Sex Hormones across the Menstrual Cycle in Pulmonary Arterial Hypertension: Adding a New Layer of Complexity

One of our most recent diagnoses is a 9-year-old male with heritable pulmonary arterial hypertension (PAH) owing to a known rare variant (mutation) in a PAH-related gene. I ached for his family during a call last night, as we discussed a new result—his currently healthy younger sister shares the same pathogenic genetic variant. Their parents described it as a “gut punch,” amplified because they know that females have a 42% lifetime risk of developing PAH among those with a genetic risk (1). Males develop PAH, no doubt. However, the strongest nongenetic risk factor for PAH, a progressively fatal pulmonary vascular disease, is female sex (2–4). Although the female predominance dates back at least to the initial reports of “primary pulmonary hypertension” by Dresdale and colleagues in the 1950s, we still do not know why incident case numbers skew toward female individuals over males (5).

But our field is making progress (please see the recent comprehensive review on this topic by Hester and colleagues [6]). Female enrichment suggests sex hormone effects, sex chromosome influences, or both contribute to PAH pathogenesis. The numerous human studies over the past decade exploring the association between PAH and variations in circulating sex hormones, with a particular focus on parent compound estrogens (estradiol [E2] and estrone [E1]), generally support the finding that individuals with PAH have higher estrogen levels. This is the case for both females and males with PAH compared with healthy subjects. Recent detailed efforts to explore not only E2 and E1 levels but also estrogen metabolites suggest that a global burden of active estrogens and associated downstream signaling are likely important.

In addition, a focus only on estrogens is likely too narrow. Recent work supports the concept that androgens are relevant to the PAH phenotype as well. In fact, androgens may contribute to, or perhaps antagonize, the PAH phenotype, with lower levels associated with not only disease but also functional outcomes (7, 8). The association of androgens appears applicable to women as well as men. Although less comprehensively studied, progesterone may also be relevant, with low progesterone levels recently associated with PAH and reduced survival among premenopausal adult females with idiopathic pulmonary arterial hypertension (IPAH); in contrast, high follicle-stimulating hormone levels were associated with IPAH status and reduced survival in that cohort (9).

However, a major challenge of studying sex hormones, particularly among women of reproductive age, is the potential variation in sex hormone levels in a given individual both in the short and long term. For example, in women of reproductive age (but not postmenopausal women), endogenous levels of estrogens fluctuate in response to the menstrual cycle. The menstrual cycle, in fact, is a crucial variable to consider: female sex hormone production and activity varies tremendously across the four phases of a typical 28-day cycle (menstrual, follicular [proliferative], ovulation, and luteal phases). Interestingly, however, estrogen levels are fairly constant for an individual woman of reproductive age when drawn at the same phase in the menstrual cycle in a given individual over 2–3 years; but, they do alter over the lifespan and with changes in body constitution (10). With regard to androgens, circulating levels of testosterone and androstenedione vary with a person’s age, body mass index, and (for women) stage of menstrual cycle in a given month (although these androgens do not vary substantially in a given day) (11, 12). For these and other reasons, the study of humans with PAH, particularly females, and other conditions with regard to sex hormones is a challenging endeavor.

In this issue of AnnalsATS, Baird and colleagues (pp. 218–228) provide the first study published investigating the issue of sex hormone variation over the course of a menstrual cycle in women of reproductive age with PAH (13). The authors provide a detailed look at a small cohort of PAH patients and controls, associating variations in sex hormones with phenotypic changes relevant to the PAH condition (13). In this small study of 28 women (8 with PAH; 20 healthy control subjects) over the course of a menstrual cycle, they reproduced many of the previously known data concerning sex hormone levels in premenopausal women with PAH (e.g., higher E2 levels and lower dehydroepiandrosterone-sulfate [DHEA-S] levels among patients with PAH) while also advancing our understanding of the complex relationship between sex hormone levels and the cardiopulmonary condition. Cyclic variation in E2, however, was lower in patients with PAH than in control subjects, such that at one time point, patient and control subject E2 levels (visit 3) actually were not different; this likely corresponded to approximately 2 days postovulation, when E2, luteinizing hormone, and follicle-stimulating hormone (FSH) should be near their lowest cyclic levels (14). Patient E2 levels were associated with indices of cardiopulmonary function, with higher E2 levels correlating with lower 6-minute-walk test distance and higher N-terminal (NT)-pro hormone BNP (NT-proBNP) levels, although there was some variation according to menstrual phase. Although less straightforward, right ventricular function as estimated by tricuspid annular plane systolic excursion tracked with E2 levels at some phases.

Not surprisingly, the androgen DHEA-S levels did not display the same degree of variation across the menstrual cycle phases, with consistent reductions in patients versus control subjects. As with E2, DHEA-S was associated with indices of cardiopulmonary status, including a striking linear relationship between DHEA-S and 6-minute-walk test distance regardless of menstrual phase. In contrast, NT-proBNP levels generally had an inverse relationship with DHEA-S levels, although this was less prominent during visit 3, which presumably occurred just after ovulation. Evaluations of additional hormone levels, such as luteinizing hormone, did not substantially differ between groups or associate with functional metrics. Explorations of microRNA (miRNA) levels relevant to PAH (e.g., miRNA-29 family) demonstrated interesting but complicated results with some modification of levels of sex hormones with variability according to menstrual phase; future studies will be warranted.

Although small, this study by Baird and colleagues is an important advance for our field. Human studies in PAH are difficult in general, overcoming issues such as reduced physical health and vigor, accessibility, willingness to participate in additional studies on top of patients’ large clinical demands, and other challenges. The additional pressure of timed study visits linked to the menstrual cycle at enhanced frequency further adds complexity. This work is an important reminder that no human is static, and for females of reproductive age, the study of sex hormones must take this into account.

But this study is not without limitations, many of which were addressed by the authors. First, circulating levels may not accurately reflect tissue-level activity, although the associations with functional parameters and miRNAs are suggestive. Second, as mentioned, accurate relay of the start of menstruation was crucial to activate study visit 1 and designation of phases, particularly in the absence of metrics to confirm the ovulatory phase; however, the graphic demonstration of E2 levels across study visits (Figure 1 of Baird and colleagues) did generally match the expected pattern of E2 across a standard 28-day menstrual cycle. Third, as a small study with several PAH subtypes, subtype differences, variations in therapeutic exposure, and the influence of time since diagnosis could not be explored. Fourth, cases had a substantially higher body mass index and report of irregular menstrual periods than control subjects. Finally, in-depth evaluations of right ventricular function, such as by cardiac magnetic resonance imaging, were not available.

Regardless, the current study advances our understanding of the interaction of sex hormones and PAH, including potential modulation of the clinical condition over the course of a menstrual cycle. And it is an important reminder that studies of sex hormone associations in PAH must account for expected cyclic variations among women of reproductive age. Whether the ratio of sex hormones, including sex hormone metabolites, corrects for this variation to some degree remains to be seen. Also, if cyclic variations modify response to therapy, such as in ongoing clinical trials of sex hormone modification involving women of reproductive age (NCT03528902 and NCT03648385), will be another important question to explore moving forward as we strive to determine the relationship between sex, sex hormones, and the PAH phenotype.