Osteopenia in HIV Infection
     Andrew Carr, MD
Associate Professor of Medicine 
     HIV, Immunology and Infectious Diseases Clinical Services Unit  
     St. Vincent's Hospital, Sydney, Australia    

Introduction from Jules Levin
It's been suggested from a number of studies that bone problems in HIV could be due to protease inhibitors, nucleosides, HIV therapy, the immune decline & dysfunction from HIV, immune restoration from HAART, ordinary risk factors and genetics. One French study reported at ICAAC reported osteopenia was related to duration of HIV infection and there was no relationship with ART experience, duration of PI use nor presence of lipodystrophy. A report on this study can be read in the last link at the end of the article below. As of yet, we don't have clear answers as to why but it appears as though bone problems in HIV are more prevalent than in people without HIV with similar characteristics. However, people with HIV may be more sedentary, which is a risk factor for osteopenia.

This article reviews:

At the end of the article are links to additional reports on osteopenia and avascular necrosis in HIV on the NATAP web site

What terms do I need to know?

           There are 5 bone terms that are likely to become standard HIV-speak, namely osteoporosis, osteopenia, t-score, z-score and bone mineral density. All 5 terms have universally accepted definitions established in part by the World Health Organisation, and which are applied in all types of bone studies. To understand these terms you will also need to understand some basic statistical terms.

           Bone mineral density (BMD) is the mass of bone in a given area of bone, and is commonly expressed as grams per square centimeter (g/cm2). BMD, like all things human, varies from person to person. The variability of most biological parameters about the population average or ýmeanţ value generally has a normal distribution (see figure). A normal distribution means that:

1.                  50% of values lie either side of the mean;

2.                  the bulk of values lie close to the mean; and

3.                  95% of values are taken as representing the normal or reference range in the studied population.


 A reference range spans the range from ˝2 to +2 standard deviation (SD) about the mean, a SD being a number representing how much all the values in a population vary about the mean. A normal distribution means that about 65% of values for a parameter will be within 1 SD of the mean (-1 SD to +1 SD), 15% between ˝2.5 SD and ˝1 SD, 15% between 1 SD and 2.5 SD, 2.5% less than ˝2.5 SD and 2.5% greater than +2.5 SD. For example, the mean CD4+ lymphocyte count in healthy adults is about 1000, most are between 800 and 1200, and about 95% of these values will fall within a reference range from about 500 to 1500.

           A t-score is a way of representing BMD in an SD format. A t-score of ˝2.7 means that an individualÝs BMD is 2.7 SD below the mean of healthy, same sex, young (20 to 40 year) adults. As you might imagine, the lower the t-score, the greater the risk of fracture over time. As BMD is higher in men than in women, a t-score of 1.7 in men will generally signify a higher BMD than a t-score of 1.7 in a woman.

           Osteoporosis is defined by a t-score less than ˝2.5. Osteopenia is defined by a t-score between ˝2.5 and ˝1. Why the distinction? This is merely a way of standardizing reporting of data in clinical studies of bone. For example, the percentage of patients with osteopenia might be reported at the beginning and end of a study evaluating a drugÝs potential benefit on BMD. Use of these 2 terms also allows for comparisons between drugs and patient populations.

           A z-score is a t-score corrected not only for sex, but also for age and race. Since bone density falls with increasing age (and is lower in women than in men), it follows that an individualÝs t-score and z-score will not be the same unless he/she is 30 years old. As one ages (which is not so bad given the alternative), the t-score will decline but the z-score may rise, fall or remain stable depending on what happens in most healthy adults of the same age and race. For example, a 65-year adult may have a t-score of ˝1.8 (low BMD, with a raised risk of fracture) but have a z-score of 0.5 (BMD similar to other adults of the same age, with no greater risk of fracture than adults of comparable age).

           Osteoporosis is not avascular necrosis (AVN; sometimes called aseptic necrosis) of bone. AVN is caused by poor circulation within a fragment of bone and results in gradual, painful death of bone. The cause of AVN is unknown and shares risk factors with osteoporosis. 

How do we get osteoporosis?

           Bone is created by cells called osteoblasts, and degraded by cells called osteoclasts. Bone is not a static organ; in any year, a good deal of bone turns over in healthy adults. Low BMD develops when the rate of new bone formation is less than the rate of old bone destruction. BMD declines as we age by about 0.5 to 1% per year because the rate of old bone loss is greater than the rate of new bone formation.

