1.85) | 1.38 (1.20–1.60) |
| Diabetes mellitus | 1.31 (1.28–1.34) | 1.17 (1.14–1.19) | 1.30 (1.23–1.38) | 1.13 (1.07–1.20) |
| Veterans with HIV infection | 1.12 (1.03–1.23) | 1.04 (0.94–1.14) | 0.98 (0.82–1.17) | 0.91 (0.76–1.10) |
| Hepatitis C infection | 1.24 (1.20–1.28) | 1.21 (1.17–1.25) | 1.36 (1.19–1.55) | 1.30 (1.13–1.49) |
| Smoking status (vs. never) | ||||
| Current | 1.17 (1.13–1.20) | 1.08 (1.05–1.12) | — | — |
| Former | 1.09 (1.06–1.12) | 1.05 (1.02–1.09) | — | — |
| Liver fibrosis (FIB-4 score >3.25 vs. ≤3.25) | 1.34 (1.28–1.40) | 1.36 (1.30–1.43) | — | — |
| eGFR, ml/min/1.73m2 (per SD, 2.9) | 0.89 (0.88–0.90) | 0.75 (0.73–0.76) | 0.82 (0.80–0.85) | 0.67 (0.65–0.70) |
| Alcohol abuse | 1.03 (1.00–1.05) | 1.01 (0.98–1.04) | 0.97 (0.86–1.10) | 0.95 (0.84–1.08) |
| Scleroderma | 1.18 (1.12–1.25) | 1.15 (1.09–1.22) | 1.25 (1.14–1.38) | 1.23 (1.11–1.36) |
| Cocaine abuse | 1.01 (0.96–1.05) | 0.98 (0.93–1.02) | 1.30 (1.05–1.60) | 1.26 (1.01–1.57) |
| Atrial fibrillation | 1.64 (1.60–1.69) | 1.51 (1.47–1.55) | 1.59 (1.51–1.67) | 1.50 (1.42–1.58) |
| Hgb, mg/dl (per SD, 1.9) | 0.84 (0.83–0.85) | 0.89 (0.88–0.90) | 0.83 (0.81–0.85) | 0.86 (0.83–0.88) |
Definition of abbreviations: BMI = body mass index; CI = confidence interval; COPD = chronic obstructive pulmonary disease; eGFR = estimated glomerular filtration rate; FIB-4 = Fibrosis-4; HF = heart failure; Hgb = hemoglobin; HIV = human immunodeficiency virus; SD = standard deviation; VA = U.S. Department of Veterans Affairs.
HF, COPD, and eGFR and are time updated. Other variables measured solely at baseline.
Mortality Risk
In the VA cohort 16,238 patients died over a median follow up period of 3.4 years (IQR, 1.4–6.4 yr). After incident pulmonary hypertension, patients with concurrent HF and COPD had an approximately 2.5-fold higher mortality rate compared with those with neither HF nor COPD, while mortality rate was higher in patients with HF alone than COPD alone. The increased hazard of mortality attributable to HF, COPD, or both in comparison with patients with neither was proportionally similar in the Vanderbilt and VA cohorts (Figure 1 and see Table E14). Mortality rates based on incident PASP demonstrate a large increase in risk between 35–45 and 45–55 mm Hg, above which rates increase only modestly (see Table E15). This finding was robust regardless of HF and COPD status (see Figure E3) or time between baseline TTE and incident pulmonary hypertension TTE (see Figure E4). When examined continuously, we found that hazard of mortality linearly increased from 35 to 45 mm Hg, after which the hazard plateaued in both the VA and Vanderbilt cohorts (Figures 2 and E5).
Figure 1.
Survival after incident pulmonary hypertension diagnosis by comorbid heart failure (HF)/chronic obstructive pulmonary disease (COPD) status. (A and B) Displayed are the Kaplan-Meier time-to-event curves for survival in both the VA cohort (A) and the Vanderbilt cohort (B), with stratification by HF and COPD status. Concurrent HF and COPD are associated with the highest unadjusted rate of mortality, followed by the presence of HF or COPD alone, with the lowest rate of mortality in individuals without HF or COPD. Results are similar in both cohorts. VA = U.S. Department of Veterans Affairs.
