Background: Those receiving tenofovir/emtricitabine (TDF-FTC) had greater bone loss compared with abacavir/lamivudine (ABC-3TC) in a randomized simplification trial (STEAL study). Previous studies associated increased bone turnover and bone loss with initiation of antiretroviral treatment, however it is unclear whether change in bone mineral density (BMD) was a result of specific drugs, from immune reconstitution or from suppression of HIV replication. This analysis determined predictors of BMD change in the hip and spine by dual-energy x-ray absorptiometry in virologically suppressed participants through week 96. Methodology/Principal Findings: Bone turnover markers (BTMs) tested were: formation [bone alkaline phosphatase, procollagen type 1 N-terminal propeptide (P1NP)]; resorption (C-terminal cross-linking telopeptide of type 1 collagen [CTx]); and bone cytokine-signalling (osteoprotegerin, RANK ligand). Independent predictors of BMD change were determined using forward, stepwise, linear regression. BTM changes and fracture risk (FRAX®) at week 96 were compared by t-test. Baseline characteristics (n = 301) were: 98% male, mean age 45 years, current protease-inhibitor (PI) 23%, tenofovir/abacavir-naïve 52%. Independent baseline predictors of greater hip and spine bone loss were TDF-FTC randomisation (p≤0.013), lower fat mass (p-trend≤0.009), lower P1NP (p = 0.015), and higher hip T score/spine BMD (p-trend≤0.006). Baseline PI use was associated with greater spine bone loss (p = 0.004). TDF-FTC increased P1NP and CTx through Wk96 (p<0.01). Early changes in BTM did not predict bone loss at week 96. No significant between-group difference was found in fracture risk. Conclusions/Significance: Tenofovir/emtricitabine treatment, lower bone formation and lower fat mass predicted subsequent bone loss. There was no association between TDF-FTC and fracture risk.