icon-    folder.gif   Conference Reports for NATAP  
 
  18th CROI
Conference on Retroviruses
and Opportunistic Infections
Boston, MA
February 27 - March 2, 2011
Back grey_arrow_rt.gif
 
 
 
Bone and Vitamin D at CROI 2011
 
 
  By Todd T. Brown, MD, PhD
 
HIGHLIGHTS:
- HCV coinfection increases fractures
- HAART/immune reconstitution & bone markers/bone mineral density
- In HIV-infected young males treated with ART, there may be a failure to achieve peak bone mass because in a study reported here & previously young men already have low bone mineral density (study reports here, associated with popper & amphetamine use)
 
[from Jules: young men infected with HIV have low BMD already plus it has been reported that after HAART initiation bone loss occurs in the first year so young men experience this double bone hit. Vitamin D levels testing & supplementation may be helpful when starting HAART or for that matter for anyone on HAART or with low vitamin D levels for sure]
 
- Tenofovir
- Switching to raltegravir was associated with increases in total body BMD, whereas those who remained on PI BMD showed no change over the 48 week interval. The between group differences were not significant.
- important to investigate the whether these potential beneficial effects on bone turnover with tesamorelin translate in improved BMD
- in HIV-infected patients with lipodystrophy, rosiglitazone was associated with decreases in the bone formation marker....pioglitazone likely has similar effect on bone metabolism....I don't believe that TZDs should be recommended for the treatment of lipoatrophy in those without diabetes mellitus. In those with diabetes and lipoatrophy, pioglitazone can be considered, but the risks and benefits need to be weighed carefully.
 
- In the ECHO trial (#79 LB), HIV-infected persons initiating a regimen of TDF/FTC/EFV had a median decrease of 2.2 ng/mL compared to those randomized to TDF/FTC/TMC278
 
Many questions remain about pathogenesis, clinical consequences, and optimal management of osteoporosis and vitamin D deficiency among HIV-infected patients. As evidenced by the increasing number and quality of abstracts at this year's CROI, issues related to bone and vitamin D are gaining prominence among HIV researchers. Although there were no major breakthroughs presented at this year's meeting, there were many studies which will help us to understand these complicated problems to improve the health of our patients. My goal here is to share some of the highlights from this year's CROI in the world of bone and vitamin D, particularly emphasizing those studies with clinical implications.
 
Fracture Risk:
 
One of the most consistent risk factors for fracture among HIV-infected patients is co-infection with HCV. A large retrospective study of US Medicare recipients used ICD-9 codes for fracture at the hip and spine to determine whether HIV/HCV co-infection was associated with an increase risk of fracture compared to either HIV or HCV alone (#914). The researchers used propensity scores to balance for other covariates across the populations. The major findings of this study were that HIV/HCV co-infection was associated with a higher risk of hip fracture compared to either HIV or HCV alone or HIV-uninfected persons. This was true for both men and women. At the spine, HIV/HCV co-infected men and women had a higher risk of fracture than HIV mono-infection or HIV-uninfected persons. The underlying causes of the increased risk of fracture with HIV/HCV co-infection are unclear. Possibilities include: 1) endocrine abnormalities related to liver dysfunction (low vitamin D, low IGF-1, low testosterone, increased catecholamines), 2) shared risk factors associated with both HCV and low bone density (eg, opiate use, compromised nutritional status), 3) enhanced immune activation in the setting of two chronic infections. Clearly, understanding the mechanisms underlying the link between HCV/HIV co-infection and fractures is going to be crucial in trying to reduce the fracture burden in this vulnerable population. As new and improved medications become available for the treatment of HCV, it will also be important to understand whether bone density increases with successful HCV treatment.
 
HCV was also a significant risk factor for fracture in an analysis of the ACTG Longitudinal Linked Randomized Trial (ALLRT) cohort (#830). This study examined all incident fractures in this group of HIV-infected persons who are followed at regular intervals after completion of an ACTG randomized clinical trial. Perhaps the most interesting finding from this study besides the confirmation of the importance of HCV co-infection was the observation that fractures were significantly more common in the first two years after ART initiation. As discussed below, ART-initiation is associated with a 2-6% decrease in BMD over the first 48-96 weeks of therapy regardless of the regimen started. These findings in ALLRT suggest that the decrease in BMD with ART initiation may be clinically relevant.
 
