Osteoporosis (bone), Frailty, and Vitamin D: CROI 2014
Todd T. Brown, MD, PhD
Division of Endocrinology, Diabetes, & Metabolism
Johns Hopkins University
With continued interest in the aging HIV population, research investigating aging-related co-morbidities, such as osteoporosis and frailty, continues to gain prominence at CROI. This year's CROI featured several important studies related to osteoporosis and bone density as well as an excellent and thought-provoking themed discussion of studies related to frailty and functional decline with aging (link below), led by Kevin Yarasheski from Washington University. There were fewer studies related to vitamin D this year. As in years past, I will highlight the studies related to osteoporosis, vitamin D, and frailty that have the highest impact on either our understanding of the pathogenesis of these problems or how we manage our patients.
PIs have More Pronounced Bone Loss with ART Initiation Than Raltegravir: Bone mineral density drops by 2-6% with ART initiation which partially depends on the regimens being used. Tenofovir disoproxil fumarate (TDF) is associated with a 1-2% greater loss than other NRTIs. In addition, some studies have suggested that PIs have also been associated with more loss compared to other 3rd drugs, but the results have been inconsistent and individual PIs have not been compared head to head. There's also some evidence from switch studies and ART initiation trials that integrase inhibitors may be associated with less bone loss compared to other medications. A5257 (ARDENT) was a large clinical trial sponsored b the AIDS Clinical Trial Group in which ART-naïve, HIV-infected persons were randomized to TDF/emtricitabine (FTC) plus either boosted atazanavir (ATV/r), boosted darunavir (DRV/r), or raltegravir (RAL) (#85, http://www.natap.org/2014/CROI/croi_30.htm). A substudy of ARDENT was also conducted in 328 persons to determine the effect of these regimens on measures of subclinical cardiovascular disease, body composition, and bone density. The bone results were presented at this year's CROI (3779LB, http://www.natap.org/2014/CROI/croi_57.htm).
As in previous studies, bone mineral density dropped in all three arms over 96 weeks. However, at both the lumbar spine and the total hip, bone loss was less with RAL compared to the PIs and there was no difference between ATV/r and DRV/r. At the total body, the picture was a little different: ATV/r showed greater bone loss than DRV/r and there was no difference in the bone loss between the DRV/r and RAL arms. The reasons and clinical significance for these latter findings are unclear. Total body BMD does indeed predict fracture in the general population, but the spine and hip sites are more clinically relevant. The findings were not due to differential dropout in the ATV/r arm since the intent to treat and as treated analyses showed similar results. Although we didn't show that data in the poster, there were no differences between the arms with respect to the change in fat mass, lean mass, or biomarkers that could explain the results. One possibility to explain the differences in the PI effect between the spine, hip, and total body is the type of bone that was being studied. The total body is about 80% cortical bone and 20% trabecular bone, whereas the spine is 20% cortical bone/80% trabecular bone and the ratio at the total hip is about 50/50. It is possible that DRV/r has less of an effect on cortical bone compared to ATV/r but we could not resolve these differences, since DXA, unlike quantitative CT, cannot distinguish between trabecular and cortical bone. Nevertheless, these findings may have some important clinical implications. For persons with elevated fracture risk initiating ART, RAL (or perhaps other integrase inhibitors) may be a better choice than a PI.
High Dose Vitamin D/Calcium Attenuates Bone Loss with TDF/FTC/Efavirenz: The biggest bone story of the meeting was also the biggest vitamin D story. If bone loss is predictably associated with the initiation of ART, what can be done about it? In another ACTG study, 165 ART-naïve persons initiating TDF/FTC/EFV were randomized to either Vitamin D3 4000 IU/d with calcium 1000 mg or placebo and followed over 48 weeks. (#133; http://natap.org/2014/CROI/croi_50.htm). While bone density dropped in both arms, the loss of BMD was less in those randomized to calcium/vitamin D (Total Hip: active: 1.46% vs placebo 3.19%, p<0.001; Lumbar Spine: active: 1.41% vs placebo 2.91%, p=0.08) . Calcium/vitamin D treatment also led to attenuated increases in bone turnover markers and parathyroid hormone, but changes in inflammatory markers (IL-6, sTNF-R1, aTNF-R2, CD14) were similar between the arms. Adverse events, including kidney stones were limited and did not differ between the arms. These findings suggest the calcium and vitamin D may be a safe and inexpensive way to reduce ART initiation-associated bone loss with TDF/FTC/EFV.
