High frequency of vitamin D deficiency in HIV-infected patients: effects of HIV-related factors and antiretroviral drugs

Journal of Antimicrobial Chemotherapy Advance Access published May 15, 2012 J Antimicrob Chemother doi:10.1093/jac/dks176 High frequency of vitamin D deficiency in HIV-infected patients: effects of HIV-related factors and antiretroviral drugs C. Allavena1*, C. Delpierre2, L. Cuzin3, D. Rey4, N. Viget5, J. Bernard6, P. Guillot7, C. Duvivier8,9, E. Billaud1 and F. Raffi1 1 Maladies Infectieuses et Tropicales, CHU Nantes, Nantes, France; 2INSERM U558, F-31073, Toulouse, France; 3COREVIH, University Hospital, Toulouse, France; 4HIV/AIDS Center, University Hospital, Strasbourg, France; 5Infectious Diseases Department, University Hospital, Tourcoing, France; 6Rhumatologie, CHU Toulouse, Toulouse, France; 7Rhumatologie, CHU Nantes, Nantes, France; 8 AP-HP, Hôpital Necker-Enfants Malades, Service de Maladies Infectieuses et Tropicales, Centre d’Infectiologie Necker-Pasteur, Paris, F-75015 France; 9Institut Pasteur, Centre Médical de l’Institut Pasteur; Centre d’Infectiologie Necker-Pasteur, Paris, F-75015 France Downloaded from http://jac.oxfordjournals.org/ by guest on May 16, 2012 *Corresponding author. Tel: +33-2-40-08-31-12; Fax: +33-2-40-08-31-81; E-mail: clotilde.allavena@chu-nantes.fr Received 13 January 2012; returned 29 February 2012; revised 27 March 2012; accepted 13 April 2012 Objectives: The aim of this study was to assess 25-hydroxyvitamin D (vitamin D) status in an HIV-infected adult population and to define HIV- and antiretroviral-related factors associated with vitamin D deficiency. Methods: Using data from a prospective cohort of HIV-infected adult patients followed in five French centres (Dat’AIDS cohort), we evaluated the prevalence of vitamin D deficiency/insufficiency (,30 ng/mL). A multiple linear regression model was used to examine risk factors for vitamin D deficiency (≤10 ng/mL). Results: Vitamin D deficiency/insufficiency was observed in 86.7% of the 2994 patients, including 55.6% with vitamin D insufficiency and 31.1% with vitamin D deficiency. In multivariate analysis, factors associated with vitamin D deficiency were current smoking [adjusted OR (aOR) 1.55], estimated glomerular filtration rate ≥90 mL/min/1.73 m2 (aOR 1.51), vitamin D measurement not performed in summer (aOR 0.27), CD4 ,350 cells/mm3 (aOR 1.37 for CD4 200 to ,350 and 1.62 for CD4 ,200 cells/mm3) and antiretroviral therapy (aOR 2.61). Gender, body mass index, age, coinfection and previous AIDS were not associated factors. In the antiretroviral-treated population (n¼ 2660), besides the same factors found in the whole population, efavirenz was the only drug to be significantly associated with deficiency, with an aOR of 1.89 (95% CI 1.45–2.47). Conclusions: Vitamin D deficiency is frequent in this HIV-infected population. Patients on antiretroviral therapy are at higher risk of vitamin D deficiency than antiretroviral-naive patients, with an increased risk in patients receiving efavirenz. No effect of the other antiretrovirals, including the latest (etravirine, darunavir, raltegravir), was found. Keywords: HIV infection, efavirenz, protease inhibitors, raltegravir Introduction Epidemiological data indicate that a low level of vitamin D is associated with an increased risk of musculoskeletal complaints, including osteomalacia and osteoporosis.1 It may also contribute to a number of pathological processes, such as immune dysfunction, inflammation, autoimmune diseases, cardiovascular diseases and malignancies.2 – 4 These effects could explain in part why, in the general population, morbidity and mortality are significantly more frequent in subjects with vitamin D deficiency.5,6 Most of the morbidities associated with a low level of vitamin D have been reported to have a high prevalence in the HIV-infected population, in particular regarding bone mineral metabolism. As some studies have reported an incidence of osteopenia and osteoporosis greater than in the general population matched for age and sex, the evaluation of bone metabolism alteration is an important issue in the HIV population.7 – 10 Some antiretroviral drugs, such as tenofovir and some protease inhibitors, are known to have an impact on bone metabolism.11 – 13 Studies focusing on vitamin D are therefore of interest because they might provide a link to the high prevalence of abnormalities of bone # The Author 2012. Published by Oxford University Press on behalf of the British Society for Antimicrobial Chemotherapy. All rights reserved. For Permissions, please e-mail: journals.permissions@oup.com 1 of 9 Allavena et al. Methods Study design Information was collected from five large HIV reference centres in France participating in the Dat’AIDS cohort (Nantes, 47812′ N 01833′ W; Strasbourg, 47840′ N 07820′ E; Lille, 50838′ N 03803′ E; Paris Pasteur, 48850′ N 2820′ E; and Toulouse, 43837′ N 1827′ E). These centres maintain prospective cohorts of all HIV-1-infected patients who provided written consent. The cohorts are implemented via a common electronic medical record (EMR).19 The patients enter the cohort when they seek care in one of the