Secondary Logo

Journal Logo

CLINICAL SCIENCE

Reduced bone mineral density among HIV-infected, virologically controlled young men

prevalence and associated factors

Shaiykova, Arnooa; Pasquet, Armelleb; Goujard, Cécilea,c; Lion, Georgesd; Durand, Emmanuele; Bayan, Tatianaa; Lachâtre, Marieb,f; Choisy, Philippeb; Ajana, Faïzab; Bourdic, Katiac; Viget, Nathalieb; Riff, Bertrandb,g; Quertainmont, Yannc; Cortet, Bernardh; Boufassa, Faroudya,*; Chéret, Antoinec,*

Author Information
doi: 10.1097/QAD.0000000000002001
  • Free

Abstract

Introduction

Combination antiretroviral therapy (cART) has radically improved the prognosis of people living with HIV (PLHIV) [1]. However, this therapeutic advance gives rise to, and permits the persistence of, several comorbidities and metabolic complications, related both to the virus and cART [2]. Among them, accelerated loss of bone mass, that is osteopenia, characterized by decreased bone density, and osteoporosis, by porous and fragile bones, has been described in PLHIV. Numerous studies have shown that the prevalence of reduced bone mineral density (BMD) is two to six-fold higher in PLHIV than HIV-negative individuals [3]. Thus, the prevalence of osteopenia varies from 22 to 66% and osteoporosis from 1 to 26.8% in PLHIV of all ages [4]. Reduced BMD contributes to an increased risk of fractures, which was three-fold higher in this population, representing a major cause of morbidity [5].

Causes of BMD reduction in PLHIV appear to be multifactorial, including those related to HIV infection and traditionally established osteoporosis risk factors, such as hypogonadism in men and menopause in women, advanced age, low BMI, physical inactivity, tobacco consumption, alcohol abuse and decreased intake of calcium and vitamin D [4,6–8]. The duration of infection, high viral load, low CD4+ cell count and exposure to some classes of antiretroviral drugs, such as nucleoside reverse transcriptase inhibitors (NRTIs) or protease inhibitors, have also been highlighted by some authors [9–11].

Most studies to evaluate reduced BMD have included heterogeneous populations composed of men and women of all ages, virologically suppressed or not. Current recommendations for the evaluation of bone disease by dual-energy X-ray absorptiometry (DXA) screening and its management in PLHIV are age-specific and concern mainly postmenopausal women and men older than 50 years [12,13]. Furthermore, most PLHIV patients living in industrialized countries are virologically suppressed [14]. To our knowledge, no studies have yet been performed in PLHIV men under 50 years, with undetectable viral load, under cART. Thus, the aim of our study was to determine the prevalence of reduced BMD and its associated factors in this population.

Materials and methods

Study design and participants

A bicentric cross-sectional survey was established in Tourcoing and Kremlin-Bicêtre Hospitals in France. Patients were consecutively included between 1 January 2013 and 30 June 2016. They were selected according to the following criteria: men aged between 18 and 50 years, treated by cART for over 6 months, with undetectable plasma HIV-1-RNA (i.e. HIV-RNA <50 copies/ml) for at least 6 last months, and free of opportunistic diseases or hepatitis B or C coinfection.

At their first clinical visit during the study period, questionnaires on demographic (including lifestyle habits) and anthropometric characteristics were completed. Blood samples were collected for measurement of hormonal (testosterone and oestradiol), metabolic (glycaemia and lipid profile), infectious (HIV-DNA and RNA), immunological (CD4+ and CD8+ cell counts) and bone remodelling parameters [serum calcium, creatinine, peptide C, 25-OH Vitamin D and turnover markers of bone formation and resorption: N-terminal propeptide of type one collagen (P1NP) and cross laps (CTX)]. Other variables, such as HIV transmission group, time since HIV diagnosis, CDC AIDS stage, nadir CD4+ cell count, current and previous cART regimens, cumulative duration on cART and medication intake affecting BMD were extracted from an electronic medical record used for the follow-up of PLHIV. The study was approved by the Lille Ethics Committee and the French Health Products Safety Agency and complied with the Helsinki Declaration. All participants gave written informed consent. This study was registered with ClinicalTrials.gov, number NCT: 2012-005231-93.

