Interventions for the treatment of decreased bone mineral density associated with HIV infection

Abstract

Background

Decreased bone mineral density (BMD) occurs more commonly in patients with HIV than in the general population, making this group more susceptible to fragility fractures. However, bone loss is under‐treated in patients with HIV.

Objectives

To assess the effects of interventions aimed at increasing bone mineral density in HIV‐infected adults.

Search methods

We searched MEDLINE, EMBASE, LILACS, The Cochrane Library, Meeting Abstracts, AIDSTRIALS, ACTIS, Current Controlled Trials, National Institutes of Health Clinical Trials Registry, and CenterWatch (search date July 2006).

Selection criteria

Randomised trials comparing any pharmacological or non‐pharmacological therapy with placebo, no treatment, or an alternative therapy, with the goal of increasing bone mineral density in adult (age 18 years or over) patients with HIV.

Data collection and analysis

Two reviewers independently assessed trial eligibility and quality, and extracted data. Where data were incomplete or unclear, conflicts were resolved with discussion and/or trial authors were contacted for further details.

Main results

Three completed randomised‐controlled studies examined the role of alendronate in patients with HIV and osteopenia or osteoporosis. When all three studies were combined, much heterogeneity was seen (p<0.0001), most likely due to different populations and interventions. A sensitivity analysis showed that in two studies without heterogeneity (p=0.11), alendronate, calcium and vitamin D improved lumbar BMD after one year when compared with calcium and vitamin D (weighted mean difference +2.65 95% confidence interval (CI) 0.80, 4.51 percent). However the alendronate group did not have less fragility fractures, relative risk (RR) 1.28 (95% CI 0.20, 8.21), or osteoporosis, RR 0.50 (95% CI 0.24, 1.01). Adverse events were not significantly different between groups, RR 1.28 (95% 0.20, 8.21). One randomised‐controlled study done in patients with AIDS wasting found that after three months, testosterone enanthane improved lumbar BMD compared to placebo by +3.70 (95% CI 0.48, 6.92) percent, but progressive resistance training did not improve lumbar BMD (+0.40 95% CI ‐2.81, 3.61 percent). No group in this study had any adverse effects.

Authors' conclusions

The very limited data reviewed showed that bisphosphonate therapy andin those with AIDS wasting syndrome, testosteronemay be safe and possibly effective methods to improve bone mineral density in HIV patients. The available studies are small, of short duration, and not powered to detect changes in WHO categories and fracture rates.

Larger studies using bisphosphonates are currently underway. The role of colecalciferol, androgen replacement in women, and growth hormone are also under investigation.

Plain language summary

Drug‐based and non‐drug‐based interventions to improve the bone mineral density in patients living with HIV

Osteoporosis is caused by bone loss, and people who have the condition are at higher risk of having a fracture. Measuring a person's bone mass density (BMD) is a way to measure his or her risk of having a fracture due to fragile bones. Decreased BMD is much more common in HIV patients than in the general population. The cause of this decrease is not certain, but it may be because of the HIV infection itself or because of the antiretroviral medications that patients with HIV take. Although patients with HIV often get fractures because of their sometimes more fragile bones, it has been shown that this bone loss is often not effectively treated in this population. This review examines the randomised controlled trials investigating treatments for bone loss in patients with HIV infection.

Three trials examined the use of the drug alendronate to improve BMD in patients with HIV. These three studies were quite different from each other in terms of the populations studied and the interventions used, but even similar studies did not always have heterogeneity. A fourth study examined the use of testosterone in male patients with HIV and AIDS wasting syndrome. The four studies in this review were limited by the fact they were all very small and lasted a short amount of time, and thus they were unable to detect prevention of fractures or changes in number of patients with osteoporosis. There were also further compromises in study design. However, the limited available data show that there may be safe and perhaps effective treatments in the form of alendronate for patients with HIV who have decreased bone mineral density and, in those with AIDS wasting syndrome, testosterone.

Larger studies further examining the issue of decreased BMD are currently underway.

Background

Osteoporosis is a state of low bone mass and microarchitectural deterioration of bone tissue, leading to bone fragility and increased risk of fracture (NIH 1993). It is defined by the presence of a fragility fracture or a t‐score of ‐2.5 or less, meaning a bone mass density (BMD) 2.5 standard deviations below that of a healthy young population (Kanis 1994). Individuals with a bone mineral density between 1.0 and 2.49 are defined as having osteopenia (Kanis 1994). BMD is the most readily quantifiable predictor of fracture risk in patients who do not have a history of fragility fracture when viewed in context with the patient's age (Kanis 2005). Increased age, prior fragility fracture, parental history of hip fractures, smoking, use of systemic corticosteroids, excess alcohol intake, and rheumatoid arthritis are other major risk factors for future fracture (Kanis 2005).

Decreased bone mineral density also occurs more commonly in HIV patients than in the general population (Bruera 2003). In some studies, the use of antiretroviral therapy, in particular those with protease inhibitors, is associated with osteopenia or osteoporosis in as many as 50% of patients (Tebas 2000; Moore 2001; Fernandez 2003). Other studies have shown that HIV‐infected patients have increased rates of osteopenia and osteoporosis regardless of treatment, suggesting that HIV infection may directly decrease BMD (Amiel 2004; Landonio 2004). Lipodystrophy, a common side effect of antiretroviral therapy, has also been occasionally associated with reduced BMD (Huang 2001; Brown 2004). Finally, overlapping risk factors such as female gender, increased age, low body mass index, steroid use, and smoking can contribute to reduced bone mineral density in these patients (Mondy 2003; Brown 2004).

