Increased prevalence of osteopenia and osteoporosis is well documented in both antiretroviral-naive and antiretroviral-treated HIV-infected persons . Chronic HIV infection is associated with immune activation, chronic inflammation, and low body mass index (BMI), which are all established risk factors for low bone mineral density (BMD). Antiretroviral treatment may also cause accelerated bone loss. A number of randomized controlled trials have demonstrated accelerated bone loss in the initial 6–12 months after highly active antiretroviral therapy (HAART) initiation followed by stabilization of BMD hereafter [2–4]. Further, prospective studies of patients on established HAART showed stable or increasing BMD [5,6]. Although there are drug and drug class differences, especially tenofovir causes a larger initial bone loss [2,4], the initial BMD loss after HAART initiation has been observed for all combinations of major drug classes in the HAART regimen [3,7–9]. Recent studies have also shown low BMD in men who have sex with men (MSM) with primary HIV infection  and in HIV-negative MSM at high risk of HIV infection  thus suggesting importance of risk factors associated with both risk of HIV infection and with low BMD.
The clinical impact of the changes in BMD remains uncertain. Most studies have found increased fracture rates among HIV-infected individuals compared with HIV-negative controls [12–15], whereas a study of premenopausal women did not . Especially the influence of the short-term bone loss associated with HAART initiation on long-term risk of fracture is not well described. Only few studies have specifically addressed the influence of HAART treatment on subsequent fracture risk. Two American studies found no association between HAART exposure and risk of fracture [14,16]; recently Bedimo et al.  found increased fracture risk in the HAART era compared with the pre-HAART era but primarily ascribed the increased risk to non-HIV-related risk factors. Some of the published studies have been hampered by small study populations, absence of information on hepatitis C status, included mainly intravenous drug users, or did not conduct separate analyses for low-energy and high-energy fractures.
We aimed to compare the overall incidence of fractures and the incidence of low-energy and high-energy fractures in antiretroviral-naive and HAART-treated HIV-infected patients with the general population during the period 1 January 1995 to 1 January 2010. Linking data from the population-based Danish HIV Cohort Study with the Danish Civil Registration System and Danish National Hospital Registry allowed us to capture all fractures diagnosed at hospital admissions and at outpatient or emergency visits for both HIV-infected persons and matched general population controls.
The adult population of Denmark is 4.3 million with an estimated HIV prevalence of 0.09% . Patients with HIV infection are treated in one of the country's eight specialized medical centres, where they are seen on an outpatient basis at intended intervals of 12 weeks. Antiretroviral treatment is provided free of charge to all HIV-infected residents in Denmark and prescribed according to international guidelines.
We used the unique 10-digit civil registration number assigned to all individuals in Denmark at birth or upon immigration to link the data sources described below.
The Danish HIV Cohort study (DHCS) is a population-based prospective nationwide cohort study of all HIV-infected individuals 16 years or older and who have been treated at Danish HIV centres since 1 January 1995 [18,19]. Patients are consecutively enrolled, and multiple registrations are avoided through the use of the unique civil registration number. Data are updated yearly and include demographics, date of HIV infection, AIDS-defining events, smoking status, date and cause of death and antiretroviral treatment. CD4+ cell counts and HIV-RNA measurements are extracted electronically from laboratory data files.
From the Danish Civil Registration System (CRS), which has stored information on all Danish residents since 1968, we extracted information on vital status, residency, immigration and emigration .
The Danish National Hospital Register (DNHR) was initiated in 1977 and contains information on all patients discharged from Danish nonpsychiatric hospitals. Since 1995 data on outpatients and emergency patients have been included as well. Each record includes the dates of admission and discharge, 1 primary and up to 19 secondary discharge diagnosis coded according to the International Classification of Diseases, 8th revision (ICD-8) until the end of 1993 and the 10th revision (ICD-10) thereafter. From this register, we extracted date of first fracture diagnosis, date of first low-energy and date of first high-energy fracture diagnosis .
HIV-infected and population control study populations
We included all HIV patients from the DHCS who received a diagnosis of HIV after the age of 16 years and lived in Denmark at the time of HIV diagnosis. For the HIV-infected patients we defined the index date as 1 January 1995 or the date of the HIV diagnosis which ever came last. From the CRS we identified five population controls for each HIV-infected patient, matched by age (month and year of birth) and sex. For the population controls, the index date was defined as the index date of the HIV-infected patients to whom they were matched. The population controls had to be alive and living in Denmark on the index date.
