HIV: An underrecognized secondary cause of osteoporosis?
Article first published online: 21 MAY 2013
Copyright © 2013 American Society for Bone and Mineral Research
Journal of Bone and Mineral Research
Volume 28, Issue 6, pages 1256–1258, June 2013
How to Cite
Brown, T. T. (2013), HIV: An underrecognized secondary cause of osteoporosis?. J Bone Miner Res, 28: 1256–1258. doi: 10.1002/jbmr.1967
- Issue published online: 21 MAY 2013
- Article first published online: 21 MAY 2013
- Accepted manuscript online: 29 APR 2013 08:30AM EST
With the advent of effective antiretroviral therapy (ART) more than 15 years ago, HIV has been transformed from a fatal illness into a manageable chronic condition and the HIV-infected population is aging into their 60s and 70s and beyond. There is concern, however, that the burden of comorbid conditions normally seen with increased prevalence with aging, such as cardiovascular disease, chronic obstructive pulmonary disease, and neurocognitive decline, is higher among HIV-infected populations and that these conditions, rather than HIV infection, pose the greatest threat to longevity and quality of life for HIV-infected patients.
Since early in the era of effective ART (sometimes referred to as highly active antiretroviral therapy [HAART]), a higher prevalence of low bone mineral density (BMD) and osteoporosis has been described among HIV-infected persons compared to their HIV-uninfected peers. There are fewer data, however, comparing fracture rates in HIV-infected patients to HIV-uninfected controls. In this issue of the Journal of Bone and Mineral Research, Güerri-Fernandez and colleagues compare fracture rates in HIV-infected and HIV-uninfected persons over a 2-year period (2007–2009) in over 1 million patients in Catalonia, Spain, representing 30% of the Spanish population. The major findings were that HIV-infected persons had over a fourfold increased risk of incident hip fracture (hazard ratio [HR], 4.72; 95% confidence interval [CI], 2.35–9.47) compared to the HIV-uninfected patients after multivariable adjustment. Similar results, albeit of a lesser magnitude, were observed for non-hip fractures (HR, 1.63; 95% CI, 1.12–2.32) or all clinical fractures (HR, 1.75; 95% CI, 1.24–2.48). This study joins other population-based studies in Denmark and the United States which have also showed that HIV-infected persons have a higher than expected risk of fracture at sites generally associated with osteoporosis.[2-5]
One of the critical issues in interpreting these results is understanding the composition of the HIV-uninfected control group. As might be predicted, the HIV-infected group differed from the control group in many ways that would be expected to increase the risk of osteoporosis and fracture, including lower body mass index (BMI), higher amount of smoking, heavy alcohol use, and liver disease. In addition, some factors that might be expected to be associated with a lower fracture risk, such as younger age and male gender, were also overrepresented in the HIV-infected persons. Although the authors quite appropriately adjusted for these confounders, it is possible that the results could have been affected by the presence of unmeasured confounders (such as opiate use, differences in nutritional status, or antidepressant use) or incomplete adjustment of known confounders (ie, residual confounding). For example, liver disease was in part accounted for by the use of the Charlson comorbidity index as a covariate, which includes the presence of liver disease and its severity. It is unclear, however, how well this index accounted for hepatitis C infection, which is highly prevalent in HIV-infected populations and is a strong and consistent predictor of osteoporosis and fracture in prior studies. Nevertheless, this study makes an important contribution to the literature regarding fracture risk in HIV-infected persons and leads to the question of whether strategies to prevent fractures in HIV-infected persons, such as screening policies or treatment thresholds, should be different than in the general population. In the U.S. National Osteoporosis Foundation 2008 guidelines, for example, HIV is not listed among those conditions that would prompt BMD screening in men between 50 and 70 years old and in postmenopausal women who are less than 65 years old. This has potential implications not only for fracture prevention but for insurance coverage for BMD testing.
If HIV is associated with a higher risk of fracture, what are the underlying causes? How much is attributable to traditional osteoporosis risk factors, such as smoking, alcohol use, low BMI, opiate use, and hypogonadism, which are found in higher prevalence in HIV-infected populations? How much does chronic HIV infection itself and its consequent increase in systemic inflammation contribute? How much is attributable to ART? Clearly, more work is needed to disentangle these factors. The study by Güerri-Fernandez and colleagues did not examine the contributions of HIV disease-related factors (eg, CD4 cell count, HIV duration, or controlled HIV replication) or specific antiretroviral agents on fracture in the HIV-infected population, but this is an important area of further inquiry. It is particularly difficult to ascertain the effect of untreated HIV infection on fracture risk, given that the various iterations of HIV treatment guidelines over the past 2 decades have recommended ART at progressively higher CD4 cell counts, thus limiting the pool of untreated HIV-infected persons.
Regarding the potential effect of ART, there are many reasons to hypothesize that initiation of ART would lead to improved bone health. Markers of systemic inflammation, such as tumor necrosis factor α (TNF-α) and interleukin 6 (IL-6), which are associated with osteoporosis and fracture in other populations, drop significantly with ART initiation. Lean body mass, which is generally associated with higher BMD and decreased fracture risk, increases when ART is started. Testosterone concentrations increase when men start ART. Despite these changes, somewhat paradoxically, BMD drops approximately 2% to 6% over the first 1 to 2 years of ART, but then remains stable thereafter and generally does not return to baseline.[6-8] This finding has been consistent across studies and is independent of the specific ART regimen used. This bone loss is associated with marked increases in bone turnover in the first 6 months of ART, with markers of bone resorption rising earlier and higher than markers of bone formation, creating a “catabolic window.”
