Beyond race and ethnicity: utilizing genetic ancestry and measures of social vulnerability in a diverse population to characterize disparities in the diagnosis of endometriosis

Since race and ethnicity are social constructs, we aimed to determine the association between genetic ancestry and endometriosis diagnosis, and to examine the interactions between ancestry, self-reported race/ethnicity, and social vulnerability to better characterize disparities in endometriosis. This is a retrospective cohort study of female patients age ≤18-51 enrolled in Penn Medicine BioBank 2008-2020. Genotype data and principal component analysis were used to categorize patients into 5 ancestral populations: African (AFR), Admixed American (AMR), East Asian (EAS), European (EUR), or South Asian (SAS). Primary outcome was endometriosis diagnosis by ICD-9/10 codes. Logistic regression was adjusted for age and body mass index (BMI). Subgroup analyses were conducted based on endometriosis subtype and social vulnerability metrics. Chi-squared tests were used to examine differences in race/ethnicity and social vulnerability by ancestry in patients with endometriosis. Among 9,914 female patients (age 35.9 ± 8.7 years, BMI 29.4 ± 8.8 kg/m2), the distribution of genetic ancestry was highly diverse: 55.7% EUR, 36.2% AFR, 3.2% EAS, 2.9% AMR, and 2.1% SAS. Race/ethnicity also varied: 52.4% White, 34.2% Black, 4.5% Hispanic, 3.8% Asian. Genetic ancestry was associated with race/ethnicity for all patients (p<0.0001). Of EUR patients with endometriosis, 94.9% identified as White, 96.7% of AFR patients identified as Black, and 100% of EAS or SAS patients identified as Asian. Greater genetic variation was seen among Hispanic patients with endometriosis, of which 32% were EUR, 29% AFR, 35% AMR, and 3% SAS. 794 patients (8.0%) had a diagnosis of endometriosis. Endometriosis was more common among women of AFR ancestry than EUR, adjusting for age and BMI (aOR 1.97, 95% CI 1.7-2.3). This was driven by patients with endometriosis of the uterus (deep or superficial disease, or adenomyosis), which was strongly associated with AFR (aOR 2.98, 95% CI 2.4-3.6) and concurrent fibroids (aOR 4.87, 95% CI 4.0-6.0). Excluding uterine disease, we found that endometriosis of other organs (e.g. ovary, peritoneum, rectovaginal septum) was less common among AFR patients, though this was not significant (aOR 0.82, 95% CI 0.6-1.1). There were no differences in endometriosis among other ancestry groups, regardless of subtype. The association between endometriosis and ancestry was independent of social vulnerability metrics, such as socioeconomic status (SES), household characteristics, and access to care, though these metrics did differ by ancestry. Of AFR patients, 88% were in the lowest SES tertile, compared with 10% of EUR patients. Genetic ancestry and race/ethnicity are closely associated in patients with endometriosis. Significant differences exist in endometriosis diagnosis based on ancestry, with AFR associated with uterine subtype and EUR associated with other organ involvement. Studies that exclude endometriosis of the uterus may overlook AFR patients with disease.