Background:  Several observational cohorts have reported a higher degree of decline in calculated glomerular filtration rates (GFR) in patients treated with tenofovir (TDF), particularly in combination with boosted protease inhibitors (PI), compared with nucleosides. 

Methods:  ESS40006 was primarily designed to compare 2 regimens of aprenavir/ritonavir (APV/r) (600/100 vs 900/100 twice daily) in subjects failing their current ART regimen. In addition, non-nucleoside reverse transcriptase inhibitor (NNRTI)-naïve subjects were assigned to receive efavirenz (EFV), abacavir (ABC), and 1 additional NRTI, while NNRTI-experienced subjects were assigned to receive TDF in place of EFV. Descriptive statistics were summarized for subjects treated with EFV or TDF. The modification of diet in renal disease (MDRD) formula was used to calculate GFR. Potential predictors of GFR decline over 48 weeks of therapy, including baseline demographic data, CDC HIV-1 classification, CD4⁺ cell count, plasma HIV-1 RNA, prior therapy, concurrent ART, weight, and clinical laboratory results were assessed using multiple regression analyses.

Results:  The median calculated GFR at baseline was comparable between the 2 groups (EFV-treated subjects [n = 38], 107.36; TDF-treated subjects [n = 76], 108.24). However, there were statistically significant differences in the median change in calculated GFR between the EFV-treated and TDF-treated subjects (median change from baseline at 24 weeks was 12.66 vs –9.85 [p <0.001] and at 48 weeks: –0.37 vs –11.07 [p = 0.004]). In the TDF group, the median reduction in calculated GFR from baseline was also statistically significant at both weeks 24 (p <0.001) and 48 (p <0.001). The only predictor of a decline in GFR in the multivariable model after adjusting for baseline GFR was TDF use.

Conclusions:  For NNRTI-experienced subjects treated with TDF in this study, a significant decline in the median calculated GFR was observed over 48 weeks of therapy. This decline was not seen in the NNRTI-naive subjects treated with EFV in the same study. TDF use was the only predictor of GFR decline using multiple regression analysis. The clinical significance of these changes in calculated GFR deserves further study.