We had three outcomes: the main outcomes were time to the first fracture at any site (ICD-10 codes: M48.4-M48.5, M84.3, S02.0–S02.9, S12.0–S12.9, S22.0–S22.9, S32.0–S32.9, S42.0–S42.9, S52.0–S52.9, S62.0–S62.9, S72.0–S72.9, S82.0–S82.9, S92.0–S92.9, T02.0–T02.9, T.10.0, T10.9, T12.0, T12.9, T14.2) and time to first low-energy fracture. The secondary endpoint was time to first high-energy fracture. Similar to the approach used by Lippuner et al. , we defined low-energy fractures as fractures possibly due to osteoporosis, typically those caused by low-energy trauma. The group comprised 28 relevant ICD-10 codes: M48.4 (vertebral fatigue fracture); M48.5 (vertebral compression fracture, not classified elsewhere); M84.3 (stress fracture, not classified elsewhere); S22.0 (fracture of the thoracic spine); S22.1 (multiple fractures of the thoracic spine); S22.3 (rib fracture); S32.0 (fracture of the lumbar spine); S32.1 (fracture of the sacrum); S32.5 (fracture of the pubis); S32.7 (multiple fractures of the lumbar spine); S32.8 (other fractures of the lumbar spine); S42.2 (fracture of the proximal humerus); S42.3 (fracture of the humerus shaft); S52.2 (fracture of the ulna shaft); S52.5 (fracture of the distal radius); S52.6 (combined fracture of the distal radius/ulna); S72.0 (fracture of the femoral neck); S72.1 (pertrochanteric fracture); S72.2 (subtrochanteric fracture); S72.4 (fracture of the femur, distal); S72.8 (fracture of the femur, other); S72.9 (fracture of the femur, no further mention); S82.1 (fracture of the tibia, proximal); S82.2 (fracture of the tibia shaft); S82.3 (fracture of the tibia, distal); S82.4 (fracture of the fibula); S82.5 (fracture of the malleolar int.); S82.6 (fracture of the malleolar ext.).
To further explore the risk of low-energy fracture, we grouped the low-energy fractures as wrist, humerus, hip, vertebral, or other low-energy fractures.
High-energy fractures were defined as fractures unlikely to be due to osteoporosis, typically those caused by high-energy trauma . High-energy fractures comprised 29 relevant ICD-10 codes: S42.0 (fracture clavicula); S42.4 (fracture humerus distal); S42.7 (multiple fractures clavicula); S42.8 (other fractures of the shoulder, upper arm); S42.9 (fracture of the shoulder, no further mention); S52.0 (fracture ulna proximal); S52.1 (fracture radius proximal); S52.3 (fracture of the radius shaft); S52.4 (combined fracture of the radius/ulna); S52.7 (multiple fractures of the forearm); S52.8 (fracture of the forearm, other); S52.9 (fracture of the forearm, no further mention); S62 (fracture of the hand and wrist); S72.3 (fracture of the femur shaft); S72.7 (multiple fractures femur); S82.0 (fracture patella); S82.7 (multiple fractures of the lower leg); S82.8 (fractures of the lower leg, other); S82.9 (fracture of the lower leg, no further mention); S92 (fracture of the foot, excluding ankle); T02.1 (fractures involving thoracic and lumbar spine and pelvis); T02.2 (multiple fractures of the upper limb); T02.3 (multiple fractures of the lower limb); T02.4 (multiple fractures of both upper limbs); T02.5 (multiple fractures of both lower limbs); T02.6 (multiple fractures of the lower and upper limbs); T02.7 (fractures involving thoracic and lumbar spine and pelvis and limbs); T02.8 (other combined fractures); T02.9 (multiple fractures, no further mention).
Patients with at least one positive hepatitis C virus (HCV)-antibody test or a positive HCV-RNA test were considered HCV-positive.
