Osteoporosis is a condition where the bones lose mass and density. It is commonly referred to as 'thinning of the bones' and occurs most commonly in older, post-menopausal women. Osteopenia refers to less severe bone mineral loss. Neither condition is an AIDS-defining condition. Until recently, osteopenia and osteoporosis were rarely seen among people with HIV.

Reduced bone mass can be diagnosed using a dual-energy X-ray absorptiometry (DEXA) scan, an X-ray based method which measures the density of different body compartments such as fat and bone. It is usually conducted on the bones of the spine.

Why does osteoporosis occur in people with HIV?

Osteoporosis is caused by a lack of bone calcium and protein, but the reasons for its appearance in relatively young, HIV-positive people, is still unexplained.

The high prevalence of osteopenia in HIV-positive people relative to HIV-negative people suggests that HIV itself contributes to thinning bones in this population (Knobel 2001; McGown 2001; Negredo 2001). However, it is not yet clear whether other factors contribute to thinning bones in HIV-positive people. There is growing evidence that protease inhibitors are not associated with osteoporosis but other antiretrovirals may contribute to this condition.

Common risk factors for osteoporosis in HIV-negative people include:

• Family history.

• Early menopause.

• Low testosterone levels in men.

• Prolonged corticosteroid or anticonvulsant use.

• Low calcium intake.

• Poor absorption of calcium.

• Smoking.

• Low body mass.

• Sedentary lifestyle.

Prevalence in HIV-positive adults

In 1999, the Royal Free Hospital in London reported two cases of sudden onset osteoporosis in young African women on anti-HIV treatment. Both women had lower back pain and their periods had been irregular. Bone scans showed a number of bones had collapsed and bone mineral analysis confirmed osteoporosis. One woman was taking d4T (stavudine, Zerit), 3TC (lamivudine, Epivir) and nevirapine (Viramune) and the other was taking ddI (didanosine, Videx / VidexEC), hydroxycarbamide (Hydrea), efavirenz (Sustiva), nelfinavir (Viracept) and saquinavir (Invirase / Fortovase). None of these medications had previously been associated with changes in bone density or hormonal levels (Stephens 1999).

Since then, several studies have established that osteopenia is relatively common, affecting about 30% of HIV-positive individuals in the investigations conducted so far. A study of Royal Free HIV patients found that 71% showed signs of thinning bones, with increased risk among those who have received antiretroviral therapy (Moore 2001). A recent study of HIV-positive women found that the prevalence of osteopenia and osteonecrosis was 2.5 times greater than in HIV-negative women, but that antiretroviral therapy did not alter the risk. Risk factors included low body mass index, a history of low body weight, low body fat percentage and infrequent menstruation, while the pattern of bone loss in menopausal women was different in HIV-positive and -negative women (Dolan 2004).

When bone mineral density reduces, the risk of fracture is increased. When the bone mineral density has fallen so low that an individual has a fracture risk four to five times higher than the average for the population, the condition is defined as osteoporosis. Osteoporosis has been reported in between 3 and 21% of clinic populations investigated (Tebas 2000; Carr 2001; Nolan 2001).

Osteonecrosis, or avascular necrosis, has also been reported in HIV-positive people. This condition refers to the death of the bone, notably in the hip joint. For more information, see the separate entry Osteonecrosis in Symptoms and illnesses: A to Z of illnesses.

Link to protease inhibitors and metabolic disorders?

Several studies have found that protease inhibitor (PI) treatment has been associated with a significantly greater incidence of osteoporosis.

For example, one study found that 21% of a group of 64 men receiving PIs, compared with 6% of an age-matched HIV-negative control group, had severe osteoporosis. There was no significant relationship between osteoporosis and fat redistribution (Tebas 2000). Fifty percent of the PI group had some evidence of reduced bone mass, compared to 29% of the control group. Although some researchers have speculated that reduced level of the male sex hormone testosterone might be linked to reduced bone mass, no link was found in this study.

An Australian group has also reported reduced bone mass in 28% of 80 patients with lipodystrophy. However, researchers do not know what is causing this high prevalence of osteoporosis given that the proportion with reduced bone mass did not increase during six months of follow-up, and switching from a PI-containing regimen to a PI-sparing regimen did not improve bone mass (Hoy 2000). The loss of bone mass was most pronounced in the legs and spine, increasing the possibility of broken bones among PI recipients.

One test tube study has shown that PIs interfere with vitamin D metabolism through their influence on the cytochrome P450 system (Dusso 2000). While vitamin D is crucial for bone formation, there is no evidence that people taking PIs have reduced levels of vitamin D. On the other hand, another test tube study has shown that some PIs may contribute to the manufacture of new bone cells which contradicts the theory that PIs are weakening bones (Nolan 2001).