           The bone that turns over the most is in the spine, as this is trabecular bone, which has a much greater surface area exposed to the blood than does cortical bone (in long bones in limbs). Therefore, factors in the blood that cause low BMD are likely to result in a fall in spinal BMD before any fall in total BMD.

How is BMD measured?

           BMD is measured by dual-energy X-ray absorptiometry (DEXA). There are 2 types of DEXA scan. One measures BMD in the spine, hip and forearm (3 common sites of fracture in patients with osteoporosis), and the other form of DEXA measures fat and lean tissue (mostly muscle) in the whole body and in various body regions (the trunk, arms and legs). When DEXA is used to measure fat and muscle, the machine can automatically generate a BMD, t-score and z-score for the whole body if the machineÝs computer software is set appropriately, as well as any regions asked of the machine. It was the use of DEXA for the study of lipodystrophy in HIV-infected adults that lead to the observation that some patients had low total body BMD. Although total BMD on soft tissue or bone DEXAs are similar, they are not identical. There may be a substantial difference between a spine BMD value measured on bone DEXA or soft tissue DEXA. BMD should ideally be determined from bone DEXA, as the values are more accurate and provide better information about risk of fracture. One point worth stressing is that a spine BMD generated on soft-tissue DEXA is a BMD for the whole spine, whereas a spine BMD on bone DEXA refers to the lumbar (lower) spine.

           Sometimes a routine X-ray report may have a comment to the effect that there is low BMD. X-ray is a very insensitive tool for detecting osteopenia, so if an X-ray shows low BMD, there is a high chance that BMD is well below average. 

What happens when we have osteoporosis?

           The lower the BMD, the greater the risk of fracture, either spontaneously or with minimal trauma. The risk of fracture in patients with osteopenia is about 2-times average, with osteoporosis is about 4 to 5-times average, and in someone with osteoporosis and a previous fracture is about 20-times average.  

           The most common sites of osteoporotic fractures are anywhere in the spine, the forearm and hips. The latter 2 fractures are secondary to falls, whereas the first generally occurs spontaneously, and leads to progressive loss of height and curvature of the spine with age. 

Risk factors for low BMD; males vs females

           Three intrinsic factors determine BMD in healthy adults: age, sex hormones and weight. Low BMD is more common as we age, but is more likely in a 60-year old women than a 60-year old man, as female sex hormones (particularly estrogen) are critical for new bone formation, and estrogen levels fall drastically after menopause. Male sex hormone (testosterone) deficiency is frequently associated with low BMD. BMD is also very dependent upon lifelong body weight, the less one weighs, the lower the BMD.

           Environmental risk factors for low BMD include immobility, smoking, high alcohol intake, diseases such as kidney failure and thyroid overactivity, and longterm treatment with cortisone (but not anabolic steroids, which increase BMD if anything). Deficiency of the hormone insulin-like growth factor type 1 (IGF-1) is increasingly recognized as a cause of low BMD, but the mechanism of this is unknown. The main factor, however, may be our inheritance; studies of identical and non-identical twins have shown that multiple genes (mostly unknown) determine BMD. Osteoporosis has been linked very rarely to mitochondrial deletions in young HIV-uninfected males with no other clinical features of mitochondrial disease, and some had raised plasma lactate (a marker of mitochondrial disease) without symptoms (5,6).

           Overall, only about 50% of low BMD can be explained by the above known risk factors. This is similar to the situation with risk factors for heart attack, and similar to the situation with HIV disease when CD4 counts but not viral load could be measured.

Bone density in HIV disease

           Osteoporosis is a recently described adverse event in HIV-infected patients. Prior to the introduction of HAART, healthy HIV-infected men generally had normal bone mineral density that was stable over time (1); no comparable data are available for HIV-infected women. This suggests that known risk factors for osteopenia that are common in HIV-infected men such as immobility, smoking, low sex hormones and increased activity of the immune system may not play a major role in HIV-related osteopenia.

           Two studies found osteopenia in 42% and 38% of HIV-infected adult outpatients receiving combination antiretroviral therapy. In one study, osteopenia was linked to protease inhibitor therapy, although potential confounding factors such as nucleoside analogue type and duration, smoking, exercise, testosterone, lean body mass and weight prior to therapy were not studied (2). The second study found a high prevalence in lipodystrophic adults recruited to a randomised study of protease inhibitor cessation (3). No difference in total BMD was seen between the two randomized groups, however, after 48 weeks. Although both these studies reported no fracture, osteoporotic fractures have been reported in two HIV-infected African women (4).