Bone Loss with ART Initiation: An Effect of Immune Recovery?: The bone loss described above with ART-initiation is accompanied by an acceleration of bone turnover, whose time course and mechanism have not been well-described. In a single arm ART initiation study of LPV/r/TDF/FTC (#78), bone turnover markers began to rise as early as two weeks after beginning ART, peaked at 12 weeks, and remained elevated after 48 weeks. Interestingly, this increase in bone turnover was associated with large increases in TNF-α and soluble receptor of NF-κB Ligand (RANKL), a factor secreted by activated T-cells, osteoblasts, and other cells of immune origin that leads to bone resorption by inducing osteoclast differentiation and activation. Accompanying this human study, the authors presented an elegant animal model of immune reconstitution (T-cell adoptive transfer into T-cell receptor null mice (TCR beta)) and similar to their human studies, found an increase in bone turnover, associated with increases in RANKL and TNF-α concentrations at early timepoints. These changes were also accompanied by profound decreases in BMD. Taken together, these data suggest that immune reconstitution plays a critical role in the loss of BMD with ART-initiation. It is interesting to note that prior studies investigating TNF-α activity with ART initiation have generally shown decreases rather than increases of these markers (usually soluble receptors of TNF-α). Similarly, RANKL has previously been shown to decrease with ART initiation. Factors accounting for these discrepancies require further investigation.
 
Another study investigating ART initiation also sought to determine the time course of bone turnover, in this case investigating both bone resorption (CTX) and bone formation (P1NP) at several timepoints after initiation of AZT/3TC/LPV/r vs NVP/LPV/r (MEDICLAS, #833). CTX rose quickly in both arms of the study and peaked by 3 months, whereas P1NP rose to a lesser extent and peaked later (at 12 months). These findings suggest that there is a period after ART initiation when bone resorption outstrips bone formation, creating a "resorptive window", and leading to net bone loss. Studies are ongoing to determine whether intervening during this window (eg bisphosphonates , vitamin D, etc) can attenuate the bone loss with ART-initiation.
 
Bone in Children and Adolescents: Peak bone mass is reached at approximately age 30 and is a major determinant of bone density in older adults. Kathy Mulligan presented a cross-sectional study of behaviorally acquired HIV-infected adolescent males enrolled in the Adolescent Trial Network (#705). Compared to HIV-uninfected controls, HIV-infected males (mean age 21 years) on ART had lower BMD and Z-scores, with a more pronounced effect in those receiving PIs. Consistent with the known effect of ART initiation on BMD, HIV-infected males na•ve to ART had similar to BMD to HIV-uninfected controls and higher BMD that the ART-treated, HIV-infected males. These data suggest that in this group of HIV-infected males treated with ART, there may be a failure to achieve peak bone mass. The magnitude of the differences was not small at some of the sites. For example, there was an 8% difference in total hip BMD between HIV-infected males receiving PIs and HIV-uninfected controls. It has been speculated that a 10% lower peak bone mass in the general population is associated with a doubling of fracture risk after age 50 (see Bonjour, Salud Publica de Mexico, 2009, for a nice review).
 
One strategy to improve peak bone mass in children and adolescents may be to supplement with vitamin D. Stephen Arpadi presented the results of a randomized clinical trial of oral cholecalciferol 100,000 units twice a month/Calcium (1 gram) vs placebo in 59 perinatally infected children and adolescents (mean age 10 years, 60% African American, 40% Hispanic) given over 2 years. While bone mineral density increased in both arms, there was no benefit of the vitamin D and calcium supplementation. The authors speculated that the lack of effect may have been a lack of adherence to calcium or failure to maintain 25 hydroxy vitamin D (25OHD) levels above 30 ng/mL. Based on these results, it seems that this vitamin D supplementation will not be a promising approach to help HIV-infected children reach their peak bone mass, at least in similar populations. Although vitamin D deficiency is extremely common in African-Americans, there is also evidence, at least in African-American adults in the general population , that the vitamin D deficiency is not associated with decreased BMD, despite higher PTH levels (see Hannan, JCEM, 2008, PMID: 17986641). The mechanisms regarding this observation are not clear, but PTH resistance in African-Americans with respect to bone has been speculated. The implications of this differential effect of race may mean that vitamin D supplementation could be less effective for bone outcomes in African-American populations. It also leads to the question of whether the definitions of vitamin D deficiency should be race-specific.
 
