Low bone mineral density (BMD) is frequently observed among HIV-infected adults and children.1,2 The etiology is poorly understood and probably multifactorial. The long-term clinical outcomes are not known. Most cases of bone loss in HIV-infected individuals result in mild osteopenia; osteoporosis is less prevalent.3-6 As HIV-infected patients live longer due to more effective antiretroviral therapy (ART), they are exposed to numerous factors over time which may have an impact on BMD, including traditional risk factors (aging, genetic predisposition, menopause, low calcium and vitamin D intake and low sun exposure, lack of weight-bearing exercise, smoking, catabolic steroid use, and weight change) and HIV-associated factors (chronic HIV infection and various and changing treatment regimens).
Cross-sectional studies in HIV-infected adults show conflicting results regarding the association of low BMD with ART.3,7-14 Most of these studies selected specific patient groups based on current treatment regimen, wasting, or lipodystrophy status, making it difficult to compare across studies or to generalize across HIV-infected subgroups. In addition, the cross-sectional design of these studies makes it impossible to know for certain if the risk factors preceded the bone loss.
Our understanding of the independent effects of host factors, medications, and HIV disease factors is limited by the lack of prospective studies in men and women with careful control for confounding by host factors. Few studies have a baseline measure of BMD before treatment. In 1 clinical trial, ART-naive patients randomized to tenofovir-based highly active antiretroviral therapy (HAART) had a greater decrease in lumbar spine BMD over time compared with patients on stavudine (d4T)-based HAART.13 A few other studies examined change in BMD or bone mineral content (BMC) over time. Mondy et al15 followed patients for 72 weeks and found an overall increase in lumbar spine (2.6%) and hip (2.4%) BMD among those with undetectable HIV viral load and with greater increases in CD4 count, but no difference by protease inhibitor (PI)-based HAART compared with nonnucleoside reverse transcriptase inhibitor-based HAART. Other studies have found decreasing BMD in patients on dual-nucleoside reverse transcriptase inhibitor,16 stable BMD among patients using nelfinavir-containing HAART,13 and slightly increasing BMD among patients using indinavir-containing HAART.13 Dolan et al17 found lower baseline BMD in HIV-infected compared with HIV-negative persons, but a similar rate of change over time between the groups. McDermott et al18 found decreases in extremity BMC in PI-based HAART users and increases in those on nelfinavir. Decreases in trunk BMC were observed among those on zidovudine and increases in those on d4T. Finally, tenofovir is associated with bone loss.19-22
To better understand the evolution and predictors of change in BMD over time, we followed a heterogeneous cohort of HIV-infected men and women in the Nutrition for Healthy Living Study for a median of 2.5 years (range 0.9-6.8 years). We performed annual assessments of total body BMD by dual-energy x-ray absorptiometry (DXA) and semiannual measurements of demographics, disease severity, body mass index (BMI), physical activity, steroid use, ART, and dietary intake.
Nutrition for Healthy Living Study is a longitudinal cohort of HIV-infected adults (18 years or older) living in Massachusetts or Rhode Island as previously reported.23-27 Those included in this analysis were seen between August 1996 and September 2003. The Institutional Review Boards at Tufts University School of Medicine and Miriam Hospital, Province, RI, approved this study, and written informed consent was obtained from each participant. Participants were seen semiannually for all data collection except DXA that was annual.
Data on demographics, dietary intake, HIV-related clinical events, ART, and steroid use (prednisone/hydrocortisone, testosterone, and growth hormone) were collected by trained interviewers and nutritionists.28 Doses were not collected on medications. Participants were able to select more than 1 category to define their race. If at least 1 of the selected categories was black, they were categorized as African American. If they only selected white, they were categorized as “white.” All other categories were combined as “other.” Injection drug use was dichotomized as ever vs. never, and smoking was defined as current vs. not current. We determined which participants were living below the US Federal poverty level (yes/no) based on household income.29 Daily caloric intake, protein, calcium, and vitamin D consumption, including supplements, were determined from 3-day food records using the Minnesota Nutrition Data System Version 4.06_34. If a 3-day food record was not kept, a 24-hour food recall was obtained by a trained nutritionist. We assessed strength training over the past 7 days using the physical activity recall instrument.30 “Strength training” was defined as any strength training in the last week.