Figure 2.
Mortality hazard by pulmonary arterial systolic pressure (PASP) at incident pulmonary hypertension diagnosis (U.S. Department of Veterans Affairs [VA] cohort). (A–D) We evaluated the relationship between PASP at the time of echocardiographic incident pulmonary hypertension diagnosis and hazard of mortality using restricted cubic splines with up to three knots in the VA cohort. We assessed this relationship in the full cohort (A), patients with heart failure (HF) and no chronic obstructive pulmonary disease (COPD) (B), patients with COPD and no HF (C), and patients with neither COPD nor HF (D). In all groups, the relative hazard of mortality increased linearly as PASP increased from 35 mm Hg up to approximately 45–50 mm Hg, above which the hazard of mortality remained stable. (E) We then stratified the full cohort by gender and found that mortality risk plateaued for men at approximately 45–50 mm Hg but continued to gradually increase for women. F = female; M = male.
Discussion
Our study substantially expands on prior pulmonary hypertension incidence and prevalence studies, using raw diagnostic information from clinical TTE examinations and assessing comorbidity specific rates in two large racially and ethnically diverse cohorts. The major findings are as follows: 1) the IRs of pulmonary hypertension in the TTE referral population are substantially higher than previously appreciated by code-based definitions; 2) both HF and COPD increase the hazard of incident pulmonary hypertension (193–225% and 52–57% increased risk, respectively), with concurrent diagnosis of HF and COPD associated with additive risk (263–307%); 3) the predominant risk factors for incident pulmonary hypertension independent of COPD and HF include age, male sex, Black race, and cardiometabolic comorbidities; and 4) incident pulmonary hypertension is associated with increased hazard of mortality, which increases with comorbid COPD or HF as well as higher PASP to approximately 45 mm Hg. This work provides important new evidence for pulmonary hypertension surveillance among at-risk individuals and identifies hemodynamic thresholds associated with the highest risk in patients with newly diagnosed pulmonary hypertension.
The only prior study to describe pulmonary hypertension IRs across World Symposium on Pulmonary Hypertension groups estimated the pulmonary hypertension incidence to be 19.8–24.1 per 100,000 patient-years in a population-wide regional Canadian healthcare database relying on ICD codes (10). We identified a pulmonary hypertension IR estimate approximately 30-fold higher among patients referred for TTE. In contrast to prior findings, our data suggest that the IR of pulmonary hypertension is stable over time, aligning with epidemiologic evidence that the IRs of HF and COPD are stable or decreasing (33–35). Code-based incident pulmonary hypertension was diagnosed in 18–19% of TTE-diagnosed cases, with IRs one-fourth to one-half as high as TTE-based rates. This is likely due to clinical underrecognition, as previously noted (7). Conversely, approximately 50% of cases identified by ICD coding were not identified by TTE, which may be due to unrecognized cases but more likely to inappropriate coding or suspicion for pulmonary hypertension that is not subsequently confirmed (7). The magnitude of difference in IRs between the studies may be explained by enrichment for cardiopulmonary comorbidities (among others) in the TTE referral population compared with the general population. A TTE referral cohort may have lower generalizability to the overall population but is highly relevant to patients seeking care and their providers. This may have particularly impact in the HF and COPD populations in which the minority of patients undergo RHC (10), a practice consistent with current pulmonary hypertension guidelines when there is little uncertainty of the underlying etiology (8). Unsurprisingly, we found minor differences in incidence rate and adjusted hazards comparing between the VA and Vanderbilt cohorts. The absolute rates of disease vary across populations; our findings must be interpreted within the context of the cohorts studied.
