icon-    folder.gif   Conference Reports for NATAP  
  20th Conference on Retroviruses and
Opportunistic Infections
Atlanta, GA March 3 - 6, 2013
Back grey_arrow_rt.gif
Update from 20th CROI: Bones, Vitamin D, Frailty
  Todd T. Brown, MD, PhD
Association Professor of Medicine and Epidemiology
Division of Endocrinology and Metabolism
Johns Hopkins University
At the 20th CROI, the center stage for co-morbidities was definitely given to cardiovascular disease (see excellent NATAP review by David Shepp). Nevertheless, the studies investigating osteoporosis, frailty, and vitamin D were relatively limited in number, but generally excellent in in quality. Here, I'll review some of these studies with a particular emphasis on those with potential clinical relevance.
CROI 2013: Cardiovascular Disease and Co-morbidities in HIV - written by David H. Shepp, MD Associate Professor of Medicine, Hofstra North Shore-LIJ School of Medicine North Shore University Hospital - Manhasset, NY (03/11/13)
Osteoporosis and Fragility Fracture: Extent of the Problem and Mechanisms
Osteoporosis: There have been multiple studies that compare bone mineral density (BMD) in HIV+ and HIV- individuals and have generally shown lower BMD in HIV-infected persons. One of the limitations of the some of the previous studies has been the HIV-uninfected control groups, which may differ in other ways from the HIV-infected groups and may introduce bias in the comparison. To overcome this limitation, the HIV UPBEAT (Understanding the Pathology of Bone Disease in HIV-infected Subjects) recruited 210 HIV-infected subjects and 264 HIV-uninfected controls from similar demographic backgrounds (Poster #817). This Irish cohort is primarily young, male (59% HIV+ vs 44% HIV-), and Caucasian (HIV+ 61% vs HIV+ 75%). Importantly, body mass index was similar by HIV-serostatus (26 vs 27 kg/m2). BMD was found to be lower in the HIV+ group at the lumbar spine, total hip and femoral neck after adjustment for demographic factors and BMI. Low BMD (Z-score < -2) was about 2 fold higher in the HIV+ group across the sites. BMD was not related to vitamin D levels or PTH. However, alkaline phosphatase, whose skeletal derived fraction is a measure of bone turnover, was higher in HIV+ subjects and partially, but not completely, explained the difference in BMD by serostatus. This study confirms previous findings regarding low BMD in HIV+ populations. We will undoubtedly be seeing more data coming out of this rich cohort regarding the effect of antiretroviral therapy, chronic inflammation/immune activation, and other factors which may help to explain the differences by HIV-status.
While the UPBEAT study showed a difference in alkaline phosphatase by HIV serostatus which was related to BMD, the role of these biomarkers is clinical practice is unclear. Since these markers are related to bone turnover, it has been hypothesized that the use of these markers may be able to identify a population that is at higher risk of bone loss (so called, "fast losers"). In the Menopause Study (#819), Sharma and colleagues reported that baseline levels of bone specific alkaline phosphatase, osteocalcin, and c-telopeptide did not predict changes in BMD over about 2 years in 246 HIV+ and 219 HIV- women. One of the limitations to these markers, which has been a problem for their clinical use, is the high degree of analytic and biological variability. Nevertheless, this study would suggest, that even on a population basis, these markers are not useful to predict bone loss, at least in this study. It should be mentioned that that this cohort was different from the UPBEAT study in that bone specific alkaline phosphatase concentrations were similar by serostatus, but osteocalcin was higher in the HIV+ women. In contrast, c-telopeptide, a marker of bone resorption (not reported in UPBEAT) was actually higher in the HIV-uninfected women.