Comorbidity was included in the analyses as a modified Charlson comorbidity index (CCI) based on discharge diagnoses registered in the DNHR prior to the index date for both populations. The CCI assigns a score between 1 and 6 to a range of diseases recorded during previous hospital contacts, with the sum of scores serving as the comorbidity measure for each patient . We captured the comorbid diseases using the ICD-8 and ICD-10 codes. AIDS-defining diagnoses were not included. We defined three modified comorbidity levels according to the CCI: low (score 0); medium (score 1–2); or high (score>2) [23,24].
Highly active antiretroviral therapy was defined as either combination antiretroviral treatment with at least three drugs that included a protease inhibitor, a non-nucleoside reverse transcriptase inhibitor (NNRTI), an integrase inhibitor, and/or abacavir; or a combination of a ritonavir-boosted protease inhibitor with an NNRTI or an integrase inhibitor.
We computed time from the index date to date of first fracture, death, lost to follow-up, emigration or 1 January 2010 whichever came first. Similarly, we computed time from index date to first low-energy or to first high-energy fracture, death, lost to follow-up, emigration or 1 January 2010 whichever came first. As a measure of relative risk of contracting a fracture we used Cox regression analysis to compute unadjusted incidence rate ratios (IRRs). For low-energy fractures we further estimated IRR adjusted for comorbidity. In Denmark, HCV infection is a marker of intravenous drug use and therefore we conducted separate analyses for HIV-monoinfected and HIV/HCV-coinfected patients .
For HIV-monoinfected patients, HIV/HCV-coinfected patients and population controls, we computed the cumulative incidence of low-energy and high-energy fractures, respectively. Further, for HIV-monoinfected patients and corresponding population controls we computed cumulative incidence of low-energy and high-energy fractures stratified by HAART treatment. In the analysis of fracture incidence before HAART initiation patients and corresponding population controls were followed from study entry until date of HAART initiation or censoring. In the analysis of fracture incidence after HAART initiation patients and corresponding controls were followed from date of HAART initiation until end of study.
Cumulative incidence of fracture was computed with death as competing risk to avoid overestimation of fracture risk . By using Schoenfeld residuals we determined that hazard ratios were constant within follow-up after HAART initiation.
To evaluate predictors of low-energy fractures for HIV-monoinfected patients on HAART we fitted a model including CD4 cell count before start of HAART, prior AIDS-defining event, sex, age (<35 years, 35–49 years, ≥50 years), race, CCI, and time of HIV diagnosis before or after 1 January 1995. We did not have complete data on smoking, therefore we estimated the effect of smoking (ever vs. never) on fracture risk in a separate model.
As tenofovir has been associated with more accelerated bone loss [2,4] and efavirenz has been associated with vitamin D deficiency [27,28] we performed analyses in which the start date of tenofovir or efavirenz was introduced as time-dependent variables from date of first exposure to the drug of interest until end of study. Tenofovir has been approved later than most other nucleoside reverse transcriptase inhibitors (NRTIs), which could introduce confounding. Abacavir has also been introduced later than other NRTIs and we undertook a similar analysis in which start date of abacavir was the time-dependent variable to allow comparison of the effect of tenofovir versus abacavir.
For low-energy fractures, incidence rates were calculated for 1000 person-years at risk (PYR) with 95% confidence intervals (CIs) for three time periods (1995–1996, 1997–2003, and 2004–2009).
SPSS statistical software (Norusis; SPSS Inc., Chicago, Illinois, USA) and R software, version 2.8.1, was used for data analysis.
Approvals and permissions
The Danish Data Protection Agency approved the establishment of the cohort study and the record linkage with CRS and DNHR.
Characteristics of the study population
The study population included 5306 HIV-infected patients and 26 530 individuals in the general population control cohort (Table 1). Patients and population controls were well matched in terms of age at index date and sex, and furthermore, they were equally distributed in terms of emigration and loss to follow-up. Median age was 37 years and the male-to-female ratio was 3 : 1. Almost two-thirds of the HIV-infected patients (62%) had received a diagnosis of HIV infection after 1 January 1995, and 78% of the HIV-infected patients started HAART during the study period. Coinfection with HCV was observed in 851 (16%) of the HIV-infected patients.
