When Incidence Leads to Precedence: A Call for Early Detection Protocols for Pulmonary Hypertension in Heart Failure and Chronic Obstructive Pulmonary Disease

Recent epidemiologic data suggest that heart failure (HF) and chronic obstructive pulmonary disease (COPD) affect 6.7 million (2.7%) and 14.1 million (4.2%) persons in the United States, respectively, leading to significant morbidity and mortality (1, 2). Pulmonary hypertension (PH), a chronic, progressive disease of the pulmonary vasculature that leads to right HF and death, commonly complicates HF and COPD. Extrapolating from cohort studies, estimates of the prevalence of PH in the setting of HF in the United States range from 2.7 to 5.4 million, and estimates of PH in COPD in the United States range from 2.8 to 11 million depending on the definition of PH employed. The clinical relevance of PH in either disease state is significant. Beyond the impact on symptoms and morbidity, PH in the context of HF is associated with a 1.5- to 2.2-fold increased risk of mortality compared with HF without PH, and PH in COPD is associated with a 1.7- to 3.7-fold increased risk of death (3–6). Further, the presence of PH in either disease state has impact on transplant eligibility and outcomes, as a pulmonary vascular resistance greater than 5 Wood units is a relative contraindication for heart transplantation and lung transplant waitlist, and post-transplant survival for patients with PH-COPD is worse than for COPD without PH (7–9).

Despite the high prevalence of PH and its impact on morbidity and mortality, there are no specific guideline recommendations for screening for PH in HF or COPD (10, 11). There are several possible reasons why no specific recommendations exist, including a low prevalence of severe PH in HF and COPD, the need for invasive right heart catheterization to confirm the diagnosis, and a lack of data supporting medical therapy directed at PH in either disease. In addition, uncertainty about the impact of mild PH on symptoms and outcomes and assumptions that PH is a complication of severe HF or COPD may lead clinicians to defer evaluation for PH.

In this issue of the Journal, Garry and colleagues (pp. 679–688) provide strong evidence to reconsider this approach (or lack thereof) to the evaluation of PH in HF and COPD (12). Using two large retrospective cohorts, one from the Veterans Health Administration (VA; n = 245,067 patients) and one from the Vanderbilt University Medical Center (VUMC; n = 117,526 patients), the investigators examine the incidence of PH in individuals referred for transthoracic echocardiography (TTE), focusing on those who had prior diagnoses of HF and COPD. Using an echo-based diagnosis of PH (right ventricular systolic pressure [RVSP] >35 mm Hg), the investigators identified incident PH in 15.9% of individuals in the VA cohort and 10.4% of individuals in the VUMC cohort. The overall incidences of PH were calculated to be 22.0 and 25.5 per 1,000 person-years (PY) in the VA and VUMC cohorts, respectively. The incidence was substantially higher for HF (49.8–56.4 per 1,000 PY) than COPD (22.6–31.4 per 1,000 PY), and even more so for those with HF and COPD (60.3–79.8 per 1,000 PY). The incidence did differ between HF phenotypes, with patients with HF with reduced ejection fraction having higher rates than those with HF with preserved ejection fraction (range, 62.7–78.2 vs. 43.9–51.0 per 1,000 PY; hazard ratio, 1.44; 95% confidence interval, 1.39–1.49). Only 19% of the VA cohort and 18% of the VUMC cohort carried a diagnosis of PH at the time of TTE; in fact, the associated incidence based on International Classification of Diseases (ICD) coding was one third to one fourth lower than the echo-based incidence. The authors also examined the association between clinical characteristics and incident PH and found similar predictors, including hypertension, diabetes, and atrial fibrillation, between the HF and COPD cohorts. Importantly, mortality risk was increased with even mildly increased RVSP and higher when comparing patients with an RVSP of 45–55 mm Hg versus those with an RVSP of 35–45 mm Hg.

Taken together, these findings offer important insights into the epidemiology of PH in HF and COPD. The incidence reported is significantly (nearly 30-fold) higher than in the only prior study to define the incidence of PH across groups; notably, this study relied on ICD codes for the diagnosis of PH (13). Although the populations included in the earlier study and the new study differ, the magnitude of difference between the incidence of PH in the general population compared with the HF and COPD cohorts emphasizes the importance of the findings of Garry and colleagues. Further, the association of incident PH with mortality and the linear association between increasing RVSP and mortality risk highlight the impact of even mild PH on survival in HF and COPD, as previously shown in a general referral population undergoing TTE (14). There are interesting observations that require further investigation, such as increased risk of incident PH among male and Black individuals compared with women and White individuals, as these findings differ from those of other epidemiologic studies (15). Most importantly, these findings demonstrate the need for routine early detection protocols in at-risk populations such as those with HF and COPD. Although echocardiography is commonly performed as part of the evaluation of HF, current guidelines do not recommend routine screening for PH as an evolving comorbidity (10). Similarly, the most recent Global Initiative for Chronic Obstructive Lung Disease report does not recommend any routine evaluation for PH, noting that PH occurs in advanced COPD and that only severe PH is associated with poor survival (11). The high incidence rate of PH and its impact on survival demonstrated in this study warrant reevaluation of these recommendations.

There are several limitations to the study of Garry and colleagues. First, the definitions of HF and COPD, the exposures of interest, were based on ICD codes, which may have limited accuracy. Further, because there were no set intervals for repeating TTE in the cohort, ascertainment bias may impact the observed incidence rate and associations. Pulmonary function test results were not reported because these were not routinely available in the VA cohort; therefore, we do not have insight into the relationship between COPD severity and incident PH or the potential relationships between incident PH and other pulmonary function markers of pulmonary vascular disease such as diffusing capacity for carbon monoxide. The authors do not incorporate additional TTE-based measures, such as right atrial area and tricuspid annular plane systolic excursion to pulmonary artery systolic pressure ratio, as recommended by current PH guidelines to inform an estimated probability of PH (16). Inclusion of these measurements may improve the accuracy of TTE-based diagnosis of PH. In addition, diagnosis of PH by TTE has significant limitations, including variable accuracy of RVSP estimates compared with invasively measured pulmonary artery systolic pressure, particularly in patients with underlying lung disease, and an inability to accurately identify hemodynamic phenotypes of PH. Given the known differences in outcomes based on the hemodynamic phenotype, i.e., precapillary, isolated postcapillary, and combined pre-/postcapillary PH in HF and COPD, future studies should include invasive assessment of patients with TTE evidence of PH to better characterize their PH to inform subsequent risk and treatment options.

Despite these limitations, the work by Garry and colleagues offer important epidemiologic evidence of: 1) a high incidence of PH in at-risk populations of individuals with HF and COPD and 2) a strong association between PH and survival in these populations. Future studies should focus on the development and implementation of standardized early detection protocols for PH in HF and COPD with a goal of reducing morbidity and mortality in these complex cardiopulmonary diseases.