Fragility fractures due to decreased BMD have been previously described, although prospective longitudinal studies are needed (McComsey 2004). Bone loss in HIV patients may also be poorly treated. In one study, only 35% of 55 patients who had suffered a fragility fracture recalled having any therapy to treat their underlying bone disease (McComsey 2004). Calcium supplementation was the most commonly reported therapy, and only 10% received antiresorptive therapy. This systematic review aims to review the current evidence on treating bone loss in patients with HIV infection.

Objectives

To assess the effects of interventions aimed at increasing bone mineral density in HIV‐infected adults.

Methods

Criteria for considering studies for this review

Types of studies

Randomised trials, irrespective of language and publication status.

Types of participants

Adults (over 18 years) who have with HIV infection and decreased bone mineral density.

Types of interventions

Intervention: any pharmacological or non‐pharmacological therapy aimed at increasing bone mineral density in patients with HIV.

Control: placebo, no treatment, or an alternative therapy.

Types of outcome measures

Primary Outcome: mean change in bone mineral density and proportion with fractures.

Secondary Outcome: proportion with adverse events and proportion with osteoporosis and/or osteopenia.

Search methods for identification of studies

See: Cochrane HIV/AIDS Group methods used in reviews.

Two reviewers searched the following databases, filtering for randomised controlled trials: 
 ·MEDLINE 
 ·EMBASE 
 ·LILIACS 
 ·The Cochrane Library (CENTRAL) 
 ·Meeting Abstracts (http://gateway.nlm.nih.gov/meetingabstracts.html)

The following keywords were used: 
 Acquired Immunodeficiency Syndrome, HIV seropositivity, HIV, human immunodeficiency virus, HIV infection, osteopenia, osteoporosis, bone mineral density, exercise, diet, nutrition, alfacalcidol, calcitriol, vitamin K, alendronate/therapeutic use, metabolic/drug therapy, osteoporosis/drug therapy, vitamin D/therapeutic use, calcium/therapeutic use, diphosphonates/therapeutic use, bone diseases/drug therapy, calcitonin, etidronic acid, zoledronic acid, sparing, switch, switching, replacement, fluoride, hormone replacement therapy, testosterone, estrogen, progesterone, selective estrogen receptor modifiers, parathyroid hormone, raloxifene, risedronate, etidronate, caffeine, salt, protein, fatty acids, dietary fibre

The following databases were searched for ongoing randomised controlled trials: 
 ·AIDSTRIALS (http://aidsinfo.nih.gov/clinical_trials/); 
 ·ACTIS (AIDS Clinical Trials Information Service at http://www.actis.org/); 
 ·Current Controlled Trials (http://www.controlled‐trials.com/); 
 ·The National Institutes of Health Clinical Trials Registry (http://clinicaltrials.gov/); 
 ·CenterWatch (http://www.centerwatch.com/).

The citations of included trials and major reviews were handsearched for additional studies.

Data collection and analysis

Two reviewers reviewed the titles and potentially relevant abstracts retrieved by our search strategy. They obtained full articles of the relevant trials that fulfilled our inclusion criteria. Relevant articles were inspected independently by two reviewers and disagreements were solved by discussion. Both reviewers assessed the methodological quality of each trial in terms of allocation concealment, blinding, and inclusion of all randomised participants. Allocation concealment was classified as adequate (Grade A), unclear (Grade B), inadequate (Grade C), or not used (Grade D). Blinding was described as open (all parties are aware of treatment), single (either the participant or the care provider is aware of the treatment given), or double (neither the participant nor care provider know which treatment is given). Inclusion of all randomised participants was rated as Grade A (intention‐to‐treat analysis was possible and there were few losses to follow‐up), Grade B (exclusions after randomisation were less than 10%), and Grade C (exclusions were greater than 10% or were widely different between groups).

Both reviewers independently extracted the data using a data‐extraction form. For each of the studies, we extracted publication status, year, trial sponsor, duration of study follow‐up, study setting, number of randomised patients in each group, age, number of female patients, HIV risk factors, baseline osteoporosis risk factors, baseline t‐scores, presence of antiretroviral therapy, and HIV disease severity as measured by CD4‐positive lymphocyte count. We crosschecked data and solved discrepancies by discussion. If needed, study authors were contacted for clarification.

Data was analysed using Review Manager 4.2.8. The weighted mean difference method using inverse variance with 95% confidence intervals was used for continuous outcomes. The Mantel‐Haenszel method odds ratios with 95% confidence intervals were used for dichotomous outcomes. Heterogeneity among trials was assessed using chi‐square and I‐square statistics. Where heterogeneity was detected, heterogeneity was explored through stratifying by the above patient characteristics, using a random‐effects model, and/or excluding outlying trials. After all eligible studies were included in the primary analysis, sensitivity analyses were planned a priori, if needed, for each of the methodological quality factors. Funnel plots were planned a priori, if needed, to estimate the treatment effect against the precision of trialsin order to estimate asymmetry because of selection bias or methodological flaws. Subgroup analyses were planned a priori, if sufficient number of studies and data existed, to be conducted for subsets of studies (setting, control group intervention), and subgroups of patients (gender, ethnicity, AIDS status).

Results

Description of studies

Four published randomised controlled trials (Fairfield 2001; Guaraldi 2004; Mondy 2005; Negredo 2005) that examined interventions aimed at improving bone loss in patients with HIV were identified. Two of these studies examined the use of alendronate, calcium, and vitamin D supplementation compared with calcium and vitamin D supplementation alone (Guaraldi 2004; Mondy 2005). One of these studies examined the use of alendronate compared with no study drug, with both groups receiving dietary counselling to ensure adequate daily calcium intake (Negredo 2005). The last study used a 2 x 2 factorial design to examine the role of testosterone versus placebo and exercise versus no exercise (Fairfield 2001).