Effect of ART on Bone:
Tenofovir Disoproxil Fumarate: TDF exposure has been association with consistent, but modest (1-2%) decreases in BMD in randomized trials. The clinical significance of this effect is not yet understood, as there have been no adequately powered studies to determine the effect of TDF on fracture incidence. There were several studies that investigated TDF exposure with respect to bone metabolism. It had previously been reported that TDF was associated with increases in parathyroid hormone (TDF Package Insert). In a cross-sectional study from a single clinic in Spain (#824), patients receiving TDF had a higher PTH, lower serum calcium, and 1,25 dihydroxy vitamin D compared to those not receiving TDF. What was particularly interesting in this study was that these differences were only observed in those who had a 25OH D concentration < 25 ng/mL, although overall the TDF groups and non-TDF groups had similar 25OHD concentrations. These findings suggest that those with vitamin D deficiency receiving TDF may be at particular risk for abnormalities in mineral metabolism. It's not clear whether these changes are clinically relevant as BMD was not measured.
 
This independent effect of TDF on PTH was also seen in another cross-sectional study (#825) in which TDF use was associated with a 3.9- fold increase in the prevalence of hyperparathyroidism, independent of serum ionized calcium and 25OHD. The mechanisms underlying this effect on PTH require more understanding. If the PTH increase is unrelated to 25OHD, other possibilities may include a TDF-effect on calcium absorption or TDF-effect on the Ca++-sensing receptor on parathyroid cells. The big caveat with these studies is their cross-sectional nature. Further prospective studies are required to better understand these effects.
 
One way to better understand the potential interaction of vitamin D and TDF is to see if the replacement of vitamin D can decrease PTH and reduce bone turnover among TDF-treated persons. This is precisely what was investigated in a randomized, placebo controlled trial among HIV-infected adolescents on stable ART in the Adolescent Trial Network, presented by Peter Havens (#80). This was a 12-week trial of D3 50,000 units monthly vs placebo, with randomization stratified by the use of TDF. Overall, there were 118 who were receiving TDF (59 Vitamin D and 59 placebo) and 85 who were not receiving TDF (43 Vitamin D; 42 placebo). Consistent with the above cross-sectional studies, at baseline, TDF-treated subjects had a higher PTH compared to those not treated with TDF. 25OHD levels were similar between the groups. After 12 weeks of treatment, 95% of those randomized to vitamin D achieved adequate 25OH D levels (mean+SD: 35±19.1 ng/mL). In those not treated with TDF, there were no changes with vitamin D supplementation in any of the outcome measures (tubular resorption of phosphate, bone alkaline phophatase (BAP, a bone formation marker), c-telopeptide (CTX; a bone resorption marker), and PTH). However, in those treated with TDF, vitamin D supplementation was associated with a significant decrease in PTH and a trend towards a decrease in BAP. An interesting observation was that this effect on PTH with vitamin D was seen regardless of whether the subject started with adequate 25OHD. Given the short duration of the trial, BMD was not measured, so it's not clear whether vitamin D supplementation in TDF-treated persons will translate into improved BMD. From this study and those presented above, it appears that TDF is associated with higher PTH levels, the clinical significance of which is unclear. This differential effect of the PTH lowering effect of vitamin D supplementation in those receiving TDF requires further exploration.
 
The other big news regarding the effect of TDF on bone was the results of two PrEP studies in men who have sex with men (MSM). In the CDC PrEP study (#93), about 200 men were randomized to TDF or placebo and BMD was assessed at baseline, 9-12 months, and 24 months. Compared those randomized to placebo, BMD declined in the TDF-exposed men to greater extent at the femoral neck (net decrease -1.1%, p=0.004), total hip ( -0.8%, p=0.003), and a trend at the lumbar spine (-0.7%, p=0.11). Similar results were seen in the iPrEx substudy (#94LB). In this RCT, 503 MSM were randomized to FTC/TDF vs placebo and followed over 48-72 weeks. At 24 weeks, there was a -0.7 to -1.0% in BMD in the FTC/TDF arm. At 48 weeks, the differences increased at the total hip, but decreased at the lumbar spine. A few points should be raised about these consistent findings regarding TDF in HIV-negative persons:
 