In patients with at least one follow up TTE examination after baseline, we found lower pulmonary hypertension IRs (65.9–69.1 per 1,000 person-years) in comparison with one prior study examining TTE-based pulmonary hypertension IRs (92.3 per 1,000 person-years) (11). Variation between the estimates may be explained by differing definitions of pulmonary hypertension (PASP ≥ 30 mmHg vs. PASP > 35 mm Hg) and the exclusion of patients with left heart disease. Our study reinforces and expands on that of Stewart and colleagues (11) through larger, U.S.-based, racially and ethnically diverse cohorts inclusive of left heart disease and clarifies the contribution of comorbidities to hazard of incident pulmonary hypertension.
Prior TTE-based studies of prevalent pulmonary hypertension within population-based initiatives have noted increased prevalence of pulmonary hypertension among older adults and patients with comorbidities (36–38). We found that HF was the strongest single factor associated with incident pulmonary hypertension, with HFrEF more strongly associated than HFpEF. Within the HF population left ventricular (LV) dysfunction may associate with higher left-sided filling pressures and thereby degree of pulmonary venous congestion (39), but data on LV filling pressures were missing for a large proportion of patients.
COPD was strongly associated with incident pulmonary hypertension, correlating with prior evidence of pulmonary hypertension prevalence in COPD estimated at 40% in RHC referral cohorts, although these may be biased toward more severe cases (40, 41). The relationship between the severity of COPD and the hazard of incident pulmonary hypertension remains an outstanding question. In patients with concurrent HF and COPD, the hazard of incident pulmonary hypertension was additive, suggesting distinct pathophysiologic processes contributing to pulmonary hypertension.
Age, male sex, and Black race were independently associated with increased hazard of incident pulmonary hypertension. Pulmonary arterial pressure increases with age and is hypothesized to occur through increased blood vessel stiffness and impaired LV diastolic function (36, 37, 39, 42). Several studies have suggested that the prevalence of pulmonary hypertension is higher in women than men across World Health Organization groups (10, 43), but findings from the ARIC (Atherosclerosis Risk in Communities) study suggest no difference in PASP change over time between sexes (39), and we found that male sex was associated with increased hazard of incident pulmonary hypertension. On the basis of these discordant findings, the population-level association between sex and pulmonary hypertension incidence remains unclear. We found an increased hazard of incident pulmonary hypertension among patients who self-identified as non-Hispanic Black race compared with White race, aligning with prior analyses in prevalent pulmonary hypertension (44, 45) but conflicting with ARIC findings of no differences in PASP change over time between Black and White participants (39). Overall, the differing findings from our cohorts and the ARIC cohort may be explained by substantial differences in patient characteristics. Cardiometabolic risk factors including atrial fibrillation, hypertension, and diabetes were associated with incident pulmonary hypertension, building on prior associations noted in prevalent pulmonary hypertension cohorts (46, 47).
We found a strong association between incident pulmonary hypertension and increased hazard of all-cause mortality, building on prior studies (37, 38, 43). We provide new granularity by modeling PASP continuously, rather than categorizing into ranges. We found that mortality hazard increases linearly with pulmonary pressure until ∼45 mm Hg, above which the hazard plateaus. Furthermore, we found that the relationship between PASP and mortality above 50 mm Hg may vary by gender, with a more significant plateau in mortality risk among men than women. This may reflect gender differences in right ventricular compensation, differing etiologies of pulmonary hypertension between the genders, or other unaccounted for coassociating factors. These findings are consistent with those of prior studies in prevalent pulmonary hypertension, with ∼45–50 mm Hg possibly representing a threshold for right ventricular dysfunction (6, 23). We contend that incident pulmonary hypertension should be recognized as a powerful prognostic indicator, especially when PASP is >45 mm Hg. Pulmonary hypertension is likely associated with mortality directly through right ventricular dysfunction and indirectly as a marker of comorbid disease severity.
Strengths and Limitations
The cohorts examined in this study are the largest to date describing risk factors and risk of mortality in patients with incident pulmonary hypertension. The use of TTE to identify pulmonary hypertension is more clinically practical than other diagnostic modalities and is reasonably accurate. Our findings were consistent in two distinct cohorts that were similar in their demographics and prevalence of comorbidities with racial and ethnic representation reflective of the U.S. population. We performed multiple sensitivity analyses to bolster the validity of our findings, accounting for time-updated comorbid diagnoses and the effect of time between TTE examinations on risk of mortality, reporting pulmonary hypertension IR over a 20-year period, defining pulmonary hypertension with multiple cutoffs, and restricting the cohort to individuals with a second TTE.