Negredo (Negredo 2005) enrolled 26 HIV‐infected patients who were on stable highly active antiretroviral therapy (HAART) and fulfilled the World Health Organization criteria of osteoporosis (t‐score of ‐2.5 or less) in a pilot study. Patients were randomly assigned to receive alendronate 70mg once a week plus dietary counselling at each study visit to ensure dietary calcium intake of 1200g/day, or dietary counselling alone. Patients were followed every 24 weeks for 96 weeks. A DEXA scan was performed at each study visit for the lumbar spine and hip. Twelve patients were allocated to receive alendronate and 14 were controls. One patient in the alendronate group was withdrawn. At baseline, the two groups were similar, except that the alendronate group was older and had lower dietary calcium intake. All patients were on HAART, but the study did not describe whether the same proportion of patients treated with protease inhibitors was in each group. The outcomes measured were proportion of patients fulfilling criteria for osteoporosis and improvement in BMD after 48 and 96 weeks.

Mondy (Mondy 2005) enrolled 31 HIV‐infected patients on stable HAART who had a lumbar spine BMD t‐score of less than ‐1 in a single‐centre pilot study. Patients were randomised to receive alendronate 70mg weekly or no alendronate for 48 weeks; calcium supplements (1000mg) and vitamin D supplements (400IU) were given to both groups. A DEXA scan was performed at baseline and at 24‐week intervals. Fifteen patients were randomised to receive 70mg of alendronate weekly with calcium and vitamin D, while 16 patients received only calcium and vitamin D. Age, gender, race, CD4 count, cigarette/alcohol use, weight, baseline body mass index, baseline BMD at the hip and lumbar spine, proportion with osteoporosis, and serum bone formation markers were not significantly different at baseline. Eighty‐seven percent of the total participants were male and the mean age was 44 ± 8.4. All patients were on HAART, with either one protease inhibitor or non‐nucleoside reverse transcriptase inhibitor, but the study did not describe if the same proportion of protease inhibitors treated patients were in each group. The primary endpoint was change in the BMD in the lumbar spine region. The study was powered to detect a 3% within‐arm difference in lumbar spine BMD at 48 weeks. Whole‐body BMD, adipose and lean tissue regional distributions, and bone metabolism markers were also obtained.

Guaraldi (Guaraldi 2004) enrolled 41 HIV‐infected patients on stable HAART who had a lumbar spine and hip t‐score of less than ‐1 in a multi‐centre study. Patients were randomised to receive alendronate 70mg weekly or no alendronate for 52 weeks; calcium 1000 mg and vitamin D 500 IU were provided to both groups. Patients were followed at 16, 32, and 52 weeks. A DEXA scan was performed at each visit for the lumbar spine and femoral neck. Eighteen participants were in the alendronate plus calcium and vitamin D group and 23 participants were in the calcium and vitamin D group. At baseline, the mean lumbar and femoral neck BMD, age, gender, body mass index, bone resorption markers, HIV viral load, and CD4 count were not significantly different between groups. Seventy‐one percent of the total participants were male. The mean age was 44.5 ± 3.6 (mean ± standard deviation) in the alendronate group and 42.5 ± 3.5 in the control group. All patients were on HAART, and had similar lengths of exposure to different classes of antiretroviral therapies, although it was unclear if the two groups had a similar proportion of patients treated with protease inhibitors during the time of the study. The outcomes measured were BMD at the lumbar spine and femoral neck, vertebral and nonvertebral fractures, and serum biochemical markers for bone resorption.

Fairfield (Fairfield 2001) enrolled 54 eugonadal (serum free testosterone >42 pmol/L) HIV‐infected men with AIDS wasting syndrome (weight <90% of ideal body weight or weight loss >10% baseline weight) between 27 and 51 years of age in a single‐centre 2x2 factorial study for 12 weeks. Patients were stratified for weight, above or below 90% ideal body weight, and randomized to receive testosterone enanthate (200mg/week intramuscularly) or placebo, and simultaneously to receive progressive resistance training three times per week with a licensed physical therapist or no training. A DEXA scan of the lumbar spine and proximal femur was performed at baseline and at 12 weeks. Four patients elected not to participate in the study before the baseline evaluation and seven did not complete the study. Twenty‐four men were randomized to testosterone and 26 men were randomized to exercise. The groups did not differ at baseline with respect to age, weight, body mass index, total body fat content, total body lean content, CD4 count, viral load, total testosterone, free testosterone, estrone, estradiol, or lumbar spine BMD. In this study, protease‐inhibitor use was not associated with significant differences in BMD or t‐scores. Seventy‐six percent of patients were receiving antiretroviral therapy and 72% were receiving HAART, but it was unclear if the groups had a similar proportion of patients treated with antiretroviral therapy during the time of the study. The average age was 38.1 ±3.5. At baseline, the average t‐scores for total hip were ‐0.65 ± 0.8 and for lumbar spine were 0.62 ±1.2. The main outcome was change in lumbar spine BMD after three months of treatment. Fasting weight and serum total and free testosterone levels, serum estradiol and estrone levels, and bone metabolism markers were also measured.

Risk of bias in included studies

Negredo (Negredo 2005) randomly assigned patients to treatment or control, but the method of allocation concealment was not explained (Grade B). The trial was an open‐label study (Grade C). Intention‐to‐treat analysis was not performed; one patient dropped out of the treatment group for an exclusion rate of 4% (Grade B).

Mondy (Mondy 2005) used pharmacy allocation and the allocation was appropriately concealed (Grade A). The trial was an open‐label study (Grade C). Intention‐to‐treat analysis was not performed; two patients dropped out of each group for an exclusion rate of 6% (Grade B).

Guaraldi (Guaraldi 2004) used closed‐envelope allocation prepared by the statistician for its single‐centre study and the allocation was appropriately concealed (Grade A). The trial was an open‐label study (Grade C). No patients dropped out of the study and intention‐to‐treat analysis was performed (Grade A).

