by Mark Mascolini, National AIDS Treatment Advocacy Project
7th International AIDS Society Conference, Kuala Lumpur, June 2013:
Each additional antiretroviral therapy (ART) regimen taken was independently associated with 0.011 g/cm(2) lower femoral neck bone mineral density (BMD) and 0.015 g/cm(2) lower lumbar spine BMD in a prospective study of 210 people with HIV infection [1]. Current or cumulative ART exposure, including tenofovir, did not affect BMD in these analyses. But in other analyses current ART and cumulative protease inhibitor (PI) use were linked to lower BMD.
In people with HIV infection, BMD declines most when they first start ART or reach an undetectable viral load with first- or second-line ART. Research also links use of tenofovir or PIs to bigger drops in BMD. A team at two Dublin institutions hypothesized that cumulative use of tenofovir, PIs, or ART would lower BMD in their analysis of a prospective cohort, Understanding the Pathology of Bone Disease in HIV-Infected Subjects (UPBEAT).
UPBEAT enrolls HIV-positive and negative people with similar demographic backgrounds. This baseline HIV-specific analysis aimed to identify independent predictors of BMD change at the femoral neck and lumbar spine, considering the following HIV-related factors: current and nadir CD4 count, viral load below or above 40 copies, injection drug use (IDU) versus non-IDU, current and cumulative use of ART, tenofovir, PIs, and nonnucleosides, number of different antiretroviral regimens, and time since HIV diagnosis. Initial multivariable linear regression models adjusted for age, gender, ethnicity, education level, and body mass index. The final model also adjusted for undetectable HIV RNA.
Among 474 people recruited from February 2011 to July 2012, 210 (44%) had HIV infection. HIV-positive people had a median age of 39 (interquartile range [IQR] 33 to 46) and a median body mass index of 26 kg/m(2) (IQR 23 to 30). Men made up more than half of the HIV-positive group (58.6%), while 34.8% were current smokers and 46.2% had completed third-level education. Whites accounted for 60.5% of the study population and blacks for 39.5%.
Half of the study group (51%) acquired HIV during heterosexual sex, 29.1% during sex between men, and 18.6% while injecting drugs. Median nadir and current CD4 counts were 212 and 471, and median time since HIV diagnosis was 5 years (IQR 2 to 8). Only 19 people with HIV (9%) had never taken antiretrovirals. Among treated people, median treatment duration was 2.9 years (IQR 0.6 to 5.10). Almost three quarters (71.4%) of currently treated people had taken tenofovir, while fewer than half (41.9%) had taken a PI.
The model adjusted for age, gender, ethnicity, education, and body mass index identified associations between femoral neck BMD and injection drug use (-0.013 g/cm(2), P = 0.02), every 10-cell higher baseline CD4 count (+0.0001 g/cm(2), P = 0.05), every 10-cell higher nadir CD4 count (+0.001 g/cm(2), P = 0.04), viral load below versus above 40 copies (-0.055 g/cm(2), P = 0.01), current ART versus no current ART (-0.058 g/cm(2), P = 0.05), each additional antiretroviral regimen (-0.013 g/cm(2), P = 0.02), and each additional year of PI exposure (-0.007 g/cm(2), P = 0.05), but not current or cumulative tenofovir exposure.
When the UPPEAT team further adjusted the model to include undetectable viral load, only number of antiretroviral regimens remained associated with BMD, with each additional regimen linked to 0.011 g/cm(2) lower BMD (95% confidence interval -0.022 to -0.0003, P = 0.05). No antiretroviral-related variable, including current or cumulative tenofovir or PI use, was associated with BMD in this analysis. Further adjustment for cumulative antiretroviral exposure or time since HIV diagnosis did not alter the association with number of antiretroviral regimen. Median femoral neck BMD dropped from nearly 1.1 g/cm(2) with no antiretroviral regimens, to about 1.0 g/cm(2) with 2 regimens, to between 0.9 and 1.0 g/cm(2) with 4 or more regimens.
The primary analysis also linked each additional antiretroviral regimen to lower lumbar spine BMD (-0.015 g/cm(2), P = 0.03). But in this analysis longer time since HIV diagnosis and injection drug use were also associated with lower lumbar spine BMD, while nonnucleoside regimens were linked to higher spine BMD.
The researchers suggested the association between number of regimens and lower BMD is “consistent with data demonstrating greater reduction in BMD associated with ART-induced HIV RNA suppression occurring with multiple ART regimens.” To further characterize potential mechanisms, they plan analyses focused on the relationship between BMD and number of virologic suppressions or rebounds.
ABSTRACT:
Cotter A, Sabin CA, Simelane S, et al. Number of different antiretroviral regimens rather than cumulative exposure to antiretrovirals associated with lower bone mineral density in HIV-positive subjects. 7th IAS Conference on HIV Pathogenesis, Treatment and Prevention, June 30-July 3, 2013, Kuala Lumpur. Abstract MOPE078.