Why Does Allostatic Load Not Differ Among Women of Diverse Sexual Orientations?

This editorial aims to provide reflection and discussion as to why women do not appear to differ by sexual orientation when assessing allostatic load (AL). AL refers to the multisystemic “wear and tear” of chronic stress in synergy with unhealthy behaviors ¹ that can be indexed with several biomarkers representing metabolic, cardiovascular, immune, and neuroendocrine functioning. ² In addition to large number of studies assessing race/ethnic differences, a small but growing literature has emerged that applies the AL model to assess sexual diversity and to indirectly assess whether stigma can impact AL and health trajectories. ^3,4

In this issue of the Journal of Women's Health, Walubita et al., ⁵ analyzed 2001–2010 as well as 2015–2016 waves from the National Health and Nutrition Examination Survey (NHANES). They focused on Black women to determine whether sexual orientation differences exist in AL indexed with nine biomarkers representing metabolic (glycosylated hemoglobin, body mass index, total and high-density lipoprotein cholesterol), cardiovascular (systolic and diastolic blood pressure, heart rate), and immune (albumin and c-reactive protein) functioning. It is worth noting that like other NHANES AL studies, this analysis did not include neuroendocrine biomarkers such as cortisol central to the primary mediators of chronic stress in the AL model. Notwithstanding, the power of the NHANES is the large sample size that allows for subsamples to be explored.

To summarize, no significant differences in AL by sexual orientation among Black women were detected. ⁵ This analysis provides a much-needed exploration of intersectionality as a guiding principle. Intersectionality states that multiple marginalized identities compound together to worsen health disparities among marginalized subgroups. Among Black women in particular, the “weathering hypothesis” has been proposed stating that Black women experience an accelerated aging process and succumb to chronic diseases earlier than other race and ethnic groups. We believe Wulubita et al.'s study makes an important contribution to these literature as well as the emerging AL literature on sexual orientation. To help expand thinking and discussion in this emerging literature, we have several points of consideration in the interests of better understanding why differences in AL are difficult to identify within sex for birth-assigned women. Note that gender identity has yet to be included in the NHANES, which limits any representation of transgender and/or nonbinary Americans.

Numerous studies use the NHANES as their primary database in the AL literature. To date, however, only a few studies have conducted analyses linking sexual orientation to AL and fewer still have applied an intersectional perspective. As such, the analysis by Walubita et al. ⁵ is highly innovative and could inspire others to follow suit, especially by expanding intersectional representation. For instance, including combinations of White and Hispanic groups, considering socioeconomics, geopolitics, and other variability as moderators rather than as covariates are all approaches that can allow intersectional analyses.

Consider the following as examples of intersectionality. To date, existing NHANES studies that compare heterosexual women with sexual minority women (SMW) report that SMW are younger, ^{6 –12} consume significantly more alcohol, ^{6 –10,12,13} consume more drugs, ^8,10 and smoke more cigarettes. ^{6 –10,13} In addition, SMW are less likely to be married, ^{6,7,10 –12} to have health insurance coverage, ^7,10,12,13 and to have completed some college education. ^7,10,13 SMW also report a significantly lower income ^6,7,10,12 and significantly more mental distress ^6,10,13 than their heterosexual counterparts. As most of these variables are well-known covariates in the AL literature, ¹⁴ they theoretically would put SMW at risk to be in the AL high-range category. In addition, many of these factors that are otherwise used as covariates are central to intersectionality and should, therefore, be considered interactively.

Despite these behavioral, relational, social, psychological, and socioeconomic disadvantages, the only study specifically comparing AL between different sexual minority subgroups using the NHANES database surprisingly reported no differences between the AL of women from all sexual orientations. ¹³ Indeed, the difference in AL by sexual orientation is only present in men, whereby bisexual men show the highest AL and gay men show the lowest AL. In the first study of sexual orientation and AL among a small convenience sample (N = 87) of Canadians, Juster et al. likewise found that sexual minority men showed lower AL than heterosexual men, while no differences in AL were found by sexual orientation for women. ¹⁵ Additional analyses in this same sample, Juster found that SMW (75% White) had higher levels of testosterone, progesterone, dehydroepiandrosterone-sulfate, and stress reactive cortisol, as well as lower low-density lopoprotein (LDL) cholesterol than heterosexual women (55% White). Many of these biomarkers were positively associated with psychosocial distress.

Why then are there no differences found in AL by sexual orientation in women? Is there perhaps some paradox among SMW when it comes to stress biomarkers that could explain this? There are significant differences within SMW subgroups when comparing the biomarkers used to calculate AL. ^9,13 In addition, there are significant differences between women of different race/ethnic groups by sexual orientation when comparing these biomarkers. ⁷ Yet, these differences are interpreted according to the research objectives and focus of the analysis at hand. Moreover, biomarker variability among different subgroups is rarely considered. In most cases with the NHANES, the objectives of a given analysis focus on cardiovascular disease risk but not on AL specifically that comprises multiple connected systems. For example, non-Hispanic Black (NHB) women who self-identify as bisexual seem to score worse on several risk factors for cardiovascular diseases compared with SMW of different race/ethnicity and compared with heterosexual women of different race/ethnicity. ⁷

Studies using the NHANES to focus on the AL model are more frequently representative of race/ethnic health disparities, but without taking sexual orientation into account until only recently with May and colleagues' work. It is worth noting that NHANES data have some issues of comparative invalidity for the sexuality questionnaires, ¹⁶ specifically regarding sexual identity and sexual behavior items. For example, a Hispanic man may not self-identify as bisexual despite having engaged in same-sex sexual behaviors. Furthermore, health disparities exist between SMW subgroups when we take into account the way sexual minority statuses are measured: women who self-identify as being a sexual minority seem to have a higher prevalence of chronic diseases ¹¹ and seem to be 5.1% older in terms of vascular integrity than their heterosexual counterparts and than women who have sex with women but who self-identify as heterosexual. ⁸ These methodological issues could potentially explain the under-representation of sexual orientation in NHANES studies and present themselves as potential confounders.

Race and ethnicity are also not uniform in the NHANES. For example, the AL of NHB seems to be higher than that of non-Hispanic Whites (NHW) even when controlling for socioeconomic status or education. ^{14,17 –20} There also seem to be differences in the biological pathway through which AL is being influenced when we compare NHB and NHW. ^17,21 Of particular importance for concepts of “weathering” is the accelerated aging some race/ethnic minority women have been shown to experience, according to some of early AL studies using NHANES analyses. ²²

To conclude, results are often contradictory or incomplete, and so they cannot confirm intersectional theory involving AL at this time. However, finding ways to better represent diverse groups and subgroups and the unique lived experiences therein may provide unique insights. Indeed, multiple methodological factors could be involved in intersectionality and the measurement of AL itself. For example, a study by Duong et al. ²³ reported 18 different ways to calculate AL in 21 studies examining AL with the NHANES database. This lack of consensus could partially explain the lack of significant differences found in the study of Walubita et al. ³ In contrast, there are numerous psychosocial variables in the NHANES that might be indexed as to provide more contextual information. Ultimately, exploring these variables among subgroups might help improve researchers' ability to detect associations with AL and health outcomes. Of particular interest for future studies is exploring the rich data on health behaviors amassed in the NHANES to understand, for example, how physical activity, nutrition, alcohol consumption, smoking, drug use, medication use, sexual behaviors, and other health behaviors contribute to AL. In addition, perhaps exploring subgroup differences in moderation analyses that consider synergies among birth assigned sex, sexual orientation, race/ethnicity, age, and other intersections could provide insights not otherwise detectable.