This is an observational study, and our findings do not imply causality but rather highlight the importance of future mechanistic work. We defined COPD and HF by ICD codes, which may vary in accuracy by dataset and are not the gold standard for diagnosis. Furthermore, we were unable include echocardiographic diastolic dysfunction in our definition of HFpEF, as the recommendations for assessment of diastolic function varied across the period of our study, with poor concordance among the multiple definitions (48). These definitions may introduce misclassification bias in our findings. Referral cohorts are inherently different than the general population, so all results must be interpreted within the context of the populations studied. Although we adjusted for a wide range of confounders, we cannot rule out the possibility of residual confounding from variables that could not be measured. Reflective of a real-world clinical population, the time between studies was not standardized, and it is possible that pulmonary hypertension was systematically identified late. Indication for TTE also varied and could not be ascertained, which also may introduce bias. Loss to follow up could not be estimated, so we assumed patients without a second TTE study to have not developed pulmonary hypertension, which provides a conservative estimate. In contrast, our sensitivity analysis requiring at least one follow up TTE examination after baseline likely overestimates IRs compared with the overall referral population. Outside of protocolized enrollment, the VA as a closed healthcare system offers the best opportunity to capture all follow up compared with other U.S. healthcare systems. TTE is not the gold standard for diagnosing pulmonary hypertension and cannot differentiate among World Health Organization groups of pulmonary hypertension (49). Guidelines recommend the incorporation of indirect TTE findings to assess the risk of pulmonary hypertension, but these data were not consistently collected over the full period of the study in our cohorts. Although this may affect individual risk classification, PASP remains an effective surrogate for pulmonary hypertension on the basis of mortality risk in a large population (6). The minimal detectable difference for PASP is not well established and has been examined only in a small cohort of patients with systemic sclerosis (50), which may limit the interpretability of our findings for individual patients with borderline PASP values, but these patients represent a small proportion of incident pulmonary hypertension cases. Our VA cohort age range is limited to patients born between 1945 and 1965 who underwent TTE between 1999 and 2020 (minimum age 34 yr, maximum age 75 yr). This may introduce bias by excluding younger patients with World Symposium on Pulmonary Hypertension group 1 pulmonary hypertension and older patients at higher risk of incident pulmonary hypertension. This limitation was not present in our Vanderbilt cohort, which yielded similar findings.
Conclusions
Incident pulmonary hypertension is markedly more common among the TTE referral population than previously recognized. Comorbid HF and COPD are strongly associated with incident pulmonary hypertension, together with demographic factors and cardiometabolic comorbidities. Individuals with incident pulmonary hypertension are at increased risk of mortality, and new pulmonary hypertension on TTE should be recognized as a significant prognostic sign. There is a substantial population of individuals at risk for pulmonary hypertension in need of preventive strategies and therapeutics.
Supplemental Materials
Acknowledgments
Acknowledgment
This work uses data provided by patients and collected by the VA as part of their care and support.
Footnotes
Supported by National Heart, Lung, and Blood Institute grants T32 HL 087738 (J.D.G.), R01 HL 163960 (E.L.B.), R01 HL 146588 (E.L.B., M.F.), and R01 HL 155278 (E.L.B.); National Institute of Diabetes and Digestive and Kidney Diseases grant R01 DK 124845 (E.L.B.); and U.S. Food and Drug Administration grant R01 FD 007627 (E.L.B.). The Veterans Affairs Cohort Study is supported by U.S. Department of Veterans Affairs grant P01 AA 029545 and grant U24 AA 020794. The views and opinions expressed in this paper are those of the authors and do not necessarily represent those of the Department of Veterans Affairs or the U.S. government.