Laboratory methods

Biological parameters were measured from fasting blood samples. Measurements were analysed centrally in the Biology Laboratory at Lille Hospital. Serum oestradiol levels and total testosterone were assessed by radioimmunoassay. Free serum testosterone was measured twice, 1 week apart, in the morning. Male hypogonadism was assessed and defined as an average free serum testosterone less 70 pg/ml, calculated with the Vermeulen equation [15]. Electrochemiluminescence immunoassay (ECLIA) was used to quantify bone-remodelling markers in plasma: peptide C, 25-OH vitamin D, P1NP and CTX. Calcium and serum creatinine levels were assessed by colorimetry in each hospital. Renal dysfunction was sought by calculating the glomerular filtration rate (GFR) using the MDRD (modification of diet in renal disease) equation. Plasma HIV-DNA and RNA levels were measured by real-time PCR and CD4+ and CD8+ cell counts by flow cytometry.

Bone mineral density assessment

BMD was assessed, at the lumbar spine (L1–L4) and left hip by DXA. Hologic Discovery A was used for patients followed at Tourcoing and Prodigy Advance GE Lunar for those followed at Kremlin-Bicêtre. Machines were verified by daily quality assurance tests. Conversion formulas, obtained from GE Healthcare Lunar, were applied to standardize BMD measurements between the two machines. Z and T scores were calculated from the site-specific DXA values, expressed in g/cm2, using normative data matched for sex, race and age. A history of low energy fractures, resulting from minimal or no recognizable trauma, was systematically sought and collected.

Outcomes

Z-score is the standard score used to interpret BMD results in premenopausal women and men under the age of 50 years [12]. We therefore used Z-scores as the primary outcome in this study. For young populations, the International Society of Clinical Densitometry (ISCD) recommends considering Z-scores of −2.0 or less as low BMD or ‘below the expected range for age’. In our main analysis, we considered a Z-score less than −1 to reflect a decrease in BMD, and defined two grades of decrease as follows: osteoporosis corresponded to the lowest Z-score for the lumbar spine or hip less than −2.0, and osteopenia to a Z-score between −1.0 and −2.0. Normal BMD was defined as a Z-score more than −1 at both sites.

We also established secondary outcomes for sensitivity analysis because of the lack of a clear definition of reduced BMD in young populations. We first focused on the comparison of low BMD (Z-score ≤−2.0) with normal BMD (Z-score >−2), as recommended by the ISCD. We then looked at the definitions of osteopenia and osteoporosis based upon T-scores: patients were considered to have osteopenia or osteoporosis if the lowest T-score at either site was between −1 and −2.5 or −2.5 or less, respectively, based on WHO criteria [16]. Finally, we analysed DXA values at the hip and lumbar spine separately, as a continuous outcome.

Statistical analyses

Baseline characteristics were described for all patients in accordance with the Z-score group. Quantitative variables are expressed as medians with interquartile ranges (IQRs) and were compared using the Kruskal–Wallis test. Categorical variables are expressed as frequencies and percentages and were compared using the Chi-square or Fisher exact test. We evaluated factors associated with osteopenia and osteoporosis, according to the Z-score, using univariate and multivariate polytomous logistic regressions. Binary and polytomous logistic regressions were used similarly in sensitivity analyses based on the low BMD and T-score derived definitions. Finally, factors associated with DXA values were evaluated in univariate and multivariable linear regressions. Variables with P value less than 0.10 in univariate analyses were included in the final models for all tests. Continuous variables that did not verify the linearity assumption or because of clinical significance were classified into categories. All models were adjusted for centre, age, physical activity and tobacco consumption. The goodness of fit of the final logistic regression models was verified by Hosmer and Lemeshow tests. All statistical tests were two-tailed and P values less than 0.05 were considered to be statistically significant. Analyses were performed using STATA 14 Software (Stata Corp, College Station, Texas, USA).

Results

Patient characteristics

Two hundred and forty participants were included in this study; 10 (4.2%) did not have a DXA performed and were therefore excluded from our analyses. For one patient, only a DXA of the lumbar spine was performed. Hence, he was excluded from the linear regression analyses for BMD of the hip. Our participants were mostly white (n = 208; 90.4%) and MSM (n = 181; 78.7%), with a median age of 43 years (IQR 36–47), and a high median CD4+ cell count [632 cells/μl (IQR 508–769)] (Table 1).

Table 1
Table 1:
Baseline characteristics of all HIV-infected patients and according to Z-score group.