centres, regardless of their HIV disease history, and all previous clinical events as well as therapeutic history are collected with appropriate dates. The EMR collects demographic details, clinical events, antiretroviral history, viral load and CD4 cell count data at regular 3 – 6 month intervals during routine clinical assessment. This system allows the use of the databases with minimal delay, apart from automatic and manual quality controls performed before any analysis. For the purpose of this cross-sectional study, we included centres in which 25% of the patients had routine 25-hydroxyvitamin D (vitamin D) levels measured between January 2007 and June 2010. Patients To evaluate the impact of antiretroviral drugs, adult patients were included if they were either antiretroviral naive or were receiving a stable antiretroviral regimen for .3 months. Patients receiving vitamin D supplementation, treatment for osteoporosis, on renal dialysis or with creatininaemia .400 mmol/L at time of the vitamin D measurement were excluded. Data collection Patient demographics (age, gender), clinical and biological variables [body mass index (BMI), creatininaemia, phosphataemia, calcaemia and parathyroid hormone (PTH)] and HIV-related parameters, including route of transmission, CDC stage, time since HIV diagnosis, CD4 cell count nadir (,200, 200 to ,350 and ≥350 cells/mm3), CD4 cell count at the time of vitamin D measurement (,200, 200 to ,350 and ≥350 cells/mm3), HCV or HBV coinfection, antiretroviral naive or experienced status, history of treatment and duration of antiretroviral therapy, were extracted from the database at the time of the vitamin D measurement. 2 of 9 The estimated glomerular filtration rate (eGFR) was obtained using the Modification of Diet in Renal Disease (MDRD) formula. An eGFR cut-off of 90 mL/min/1.73 m2 was chosen in our study with respect to the definition of stage 2 chronic kidney disease.20 Measurement of vitamin D and definitions Serum vitamin D, which is the best indicator of vitamin D status, was measured using radioimmunoassay (DiaSorin, normal range 30– 100 ng/mL) in four centres and electrochemoluminescence (Cobas Roche, normal range 30–80 ng/mL) in one centre (10% of the subjects). A validated quality/control process was performed in all the centres. Low-level vitamin D concentration was defined as ,30 ng/mL. Vitamin D insufficiency and deficiency were defined as vitamin D between .10 and ,30 ng/mL, and ≤10 ng/mL, respectively.21 Statistical analysis Characteristics of the population and the prevalence of vitamin D deficiency are presented for the whole population and separately in antiretroviral-naive and -treated patients. Variables were described using the median and IQR for quantitative variables and proportions for qualitative variables. Then we performed a logistic regression model to identify variables associated with vitamin D deficiency (≤10 versus .10 ng/mL). Variables with a P value ,0.20 in univariate analysis were included in the logistic regression model, except for age, which was forced in the model. Treatment was considered in three steps in separate models: first as a binary variable (treated or naive); then, in case of significance, as a combination of drugs [two nucleoside reverse transcriptase inhibitors (NRTIs)+one boosted PI, two NRTIs+one non-nucleoside reverse transcriptase inhibitor (NNRTI) or other] and drug by drug among only antiretroviral-treated patients. The x2 test was used to test categorical variables and the Kruskal– Wallis test was used for continuous variables. Associations were assessed using adjusted ORs (aORs) and corresponding 95% CIs for logistic regression. The season of vitamin D measurement was adjusted for logistic regression in the multivariate models. For the subsample with a measure of PTH at the same time as the measure of vitamin D, the median and IQR of PTH were estimated according to the level of vitamin D, eGFR and treatment. Statistical analyses were performed using SAS version 9.1 (SAS Institute, Cary, NC, USA). Results Five centres participated in the study, contributing a total of 7065 eligible participants. Of these, 3570 patients had at least one measurement of vitamin D level determined between January 2007 and June 2010. Among these, 508 had received antiretroviral treatment for ,3 months and were therefore excluded. Respectively, 60 patients had a history of osteoporosis and 8 a history of severe renal impairment, and were also excluded. The study thus included 2994 patients, 334 being antiretroviral naive and 2660 being treated with a stable regimen for .3 months. Patients had a median age of 45 years, were predominantly male (72.3%), with a median nadir CD4 count of 203 cells/mm3 (IQR 101 –303) and median time since HIV diagnosis of 11.6 years (IQR 5.5 –17.6), and 18% were HBV or HCV coinfected. No vitamin D gradient was found between sites located at higher and lower latitudes, at 508 N and 438 N, respectively. Baseline patient characteristics at the time of vitamin D measurement are shown in Table 1. These patients Downloaded from http://jac.oxfordjournals.org/ by guest on May 16, 2012 mineral metabolism in the