HIV RNA (log10 copies/mL) was measured by the Roche Amplicor Monitor reverse transcriptase-polymerase chain reaction assay (Roche Molecular Systems, Somerville, NJ), with a lower detection limit of 400 copies per milliliter.
Subjects were weighed (kg) fully dressed, but without shoes, heavy clothing, or objects, before eating or drinking (minimum 5-hour fast). Height (cm) was measured without shoes by stadiometer. BMI was calculated as weight divided by height squared (kg/m2). Wasting was defined as: BMI <20 kg/m2, weight loss >10% body weight since enrollment, or weight loss >5% body weight maintained at least 6 months. Ever wasting was any episode of wasting since entering the study.
Yearly DXA scans were performed on the Hologic QDR-2000 machine (Hologic, Waltham, MA). Hologic 2000 software computed total body BMD (total BMD, g/cm2).
Baseline characteristics are described using the median (25th, 75th) for continuous variables and the number (%) for categorical variables by sex. Total BMD at baseline was calculated at the median age for men, premenopausal women, and menopausal women by race (white/Asian and African American).
Unit of Analysis-DXA Interval
For each participant, consecutive DXA visits were paired to form DXA intervals. First to second DXA, second to third DXA, etc. The length of time within each DXA interval was determined. For this analysis, all intervals ranging from 9 to 24 months were included on all participants and longer or shorter intervals were excluded. Thus, each participant with ≥1 DXA interval (at least 2 DXA visits) that conformed to the above criteria was included. Two intervals were excluded because strength training was missing. We report participant characteristics at the beginning of the first DXA interval (baseline) in Table 1 and 2.
Dependent Variable (Outcome)
The dependent variable was the percent change in total BMD across a DXA interval. It was calculated as follows: (total BMD at the end of the interval minus total BMD at the beginning of the interval)/total BMD at the beginning of the interval.
We evaluated several variables for their relationship to the outcome (percent change in total BMD). These variables were measured at the beginning of a DXA interval. The variables were menopause (premenopausal and postmenopausal), disease severity (HIV viral load, CD4, and albumin), dietary intake (calcium and vitamin D) with and without supplements, body composition (BMI and weight loss), exercise (strength training), prednisone/hydrocortisone, testosterone/growth hormone, and use of HAART or any ART. Each woman was asked, “Have you been through menopause (the change of life).” If the answer was “yes,” she was categorized as postmenopausal, if the answer was “no,” she was categorized as premenopausal. She was also asked if she had had a hysterectomy (yes/no). We made 2 dummy variables to compare males with premenopausal women and to compare postmenopausal women with premenopausal women. Thus, premenopausal women were the reference group. Each ART agent was examined as months of continuous use up until the beginning of the interval. The beginning of continuous use of an ART agent may have begun before the baseline visit, and thus continuous use is the total use from baseline until the beginning of the interval plus use before the study began if it was uninterrupted at baseline. HAART was defined as use of at least 3 medications from 2 or more classes (nucleoside reverse transcriptase inhibitor, nonnucleoside reverse transcriptase inhibitor, and PI) or 2 nucleoside reverse transcriptase inhibitors.
Repeated Measures Analysis of Change in Total BMD
We evaluated the association between each independent variable and the outcome using generalized estimating equations for repeated measures regression analysis because there are multiple records per person. The Toeplitz 2 covariance structure was used because the outcome is change. The basic models contained the determinant of interest, and they were adjusted for the dummy variables of male and postmenopausal women (premenopausal women was the reference group) and time in interval (time between DXA visits). The robust variance was used to calculate 95% confidence intervals (95% CIs). Significant variables at P < 0.20 in the basic models were entered into multivariate models. The final multivariate model included significant variables at P ≤ 0.05 plus time between DXA visits, age, race, the male and menopause dummy variables, smoking, and number of years on “other ART.” Other ART indicates the maximum time of all the other ART agents that a participant was taking, not including the ones in the multivariate model. This term was included because patients are rarely on 1 agent alone. All analyses were performed in SAS 9.2 (SAS Institute, Carey, NC). The assumption of linearity between each continuous determinant (CD4, viral load, albumin, BMI, and duration of each ART) and the outcome (percent change in total BMD) was assessed using restricted cubic splines.31 Because none of these variables had large deviations from normality, they were used as continuous variables. A missing indicator was used for albumin and menopause, for which there were 35 and 18 intervals missing, respectively. Two additional variables were included, 1 for catabolic steroid use (hydrocortisone and prednisone) and 1 for anabolic steroids (testosterone and growth hormone).
