icon- folder.gif   Conference Reports for NATAP  
  16th CROI
Conference on Retroviruses and Opportunistic Infections Montreal, Canada
February 8-11, 2009
Back grey_arrow_rt.gif
Bone/Osteoporosis CROI 2009 Update
  CROI 2009 Montreal Feb 8-12
Todd Brown, MD, PhD
Assistant Professor of Medcine
Division of Endocrinology and Metabolism
Johns Hopkins University
Baltimore, MD
Osteoporosis is a major source of morbidity and mortality associated with aging. Among HIV-infected patients, it has been established that the prevalence of low bone mineral density (BMD) and fragility fracture is higher compared to HIV-uninfected control populations. The pathogenesis of reduced BMD among HIV-infected populations is multi-factorial with contributions from the effects of chronic HIV-infection, some antiretroviral therapies, and traditional osteoporosis risk factors which are more prevalent among HIV-infected populations.
Many aspects regarding the risk factors and pathophysiology of osteoporosis, and its optimal management in HIV-infected patients, however, require further clarification. With an aging HIV-population, increasing attention is being focused on this important problem, which is evidenced by the increasing quality and quantity of the studies examining BMD issues among HIV-infected patients at the 16th CROI in Montreal. This review will summarize some of the studies presented at this year's CROI, particularly emphasizing those that have clinical relevance.
HIV Effects on Bone: Bone is constantly being remodeled though the actions of osteoclasts which resorb bone and osteoblasts which lay down new bone. If bone resorption (Bone resorption is the process by which osteoclasts break down bone) outstrips bone formation, then there is net bone loss, resulting in osteoporosis. It has been previously observed that HIV is associated with impairment in osteoblast function. In an in vitro study of human osteoblasts (#759), heat-inactivated HIV administration resulted in osteoblast apoptosis, attributable to the effects of the HIV protein, gp120. Importantly, there was no evidence of HIV-infection of the osteoblasts and the apoptotic effect of gp120 was mediated by secreted TNF_. These results show an interaction between viral and inflammatory factors in osteoblast apoptosis, which have important implications in the pathogenesis of bone loss in those with HIV-infection.
BMD and the Immunologic Response to HIV-infection: Low CD4 cell count has been identified as an independent risk factor for low BMD in both cross-sectional and longitudinal studies. The mechanism underlying this association is uncertain. In a cross-sectional analysis of 78 HIV-infected patients (#752), those who had low BMD were found to have a lower mean percentage of CD8+CD127+ cells, compared to those with normal BMD, suggesting that low BMD is associated with enhanced T-cell activation and senescence. This gives us more specific information regarding immunologic abnormalities and bone mineral density, but requires longitudinal studies to investigate a potential causal relationship.
Antiretroviral Factors Associated with Low BMD: Tenofovir: Multiple studies, including the GS903 randomized trial, have previously established an independent effect of tenofovir (TDF) on BMD. In the STEAL study (#576), investigators in Sydney provided important information regarding the effect of TDF vs abacavir (ABC) on BMD. In this study, 357 ART-treated HIV-infected patients with well-suppressed HIV were randomized to fixed dose TDF/FTC or ABC/3TC and followed for 96 weeks. Although the two approaches showed similar virologic efficacy, those who were randomized to TDF had a statistically significant decrease in BMD compared to those randomized to ABC (mean T-score difference 0.16 (95%CI 0.08, 0.23; p<0.0001)). These results extend upon the results of the GS903 study which showed a significant difference in the decrease in BMD in TDF vs d4T regimens with ART initiation. However, there are important differences. In this study, the change in BMD occurred in patients whose HIV was already suppressed. This helps to disentangle the effects of ART-initiation (about a 2-3% decrease in BMD regardless of the regimen) and the specific effects of TDF. Furthermore, the study compared BMD changes in the two most commonly used nucleoside/tide backbones using a randomized design. The clinical significance and the mechanism underlying the effect of TDF on BMD require clarification, although some insight into mechanism can be gleaned from other studies presented at the conference.
