The study was approved by the Danish Data Protection Agency, the Danish Medicines Agency and the Regional Ethical Committees. The study was conducted and monitored according to good clinical practice (GCP). All patients provided informed written consent. The ClinicalTrial.gov identifier was NCT00135460. Antiretroviral-naïve

HIV-infected patients who were at least 18 years of age were eligible for inclusion in the study when the treating physician found indications for antiretroviral treatment. National criteria for initiating HAART were HIV-related disease, acute HIV infection, pregnancy, or CD4 cell count <300 cells/μL [12]. The exclusion criteria were pregnancy or breastfeeding, ongoing illicit drug use, serum creatinine concentrations above 200 μmol/L, alanine aminotransferase or aspartate aminotransferase values more than Selleck Antiinfection Compound Library five times the upper normal limit, or ongoing medical treatment with drugs having a clinically significant interaction with lopinavir, ritonavir or efavirenz. BMD and T-scores were measured at the lumbar spine (lumbar vertebrae 1–4) and femoral neck using DEXA scanning. Two centres used Norland XR

36 (Norland Corporation, Fort Atkinson, WI) and one centre used Hologic (Hologic Inc., Bedford, MA). The scanners were calibrated daily against a standard calibration block to avoid drift and shift. Trained radiographic personnel blinded to the treatment arm read the DEXA scans at each centre. For HDAC cancer each patient, only scans from the same scanner were analysed. As mentioned, randomization was stratified by centre and therefore also by scanner. Osteoporosis was defined as recommended by the National Osteoporosis Foundation according to the T-score [13]. The T-score is the difference between a person’s BMD and the mean BMD of a young (20–30-year-old)

race- and gender-matched reference population divided by the standard deviation of the group. Patients with at least one of the two T-scores (spine and femoral neck) <−2.5 were defined as having osteoporosis. FER Patients with at least one of the two T-scores <−1 were categorized as having low BMD. We included all patients with baseline BMD measurements and at least one follow-up BMD measurement. Baseline data are presented as medians and interquartile ranges (IQRs). Differences between groups were compared using the Mann–Whitney U-test and χ2 test as appropriate. We used Cox regression analyses to compare time to discontinuation of randomized treatment. The evolution of BMD is presented as the mean percentage change from baseline with 95% confidence intervals (CIs). Data were analysed for the intention-to-treat (ITT) population regardless of whether patients had stayed on randomized treatment or not. Furthermore, we conducted an ‘on-class’ analysis including both patients still on randomized treatment and patients who switched one or more drugs, respecting the assigned NRTI-sparing or PI-sparing arm (e.g.