Estimates of Change in Total BMD
For the basic models, we estimated the percent change in total BMD over 12 months at selected levels of the determinant of interest. For example, the estimate for strength training in males from the basic model is percent change in total BMD = intercept + β1time in interval + β2Male + β3strength training. For the multivariate model, we estimated the percent change in total BMD for selected levels of all variables of interest in the model, adjusted for age, race, and smoking, the male and postmenopausal women dummy variables, and time on other ART. All estimates assume 3 years of other ART use and were calculated separately for males, premenopausal women, and postmenopausal women. Age, BMI, and albumin were mean centered at 45 years, 25 kg/m2, and 4.0 g/dL, respectively, so that the intercept for each model reflects the mean value, not zero, for each of these variables. For example, the estimate for a male on d4T would be percent change in total BMD = intercept + β1(time in interval = 1 year) + β2age + β3African American + β4male + β5smoker + β6albumin + β7strength training + β8BMI + β9Time on d4T + β10Time on Other ART. For race, smoking status, and strength training, we used the values for white (African American = 0), smoker = 0, and strength training = 0 for all estimates, except when we made estimates for strength training, we then used the value 1 for strength training. For estimates of men, premenopausal females, and postmenopausal females, age was specified as 45, 40, and 51, respectively, based on the median age in each group over all the intervals. The P values given in Tables 3 and 4 test the hypothesis that there is no association between change in total BMD and the continuous or categorical row variable, using the robust Wald test. For some ART agents, strength training and types of steroid use, there were few or no premenopausal or postmenopausal women who used the agent or practiced strength training. In those cases, no estimate was given in Table 3 or Table 4 for the variable.
A total of 799 intervals were available for 379 individuals. The median (25th, 75th) number of intervals included per person was 2 (1, 3), range 1-6. The date at the beginning of the first interval ranged from August 1996 until August 2002. The median (25th, 75th) number of months within the DXA intervals was 12.8 (12.0-16.5). The median total time of follow-up from first DXA visit to last DXA visit was 2.5 years (interquartile range 1.2-3.8; range 0.9-6.8 years).
There were 283 men, 76 premenopausal women, and 20 postmenopausal women in the study with a median age of 43, 38, and 49 years, respectively. Three of the 20 postmenopausal women became postmenopausal during follow-up, and 8 of 20 reported ever having a hysterectomy. Nine premenopausal women reported ever having a hysterectomy. Compared with women in this study, men were older; more likely to be white; less likely to be obese; reported higher intakes of calories, calcium, and vitamin D; and had lower CD4 cell counts (Table 1). Intake was inadequate for vitamin D (<10 μg) and calcium (<1000 mg) for 43% and 46% of men, 63% and 71% of premenopausal women, and 53% and 82% of postmenopausal women, respectively. At the end of follow-up (data not shown), the median age of the postmenopausal women, premenopausal women, and men in our cohort was 51, 41, and 46 years, respectively. Use and duration of use until the beginning of the first DXA interval are shown in Table 2 for different ART agents and HAART regimens, stratified by sex.
Total BMD at Baseline by Sex, Race, and Menopausal Status
At baseline, the estimated total BMD was 1.12 g/cm2 for non-African American men (includes white, Hispanic, and Asian) and 1.20 g/cm2 for African American men (P < 0.0001) (data not shown). For premenopausal women, total BMD was 1.09 g/cm2 for non-African American women and 1.13 g/cm2 for African American women (P = 0.04). Finally, among postmenopausal women, total BMD was 1.04 for non-African American women and 1.08 g/cm2 for African American women (P = 0.41).