Bone Turnover in HIV+ and HIV-negative: One of the potential pathways to explain difference in bone turnover and BMD by serostatus is alteration in osteoprotegerin (OPG) and receptor activator of NFαB Ligand (RANKL). Secreted RANKL binds to RANK on the cell surface of osteoclast precursors, leading to osteoclast activation and bone resorption. OPG is also secreted to bind to RANKL, thereby preventing the interaction of RANK and RANKL and slowing bone resorption. These markers were first described as being products of osteoblasts and governing communication between osteoblasts and osteoclasts. However, it's now clear that other cells produce both OPG and RANKL, including B cells. It has been hypothesized that B cell dysregulation of OPG and RANKL production may play a role in HIV-related bone loss, as was suggested by the HIV-1 transgenic rat (Vikulina, PNAS, 2010). In a follow-up to this work, Titanji and the Emory team (#821) showed that markers of bone resorption (CTX) were higher in untreated HIV+ persons compared to HIV-negative controls, though osteocalcin, a marker of bone formation was similar. Of note, compared to the HIV-negative controls, intracellular B-cell expression of OPG was lower in HIV+ persons, but RANKL expression was higher. These findings suggest that the ratio of these two factors may be dysregulated in HIV-infection, which may lead to osteoclast activation, bone resorption, and bone loss, but this was not specifically investigated in this study. Interestingly, B cell expression of these markers may be different than what is measured in the peripheral blood. In the MACS Cardiovascular substudy, we found that plasma concentrations of OPG were higher and RANKL concentrations were lower in HIV-infected men, compared to HIV-uninfected men (#792). It should be noted that the HIV+ men in the MACS were well-suppressed on ART and this may affect OPG and RANKL levels. Also, we did not measure B-cell expression of these markers and the Emory study did not measure peripheral levels, so it's difficult to make direct comparisons. What's clear however it that the OPG/RANKL system is altered in HIV-infected persons which may have consequences for bone health. It's a complicated system however and care is needed when interpreting results.
Fracture: THE Osteoporosis Outcome:
When we're talking about bones, there's really only one outcome: fracture. Over the last few years there has been a steady stream of published or presented studies that have shown, with a few exceptions, that fracture rates in HIV-infected persons is higher than HIV-uninfected persons. Warriner and colleagues (#820) used a Medicare database to compare fracture rates (hip, spine, femur, pelvis, tibia/fibula, ankle, clavicle, humerus, radius/ulna, carpal bones, foot, rib) in HIV-infected and HIV-uninfected persons between 1999-2010. In persons < 65 and those > 65 years (total n > 2M) , fracture rates were about 50% higher in HIV+ vs HIV- after adjustment for age, gender, year, U.S. region, race, fracture in baseline, Hepatits C, and comorbidity (< 65 y: RR 1.55 (95% CI: 1.44 , 1.67 ) ; >65 y: 1.52 (95% CI: 1.34 , 1.73 )). These data provide further evidence that HIV-infected persons are more susceptible to fracture than their HIV-uninfected counterparts and may suggest that screening should be more aggressive than in the general population. It would be interesting to know the breakdown of fracture types, particularly whether the same results were seen for hip fracture alone, which has the most devastating consequences in terms of both morbidity and mortality.
Don't Forget Occult Spine Fractures! Over half of spine fractures are clinically silent. However, the presence of these fractures does indeed markedly increase the risk of a subsequent clinical fractures. In a comparison of 175 HIV+ and 120 age-matched HIV-uninfected patients (#822), 30% of the HIV+ had subclinical (or morphometric) vertebral fractures vs 4% of the HIV-uninfected controls. Notably, the prevalence of morphometric fracture among HIV+ was as common in those with osteopenia as in those with osteoporosis. This finding has important clinical implications. When assessing fracture risk and potential for subsequent bone treatment in those with osteopenia, plain films of the thoracic and lumbar spine (or Vertebral Fracture Assessment if available on your DXA machine) are an important tool, in that the presence of a morphometric fracture may push you to treat. Given the high prevalence of these fractures in this HIV population, this simple testing may be quite high yield.
