Bone Density Decline No Faster With HIV and Continuing Treatment
Bone mineral density (BMD) was lower with than without HIV infection in a 384-person cohort in Ireland. But, through three years of follow-up, BMD declined at a similar rate in people with and without HIV. Recently starting antiretroviral therapy (ART) predicted BMD loss, but BMD remained relatively stable with continuing ART.
Prevalent low BMD in HIV populations may reflect classic risk factors, ART and the ongoing inflammation of HIV infection itself. Randomized trials indicate that BMD falls 2% to 6% in the first two years of ART, regardless of regimen, but then bone density stabilizes. To get a better understanding of BMD changes over time in HIV-positive and negative adults, HIV UPBEAT investigators in Ireland conducted this study.
This three-year, single-site, observational cohort enrolled HIV-positive and negative people with similar demographic backgrounds from February 2011 through July 2012. At a baseline visit then annually, participants had dual x-ray absorptiometry (DXA) scans at the femoral neck, total hip and lumbar spine. To assess BMD rate of change with versus without HIV, the researchers compared BMD values at all three anatomical sites with mixed effects regression models.
The longitudinal analysis included 384 cohort members, 176 (46%) with HIV infection. Among all participants, 120 contributed two annual DXA scans and 264 contributed three. Median follow-up was 103 weeks. The HIV group was younger (median 39 versus 43 years, P = .04) and included a higher proportion of men (61% versus 46%, P = .003) and a higher proportion of persons with African ethnicity (39% versus 18%, P < .001). Most HIV-positive participants (88%) were taking ART, most (83%) were taking tenofovir disoproxil fumarate (TDF, Viread) and median treatment duration was 2.9 years. Three-quarters of the HIV group (78%) had a viral load at or below 40 copies/mL.
Compared with HIV-negative cohort members, those with HIV had significantly lower median BMD at the baseline measure at the femoral neck (median 1.024 versus 1.055 g/cm2, P = .003), total hip (1.061 versus 1.107 g/cm2, P < .003) and lumbar spine (1.164 versus 1.238 g/cm2, P = .001). Through three years of follow-up, BMD declined in both the HIV-positive and negative groups with no significant difference between groups in analyses adjusted for age, sex, ethnicity, smoking and body mass index. In the HIV-positive group, annual BMD declines were significant at the femoral neck (-0.0063 g/cm2 per year, P = .001) and total hip (-0.0044 g/cm2 per year, P < .001) but not the lumbar spine (-0.0024 g/cm2 per year, P = .25). In the HIV-negative group, annual BMD declines were also significant at the femoral neck and total hip but not the lumbar spine.
Multivariable mixed effects models including only people with HIV determined that age over 30 years was independently associated with greater BMD decline at the femoral neck (-0.013 g/cm2 per year, P = .02), as was Caucasian race (-0.011 g/cm2 per year, P = .006).
The multivariable models did not link BMD change over time to use of TDF or protease inhibitors. Starting ART within three months of enrollment predicted greater BMD loss at all anatomical sites (all P < .001). But not taking ART during follow-up predicted greater BMD loss. Specifically, compared with being on continuous ART or starting ART more than three months before entering the cohort, starting ART less than three months before or after enrolment and not being on ART during study follow-up were independently associated with greater BMD loss at the femoral neck (-0.016 g/cm2 per year, P = .002; and -0.019 g/cm2 per year, P = .012).
In summary, the researchers explained, starting ART in the previous three months was the only factor their multivariate model tied to greater BMD loss at all three sites. They found "no evidence to support an effect of ART itself, or specific components such as TDF or protease inhibitor, in accelerating bone loss." These findings are consistent with previous research indicating that most BMD loss with ART occurs within the first year of treatment. Furthermore, although HIV-positive people had lower BMD than HIV-negative people when follow-up began, BMD decline over time did not differ between people with or without HIV.