Estimates of Percent Change in Total BMD Over DXA Intervals by Sex and Menopausal Status
Estimates of the percent change in total BMD over 1 year are shown in Table 3 for each selected level of the determinant in each basic model. They are presented for men, premenopausal women, and postmenopausal women separately.
Estimates of the percent change in total BMD over 1 year for selected levels of each determinant in the multivariate model adjusted for time in interval, age, race, smoking, and the sex and menopausal dummy variables are shown in Table 4. Total BMD did not vary significantly by age, race, and time in the interval after adjusting for other variables. Premenopausal women tended to have less bone loss than men (P = 0.065), whereas postmenopausal women lost almost 3 times the BMD than premenopausal women (P = 0.012). For example, for participants on other ART for at least 3 years, on average, men lost −0.57% per year (95% CI −1.00 to −0.14), premenopausal women lost −0.28% per year (95% CI −0.71 to −0.15), and postmenopausal women lost −1.27% per year (95% CI −2.07 to −0.47). The P values for all other variables of interest are shown in Table 3. There were no significant interactions between sex or menopausal status and other terms in the model.
The adjusted estimates described in this paragraph are for men, and the estimates for women can be seen in Table 3. Higher albumin, higher BMI, and strength training were associated with less bone loss, after adjusting for other determinants of change in BMD. For example, the estimated 1-year percent change in total BMD in a male with an albumin level of 3.5 mg/dL was −0.76% per year (95% CI −1.23 to −0.30) and −0.57% per year (95% CI −1.00 to −0.14) for an albumin of 4.0 mg/dL. For men with BMI of 20, 25, and 30 kg/m2, the estimated percent change in total BMD was −0.78% per year (95% CI −1.21 to −0.36), −0.57% per year (95% CI −1.00 to −0.14), and −0.36% per year (95% CI −0.85 to 0.13), respectively. On average, men who practiced strength training had stable BMD (−0.15% per year, 95% CI −0.52 to 0.21), whereas those who did not practice strength training had a decrease in total BMD (−0.57% per year, 95% CI −1.00% to −0.14%). Men on no ART had a decline in BMD −0.37% per year (95% CI −0.76% to −0.02%). Men on prednisone/hydrocortisone lost 2 times the amount of BMD than nonusers. Decreases in total BMD were greater with increasing time on didanosine (ddI) and with tenofovir use. Use vs. nonuse of tenofovir (P = 0.003) was included in the model as it was more significant than duration of use (P = 0.10). The estimated percent change in total BMD in men was −0.84% per year (95% CI −1.32 to −0.35) for men on ddI for 1 year and −1.37% per year (95% CI −2.10 to −0.64) for men on ddI for 3 years. The estimated change in total BMD for men on tenofovir was −2.04% (95% CI −3.00 to −1.08). In contrast, long durations of d4T and saquinavir were associated with less loss in total BMD.
Most participants who were on these agents at the beginning of the interval were on it throughout the interval (percent on throughout interval): d4T (65%), tenofovir (77.3%), saquinavir (71.3%), and ddI (69.4%). There was no significant difference in the percent change in total BMD by CD4 cell count, HIV viral load, years known HIV-infected, smoking, testosterone/growth hormone, calcium or vitamin D intake, hormonal contraceptive use or other individual ART.
In this study, we followed an ethnically diverse group of HIV-infected men and women to study the effect of various treatment regimens on change in total body BMD over time with detailed adjustment for known risk factors. There were 3 main findings. First, postmenopausal women had more bone loss than premenopausal women or men. Second, higher albumin, higher BMI, and strength training were all associated with less bone loss. Third, use of tenofovir, ddI, or catabolic steroids was associated with more loss and use of d4T or saquinavir was associated with less loss, compared with nonusers. Markers of HIV disease severity (lower CD4 cell count and higher HIV viral load), history of wasting, dietary intake of calcium, testosterone/growth hormone, smoking, and race were not independently associated with changes in BMD.