There were a few important issues that were not explored in this trial. First, it's unclear whether the effect of calcium/vitamin D would be similar with the initiation of other ART regimens. TDF/FTC/EFV was chosen given its popularity, the known effects of TDF on bone, and the negative effects of EFV on vitamin D metabolism. It will be important to test the effect of calcium/vitamin D with the initiation of other regimens. Second, many questions remain about the optimal dose and duration. We chose a high dose of vitamin D to optimize the chances that an effect would be seen, since this was the first trial of its kind. It's unclear whether a lower dose of either of these supplements would have had similar effects, whether the observed effect was due to calcium, vitamin D, or the combination, how durable the effects are, and whether BMD would decrease after the calcium/vitamin D has been discontinued or the dosage is reduced after the initial 48 weeks. These questions deserve further investigation.
Nevertheless, this study is the first to date to report an effective intervention to curb bone loss during this period and may be a good strategy for clinicians to employ during ART initiation.
Statins and Bone: Numerous studies in the general population have clearly shown that statins reduce the risk of cardiovascular events. Statins may also affect other organ systems including bone. Both statins and bisphosphonates impair the mevalonate pathway. In the osteoclast, bisphosphonates inhibit an enzyme which is further down the mevolonate pathway than HMG-coA reductase, farnesyl-PP synthase. This enzyme is important for proper osteoclast bone resorption. In addition, in vitro evidence suggests that statins may increase bone formation through the effect of Bone Morphogenic Protein-2. Thus, there is rationale to investigate whether statins may impact bone metabolism. In the SATURN study (http://www.natap.org/2014/CROI/croi_44.htm), 147 HIV-infected persons on ART with HIV RNA < 1000 cp/ml, LDL <130 mg/dL, and evidence of heightened immune activation (CD8+CD38+HLA-DR+>19% or hsCRP>2 mcg/mL) were randomized to rosuvastatin or placebo and followed over 48 weeks. About 25% of the population had osteopenia at the total hip or lumbar spine at baseline (NB: this is lower than the prevalence estimates of osteopenia in other HIV-infected cohorts). Over the course of the study, BMD increased by 0.6% in the rosuvastatin group compared to a decrease of -0.6% in the placebo group (p=0.017). There was no difference between the arms in the change in spine BMD. The intervention was well tolerated, but statin treatment was associated with a 72% increase in HOMA-IR (p=0.0015 vs baseline, vs 14.5% in the placebo group, p= 0.007), a marker of insulin resistance, although only one participant in the placebo group developed diabetes. These findings echo studies from the general population describing the increased risk of diabetes in those on statins. The study highlights the fact that glucose should be monitored in statin-treated patients. However, since, in the general population, statins reduce CV events similarly in those with and without DM, clinicians should not avoid prescribing statins in those with or at risk for diabetes for fear of worsening glycemic control. Regarding the possible effect on BMD, it's unclear whether this modest increase in BMD would be any different in a population that was selected based on low BMD and it remains to be determined whether these increases in hip BMD would be associated with reduced fracture risk. The studies that have investigated the reduction of fracture risk in the general population with statins have been disappointing.
Bisphosphonate Dosing: Bisphosphonates are the first line treatment for osteoporosis in the HIV-infected patient and in the general population. Among bisphosphonates, zoledronic acid is probably the most effective at decreasing fracture risk, and, because it is given IV (yearly), it does not have the same problems as oral bisphosphonates regarding medication adherence and persistence. Some data from the general population suggest that less frequent dosing may be equally as effective yearly dosing. Negredo and colleagues randomized 31 HIV-infected persons with low BMD in a 2:1 fashion to either zoledronic acid or no treatment (http://natap.org/2014/CROI/croi_63.htm). Those who received zoledronic acid either received annual dosing over 96 weeks (i.e., two doses) or biennial dosing (i.e, one dose at baseline). Both zoledronic acid arms showed an increase in both spine and hip BMD over the interval compared to the untreated control group, but there was no difference between the two dosing regimens of zoledronic acid at end of 96 weeks. These findings suggest that the bisphosphonates could potentially be given less frequently in this population, thereby reducing toxicity and cost. While these results are intriguing, it should be noted that it is unclear whether this strategy would lead to the same protection against fractures. Until more data become available, clinicians should continue to give yearly treatment.