HIV-infected population and objectivize the effect that antiretroviral drugs may have on vitamin D levels. Recently, results from the EuroSIDA cohort and the ICONA study demonstrated that vitamin D deficiency was independently associated with a higher risk of death and AIDS events.14,15 The SUN study showed that vitamin D insufficiency and deficiency were highly prevalent in an HIV-infected adult population, although less prevalent than in the general population, and were associated with exposures to certain antiretroviral drugs, such as efavirenz and ritonavir.16 Studies on the relationship between vitamin D status, HIV infection and the impact of antiretroviral therapies have been performed on cohorts of HIV-infected patients, but mostly in heterogeneous populations or using a limited sample size.15,17,18 In fact, the effects of individual protease inhibitors (PIs) on vitamin D level, in particular vitamin D deficiency, are still unclear. This study assessed vitamin D status and aimed to define HIV- and antiretroviral-related factors associated with vitamin D deficiency in a large French HIV-infected adult population. JAC Vitamin D deficiency and antiretroviral drugs Table 1. Baseline characteristics of the population at the time of measurement of vitamin D before any supplementation Antiretroviral-naive group (n¼334) Antiretroviral-treated group (n¼2660) P Total (n¼2994) Age (years), median (IQR) Gender, male, n (%) BMI, median (IQR) 38.1 (31.4–46.5) 242 (72.5) 23.0 (20.8–25.9) 45.8 (40.3– 52.2) 1924 (72.3) 23.4 (21.2– 25.9) HIV transmission risk category homosexual heterosexual intravenous drug user others unknown 167 (50.0) 133 (39.8) 8 (2.4) 6 (1.8) 20 (6.0) 1175 (44.2) 1018 (38.3) 190 (7.1) 126 (4.7) 151 (5.7) 20 (6.0) 7 (2.1) 150 (44.9)/138 (41.3) 43.7/24.0/11.7/20.7 524 (19.7) 615 (23.1) 1012 (38.1)/1284 (48.3) 47.0/17.3/13.6/22.1 ,0.0001* ,0.0001* 0.04* 0.03* 544 (18.2) 622 (20.8) 1162 (38.8)/1422 (47.5) 46.6/18.1/13.4/21.99 1.7 (0.4– 5.1) 13.0 (6.7 –18.1) ,0.0001** 11.6 (5.5– 17.6) 402 (301– 525) 492 (368– 681) 186 (89– 274) 569 (412–753) ,0.0001** ,0.0001** 203 (101– 303) 561 (406– 747) Nadir CD4, cells/mm3, n (%) ,200 200 to ,350 ≥350 29 (8.7) 97 (29.0) 208 (62.3) 1434 (53.9) 918 (34.5) 308 (11.6) ,0.0001** 1463 (48.9) 1015 (33.9) 516 (17.2) CD4, cells/mm3, n (%) ,200 200 to ,350 ≥350 19 (5.9) 52 (16.0) 253 (78.1) 130 (5.0) 297 (11.5) 2157 (83.5) 0.04** 149 (5.1) 349 (12.0) 2410 (82.9) eGFR, mL/min/1.73 m2, n (%) ,90 ≥90 90 (28.6) 225 (71.4) 1178 (46.4) 1360 (53.6) ,0.0001 1268 (44.4) 1585 (55.6) 0.001 45.1 (39.2–51.7) 2166 (72.3) 23.3 (21.1–25.9) 1342 (44.8) 1151 (38.5) 198 (6.6) 132 (4.4) 171 (5.7) *x2 test. **Kruskal–Wallis test. were comparable to patients included in the French cohort in terms of age, gender, HIV transmission risk category, time since HIV diagnosis and nadir CD4 cell count.19 A low level of vitamin D was observed in 86.7% of the whole population, including 55.6% with vitamin D insufficiency and 31.1% with vitamin D deficiency. Serum vitamin D, calcium, phosphate and PTH levels are shown in Table 2. Median values of calcaemia, phosphataemia and serum PTH were in the normal range. In univariate analysis, factors significantly associated with vitamin D deficiency were current smoking, eGFR ≥90 mL/min/ 1.73 m2, measurement not performed in summer, nadir CD4 ,200 cells/mm3, current CD4 ,350 cells/mm3, prior AIDS and ongoing antiretroviral therapy. Age, time since HIV diagnosis, gender, BMI and hepatitis coinfection were not associated with vitamin D deficiency (Table 3). In multivariate analysis, factors associated with vitamin D deficiency were current smoking [aOR 1.55 (1.26 –1.90)], current CD4 ,350 cells/mm3 [aOR 1.37 (1.04 –1.82) for CD4 between 200 and ,350 and aOR 1.62 (1.07 –2.45) for CD4 ,200 cells/mm3], eGFR ≥90 mL/min/1.73 m2 [aOR 1.51 (1.25 – 1.83)] and antiretroviral therapy [aOR 2.61 (1.79 –3.81)] (Table 3). Compared with a vitamin D measurement performed in autumn, measurement performed in summer was associated with a lower probability of vitamin D deficiency [aOR 0.27 (0.18 – 0.40)] and measurement in winter with a higher probability of vitamin D deficiency [aOR 1.95 (1.54 –2.47)]. Prior AIDS was not associated with vitamin D deficiency (Table 3). As antiretroviral treatment was associated with deficiency, analyses were then conducted in treated patients. The 2660 antiretroviral-treated subjects had been receiving treatment for a median of 9.6 years (IQR 4.5–13.1) and had been on their current regimen for a median of 18 months (IQR 9 –35); 34.6% of the patients received a combination with two NRTIs + one NNRTI and 40.2% a combination with two NRTIs+one boosted PI. 3 of 9 Downloaded from http://jac.oxfordjournals.org/ by guest on May 16, 2012 Coinfection, n (%) CDC stage C, n (%) Smokers/non- or ex-smokers, n (%) Measurement in winter/spring/ summer/autumn, % Time since HIV diagnosis, years, median (IQR) Nadir CD4, cells/mm3, median (IQR) CD4, cells/mm3, median (IQR) ,0.0001** 0.94* 0.24** Allavena et al. Table 2. Serum vitamin D, calcium, phosphate and PTH levels Antiretroviral-naive group (n¼334) Antiretroviral-experienced group (n¼2660) ≥30 ng/mL, n (%) .10 to ,30 ng/mL, 59 (17.7%) 206 (61.7%) 339 (12.8%) 1458 (54.8%) ≤10 ng/mL, n (%) level, ng/mL, median 69 (20.6%) 17.7 (11.1– 27.0) 863 (32.4%) 14.7 (8.9– 23.0) Phosphate, mmol/L, median (IQR) Antiretroviral-naive group (n ¼276) 0.99 (0.86– 1.12) Vitamin D Vitamin D n (%) Vitamin D Vitamin D (IQR) P ,0.0001* Total (n¼2994) 398 (13.3%) 1664 (55.6%) ,0.0001** 932 (31.1%) 15.0 (9.0– 23.2) Antiretroviral-experienced group (n¼ 2225) 0.94 (0.82–1.08) P Total (n¼2501) Antiretroviral-experienced group (n¼ 2202) 2.38 (2.30–2.45) P Calcium, mmol/L, median (IQR) Antiretroviral-naive group (n ¼274) 2.36 (2.29– 2.45) Antiretroviral-experienced group (n¼ 971) 40.5 (31.0–56.0) P Total (n¼1067) PTH, ng/L, median (IQR) Antiretroviral-naive group (n ¼96) 29.5 (23.5– 38.9) ,0.0001 39.5 (30.0–54.2) 0.0004** 0.95 (0.83–1.09) Total (n¼2476) 2.38 (2.30–2.45) *x2 test. **Kruskal–Wallis test. In univariate analysis, besides the same factors found in the whole study population, a higher median duration of the current treatment, being treated with two NRTIs+ one NNRTI and receiving a regimen including zidovudine, lamivudine, emtricitabine, abacavir, efavirenz or ritonavir were statistically associated with vitamin D deficiency. A regimen including tenofovir or darunavir was associated with a lower risk of vitamin D deficiency (Table 4). In multivariate analysis, vitamin D deficiency was significantly associated with an NNRTI-based regimen compared with a boosted PI-based regimen [aOR 1.53 (1.22 –1.92)]. The only antiretroviral drug associated with a vitamin D deficiency was efavirenz [aOR 1.89 (1.45 –2.47)] (Table 4). No association was found between vitamin D deficiency and the other available NNRTIs (nevirapine and etravirine), any nucleoside analogue, including zidovudine and tenofovir, any boosted protease inhibitor or raltegravir. By considering the duration of exposure to antiretroviral drugs, we observed the same associations as in our model using antiretroviral drugs as a binary variable (data not shown), the duration of exposure to efavirenz being the only drug exposure associated with an increased risk of deficiency [aOR 1.02 (1.01 –1.03)]. Among the 2994 patients, 1067 had a measure of PTH at the time of vitamin D measurement. These patients were comparable to patients without PTH measurement except for sites, hepatitis coinfection, antiretroviral status and season of vitamin D measurement. Median PTH level was 39.5 ng/L (IQR 30 –54.2). PTH levels were significantly associated with vitamin D: median PTH was 34.5 ng/L (IQR 26.0–45.0) among patients with vitamin D ≥30 ng/mL, 39 ng/L (IQR 30.0 –54.0) among patients with vitamin D insufficiency and 43 ng/L (IQR 31.6 – 58.0) among patients with vitamin D deficiency (P,0.0001). Median PTH was higher in patients with eGFR ,90 mL/min/ 1.73 m2 than in patients with an eGFR ≥90 mL/min/1.73 m2 4 of 9 [41.7 ng/mL (IQR 31 –58) and 38 ng/mL (IQR 29 –51), respectively (Kruskal–Wallis test P¼ 0.0003)]. Median PTH was significantly higher in antiretroviral-naive patients than in antiretroviraltreated patients, even after adjusting for vitamin D concentration (Table 2). Among antiretroviral-treated patients with a PTH measurement (n¼ 971), no effect of drug classes (2 NRTIs + 1 boosted PI versus 2 NRTIs + 1 NNRTI) and no effect of tenofovir on PTH concentration were found in our study (data not shown). Discussion The present study documents the high prevalence of vitamin D insufficiency and deficiency in a French cohort of HIV-infected adult subjects. Overall, 86% of the study population had a low level of vitamin D and 31% a vitamin D deficiency. This result is concordant with the Swiss cohort, where vitamin D deficiency occurred in 14%–52% of 211 HIV-infected subjects according to the season of the measurement.22 It seems more frequent than in an Italian HIV-infected population (ICONA study), in which vitamin D deficiency was documented from 6% to 9% according to the season, and in the general population, in which it has been evaluated as 14% in France and ,10% in the USA.14,16,23,24 However, comparison with the general population would need to be re-evaluated in a new age- and gender-matched study since recent data from the NHANES studies have shown deterioration of vitamin D status over the last decade in the general population.25 Consistent with earlier reports in the general population, smoking and measurement in winter were strongly related to low levels of vitamin D. Exposure to antiretroviral therapy was also associated with vitamin D deficiency. However, in the multivariate analysis, BMI, time since HIV diagnosis and previous AIDS did not affect the level of vitamin D; Downloaded from http://jac.oxfordjournals.org/ by guest on May 16, 2012 0.05** JAC Vitamin D deficiency and antiretroviral drugs Table 3. Factors associated with vitamin D deficiency in the total population Univariate analysis (n¼2994) vitamin D deficiency Multivariate analysis (n¼2816), aOR (95% CI) no (n ¼2062) yes (n ¼932) P 45 (39–52) 11.5 (5.3 –17.5) 23.4 (21.2–25.9) 45 (40– 52) 11.8 (5.7–17.8) 23.2 (21.0–26.0) 0.42** 0.14** 0.42** 1502 (69.3) 560 (67.6) 664 (30.7) 268 (32.4) 0.37* 365 (70.7) 737 (72.6) 960 (65.6) 