The advances in the treatment of HIV-infected individuals, particularly with the use of highly active anti-retroviral therapy (HAART), have significantly decreased the morbility and mortality caused by this infection. The long-term survival of the treated patients has revealed several metabolic complications, such as lipodystrophy, insulin resistance, diabetes, dyslipidemia and, more recently, alterations in phosphocalcic metabolism.
Paton et al. showed, for the first time, a decrease in bone mineral density (BMD) ; since then a growing number of authors have reported different alterations in the bone mineral metabolism in HIV patients. Today, there is reasonable agreement about the frequent association between osteopenia and HIV infection.
A variable pattern for the biochemical parameters controlling bone formation and resorption have been described: a study by Teichmann et al.  showed in HIV-seropositive patients a decrease in the formation markers with an increase in the bone resorption markers, regardless of variables such as the type and duration of treatment therapy. Tebas et al.  reported an increase in the bone formation and resorption markers in patients on protease inhibitor (PI) therapies, suggesting an increased bone turnover. Non-PI treated patients, however, were not included in this study. The role of PIs in the metabolism of phosphocalcic markers is also controversial, several authors have found a significant increase in bone formation markers in patients treated with PI, suggesting a beneficial effect of this therapy .
In a recently published study, Nolan et al.  concluded that there is an increase in BMD in patients taking indinavir or nelfinavir, suggesting that the osteopenia observed in treatments that use different PI may be due to their own metabolism. In a study where neither therapeutic regimens nor administration times were considered, Huang et al.  found a relationship, in HIV-infected men, between the decrease of BMD and the increase in abdominal visceral fat (lipodystrophy).
Carr et al.  did not find a relation between decreased BMD and use of PIs or therapy duration, linking osteopenia to both lactic acidosis and low patient weight at the beginning of therapy.
The etiology and pathogenesis of osteopenia in HIV infection has not been fully elucidated. Multiple factors have been involved, including the direct effect of the virus upon osteogenic cells [8–10]; persistent activation of pro-inflammatory cytokines, especially the tumor necrosis factor-alpha (TNF-α) [11–13]; alterations in the metabolism of vitamin D and its derivatives [13,14]; opportunistic and/or chronic diseases associated with HIV ; and mitochondrial abnormalities related to lactic acidemia .
In a study carried out at the Massachusetts General Hospital, a relation was established between the decreased BMD at the lumbar spine and the increased abdominal fat in men with lipodystrophy . This finding was not confirmed by Cooper's team, who compared HIV-seropositive patient groups with and without lipodystrophy, and found similar BMD values in both groups .
Although the scope of these findings is still uncertain, there are few reports of pathologic fractures associated with osteopenia in HIV-infected patients [16,17]. The role of HAART, especially regimens containing PI, on bone metabolism and on BMD in HIV-seropositive patients is controversial. In one of the first studies, Tebas et al. , showed a relationship between the use of PIs and a decrease in BMD, and other authors have found PIs linked to the diagnose of osteopenia [19–21]. However, these findings still need to be confirmed by other groups. When studying a patient with lipodystrophy, Hoy et al.  found that BMD remained unaltered after replacing PI-containing regimens with a non-PI therapy during 48 weeks. Moyle et al.  studied HIV-seropositive patients on PI and non-PI therapies, and found similar BMD values in both groups, which suggests a direct effect of the HIV infection on BMD. More recent studies have found a frequent occurrence of osteopenia among HIV-infected patients, but no significant differences among groups treated with different therapeutic regimens; thus, the alterations found could not be attributed to the regimens themselves [7,24–26]. Moreover, in a longitudinal study, Nolan et al.  did not find significant BMD changes in patients on HAART, and pointed out that indinavir therapy can be linked to an increase of BMD over time.
The objective of this work was to identify and describe possible bone metabolic alterations in HIV-infected patients, and to evaluate the effects of different therapeutic regimens on such changes.
Materials and methods
A total of 142 outpatients (113 male, 29 female) aged 20–45 years, were divided into four groups: group A, (n = 33) HIV-infected seropositive patients, antiretroviral-naïve; group B1, (n = 36) HIV-infected patients, on combined antiretroviral therapy for over 1 year without PIs; group B2, (n = 42) HIV-infected patients, on combined antiretroviral therapy for over 1 year with PIs, and a control group; group C, (n = 15) healthy, non-HIV-infected individuals.
All patients were asymptomatic and have had a confirmed positive serology at least for 2 years.