Many of the important risk factors for loss of BMD in our cohort are established risk factors for bone loss in the general population. These include menopause, low BMI, lack of strength training, and prednisone/hydrocortisone use. In the general population, peak bone mass is achieved between 20 and 25 years of age in both sexes32 and decreases slowly after 35 years of age33 with accelerated loss during the perimenopausal years.33,34 Menopausal women had greater loss than premenopausal women and men in our cohort. In contrast, men tended to have more loss than women who had not reached menopause and less bone loss than postmenopausal women. With larger numbers of premenopausal and postmenopausal women, we would have had more power to detect differences. The reasons for suggested differences are not clear. The premenopausal women were on average 5 years younger than the men. Many of these are probably not perimenopausal. Age was not an independent risk factor, and the rate of change over time did not differ across race. For each sex, total BMD was higher in African Americans compared with other races. This corresponds to population surveys in which blacks have higher BMD than other races. Blacks also have a lower rate of hip fracture than whites, Hispanics, and Asians.35 We did not have data on fractures.
A recent meta-analysis in HIV showed that much of the difference in BMD between HIV-positives and HIV-negatives disappeared after adjustment for weight.36 In the general population, positive correlations of BMD with weight, frame size,37 or BMI33,38 are thought to be largely due to the effect of mechanical force on bone formation. Within our HIV-infected patients, BMI was an important factor for bone loss but a past episode of wasting did not predict future bone loss. Although wasting still occurs in HAART users,39,40 episodes of wasting are less severe and may explain the lack of association in our study.41
Exercise training of various types among persons with HIV, including aerobic exercise and resistance exercise, has been shown to improve muscle mass,42,43 strength,44,45 physical functioning,43,46,47 and a variety of cardiovascular and metabolic parameters.48 This is the first report we are aware of that indicates that strength training may also reduce bone loss in HIV-infected patients. Maintaining adequate weight and engaging in physical activity may help to preserve and perhaps improve BMD in HIV-infected patients.
Participants who used catabolic steroids such as prednisone or hydrocortisone also had greater bone loss than those who did not, whereas testosterone and growth hormone did not explain gain in BMD. Vitamin D insufficiency is an important risk factor for osteoporosis.49 In our cohort, 43% of men, 63% of premenopausal women, and 53% of postmenopausal women had inadequate vitamin D intake. Low vitamin D intake was not a significant predictor of bone loss in men (P = 0.77) or women (P = 0.11). Dietary and supplement intake may not adequately reflect body levels as patients with HIV may have other complications, such as malabsorption or chronic diarrhea, which may prevent absorption of dietary nutrients, and renal disease may inhibit 1,25 vitamin D production. Moreover, there may not be enough variation in intake in either males or females in our cohort to detect an effect on bone.
Common markers of HIV disease severity, including CD4 and viral load, were not predictive of bone loss in our cohort of HIV-infected adults. However, our patients with low albumin were at increased risk of bone loss. This has been observed by others.50 Although the mechanism is unclear, low albumin predicts decreasing CD4, development of AIDS, and death51-56 and probably reflects a combination of factors including poor nutrition, liver disease, concurrent infections, and proinflammatory cytokines.
There are few relevant animal studies or clinical trials showing the effect of specific antiretroviral agents on bone mineralization. Our finding that participants on tenofovir had greater decreases in total BMD over time is supported by previous studies.19,20-22 To our knowledge, there are no other reports of the potentially deleterious effects of ddI or the potentially protective effects of saquinavir on BMD. There are a few studies on the association between d4T and BMD. In a previous study in our cohort, McDermott et al18 observed increases in trunk BMC in those on d4T. Gallant et al21 reported decreases of 1.0% in the spine and 2.4% in the hip over 3 years in antiretroviral-naive patients who began d4T-based HAART. The authors did not comment on the statistical significance of these estimates. We estimate a rate of approximately 0.3% increase over 3 years in total body BMD in men using d4T for at least 3 years. In contrast, to the former study, our participants were not antiretroviral-naive. In addition, we could not determine why each individual started, continued, or discontinued a particular treatment. Finally, total body BMD, which is primarily cortical bone, is less metabolically active and less-sensitive to acute changes in medication and health compared with the spinal BMD. Total BMD reflects the longer term impact of risk factors and may change more slowly. Thus, factors associated with hip and spine may differ from those that affect total BMD.57-59 Future studies are needed to understand how each bone site predicts fractures and health outcomes in this population.