Fairfield (Fairfield 2001) concealed allocation to the investigator (Grade A). The trial was a double‐blind placebo‐controlled study (Grade A). Intention‐to‐treat analysis was not performed; four patients dropped out before the baseline evaluation and seven men dropped out of the study before the end of the study evaluation for an exclusion rate of 20% (Grade C).

Effects of interventions

Bisphosphonates 
 We found three completed randomised controlled studies that examined the role of alendronate in patients with HIV and osteopenia or osteoporosis. All studies were missing some data and the authors were contacted for additional data (Guaraldi 2004; Mondy 2005; Negredo 2005). Two authors responded (Guaraldi 2004; Mondy 2005). One study (Negredo 2005) did not include a measure of variance, and standard deviations were imputed from the p‐values presented.

Bone Mass Density After One Year, Comparison Table 1, Outcomes 01 to 04 
 When all three studies are combined, a considerable amount of heterogeneity was seen (p<0.0001). The average treatment effect using the random‐effects model in the lumbar BMD was +4.00 (95% confidence interval (CI) ‐0.11, 8.11) percent, femoral neck BMD +3.19 (95% CI ‐1.79, 8.17) percent, total hip BMD 0.49 (95% CI ‐0.15, 1.13) percent, and trochanter BMD ‐0.27 (95% CI ‐0.81, 0.27) percent.

Fractures, Osteoporosis Category, and Adverse Events After One Year, Comparison Table 1, Outcomes 05 to 08 
 Alendronate, calcium and vitamin D did not decrease the number of fragility fractures after one year, relative risk (RR) 1.28 (95% CI 0.20, 8.21). Alendronate containing therapy also did not decrease the number of patients with osteoporosis after one year, RR 0.50 (95% CI 0.24, 1.01). Adverse events were not significantly different between groups in the meta‐analysis RR 1.28 (95% 0.20, 8.21). No treatment‐limiting adverse events occurred in the studies. There was no significant heterogeneity for the results in these comparisons (p=0.11 to p=0.80).

Bone Mass Density After Two Years, Comparison Table 1, Outcome 09 
 Alendronate and calcium intake counselling had a significant weighted mean difference in lumbar bone mineral density of +12.79 (95% CI 0.71, 24.87) percent after two years when compared with calcium‐intake counselling only (Negredo 2005). 
 
 Bone Mass Density After One Year, Sensitivity Analysis, Comparison Table 1, Outcomes 10 to 13 
 The two studies that examined the effect of alendronate, calcium, and vitamin D compared with calcium and vitamin D (Guaraldi 2004; Mondy 2005) were combined in a sensitivity analysis using a fixed‐effects model. Alendronate, calcium, and vitamin D had a significant weighted mean difference in lumbar BMD of +2.65 (95% CI 0.80, 4.51) percent after one year, but did not improve femoral neck BMD (+0.32, 95% CI ‐2.70, 3.34 percent). Mondy (Mondy 2005) also found no significant difference in total hip bone mineral density (+0.45 95% CI ‐0.20, 1.10 percent), and trochanter bone mineral density (‐0.27 95% CI ‐0.82, 0.28 percent) after one year. There was no significant heterogeneity for the results in these comparisons (p=0.11 to p=0.73). This sensitivity analysis also excluded the only study containing imputed data in the meta‐analysis and patients with a t‐score of less than ‐2.5.

Due to the very limited number of studies available for analysis and their variation in design and methodology, sensitivity analyses for each of the methodological quality factors do not add additional information and are not presented. Funnel plots are not available in random‐effects models and also are not presented for these studies.

Ongoing studies 
 No completed studies were found for other bisphosphonates. One ongoing study was found for zoledronate (Huang 2006) and two ongoing studies were found for alendronate (McComsey 2006; Rozenberg 2006). 
 
 Hormone therapy 
 Bone Mass Density After Three Months, Comparison Table 02, Outcomes 01 and 02 
 We found one randomised controlled study (Fairfield 2001) that examined the role of testosterone in male patients with HIV and AIDS wasting syndrome (comparison Table 2). These patients were found to have a below‐average t‐score in their total hip and lumbar spines. Compared to placebo, testosterone enanthane intramuscular injection (200mg/week) improved lumbar bone mineral density by +3.70 (95% CI 0.48, 6.92) percent after three months. Neither group had any adverse events. There were no completed studies found that examined the effect of female hormone replacement therapy, raloxifene, parathyroid hormone, or calcitonin. There is an ongoing study examining the effects of androgen replacement in female patients with HIV, relative testosterone deficiency, and AIDS wasting syndrome (Grinspoon 2006a). Another study is investigating the role of low‐dose growth hormone administration in patients with HIV who have reduced growth‐hormone secretion and increased visceral adiposity (Grinspoon 2006b). 
 
 Exercise 
 Bone Mass Density After Three Months, Comparison Table 03, Outcomes 01 and 02 
 We found one randomised controlled study (Fairfield 2001) that examined the role of testosterone in patients with HIV and AIDS wasting syndrome (comparison Table 3). These patients were found to have a below‐average t‐score in their total hip and lumbar spines. Compared to no therapy, progressive resistance training did not improved lumbar bone mineral density (0.40 95% CI ‐2.81, 3.61) percent after three months. Neither group had any adverse events.

Diet 
 No completed randomised controlled studies were found investigating the role of vitamin D, calcium, caffeine, salt, protein, fatty acids, dietary fibre, or micronutrients. A study is investigating the effect of colecalciferol on bone mineral density in HIV‐infected patients (van der Ven 2006).

Changes in HIV Therapy 
 No completed randomised controlled studies were found investigating the role of switching from one class of antiretroviral therapy to another to improve bone mineral density.

