Through seven years of follow-up, bone mineral density (BMD) declined significantly more in antiretroviral-treated U.S. trial participants with HIV than in age-similar HIV-negative adults in two longitudinal cohorts. In people with HIV infection, the greatest drop came during the first 96 weeks of antiretroviral therapy, but after that BMD at the lumbar spine continued to fall at a faster rate with HIV than in HIV-negative controls, according to the study presented at AIDS 2016.
Declining BMD in antiretroviral-treated people is a well-appreciated phenomenon. But long-term BMD changes in people taking contemporary antiretroviral regimens, especially compared with HIV-negative adults the same age, remain poorly understood. AIDS Clinical Trials Group (ACTG) investigators addressed that issue by contacting participants in a 2005-2007 metabolic substudy, remeasuring their BMD and comparing findings with longitudinal changes in two HIV-negative cohorts.
In 2005-2007 ACTG A5202 randomly assigned participants to atazanavir/ritonavir or efavirenz (Sustiva, Stocrin) plus tenofovir/emtricitabine (Truvada) or abacavir/lamivudine (Epzicom, Kivexa). Participants in ACTG A5224s, a metabolic substudy of A5202, underwent dual energy x-ray absorptiometry (DXA) scans at baseline and week 96. For the new analysis, ACTG investigators contacted A5224s participants in 2013-2014 and asked them to have follow-up whole-body and site-specific DXA scans. They gathered seven-year DXA data on HIV-negative men in the Boston Area Community Health/Bone Survey and five-year DXA data on HIV-negative women in the Women's Interagency HIV Study (WIHS). They excluded data on participants outside the sex-specific age window of HIV-positive participants (20 to 64 years for men and 23 to 55 years for women).
Read: Higher Rates of Five Major Non-AIDS Comorbidities After Age 50
The ACTG team used repeated measures analysis to compare percentage change in spine and hip BMD per year in HIV-positive versus HIV-negative people. These analyses considered two periods, an early period covering the first 96 weeks of follow-up and a late period covering week 96 to the end of follow-up. The analyses adjusted for age, sex, race, body mass index, self-reported physical activity, cigarette and alcohol use and use of concomitant medications that may affect BMD.
The researchers focused on 97 of 269 participants in ACTG A5224s (86% men) and 614 HIV-negative controls (87% men). Age at first DXA was significantly younger in the HIV group (median 40 versus 46 years, P < .001), and body mass index was lower in the HIV group at first DXA (median 24 versus 28 kg/m2, P < .001) and final DXA (27 versus 29 kg/m2, P < .001). One-third of participants in the HIV-positive and negative groups were black, but the HIV group had a higher proportion of whites (47% versus 35%) and a lower proportion of Hispanics (14% versus 31%) (P < .001). Median time between first and final DXA measured 7.6 years in the HIV group and 6.9 years in HIV-negative controls.
From baseline to week 96, HIV-positive people had a significantly greater lumbar spine BMD decline than controls in an adjusted analysis: -0.76% versus -0.09% per year (P = .001). Adjusted annual rates of decline in hip BMD through week 96 were -1.56% for the HIV group and -0.31% for controls (P < .001). After week 96 the drop in BMD slowed in both study groups. But spine BMD continued to fall in the HIV group while stabilizing in the control group: From week 96 through the end of follow-up, adjusted change in spine BMD measured -0.25% in the HIV group and 0.09% in controls (P = .008). During this later period annual change in hip BMD was -0.31% in both the HIV group and controls.
Through follow-up week 96, three variables independently predicted greater lumbar spine BMD loss in people with HIV: randomization to tenofovir/emtricitabine (P = .028), lower baseline CD4 count (P = .005) and higher baseline viral load (P = .029). Two factors independently predicted greater hip BMD loss through week 96: lower time-updated lean body mass (P = .001) and time-updated use of medications that decrease BMD (P = .03). From week 96 to the end of follow-up in the HIV group, only lower time-updated total lean body mass independently predicted greater lumbar spine BMD loss (P < .001). Two factors independently predicted greater hip BMD loss in the later period: concomitant medications that lower BMD (P < .001) and lower time-updated lean body mass (P < .001). Cumulative exposure to tenofovir or a protease inhibitor did not affect BMD change in the late study period.
The researchers noted that "the difference in the rate of change in BMD between infected individuals and uninfected individuals is relatively modest." But the greater BMD loss with HIV "could be compounded in persons with preexisting low BMD or if the loss continues beyond the evaluated period." Because people who take antiretroviral therapy for decades may run an increased risk of fragility fractures, the authors propose that "interventions to maintain or increase lean body mass may benefit long-term skeletal health."