Effect of ART Initiation on Bone: From multiple studies over the past 10 years, we know that the period after ART initiation is particularly bad for the bone. Two ACTG studies looked at this. In a combined analysis of 3 large ACTG trials (#823), we found that lower CD4 cell count at ART initiation (but not the change in CD4 cell count) was associated with a steeper drop in total BMD in the first 96 weeks on ART. This effect was particularly strong in those with a baseline CD4 cell count < 50, but was also seen in those with a CD4 between 50 and 500 (vs > 500). These finding suggest that initiation of ART at a higher CD4 cell count may attenuate ART-related bone loss. There were a few other interesting findings. Not surprisingly, TDF use was associated with a steeper BMD drop (see below), but PI use also showed a significant association. Previous investigations of PI vs non-PI effects on bone have showed differences in the lumbar spine, but not at the total body. Our findings suggest that the total body is vulnerable too, although the effect size is smaller. The total body is about 80% cortical bone and 20% trabecular bone, whereas the spine is just the opposite. Whether PIs particularly affect trabecular bone and what the mechanisms are requires clarification.
One of the paradoxes about the bone loss with ART initiation has been that this bone loss occurs at a time when lean mass is increasing, which should be generally good for bone. What's more is that lean mass increases the most in those who are the sickest at baseline (i.e those with the lowest CD4 cell counts). To assess the relationship between lean mass and BMD with ART initiation, Erlandson et al (#826) used the ACTG 5224s data to show that: 1) lean mass increased in all arms of the study (though ATV/r > EFV), 2) lean mass increases over 96 weeks were associated with lower CD4 cell count and higher HIV-RNA, 3) smaller increases in lean body mass were associated with a greater decrease in BMD over 96 weeks in the hip, but not the spine. These findings suggest that if lean mass could be maximally increased with ART initiation, the decrease in BMD with ART initiation could be attenuated.
Specific ART Effects on Bone: The Tenofovir Story Gets More Interesting: It has been shown that TDF has negative effects on BMD and has been associated with an increased risk of fracture, at least in the VA cohort (Bedimo, AIDS, 2012). One of the big management questions is whether a TDF-treated person with low BMD would benefit from switching off. Last year at CROI, Bloch et al showed in a single arm study that switching from TDF to raltegravir was associated with a significant increase in BMD. Although there was no comparison arm, these findings suggested that the switch to RAL may be a good strategy for those with known low BMD on TDF. Negredo et al (#824) presented the results of a small, but interesting RCT assessing the effect of switching off of TDF to a regimen of ABC (3rd drug was mixed 30% PI, 64% NNRTI). It showed that those who switched to ABC had an increase in the femoral neck of 1.98% over 48 weeks vs no change in the TDF group (p=0.09). At the spine, there was no change in the ABC group, but a significant decrease in the TDF group (p=0.05). Taken together, these findings suggest that a switch off TDF may improve BMD in those with osteopenia or osteoporosis. . Given the questions regarding the long term safety of bisphosphonates, the first line osteoporosis treatment, a switch off of TDF may represent a reasonable management strategy for those with T-scores around -2.5 without a previous fracture. Clearly, further data are needed.
Perhaps the biggest bone story at the meeting had to do with tenofovir alafenamide (TAF), a tenofivir prodrug that is activated in lymphoid cells by cathepsin A. Compared to TDF, peripheral concentrations of tenofovir are much lower with TDF which had been hypothesized to have fewer renal and bone toxicities. In a Phase 2 trial, TAF was compared to TDF (along with elvitegravir, cobicistat, and emtricitabine) in ART-naïve, HIV-infected patients. In general, the drug was effective and well tolerated. From a bone perspective, TAF clearly outperformed TDF over 24 weeks with -0.8% decrease in the spine (vs -2.5% decrease for TDF) and a -0.3% decrease in the hip (vs a -2.0% decrease for TDF). While these findings appear promising, it will be good to get longer term data and also get a better understanding of the tenofovir concentrations in osteoblast and osteoclasts and their precursors. Given the switch discussion above, it will be important to investigate the effect of switching to TAF in those who have low BMD on TDF.