Osteopenia in HIV-infected men: association with increasing age, lower weight pre-antiretroviral therapy, and lactic acidemia
     Andrew Carr1, John Miller2, David A. Cooper1,2
     St VincentÝs Hospital1, and National Centre in HIV Epidemiology and Clinical Research2, Sydney, Australia 

In: Abstracts of the 2nd International Workshop on Adverse Drug Reactions and Lipodystrophy in HIV, Toronto, Canada, 23-25 September, 2000.

Introduction:    Lactic acidemia is a well-known mitochondrial toxicity of nucleoside analog therapy and has been linked with osteoporosis. To address the possibility that HIV-associated osteopenia might have a mitochondrial pathogenesis, bone density (t-score, z-score, and total BMD) was assessed in 221 otherwise well, HIV-infected men (mean age 43 years) recruited to a lipodystrophy (LD) prevalence survey. The following parameters were evaluated for any association with BMD: age, HIV duration, presence of AIDS, smoking, current exercise level (graded as sedentary, mild moderate, high), types and durations of all antiretroviral therapies, weight (including weight prior to commencing antiretroviral therapy), symptoms and signs associated with lactic acidemia (fatigue, nausea, weight loss of at least 3 kg during the preceding 3 months), CD4 count, HIV RNA load, fasting metabolic parameters (lipid, glycemic, lactate, liver enzymes, testosterone), and LD (physical examination, DEXA and abdominal CT scan).

Results:     Osteoporosis was found in 3%, and osteopenia in a further 20%, although no patient had a documented fracture since being diagnosed with HIV infection. The only factors independently associated with osteopenia or osteoporosis were higher lactate (odds ratio [OR] 2.39 [95% CI 1.39 to 4.11] per 1 mmol/l increase; p=0.002) and lower weight prior to commencing antiretroviral therapy (OR 1.06 [95% CI 1.02 to 1.11] per 1 kg decrease; p=0.006). ). There was no independent association with any other parameter, including type or duration of any antiretroviral drug or drug class, or with lipodystrophy at any site. Low spinal (but not low total) BMD was associated with both lactic acidemia and duration of nucleoside analogue therapy. Conclusion:           Osteopenia in HIV-infected men appears more likely with lower body mass prior to antiretroviral therapy and the cumulative duration and magnitude of nucleoside analog-induced lactic acidemia. 

The above data come from the largest and most comprehensive study attempting to identify factors associated with low BMD in HIV-infected men. All the analyses suggest that duration and magnitude of lactic acidemia are linked somehow to low BMD, especially in the spine. But how? There are at least 2 possible explanations. One is a direct effect of nucleoside analogs on osteoblast mitochondria. The other possibility is that hydroxyapatite (the principle bone salt that provides the compressional strength vertebrate bone) is being leached from bone to buffer the acid load of chronic lactic acidemia.

            What donÝt we know from this study? The study cannot demonstrate a cause and effect relationship between increased lactate and low BMD because the study was cross-sectional not prospective. Also, the study has not evaluated women (in whom osteopenia is likely to be more common), children, or various racial groups. Only 13 patients were receiving abacavir, and so no conclusion about its impact can be made. Although abacavir is being promoted as a clean drug in this respect, it is worth noting that a presentation from the Swiss HIV Cohort Study at the 13th World AIDS Conference in Durban linked abacavir with lactic acidemia.


           We certainly donÝt know whether low BMD in our patients will result in increased fracture rates over time, in the same way we donÝt know the metabolic changes of lipodystrophy will lead to increased rates of heart attack and stroke. The predominant associations with low spinal BMD suggest that osteoporosis will lead to spinal fractures before fractures in long bones such as the hip or forearm, but time will tell. 

Should lactate or BMD be measured routinely?

           Lactate is not routinely measured as until now it has only been linked to lactic acidosis that is both very rare (about 1 in 1000) and unpredictable. Asymptomatic, low-level lactic acidemia occurs in 12 to 20% of nucleoside analogue recipients (7,16) and is now linked to osteopenia. Measurement of lactate in nucleoside analogues recipients without symptoms of lactic acidemia who have a fracture, osteoporosis or perhaps even other risk factors for osteoporosis should be considered. As low BMD remains largely an asymptomatic phenomenon, there appears little reason to measure BMD routinely. 

What should be done if I have reduced BMD?

           Investigation and treatment of low BMD should probably be no different than in the general population, except of course that patients with advanced HIV disease are of course far more likely to run into trouble with progressive AIDS than with fractures.

