Hepatitis B and C co-infection are independent predictors of progressive kidney disease in HIV-positive, antiretroviral-treated adults

Abstract

Chronic kidney disease (CKD) is an important cause of morbidity and mortality in HIV-positive individuals. Hepatitis C (HCV) co-infection has been associated with increased risk of CKD, but prior studies lack information on potential mechanisms. We evaluated the association between HCV or hepatitis B (HBV) co-infection and progressive CKD among 3,441 antiretroviral-treated clinical trial participants. Progressive CKD was defined as the composite of end-stage renal disease, renal death, or significant glomerular filtration rate (eGFR) decline (25% decline to eGFR <60 mL/min/1.73 m(2) or 25% decline with a baseline <60). Generalized Estimating Equations were used to model the odds of progressive CKD. At baseline, 13.8% and 3.3% of participants were co-infected with HCV and HBV, respectively. Median eGFR was 111, and 3.7% developed progressive CKD. After adjustment, the odds of progressive CKD were increased in participants with HCV (OR 1.72, 95% CI 1.07-2.76) or HBV (OR 2.26, 95% CI 1.15-4.44). Participants with undetectable or low HCV-RNA had similar odds of progressive CKD as HCV seronegative participants, while participants with HCV-RNA >800,000 IU/ml had increased odds (OR 3.07; 95% CI 1.60-5.90). Interleukin-6, hyaluronic acid, and the FIB-4 hepatic fibrosis index were higher among participants who developed progressive CKD, but were no longer associated with progressive CKD after adjustment. Future studies should validate the relationship between HCV viremia and CKD.

Trial registration: ClinicalTrials.gov NCT00027352; NCT00004978.

Figure 1. Cumulative proportion of participants with progressive kidney disease.

Figure 2. Association of Hepatitis B and C viremia and Hepatitis C genotype with progressive kidney disease.

Final models were adjusted for trial, age, race, infection through same sex exposure, prior AIDS diagnosis, CD4, CD4 nadir, previous exposure to antiretrovirals, previous exposure to indinavir, treatment with antihypertensives, treatment with lipid lowering therapy, estimated glomerular filtration rate, and year of randomization, all measured at randomization into the parent trial. As appropriate, analyses were also adjusted for ¹Hepatitis C (HCV) antibody status or ² Hepatitis B surface antigen (HBsAg) status. Participants with positive HBsAg were further stratified by the presence or absence of detectable HBV DNA, with a cutoff of <357 IU/ml. The number of patients (events) for HBsAg negative, positive with HBV DNA <357 IU/mL, and positive with HBV DNA ≥357 IU/ml were 3321 (113), 44 (6), and 70 (8), respectively. Participants with positive HCV antibody were further stratified into those with undetectable HCV RNA, low HCV RNA (≤800,000 IU/ml), and high HCV RNA (>800,000 IU/ml). The number of patients (events) for HCV antibody negative, positive with undetectable HCV RNA, low HCV RNA, and high HCV RNA were 2963 (102), 108 (4), 212 (9), and 151 (12), respectively. The analysis of HCV genotype was limited to participants with positive HCV antibody and known HCV genotype, stratified as genotype 1 (258 patients, 18 events) or other (95 patients, 3 events).

Figure 3. Association of Hepatitis C viremia with progressive kidney disease.

Hepatitis C (HCV)-RNA was further stratified by quartiles to determine whether there is clinically useful threshold HCV-RNA level associated with increased risk of progressive kidney disease. The number of patients (events) included for HCV antibody negative and for increasing quartiles of HCV-RNA were 2963 (102), 108 (4), 154 (7), 102 (7), and 107 (7), respectively. Final models were adjusted for trial, age, race, infection through same sex exposure, prior AIDS diagnosis, CD4, CD4 nadir, previous exposure to antiretrovirals, previous exposure to indinavir, treatment with antihypertensives, treatment with lipid lowering therapy, estimated glomerular filtration rate, year of randomization, and Hepatitis B surface antigen status, all measured at randomization into the parent trial.