Greater decrease in bone mineral density with protease inhibitor regimens compared with nonnucleoside reverse transcriptase inhibitor regimens in HIV-1 infected naive patients

To evaluate the change in bone mineral density (BMD) at specific sites in patients initiating antiretroviral therapy in a substudy of the ANRS 121 trial. Antiretroviral-naive patients were randomized (2: 1: 1) into three treatment strategy arms: a nonnucleoside reverse transcriptase inhibitor (NNRTI) and a boosted protease inhibitor (PI/r), a PI/r and two nucleoside reverse transcriptase inhibitors (NRTIs) or an NNRTI and NRTIs. Hip and lumbar spine standardized BMD were evaluated at baseline and week 48 by dual X-ray absorptiometry by a central reading laboratory. Seventy-one patients were enrolled: 36 in the PI/r and NNRTI, 19 in the PI/r and NRTIs and 16 in the NNRTI and NRTIs arms. Baseline characteristics were [median (interquartile range)]: male (77%), age 40 years (33-49), 69% white, 58% smokers, BMI 23 kg/m2 (21-24), CD4 cell count 219 cells/microl (144-285). In the arms with NRTIs, 86% of patients received zidovudine/lamivudine. At baseline, 31% had osteopenia and 3% had osteoporosis. At week 48, there was a mean change in BMD of -4.1 +/- 3.9% at lumbar spine and -2.8 +/- 4.7% at hip (both P< or = 0.001). The decrease of BMD at lumbar spine was significantly worse in the PI/r and NNRTI arm (-4.4 +/- 3.4%) and in the PI/r and NRTIs arm (-5.8 +/- 4.5%) compared with the NNRTI and NRTIs arm (-1.5 +/- 2.9%), P = 0.007 and P = 0.001, respectively. BMD was impaired in 34% of patients, before starting any antiretrovirals. After 1 year, the decrease in lumbar spine BMD was more pronounced in patients receiving either PI/r-containing regimen compared with NNRTI and NRTIs. BMD at specific sites should be monitored during lifelong antiretroviral therapy.