An analysis of more than 10,000 patients enrolled in 13 randomised studies of PIs found no difference between patients receiving PIs and patients receiving other regimens in the rate of fractures. Approximately 2% in each group reported fractures, and very few of these were compression fractures, the type associated with osteoporosis (Struble 2001).

Other possible risk factors

Despite these findings, evidence is starting to suggest that the duration of HIV infection is the crucial factor in osteopenia (Knobel 2001; Lawal 2001; McGowan 2001; Negredo 2001). For example, a French study of 85 consecutive patients attending an HIV clinic found that 20% of treatment-naive patients had osteopenia, and 45% of PI-naive patients had osteopenia. Only the duration of HIV infection was significantly associated with osteopenia (Allavena 2000).

Similarly, a study conducted by the Chelsea and Westminster Hospital, London, found no significant difference in bone mineral density between 52 untreated HIV-positive patients, 22 men on PI therapy and ten men receiving NRTIs only. Indeed, antiretroviral therapy of any sort seemed to reduce the severity of bone mineral loss when patients were matched for duration since HIV diagnosis, and there was a trend towards reduced risk in those with lower viral load regardless of risk. These findings suggest that bone mineral loss may be a consequence of long-term HIV infection and immune activation (Moyle 2000).

A number of factors, other than PI treatment, may be associated with reduced bone mineral density:

• More rapid loss of subcutaneous fat correlated with a greater reduction in bone mineral density in 171 patients, regardless of PI therapy. However, indinavir (Crixivan) therapy was associated with increased bone mineral density compared to nelfinavir (Nolan 2001). In addition, research to date has not shown a consistent association between lipodystrophy and bone thinning (Huang 2001; Tebas 2001).

• Higher lactate levels in the blood. Lactic acidemia has been associated with current ddI or d4T treatment and the magnitude of CD4 cell count increases since starting treatment (Carr 2001). The authors suggested that bone-derived calcium may be used to buffer high levels of acid in the blood. However, other research has failed to confirm this association (Claxton 2001).

• Lower weight prior to starting antiretroviral therapy.

Treatment of osteoporosis

Increasing calcium intake may have a modest effect on bone loss. One study in post-menopausal women showed that women with a higher calcium intake lost 1 to 2% less bone mass compared to a control group over two years of follow-up.

An adequate intake of calcium is at least 1500mg per day, which can be obtained from eating two or three servings of dairy products.

Exercise is thought to improve bone density and bone strength and is generally recommended in the elderly at increased risk of fractures due to osteoporosis.

Vitamin D supplementation has been shown to reduce the rate of hip fractures in the elderly, but is only likely to be necessary or appropriate in people who live in parts of the world where exposure to the sun is limited for more than half the year, such as northern Europe, the northern United States and Canada. Vitamin D is synthesised by the body as a result of exposure to ultraviolet light, but supplementation at high levels could lead to toxicity.

In a recent presentation, one of the leading researchers on osteoporosis in HIV suggested that the only interventions currently supported by evidence are to ensure an adequate calcium intake (approximately 1500mg per day), and to ensure a vitamin D intake of between 400 and 1000IU per day.

Anabolic steroids have been shown to increase spinal bone density by 2 to 3% after treatment but no effect on the rate of fractures has been seen. However, a study of the steroid oxandrolone had no effect on bone mineral density among a small group of HIV-infected individuals (Lawal 2001b). Growth hormone also had no effect on bone mineral content in HIV-infected men with lipodystrophy (Lawal 2001c).

In severe osteoporosis, several drugs have been tested to reduce the risk of fractures. A vitamin D analogue called calcitriol (Rocaltrol / Calcijex) has been shown to reduce fracture rates in post-menopausal women. Disodium etidronate (Didronel), calcitonin (Forcaltonin / Miacalcic) and alendronic acid (Fosamax) have all been approved for the treatment of osteoporosis in post-menopausal women.

However, few studies have investigated whether these treatments are effective in HIV-positive people diagnosed with osteoporosis. One study comparing vitamin D and calcium supplementation with or without alendronic acid in 31 HIV-infected people with osteopenia found that spinal bone mineral density improved by 5% in the alendronic acid group and 1% in the supplement-only group at 48 weeks (Mondy 2003). Another study using a similar protocol in 40 HIV-infected people with osteopenia, found that alendronic acid was significantly more effective in reducing bone turnover, although two people in each group experienced fractures due to minimal trauma (Guaraldi 2004). A small, non-randomised study of vitamin D and calcium supplements in HIV-infected children with osteoporosis also found little benefit from this type of supplementation (McComsey 2003).

