HIV, nephrotoxic medications, and chronic kidney disease: Prevalence, risk factors, and mediation analyses among people with and without HIV enrolled in the Multicenter AIDS Cohort Study (MACS)/ Women’s Interagency HIV Study (WIHS) combined cohort study

Abstract

Background Chronic kidney disease (CKD) affects over 37 million adults in the United States, and people living with HIV (PLWH) are at greater risk for progression to end-stage kidney disease. Although both conditions are common among PLWH, the potential pathways through which depression and use of medications with nephrotoxic potential may influence CKD development remain underexplored. We evaluated the relationships of depression and nephrotoxic medication use with CKD prevalence among PLWH, and investigated the potential mediating effects of these factors on the pathway to CKD among PLWH. Methods We analyzed data from the Multicenter AIDS Cohort Study (MACS)/Women’s Interagency HIV Study (WIHS) Combined Cohort Study (MWCCS). Baseline CKD prevalence was estimated using Visit 101 only (November 2020-September 2021). To examine associations with CKD over time, we fit generalized estimating equations (GEE) using repeated observations from Visits 101–103 (November 2020-March 30, 2023) with a Poisson distribution and log link to estimate relative risks (RR) and account for within-participant correlation. Counterfactual-based causal mediation analyses were conducted using Visit 101–103 data to evaluate whether elevated depressive symptoms (CES-D ≥ 16) or nephrotoxic medication use mediated the HIV-CKD association while adjusting for baseline confounders. Results Among 2,530 participants [1,622 PLWH and 908 people living without HIV (PLWoH)], CKD prevalence was higher in PLWH (18.1%) compared to PLWoH (9.7%). In univariate repeated-measures GEE models, HIV serostatus (RR = 1.37, 95% CI: 1.28–1.48, p < 0.0001) and Nephrotoxic medication use (RR = 1.49, 95% CI: 1.30–1.71, p < 0.0001) were significantly associated with higher CKD risk. Several covariates were also associated with CKD in univariate GEE models, including age (RR = 1.03, 95% CI: 1.03–1.03, p < 0.0001), non-Hispanic Black (RR = 1.19, 95% CI: 1.11–1.27, p < 0.0001) compared to non-Hispanic White, diabetes (RR = 1.26, 95% CI: 1.17–1.35, p < 0.0001), and higher income (RR = 0.99, 95% CI: 0.98–1.00, p = 0.005). Depressive symptoms were not associated with CKD in mediation-model adjusted analyses and did not mediate the HIV-CKD association. Mediation analysis indicated that nephrotoxic medication use accounted for a small but significant proportion of the HIV-CKD association (indirect effect OR = 1.02, 95% CI: 1.00–1.03, p = 0.02). Conclusions While it is well established that PLWH have a higher prevalence of CKD compared to PLWoH, our findings suggest that nephrotoxic medication use may modestly amplify this risk. Although most of the risk appears to be attributable to the direct effects of HIV, these medications represent a modifiable contributor. PLWH receiving such treatments may benefit from closer kidney function monitoring. Future research should evaluate psychosocial contributors to CKD using designs that incorporate clinical depression diagnosis and treatment data to clarify depression-related pathways.