Author Contributions: J.D.G., S.K., S.E., M.F., and E.L.B. contributed to study conception and design. J.A. and C.A. performed data acquisition. S.K. performed data analysis. J.D.G. drafted the manuscript. All authors participated in data interpretation, critically reviewed the manuscript, and approved the final manuscript submitted for publication.
A data supplement for this article is available via the Supplements tab at the top of the online article.
Author disclosures are available with the text of this article at www.atsjournals.org.
References
- 1. Hoeper MM, Humbert M, Souza R, Idrees M, Kawut SM, Sliwa-Hahnle K, et al. A global view of pulmonary hypertension. Lancet Respir Med . 2016;4:306–322. doi: 10.1016/S2213-2600(15)00543-3. [DOI] [PubMed] [Google Scholar]
- 2. Strange G, Stewart S, Celermajer DS, Prior D, Scalia GM, Marwick TH, et al. NEDA Contributing Sites Threshold of pulmonary hypertension associated with increased mortality. J Am Coll Cardiol . 2019;73:2660–2672. doi: 10.1016/j.jacc.2019.03.482. [DOI] [PubMed] [Google Scholar]
- 3. Rosenkranz S, Gibbs JSR, Wachter R, De Marco T, Vonk-Noordegraaf A, Vachiéry JL. Left ventricular heart failure and pulmonary hypertension. Eur Heart J . 2016;37:942–954. doi: 10.1093/eurheartj/ehv512. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 4. Seeger W, Adir Y, Barberà JA, Champion H, Coghlan JG, Cottin V, et al. Pulmonary hypertension in chronic lung diseases. J Am Coll Cardiol . 2013;62:D109–D116. doi: 10.1016/j.jacc.2013.10.036. [DOI] [PubMed] [Google Scholar]
- 5. Nathan SD, Barbera JA, Gaine SP, Harari S, Martinez FJ, Olschewski H, et al. Pulmonary hypertension in chronic lung disease and hypoxia. Eur Respir J . 2019;53:1801914. doi: 10.1183/13993003.01914-2018. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 6. Jankowich M, Maron BA, Choudhary G. Mildly elevated pulmonary artery systolic pressure on echocardiography: bridging the gap in current guidelines. Lancet Respir Med . 2021;9:1185–1191. doi: 10.1016/S2213-2600(21)00072-2. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 7. Maron BA, Choudhary G, Khan UA, Jankowich MD, McChesney H, Ferrazzani SJ, et al. Clinical profile and underdiagnosis of pulmonary hypertension in US veteran patients. Circ Heart Fail . 2013;6:906–912. doi: 10.1161/CIRCHEARTFAILURE.112.000091. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 8. Humbert M, Kovacs G, Hoeper MM, Badagliacca R, Berger RMF, Brida M, et al. ESC/ERS Scientific Document Group 2022 ESC/ERS guidelines for the diagnosis and treatment of pulmonary hypertension. Eur Respir J . 2023;61:2200879. doi: 10.1183/13993003.00879-2022. [DOI] [PubMed] [Google Scholar]
- 9. Huston JH, Maron BA, French J, Huang S, Thayer T, Farber-Eger EH, et al. Association of mild echocardiographic pulmonary hypertension with mortality and right ventricular function. JAMA Cardiol . 2019;4:1112–1121. doi: 10.1001/jamacardio.2019.3345. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 10. Thiwanka Wijeratne D, Lajkosz K, Brogly SB, Diane Lougheed M, Jiang L, Housin A, et al. Increasing incidence and prevalence of World Health Organization groups 1 to 4 pulmonary hypertension: a population-based cohort study in Ontario, Canada. Circ Cardiovasc Qual Outcomes . 2018;11:e003973. doi: 10.1161/CIRCOUTCOMES.117.003973. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 11. Stewart S, Chan Y-K, Playford D, Harris S, Strange GA. Incident pulmonary hypertension in 13 488 cases investigated with repeat echocardiography: a clinical cohort study. ERJ Open Res . 2023;9:00082-2023. doi: 10.1183/23120541.00082-2023. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 12. Link J, Glazer C, Torres F, Chin K. International Classification of Diseases coding changes lead to profound declines in reported idiopathic pulmonary arterial hypertension mortality and hospitalizations: implications for database studies. Chest . 2011;139:497–504. doi: 10.1378/chest.10-0837. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 13. Mathai SC, Mathew S. Breathing (and coding?) a bit easier: changes to International Classification of Disease coding for pulmonary hypertension. Chest . 2018;154:207–218. doi: 10.1016/j.chest.2018.04.004. [DOI] [PubMed] [Google Scholar]