Seventy-five patients [32.6% (confidence interval, CI, 26.8–39.0)] met the definition of osteopenia and 36 [15.7% (CI, 11.5–21.0)] osteoporosis based on Z-scores. For binary Z-score outcome, 194 (84.4%) individuals were considered to have a normal BMD. On the basis of T-score, 107 men [46.5% (CI, 40.1–53.0)] were considered to have osteopenia, and 23 (21.5%) of them were considered to have a normal BMD by Z-score. Thirteen (12.2%) of the 24 men [10.4% (CI, 7.0–15.1)] with T-score placing them in the osteoporosis group were also classified as having osteoporosis by Z-score (Table 2).

Table 2
Table 2:
Distribution of osteopenia/osteoporosis definitions according to T and Z scores.

On the basis of the Z-scores, the groups were comparable with respect to alcohol use (n = 21; 9.3%), tobacco consumption (n = 93; 40.4%) and patient-reported regular physical activity (n = 132; 57.4%). Levels of 25-OH vitamin D and serum creatinine did not significantly differ between groups and were within the normal range. Two patients presented moderate renal impairment, that is GFR of 60 ml/min per 1.73 m2 or less and two a history of traumatic fractures, both at the femoral neck. Twenty men (8.7%) could be considered as having hypogonadism, assuming free serum testosterone concentrations. We observed a decrease in BMI corresponding to BMD classification, despite all patients having a BMI in the normal range. Median serum oestradiol levels were significantly lower in the osteoporotic group. The time since the diagnosis of HIV-infection and the duration of cART were significantly lower in osteopenic, with medians of 5.0 (IQR 2–11) and 3.3 (IQR 1.3–6.0) years, respectively, than nonosteopenic patients.

Factors associated with osteopenia and osteoporosis (Z-score)

The following factors were significantly associated with reduced BMD by univariate polytomous logistic regression: BMI, total lean mass less than 61 kg, serum peptide C, serum oestradiol and duration of cART. Duration of NRTI therapy and exposure to NNRTI were also significantly associated with osteopenia, but not osteoporosis. Similarly, when focusing on antiretroviral drugs separately, there were significant associations with duration of lamivudine or emtricitabine therapy (3TC/FTC), of lopinavir therapy (LPV) and exposure to efavirenz (EFV), whilst tenofovir disoproxil fumarate (TDF) was marginally associated with osteopenia.

In multivariate models (Table 3), a decrease by 1 kg/m2 of BMI was significantly associated with both osteopenia [adjusted odds ratio (aOR) = 1.16, P = 0.01] and osteoporosis (aOR = 1.24, P = 0.01). A decrease of 5 pg/ml oestradiol was independently associated only with osteoporosis (aOR = 1.32, P = 0.05). There was a protective effect against osteopenia for a 1-year increase in the duration of cART (aOR = 0.87, P < 0.01), which was even greater when the duration was more than 3 years (aOR = 0.44, P = 0.02). In the model for total lean mass, the risk of osteopenia was approximately three-fold higher when the lean mass was less than 61 kg (aOR = 2.98, P < 0.01), but no factor was associated with the risk of osteoporosis.

Table 3
Table 3:
Factors associated with reduced bone mineral density in three multivariate logistic regression models.

Four different models were generated to further explore which antiretroviral drugs were associated with reduced BMD (Fig. 1). After adjustments, the association between duration of TDF treatment and osteopenia was no longer significant (P = 0.18). However, similar to cART, the duration of lamivudine or emtricitabine (P = 0.04) or lopinavir (P < 0.01) treatment, or exposure to efavirenz (P = 0.02) showed a protective effect against osteopenia and did not change other associations.

Fig. 1
Fig. 1:
Multivariate logistic regression models with four different antiretroviral therapies.(a) Model 1: BMI, serum oestradiol, duration on tenofovir disoproxil fumarate (TDF) (years); (b) Model 2: BMI, serum oestradiol, duration on lamivudine (3TC) or emtricitabine (FTC) (years); (c) Model 3: BMI, serum oestradiol and exposure to efavirenz (EFV); (d) Model 4: BMI, serum oestradiol and duration on lopinavir (LPV) (years). All models were adjusted for centre, age, physical activity and tobacco consumption. Odds ratios with confidence intervals are shown as squares and circles, for osteopenia and osteoporosis, respectively.

Sensitivity analyses

Low bone mineral density

In univariate and adjusted multivariate models, the factors associated with low BMD in binary logistic regression were identical to those associated with osteoporosis in primary analysis: BMI (aOR = 1.17, P = 0.03) and oestradiol (aOR = 1.37, P = 0.02).