Discussion

The very limited data reviewed showed that bisphosphonate therapy may be a safe and possibly effective method to improve bone mineral density in HIV patients with bone loss. The three published studies examining alendronate were small, lasted only one to two years, and were not powered to detect small differences in bone mineral density. The power and length of treatment were also insufficient to determine improvements in WHO category status change and fracture rates. All three were open‐label studies, and two did not use intention‐to‐treat analysis (Mondy 2005; Negredo 2005).

Considerable heterogeneity occurred among the three alendronate studies (p<0.0001), most likely due to different populations (t‐score less than 2.5 in one study, and t‐score less than 1.0 in two studies) and different interventions (alendronate with calcium intake counselling versus calcium intake counselling alone in one study, and alendronate with calcium and vitamin D supplementation versus calcium and vitamin D supplementation alone in two studies). Imputed data and methodological quality differences may also be contributing factors. In the two studies examining the effect of alendronate, calcium, and vitamin D compared to calcium and vitamin D alone, heterogeneity was not present and an improvement in lumbar BMD was demonstrated after one year.

The use of testosterone in male patients with HIV and AIDS wasting syndrome, but not all living with osteoporosis or osteopenia, was examined in one study and significantly improved bone mineral density after only three months (Fairfield 2001). In the same study, exercise did not improve bone mineral density but a power analysis was not performed. Intention‐to‐treat analysis was also not done. Improvements in WHO category status change and fracture rates were not reported and there was likely insufficient power and length of treatment to detect these measurements for either testosterone or exercise. Although no adverse events were reported, adverse effects due to long‐term use of testosterone are not known for this setting.

No randomised controlled trials were found examining the role of changes in HIV therapy on bone loss. Given that certain antiretroviral therapies may increase risk of decreased bone mineral density, the four trials included in this review did not provide data showing whether the antiretroviral therapies were similar between groups. It is unclear if differences in antiretroviral therapies between groups were a confounding factor in the bone mineral density changes.

Larger studies using bisphosphonates are currently underway (Huang 2006; McComsey 2006; Rozenberg 2006). Trials are also underway to examine the role of androgen replacement in women and growth hormone in men and women (Grinspoon 2006a; Grinspoon 2006b). One trial is examining the role of colecalciferol (van der Ven 2006).

Authors' conclusions

Implications for practice.

Limited data show that there may be safe and perhaps effective treatments in the form of alendronate for patients with HIV who have decreased bone mineral density and, in those with AIDS wasting syndrome, testosterone. These data are limited because of their small size, relatively short study duration, and some compromises in study design.

Implications for research.

Long‐term efficacy and safety data regarding the use of bisphosphates and testosterone in larger studies will verify the results of these pilot studies. Other interventions to treat and prevent bone loss in patients with HIV need further investigation.

What's new

Date	Event	Description
29 October 2008	Amended	Converted to new review format.

History

Protocol first published: Issue 1, 2006
 Review first published: Issue 2, 2007

Date	Event	Description
12 February 2007	New citation required and conclusions have changed	Substantive amendment

Data and analyses

Comparison 1. Alendronate compared to regular treatment.