Subjects excluded from the study include those with a known history of bone or chronic metabolic diseases; long bed rest periods (> 6 months); renal or liver failure; patients on treatment with drugs affecting the bone metabolism such as corticosteroids, levothyroxine, lithium, etc.; individuals with endocrinopathies (hypogonadism, hyper- and hypothyroidism, hyper- and hypocortisolism); post-menopausal women with active opportunistic infections; patients with evidence of neoplasia, chronic diarrhea, prior weight changes (greater to 10% body weight) or absorption dysfunction.
Patients were selected according to inclusion–exclusion criteria, by exhaustive interviewing and complete physical examinations. The interview questionnaire was designed to determine prior pathological, toxic, epidemiological histories, as well as social situation and family history, medications intake, physical activity and eating habits. In order to assess the presence of related endocrinopathies, thyroid stimulating hormone (TSH), free urinary cortisol, testosterone (in men), and estradiol (in women) were determined. To rule out other pathologies blood and urine chemistry battery tests, including blood glucose, urea, creatinine, cholesterol, triglycerides, high-density lipoprotein (HDL) cholesterol, uric acid, glutamic-pyruvic transaminase (GPT) and glutamic oxaloacetic transaminase (GOT) were measured, BMD associated to the body composition was determined in all patients by dual energy X-ray absorptiometry in total body, lumbar spine (L1–L4) and proximal femur (HOLOGIC QDR 4500w; Hologic Inc., Bedford, Massachusetts, USA). Additional tests included the determination of osteocalcin, as a bone formation marker; d-pyridinoline, as bone resorption marker; parathyroid hormone (THP) (Chemiluminescence – Immulite; Diagnostic Products Corporation, Los Angeles, California, USA); calcemia, phosphatemia and calciuria (conventional chemical method). HIV patients were characterized by determining the plasma HIV RNA, the CD4 cell count, the disease progression, and antiretroviral used and duration of treatment. All tests were performed at the same reference center, with internal and external quality controls using the same kits in order to avoid methodological discrepancies.
Subjects were categorized as ‘osteopenia’ and ‘osteoporosis’ according to the BMD results using the t-score (results were compared to those of young subjects, general population) and following the criteria recommended by the World Health Organization . To determine osteoporosis odd ratio by the logistic regression analysis (see Table 3 results), patients with a z-score lower than −2 in L1–L4 were considered. In order to estimate the fat distribution (body composition by densitometry), the appendiceal/central fat index was determined using the following formula: trunk fat mass divided by total sum of the four limbs fat mass. To evaluate the variability of the method, a second densitometry was blindly performed (operator did not know what kind of individuals were tested) on 15 subjects chosen randomly (10 patients and five controls); the results obtained were less than 2% in all cases. The statistical data analysis was performed using the Kruskal–Wallis, the Wilcoxon and Spearman's correlation coefficient tests. P values lower than 0.05 indicated statistically significant differences among groups.
The characterization of the HIV infected patients, per group, and control individuals included in this study are described in Table 1.
Body mass index (BMI) and weight were slightly lower in HIV-infected patients than in controls. The antiretroviral-naive patient group (group A) had the shorter infection time and slighter elevated plasma HIV RNA than patients on treatment (groups B1 and B2).
Table 2 shows the BMD results obtained at the different bone sites for HIV-positive patients and control subjects. The BMD, expressed both in absolute values and standard deviation score (t and z score), was statistically lower, in every case, in HIV-seropositive patients in comparison with the healthy controls. Different treatment therapies revealed similar BMD results. The percent of calcium (calciuria) present in HIV-seropositive patients was lower than in controls; again no differences were observed among groups A, B1 and B2. Comparable results were obtained for the rest of the measurements.
Figure 1 shows the percentage of subjects with osteopenia and osteoporosis – both in patients on different treatment regimens and in controls – in the femoral neck and lumbar spine respectively (t-score). There was a significantly higher percentage of osteopenia and osteoporosis in HIV-patients than in controls (P < 0.0001), with no differences among patient groups. When categorizing subjects using the z-score (adjusted for same sex and age subjects) (Fig. 2), the percentage of patients with osteopenia and osteoporosis remained significantly higher in HIV-seropositive patients than in controls (P < 0.002).
There was significant correlation between years of infection and BMD in femoral neck (r = −0.26; P = 0.01), lumbar spine (r = −0.25; P = 0.02) and total body (r = −0.26, P = 0.01). The rest of the variables did not correlate with BMD.
Table 3 shows the logistical regression analysis for osteoporosis as a dependent variable, and the plasma HIV RNA, years of infection, CD4 cell counts and age, as independent variables (regressor). The odds ratio was significantly higher for the years of infection, without differences against the other variables.
