Bone mineral loss in HIV-positive patients receiving antiretroviral therapy

Tebas and colleagues [1] employed dual-energy X-ray absorptiometry to measure bone mineral density (BMD) in a cross-sectional analysis of 112 male subjects, consisting of: (i) HIV-infected subjects treated with highly active antiretroviral therapies (HAART) containing a HIV protease inhibitor (PI); (ii) PI-naïve HIV-infected subjects treated with nucleoside reverse transcriptase inhibitors (NRTI); and (iii) seronegative controls. The authors concluded that osteopenia and osteoporosis were associated with PI-containing therapies. However, it was stated that a clear association between decreased BMD and total PI exposure needed to be interpreted with caution, because the test numbers were small and the baseline BMD values before HAART were unknown. We agree with the authors’ statement that further studies comparing antiretroviral regimens with and without PI are required to establish conclusively the role of the various components in the loss of BMD. Furthermore, we agree that the decrease in BMD may be related to metabolic complications of lipodystrophy and dyslipidemia associated with HAART. However, because no data have been reported to date comparing the effects of individual PI in causing the loss of BMD, we feel it is premature for Tebas et al. [1] to imply that this adverse event is a class-specific effect of PI usage. Metabolic complications were thought initially to be caused solely by HIV PI, but now HIV infection and multiple classes of antiretroviral agents may be implicated, including select NRTI, non-NRTI or hydroxyurea. For example, in the absence of PI therapy, lipodystrophy was observed in 63% of patients taking stavudine and 19% taking zidovudine after 14 months [2]. This suggests that there are distinct differences in the mechanism of action and degree of culpability among the various member compounds within a given class of drugs to affect these complications. Consistent with this suggestion, ritonavir therapy elevates serum lipid levels more than other currently available PI [3]. Likewise, indinavir is the only PI to alter retinoid signaling in vitro[4]. Moreover, in various experimental model systems we have observed clear differences between the various PI in their ability to affect fat metabolism [4]. In addition, PI and NRTI may interact to accelerate the onset of lipodystrophy [5]. Emerging data thus suggest distinct differences within a class and between classes of antiretroviral therapies on affecting metabolic complications. In our opinion, a major limitation of the study by Tebas et al. [1] was its failure to discriminate between the effects of individual PI on decreased BMD. In particular, indinavir is the only PI known to alter the activity of osteoblast alkaline phosphatase in vitro[4], a potential mechanism affecting BMD after therapy. A second limitation is that Tebas et al[1] did not address adequately the role of previous or current NRTI on the loss of BMD. Because the duration of stavudine therapy affects lipodystrophy [2], we feel strongly that the issue of previous exposure to NRTI must be controlled before suggesting that PI cause osteopenia and osteoporosis. Likewise, the effects of the additional components comprising HAART need to be determined. In closing, further work is required to compare and discriminate between individual PI and address the role of non-PI-containing regimens on decreased BMD. The observation that the HIV plus PI patient population is 4 years older than the PI-naive population [1] raises the question of whether there are other differences in patient histories between both groups. Until this is explored, we feel it unwise and misleading to imply that all PI cause the same adverse effects on BMD or that osteopenia and osteoporosis are class-specific effects caused by the use of PI. James E. Weiel James M. Lenhard