FRAX and Low BMD Predict Fracture in HIV-infected Patients: Two studies in the HOPS and SUN cohorts provided important data regarding the relationship between surrogates of fracture risk and actual fracture risk (http://www.natap.org/2014/CROI/croi_58.htm; http://www.natap.org/2014/CROI/croi_56.htm). DXA-derived BMD is an excellent tool to assess fracture risk. Similarly, FRAX, which gives the 10 year risk of fracture based on several clinical risk factors for fracture, with or without BMD (think Framingham Risk Score for fracture), is also an excellent clinical tool to risk stratify patients. Both of these measurements have been assumed to be associated with actual fractures in HIV-infected persons, but there was a lack of data. These studies clearly showed that both of these measures were statistically associated with the fracture rate. For BMD, those with osteoporosis at the hip (about 3% of the study population of 1008) had over a three-fold risk of having an incident fracture (HR 3.03, 95% CI: 1.46-6.29), whereas those with osteopenia had no increased risk compared to those with normal BMD. One limitation was that the study used only hip T-score to categorize the population into osteoporosis/osteopenia/normal, rather than both the hip and lumbar spine T-score. In addition, the osteopenia group is quite heterogeneous and exploring various cutpoints within the osteopenia category would have been useful. In this way, clinicians would be able to determine the clinical relevance of low spine BMD or a specific T-score within the broad category of osteopenia.
For FRAX, a score ≥ 3% was associated with a greater than two-fold increase in any incident fracture (HR 2.31, 95% CI: 1.54-3.46). HIV-infected persons in this higher risk category were more likely to be white, male, MSM, and have a lower CD4 cell count. The major limitation was the study focused on all incident fractures, rather than just the fractures that are used in FRAX (forearm, upper arm, vertebral bodies, hip) since the numbers were small in the latter group. Therefore, the authors were unable to determine whether FRAX underestimates observed fracture risk in HIV-infected persons. This has been hypothesized since other clinical risk factors (eg ART regimen, CD4 cell count) may be important in the pathophysiology of fracture risk, but are not included in FRAX. It will be important for clinicians to know whether they can rely on FRAX in their HIV-infected patients to guide screening and treatment decisions.
Bones in Special Populations:
TDF-Exposed Infants: The safety of TDF exposure in utero is not clear. Given the known detrimental effect on bone in HIV-infected persons and in HIV-uninfected persons receiving PrEP, there is concern that TDF may also affect the skeleton of developing fetus in a TDF treated mother. In Pediatric HIV/AIDS Cohort Study (PHACS), the bone mineral content was compared in HIV-uninfected newborn infants of mothers who were exposed to at least 8 weeks of TDF during pregnancy (n= 74) vs those born to mothers who took a non-TDF-containing regimen (n=69) (http://natap.org/2014/CROI/croi_64.htm). The main finding was that TDF-exposed newborns had BMC approximately 0.5 standard deviations lower that those without TDF-exposure, which was unchanged after multivariable adjustment. It will be critical to follow these infants longitudinally to determine whether this difference persists or if BMC returns to normal. If the difference persists, consideration will need to be given to either trying to avoid TDF during pregnancy or investigate preventative measures such as vitamin D and calcium supplementation that can be given to TDF-treated mothers.
Perinatal Infection: Mike Yin has previously shown that young men who were infected though perinatal transmission had multiple abnormalities of bone density, structure and strength (Yin, AIDS, 2013) compared to age matched controls. These abnormalities reduce peak bone mass and strength and may lead to later fragility fractures. In a new analysis (http://www.natap.org/2014/CROI/croi_101.htm), he and his co-authors from Columbia tested the hypothesis that these men would show increased senescence of osteoblast precursors due to the effects of chronic inflammation, thereby leading to a bone formation defect. They also looked at bone parameters and osteoblast precursors in men with HIV acquired during adolescence. They found that perinatally HIV-infected men had lower indices of bone density and strength compared to adolescence-infected men. Perinatally HIV-infected men also had lower numbers of circulating osteoblasts precursors and shorter telomere lengths, suggestive of increased replicative senescence. Both HIV groups had lower bone strength parameters and osteoblast precursors compared to the HIV-uninfected control population. These findings suggest that early infection with HIV may lead to impaired osteoblast function which may have major implications for the bone health of these men later on in life.
