Abstract
BACKGROUND: Sex and social determinants of health predict stroke, yet few studies focus exclusively on women at high social risk. Understanding associations between stroke and modifiable risk factors that are disproportionately common in high-risk populations may aid in tailoring primary and secondary prevention services. We sought a better understanding of the association between polysubstance use, with an emphasis on stimulant use, and stroke in women who experience homelessness and unstable housing. METHODS: We recruited women who experience homelessness and unstable housing from San Francisco community-based venues to participate in a cohort study. We identified stroke from electronic health records during the 2.5-year study period and prospectively for up to 2 years after study completion. We then assessed associations with baseline study factors, including social determinants of health, toxicology-confirmed use of multiple substances, and traditional stroke risk factors. We used multiple logistic regression and the Akaike Information Criterion to construct the most parsimonious adjusted model to determine independent relationships between study factors and stroke risk. RESULTS: Of 245 participants, 238 had a complete data set and were included. The median participant age was 53.4 (interquartile range, 45.4-59.4) years. Nine (3.8%) participants had documented strokes. After adjusting for age, hyperlipidemia, and opioid use, cocaethylene-a biomarker signaling alcohol and cocaine concurrent co-use-had a particularly strong association with stroke (odds ratio, 3.22 [95% CI, 1.02-10.20]). CONCLUSIONS: Cocaethylene and opioids were strongly associated with stroke among women who experience homelessness and unstable housing. This suggests 2 possible opportunities for prevention in an environment where opioid use is already closely monitored. First, women reporting cocaine and alcohol co-use should be counseled about the especially harmful effects of this substance combination on their cerebrovascular health. Second, the utility of cocaethylene to assess stroke risk in high-risk populations may be currently underestimated.