151 (29.3) 278 (27.4) 503 (34.4) 0.0006* reference 0.70 (0.52– 0.95) 0.91 (0.67– 1.23) CD4 cell count, cells/mm3, n (%) ≥350 200 to ,350 ,200 1689 (70.1) 219 (62.8) 90 (60.4) 721 (29.9) 130 (37.3) 59 (39.6) 0.002* reference 1.37 (1.04– 1.82) 1.62 (1.07– 2.45) Previous AIDS, n (%) no yes 1657 (69.9) 405 (65.1) 715 (30.1) 217 (34.9) 0.02* reference 1.16 (0.92– 1.46) Coinfection with HCV and/or HBV, n (%) no yes 1700 (69.4) 362 (66.5) 750 (30.6) 182 (33.5) 0.19* reference 0.98 (0.77– 1.25) Tobacco smoking status, n (%) non- or ex-smokers smokers unknown 1033 (72.6) 754 (64.9) 275 (67.1) 389 (27.4) 408 (35.1) 135 (32.9) ,0.0001* reference 1.55 (1.26– 1.90) 1.21 (0.91– 1.61) 184 577 128 43 ,0.0001 reference 1.95 (1.54– 2.47) 0.80 (0.59– 1.08) 0.27 (0.18– 0.40) Age, years, median (IQR) Duration since HIV diagnosis, years, median (IQR) BMI, kg/m2, median (IQR) Gender, n (%) male female Season, n (%) autumn winter spring summer 473 (71.2) 818 (58.6) 413 (76.3) 358 (89.3) (28.0) (41.4) (23.7) (10.7) Antiretroviral naive, n (%) yes no 265 (79.3) 1797 (67.6) 69 (20.7) 863 (32.4) ,0.0001* reference 2.61 (1.79– 3.81) eGFR, mL/min/1.73 m2, n (%) ,90 ≥90 934 (73.7) 1027 (64.8) 334 (26.3) 558 (35.2) ,0.0001* reference 1.51 (1.25– 1.83) Downloaded from http://jac.oxfordjournals.org/ by guest on May 16, 2012 Nadir CD4 cell count, cells/mm3, n (%) ≥350 200 to ,350 ,200 1.01 (0.99– 1.01) 1.00 (0.98– 1.01) *x2 test. **Kruskal–Wallis test Logistic regression model adjusted also on care centre. nor did ageing, although it is a well-recognized factor for vitamin D deficiency in the general population. This may be due to the fact that patients included in HIV cohorts are younger than the general population, with very few patients over 60. Our results focused on vitamin D deficiency ≤10 ng/mL because this threshold has been shown to be more frequently associated with clinical manifestations and should have clinical relevance.14,15,24 Epidemiological studies have indeed demonstrated an association between vitamin D deficiency and an overall increased risk of mortality in healthy populations.1,5 In an HIV-infected adult population, analysis of the EuroSIDA cohort confirmed the clinical relevance of vitamin D levels 5 of 9 Allavena et al. Table 4. HIV- and antiretroviral-related factors associated with vitamin D deficiency in the antiretroviral-treated population (logistic regression model: model 1, treatment included in the model as combined antiretroviral therapy regimen; and model 2, treatment included in the model drug by drug) Univariate analysis (n ¼ 2660) Multivariate analysis (n ¼ 2504) vitamin D deficiency no (n ¼ 1797) Age, years, median (IQR) Duration since HIV diagnosis, years, median (IQR) Duration of treatment, years, median (IQR) Duration of current treatment, months, median (IQR) BMI, kg/m2, median (IQR) Gender, n (%) male female 45.8 13.0 9.8 18.0 23.4 (40.3 –52.2) (6.9 –18.1) (4.5 –13.1) (9.0 –34.0) (21.2 –25.8) yes (n ¼ 863) 45.7 13.0 9.3 20.0 23.3 (40.2 –52.2) (6.6 – 18.1) (4.5 – 13.2) (9.0 – 37.0) (21.2 –26.0) aOR (95% CI) P model 1 model 2 0.78** 0.68** 0.72** 0.01** 0.68** 1.00 (0.99 –1.01) 1.00 (0.99 –1.01) 1.00 (1.00 –1.01) 1.00 (0.99 –1.01) 615 (32.0) 248 (33.7) 0.39* 209 (67.9) 652 (71.0) 936 (65.3) 99 (32.1) 266 (29.0) 498 (34.7) 0.01* reference 0.81 (0.58 –1.14) 1.06 (0.76 –1.48) reference 0.81 (0.58 –1.13) 1.07 (0.77 –1.50) Baseline CD4 cell count, cells/mm3, n (%) ≥350 200 to ,350 ,200 1494 (69.3) 174 (58.6) 73 (56.2) 663 (30.7) 123 (41.4) 57 (43.8) ,0.0001* reference 1.51 (1.12 –2.03) 2.05 (1.32 –3.19) reference 1.53 (1.13 –2.06) 2.01 (1.29 –3.13) Prior AIDS, n (%) no yes 1396 (68.3) 401 (65.2) 649 (31.7) 214 (34.8) 0.16* reference 1.18 (0.93 –1.50) reference 1.18 (0.93 –1.49) Hepatitis coinfection, n (%) no yes 1450 (67.9) 347 (66.2) 686 (32.1) 177 (33.8) 0.47* Tobacco smoking status, n (%) non- or ex-smokers smokers unknown 915 (71.3) 640 (63.2) 242 (66.5) 369 (28.7) 372 (36.8) 122 (33.5) 0.0002* reference 1.56 (1.26 –1.93) 1.22 (0.90 –1.66) reference 1.57 (1.27 –1.95) 1.20 (0.89 –1.63) Season, n (%) autumn winter spring summer 414 717 345 321 174 532 116 41 (29.6) (42.6) (25.2) (11.3) ,0.0001* reference 1.93 (1.50 –2.47) 0.78 (0.56 –1.07) 0.25 (0.16 –0.38) reference 1.93 (1.50 –2.47) 0.75 (0.54 –1.04) 0.25 (0.16 –0.38) eGFR, mL/min/1.73 m2, n (%) ,90 ≥90 865 (73.4) 845 (62.1) 313 (26.6) 515 (37.9) ,0.0001* reference 1.54 (1.26 –1.88) reference 1.51 (1.23 –1.85) cART regimen 2 NRTIs + 1 boosted PI 2 NRTIs + 1 NNRTI other 754 (70.6) 573 (62.3) 470 (69.9) 314 (29.4) 347 (37.7) 202 (30.1) 0.0001* reference 1.53 (1.22 –1.92) 0.96 (0.75 –1.23) 1612 (68.5) 185 (60.5) 742 (31.5) 121 (39.5) 0.005* reference 1.30 (0.88 –1.92) lamivudine, n (%) no yes 1036 (70.8) 761 (63.6) 427 (29.2) 436 (36.4) ,0.0001* reference 1.15 (0.77 –1.71) abacavir, n (%) no yes 1219 (69.6) 578 (63.6) 532 (30.4) 331 (36.4) 0.002* reference 1.26 (0.90 –1.76) Nadir CD4 cell count, cells/mm3, n (%) ≥350 200 to ,350 ,200 Antiretroviral contained in the regimen zidovudine, n (%) no yes (70.4) (57.4) (74.8) (88.7) Continued 6 of 9 Downloaded from http://jac.oxfordjournals.org/ by guest on May 16, 2012 1309 (68.0) 488 (66.3) JAC Vitamin