Switching from a PI-based to a PI-sparing regimen has not been shown to improve bone mineral density after 48 weeks (Hoy 2000; Claxton 2001), suggesting either that improvement of the problem could take longer, or that PIs are not the cause of the problem. The condition did not worsen during the follow-up period, regardless of whether people stayed on PI therapy or switched to an non nucleoside reverse transcriptase inhibitor (NNRTI)-containing regimen.

Research into osteoporosis: HIV or HAART?

Anastos (2004) measured bone mineral density in 88 HIV-negative and 184 HIV-positive women (90 not on HAART, and 94 on HAART) enrolled in the Womens Interagency HIV Study. Bone mineral density was 6 to 8% lower in the HIV-positive than in the HIV-negative women (p <0.03) with similar bone mineral density in the HIV+ women taking or not taking HAART. The prevalence of osteopenia/osteoporosis was 6.4% in the HIV-negative women, 18.9% in the HIV-postive women not on HAART and 20.4% in women receiving HAART (adjusted OR 3.15, p = 0.027 in all HIV-positive vs. HIV-negative women). Longer nevirapine use was significantly associated with higher bone mineral density, and longer abacavir use with lower bone mineral density. White race, lower body mass index and self-reported postmenopausal status were also independently associated with lower bone mineral density.

Dolan (2004) measured bone density using DEXA scans of the lumbar spine, femur and hip in 84 HIV-positive and 63 HIV-negative women, with an average age of 41. In both groups, around one third were caucasian and 14% were Hispanic, but more of the HIV-positive patients were African American (36 vs. 27%). Mean body mass index (BMI) was 26-27kg/m². The HIV-positive women had less body fat and a history of lower body weight. 93% of the HIV-positive women had used antiretroviral therapy. During the study, 80% were on NRTIs, 42% on PIs and 27% on NNRTIs. HIV-positive women had lower bone density in the lumbar spine (1.02 vs. 1.07g/cm², p = 0.03), femoral neck (0.82 vs. 0.87g/cm², p = 0.01) and hip (0.93 vs. 0.99g/cm², p = 0.004), and a higher rate of osteopenia (54 vs. 30%) and osteonecrosis (10 vs. 5%). After adjustment for BMI, menstrual status, age and race, HIV-positive women were 2.5 times more likely to have low bone mineral density. Among HIV-positive women, risk factors were: current BMI, history of low body weight, lower boy fat percentage, oligomenorrhoea (<3 periods in the previous 3 months) and menopause, although the pattern of bpne loss was different in HIV-positive and -negative women. There was no effect of past or current antiretroviral therapy.

Negredo (2001) conducted a cross-sectional comparative study of 97 HIV-infected patients. 49 had received long-term antiretroviral therapy, 24 were treatment-naive and 24 had taken treatment for less than 6 months. DEXA scans looked at the spine, neck and femur. 83.6% of the heavily pretreated group had thinning bones compared to 58.3% of the naive/lightly treated group. A third of the treatment naive people had osteopenia. Authors concluded that PIs may play a role but that other factors (duration of HIV infection, nutritional status) may also play a role.

Lawal (2001) compared bone mineral content and density of two groups of HIV-infected people involved in clinical trials in 1993 (pre-HAART) and 1998 who had DEXA scans, and matched controls. The first group were 36 malnourished men and the second were 19 men and 3 women with body fat changes. Whole body scans were conducted which showed no difference between the two HIV-infected groups. However, both HIV-infected groups had significantly lower total bone mineral content, total bone calcium and total bone density than their matched controls.

McGowan (2001) conducted DEXA scans on the first 300 participants enrolled in GS-99-903 - an international comparative study of tenofovir or d4T plus efavirenz and 3TC in people who have never taken anti-HIV drugs. Results on the first 151 showed 24% had osteopenia and 1% had osteoporosis, suggesting that chronic HIV-infection is associated with high prevalence of reduced bone mineral density.

Knobel (2001) compared 80 HIV-infected people (26 on no treatment, 37 on HAART including a PI and 17 on HAART without a PI) with HIV-negative matched controls. DEXA showed neck and spine density was lower among HIV-infected people. Results were not affected by treatment, viral load, CD4 count, or duration of therapy. Weight was correlated with lower bone density. Osteopenia rates were: 25% treatment naive; 40% PI-treated; 33% non-PI HAART, and 16% uninfected. Only total HIV positive versus negative rates of osteopenia reached statistical significance (34.5% vs 16%, P=0.03).