Total number of fractures
We observed 806 fractures in the HIV-infected cohort during 38 456 PYR [incidence rate of 21.0/1000 PYR (95% CI 19.8–22.2)], and 3312 fractures in the general population control cohort during 245 315 PYR [incidence rate of 13.5/1000 PYR (95% CI 13.1–13.9)]. HIV-infected patients had increased fracture risk compared with population controls (IRR of 1.5). Fracture risk was increased in both HAART-naive (IRR of 1.4) and HAART-exposed patients (IRR of 1.6). Compared with corresponding population controls both HIV-monoinfected patients (IRR of 1.3) and HIV/HCV-coinfected patients (IRR of 2.9) had increased risk of fracture. Table 2 displays IRRs with 95% CI.
For HIV-monoinfected patients the incidence rate of low-energy fracture was 7.4/1000 PYR (95% CI 6.7–8.2), for HIV/HCV-coinfected patients the incidence rate was 17.7/1000 PYR (15.3–20.5), and for the population controls cohort the incidence rate was 4.8/1000 PYR (4.6–5.0). Compared with the population controls both HIV-monoinfected patients (IRR of 1.6) and HIV/HCV-coinfected patients (IRR of 3.8) had increased risk of low-energy fracture. Figure 1 displays cumulative incidence of low-energy fractures for the three groups.
All subgroups of low-energy fractures contributed to the increased risk of low-energy fracture observed in HIV-monoinfected patients. For HIV/HCV-coinfected patients the relative risk of hip fractures was substantially higher than risk of other low-energy fractures (IRR of 16.0), whereas vertebral fractures seemed under-represented (Table 2).
For HIV-infected patients the incidence rate of high-energy fracture was 9.5/1000 PYR (95% CI 8.7–9.5), for HIV/HCV-monoinfected patients the incidence rate was 22.7/1000 PYR (19.9–25.9), and for population controls the incidence rate was 8.7/1000 PYR (8.4–9.0). Thus HIV-monoinfected patients and population controls had comparable risks of high-energy fractures, whereas HIV/HCV-coinfected patients had increased risk of high-energy fractures compared with population controls (IRR of 2.4) (Fig. 1).
Low-energy fractures in HAART-naive versus HAART-treated HIV-monoinfected patients
Figure 2 displays cumulative incidence of low-energy fractures stratified by time before and after HAART exposure in HIV-monoinfected patients and corresponding population controls. As illustrated, there was no significant difference in fracture risk between HAART-naïve patients and population controls. In contrast, HAART-exposed patients had increased risk of low-energy fracture [IRR of 1.8 (95% CI 1.5–2.1)] compared with population controls. After adjusting for comorbidity the IRR was 1.6 (95% CI 1.4–1.9). A plot of Schoenfeld residuals proved that the risk of low-energy fracture was constant over time after HAART initiation (data not shown).
Risk factors for low-energy fractures in HAART-exposed HIV-monoinfected patients
For HIV-monoinfected patients, we estimated the effect of age, sex, an AIDS-defining diagnosis prior to HAART, CD4 cell count and race in univariate and multivariate analyses restricted to time after HAART initiation (Table 3). We found no significant association between CD4 cell count, prior AIDS diagnosis or sex and risk of low-energy fracture. Caucasian race, increasing age and medium or high comorbidity score was associated with increased fracture risk in both univariate and multivariate analyses. We had only information on smoking in 69% of the patients (of whom 67% were current or former smokers) and therefore evaluated the effect of smoking in a separate model. Smoking was associated with increased fracture risk, unadjusted IRR of 2.0 and adjusted IRR of 2.0.
Use of tenofovir did not increase the risk of fracture [IRR of 1.2 (95% CI 0.8–1.7)], neither did the use of efavirenz [IRR of 1.1 (0.8–1.4)]. The IRR for use of abacavir was 0.9 (0.7–1.2).
Temporal trends in incidence rates of low-energy fractures
For HIV-monoinfected patients the incidence rate increased from 3.5/1000 PYR (95% CI 2.2–5.7) for the pre-HAART period (1996–1997) to 8.3/1000 PYR (95% CI 7.2–9.7) for the period 1997–2003 and 7.5/1000 PYR (95% CI 6.5–8.6) for the period 2004–2009.
For HIV/HCV-infected patients the incidence rate remained stable during the three periods: 16.7/1000 PYR (11.1–25.3), 17.9/1000 PYR (14.4–22.2) and 17.8/1000 PYR (14.1–22.4), respectively.
