Background:  Tenofovir (TDF), atazanavir (ATV), and other ARV have been associated with renal impairment, but the extent of such adverse reactions in HIV⁺ persons with normal baseline renal function is unknown.

Methods:  From 2004 D:A:D participants with Cockcroft-Gault-estimated glomerular filtration rates (eGFR) ≥90 mL/min were followed to confirmed (≥3 months apart) eGFR ≤70 (ceGFR <70; threshold below which protective ARV switches may occur), confirmed eGFR ≤60 (CKD) or last eGFR. Discontinuation rates of individual ARV according to latest eGFR were assessed. Poisson regression models, adjusting for renal and HIV-related risk factors, were used.

Results:  Of 22,603 persons included, 468 (2.1%) progressed to ceGFR ≤70 (incidence rate 4.78/1000 person-years [95%CI 4.35 to 5.22]) and 131(0.6%) to CKD (1.33/1000 person-years [1.10 to 1.56]) during a median follow-up of 4.5 years (IQR 2.7 to 6.1). Persons with eGFR levels of 60 to 70 had significantly higher rates of discontinuing TDF, but not of other ARV (adjusted IR ratio 1.72 [1.38 to 2.14]) compared to persons with eGFR ≥90. Cumulative TDF and ATV/r use were independently associated with increased rates of ceGFR ≤70, but not CKD, whereas lopinavir/ritonavir (LPV/r) use was associated with both endpoints (see the table). Inconsistent trends were seen for abacavir. Analyses censoring for concomitant use of each of the ARV were highly consistent. Fitting ARV according to current or prior use, those currently on TDF, LPV/r, and ATV/r experienced an increased rate of ceGFR ≤70 with increased exposure, whereas >1 year after drug discontinuation the rates approached 1. Other predictors of ceGFR ≤70 were age (aIRR 2.63/10 years [2.33 to 2.96]), diabetes (1.54 [1.06 to 2.23]), hepatitis B virus (1.56 [1.10 to 2.22]), hepatitis C virus (1.47 [1.09 to 2.00]), current CD4 count (0.75/doubling [0.69 to 0.82]), and prior AIDS (1.38 [1.13 to 1.69]). There were no significant interactions between ARV and these predictors. No information on other nephrotoxic drugs or proteinuria was included.

Conclusions:  In HIV⁺ persons with normal baseline eGFR, cumulative use of TDF, ATV/r, and LPV/r was associated with fast eGFR decline. The rates of ARV associated progressions were low over 4.5 years’ follow-up, but may represent significant issues for lifelong use should they remain constant or increase with age. Increased discontinuation of TDF in people with decreasing eGFR suggests that drug switches may have prevented further deterioration to CKD. Closer monitoring of renal function may be appropriate for persons on ATZ/r and LPV/r.