D deficiency and antiretroviral drugs Table 4. Continued Univariate analysis (n ¼ 2660) Multivariate analysis (n ¼ 2504) vitamin D deficiency no (n ¼ 1797) yes (n ¼ 863) aOR (95% CI) P model 1 model 2 888 (65.0) 909 (70.3) 478 (35.0) 385 (29.7) 0.004* reference 0.82 (0.55 –1.22) emtricitabine, n (%) no yes 999 (65.7) 798 (70.1) 522 (34.3) 341 (29.9) 0.02* reference 1.15 (0.73 –1.81) didanosine, n (%) no yes 1735 (67.5) 62 (68.1) 834 (32.5) 29 (31.9) 0.91* nevirapine, n (%) no yes 1439 (67.5) 358 (67.8) 693 (32.5) 170 (32.2) 0.89* efavirenz, n (%) no yes 1524 (70.3) 273 (55.4) 643 (29.7) 220 (44.6) ,0.0001* etravirine, n (%) no yes 1744 (67.4) 53 (73.6) 844 (32.6) 19 (26.4) 0.27* ritonavir, n (%) no yes 1072 (64.6) 725 (72.5) 588 (35.4) 275 (27.5) ,0.0001* lopinavir, n (%) no yes 1519 (67.1) 278 (70.2) 745 (32.9) 118 (29.8) 0.22* amprenavir, n (%) no yes 1663 (67.3) 134 (70.5) 807 (32.7) 56 (29.5) 0.36* saquinavir, n (%) no yes 1725 (67.3) 72 (74.2) 838 (32.7) 25 (25.8) 0.15* reference 0.81 (0.44 –1.50) atazanavir, n (%) no yes 1408 (66.8) 389 (70.6) 701 (33.2) 162 (29.4) 0.09* reference 1.02 (0.72 –1.45) darunavir, n (%) no yes 1584 (66.4) 213 (78.0) 803 (33.6) 60 (22.0) ,0.0001* reference 0.87 (0.54 –1.40) raltegravir, n (%) no yes 1607 (67.0) 190 (72.8) 792 (33.0) 71 (27.2) 0.06* reference 0.84 (0.58 –1.22) reference 1.89 (1.45 –2.47) reference 0.97 (0.70 –1.34) Sample too small for zalcitabine (n¼0), stavudine (n¼11), delavirdine (n ¼0), indinavir (n ¼7), nelfinavir (n¼2), tipranavir (n¼4) and enfuvirtide (n ¼4). *x2 test. **Kruskal–Wallis test. Model adjusted also on care centre. 7 of 9 Downloaded from http://jac.oxfordjournals.org/ by guest on May 16, 2012 tenofovir, n (%) no yes Allavena et al. 8 of 9 patients had normal or mildly impaired renal function. However, PTH was inversely correlated with a lower eGFR, as expected. We expect that in renal disease stage 2 and 3, a mild impairment of the enzyme 1a-hydroxylase could lead to an accumulation of vitamin D. Data on 1,25-hydroxyvitamin D were missing in our study and could have been useful in explaining these results. Our study has several limitations. Some data that could represent potential confounding factors, such as sun exposure, socioeconomic factors, ethnicity and dark-coloured skin, were not routinely captured in our database, Measurement of 1,25-hydroxyvitamin D could also be informative for better interpretation of some results but it was not routinely prescribed. Due to the data collection method, it is possible that the clinician did not fully capture over-the-counter multi-vitamin supplements containing vitamin D, or even vitamin D supplementation provided by another physician. In the same way, some characteristics that we did not capture could play a confounding role, particularly regarding the unexpected association between eGFR and vitamin D concentration. As our sample represents 42.4% of the eligible population, a selection bias is always possible, limiting the applicability of our results to the whole HIV population of the five centres participating in the study. In summary, our study confirms the high frequency of vitamin D deficiency in HIV-infected patients. It is very challenging to separate HIV-related factors, including antiretroviral-related factors, from the confounding risk factors, which are over-represented in the HIV-infected population. Apart from risk factors that are well known in the general population, antiretroviral therapy, and particularly the use of efavirenz but not tenofovir, increased the probability of vitamin D deficiency. None of the newer antiretroviral drugs seems to have an impact on vitamin D status. Further studies are needed to evaluate the impact of switching antiretroviral treatment and/or vitamin D supplementation on vitamin D levels and their clinical consequences in an HIV-infected population. Acknowledgements We wish to thank the patients, and the members of the Dat’AIDS group: B. Marchou, P. Massip, E. Bonnet, M. Obadia, M. Alvarez, L. Porte, L. Cuzin, M. Chauveau and I. Lepain (Toulouse); C. Allavena, E. Billaud, C. Biron, B. Bonnet, S. Bouchez, D. Boutoille, C. Brunet-François, N. Feuillebois, T. Jovelin, O. Monoury, P. Morineau, F. Raffi, V. Reliquet, H. Hue and D. Brosseau (Nantes); F. Ajana, I. Alcaraz, V. Baclet, Ph. Choisy, S. Dassonneville, H. Gueroumi, M. Marien, H. Melliez, B. Riff, X. de la Tribonnière, M. Valette, N. Viguet and Y. Yazdanpanah, (Tourcoing); C. Duvivier (Paris); and D. Rey, M. L. Partisani, C. Cheneau, P. Fischer and E. Ebel (Strasbourg). Funding This study was carried out as part of our routine work. Transparency declarations C. A. is a board member of Gilead and Tibotec, and has received travel grants from ViiV Healthcare, BMS, Gilead and Tibotec. L. C. is a board member of BMS and Gilead, and has received travel grants from BMS, Downloaded from http://jac.oxfordjournals.org/ by guest on May 16, 2012 ,12 ng/mL with an increased risk of an AIDS event or death in these patients.15 In our study patients with CD4 ,200 cells/mm3 had a 2-fold higher risk of vitamin D deficiency compared with patients with CD4 ,350 cells/mm3. These results are consistent with the