Longitudinal studies offer a unique opportunity to examine changes in BMD over time. Nevertheless, in contrast to a clinical trial, it is not always possible to disentangle the effects of HIV disease and therapy in treatment-experienced patients. That is, patients may continue an agent or switch to another agent based on CD4, viral load, viral resistance, agent-related symptoms, metabolic disturbances, or interaction with concomitant medications. Also, there are counteracting effects. Sicker patients may be less physically active, have more weight loss, and have increased cytokine activity and hypogonadism, whereas healthier patients may experience the metabolic side effect of specific ART, including mitochondrial toxicity. We measured agent use before changes in BMD, so we feel that temporal bias is unlikely to explain our findings. In addition, in contrast to many studies in HIV, we included variables for disease activity, diet, exercise, and steroid use. Some of the regimens in our analysis may not be used to treat HIV disease today. Similar analyses in current cohorts will shed light on the new ART agents. Finally, we lacked biochemical measures such as estrogen or testosterone levels, markers of bone mineralization, and chronic inflammation that may further explain our findings.
Our findings suggest that in the context of HIV infection, maintaining adequate weight and nutritional status and performing strength training exercises may help to combat HIV-associated bone loss. Overall, we observed small changes in total BMD, and the risk of fractures is not known. Further study is needed to assess the effects of individual antiretroviral medications on bone health, especially among postmenopausal women, and to assess fracture rates. Patients on specific agents, such as tenofovir, ddI, or catabolic steroids, may require close monitoring of BMD.
This article is dedicated to the memory of Abby Shevitz, MD, MPH. We wish to thank Jinyong Huang for his contributions to the data analysis, Sally Skinner for checking the programs, and Christine Wanke for her valuable comments on HIV disease and treatment.
1. Glesby MJ. Bone disorders in human immunodeficiency virus infection. Clin Infect Dis
2. Thomas J, Doherty SM. HIV
infection-a risk factor for osteoporosis. J Acquir Immune Defic Syndr
3. Fairfield WP, Finkelstein JS, Klibanski A, et al. Osteopenia in eugonadal men with acquired immune deficiency syndrome wasting syndrome. J Clin Endocrinol Metab
4. Moore AL, Vashisht A, Sabin CA, et al. Reduced bone mineral density
-positive individuals. AIDS
5. Knobel H, Guelar A, Vallecillo G, et al. Osteopenia in HIV
-infected patients: is it the disease or is it the treatment? AIDS
6. Loiseau-Peres S, Delaunay C, Poupon S, et al. Osteopenia in patients infected by the human immunodeficiency virus. A case control study. Joint Bone Spine
7. Arnsten JH, Freeman R, Howard AA, et al. HIV
infection and bone mineral density
in middle-aged women. Clin Infect Dis
8. Brown TT, Ruppe MD, Kassner R, et al. Reduced bone mineral density
in human immunodeficiency virus-infected patients and its association with increased central adiposity and postload hyperglycemia. J Clin Endocrinol Metab
9. Carr A, Miller J, Eisman JA, et al. Osteopenia in HIV
-infected men: association with asymptomatic lactic acidemia and lower weight pre-antiretroviral therapy
10. Huang JS, Wilkie SJ, Sullivan MP, et al. Reduced bone density in androgen deficient women with acquired immune deficiency syndrome wasting. J Clin Endocrinol Metab
11. Huang JS, Rietschel P, Hadigan CM, et al. Increased abdominal visceral fat is associated with reduced bone density in HIV
-infected men with lipodystrophy. AIDS
12. McDermott AY, Shevitz A, Knox T, et al. Effect of highly active antiretroviral therapy
on fat, lean, and bone mass in HIV