- 14. Sarkar S, Esserman DA, Skanderson M, Levin FL, Justice AC, Lim JK. Disparities in hepatitis C testing in US veterans born 1945–1965. J Hepatol . 2016;65:259–265. doi: 10.1016/j.jhep.2016.04.012. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 15. To T, Zhu J, Gray N, Feldman LY, Villeneuve PJ, Licskai C, et al. Asthma and chronic obstructive pulmonary disease overlap in women: incidence and risk factors. Ann Am Thorac Soc . 2018;15:1304–1310. doi: 10.1513/AnnalsATS.201802-078OC. [DOI] [PubMed] [Google Scholar]
- 16. Gershon AS, Wang C, Guan J, Vasilevska-Ristovska J, Cicutto L, To T. Identifying individuals with physician diagnosed COPD in health administrative databases. COPD . 2009;6:388–394. doi: 10.1080/15412550903140865. [DOI] [PubMed] [Google Scholar]
- 17. Crothers K, Rodriguez CV, Nance RM, Akgun K, Shahrir S, Kim J, et al. Accuracy of electronic health record data for the diagnosis of chronic obstructive pulmonary disease in persons living with HIV and uninfected persons. Pharmacoepidemiol Drug Saf . 2019;28:140–147. doi: 10.1002/pds.4567. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 18. McGinnis KA, Brandt CA, Skanderson M, Justice AC, Shahrir S, Butt AA, et al. Validating smoking data from the Veteran’s Affairs Health Factors dataset, an electronic data source. Nicotine Tob Res . 2011;13:1233–1239. doi: 10.1093/ntr/ntr206. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 19. Butt AA, McGinnis K, Rodriguez-Barradas MC, Crystal S, Simberkoff M, Goetz MB, et al. Veterans Aging Cohort Study HIV infection and the risk of diabetes mellitus. AIDS . 2009;23:1227–1234. doi: 10.1097/QAD.0b013e32832bd7af. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 20. McGinnis KA, Justice AC, Bailin S, Wellons M, Freiberg M, Koethe JR. High concordance between chart review adjudication and electronic medical record data to identify prevalent and incident diabetes mellitus among persons with and without HIV. Pharmacoepidemiol Drug Saf . 2020;29:1432–1439. doi: 10.1002/pds.5111. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 21. Chobanian AV, Bakris GL, Black HR, Cushman WC, Green LA, Izzo JL, et al. National High Blood Pressure Education Program Coordinating Committee Seventh report of the Joint National Committee on Prevention, Detection, Evaluation, and Treatment of High Blood Pressure. Hypertension . 2003;42:1206–1252. doi: 10.1161/01.HYP.0000107251.49515.c2. [DOI] [PubMed] [Google Scholar]
- 22. Fultz SL, Skanderson M, Mole LA, Gandhi N, Bryant K, Crystal S, et al. Development and verification of a “virtual” cohort using the National VA Health Information System. Med Care . 2006;44:S25–S30. doi: 10.1097/01.mlr.0000223670.00890.74. [DOI] [PubMed] [Google Scholar]
- 23. Maron BA, Hess E, Maddox TM, Opotowsky AR, Tedford RJ, Lahm T, et al. Association of borderline pulmonary hypertension with mortality and hospitalization in a large patient cohort: insights from the Veterans Affairs Clinical Assessment, Reporting, and Tracking program. Circulation . 2016;133:1240–1248. doi: 10.1161/CIRCULATIONAHA.115.020207. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 24. Crothers K, Huang L, Goulet JL, Goetz MB, Brown ST, Rodriguez-Barradas MC, et al. HIV infection and risk for incident pulmonary diseases in the combination antiretroviral therapy era. Am J Respir Crit Care Med . 