Moore (2001) reported interim data from a study of prevalence of bone density among HIV-infected patients attending the Royal Free Hospital. Of 72 HIV-infected individuals, 50-66% showed signs of reduced bone density. Average age was 39 years, 35% were current smokers, average time since HIV diagnosis was 4.9 years and 66% had ever taken antiretroviral therapy. Adjusted odds ratio indicated that a history of antiretroviral therapy was associated with reduced bone density (OR3.08, p=0.05).

Tebas (2000) enrolled 122 individuals (112 men and 10 women) receiving PI-based HAART (n=64) and a control group from the general population (n=22), and HIV-infected individuals not exposed to protease inhibitor treatment (n=36), but excluded women from the analysis because the number recruited was too small for comparative purposes. The PI-treated group had a median exposure of 104 weeks, and had a higher mean age than the two other groups (41 years vs 33-37) (p=0.001). DEXA scanning of the lumbar spine (L1-L4), the proximal femur and the whole body was conducted to determine bone mineral density. Bone mineral density was scored using two measures: t (measured bone mineral density - population mean BMD at age 30 / standard deviation (BMD at age 30), and z (measured BMD - population mean BMD for matched age subject/standard deviation (BMD at same age). World Health Organisation definitions of osteopenia and osteoporosis were used, where t>-1 = normal and t< -2.5 = osteoporosis. A t-score < -2 is associated with a four to five fold greater risk of fracture. A z-score of <-2 was also classified as osteoporosis. PI recipients had lower lumbar spine and proximal femur BMD as measured by t and z scores, and 50% of this group were classified as osteoporotic (p=0.02), compared with 6% of the HIV-negative control group and 11% of the HIV-positive no-PI group. PI recipients were more than twice as likely (RR=2.19) to be osteoporotic when compared with HIV-positive individuals not receiving PI treatment.

Stuble (2001) conducted a meta-analysis of more than 10,000 patients enrolled in 13 randomised studies and found no difference between patients receiving protease inhibitors and patients receiving other regimens in the rate of fractures. Approximately 2% in each group reported fractures.

Possible association with fat and metabolic disorders

Bone thinning in children

Ramos (2004) studied 35 vertically HIV-infected children with a median age of 129 months over a median interval of 13 months. Thirty children were taking HAART (28 PI-based, 2 PI-sparing regimen with NNRTI), two taking dual NRTIs and three no therapy. Median time on HAART at first DEXA scan was 62 months, and 73% of HAART-treated children had an undetectable viral load. At first DEXA scan, 40% were osteopenic, and there were no significant differences in the proportion of children with osteopenia (45%) at the second DEXA scan. There was a trend to a higher decrease of bone mineral density in children treated with PI-containing regimen (p = 0.06). Markers of bone resorption (urine deoxypyridinoline, NTx, calcium/creatinin) were significantly greater in osteopenic children.

McComsey (2003) studied bone mineral density in 22 children on HAART and 1 child who had never taken antiretrovirals. 48% (11/23) were osteoporotic and 74% (17/23) were osteopenic. Calcium and vitamin D supplements were given for a median duration of 9 months. There was a non-significant trend to improved bone mineral density with z scores improving from -2.7 pre-supplementation to -2.1 after supplementation (p = 0.13).

Arpadi (2002) studied 51 HIV-infected children by DEXA scan and found that bone mineral density increased with age, even after controlling for body size, race/ethnicity, gender and bone area. However, bone mineral in HIV-positive children is lower than in HIV-negative controls, and the disparity increases with age. PI treatment was not associated with decreased mineral density.

Mora (2001) compared bone formation and thinning in 40 HIV-infected children and controls. Total and spine bone density was lower in the 35 HIV-infected children on HAART compared to untreated children and healthy controls. Children on HAART with lipodystrophy had lower bone density than untreated children, while children on HAART without lipodystrophy had bone density between the other two groups.

References

Allavena C et al. Osteopenia and osteoporosis in HIV-infected patients: role of antiretroviral therapy (ART)? 40th Interscience Conference on Antimicrobial Agents and Chemotherapy, Toronto, abstract 1304, 2000.

Anastos K et al. The association of bone mineral density with HIV infection and antiretroviral treatment in women. Eleventh Conference on Retroviruses and Opportunistic Infections, San Francisco, abstract 744, 2004.

Arpadi SM et al. Bone mineral content is lower in prepubertal HIV-infected children. Journal of Acquired Immune Deficiency Syndromes 29(5): 450-454, 2002.

Carr A et al. Osteopenia in HIV-infected men: association with asymptomatic lactic acidemia and lower weight pre-antiretroviral therapy. AIDS 15(6): 703-709, 2001.