· Effect size: The effect on BMD was modest in both studies. One concern is the extent to which poor drug compliance may have reduced TDF exposure, thereby attenuating the bone effects. Both studies are currently investigating this issue. Second, although the mean changes were relatively modest, there was a notable proportion who had a potentially clinically significant decrease in BMD with TDF exposure. In the iPrEx study, 9% in the FTC/TDF arm had a decrease >5% at the lumbar spine vs 3% in the placebo group, although there was no difference in the percentage with >5% loss at the total hip. In the CDC PrEP, 13% vs 6% had > 5% loss at the femoral neck (p=0.3), with similar proportions between the arms at the other sites. This is quite a young population who should have had stable BMD over the interval. I am particularly concerned about those "in the tail" of the distribution as these people are the most likely to have negative clinical outcomes, particularly since many have not reached their peak bone mass. More work needs to be done to determine what clinical, biochemical, and genetic factors can identify those who will go on to have a large decrease of BMD on TDF.
 
· Lack of Fractures: In both of these studies, the incidence of fractures was no different between the arms. Of course, neither study was not designed or powered to investigate fracture incidence.
 
· Possible Reversibility: The iPrEx study has an off-study drug component, but these data have not yet been presented. It will be very interesting to see whether BMD returns to baseline with TDF discontinuation (similar to what is seen after stopping Depo-Provera). A related issue that is whether each repeated exposure to TDF will have a negative effect on bone.
 
· Low baseline BMD: In both study populations, the baseline BMDs were unexpectedly low. In the CDC study, lower BMD was related to popper use and amphetamine use. Further work needs to be done in MSMs to understand to lower BMDs.
 
Tenofovir- Prodrug: It has been hypothesized that the toxicities related to TDF may be related in part to the high plasma levels required to suppress HIV replication. Results were presented of a dose-finding trial of GS7340 (#152LB), a tenofovir (TFV) pro-drug which is targeted to mononuclear cells and can achieve higher intracellular potency at lower plasma concentrations compared to TDF. This specificity depends on the presence of cathepsin A, an enzyme that is expressed by PBMCs and is responsible for the intracellular conversion of GS7340 to TFV. Indeed, the TFV plasma concentrations were 58-88% lower with GS7340 compared to TDF, whereas HIV RNA suppression was greater. When considering the potential bone effects, it should be noted that osteoclasts and osteoblasts share a common lineage with PBMCs and both types of bone cells express cathepsin A. Therefore, if the bone effects of TDF (and thus TFV) are mediated by direct effects on osteoblasts and osteoclasts, it is possible that GS9730 will have similar bone effects to TDF, potentially magnified by the higher intracellular concentrations of TFV in these cell types.
 
PI → Raltegravir Switch: Certain PIs have been associated with lower bone mineral density with ART initiation. In the SPIRAL LIP study (#845), HIV-infected persons on suppressive ART were randomized to continue their PI-containing ART regimen or switch the PI for raltegravir. Switching to raltegravir was associated with increases in total body BMD, whereas those who remained on PI BMD showed no change over the 48 week interval. The between group differences were not significant. Spine and hip BMD were not measured in this study. These sites are not only more clinically relevant than total body BMD, but also are more sensitive to changes. Based on this study, it would be reasonable to carry out further studies to examine the PI to raltegravir switch on BMD at the spine and the hip. Little is known about raltegravir and bone, so this kind of switch strategy may be important, especially for older persons with lower BMD.
 
Non-ART Medications on Bone: The Good and the Bad: Growth hormone has a known anabolic effect on bone. In an randomized, placebo controlled trial (#834), tesamorelin (the growth hormone releasing hormone analogue recently approved in the US for the treatment of central lipohypertrophy in HIV) was shown to increase both markers of bone formation and bone resorption after 26 weeks of therapy compared to placebo, with a greater relative increase in bone formation. The change in these markers was associated with change in IGF-1. It will be important to investigate the whether these potential beneficial effects on bone turnover with tesamorelin translate in improved BMD (as is generally seen with GH treatment in other populations), especially since central lipohypertrophy in HIV has been associated with lower BMD in some studies.
 