Osteopenia and osteoporosis based upon T-score

In the multivariate analyse, a decrease by 1 kg/m2 of BMI was associated with osteopenia (aOR = 1.17, P = 0.03) and osteoporosis (aOR = 1.59, P < 0.001), based on the T-score. Serum oestradiol was found to be of borderline significance (aOR = 1.31, P = 0.09) in the osteoporotic group and exposure to efavirenz (P = 0.05) in the osteopenic group. In a second multivariate model, the risk of osteopenia (aOR = 4.03, P < 0.001) or osteoporosis (aOR = 4.41, P = 0.01) was four-fold higher when the lean mass was less than 61 kg.

Dual-energy X-ray absorptiometry (g/cm2) values at the hip and lumbar spine

In multivariable models, only BMI decrease [regression coefficient (β) = −0.02, P < 0.001] and lower total lean mass (β = −0.08, P < 0.001) were independently associated with BMD reduction at the hip. In contrast, BMI (β = −0.02, P < 0.001), lean mass (β = −0.06, P < 0.001), duration of cART (β = 0.005, P = 0.01), particularly the duration of lamivudine or emtricitabine treatment (β = 0.007, P = 0.001) and exposure to efavirenz (β = 0.04, P < 0.01), were independently associated with BMD at the lumbar spine in different multivariable adjusted models.

Discussion

This study is the first to show an unexpectedly high prevalence of reduced BMD in treated PLHIV men under 50 years of age with undetectable viral load. Among them, 32.6% were considered to have osteopenia and 15.7% osteoporosis, and these prevalences are similar to those reported for PLHIV of all ages [4]. Epidemiological published data in young general population are uncommon because idiopathic osteopenia and osteoporosis are very rare. Nevertheless, the sporadic cases of reduced BMD described in young patients were mainly attributed to chronic diseases, to an accidental discovery after a fracture or to some medications [17,18].

The choice of definitions for osteopenia and osteoporosis for this study was challenging given the relatively young age of our population. Our Z-score derived definitions are not standard clinical definitions. However, they made it possible for us to analyse and present our data in three groups according to Z-score, as we do for elderly populations using T-scores, to identify those likely to have the poorest bone health in our population. We therefore considered a Z-score less than −1 to reflect a decrease in BMD. Several other studies on young populations have made use of T-scores [9,19], whereas others used Z-scores, with a threshold of −2 or less defining a low BMD [20–22]. The lack of direct comparability between our results and those of published studies led us to carry out sensitivity analyses, using T-scores and low BMD defined as a Z-score of −2 or less as a binary outcome. We observed that the prevalences of osteopenia and osteoporosis differed according to the score used. The prevalence of a reduced BMD seemed to be higher for T-score, particularly for the osteopenic group. Regardless of the score applied, the prevalences of osteopenia and osteoporosis were high in this population. Our findings therefore clearly indicate that early bone demineralization occurs in young treated PLHIV men, possibly indicating premature ageing of this population [23]. Our definition is, therefore, likely to be useful for future studies in such populations.

Some associated factors were similar to those found in studies of the HIV-negative elderly population. This was notably the case for serum oestradiol, the low level of which was significantly associated with low BMD and osteoporosis regardless of score used. The importance of the role of oestradiol on BMD has often been described in postmenopausal women and has already been observed in elderly men of the general population [24,25]. With regard to PLHIV men, it has been hypothesized that hypogonadism, defined by low total testosterone levels, is more strongly associated with bone density than oestrogen levels [26]. However, Santi et al.[27] showed that serum oestradiol at least 27 pg/ml may be protective for bone health, even in PLHIV men. Serum testosterone levels, and thus hypogonadism, were not associated with bone density [27], similar to our study. Our results strengthen the hypothesis of a protective effect of serum oestradiol on BMD in young men living with HIV. The fact that this association was found in young individuals once again suggests premature ageing of PLHIV [23].

BMI, total lean mass and duration of cART were similarly associated with DXA values at the lumbar spine, indicating at least osteopenia according to the Z-score. The duration of cART was not found to be associated with reduced BMD in analyses based on the T-score derived definition or with low BMD. These findings support our choice of definition and highlight the importance of categorizing Z-score into three groups.