Bones in Kids and Adolescents:
Lower peak bone mass (bone mass at about age 25-30) is highly predictive of osteoporosis and fragility fracture later on in life. It's critical therefore to understand whether bone health is compromised in HIV+ children and adolescents and determine the risk factors for lower BMD, such that preventative measures can be implemented. Two studies at CROI examined this question. Yin et al (#818) conducted a cross-sectional study of 30 HIV+ and 15 HIV-uninfected young men (ages 20-25 years). Of the HIV-uninfected men, 15 perinatally-infected and 15 were infected during adolescence; all were ART-treated. All men underwent conventional DXA scanning to determine areal BMD. At all sites investigated (spine, hip, radius), the HIV-infected men had Z-scores that were 0.4-1.3 lower than the HIV-uninfected group. HIV-infected men also had higher levels of CTX, a marker of bone resorption. It should be noted that the groups were well matched for BMI and fat mass, important potential confounders for BMD.
All study participants also underwent more specialized assessment of bone health. Conventional DXA is a good measurement, but it explains only about 50% of fracture risk. Part of this limitation is related to skeletal characteristics that also lead to bone fragility that conventional DXA does not measure. To overcome these limitations, new imaging techniques have been developed. High resolution peripheral CT scanning (XtremeCT ) is one such technique which measures not only bone mineral density, but can also resolve differences between cortical and trabecular bone and give detailed information regarding bone microstructure. These later measurements are thought to assess bone strength, independent of bone density. In the Yin study, HIV+ men also had lower trabecular BMD (distal radius and tibia) and lower cortical thickness. In addition, bone strength parameters, such as trabecular number, thickness, separation and stiffness also differed by HIV-serostatus. Interestingly, there were no differences between the two HIV-infected groups, despite much longer ART duration in those who were perinatally infected. Taken together, these findings suggest that bone density and bone strength are compromised in HIV-infected men nearing peak bone mass which has important health implications for this group. Whether similar results are seen in young HIV-infected women is not known.
Examining the factors which lead to lower BMD in HIV-infected children and adolescents is also critical. Jimenez (#964) examined BMD in 24 perinatally HIV-infected children and adolescents. Low BMD (Z-score < -1) was found in 38% of subjects. Low BMD was associated with lower nadir CD4 cell count and longer time with uncontrolled viremia, but not cellular markers of immune activation/senescence or serum markers of inflammation (IL-6, hsCRP) or monocyte activation. Tenofovir exposure was not associated with low BMD.
Vitamin D at CROI: Compared to previous years, this year's CROI was relatively vitamin D deficient. There were only a handful of studies that examined risk factors for vitamin D deficiency, associations between vitamin D deficiency and outcomes, and the effects of vitamin D supplementation. In the SMART study, BMD declined in those who continued on ART, but stabilized or increased in those who were randomized to CD4-guided treatment interruption. The mechanisms underlying this difference are unclear. Given that ART may affect vitamin D metabolism (true for EFV, but speculated for PIs), the SMART investigators (#827) examined whether the ART-interruption group differed from the continuous ART group in the change in 25OHD over the first year of the study and whether these differences in vitamin D could account for the differences in BMD observed between the two groups. Compared to the continuous ART group, 25D increased in the ART-interruption group (difference of 4.8 ng/mL), driven by those who were using ART at baseline and then stopped. Similar findings were observed for 1,25 OH2 D. Interestingly, PI use, rather than EFV use, at baseline showed significant correlations with the change in 25OHD. The increase in 25OHD in the ART-interruption arm was associated with increases in BMD, but the decrease in 25OHD was not related to the decrease in BMD in the continuous ART group. These findings regarding the effect of ART interruption on 25OHD levels are interesting, but it's difficult to say for sure if they changes are indeed driving the differences in BMD. The PI effect is novel and raises the question about the effect of PI use offset on vitamin D metabolism and whether it's related to the effect of the PI per se or if the effect is mediated by other factors also related to the offset of PI use. As the investigators point out, SMART was not designed to look at effects of specific drugs. Switch studies, in those who have suppressed viremia, may be useful to isolate a specific PI effect on vitamin D metabolism.