-seropositive men and women. Am J Clin Nutr
13. Nolan D, Upton R, McKinnon E, et al. Stable or increasing bone mineral density
-infected patients treated with nelfinavir or indinavir. AIDS
14. Tebas P, Yarasheski K, Henry K, et al. Evaluation of the virological and metabolic effects of switching protease inhibitor combination antiretroviral therapy
to nevirapine-based therapy for the treatment of HIV
infection. AIDS Res Hum Retroviruses
15. Mondy K, Yarasheski K, Powderly WG, et al. Longitudinal
evolution of bone mineral density
and bone markers in human immunodeficiency virus-infected individuals. Clin Infect Dis
16. Tsekes G, Chrysos G, Douskas G, et al. Body composition changes in protease inhibitor-naive HIV
-infected patients with two nucleoside reverse transcriptase inhibitors. HIV Med
17. Dolan SE, Kanter JR, Grinspoon S. Longitudinal
analysis of bone density in HIV
-infected women. J Clin Endocrinol Metab
18. McDermott AY, Terrin N, Wanke C, et al. CD4+ cell count, viral load, and highly active antiretroviral therapy
use are independent predictors of body composition alterations in HIV
-infected adults: a longitudinal
study. Clin Infect Dis
19. Van Rompay KKA, Brignolo LL, Meyer DJ, et al. Biological effects of short-term or prolonged administration of 9-[2-(phosphonomethoxy)propyl]adenine (tenofovir) to newborn and infant rhesus macaques. Antimicrob Agents Chemother
20. Castillo AB, Tarantal AF, Watnik MR, et al. Tenofovir treatment at 30 mg/kg/day can inhibit cortical bone mineralization in growing rhesus monkeys (macaca mulatta). J Orthop Res
21. Gallant JE, Staszewski S, Pozniak AI, et al. Efficacy and safety of tenofovir DF vs stavudine in combination therapy in antiretroviral-naive patients. JAMA
22. Guest JL, Rimland D, Patterson BY, et al. Tenofovir-induced nephrotoxicity in the first year of therapy. Paper presented at: Conference on Retroviruses and Opportunistic Infections; 2006; Denver, CO.
23. Shevitz AH, Knox TA, Spiegelman D, et al. Elevated resting energy expenditure among HIV
-seropositive persons receiving highly active antiretroviral therapy
24. Silva M, Skolnik PR, Gorbach SL, et al. The effect of protease inhibitors on weight and body composition in HIV
-infected patients. AIDS
25. Wilson IB, Roubenoff R, Knox TA, et al. Relation of lean body mass to health-related quality of life in persons with HIV
. J Acquir Immune Defic Syndr
26. Forrester JE, Spiegelman D, Tchetgen E, et al. Weight loss and body-composition changes in men and women infected with HIV
. Am J Clin Nutr
27. Knox TA, Spiegelman D, Skinner SC, et al. Diarrhea and abnormalities of gastrointestinal function in a cohort of men and women with HIV
infection. Am J Gastroenterol
28. Jacobson DL, Knox TA, Spiegelman D, et al. Prevalence, evolution and risk factors for fat atrophy and fat deposition in a cohort of HIV
-infected men and women. Clin Infect Dis
29. Federal Register. Annual Update of the HHS Poverty Guidelines
30. Sallis JF, Haskell WL, Wood PD, et al. Physical activity assessment methodology in the five-city project. Am J Epidemiol
31. Durrleman S, Simon R. Flexible regression models with cubic splines. Stat Med
32. Nguyen TV, Maynard LM, Towne B, et al. Sex differences in bone mass acquisition during growth: the fels longitudinal
study. J Clin Densitom
33. Riggs BL, Melton LJ. Osteoporosis: Etiology, Diagnosis and Management
. Vol 2. Philadelphia, PA: Lippincott-Raven; 1995.
34. Favus MJ, ed. Primer on the Metabolic Bone Disorders and Disorders of Mineral Metabolism
. 5th ed. Washington, DC: American Society for Bone and Mineral Research; 2003.
35. Cooper C. Epidemiology of osteoporosis. In: Favus MJ, ed. Primer on the Metabolic Bone Diseases and Disorders of Mineral Metabolism
. Vol 5. Washington, DC: American Society for Bone and Mineral Research; 2003:311.