One potential mechanism for the effect of TDF on BMD is a direct effect on bone turnover. In a longitudinal observational study (#760), bone turnover was investigated in 44 HIV-infected patients initiating TDF-containing regimens and 43 initiating non-TDF regimens (91% AZT). Before ART initiation, bone resorption activity was high, as evidenced by the high proportion of patients (94%) with CTx (a marker of bone resorption) above the normal range. In contrast, osteoblast activity, as measured by serum osteocalcin, was relatively suppressed, consistent with a direct effect of HIV on osteoblasts (as seen in abstract #759 described above). With ART initiation, bone turnover accelerated markedly in both TDF-treated and non-TDF treated patients, with larger increases in bone resorption ( median 41%), compared to bone formation (median 33%), suggesting an imbalance in osteoclast and osteoblast activity, in favor bone resorption. Interestingly, in multivariate models, only advanced HIV-disease at baseline was associated with increases in bone resorption (percent change in CTx) whereas TDF-exposure, PI-exposure, and higher levels of baseline inflammation (sTNFR1) were associated with greater increases in bone formation. Although speculative at this point, these findings may suggest that TDF and PIs may increase osteoblast activity in a system with already active bone resorption, thereby greatly accelerating bone turnover. However, because of relatively higher levels of bone resorption, the net effect is bone loss. Further studies using a randomized design and concomitant measurement of bone mineral density are required to better understand these effects.
Another potential mechanism for the effect of TDF on BMD is a direct effect on phosphate homeostasis. TDF has been associated with proximal tubule dysfunction and phosphate wasting, which can result in osteomalacia (reduced bone mineralization). The Swiss Cohort reported a high prevalence of proximal tubule dysfunction (22%-50%), with the highest prevalence in those receiving TDF (#743). The most sensitive marker of proximal tubule dysfunction was the fractional excretion of phosphate. This study raises the important question about using fractional excretion of phosphate as a screening tool for proximal tubule dysfunction. Prospective studies are required to link urinary phosphate wasting with TDF and BMD loss.
Effect of TDF in Children: One of the major risk factors for osteoporosis later in life is a failure to reach bone mass during young adulthood. Data from the Pediatric HIV/AIDS Cohort study (#911) demonstrated a ~ 20% prevalence of low BMD (Z-score ≦ -1.5) in of 249 children (median age 12 years), suggesting that among some HIV-infected children, bone mass accrual is affected. Because of the known effects of TDF on BMD in adults, the effect of TDF administration on BMD in children has been actively investigated.
In a 4-year longitudinal study (#913), bone mass (lumbar spine and total body) in 23 HIV infected children was measured after having switched from d4T/3TC/PI to TDF/3TC/EFV. These measurements were then compared to the bone mass in 194 healthy HIV-negative subjects. Bone mineral content in the HIV-infected children was lower at baseline and at 4 years, but no significant changes from baseline were observed. While these results argue against a specific effect of TDF on BMD in this small sample, BMD remained lower in HIV-infected children compared to HIV-negative controls and low bone mass was not restored.
Even among some HIV-infected children in whom bone mass accrual appears to be affected by TDF, the effect on BMD may wane over time. Of 37 children followed in the NIH clinic (#912), overall the average Z-score in the lumbar spine was -0.43 +/- 1.06. Overall, those exposed to TDF had similar Z-scores to those who were not exposed to TDF. However, close examination of the BMD trajectory in two TDF-treated patients provides some interesting insights. Although in the early pre-pubertal years, bone mass accrual appeared to be attenuated. By mid-puberty, one child had a rapid increase in BMD despite remaining on TDF, while the other only recovered bone after TDF discontinuation. These case reports demonstrate that losses of bone mass with TDF can be reversible.
Bone Mineral Density in Men: The sites measured by DXA differ not only in location, but in the proportion of trabecular and cortical bone. The spine is approximately 75% trabecular bone, whereas the proportion of trabecular bone in the hip is lower (25%-50%). Both decreases in cortical and trabecular bone contribute to osteoporosis in the general population. Among HIV-infected patients, it is not clear whether one site or one type of bone is preferentially affected compared to another.
In a cross-sectional study (#755), younger HIV-infected males (median age 43) were found to have lower hip T-scores, but not spine T-scores, compared to unselected age-matched HIV-uninfected men referred for DXA, which suggests that the hip site may be preferentially affected in HIV-infected men. Further studies comparing HIV-infected men and women to matched HIV-negative controls using techniques that can quantify trabecular and cortical bone are required to understand the type of bone that may be lost in HIV-infected patients.