2011;183:388–395. doi: 10.1164/rccm.201006-0836OC. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 25. Patterson OV, Freiberg MS, Skanderson M, Fodeh JS, Brandt CA, DuVall SL. Unlocking echocardiogram measurements for heart disease research through natural language processing. BMC Cardiovasc Disord . 2017;17:151. doi: 10.1186/s12872-017-0580-8. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 26. Wells QS, Farber-Eger E, Crawford DC. Extraction of echocardiographic data from the electronic medical record is a rapid and efficient method for study of cardiac structure and function. J Clin Bioinforma . 2014;4:12. doi: 10.1186/2043-9113-4-12. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 27. Lam CSP, Roger VL, Rodeheffer RJ, Borlaug BA, Enders FT, Redfield MM. Pulmonary hypertension in heart failure with preserved ejection fraction: a community-based study. J Am Coll Cardiol . 2009;53:1119–1126. doi: 10.1016/j.jacc.2008.11.051. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 28. Choudhary G, Jankowich M, Wu WC. Elevated pulmonary artery systolic pressure predicts heart failure admissions in African Americans Jackson Heart Study. Circ Heart Fail . 2014;7:558–564. doi: 10.1161/CIRCHEARTFAILURE.114.001366. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 29. Heidenreich PA, Bozkurt B, Aguilar D, Allen LA, Byun JJ, Colvin MM, et al. 2022 AHA/ACC/HFSA guideline for the management of heart failure: a report of the American College of Cardiology/American Heart Association Joint Committee on Clinical Practice Guidelines. Circulation . 2022;145:e895–e1032. doi: 10.1161/CIR.0000000000001063. [DOI] [PubMed] [Google Scholar]
- 30. O’Leary JM, Assad TR, Xu M, Farber-Eger E, Wells QS, Hemnes AR, et al. Lack of a tricuspid regurgitation Doppler signal and pulmonary hypertension by invasive measurement. J Am Heart Assoc . 2018;7:e009362. doi: 10.1161/JAHA.118.009362. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 31. van Buuren S, Groothuis-Oudshoorn K. mice: multivariate imputation by chained equations in R. J Stat Soft . 2011;45:1–67. [Google Scholar]
- 32.Little RJA, Rubin DB. Statistical analysis with missing data. 2nd ed. Hoboken, NJ: John Wiley; 2002. [Google Scholar]
- 33. Khera R, Kondamudi N, Zhong L, Vaduganathan M, Parker J, Das SR, et al. Temporal trends in heart failure incidence among Medicare beneficiaries across risk factor strata, 2011 to 2016. JAMA Netw Open . 2020;3:e2022190. doi: 10.1001/jamanetworkopen.2020.22190. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 34. Tsao CW, Lyass A, Enserro D, Larson MG, Ho JE, Kizer JR, et al. Temporal trends in the incidence of and mortality associated with heart failure with preserved and reduced ejection fraction. JACC Heart Fail . 2018;6:678–685. doi: 10.1016/j.jchf.2018.03.006. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 35. Ford ES, Croft JB, Mannino DM, Wheaton AG, Zhang X, Giles WH. COPD surveillance—United States, 1999–2011. Chest . 2013;144:284–305. doi: 10.1378/chest.13-0809. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 36. D’Andrea A, Naeije R, Grünig E, Caso P, D’Alto M, Di Palma E, et al. Echocardiography of the pulmonary circulation and right ventricular function: exploring the physiologic spectrum in 1,480 normal subjects. Chest . 2014;145:1071–1078. doi: 10.1378/chest.12-3079. [DOI] [PubMed] [Google Scholar]