In our study, lower BMI was significantly associated with the risk of osteopenia and osteoporosis (regardless of the score used) and low BMD [4,6], despite the values obtained being in the normal range. Our findings support that BMI, and thus weight variation, is a strong predictor of BMD.

We also found that lower total lean mass was significantly associated with the risk of osteopenia, but found no association with fat mass. HIV infection itself and some cART regimens might affect body composition and the distribution of fat and lean mass and thus contribute to premature ageing of PLHIV [6]. The fact that we observed this association in the osteopenic group only may be explained by a significantly shorter duration of HIV infection and cART exposure than in osteoporotic patients. Indeed, PLHIV individuals had significantly greater increases in lean mass and total fat during the first 2 years after cART initiation than untreated HIV-negative controls as shown by Grant et al.[28].

Moreover, a longer duration of cART had a protective effect on BMD in our osteopenic group. Several randomized controlled trials have shown that accelerated bone mass loss occurs mainly between 6 and 12 months after cART initiation, followed by stabilization of BMD [9]. Longitudinal studies have been conducted to evaluate the long-term effect of cART on the progression of BMD in PLHIV [29]. Patients with untreated HIV infection tend to have decreased bone turnover, which then increases after the first 2 years after cART initiation and immunological recovery, before stabilizing thereafter [29]. In our study, the risk of osteopenia was higher for a duration of cART of less than 3 years. Our osteopenic patients were probably enrolled in the study during the ‘BMD recovery phase’ (the median duration of cART was approximately 3 years). In contrast, the association was not significant for patients with osteoporosis, probably because of the longer duration of HIV infection and cART exposure, which may correspond to the ‘BMD stabilization phase’.

When we considered antiretroviral agents separately, the duration of lamivudine or emtricitabine, lopinavir therapy and exposure to efavirenz showed the same associations. In contrast, the duration of TDF exposure was not associated with osteopenia, even though TDF may affect kidney function and bone density [30]. Several studies have highlighted the relationship between bone demineralization and renal impairment via phosphate metabolism [30,31]. Only two patients in our sample had moderate renal impairment. In addition, TDF was always combined with other drugs, particularly efavirenz and emtricitabine, which showed a protective effect against osteopenia. In addition, Bedimo et al.[31] showed that although EFV + FTC + TDF was associated with BMD reduction, which was less pronounced than with other TDF-containing regimens, such as protease inhibitor or cobicistat, and presented a more favourable renal and bone safety profile. The absence of a significant association between reduced BMD and TDF was probably offset by the observed protective effect of the other drugs.

Chronic inflammation caused by HIV infection itself is generally associated with increased bone resorption in patients with high plasma viral load. A direct effect of HIV on osteogenic cells, persistent activation of pro-inflammatory cytokines and alterations in vitamin D metabolism are observed in cART-naive individuals [32]. In contrast to other studies, our participants had normal vitamin D concentrations, high CD4+ cell counts and undetectable HIV-RNA, with probably low-level immune activation. Thus, the frequency of reduced BMD remained high, even with a favourable immunovirological status. However, immunological recovery with cART is essential for the restoration of bone health [29].

The cross-sectional design, which is the main limitation of our study, will nevertheless provide a means of exploring associated factors in depth in future longitudinal studies. In addition, 90% of our patients were treated white men. It will thus be necessary to set up a study with a relatively large representative sample of all PLHIV, including premenopausal women. Moreover, our results are not representative of patients with lower CD4+ cell counts and higher BMI, two factors likely to worsen bone health. Finally, as we wished to investigate the reduction of BMD in healthy PLHIV, we excluded patients with hepatitis coinfections from this study. However, hepatitis B or C coinfections may worsen the impact on BMD, and it would be interesting for subsequent studies to take them into account.

Finally, our study highlights the significant prevalence of reduced BMD in nonelderly PLHIV, suggesting premature aging of PLHIV. Although the number of fractures was negligible in our patients, the observed reduced BMD could lead to an increased risk of fracture and worsening of quality of life as they age. The evaluation and management of bone disease in PLHIV is recommended for postmenopausal women and men older than 50 years. However, our study raises the question of extending the recommendations for BMD assessment to PLHIV less than 50 years. It will thus be preferable to evaluate BMD after it stabilizes and waiting for at least 3 years of effective cART.