Bone Mineral Density in Women: Because the majority of fragility fractures occur in post-menopausal women in the general population, post-menopausal HIV-infected patients are a particularly important sub-population in whom to characterize BMD. Yin et al (#758) presented data on 99 HIV-infected and 101 HIV-uninfected post-menopausal women in New York City. Even after adjustment for traditional risk factors, T-scores (spine and hip) were lower in HIV-infected women compared to HIV-uninfected controls. Furthermore, measures of bone resorption were higher among HIV-infected women, suggesting a high bone turnover state. It is unclear the extent to which these differences in BMD and bone turnover will translate into a higher fracture risk among post-menopausal HIV-infected women.
Similar results were seen among a large sample of middle-age women (n=464) with or at risk for HIV-infection in New York City (#757), in whom a baseline cross-sectional evaluation showed lower BMD among HIV-infected women. At a subsequent DXA evaluation (minimum 18 months), the change in BMD did not differ by HIV-status, which in accordance with other published studies showing relative stability of BMD in ART-treated persons. Interestingly, those women reporting opiate use were found to have a greater decrease in BMD over time, confirming similar findings observed in older men.
The Aquitaine Cohort presented a cross-sectional study of younger HIV-infected women (#978), confirming the association in the general population between low lean body mass and lower BMD. It is unclear whether this association is due to reduced physical activity, poor nutritional status, mechanical effects of muscle on bone, or hormonal factors which may contribute to an increase in bone loss.
One potential hormonal deficiency among HIV-infected women which may contribute to both low lean body mass and reduced BMD is testosterone (T) deficiency. Although low T levels are common among HIV-infected women, routine T replacement is not currently recommended in women, due to safety concerns. Looby et al (#976) reported the results of a small (n=25), randomized trial of HIV-infected women with low T who were randomized to either T (300 mcg transdermal twice weekly) or placebo over 18 months. Among those randomized to T, the mean T concentration at 18 months was 7.9 +/- 1.8 pg/mL (nl range 1.1-6.3 pg/mL) and BMD increased over the study interval compared to placebo, but clinical adverse effects were similar between groups. This is an important proof-of-concept trial, but further studies need to be carried out before T replacement in women can be used safely in the clinical setting.
Risk Factors for Low Bone Mineral Density: Medications other than antiretrovirals may contribute to the high prevalence of reduced BMD among HIV-infected patients. Thiazolidenediones (glitazones) are anti-diabetic medications, which have also shown some promise in the treatment of lipoatrophy, although the findings have not been consistent. Recently, studies in the general population have shown that glitazones are associated with reduced BMD and increased fracture risk, likely through their effects of decreasing the numbers of osteoblasts, via PPAR-y activation.
Marisa Tungsirispat presented results from a 48-week, randomized, placebo-controlled trial of rosigltiazone (4 mg twice daily) in HIV-infected with lipoatrophy not taking thymidine analogs, which showed a significant improvement in limb fat compared to placebo (~0.65 kg) (#42LB). The authors also showed that there were no differences in the change in total body BMD between the arms. While this is somewhat reassuring, only total BMD was measured as an outcome. It is unclear if rosiglitazone would have an effect on BMD, if sites with more trabecular bone, which is more metabolically active (eg spine), were measured in addition to total body BMD.
In the Modena Cohort, hepatits co-infection in women, but not men was associated with a significantly higher prevalence of low bone mineral density (Z-score ≦ -2.0) compared to those with HIV mono-infection, after adjustment for traditional osteoporosis risk factors and other imbalances between the groups. It is unclear if this difference is due to specific effects of viral hepatitis or risk factors that are more prevalent among co-infected women, such as nutritional deficits, opiate use, or menstrual abnormalities.
Screening Concerns: Screening strategies for osteoporosis among HIV-infected patients have not been established. In the general population in the US, the National Osteoporosis Foundation recommends screening with DXA in: (1) men ≥ 70 years and women ≥ 65 years, and (2) younger post-menopausal women and men > 50 who have additional osteoporosis risk factors. Although HIV is not specifically identified as an osteoporosis risk factor in the NOF guidelines, a strong argument can be made for screening HIV-infected men and women > 50 years. Other screening strategies can be used.