Resource Limited Settings: There were several studies which investigated the prevalence of osteoporosis among HIV-infected persons in resource limited settings. Wandera (http://www.natap.org/2014/CROI/croi_65.htm) measured BMD by DXA in 181 HIV-infected patients who had failed their first line ART-regimen and were about to embark on their second line regimen. They found that 8% had a lumbar spine T-score ≤-2.5 and that low BMD was associated with lower body weight. Given that the median age was 35 years, the prevalence of low BMD was notable and deserves follow-up. It will be particularly interesting to investigate the effect of the second line regimen on BMD. In other populations, initiation of a second ART regimen brings about another period of bone loss (Mallon, AIDS, 2013). In the upcoming years, it will be also important to determine whether HIV-infected patients in resource limited settings have higher risk of fracture compared to the underlying background population.
DXA is the clinical standard for the assessment of BMD. However, DXA machines are expensive and not readily available in resources limited settings. Arpadi and colleagues correlated conventional DXA measurements with heel quantitative ultrasound (QUS)( http://www.natap.org/2014/CROI/croi_67.htm). This technique has the advantage over DXA in that it uses a portable, lower cost machine and it therefore an attractive option in resource limited settings. In 47 HIV-infected children and adolescents in South Africa, a measurement of QUS, called broadband ultrasound attenuation showed moderate correlation with total body and spine DXA. Further research is needed in older populations in RLS and to determine the associations of these QUS measurements with fracture risk.
HCV/HIV co-infection: Among HIV-infected patients, HCV co-infection has been one of the most consistent risk factors for low bone mineral density and fracture. It is unclear, however, whether this is due to HCV infection per se, behavioral factors that are associated with HCV infection acquisition (eg drug use, poor nutrition), or consequences of chronic liver disease. To help better characterize the relationship between HCV infection and low BMD, Cutrell and colleagueshttp://natap.org/2014/CROI/croi_162.htm enrolled 298 male veterans into 4 groups: HIV/HCV co-infection, HCV monoinfection, HIV monoinfection, and Controls. All HIV-infected men had well controlled HIV on ART and all HCV infected men had not received HCV treatment. Both HIV infection and HCV infection were associated with lower BMD at the total hip and femoral neck, but not at the lumbar spine. Interestingly, while HIV infection was associated with increases in bone turnover markers, the picture with HCV was more mixed. CTX (a marker of bone resorption) was no different than controls, bone specific alkaline phosphatase (a marker of bone formation) was higher than controls, and osteocalcin (another marker of bone resorption) tended to be lower than controls. Higher levels of both osteoprotegerin and RANKL were related to HCV, but not HIV infection. Low BMD among the HIV-infected men was related to TDF use. Taken together, these findings suggest that HCV and HIV/ART both contribute to low BMD, but the mechanisms may be different. Further prospective studies are needed to better understand the relative effect of these chronic infections on bone metabolism and the effects of treatment.
Frailty and Functional Impairment: Aging is associated with a decline in physical function which generally occurs in the eighth or ninth decade of life. There is concern that this process may occur earlier in HIV-infected populations. However, it's unclear whether functional decline in HIV-infected persons is related to HIV itself, other co-morbidities and co-infections, legacy effects of long term use of more toxic ART regimens, or other behavioral factors (alcohol, smoking, drug use) that may be more prevalent in HIV-infected persons. There is also some concern that the HIV-uninfected control populations that have been used for comparison have not been well matched to the HIV-infected populations in some studies, resulting in biased estimates of the HIV effects. All these issue were discussed in the Frailty/Functional Impairment theme discussion on Wednesday March 5 (click here for a link to the webcast http://www.croiwebcasts.org/portal) along with the presentation of some of the latest science in this area. Links to the Themed Discussion: Frailty and Functional Impairment are at the end of this report.
Meredith Greene and her colleagues at UCSF compared the prevalence of 4 geriatric syndromes in HIV-infected persons over 50 years old (median age 57 years) to HIV-uninfected controls, including (1) falls, (2) urinary incontinence, (3) functional impairment, and (4) frailty (#766; http://natap.org/2014/CROI/croi_66.htm). While the age adjusted prevalence of falls (self-report in the last 12 months) were similar by HIV status, the age-adjusted prevalences of urinary incontinence, functional impairment, and frailty were all higher in the HIV-infected groups. Lower nadir CD4 cell count and the presence of co-morbidities were associated with the presence of two or more geriatric syndromes. These findings do suggest that geriatric syndromes are occurring in relatively young HIV-infected persons. The underlying etiologies, particularly those which can be treated or prevented, will be important to identify.
