We aimed to compare the evolution of estimated glomerular filtration rate (eGFR) in HIV-, HIV-HBV- and HBV-infected patients treated with tenofovir disoproxil fumarate (TDF). Three groups of patients receiving TDF > 12 months were recruited: 194 HIV-infected patients, 85 HIV-HBV-coinfected patients and 50 HBV-infected patients. eGFR was estimated using the Modification of the Diet in Renal Disease (MDRD) equation. Multivariate regression models were constructed to estimate factors associated with eGFR decrease from baseline. A total of 329 patients were studied. Median follow-up was 2.7 years. Median eGFR decrease was -4.9 (-16.6 to +7.2) mL/min/1.73 m(2) . After multivariate stepwise regression analysis, age (P = 0.0002), non-African origin (P < 0.0001), baseline eGFR (P < 0.0001) and TDF duration (P = 0.02) were associated with eGFR decrease in the whole population, while hypertension, diabetes and type of infection were not. Age (P < 0.0001), non-African origin (P = 0.0004), baseline eGFR (P < 0.0001) and TDF duration (P = 0.007) remained associated with eGFR decline in HIV and HIV-HBV-infected patients, while other variables including HIV risk factor, CDC stage, CD4 and HIV-RNA levels were not. Age (P = 0.03), non-African origin (P = 0.004), baseline eGFR (P < 0.0001) and baseline HBV-DNA > 2000 IU/mL (P = 0.04) were associated with eGFR decline in HBV and HIV-HBV-infected patients, while other variables including HBV risk factor and fibrosis stage were not. Estimated glomerular filtration rate decline under TDF therapy appears mainly associated with older age, non-African origin, higher baseline eGFR and longer TDF administration but not with the type of viral infection. Regular follow-up of renal function, especially tubular function is recommended during TDF therapy.