If ART initiation is bad for the bone, would interruption of ART improve bone health? The study by Hoy and colleagues in this issue of the JBMR uses unique data from a randomized controlled trial to address this question. In the mid-2000s, there was concern that the toxicities of ART were a major limitation to care, and by giving ART only when the CD4 cell count was low and stopping it when the CD4 cell count drifted above some “safe” level, some of the ART-related toxicities and costs could be minimized without the patient developing opportunistic infections and other complications of immunosuppression. The Strategies for Management of Antiretroviral Therapy (SMART) study addressed this question directly by randomizing over 5400 patients to either delayed initiation/CD4 cell-guided interruption (starting ART if CD4 cell count dropped below 250 cells/mL, and stopping once it reached over 350 cells/mL) or continuous ART. This landmark study not only showed that CD4-guided interruption was associated with more AIDS events, but somewhat surprisingly, there was an excess risk of non–AIDS-related events, including cardiovascular disease. Largely on the basis of this study, the practice of structured treatment interruption has been abandoned. Once a person starts on ART, he should continue.
The design of SMART offered an excellent opportunity to understand the effect of ART interruption on bone metabolism. In a previous publication, SMART substudy investigators showed that, in contrast to those who were taking continuous ART who showed progressive declines in BMD, those who interrupted or deferred ART showed no decrease in BMD or, in the case of the spine, a statistically significant uptick in BMD. The study by Hoy and colleagues expands on these findings by measuring markers of bone turnover (cross-linked C-telopeptide [CTX], N-terminal propeptide of type I collagen [P1NP]), bone regulation (osteoprotegerin [OPG], receptor activator of NF-κB ligand [RANKL]), and systemic inflammation to determine the effect of ART interruption on these markers and whether early changes in this markers could predict future decreases in BMD.
Mirroring the effects of ART initiation, ART interruption was associated with decreases in bone turnover markers and increases in inflammatory markers. RANKL, which generally decreases with ART initiation, was shown to increase with ART interruption. Furthermore, changes these markers (larger decreases in bone turnover markers and larger increases in RANKL) by month 4 predicted increases in BMD at month 12. The clinical implications of these findings are unclear, because ART interruption is not a therapeutic option given the clear benefits of ART on both AIDS and non-AIDS events. However, this study does add important information regarding the interaction between ART and bone metabolism, providing evidence to support the paradoxical negative effect of ART on bone metabolism.
Why, then, is ART bad for the bone? There is no question that certain ART medications have specific negative effects on bone metabolism. Tenofovir, one of the most popular HIV medications worldwide, has been shown to decrease BMD by about 1% to 2% versus comparators in randomized controlled trials and to increase the risk of fracture in a large observational study.[6, 8, 12] Certain protease inhibitors have been shown to decrease BMD (atazanavir/ritonavir) and have been associated with an increased risk of fracture (lopinavir/ritonavir).[8, 12] The underlying mechanisms for these drug effects on bone metabolism have not been elucidated.
However, the decreases in BMD with ART initiation and the stabilization of BMD with ART interruption appear to be more fundamentally related to immune changes with treatment onset and treatment offset, rather than specific effects of one or more medications. The precise players involved have yet to be identified. Is it the T cell, the primary target of HIV, which has shown to be essential in elegant models of estrogen deficiency and parathyroid hormone–induced bone loss? Is it the B cell, whose dysfunction in HIV infection is known and whose role in HIV-related bone loss has been proposed by Ofotokun and colleagues? Are there as-yet-described mediators that protect bone in the setting of untreated HIV infection? By addressing these questions, we can better understand how HIV and its treatment may affect bone, which may have implications for other osteoimmune interactions.
For now, clinicians should be aware of the increased fracture risk in HIV populations and work toward implementing measures for primary prevention among older HIV-infected patients, including behavioral changes (such as smoking cessation, alcohol reduction, and weight bearing exercise), identification of contributing comorbid conditions (eg, hypogonadism), fall risk assessment, and BMD screening. In patients with established osteoporosis, consideration should be given to switching to an ART regimen with the least bone toxicity. Until more data become available, treatment guidelines should follow those established for the general population. The success of ART has resulted in a “graying” of the HIV population. Primary prevention measures to reduce fracture risk will help ensure that HIV-infected persons are graying successfully.
Dr Brown has served as a consultant for BMS, GSK, Merck, Abbott, Gilead, ViiV Healthcare, and has received research funding from Merck and GSK.
- 8Bone mineral density and fractures in antiretroviral-naive persons randomized to receive abacavir-lamivudine or tenofovir disoproxil fumarate-emtricitabine along with efavirenz or atazanavir-ritonavir: AIDS Clinical Trials Group A5224s, a substudy of ACTG A5202. J Infect Dis. 2011; 203:1791–801., , , , , , , , ,