At the present time, osteopenia is frequently linked to HIV infection. Its etiology is not fully understood. Similarly to the phenomenon of fat redistribution (lipodystrophy), osteopenia was initially attributed to the use of PIs; however, additional research efforts did not prove this relationship. These days, the role of the many antiretroviral regimens used in the treatment of HIV-infected individuals on the metabolism of phospho-calcium is controversial, to such an extent that some authors believe these regimens are responsible for the decreased BMD observed, whereas others suggest the opposite.
Tebas et al.  studied a population of 112 HIV-infected men on HAART, divided into two groups (on PI and non-PI therapy), and compared both with a third group of non-HIV-infected individuals, concluding that those on PI had a higher incidence of osteopenia and osteoporosis. They did not find any relation between ostopenia and osteoporosis and fat redistribution, suggesting that they are independent adverse effects. However, Tebas et al. did not compare them with a group that were not receiving antiretroviral therapy; in addition the densitometries were not performed at the same time, some were done at the beginning of therapy and other during treatment; furthermore the lumbar spine BMD data was extrapolated from total body densitometries and the minimum exposure time to PIs was 16 weeks, which might not have been enough to evaluate the impact of PI on BMD. All these issues and especially the length of the PI treatment could explain the differences found in our study. In our research, one of the patient inclusion criterion was a minimum of 24 months on PI therapy, which would enable us to evaluate the effects of the therapy.
The fat distribution analysis, demonstrated that the relationship between appendiceal to/central fat mass index was higher in patients on treatment, regardless of treatment type. The naive group's response was similar to that of groups on treatment.
In April, 2001, Knobel et al.  found osteopenia and osteoporosis both in patients on HAART treatment and in therapy-naive patients. The HIV-patients group showed significant differences from the non-HIV, healthy control group with respect to BMD, with a similar percentage of osteopenia and osteoporosis in HIV-infected individuals, both on therapy and therapy-naive. These findings match the results of our research. However, Knobel's team did not find a positive correlation between the years of infection and the BMD. A minimum infection time was not considered, though, which may mask the direct effect of the HIV infection on phosphocalcic metabolism.
These differences regarding the role of therapeutic agents in BMD could be attributed to the use of different regimens, without considering – in some cases – minimum disease evolution time and minimum therapy duration.
In this study, we did an exhaustive selection of patients to avoid patients with other risks that would have masked the results of the phospho-calcium metabolism. All HIV-infected patients had a documented time of infection, 2 years or more, and a well-controlled therapy duration, 1 year as a minimum. We also included no treated HIV-infected (treatment naive) and healthy, HIV seronegative individuals. The lack of statistically significant differences in BMD between patient groups may suggest that the decrease may not be due to the effect of the drugs used. However, the positive correlation between years of infection and BMD in all bone sites under study suggests a direct or indirect effect of the virus. The activation of pro-inflammatory factors, indirect effect of HIV, with known effects on phosphocalcic metabolism would be one of the proposed mechanisms., In this study these cytokines and/or chemokines were not studied; therefore, we can not make any inference about the etiology of the alterations found in bone metabolism.
In summary, we have found that the BMD was significantly lower in HIV-seropositive patients when compared with healthy seronegative individuals, with no significant differences among patient groups on different therapeutic regimens. Therefore, the alterations found could not be attributed to the therapeutic therapies themselves. Bone formation and resorption markers were similar among all the groups studied. The positive correlation between years of infection and BMD in all bone sites studied, suggests a deleterious effect of the virus on bone metabolism, independent of antiretroviral therapy.
The authors gratefully acknowledge the help provided by Luis Kremer Ph.D., Miguel Gravotta Ph.D. and Laura Nieto Ph.D. for their collaboration and Marta Leon Monzon Ph.D. and Adriana Bistoni Ph.D. for advice given during the preparation of this manuscript.
Sponsorship: Supported in part by grants from MSD (MSG/CRX-ARG-LUN09-99).
1. Paton NI, Macallan DC, Griffin GE, Pazianas M. Bone mineral density in patients with human immunodeficiency virus infection. Calcif Tissue Int
2. Teichmann J, Stephan E, Discher T, Lange U, Federlin K, Strake H, et al
. Changes in calcitropic hormones and biochemical markers of bone metabolism in patients with human immunodeficiency virus infection. Metabolism
3. Tebas P, Yarasheski M, Claxton S. Serum and urine markers of bone mineral metabolism in HIV-infected patients taking protease inhibitor-containing potent antiretroviral therapy. Antiviral Ther
(Suppl 5):18 [abstract 029].
