Deconstructing the Melting Pot in Pulmonary Arterial Hypertension

Pulmonary arterial hypertension (PAH) is a disease that is typically fatal within 5–7 years of diagnosis for most subjects and occurs in all ancestral populations (1). Since the initial discovery of BMPR2 (bone morphogenetic protein receptor type 2) mutation as a cause of PAH, countless publications have further expanded the genetics of PAH, including discoveries of other causative genes and the role of common gene variant associations (2–4). Yet, few studies have comprehensively explored how ancestry, race, or ethnicity plays a role in PAH development and response to therapy. The lack of such studies is striking given the intense focus on providing personalized care to patients with PAH. Of course, studies of discrete populations, such as minority groups, are challenging to perform in rare diseases given the small numbers of subjects.

Nearly a decade ago, Gabler and colleagues conducted a pooled analysis of data from placebo-controlled trials of the use of endothelin receptor antagonists in >1,000 participants with PAH, and uncovered variations in response to endothelin receptor antagonists related to sex and self-reported race (5). Race-based comparisons focused on black versus white individuals showed a difference in placebo-adjusted beneficial treatment response, favoring white individuals by considerable effect sizes. However, this difference did not meet statistical significance. Although other racial groups were not explored, the study was an important reminder that variations in treatment response may occur among individuals of different racial and ethnic groups. A few subsequent studies reported the impact of self-reported African ancestry; overall, there appears to be a higher degree of severity and perhaps a reduced treatment response among those who self-report as black (6–10).

The self-reported data suggesting that black individuals have poorer outcomes are troubling and raise more questions than answers. More broadly, this also highlights our inadequate understanding, even to this date, of the role of racial, ethnic, and ancestral factors in PAH and pulmonary hypertension. As a reminder, in biomedical research, the term “race” refers to an individual’s appearance. In contrast, “ethnicity” suggests a communal participation in a group (or groups) of individuals who share specific cultural traits, including traditions, language, social practices, and even geopolitical factors (11). Unfortunately, these categorizations typically rely on an individual participant’s self-report or assignment by investigators, both of which are highly prone to error, especially when compared with DNA-based determination (12). Errors may arise from an individual’s lack of knowledge about his or her actual background and alignment with an ethnic group that has a genetically admixed background (11). As a result, recent studies found that self-reported race is likely to be incorrect and/or have a higher degree of admixture than would be assumed (13, 14). Self-identified Hispanics in New York City, for example, were shown to be 29% European, 26% African, and 45% Native American by ancestry-informative markers (15).

In this issue of the Journal, Karnes and colleagues (pp. 1407–1415) explore the role of race and ethnicity in PAH (16). They determined ancestry in patients with PAH using genetic markers by incorporating information contained in several large U.S. datasets. Self-reported race and ethnicity were combined to form selected groups of non-Hispanic white (NHW), non-Hispanic African American (NHAA), and Hispanic subjects. Reasonably consistent with prior studies, NHWs were 97% European, NHAAs were 82% African (16% European), and Hispanics were 85% European, 36% Native American, and 7% African by genetically determined ancestry. Although the groups were similar by most standard PAH-relevant comparisons, Hispanics were younger, with a higher mean pulmonary artery pressure and pulmonary vascular resistance but lower concurrent use of prostacyclin analogs. Intriguingly, in survival analyses, self-reported Hispanic status in both the PAH Biobank and Allegheny Health Network cohorts was associated with a statistically significantly improved transplant-free survival. It does not appear that survival was modified by genetic ancestry according to the a priori level of statistical significance.

To strengthen their study, the authors evaluated a distinct database of patients with PAH in self-reported or hospital-assigned categories. This analysis included data for 8,829 NHWs, 2,628 NHAAs, 1,524 Hispanics, 403 Asians, and 185 Native Americans from the U.S. National Inpatient Sample database. Regression analyses of these data showed that both Hispanic status and Native American status were protective of inpatient mortality versus NHW status. In contrast, NHAA status conferred increased mortality compared with NHW status, consistent with prior concerns about poorer outcomes for African Americans with PAH (5–9).

This study has multiple strengths, including the number of subjects studied, the integration of self-report and genetic identification, and the additional value of an extensive inpatient database. Given the importance of distinguishing race, ethnicity, and ancestry, the ability to assess ancestry-informative markers added a layer of depth and supported the authors’ conclusion that Native American ancestry may contribute to Hispanics’ protection against deleterious outcomes. The large sample size allowed the authors to examine minority groups that are typically excluded or insufficiently studied in large registries of PAH.

However, this work identifies several possible future lines of investigation that could further clarify these data. The inclusion of several consented PAH cohorts increased the sample size, which was needed, but also introduced phenotypic heterogeneity. For example, the PAH Biobank only included idiopathic and heritable PAH, whereas the Allegheny Health Network, Arizona, and Stanford cohorts included all forms of PAH. The years in which the studies were conducted, the duration of follow-up, the relationship between diagnosis and enrollment, and other factors varied among the cohorts. Furthermore, researchers establishing cohorts derived from tertiary clinical care centers may inadvertently select a certain type of subject (for example, those who are particularly ill), excluding subjects from a given category. It is hoped that a large cohort study that can provide more phenotypic homogeneity, as well as a more in-depth degree of functional (e.g., imaging studies of right ventricular function and changes over time) and multi-“omic” assessments, will emerge to explore race, ethnicity, and ancestry in more detail. Also, although the National Inpatient Sample database is a useful addition to such research, it has significant limitations, including a lack of biospecimens, reliance on accurate identification of PAH from International Classification of Diseases, Ninth Revision, Clinical Modification and International Classification of Diseases, Tenth Revision, Clinical Modification data, and lack of outpatient data. Finally, identification of differences according to race/ethnicity may reflect lifestyle, options for subspecialty care, or other choices that are irrelevant to genetic susceptibility (17).

Regardless, the current study highlights the need to carefully assess differences between and among groups stratified by self-reporting as well as according to genetic markers in PAH. If the findings are replicated, the next steps would involve careful determinations of why differences in clinical outcomes exist and how they may be corrected. The solutions may be complicated and vary depending on the cause—ethnic and ancestral differences have very different underlying causes and thus would require different approaches for alteration. However, understanding subpopulation and, ultimately, individual genetic compositions will be crucial for the next generation of cohort studies, clinical trials, and therapies for PAH.