Background
Women with HIV (WWH) are at increased risk for developing diabetes mellitus (DM) compared to women without HIV (WWOH). In WWOH, the transition to menopause is associated with metabolic changes that increase DM risk. Whether HIV and menopause are associated with faster increases in insulin resistance, and whether switching to an integrase strand transfer inhibitor (INSTI) modifies this association, is unknown.
Methods
From 2006-2019, 373 virally suppressed WWH and 300 WWOH from the Women’s Interagency HIV Study who were non-pregnant and did not have hepatitis C virus infection or DM at baseline were included in the analysis. Participants were WWH who switched to an INSTI (INSTI+), WWH who did not switch (INSTI-), and WWOH from a similar time window, or index visit. Linear mixed effect models first evaluated the change in insulin resistance estimated through log HOMA-IR across five years of follow-up by HIV status and time-varying menopausal phase. We then compared trajectories by INSTI group. We used anti-Müllerian hormone, a biomarker of ovarian reserve, to categorize women into premenopause, early peri-, late-peri-, and postmenopause. Models were adjusted for demographics, time-varying waist circumference, socio-behavioral factors, comorbidities, and HIV-related factors.
Results
At the index visit, 67% identified as Black and 37% were premenopausal, 11% early peri-, 25% late peri-, and 27% postmenopausal. Compared to WWOH, WWH were older (median age: 47 vs. 44 years) and had higher median HOMA-IR (1.9 vs 1.7 mg/dL), respectively. Among WWH, 76% were on tenofovir DF at the index visit. Figure shows HOMA-IR trajectories by HIV and INSTI status in late peri- and postmenopause. When adjusted for waist circumference, there was little difference in HOMA-IR trajectories by HIV status in premenopausal women. By contrast, in late perimenopause and menopause, WWH had 6.0% (95%CI:-0.2,12.6) and 8.6% (95%CI:1.8,15.8) faster increases in HOMA-IR/year, respectively, compared to WWOH. When compared to INSTI-, the percent change in HOMA-IR/year for INSTI+ was -3.1 (95%CI:-11.5,6.1) and -1.3 (95%CI:-9.1,7.2), in late perimenopause and postmenopause, respectively.
Conclusions
WWH in late perimenopause and menopause have faster increases in insulin resistance when compared to WWOH, even after adjustment for waist circumference. Switching to an INSTI did not further accelerate increases in insulin resistance. Our findings suggest that DM screening and prevention in midlife WWH is imperative.
Table or Figure