Acknowledgements

We thank the entire staff of the Infectious Diseases and Internal Medicine Departments of Tourcoing and Bicêtre Hospitals: P. Cornavin, M. Mole, S. Vandamme, E. Teicher, T. Huleux, I. Alcaraz, E. Senneville and A. Meybeck. We also thank the patients for their participation in the study.

A.C. and A.P. were the chief investigators and designed and developed the protocol with M.L., A.C., C.G., B.R., Y.Q., A.P., N.V. and F.A. enrolled patients. F.B., A.P., K.B., A.S., T.B., P.C. and A.C. coordinated the data collection and regulatory requirements. A.S., F.B. and T.B. were responsible for the data analysis and A.C. and B.C. their interpretation. G.L. and E.D. were responsible for performing the DXA. A.S., F.B. and A.C. wrote the manuscript and all authors reviewed, revised and approved the final version.

This work was supported by the French National Agency for Research on AIDS and Viral Hepatitis, COREVIH Nord-Pas-de-Calais and MSD.

Conflicts of interest

A.C. reports grants from Merck, ViiV and personal fees from Gilead and Janssen. C.G. declares to have received conference fees and other financial support from Janssen, Gilead, ViiV and MSD, unrelated to the submitted work. All the other authors declare that they have no conflict of interest.

