Background:  Tenofovir DF (TDF) has demonstrated renal safety in clinical trials; however, case reports and observational studies have suggested that TDF may be nephrotoxic. In a previous analysis by our group of principally ART-experienced patients, 6 months’ treatment with TDF was associated with a greater decline in estimated glomerular filtration rate (GFR) (p <0.01) than alternative nucleoside reverse transcriptase inhibitors (NRTI). We have conducted a more extensive analysis of ART-experienced and -naïve patients comparing GFR in TDF and NRTI-receiving patients over 2 years.

Methods:  We analyzed data from an observational clinical cohort, comparing all patients with an estimated GFR >50 mL/minute/1.73 m² (MDRD equation) who initiated TDF (n = 565 ART-experienced, 62 ART-naive) or any alternative NRTI  (211 ART-experienced, 100 ART-naive) after January 1, 2001. We assessed absolute and percentage changes in GFR from start of therapy to maximum time of 2 years. The lowest GFR that occurred during therapy (confirmed by 2 measures) was used to calculate change in GFR.

Results:  There was no difference in median duration of therapy (TDF = 506 vs NRTI = 519 days; p = 0.25), baseline GFR (TDF = 115 vs NRTI = 114 mL/minute/1.73 m²; p = 0.27), nor number of creatinine measures during treatment (TDF = 10 vs NRTI = 9; p = 0.20). ART-experienced patients or TDF had greater declines in GFR than NRTI (15 vs 11 mL/min/1.73 m², p <0.01; 14.3% vs 8.8% decline p <0.001). The median time to lowest GFR was 132 days for NRTI and 138 days for TDF (p = 0.23). In contrast, ART-naïve patients on TDF had similar declines in GFR as those on NRTI (15 vs 14 mL/minute/1.73 m², p <0.10; 13.1% vs 13.8% decline, p <0.90). The median time to lowest GFR was 122 days for NRTI and 96 days for TDF (p = 0.05). Other factors associated with GFR decline were CD4 <50 cells/mm³, diabetes, and age >45 years. Adjusting for these variables and baseline GFR, age, race, sex, hypertension, and use of other potentially nephrotoxic drugs in a multivariate analysis, TDF remained associated with a greater decline in GFR (–7.0%, p <0.001) in ART-experienced, but not ART-naïve patients.

Conclusions:   Our data suggest that long-term use of TDF is associated with a modest decline in GFR that occurs in the first 6 months of therapy in ART-experienced patients. This relative decline was not observed in ART-naïve patients. The reasons for the GFR decline in ART-experienced vs -naïve patients is currently unknown. However, the decline is small in magnitude, does not appear progressive, and may not have an important clinical effect on most patients.