Jon C Mills, Brian W Pence, Jonathan V Todd, Angela M Bengtson, Tiffany L Breger, Andrew Edmonds, Robert L Cook, Adebola Adedimeji, Rebecca M Schwartz, Seble Kassaye, Joel Milam, Jennifer Cocohoba, Mardge Cohen, Elizabeth Golub, Gretchen Neigh, Margaret Fischl, Mirjam-Colette Kempf, Adaora A Adimora

Research linking depression to mortality among people living with HIV (PLWH) has largely focused on binary "always vs. never" characterizations of depression. However, depression is chronic, and is likely to have cumulative effects on mortality over time. Quantifying depression as a cumulative exposure may provide a better indication of the clinical benefit of enhanced depression treatment protocols delivered in routine HIV care settings. Women living with HIV (WLWH), naïve to antiretroviral therapy (ART), from the Women's Interagency HIV Study (WIHS) were followed from their first visit in or after 1998 for up to 10 semi-annual visits (5 years). Depressive symptoms were assessed using the Center for Epidemiologic Studies Depression (CES-D) scale. An area under the curve approach was used to translate CES-D scores into a time-updated measure of cumulative days with depression (CDWD). We estimated the effect of CDWD on all-cause mortality using marginal structural Cox proportional hazards models. Overall, 818 women contributed 3,292 woman-years over a median of 4.8 years of follow-up, during which the median (interquartile range) CDWD was 366 (97-853). Ninety-four women died during follow-up (2.9 deaths/100 woman-years). A dose-response relationship was observed between CDWD and mortality. Each additional 365 days spent with depression, consecutively or accumulated intermittently, increased mortality risk by 72% (hazard ratio=1.72, 95% confidence interval=1.34-2.20). In this sample of WLWH, increased CDWD elevated mortality rates in a dose-response fashion. More frequent monitoring and enhanced depression treatment protocols designed to reduce CDWD may interrupt the accumulation mortality risk among WLWH.