References

1. Deeks SG, Lewin SR, Havlir DV. The end of AIDS: HIV infection as a chronic disease. Lancet 2013; 382:1525–1533.
2. Antiretroviral Therapy Cohort Collaboration. Causes of death in HIV-1—infected patients treated with antiretroviral therapy, 1996-2006: collaborative analysis of 13 HIV cohort studies. Clin Infect Dis 2010; 50:1387–1396.
3. Brown TT, Qaqish RB. Antiretroviral therapy and the prevalence of osteopenia and osteoporosis: a meta-analytic review. AIDS Lond Engl 2006; 20:2165–2174.
4. Stone B, Dockrell D, Bowman C, McCloskey E. HIV and bone disease. Arch Biochem Biophys 2010; 503:66–77.
5. Prieto-Alhambra D, Güerri-Fernández R, De Vries F, Lalmohamed A, Bazelier M, Starup-Linde J, et al. HIV infection and its association with an excess risk of clinical fractures: a nationwide case–control study. J Acquir Immune Defic Syndr 2014; 66:90–95.
6. Kooij KW, Wit FWNM, Bisschop PH, Schouten J, Stolte IG, Prins M, et al. Low bone mineral density in patients with well suppressed HIV infection: association with body weight, smoking, and prior advanced HIV disease. J Infect Dis 2015; 211:539–548.
7. Overton ET, Chan ES, Brown TT, Tebas P, McComsey GA, Melbourne KM, et al. Vitamin D and calcium attenuate bone loss with antiretroviral therapy initiation: a randomized trial. Ann Intern Med 2015; 162:815–824.
8. Compston J. HIV infection and bone disease. J Intern Med 2016; 280:350–358.
9. Hansen AB, Obel N, Nielsen H, Pedersen C, Gerstoft J. Bone mineral density changes in protease inhibitor-sparing vs. nucleoside reverse transcriptase inhibitor-sparing highly active antiretroviral therapy: data from a randomized trial bone mineral density changes in highly active antiretroviral therapy. HIV Med 2011; 12:157–165.
10. McComsey GA, Kitch D, Daar ES, Tierney C, Jahed NC, Tebas P, et al. Editor's choice: bone mineral density and fractures in antiretroviral-naive persons randomized to receive abacavir-lamivudine or tenofovir disoproxil fumarate-emtricitabine along with efavirenz or atazanavir-ritonavir: AIDS Clinical Trials Group A5224s, a substudy of ACTG A5202. J Infect Dis 2011; 203:1791–1801.
11. Grant PM, Kitch D, McComsey GA, Dube MP, Haubrich R, Huang J, et al. Low baseline CD4+ count is associated with greater bone mineral density loss after antiretroviral therapy initiation. Clin Infect Dis 2013; 57:1483–1488.
12. Cosman F, de Beur SJ, LeBoff MS, Lewiecki EM, Tanner B, Randall S, et al. Clinician's guide to prevention and treatment of osteoporosis. Osteoporos Int 2014; 25:2359–2381.
13. Brown TT, Hoy J, Borderi M, Guaraldi G, Renjifo B, Vescini F, et al. Recommendations for evaluation and management of bone disease in HIV. Clin Infect Dis 2015; 60:1242–1251.
14. Mary-Krause M, Grabar S, Lièvre L, Abgrall S, Billaud E, Boué F, et al. Cohort profile: French hospital database on HIV (FHDH-ANRS CO4). Int J Epidemiol 2014; 43:1425–1436.
15. Vermeulen A, Verdonck L, Kaufman JM. A critical evaluation of simple methods for the estimation of free testosterone in serum. J Clin Endocrinol Metab 1999; 84:3666–3672.
16. Assessment of fracture risk and its application to screening for postmenopausal osteoporosis. Report of a WHO Study Group. World Health Organ Tech Rep Ser 1994; 843: 1–129.
17. Khosla S, Lufkin EG, Hodgson SF, Fitzpatrick LA, Melton LJ. Epidemiology and clinical features of osteoporosis in young individuals. Bone 1994; 15:551–555.
18. Ferrari S, Bianchi ML, Eisman JA, Foldes AJ, Adami S, Wahl DA, et al. Osteoporosis in young adults: pathophysiology, diagnosis, and management. Osteoporos Int 2012; 23:2735–2748.
19. Tsai MS, Hung CC, Liu WC, Chen KL, Chen MY, Hsieh SM, et al. Reduced bone mineral density among HIV-infected patients in Taiwan: prevalence and associated factors. J Microbiol Immunol Infect 2014; 47:109–115.
20. Harris VW, Brown TT. Bone loss in the HIV-infected patient: evidence, clinical implications, and treatment strategies. J Infect Dis 2012; 205 (Suppl 3):S391–S398.
21. Casado JL, Bañon S, Andrés R, Perez-Elías MJ, Moreno A, Moreno S. Prevalence of causes of secondary osteoporosis and contribution to lower bone mineral density in HIV-infected patients. Osteoporos Int 2014; 25:1071–1079.
22. Paccou J, Viget N, Drumez E, Cortet B, Robineau O. Prevalence and risk factors for low bone mineral density in antiretroviral therapy-naive HIV-infected young men. Med Mal Infect 2018.
23. Pathai S, Bajillan H, Landay AL, High KP. Is HIV a model of accelerated or accentuated aging?. J Gerontol A Biol Sci Med Sci 2014; 69:833–842.
24. Rochira V, Kara E, Carani C. The endocrine role of estrogens on human male skeleton. Int J Endocrinol 2015; 2015:165215.
25. Lormeau C, Soudan B, d’Herbomez M, Pigny P, Duquesnoy B, Cortet B. Sex hormone-binding globulin, estradiol, and bone turnover markers in male osteoporosis. Bone 2004; 34:933–939.
26. Rochira V, Guaraldi G. Hypogonadism in the HIV-infected man. Endocrinol Metab Clin North Am 2014; 43:709–730.
27. Santi D, Madeo B, Carli F, Zona S, Brigante G, Vescini F, et al. Serum total estradiol, but not testosterone is associated with reduced bone mineral density (BMD) in HIV-infected men: a cross-sectional, observational study. Osteoporos Int 2016; 27:1103–1114.
28. Grant PM, Kitch D, McComsey GA, Collier AC, Bartali B, Koletar SL, et al. Long-term body composition changes in antiretroviral-treated HIV-infected individuals. AIDS Lond Engl 2016; 30:2805–2813.
29. Bolland MJ, Grey A, Reid IR. Skeletal health in adults with HIV infection. Lancet Diabetes Endocrinol 2015; 3:63–74.
30. Hamzah L, Samarawickrama A, Campbell L, Pope M, Burling K, Walker-Bone K, et al. Effects of renal tubular dysfunction on bone in tenofovir-exposed HIV-positive patients. AIDS Lond Engl 2015; 29:1785–1792.
31. Bedimo R, Rosenblatt L, Myers J. Systematic review of renal and bone safety of the antiretroviral regimen efavirenz, emtricitabine, and tenofovir disoproxil fumarate in patients with HIV infection. HIV Clin Trials 2016; 17:246–266.
32. Erlandson KM, O’Riordan M, Labbato D, McComsey GA. Relationships between inflammation, immune activation and bone health among HIV-infected adults on stable antiretroviral therapy. J Acquir Immune Defic Syndr 2014; 65:290–298.

* Both Faroudy Boufassa and Antoine Chéret contributed equally to this work.

Keywords:

bone mineral density; combination antiretroviral therapy; HIV; young men

Copyright © 2018 Wolters Kluwer Health, Inc.