EuroSIDA cohort as immunodepression is associated with AIDS clinical manifestation and death. Our study demonstrates that antiretroviral-treated patients are at higher risk of vitamin D deficiency and is, to our knowledge, the largest to evaluate the impact of each commonly used antiretroviral drug, including the most recently available molecules (etravirine, darunavir and raltegravir). Neither raltegravir nor any PI was related to the vitamin D level. An effect of NNRTIs on vitamin D has been found in some previous studies with a small sample size that precluded the inclusion of individual NNRTIs in the multivariate analysis.22,26 – 28 Our study clearly shows that, within the NNRTI class, only efavirenz is a risk factor for vitamin D deficiency. The negative effect of efavirenz on vitamin D has been suggested in the MONET trial, in which patients showed a higher increase in vitamin D over 96 weeks when they stopped efavirenz or zidovudine rather than other antiretrovirals. In the Phase 3 ECHO trial a significant decrease in vitamin D occurred in patients treated with efavirenz but not with rilpivirine over 48 weeks.29,30 Efavirenz has been shown to reduce the expression of cytochrome P450 2RI, one of the enzymes involved in the 25-hydroxylation of vitamin D3, and to induce cytochrome 24, which converts vitamin D and calcitriol into their inactive metabolites.26,28,31 Although a high incidence of osteopenia/osteoporosis has been observed in patients receiving a PI-based therapy, a cause-and-effect relationship has not been clearly established. In vitro ritonavir impaired 25-hydroxylase activity in a human hepatocyte cell line and inhibited enzymes for calcitriol synthesis and catabolism in human macrophages, but no dramatic decrease in vitamin D levels was observed in patients treated with PI.28 In our study, vitamin D level was not modified in patients receiving a PI-based regimen. The effect of NRTI drugs on vitamin D metabolism has never been studied but interactions are unlikely as NRTIs are not known to be cytochrome P450 modifiers. Tenofovir is known to induce a proximal renal tubular dysfunction that could reduce the effect of 1a-hydroxylase in the kidney, which metabolizes vitamin D into the active form of 1,25-hydroxyvitamin D, leading to vitamin D accumulation. In concordance with the SUN study, we found a statistical negative association between vitamin D level and tenofovir using a linear regression analysis (data not shown).16 The association between tenofovir and less frequent vitamin D deficiency did not persist in the multivariate analysis, which could reflect either a mild effect of tenofovir or a confounding effect of others drugs. Several cases of hyperparathyroidism have recently been reported in HIV-infected patients treated with tenofovir.32,33 No hyperparathyroidism was seen in our study even in patients receiving tenofovir. Vitamin D deficiency was associated with higher PTH concentrations that nonetheless remained within the normal range. Furthermore, no association was shown between PTH and tenofovir. Although not statistically significant, the results of the Swiss cohort are similar to those reported here.22 Surprisingly, but concordantly with the SUN study, vitamin D deficiency was negatively associated with an eGFR ≥90 mL/min/1.73 m2.16 Of note, a large majority of Vitamin D deficiency and antiretroviral drugs JAC 15 Viard JP, Souberbielle JC, Kirk O et al. Vitamin D and clinical disease progression in HIV infection: results from the EuroSIDA study. AIDS 2011; 25: 1305– 15. MSD, ViiV HealthCare and Tibotec. D. R. is a board member of BMS and Gilead. C. Duvivier has received consulting fees or an honorarium from MSD, BMS and Gilead, is a board member of MSD and has received travel grants from MSD, BMS and Gilead. E. B. is a board member of Gilead, BMS, MSD and Abbott, and has received grants from Gilead, Abbott, Janssen, BMS, MSD, ViiV Healthcare, Roche, Boehringer-Ingelheim and Nephrotek. F. R. has received research funding or honoraria from or consulted for Abbott, Boehringer-Ingelheim, BMS, Ferrer, Gilead Sciences, GlaxoSmithKline, Janssen-Cilag, MSD, Pfizer, Splicos and ViiV Healthcare. C. Delpierre, N. V., J. B. and P. G.: none to declare. 16 Dao CN, Patel P, Overton ET et al. Low vitamin D among HIV-infected adults: prevalence of and risk factors for low vitamin D levels in a cohort of HIV-infected adults and comparison to prevalence among adults in the US general population. Clin Infect Dis 2011; 52: 396–405. Author contributions 18 Welz T, Childs K, Ibrahim F et al. Efavirenz is associated with severe vitamin D deficiency and increased alkaline phosphatase. AIDS 2010; 24: 1923– 8. All authors read and contributed to the manuscript. C. A. and C. Delpierre wrote the first draft of the article with critical review by all the authors. D. R., L. C., N. V., J. B., P. G., C. Duvivier, E. B. and F. R. participated in the study design and analysis plan, reviewed the study data reports and approved the final manuscript. 1 Dawson-Hughes B, Heaney RP, Holick MF et al. Estimates of optimal vitamin D status. Osteoporos Int 2005; 16: 713– 6. 