For the population control cohort the incidence rates were 6.2/1000 PYR (5.4–7.1) for 1995–1996, 4.6/1000 PYR (4.3–4.9) for 1997–2003, and 4.7/1000 PYR (4.4–5.0) for 2004–2009.
In this nationwide population-based cohort study we found that HIV-monoinfected patients had increased risk of low-energy fractures driven by the period after HAART exposure compared with population controls. In contrast, we observed similar risks of high-energy fractures in HIV-monoinfected patients and controls.
The risk of fractures is determined by a combination of bone strength and a relevant trauma . The fracture risk in HIV/HCV-coinfected patients with considerably increased risk of both low-energy and high-energy fractures was substantially different from that observed in HIV-monoinfected patients We have previously shown that HIV/HCV-coinfected patients had a poorer prognosis irrespective of recorded HIV transmission group , and further that HCV coinfection was a very sensitive marker of past or ongoing intravenous drug use and of lifestyle-related risk factors [25,30]. Thus the markedly increased fracture risk in HIV/HCV-coinfected patients may be associated with consequences of intravenous drug use or increased alcohol use including increased fall or trauma risk . To reduce confounding by lifestyle-related factors and more precisely explore the influence of HIV infection and HAART exposure on risk of fracture we therefore performed analyses restricted to HIV-monoinfected patients.
HIV-monoinfected individuals had increased risk of low-energy fractures compared with population controls, and importantly this risk was only increased in HAART-exposed patients. Our study is the first to highlight an association between HAART exposure and subsequent fracture risk. Only a few previous studies have analysed data on HAART initiation and subsequent fracture risk; Yin et al.  found no association with cumulative use of HAART and fracture risk; Young et al.  found that HAART exposure was associated with a hazard rate of 1.39 for fracture, but their result was not statistically significant.
As in all observational studies we can identify associations but cannot attribute causality. We found an association between HAART exposure and increased risk of low-energy fracture but cannot determine whether this increased risk is induced by the alterations in BMD observed after HAART initiation or by differences between HAART-treated and HAART-naive patients. Use of HAART may be a marker of duration and/or severity of HIV infection, and the two groups may therefore have different immunological profiles and response to HIV infection. Previous studies have shown association between low CD4 cell count and increased BMD loss after start of HAART [3,7]. Whereas one study  found an association between low CD4 cell count and fracture risk, others did not [13,16]. In our study, neither pre-HAART CD4 cell count nor pre-HAART AIDS-defining diagnosis was significantly associated with subsequent fracture risk.
Adjusting for comorbidity attenuated the difference in fracture risk between HIV-infected patients and controls. We were not able to adjust for other traditional risk factors for osteoporosis such as low BMI, smoking and steroid exposure that may also be more common in HIV-infected patients than in general population controls. Among HAART-exposed patients smoking was associated with a two-fold increased risk of low-energy fracture. Thus, increased prevalence of smoking among HIV-infected patients  may explain part of the excess risk. Concerns have been raised about the long-term effect of tenofovir on bone strengths. However, similar to three other studies we found no association between use of tenofovir and fracture risk [4,13,16]. As in most other studies, Caucasian race and increasing age were associated with increased fracture risk.
The strengths of our study include the use of population-based nationwide cohorts; long and nearly complete follow-up; access to detailed data on fracture and comorbidity in DNHR; and for the HIV patients clinical information on use of antiretroviral treatment, AIDS-defining events, CD4 cell count and viral load. Our study also had limitations. Fracture data were incomplete before 1995 as the DNHR only included data on outpatients and emergency patients since 1995. Therefore we could not exclude persons with low-energy fracture before the index date.
We used a wider definition of low-energy fractures than the major osteoporotic fractures at spine, forearm, hip and shoulder , and therefore we also present relative risks for each type of low-energy fracture. Another potential shortcoming is inaccuracies in fracture diagnoses reported to DNHR. However, the positive predictive value of the registry diagnoses is generally high (70–99%) , and the positive predictive value of hip fractures in DNHR has been shown to be as high as 93% .