Outcome or subgroup title	No. of studies	No. of participants	Statistical method	Effect size
1 Percent lumbar bone mineral density change after one year ‐ primary analysis	3	97	Mean Difference (IV, Random, 95% CI)	4.00 [‐0.11, 8.11]
1.1 Alendronate, vitamin D and calcium compared with vitamin D and calcium	2	72	Mean Difference (IV, Random, 95% CI)	2.11 [‐1.24, 5.47]
1.2 Alendronate and calcium intake counselling compared with calcium intake counselling	1	25	Mean Difference (IV, Random, 95% CI)	7.61 [6.28, 8.94]
2 Percent femoral neck bone mineral density change after one year ‐ primary analysis	3	97	Mean Difference (IV, Random, 95% CI)	3.19 [‐1.79, 8.17]
2.1 Alendronate, vitamin D and calcium compared with vitamin D and calcium	2	72	Mean Difference (IV, Random, 95% CI)	0.32 [‐2.70, 3.34]
2.2 Alendronate and calcium intake counselling compared with calcium intake counselling	1	25	Mean Difference (IV, Random, 95% CI)	5.96 [5.36, 6.56]
3 Percent total hip bone mineral density change after one year ‐ primary analysis	3	97	Mean Difference (IV, Random, 95% CI)	0.48 [‐0.16, 1.11]
3.1 Alendronate, vitamin D and calcium compared with vitamin D and calcium	2	72	Mean Difference (IV, Random, 95% CI)	0.44 [‐0.20, 1.08]
3.2 Alendronate and calcium intake counselling compared with calcium intake counselling	1	25	Mean Difference (IV, Random, 95% CI)	2.47 [‐2.02, 6.96]
4 Percent trochanter bone mineral density change after one year ‐ primary analysis	3	97	Mean Difference (IV, Random, 95% CI)	‐0.27 [‐0.80, 0.27]
4.1 Alendronate, vitamin D and calcium compared with vitamin D and calcium	2	72	Mean Difference (IV, Random, 95% CI)	‐0.26 [‐0.81, 0.28]
4.2 Alendronate and calcium intake counselling compared with calcium intake counselling	1	25	Mean Difference (IV, Random, 95% CI)	‐0.41 [‐3.31, 2.49]
5 Proportion with new fractures after one year	3	74	Risk Ratio (M‐H, Fixed, 95% CI)	1.28 [0.20, 8.21]
5.1 Alendronate, vitamin D and calcium compared with vitamin D and calcium	2	72	Risk Ratio (M‐H, Fixed, 95% CI)	1.28 [0.20, 8.21]
5.2 Alendronate and calcium intake counselling compared with calcium intake counselling	1	2	Risk Ratio (M‐H, Fixed, 95% CI)	0.0 [0.0, 0.0]
6 Proportion with osteoporosis after one year	3	97	Risk Ratio (M‐H, Fixed, 95% CI)	0.50 [0.24, 1.01]
6.1 Alendronate, vitamin D and calcium compared with vitamin D and calcium	2	72	Risk Ratio (M‐H, Fixed, 95% CI)	1.02 [0.32, 3.26]
6.2 Alendronate and calcium intake counselling compared with calcium intake counselling	1	25	Risk Ratio (M‐H, Fixed, 95% CI)	0.29 [0.11, 0.78]
7 Proportion with treatment‐limiting adverse reaction after one year	3	97	Risk Ratio (M‐H, Fixed, 95% CI)	0.0 [0.0, 0.0]
7.1 Alendronate, vitamin D and calcium compared with vitamin D and calcium	2	72	Risk Ratio (M‐H, Fixed, 95% CI)	0.0 [0.0, 0.0]
7.2 Alendronate and calcium intake counselling compared with calcium intake counselling	1	25	Risk Ratio (M‐H, Fixed, 95% CI)	0.0 [0.0, 0.0]
8 Proportion with any adverse reaction after one year	3	97	Risk Ratio (M‐H, Fixed, 95% CI)	1.28 [0.20, 8.21]
8.1 Alendronate, vitamin D and calcium compared with vitamin D and calcium	2	72	Risk Ratio (M‐H, Fixed, 95% CI)	1.28 [0.20, 8.21]
8.2 Alendronate and calcium intake counselling compared with calcium intake counselling	1	25	Risk Ratio (M‐H, Fixed, 95% CI)	0.0 [0.0, 0.0]
9 Percent lumbar bone mineral density change after two years	1	25	Mean Difference (IV, Fixed, 95% CI)	12.79 [0.71, 24.87]
9.2 Alendronate and calcium intake counselling compared with calcium intake counselling	1	25	Mean Difference (IV, Fixed, 95% CI)	12.79 [0.71, 24.87]
10 Percent lumbar bone mineral density change after one year ‐ Alendronate, vitamin D and calcium	2	72	Mean Difference (IV, Fixed, 95% CI)	2.65 [0.80, 4.51]
10.1 Alendronate, vitamin D and calcium compared with vitamin D and calcium	2	72	Mean Difference (IV, Fixed, 95% CI)	2.65 [0.80, 4.51]
11 Percent femoral neck bone mineral density change after one year ‐ Alendronate, vitamin D and calcium	2	72	Mean Difference (IV, Fixed, 95% CI)	0.32 [‐2.70, 3.34]
11.1 Alendronate, vitamin D and calcium compared with vitamin D and calcium	2	72	Mean Difference (IV, Fixed, 95% CI)	0.32 [‐2.70, 3.34]
12 Percent total hip bone mineral density change after one year ‐ Alendronate, vitamin D and calcium	2	72	Mean Difference (IV, Fixed, 95% CI)	0.44 [‐0.20, 1.08]
12.1 Alendronate, vitamin D and calcium compared with vitamin D and calcium	2	72	Mean Difference (IV, Fixed, 95% CI)	0.44 [‐0.20, 1.08]
13 Percent trochanter bone mineral density change after one year ‐ Alendronate, vitamin D and calcium	2	72	Mean Difference (IV, Fixed, 95% CI)	‐0.26 [‐0.81, 0.28]
13.1 Alendronate, vitamin D and calcium compared with vitamin D and calcium	2	72	Mean Difference (IV, Fixed, 95% CI)	‐0.26 [‐0.81, 0.28]

1.1. Analysis.

Comparison 1 Alendronate compared to regular treatment, Outcome 1 Percent lumbar bone mineral density change after one year ‐ primary analysis.

1.2. Analysis.

Comparison 1 Alendronate compared to regular treatment, Outcome 2 Percent femoral neck bone mineral density change after one year ‐ primary analysis.

1.3. Analysis.

Comparison 1 Alendronate compared to regular treatment, Outcome 3 Percent total hip bone mineral density change after one year ‐ primary analysis.

1.4. Analysis.

Comparison 1 Alendronate compared to regular treatment, Outcome 4 Percent trochanter bone mineral density change after one year ‐ primary analysis.

1.5. Analysis.

Comparison 1 Alendronate compared to regular treatment, Outcome 5 Proportion with new fractures after one year.

1.6. Analysis.

Comparison 1 Alendronate compared to regular treatment, Outcome 6 Proportion with osteoporosis after one year.

1.7. Analysis.

Comparison 1 Alendronate compared to regular treatment, Outcome 7 Proportion with treatment‐limiting adverse reaction after one year.

1.8. Analysis.

Comparison 1 Alendronate compared to regular treatment, Outcome 8 Proportion with any adverse reaction after one year.

1.9. Analysis.

Comparison 1 Alendronate compared to regular treatment, Outcome 9 Percent lumbar bone mineral density change after two years.

1.10. Analysis.

Comparison 1 Alendronate compared to regular treatment, Outcome 10 Percent lumbar bone mineral density change after one year ‐ Alendronate, vitamin D and calcium.

1.11. Analysis.

Comparison 1 Alendronate compared to regular treatment, Outcome 11 Percent femoral neck bone mineral density change after one year ‐ Alendronate, vitamin D and calcium.

1.12. Analysis.

Comparison 1 Alendronate compared to regular treatment, Outcome 12 Percent total hip bone mineral density change after one year ‐ Alendronate, vitamin D and calcium.

1.13. Analysis.

Comparison 1 Alendronate compared to regular treatment, Outcome 13 Percent trochanter bone mineral density change after one year ‐ Alendronate, vitamin D and calcium.

Comparison 2. Testosterone compared to no treatment.