36. Bolland MJ, Grey AB, Gamble GD, et al. Clinical review: low body weight mediates the relationship between HIV
infection and low bone mineral density
: a meta-analysis. J Clin Endocrinol Metab
37. Chumlea WC, Wisemandle W, Guo SS, et al. Relations between frame size and body composition and bone mineral status. Am J Clin Nutr
38. Bakker I, Twisk JW, Van Mechelen W, et al. Fat-free body mass is the most important body composition determinant of 10-yr longitudinal
development of lumbar bone in adult men and women. J Clin Endocrinol Metab
39. Wanke CA, Silva M, Knox TA, et al. Weight loss and wasting remain common complications in individuals infected with human immunodeficiency virus in the era of highly active antiretroviral therapy
. Clin Infect Dis
40. Tang AM, Jacobson DL, Spiegelman D, et al. Increasing risk of 5% or greater unintentional weight loss in a cohort of HIV
-infected patients, 1995 to 2003. J Acquir Immune Defic Syndr
41. Lawal A, Engelson ES, Wang J, et al. Equivalent osteopenia in HIV
-infected individuals studied before and during the era of highly active antiretroviral therapy
42. Driscoll SD, Meininger GE, Lareau MT, et al. Effects of exercise
training and metformin on body composition and cardiovascular indices in HIV
-infected patients. AIDS
43. Shevitz AH, Wilson IB, McDermott AY, et al. A comparison of the clinical and cost-effectiveness of 3 intervention strategies for AIDS wasting. J Acquir Immune Defic Syndr
44. Dolan SE, Frontera W, Librizzi J, et al. Effects of a supervised home-based aerobic and progressive resistance training regimen in women infected with human immunodeficiency virus: a randomized trial. Arch Intern Med
45. Perez-Moreno F, Camara-Sanchez M, Tremblay JF, et al. Benefits of exercise
training in spanish prison inmates. Int J Sports Med
46. Roubenoff R, Wilson IB. Effect of resistance training on self-reported physical functioning in HIV
infection. Med Sci Sports Exerc
47. Fillipas S, Oldmeadow LB, Bailey MJ, et al. A six-month, supervised, aerobic and resistance exercise
program improves self-efficacy in people with human immunodeficiency virus: a randomised controlled trial. Aust J Physiother
48. Driscoll SD, Meininger GE, Ljungquist K, et al. Differential effects of metformin and exercise
on muscle adiposity and metabolic indices in human immunodeficiency virus-infected patients. J Clin Endocrinol Metab
49. Bischoff-Ferrari HA, Giovannucci E, Willett WC, et al. Estimation of optimal serum concentrations of 25-hydroxyvitamin D for multiple health outcomes. Am J Clin Nutr
50. Fernandez-Rivera J, Garcia R, Lozano F, et al. Relationship between low bone mineral density
and highly active antiretroviral therapy
including protease inhibitors in HIV
-infected patients. Hiv Clin Trials
51. Melchior JC, Niyongabo T, Henzel D, et al. Malnutrition and wasting, immunodepression, and chronic inflammation as independent predictors of survival in HIV
-infected patients. Nutrition
52. Brechtl JR, Patrick PA, Visintainer P, et al. Predictors of death within six months in patients with advanced AIDS. Palliat Support Care
53. Olawumi HO, Olatunji PO. The value of serum albumin in pretreatment assessment and monitoring of therapy in HIV
/AIDS patients. HIV Med
54. Mehta SH, Astemborski J, Sterling TR, et al. Serum albumin as a prognostic indicator for HIV
disease progression. AIDS Res Hum Retroviruses
55. Bonarek M, Morlat P, Chene G, et al. Prognostic score of short-term survival in HIV
-infected patients admitted to medical intensive care units. Int J STD AIDS
56. Afessa B, Green B. Clinical course, prognostic factors, and outcome prediction for HIV
patients in the ICU. The PIP (pulmonary complications, ICU support, and prognostic factors in hospitalized patients with HIV
) study. Chest
57. Blake GM, Fogelman I. The role of bone density measurements in the evaluation of new treatments for osteoporosis. Curr Pharm Des
58. Blake GM, Herd RJ, Patel R, et al. The effect of weight change on total body dual-energy X-ray absorptiometry: results from a clinical trial. Osteoporos Int
59. Blake GM, Knapp KM, Spector TD, et al. Predicting the risk of fracture at any site in the skeleton: are all bone mineral density
measurement sites equally effective? Calcif Tissue Int