Abstract Body

It is well established that antiretrovirals (ARVs) affect markers of bone turnover, but less is known about their effect on risk of fractures and femoral osteonecrosis. We hypothesized that exposure to ARVs including tenofovir (TDF) would increase the risk of both outcomes. 

EuroSIDA participants were prospectively followed from baseline (Jan 2004) until last visit or death to assess fractures and femoral osteonecrosis. Poisson regression was used to identify clinical, laboratory and demographic factors associated with either bone endpoint. Ever, current and cumulative exposures to each ARV were added to the multivariate model.

During 86118 person-years of follow up (PYFU) among 11820 eligible persons (median age 41y, 75% male, 86% white, median baseline CD4 440/mm3 and 70.4% virologically suppressed), there were 618 incident fractures (incidence/1000PYFU 7.2; 95%CI 6.6-7.7) and 89 incident cases of osteonecrosis (1.0;0.8-1.3). After adjustment, higher risk of fractures was associated with older age, white race,  lower BMI, IV drug use, lower baseline CD4, HCV-coinfection, prior osteonecrosis, prior fracture, recent non-AIDS cancer and recent cardiovascular disease (last 12 months) (Figure). The crude incidence of fracture was 8.1/1000PYFU (7.3-8.9) in those ever exposed to TDF compared to 4.7 (4.1-5.4) in those never exposed; corresponding figures for persons currently on and off TDF were 7.8 (6.8-8.7) and 5.6 (5.0-6.3). After adjustment, persons who had ever used TDF (1.40; 1.15-1.70) or who were currently on TDF (1.25; 1.05-1.49) had a significantly higher incidence of fractures. There was no association between longer exposure to TDF and fractures (1.02/5y exposure; 0.94-1.25). No other ARV was associated with fractures (all p>0.1). Risk of osteonecrosis was associated with white race, lower nadir CD4, prior osteonecrosis, prior fracture and prior AIDS (Figure). Persons who had ever used didanosine, indinavir, saquinavir, lopinavir/r, or TDF had higher risk of osteonecrosis, but this association was no longer observed in models adjusted for confounders (not shown).

In HIV infection, host factors, HIV-specific variables and co-morbidities contribute to risk of fractures and osteonecrosis. TDF but not other ARVs was an independent risk factor for fractures. There was no association between the use of any of the ARVs and risk of osteonecrosis.