Outcome or subgroup title	No. of studies	No. of participants	Statistical method	Effect size
1 Percent lumbar bone mineral density change after one year	1	43	Mean Difference (IV, Fixed, 95% CI)	3.70 [0.48, 6.92]
2 Proportion with adverse reaction after one year	1	43	Risk Ratio (M‐H, Fixed, 95% CI)	0.0 [0.0, 0.0]

2.1. Analysis.

Comparison 2 Testosterone compared to no treatment, Outcome 1 Percent lumbar bone mineral density change after one year.

2.2. Analysis.

Comparison 2 Testosterone compared to no treatment, Outcome 2 Proportion with adverse reaction after one year.

Comparison 3. Progressive resistance training compared to no treatment.

Outcome or subgroup title	No. of studies	No. of participants	Statistical method	Effect size
1 Percent lumbar bone mineral density change after one year	1	43	Mean Difference (IV, Fixed, 95% CI)	0.4 [‐2.81, 3.61]
2 Proportion with adverse reaction after one year	1	43	Risk Ratio (M‐H, Fixed, 95% CI)	0.0 [0.0, 0.0]

3.1. Analysis.

Comparison 3 Progressive resistance training compared to no treatment, Outcome 1 Percent lumbar bone mineral density change after one year.

3.2. Analysis.

Comparison 3 Progressive resistance training compared to no treatment, Outcome 2 Proportion with adverse reaction after one year.

Characteristics of studies

Characteristics of included studies [ordered by study ID]

Fairfield 2001.

Methods	Randomized, placebo‐controlled, 2x2 factorial design
Participants	54 HIV infected eugonadal men with the AIDS wasting syndrome. Between ages 27 and 51. No recent opportunistic infection, no contraindication to exercise, no intractable diarrhea, no substance abuse, no recent protease inhibitor initiation, no prostate disease, no bipolar disorder, no metabolic or hormonal disorders, and blood dyscrasias.
Interventions	testosterone enanthate 200mg/week intramuscularly versus placebo, and simutaneously to receive dynamic progressive resistance training three times per week or no training over 12 weeks.
Outcomes	Change in BMD in the lumbar spine
Notes
Risk of bias
Bias	Authors' judgement	Support for judgement
Allocation concealment?	Low risk	A ‐ Adequate

Guaraldi 2004.

Methods	Randomized, open‐label
Participants	41 HIV infected individuals with osteopenia or osteoporosis. On stable antiretroviral therapy. No secondary causes of osteoporosis, no cardiac or renal diseases, no cancer, no peptic ulcer or esophageal disease.
Interventions	70mg of alendronate weekly or no alendronate therapy. 1000mg calcium supplements and 500IU vitamin D supplements were provided to all study participants.
Outcomes	Change in BMD in the lumbar spine and femur, fractures, serum and urine bone markers.
Notes
Risk of bias
Bias	Authors' judgement	Support for judgement
Allocation concealment?	Unclear risk	B ‐ Unclear

Mondy 2005.

Methods	Randomized, open‐label
Participants	31 HIV infected individuals with osteopenia or osteoporosis. On stable antiretroviral medication, no secondary metabolic bone disease or contraindications to alendronate, no recent prolonged bedrest, no alcohol abuse, no opportunistic infection, no cancer, no peptic ulcer or esophageal disease, and no life expectancy of <1 year.
Interventions	70mg weekly of alendronate versus no study drug. Both groups received 1000mg daily of calcium supplements and 400IU daily of vitamin D supplements during the study.
Outcomes	Change in BMD in the lumbar spine, safety and tolerability of alendronate, serum and urine bone markers at 48 weeks.
Notes
Risk of bias
Bias	Authors' judgement	Support for judgement
Allocation concealment?	Low risk	A ‐ Adequate

Negredo 2005.

Methods	Randomized, open‐label
Participants	26 HIV infected individuals with osteoporosis. On stable antiretroviral therapy. No secondary cause of osteoporosis.
Interventions	70mg weekly of alendronate versus no study drug. Both groups received dietary counselling to ensure adquate daily calcium intake of 1200mg.
Outcomes	Change in BMD in the the lumbar spine and femur, serum bone markers, incidence of osteoporosis
Notes
Risk of bias
Bias	Authors' judgement	Support for judgement
Allocation concealment?	Unclear risk	B ‐ Unclear

Characteristics of ongoing studies [ordered by study ID]

Grinspoon 2006a.

Trial name or title	Androgen Effects in HIV‐Infected Women
Methods
Participants	HIV‐infected, female (age 18‐50) with androgen deficiency (free testosterone < 3 pg/mL) and AIDS wasting syndrome. Stable antiretroviral regimen, effective contraception, no recent history of hormone therapy, no pregnancy, stable metabolic function, no anemia, normal FSH, no breast cancer, no sleep apnea, and no cardiovascular disease.
Interventions	Patients will be randomly assigned to receive testosterone 300 micrograms twice a week or identical placebo
Outcomes	Lean body mass, strength, bone density, quality of life, neurocognitive function, safety over 18‐months.
Starting date	November 1, 2004
Contact information	Steven K Grinspoon, MD, Mass General Hospital, Boston, Massachusetts, 02114, United States
Notes	ClinicalTrials.gov Identifier: NCT00095212

Grinspoon 2006b.