Short el al (#753) presented a cross-sectional study examining the utility of peripheral DXA in identifying HIV-infected patients who have osteoporosis based on central DXA. The advantage of peripheral DXA is that it is low-cost and could be used conveniently in a clinical setting. Therefore, it may be useful as part of a staged screening strategy.
Of the 168 patients enrolled in the study (median age 45, range 20-85), 22 (13%) were found to have osteoporosis with good agreement between peripheral and central DXA. Using a cutpoint of a T-score of -1 on the peripheral DEXA, sensitivity for the identification of osteoporosis was 95% with a specificity of 34% and a negative predictive value of 98%. This means that a negative peripheral DXA would be reassuring and therefore no further screening would be required. This is potentially an important strategy, particularly in those who do not meet current screening guidelines (i.e age < 50 yrs). It would be useful to know the performance of peripheral DXA specifically in this population of younger patients in whom central DXA would otherwise not recommended. In addition, it would be helpful to know if peripheral DXA gives any additional information beyond established clinical risk factors for low BMD, such as low body weight or hypogonadism.
Treatment Issues: Vitamin D deficiency is common in both HIV-infected patients and the general population and may contribute to reductions in BMD. The optimal replacement is unknown. In the study described in abstract # 756, 74 HIV-infected men were examined, 51 of whom had either Vit D deficiency or insufficiency (25 OH Vit D< 30 ng/mL). Replacement strategies with vitamin D3 (cholecalciferol) varied according to baseline 25 OH Vit D: < 10 ng/mL, 2800 IU/day; 10-20 ng/mL, 1800 IU/day; 20-30 ng/dL, 800 IU/day. After 16 weeks of replacement, the median increase in 25 OH Vitamin D was 7.2 ng/mL, in those who received supplementation. Perhaps more important were the following observations. Of the 51 patients who received vitamin D replacement, 31 (61%) reported less than 100% adherence and 12 (24%) reported 100% non-adherence. Furthermore, among those who were adherent, only 40% achieved 25 OH vitamin D levels in the normal range (> 30 ng/mL). The important take-home messages for the clinician are that: (1) adherence to prescribed daily Vitamin D is poor, and (2) higher doses of vitamin D than were used in this study are necessary to achieve adequate vitamin D levels in deficient patients.
Many experts recommend very high-dose initial repletion with subsequent daily vitamin D replacement for maintenance. In my practice, I generally give ergocalciferol (vitamin D2) 50,000 units 2-3 times per week over 8-12 weeks, depending on the degree of vitamin D deficiency. After repletion, once monthly ergocalciferol (50,000 units) or daily cholecalciferol (800-1000 IU) is generally adequate to maintain vitamin D in the normal range.
Conclusions: Multiple studies were presented at this year's CROI to help clarify the pathogenesis, risk factors, and optimal management of osteoporosis in HIV-infected patients. Many questions, however, remain unanswered. In particular, further investigation into the role of HIV infection, the immunologic host response, and the role of antiretroviral therapies is critically important. In addition, further data regarding the incidence of fragility fracture will be important to help establish guidelines regarding BMD screening in HIV-infected patients and determine whether management recommendations the general population should apply to those with HIV. An increasing incidence of fragility fractures can be expected among HIV-infected patients in the coming years as the population ages. The development of strategies to minimize the risk of fragility fracture, while balancing medication toxicities and cost-effectiveness, will be crucial to optimize the care of the HIV-infected patient. (from Jules: I think everyone with HIV should receive a baseline dexa scan since the rates of bone loss at a young age are so high among HIV+, and because HIV+ individuals appear to have such a high prevalence of risk factors: smoking cigarettes, alcohol, sedentary, poor diet, etc. Identifying low BMD or testing in general would provide an opportunity to have a discussion about the issue with patients. But clinicians would also have to receive education about screening and other issues including vitamin D and calcium cupplementation. Dexa scan screening is relatively inexpensive, about $150. We don't necessarily need 3 viral load and 3 CD4 counts yearly for all patients, dexa scans should be carved out of this. Financial considerations should not prevent addressing important individual