Thiazolidinediones have been investigated for the treatment of lipoatrophy with mixed results. One concern about this class of medications is their effect on bone density and fracture risk. PPAR gamma is thought to control the transformation of mesenchymal stem cell in the bone marrow into either osteoblasts or adipocytes. PPAR activation through TZDs is associated with decreased osteoblast formation, in favor of adipocytes. In a randomized placebo controlled trial of rosiglitazone in HIV-infected patients with lipodystrophy, rosiglitazone was associated with decreases in the bone formation marker, P1NP, which correlated with increases in limb fat. There was no difference in bone resorption or total BMD by treatment arm. Although this trial used rosiglitazone, which is effectively not available because of increased cardiovascular risk, pioglitazone likely has similar effect on bone metabolism, based on studies from the general population. Further studies in this area should used site-specific BMD. Given the modest effect of TZDs on lipoatrophy and these long-term effects on bone, I don't believe that TZDs should be recommended for the treatment of lipoatrophy in those without diabetes mellitus. In those with diabetes and lipoatrophy, pioglitazone can be considered, but the risks and benefits need to be weighed carefully.
 
Epidemiology of Vitamin D Deficiency:
 
Effect of ART on Vitamin D: EFV has been associated with lower 25OHD concentrations in cross-sectional and prospective studies. In the ECHO trial (#79 LB), HIV-infected persons initiating a regimen of TDF/FTC/EFV had a median decrease of 2.2 ng/mL compared to those randomized to TDF/FTC/TMC278, a novel NNRTI, and nearly a two-fold increase in those who developed severe vitamin D deficiency. It is speculated that the effect of EFV on 25OH D is mediated by the induction CYP24A, the enzyme that converts 25OHD to its inactive metabolite 24,25 dihydroxy vitamin D, but this requires confirmation in further clinical studies. Also, critical to our understanding of this effect is whether it leads to adverse skeletal (osteoporosis, osteomalacia, and fracture) and non-skeletal outcomes (CVD, dyslipidemia, increased inflammation, and insulin resistance). Also important is whether EFV-treated patients require a higher dose of supplementation to maintain the 25OHD in the target range (generally > 30 ng/mL).
 
In a large French cohort study (#826), ART was also investigated as a possible contributor to low vitamin D. Consistent with the ECHO trial and other studies, EFV was associated with lower 25OD, as was AZT. Interestingly TDF was associated with higher 25OHD (difference vs non-TDF; 3.4 ng/mL). Whether this relationship is causal requires evaluation in prospective trials.
 
Vitamin D and Non-skeletal Endpoints: Low vitamin D concentrations have been linked to glucose abnormalities in the general population. In the Modena Cohort (#827), those who reported consistently taking vitamin D supplements were over 80% less likely of developing diabetes over an average of 2.4 years of follow-up, in multivariate models. Although provocative, this finding requires confirmation in a prospective, randomized controlled trial before vitamin D can be recommended for the prevention of diabetes mellitus, as persons who were taking vitamin may have had other traits (more physical activity, better diet, etc) which may have influenced diabetes incidence and were incompletely adjusted for in multivariate analysis.
 
Observational data in the general population have shown an association between vitamin D deficiency and cardiovascular disease and endothelial function may improve with vitamin D supplementation. In a prospective, randomized, double-blind, placebo-controlled trial (#829), 45 HIV-infected adults on stable ART with 25OHD concentrations ≤20 ng/mL were randomized to vitamin D3 4000 IU/d or placebo over 12 weeks and the change in flow-mediated brachial artery dilation (FMD) was assessed. Although FMD was lower than expected in both arms prior randomization, vitamin D supplementation showed no effect on endothelial function. There are some caveats with the interpretations of the findings. Although 25OHD increased significantly in those who were randomized to vitamin D supplementation (average change +4.6 ng/mL), the increase was much less than anticipated. Generally, 25OHD should increase by 1 ng/mL for each 100 IU of D3 supplemented over 12 weeks, so in this case 40 ng/mL. Reported compliance in the trial was excellent and the vitamin D product contained the reported amount of vitamin D within a tight margin. The reasons for this failure of a high dose of vitamin D to increase 25OH adequately are unclear. The use of an EFV containing ART regimen was investigated as a potential contributor and there was a hint of an EFV effect, but it was not statistically significant. Whether FMD will improve if 25OH is replaced to adequate levels remains an open question.