Biomarkers influence kidney function estimates more so than race among persons with HIV

We sought to understand the results of using different estimating equations (with and without a "race" category) and the addition of cystatin C as a biomarker on Chronic Kidney Disease (CKD) stage estimates among persons with HIV (PWH), for whom CKD is an important comorbidity.

Biomarkers were measured in this cross-sectional single site U.S. clinic-based study from 2014 to 2016. Other laboratory and clinical data were abstracted from the electronic health record based on the last recorded value proximal to the study visit. Chronic Kidney Disease Epidemiology Collaboration (CKD-EPI) glomerular filtration estimating equations with and without cystatin C were applied to categorize CKD stage, and staging agreement was assessed using the difference of proportions test. Multivariable regression analyses evaluated factors associated with CKD stage, and the Breslow-Day test evaluated whether race served as an effect modifier.

Among 306 PWH, the median age was 48.2 years, 86 (28%) were female, 185 (61%) were Black, 91 (30%) Caucasian, 13 (4%) Latinx, and 46 (15%) had hepatitis C virus (HCV) co-infection. The median CD4 + T-lymphocyte count was 659/mm³, 299 (98%) were on ART, and 228 (75%) had HIV VL < 20 c/mL. Using the 2009 and 2012 CKD-EPI equations (including race), more individuals were categorized as having normal kidney function (Stage 1) with inclusion of cystatin C than creatinine alone (73% vs. 55%, p < 0.00001); fewer individuals were classified in CKD stages III-V using both cystatin C and creatinine than creatinine alone, though this did not meet statistical significance (8% vs. 12%, p = 0.14). Using 2021 equations (excluding race) a larger proportion were classified as normal kidney function with inclusion of cystatin C than creatinine alone (74% vs. 49%, p = 0.00001); fewer were categorized as CKD III-V with inclusion of cystatin C than creatinine alone (8% vs. 13%, p = 0.03). Multivariable linear regression identified age (β=-0.75, p < 0.0001) and tobacco use (β=-4.10, p = 0.03) as factors associated with kidney function. Race was not an effect modifier in our analyses based on the Tarone adjusted Breslow-Day test.

Among PWH, cystatin C shifted estimates of kidney function towards normal and resulted in shifts in kidney function categorization much more so than the race effect. As some antiretrovirals raise creatinine without affecting GFR, cystatin C is an important biomarker to confirm diminished kidney function among persons with HIV.