In conclusion, we found that HIV-infected patients had increased risk of fracture at any site. HIV-infected patients without HCV-coinfection had increased risk of low-energy fractures but not of high-energy fractures. The risk of low-energy fracture was associated with HAART exposure but also with traditional risk factors such as age, comorbidity and smoking. The risk associated with HAART was moderate and did not increase over time. Future research should explore mechanisms behind the HAART-associated initial BMD loss and investigate possible interventions in patients at high risk of osteoporosis.
The authors wish to thank Lars Haukali Omland for assistance with computing Charlson Comorbidity Index and cumulative incidences.
We thank the NOVO Nordic Foundation, and the Clinical Institute of Copenhagen University for financial support.
A.H., J.G., and N.O. designed the study and analysed and interpreted the data; A.H. wrote the manuscript; J.G., G.K., C.L., C.P., G.P. and N.O. revised the manuscript critically; J.G., G.K., C.L., C.P., G.P. and N.O. collected the data.
Conflicts of interest
Potential conflicts of interest: A.-B.E.H. has received travel grants from Bristol-Myers Squibb and Gilead.
J.G. has participated in advisory boards or has received grants or support from Abbott, Merck Sharp & Dohme, ViiV Healthcare, Gilead and Janssen.
C.P. has received honoraria for educational activities from Abbott and Merck Sharp & Dohme.
N.O. has received research funding from Roche, Bristol-Myers Squibb, Merck Sharp & Dohme, GlaxoSmithKline, Abbott, Boehringer Ingelheim, Janssen-Cilag and Swedish Orphan Drugs.
All other authors: no conflicts of interest.
1. Brown TT, Qaqish RB. Antiretroviral therapy and the prevalence of osteopenia and osteoporosis: a meta-analytic review 1
2. Gallant JE, Staszewski S, Pozniak AL, DeJesus E, Suleiman JM, Miller MD, et al. Efficacy and safety of tenofovir DF vs. stavudine in combination therapy in antiretroviral-naive patients: a 3-year randomized trial
3. 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
. HIV Med
4. McComsey GA, Kitch D, Daar ES, Tierney C, Jahed NC, Tebas P, et al. 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 A5224 s, a substudy of ACTG A5202
. J Infect Dis
5. Bolland MJ, Grey AB, Horne AM, Briggs SE, Thomas MG, Ellis-Pegler RB, et al. Bone mineral density remains stable in HAART-treated HIV-infected men over 2 years 1
. Clin Endocrinol (Oxf)
6. Nolan D, Upton R, McKinnon E, John M, James I, Adler B, et al. Stable or increasing bone mineral density in HIV-infected patients treated with nelfinavir or indinavir
7. Brown TT, McComsey GA, King MS, Qaqish RB, Bernstein BM, da Silva BA. Loss of bone mineral density after antiretroviral therapy initiation: independent of antiretroviral regimen. J Acquir Immune Defic Syndr
8. Duvivier C, Kolta S, Assoumou L, Ghosn J, Rozenberg S, Murphy RL, et al. Greater decrease in bone mineral density with protease inhibitor regimens compared with nonnucleoside reverse transcriptase inhibitor regimens in HIV-1 infected naive patients
9. van Vonderen MG, Lips P, van Agtmael MA, Hassink EA, Brinkman K, Geerlings SE, et al. First line zidovudine/lamivudine/lopinavir/ritonavir leads to greater bone loss compared to nevirapine/lopinavir/ritonavir
10. Grijsen ML, Vrouenraets SM, Steingrover R, Lips P, Reiss P, Wit FW, et al. High prevalence of reduced bone mineral density in primary HIV-1-infected men
11. Mulligan K, Glidden D, Gonzales P, Ramirez-Cardich M-E, Liu A, Namwongprom S, et al. Effect of FTC/TDF on bone mineral density in seronegative men from 4 continents: DEXA results from the iPrEx Study
[Abstract]. 18th Conference on Retroviruses and Opportunistic Infections Boston
, MA, USA, February 2011.