These geriatric syndromes may also be associated with mortality and are related to systemic inflammation. In an analysis in the ALIVE cohort (#762 ;http://natap.org/2014/CROI/croi_61.htm), a population of injection drug users in Baltimore, Damani Piggott investigated the relationship between frailty and systemic inflammation index ((log IL-6 + 2*log sTNFR1)/3) which has been used in the InChianti study, a geriatric epidemiologic study in Italy. In a multivariable model, frailty was significantly associated with the inflammatory index score (aOR 2.00; 95% CI, 1.32-3.01). The next step was to determine whether this inflammatory index was associated with mortality independent of frailty. In models adjusting for sociodemographics, comorbidity, HIV and frailty status, the inflammatory index score was independently associated with increased mortality risk (aHR 2.90; 95% CI, 2.18-3.84) and the effect was similar in those with and without HIV-infection. It should be noted that the overwhelming majority of this cohort are also chronically infected with HCV which may have contributed to the degree of systemic inflammation. These findings suggest that systemic inflammation, while related to frailty, is also an independent risk factor for mortality.
In another study, Kristine Erlandson and her colleagues at the University of Colorado found that objective measures of physical performance, such as timed chair stands and 400 meter walk, were associated with higher physical and mental quality of life as measured by the SF-36 (#767; http://www.natap.org/2014/CROI/croi_85.htm). Although causality cannot be determined by this cross-sectional study, it leaves open the possibility that exercise programs aims to increase physical performance may also improve physical and mental quality of life.
Exercise as an Intervention in Aging HIV populations: One of my favorite studies from this year's CROI was a small pilot study to understand the effect of 12 weeks of brisk walking (1 hour three times a week) on markers of inflammation in 35 HIV-infected persons on ART (http://www.natap.org/2014/CROI/croi_86.htm). In those who completed the intervention and had available data, there were significant declines in soluble markers of inflammation (d-dimer, IL-6, hsCRP, IL-18) (n=25) and cellular markers of immune activation (HLA-DR+, CD38+ on CD4+ and CD8+cells) (n=16), as well as performance on the 6 minute walk test and cholesterol parameters. There are many strategies that are being tested to decrease inflammation and immune activation with the goal of decreasing non-AIDS co-morbidities. This study provides some initial data that exercise may also be an effective strategy to decrease inflammation and has the added advantage of being low cost with few negative side effects. The greatest challenge for this study is what also diminshes the impact of exercise recommendations in real life: it's hard to encourage people to stick with it. There is no question that the efficacy of exercise interventions needs to be evaluated in HIV-infected patients, including those with frailty and functional limitations. Along with this, behavioral strategies to encourage adherence and persistence will also be critical to investigate.
Note from Jules: the posters for each of these talks contain all the data & are discussed in this report with links to the full reports of the poster contents on the NATAP website. This subject of Frailty, Functional Impairment, Comorbidities & Mortality are very important for clinical patient care. Very soon over 50% of HIC+ patients will be >50 years old & a significant percent >60 years old. It is not possible to know exactly the impact on the overall patient population of these syndromes, but it is very likely in my opinion that a significant proportion of patients will be sorely impacted. In the general population following a bone fracture in older individuals mortality increases. I don't think we as a community are adequately prepared for the potential impact. Studies that estimate survival rates of HIV+ cannot possibly include the potential impact because it is an unknown.
Links to -
Themed Discussion: Frailty and Functional Impairment
- Discussion Leader
Kevin E. Yarasheski
Washington University School of Medicine, St Louis, MO, United States
- Geriatric Syndromes Are Common Among Older HIV-Infected Adults
Medicine-Geriatrics, University of California San Francisco, San Francisco, CA, United States and San Francisco VA Medical Center, San Francisco, CA, United States
- Physical Function Impairment On Quality of Life Among Persons Aging With HIV Infection
Kristine M. Erlandson
University of Colorado Anschutz Medical Campus, Aurora, CO, United States
- Frailty, Inflammation and Mortality Among Aging HIV-Infected and At-Risk Injection Drug Users
Damani A. Piggott
The Johns Hopkins University, Baltimore, MD, United States
- Brisk Walking Improves Inflammatory Markers in cART-Treated Patients
Infectious Diseases, San Raffaele Scientific Institute, Milano, Italy
- Questions and Answers