Chang K et al. Does race influence bone mineral density in HIV-infected individuals receiving HAART? Eighth Conference on Retroviruses and Opportunistic Infections, Chicago, abstract 630, 2001.

Claxton S et al. Circulating leptin and lactate levels are not associated with osteopenia in HIV-infected men. Eighth Conference on Retroviruses and Opportunistic Infections, Chicago, abstract 634, 2001.

Dolan S et al. Reduced bone density in HIV-infected women. AIDS 18: 475-483, 2004.

Dusso A et al. Protease inhibitors inhibit in vitro conversion of 25 (OH)- vitamin D to 1,25 (OH)2-vitamin D. Antiviral Therapy 5 (Supp 5): 19, 2000.

Galindo MJ et al. Factors related to the osteopenia of HIV-infected patients. 40th Interscience Conference on Antimicrobial Agents and Chemotherapy, Toronto, abstract 1303, 2000.

Guaraldi G et al. Alendronate Reduces Bone Turnover in HIV-associated Osteopenia and Osteoporosis. Eleventh Conference on Retroviruses and Opportunistic Infections, San Francisco, abstract 742, 2004.

Hoy J et al. Osteopenia in a randomised multicenter study of protease inhibitor substitution in patients with the lipodystrophy syndrome and well-controlled HIV viremia. Seventh Retroviruses Conference, San Francisco, abstract 208, 2000.

Huang JS et al. Increased abdominal visceral fat is associated with reduced bone density in HIV-infected men with lipodystrophy. AIDS 15(8): 975-982, 2001.

Knobel H et al. Osteopenia in HIV-infected patients: is it the disease or is it the treatment? AIDS 15(6): 807-808, 2001.

Lawal A et al. Equivalent osteopenia in HIV-infected individuals studied before and during the era of highly active antiretroviral therapy. AIDS 15(2): 278-280, 2001.

Lawal A et al. Effect of oxandrolone upon bone mineral density content in malnourished HIV+ patients. Eighth Conference on Retroviruses and Opportunistic Infections, Chicago, abstract 636, 2001b.

Lawal A et al. Effect of growth hormone on osteopenia HIV+ patients. Eighth Conference on Retroviruses and Opportunistic Infections, Chicago, abstract 635, 2001c.

McComsey GA et al. The effect of calcium and vitamin D on bone mineral density in HIV-infected children with osteoporosis. Tenth Conference on Retroviruses and Opportunistic Infections, abstract 779, 2003.

McGowan I et al. Assessment of bone mineral density (BMD) in HIV-infected antiretroviral-therapy-naive patients. Eighth Conference on Retroviruses and Opportunistic Infections, Chicago, abstract 628, 2001.

Mondy K et al. Alendronate, vitamin D, and calcium for the treatment of osteopenia/osteoporosis associated with HIV infection. Tenth Conference on Retroviruses and Opportunistic Infections, abstract 134, 2003.

Moore AL et al. Reduced bone mineral density in HIV-positive individuals. AIDS 15(13): 1731-1733, 2001.

Mora S et al. Bone mineral loss through increased bone turnover in HIV-infected children treated with highly active antiretroviral therapy. AIDS 15(14): 1823-1829, 2001.

Moyle G et al. Osteopenia: a consequence of HIV not HAART? Antiviral Therapy 5 (Supp 5): 43, 2000.

Negredo E et al. Bone mineral density in HIV-infected patients. Eighth Conference on Retroviruses and Opportunistic Infections, Chicago, abstract 626, 2001.

Nolan D et al. Stable or increasing bone mineral density in HIV-infected patients treated with nelfinavir or indinavir. AIDS 15(10): 1275-1280, 2001.

Ramos JT et al. Stability of osteopenia in HIV-infected children over time. Eleventh Conference on Retroviruses and Opportunistic Infections, San Francisco, abstract 745, 2004.

Stephens E et al. Symptomatic osteoporosis in two young HIV-positive African women. AIDS 13(18): 2605-2606, 1999.

Struble K et al. Bone fracture rates in HIV+ patients receiving PI vs non-PI containing regimens. 41st Interscience Conference on Antimicrobial and Antiviral Chemotherapy, Chicago, abstract I-1329, 2001.

Tebas P et al. Lack of association between visceral adiposity and osteopenia in HIV-infected individuals. Eighth Conference on Retroviruses and Opportunistic Infections, Chicago, abstract 633, 2001.

Tebas P et al. Accelerated bone mineral loss in HIV-infected patients receiving potent antiretroviral therapy. AIDS 14:4, F63-F67, 2000.