In a previous study, low vitamin D levels in HIV-infected mothers in Tanzania were associated were associated with increased adverse heath outcomes in in HIV-exposed infants, suggesting that vitamin D supplementation might be able to decrease these outcomes (Finkelstein, Pediatric Infectious Disease Journal, 2012). These women were not treated with ART however. In a case-control study of infants born to HIV-infected mothers who initiated ART in pregnancy (#920), Powis reported that low maternal vitamin D levels were not associated in with adverse health outcomes (infant wasting, stunting, or underweight) at one-year of life. These findings suggest that low maternal vitamin D status may not be a driver of poor infant health outcomes, if the mother is ART-treated.
To really prove that vitamin D status is the cause of adverse health outcomes, randomized supplementation trials are needed. No large scale trials have been done to date in HIV infection, but smaller pilot studies are emerging. In another study from Botswana (#965), 60 HIV infected children were randomized to receive D3 4000 IU/d or 7000 IU/d over 12 weeks. 25OHD levels increased similarly in both groups (4000 IU/d: 37 to 55 ng/mL vs 7000 IU/d : 35 to 57 ng/mL). There was a suggestion of improved growth status (weight Z score in the 4000 IU/d).
Frailty Studies at CROI: Frailty is an important geriatric syndrome that is associated with morbidity and mortality in older individuals. Some studies have suggested that older HIV-infected individual have an earlier occurrence of the frailty phenotype compared to their HIV-uninfected peers. The factors underlying this effect are unclear. In the Multicenter AIDS Cohort Study (MACS), HIV-infected men who had the frailty phenotype (3 of the following 5 characteristics: weight loss, low grip strength, exhaustion, slow walking speed, and low physical activity) had higher levels of IL-6, TNF-alpha, sTNFRII, sCD14, and T cell activation, compared to robust HIV-infected men, after multivariable adjustment. No such differences were seen between frail and robust HIV-uninfected men. Given the cross-sectional nature of the study, it's difficult to know whether the increase inflammation led to the frailty status, but the findings should spur further longitudinal investigation.
One potential source of systemic inflammation in frail patients is chronic infection with CMV. In a case-control study, Erlandson (#802) found that HIV-infected persons with impaired physical function had elevated CMV IgG compared to those with normal function, adjusted for CD4 cell count, age, and smoking. CMV IgG was associated with IL-6, but not hsCRP, TNF-alpha, CD8/CD38+/DR+ % or sCD14. These intriguing cross-sectional findings require confirmation in longitudinal studies to establish temporality and may pave the way for interventions to decrease CMV reactivation.
The frailty syndrome may be associated with cognitive dysfunction. In the MACS (#444), those with HIV Associated Neurocognitive Disorder had an increased risk of developing the frailty phenotype (Odds Ratio 2.2, 95% CI: 1.1, 4.9) after adjustment for HIV duration, CD4 cell count, detectable HIV-RNA, and HAART use. Impairment in all cognitive domains (speed, motor function, executive function, working memory, memory and learning) were associated with incident frailty.
One common geriatric outcome, related to frailty, is falling. Falls are a major cause of morbidity and HIV-infected persons have multiple risk factors for falls, such as polypharmacy, neuropathy, cognitive decline, and low muscle mass. In a combined MACS/WIHS substudy (#801), we found that middle age HIV-infected men and women had a similar number of self-reported falls in the previous 12 months compared to their uninfected peers. Interestingly, clinical symptoms (feeling off balance, floating, lightheadedness, faintness) were more common in HIV-infected men and women and strongly associated with falls. If persistent, these symptoms may put HIV-infected patients at increased risk of falls with aging.