12. Triant VA, Brown TT, Lee H, Grinspoon SK. Fracture prevalence among human immunodeficiency virus (HIV)-infected versus non-HIV-infected patients in a large U.S. healthcare system
. J Clin Endocrinol Metab
13. Womack JA, Goulet JL, Gibert C, Brandt C, Chang CC, Gulanski B, et al. Increased risk of fragility fractures among HIV infected compared to uninfected male veterans
. PLoS One
14. Young B, Dao CN, Buchacz K, Baker R, Brooks JT. Increased rates of bone fracture among HIV-infected persons in the HIV Outpatient Study (HOPS) compared with the US general population, 2000–2006
. Clin Infect Dis
15. Arnsten JH, Freeman R, Howard AA, Floris-Moore M, Lo Y, Klein RS. Decreased bone mineral density and increased fracture risk in aging men with or at risk for HIV infection
16. Yin MT, Shi Q, Hoover DR, Anastos K, Sharma A, Young M, et al. Fracture incidence in HIV-infected women: results from the Women's Interagency HIV Study
17. Bedimo R, Zhang S, Drechsler H, Tebas R, Maalouf N. Osteoporotic fracture risk associated with cumulative exposure to tenofovir and other antiretroviral agent
s [Abstract]. 6th IAS Conference on HIV Pathogenesis, Treatment and Prevention
, 2011 Rome Italy 2011.
18. Obel N, Engsig FN, Rasmussen LD, Larsen MV, Omland LH, Sorensen HT. Cohort profile: the Danish HIV cohort study
. Int J Epidemiol
19. Lohse N, Hansen A, Jensen-Fangel S, Kronborg G, Kvinesdal B, Larsen CS, et al. Demographics of HIV-1 infection in Denmark: results from The Danish HIV Cohort Study
. Scand J Infect Dis
20. Frank L. Epidemiology. When an entire country is a cohort
21. Andersen TF, Madsen M, Jørgensen J, Mellemkjær L, Olsen JH. The Danish Hospital Register: a valuable source of data for modern health sciences
. Dan Med Bull
22. Lippuner K, Golder M, Greiner R. Epidemiology and direct medical costs of osteoporotic fractures in men and women in Switzerland
. Osteoporos Int
2005; 16 (Suppl 2):S8–S17.
23. Charlson ME, Pompei P, Ales KL, MacKenzie CR. A new method of classifying prognostic comorbidity in longitudinal studies: development and validation 3
. J Chronic Dis
24. Quan H, Sundararajan V, Halfon P, Fong A, Burnand B, Luthi JC, et al. Coding algorithms for defining comorbidities in ICD-9-CM and ICD-10 administrative data
. Med Care
25. Hansen AB, Gerstoft J, Kronborg G, Pedersen C, Sorensen HT, Obel N. Mortality in siblings of patients coinfected with HIV and hepatitis C virus 1
. J Infect Dis
26. Satagopan JM, Ben-Porat L, Berwick M, Robson M, Kutler D, Auerbach AD. A note on competing risks in survival data analysis
. Br J Cancer
27. Brown TT, McComsey GA. Association between initiation of antiretroviral therapy with efavirenz and decreases in 25-hydroxyvitamin D
. Antivir Ther
28. Welz T, Childs K, Ibrahim F, Poulton M, Taylor CB, Moniz CF, et al. Efavirenz is associated with severe vitamin D deficiency and increased alkaline phosphatase
29. Weis N, Lindhardt BO, Kronborg G, Hansen AB, Laursen AL, Christensen PB, et al. Impact of hepatitis C virus coinfection on response to highly active antiretroviral therapy and outcome in HIV-infected individuals: a nationwide cohort study
. Clin Infect Dis
30. Hansen AB, Lohse N, Gerstoft J, Kronborg G, Laursen A, Pedersen C, et al. Cause-specific excess mortality in siblings of patients co-infected with HIV and hepatitis C virus
. PLoS One
31. Saves M, Chene G, Ducimetiere P, Leport C, Le Moal G, Amouyel P, et al. Risk factors for coronary heart disease in patients treated for human immunodeficiency virus infection compared with the general population
. Clin Infect Dis
32. Rachner TD, Khosla S, Hofbauer LC. Osteoporosis: now and the future
33. Nickelsen TN. Data validity and coverage in the Danish National Health Registry. A literature review
. Ugeskr Laeger
34. Hoidrup S, Gronbaek M, Gottschau A, Lauritzen JB, Schroll M. Alcohol intake, beverage preference, and risk of hip fracture in men and women. Copenhagen Centre for Prospective Population Studies
. Am J Epidemiol