4. Aukrust P, Haug CJ, Ueland T, Lien E, Muller F, Espevik T, et al
. Decreased bone formative and enhanced resortive markers in human immunodeficiency virus infection: indication of normalization of the bone-remodeling process during highly active antiretroviral therapy. J Clin Endocrinol Metab
5. Nolan D, Upton R, McKinnon E. Stable or increasing bone mineral density in HIV-infected patients treated with nelfinavir or indinavir. AIDS
6. Huang JS, Rietschel P, Hadigan CM. Increased abdominal visceral fat is associated with reduced bone density in HIV-infected men with lipodystrophy. AIDS
7. Carr A, Miller J, Eisman J, Cooper D. Osteopenia in HIV-infected men: association with asymptomatic lactic acidemia and lower weight pre-antiretroviral therapy. AIDS
8. Abu-Amer Y, Tondravi MM. NF – [Kappa] B and bone: the breaking point. Nat Med
9. Levy JA. Pathogenesis of HIV infection. Microbiol Rev
10. Buck BE, Resnick L, Shah SM, Malinin T. HIV cultured from bone: Implication for transplantation. Clin Orthop
11. Huang CJ, Muller F, Aukrust P, Froland SS. Subnormal serum concentration of 1.25 dihydroxyvitamin D3 in HIV infection: correlation with degree of immunodeficiency and survival. J Infect Dis
12. Huang CJ, Muller F, Aukrust P, Froland SS, Lien E. Disseminated Mycobacterium aviumcomplex in AIDS: immunopathogenic significance of an activated tumor necrosis factor system and depressed serum levels of 1,25 dihydroxyvitamin D. J Infect Dis
13. Haug CJ, Aukrust P, Haug E, Morkrid L, Muller F, Froland SS, et al
. Severe deficiency of 1.25-dihydroxyvitamin D3 in human immunodeficiency virus infection: association with immunological hyperactivity and only minor changes in calcium homeostasis. J Clin Endocrinol Metab
14. Dusso A, Vidal M, Powderly WG,Yarashesky KE, Tebas P. Protease inhibitors inhibit in vitro conversion of 25[OH]-vitamin D to 1,25[OH]2-vitamin D. Antiviral Ther
(Suppl 5):19 [abstract 030].
15. Clerici M, Trabattoni D, Piconi S. A possible role for cortisol/anticortisol imbalance in the progression of HIV. Psychoneuroendocrinology
16. Estephens EA, Das R, Madge S. Symptomatic osteoporosis in two young HIV-positive African women. AIDS
17. Guaraldi G, Ventura P, Albuzza M. Pathological fractures in AIDS patients with osteopenia and osteoporosis induced by antiretroviral therapy. AIDS
18. Tebas P, Powderly WG, Claxton SH, Marin D, Tantisiriwat W, Teitelbaum S, et al
. Accelerated bone mineral loss in HIV-infected patients receiving potent antiretroviral therapy. AIDS
19. Moore AL, Vashisht A, Sabin C, Mocroft A, Madge S. Reduced bone mineral density in HIV-positive individuals. AIDS
20. Nolan D, Upton R, McKinnon E. Longitudinal analysis of bone mineral density [DMO] in HIV-infected patients treated with HAART: change in subcutaneous fat; with an additional independent effect of indinavir therapy to increase DMO. Antiviral Ther
21. Hoy J, Hudson J, Law M, Cooper DA. Osteopenia in a randomized multicenter study of protease inhibitor substitution in patients with the lipodystrophy syndrome and well – controlled HIV viremia. Seventh Conference on Retroviruses and Opportunistic Infections
. San Francisco, January–February 2000 [abstract 59].
22. Hoy J, Hudson J, Cooper DA. Osteopenia in a randomized multicenter study of protease inhibitor substitution in patients with the lipodystrophy syndrome extended follow-up 48 weeks. Antiviral Ther
(Suppl. 5):42. [abstract P32].
23. Moyle GJ,Newey C, Baldwin C, Torti C, Mandalia S. Osteopenia: a consequence of HIV not HAART? Antiviral Ther
(Suppl. 5):43. [abstract P32].
24. Lawal A, Engelson ES, Wang J. Equivalent osteopenia in HIV-infected individuals studied before and during the era of highly active antiretroviral therapy. AIDS
25. Knobel H, Guelar A, Vallecillo G, Nogues X, Diez A. Osteopenia in HIV-infected patients: is it the disease or is it the treatment?. AIDS
26. Carr A. Osteopenia in HIV infection. AIDS Clin Care
27. World Health Organization. Assessment of Fracture Risk and its Application to Screening for Postmenopausal Osteoporosis.
Technical report series. Geneva: World Health Organization; 1994.
© 2003 Lippincott Williams & Wilkins, Inc.