- 37. Lam CSP, Borlaug BA, Kane GC, Enders FT, Rodeheffer RJ, Redfield MM. Age-associated increases in pulmonary artery systolic pressure in the general population. Circulation . 2009;119:2663–2670. doi: 10.1161/CIRCULATIONAHA.108.838698. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 38. Strange G, Playford D, Stewart S, Deague JA, Nelson H, Kent A, et al. Pulmonary hypertension: prevalence and mortality in the Armadale echocardiography cohort. Heart . 2012;98:1805–1811. doi: 10.1136/heartjnl-2012-301992. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 39. Zierath R, Claggett B, Arthur V, Yang Y, Skali H, Matsushita K, et al. Changes in pulmonary artery pressure late in life. J Am Coll Cardiol . 2023;82:2179–2192. doi: 10.1016/j.jacc.2023.09.821. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 40. Zhang L, Liu Y, Zhao S, Wang Z, Zhang M, Zhang S, et al. The incidence and prevalence of pulmonary hypertension in the COPD population: a systematic review and meta-analysis. Int J Chron Obstruct Pulmon Dis . 2022;17:1365–1379. doi: 10.2147/COPD.S359873. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 41. Cook DP, Xu M, Martucci VL, Annis JS, Aldrich MC, Hemnes AR, et al. Clinical insights into pulmonary hypertension in chronic obstructive pulmonary disease. Pulm Circ . 2022;12:e12006. doi: 10.1002/pul2.12006. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 42. Fayngersh V, Drakopanagiotakis F, McCool FD, Klinger JR. Pulmonary hypertension in a stable community-based COPD population. Lung . 2011;189:377–382. doi: 10.1007/s00408-011-9315-2. [DOI] [PubMed] [Google Scholar]
- 43. Choudhary G, Jankowich M, Wu WC. Prevalence and clinical characteristics associated with pulmonary hypertension in African-Americans. PLoS One . 2013;8:e84264. doi: 10.1371/journal.pone.0084264. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 44. Yang BQ, Assad TR, O’Leary JM, Xu M, Halliday SJ, D’Amico RW, et al. Racial differences in patients referred for right heart catheterization and risk of pulmonary hypertension. Pulm Circ . 2018;8:2045894018764273. doi: 10.1177/2045894018764273. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 45. Brittain EL, Nwabuo C, Xu M, Gupta DK, Hemnes AR, Moreira HT, et al. Echocardiographic pulmonary artery systolic pressure in the Coronary Artery Risk Development in Young Adults (CARDIA) study: associations with race and metabolic dysregulation. J Am Heart Assoc . 2017;6:e005111. doi: 10.1161/JAHA.116.005111. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 46. Pugh ME, Robbins IM, Rice TW, West J, Newman JH, Hemnes AR. Unrecognized glucose intolerance is common in pulmonary arterial hypertension. J Heart Lung Transplant . 2011;30:904–911. doi: 10.1016/j.healun.2011.02.016. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 47. Robbins IM, Newman JH, Johnson RF, Hemnes AR, Fremont RD, Piana RN, et al. Association of the metabolic syndrome with pulmonary venous hypertension. Chest . 2009;136:31–36. doi: 10.1378/chest.08-2008. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 48. Rasmussen-Torvik LJ, Colangelo LA, Lima JAC, Jacobs DR, Rodriguez CJ, Gidding SS, et al. Prevalence and predictors of diastolic dysfunction according to different classification criteria: the Coronary Artery Risk Development in Young in Adults study. Am J Epidemiol . 2017;185:1221–1227. doi: 10.1093/aje/kww214. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 49. D’Alto M, Bossone E, Opotowsky AR, Ghio S, Rudski LG, Naeije R. Strengths and weaknesses of echocardiography for the diagnosis of pulmonary hypertension. Int J Cardiol . 2018;263:177–183. doi: 10.1016/j.ijcard.2018.04.024. [DOI] [PubMed] [Google Scholar]
- 50. Mukherjee M, Mercurio V, Balasubramanian A, Shah AA, Hsu S, Simpson CE, et al. Defining minimal detectable difference in echocardiographic measures of right ventricular function in systemic sclerosis. Arthritis Res Ther . 2022;24:146. doi: 10.1186/s13075-022-02835-5. [DOI] [PMC free article] [PubMed] [Google Scholar]
Associated Data
This section collects any data citations, data availability statements, or supplementary materials included in this article.