Trial name or title	Physiologic Growth Hormone Effects in HIV Lipodstrophy
Methods
Participants	Male and female patients age 18‐60 with HIV on stable antiretroviral regimen, having a waist‐to‐hip ratio >0.90 for men and >0.85 for women, a simulated peak GH response to arginine/GHRH of less than 7.5 mcg/dL, increased abdominal girth, relative loss of fat in the extremities or relative loss of fat in the face. No recent use of Megace, anti‐diabetic agents, GH, other anabolic agents, glucocorticoid, no diabetes mellitus, no severe chronic illness, no anemia, no hypercreatininemia, no increased prostate specific antigen, no pregnancy, contraception use, no carpal tunnel syndrome and no active malignancy.
Interventions	Patients will be randomly assigned to receive low dose growth hormone therapy or placebo at study entry, with crossover design.
Outcomes	Bone density, insulin‐like Growth Factor I, visceral adiposity, lean body mass, total and regional body fat, glucose and insulin, and safety parameters
Starting date	January 4, 2005
Contact information	Steven K Grinspoon, MD, MGH, Boston, Massachusetts, 02114, United States
Notes	ClinicalTrials.gov Identifier: NCT00100698

Huang 2006.

Trial name or title	Bisphosphonate Therapy for HIV Infected Adults With Decreased Bone Mineral Density
Methods
Participants	HIV‐infected males and females, ages 18‐70, with t‐score ‐1.5 to ‐3.5 and no more than one fracture that is asymptomatic. HIV viral load of less than 5000 copies/ml and CD4 count of more than 100cells/mm3, no contraception, no previous treatment with bisphosphonates or fluoride, no recent use of supraphysiologic systemic estrogen/androgen therapy or corticosteroid therapy, on stable antiretroviral regimen, no significant liver or kidney disease, no anemia, no hypocalcaemia, no endocrinologic disorders, no cancer or chemotherapy or radiotherapy, no osteomyelitis, no recent dental procedures, no pregnancy or breastfeeding.
Interventions	Patients will be randomly assigned to receive zoledronate or placebo at study entry
Outcomes	Bone density at screening, study entry, week 2, and months 3, 6, 9, and 12.
Starting date	February 4, 2005
Contact information	Jeannie S. Huang, MD, MPH, University of California, San Diego, San Diego, California, 92103, United States
Notes	ClinicalTrials.gov Identifier: NCT00102908

McComsey 2006.

Trial name or title	The Effect of Alendronate, Calcium, and Vitamin D on Bone Mineral Density in HIV Infected Patients
Methods
Participants	HIV‐1 infection. Decreased BMD in the lumbar spine. CD4 cell count 100 cells/mm3 or more, and viral load of 5000 copies/ml or less, stable antiretroviral regimen, stable hormonal and metabolic status, absence of esophageal disorders, recent fracture, recent use of glucocorticosteroids, allergy to alendronate, drug or alcohol dependence, hepatitis C virus infection and recent use of anabolic steroids.
Interventions	Patients will be randomly assigned to receive either alendronate or placebo. All patients will receive calcium and vitamin D.
Outcomes	Bone density by DEXA scan at weeks 2, 12, 24, 36, and 48.
Starting date	May 22, 2003
Contact information	Grace McComsey, MD, Study Chair, Division of Infectious Diseases, Case Western Reserve University
Notes	ClinicalTrials.gov Identifier: NCT00061256

Rozenberg 2006.

Trial name or title	Efficacy of Alendronate Versus Placebo in the Treatment of HIV‐Associated Osteoporosis
Methods
Participants	Patients 18 years and above, not pregnant, premenopausalwomen, femur or fumbar BMD t‐score 2.5 or less, HIV infection for at least 5 years, CD4 cell count over 50/mm3, Karnofsky score over or equal to 70, no osteoporosis resulting from a cause other than HIV, BMI above 18, no severe lung failure, no chronic alcohol intoxication, no opportunistic infection, no gastric ulcer disease, no malignancy in the previous 5 years, no cytotoxic chemotherapy or cytokine therapy, no liver cirrhosis and no breast feeding
Interventions	Patients will be randomly assigned to receive alendronate or a placebo for 2 years. The recommended adequate intake of calcium and vitamin D will be given to all participants.
Outcomes	Lumbar t‐score on DEXA scan at months 0, 12 and 24 Femoral t‐score at months 0, 12 and 24 Evolution of bone metabolism markers Occurrence of fractures Tolerance of alendronate Prevalence of osteopenia and osteoporosis
Starting date	July 11, 2005
Contact information	Sylvie Rozenberg, MD, Principal Investigator, Hopital Pitie‐Salpetriere Paris service de Rhumatologie Dominique Costagliola, Study Chair, Inserm U720
Notes	ClinicalTrials.gov Identifier: NCT00120757

van der Ven 2006.

Trial name or title	The Prevalence of Vitamin D Deficiency and Effects of Vitamin D Supplementation in HIV‐1 Infected Patients
Methods
Participants	Male and female adult (>18 years) patients with HIV, with hypovitaminosis D. No hypercalcemia, no renal disorders, no liver disorders, no pregnancy, and no drug or alcohol abuse.
Interventions	Patients will be randomly assigned to receive supplementation of colecalciferol (2000 IU daily) or placebo.
Outcomes	Bone density, normalization of vitamin D levels, immune and adipocyte function at 12 weeks.
Starting date	March 22, 2006
Contact information	André JAM van der Ven, MD, PhD, Radboud University Nijmegen Medical Center, Nijmegen, P.O. BOX 9101, Netherlands
Notes	ClinicalTrials.gov Identifier: NCT00306410

Contributions of authors

Daren Lin searched for titles, obtained full articles, assessed the methological quality of studies, extracted data, analysed data and helped draft the review manuscript. Michael Rieder coordinated the review, searched for titles, assessed the methological quality of studies, extracted data, analysed data and helped draft the review manuscript.

Sources of support

Internal sources

• Dr. Rieder holds the GSK‐CIHR Research Chair at the University of Western Ontario, Canada.

• Mr. Lin received the Summer Research Training Program Fellowship at the University of Western Ontario, Canada.

• Robarts Research Institute, Canada.

External sources

• Canadian Institutes of Health Research, Canada.

• Children's Health Research Institute, Canada.

Declarations of interest

None.

Edited (no change to conclusions)