2 Michos ED, Melamed ML. Vitamin D and cardiovascular disease risk. Curr Opin Clin Nutr Metab Care 2008; 11: 7– 12. 3 Souberbielle JC, Body JJ, Lappe JM et al. Vitamin D and musculoskeletal health, cardiovascular disease, autoimmunity and cancer: recommendations for clinical practice. Autoimmun Rev 2011; 9: 709–15. 4 Tsoukas CD, Provvedini DM, Manolagas SC. 1,25-Dihydroxyvitamin D3: a novel immunoregulatory hormone. Science 1984; 224: 1438–40. 19 Pugliese P, Cuzin L, Cabie A et al. A large French prospective cohort of HIV-infected patients: the Nadis Cohort. HIV Med 2009; 10: 504–11. 20 Levey AS, Eckardt KU, Tsukamoto Y et al. Definition and classification of chronic kidney disease: a position statement from Kidney Disease: Improving Global Outcomes (KDIGO). Kidney Int 2005; 67: 2089– 100. 21 European AIDS Clinical Society. EACS Guidelines, Version 6.0. 2011. www.europeanaidsclinicalsociety.org (03 May 2012, date last accessed). 22 Mueller NJ, Fux CA, Ledergerber B et al. High prevalence of severe vitamin D deficiency in combined antiretroviral therapy-naive and successfully treated Swiss HIV patients. AIDS 2010; 24: 1127– 34. 23 Chapuy MC, Preziosi P, Maamer M et al. Prevalence of vitamin D insufficiency in an adult normal population. Osteoporos Int 1997; 7: 439–43. 24 Yetley EA. Assessing the vitamin D status of the US population. Am J Clin Nutr 2008; 88: 558S–64S. 5 Melamed ML, Michos ED, Post W et al. 25-hydroxyvitamin D levels and the risk of mortality in the general population. Arch Intern Med 2008; 168: 1629– 37. 25 Ganji V, Zhang X, Tangpricha V. Serum 25-hydroxyvitamin D concentrations and prevalence estimates of hypovitaminosis D in the U.S. population based on assay-adjusted data. J Nutr 2012; 142: 498–507. 6 Mehta S, Giovannucci E, Mugusi FM et al. Vitamin D status of HIV-infected women and its association with HIV disease progression, anemia, and mortality. PLoS One 2010; 5: e8770. 26 Brown TT, McComsey GA. Association between initiation of antiretroviral therapy with efavirenz and decreases in 25-hydroxyvitamin D. Antivir Ther 2010; 15: 425–9. 7 Amiel C, Ostertag A, Slama L et al. BMD is reduced in HIV-infected men irrespective of treatment. J Bone Miner Res 2004; 19: 402– 9. 27 Pasquet A, Viget N, Ajana F et al. Vitamin D deficiency in HIV-infected patients: associated with non-nucleoside reverse transcriptase inhibitor or efavirenz use? AIDS 2011; 25: 873– 4. 8 Brown TT, Qaqish RB. Antiretroviral therapy and the prevalence of osteopenia and osteoporosis: a meta-analytic review. AIDS 2006; 20: 2165– 74. 9 Cazanave C, Dupon M, Lavignolle-Aurillac V et al. Reduced bone mineral density in HIV-infected patients: prevalence and associated factors. AIDS 2008; 22: 395–402. 10 Thomas J, Doherty SM. HIV infection—a risk factor for osteoporosis. J Acquir Immune Defic Syndr 2003; 33: 281–91. 11 Duvivier C, Kolta S, Assoumou L 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. AIDS 2009; 23: 817–24. 12 Gallant JE, Staszewski S, Pozniak AL et al. Efficacy and safety of tenofovir DF vs stavudine in combination therapy in antiretroviral-naive patients: a 3-year randomized trial. JAMA 2004; 292: 191–201. 13 Madeddu G, Spanu A, Solinas P et al. Bone mass loss and vitamin D metabolism impairment in HIV patients receiving highly active antiretroviral therapy. Q J Nucl Med Mol Imaging 2004; 48: 39– 48. 14 Vescini F, Cozzi-Lepri A, Borderi M et al. Prevalence of hypovitaminosis D and factors associated with vitamin D deficiency and morbidity among HIV-infected patients enrolled in a large Italian cohort. J Acquir Immune Defic Syndr 2011; 58: 163– 72. 28 Childs K, Welz T, Samarawickrama A et al. Effects of vitamin D deficiency and combination antiretroviral therapy on bone in HIV-positive patients. AIDS 2012; 26: 253– 62. 29 Fox J, Peters B, Prakash M et al. Improvement in vitamin D deficiency following antiretroviral regime change: results from the MONET trial. AIDS Res Hum Retroviruses 2011; 27: 29–34. 30 Wohl D, Doroana M, Orkin C et al. Change in vitamin D levels smaller and risk of development of severe vitamin D deficiency lower among HIV-1-infected, treatment-naı̈ve adults receiving TMC278 compared with EFV: 48-week results from the Phase III ECHO trial. In: Abstracts of the Eighteenth Conference on Retroviruses and Opportunistic Infection, Boston, MA, USA, 2011. Abstract 79LB. Foundation for Retrovirology and Human Health, Alexandria, VA, USA. 31 Fabbriciani G, Pirro M, Scarponi AM et al. Osteoporosis in men: other secondary causes. Joint Bone Spine 2009; 76: 579– 80; author reply 580. 32 Pocaterra D, Carenzi L, Ricci E et al. Secondary hyperparathyroidism in HIV patients: is there any responsibility of highly active antiretroviral therapy? AIDS 2010; 25: 1430– 3. 33 Rosenvinge MM, Gedela K, Copas AJ et al. Tenofovir-linked hyperparathyroidism is independently associated with the presence of vitamin D deficiency. J Acquir Immune Defic Syndr 2011; 54: 496–9. 9 of 9 View publication stats Downloaded from http://jac.oxfordjournals.org/ by guest on May 16, 2012 References 17 Adeyemi OM, Agniel D, French AL et al. Vitamin D deficiency in HIV-infected and HIV-uninfected women in the United States. J Acquir Immune Defic Syndr 2011; 57: 197– 204.