           Patients found to have low BMD should be referred to an endocrinologist. The lower the BMD, the younger the patient, the earlier the HIV disease, and the more risk factors for low BMD (above) and for fractures (eg. falls), the more likely treatment would be recommended. It may be prudent to address modifiable risk factors such as smoking, alcohol abuse, immobility and hypogonadism. Nucleoside analog therapy should not be altered until it is certain that:

1.                  nucleoside analogs cause osteopenia (from prospective and pathologic studies);

2.                  nucleoside analog cessation can lead to improved BMD; or

3.                  BMD can be increased by any therapy in the face of ongoing nucleoside analog therapy.

Well-conducted studies are clearly needed; the sooner we get underway, the sooner we will know what to do.

Comments from Jules Levin: Diet, exercise, and vitamin supplementation may help prevent osteopenia. Adequate intake of vitamin D and calcium through diet and vitamin supplementation may help preventosteopenia. As well, weight bearing exercise may also help. As mentioned by Carr in the article, life long low body weight and lean body mass may be a risk factor for reduced BMD. And the genetics of family background may predispose a person for BMD loss.

Additional Articles To Read:

Osteonecrosis (avascular necrosis) in HIV: A Case-Control Study
in full  PDF version
     -- What is Osteonecrosis, risk factors?
     -- Background on Lipids and Corticosteroids and Avascular Necrosis
Decreased Bone Mineral Density loss After HCV Therapy with Interferon+Ribivarin

Review of Studies on Bone Problems at ICAAC and Lipodystrophy Workshop
     -- Potential Causes or Mechanisms of Action Leading to Bone Problems  
     -- Protease Inhibitors May Inhibit In Vitro Conversion To Vitamin D
     -- Australian Study Reports on the Prevalence of Bone Mineral Density (BMD) Loss in 171 Australian patients and a Correlation of Change in Subcutaneous Fat With BMD Loss
     -- Osteopenia May Be Due to HIV Rather Than HIV Antiretroviral Drugs
     -- French Research Group Suggest Duration of HIV is a Factor in Developing Osteopenia,
not HIV meds nor lipodystrophy

References of interest 

1.                 Kanis JA, Delmas P, Burckhardt P, Cooper C, Torgerson D. Guidelines for diagnosis and management of osteoporosis. The European Foundation for Osteoporosis and Bone Disease. Osteoporosis Int 1994; 4: 325-331. 

2.                 Paton NJ, Macallan DC, Griffin GE, Pazianas M. Bone mineral density in patients with human immunodeficiency virus infection. Calcif Tissue Int 1997; 61: 30-2.

3.                 Tebas P, Powderly WG, Claxton S, et al. Accelerated bone mineral loss in HIV-infected patients receiving potent antiretroviral therapy. AIDS 2000; 14: F63-7.

4.                 Hoy J, Hudson J, Law M, Cooper DA. Osteopenia in a randomised, multicentre study of protease inhibitor substitution in patients with lipodystrophy syndrome and well-controlled HIV viraemia: extended follow-up to 48 weeks. In: Abstracts of the 2nd International Workshop on Adverse Drug Reactions and Lipodystrophy in HIV, Toronto, September, 2000 (abstract).

5.                 Stephens EA, Das R, Madge S, Barter J, Johnson MA. Symptomatic osteoporosis in two young HIV-positive African women. AIDS 1999; 13: 2605-6.

6.                 Varanasi SS, Francis RM, Berger CE, Papiha SS, Datta HK. Mitochondrial DNA deletion associated oxidative stress and severe male osteoporosis. Osteoporos Int 1999; 10: 143-9.

7.                 Papiha SS, Rathod H, Briceno I, Pooley J, Datta HK. Age-related somatic mitochondrial DNA deletions in bone. J Clin Pathol 1998; 51: 117-20.

8.                 Carr A, Miller J, Law M, Cooper DA. A syndrome of lipoatrophy, lactic acidaemia and liver dysfunction associated with HIV nucleoside analogue therapy: contribution to protease inhibitor-related lipodystrophy syndrome. AIDS 2000; 14: F25-32.

9.                 Boubaker K, Sudre P, Flepp M, et al. Hyperlactatemia and antiretroviral therapy in the Swiss HIV Cohort Study. In: Program and Abstracts of the 7th Conference on Retroviruses and Opportunistic Infections, San Francisco, January, 2000 (abstract 57).