Vitamin D, in its biologically active form calcitriol (1,25(OH)2D), plays an important role in calcium homeostasis, regulation of bone turnover, innate immune function, cell proliferation, and cell differentiation . Vitamin D deficiency (VDD) has been associated with numerous adverse health outcomes, including rickets, osteomalacia, cardiovascular mortality, cancer, autoimmune disease, and multiple sclerosis . Prior to the availability of combination antiretroviral therapy (cART), VDD in HIV-infected patients was associated with reduced survival .
High rates (29–57%) of VDD in HIV-infected patients have recently been reported [3–5]. Two of these studies suggested an association between low vitamin D levels and exposure to nonnucleoside reverse transcriptase inhibitor (NNRTI)-containing cART, though their sample size did not allow specific drugs to be implicated [4,5].
In 2008, the prevalence of VDD in our large, ethnically diverse HIV cohort was assessed as part of routine clinical care. The aims of the present study were to evaluate the associations between exposure to specific antiretroviral drugs and severe VDD and to explore the effects of VDD and antiretroviral drug exposure on serum alkaline phosphatase (ALP) as a surrogate marker of bone turnover.
Between June and December 2008, vitamin D status was assessed in all HIV-positive patients attending King's College Hospital, London, UK. Calcidiol (25(OH)D), the circulating form of vitamin D and the best indicator of total body vitamin D , was measured in all patients in addition to the regular blood panel, which included renal, bone, liver and lipid profiles, full blood count, CD4 cell count, and HIV RNA level. Prior to June 2008, routine clinical care did not include assessment of vitamin D status or provision of vitamin D supplements. All patients found to have VDD as part of the present evaluation were offered vitamin D supplementation. In accordance with local regulations, ethical approval for this study was not required, as all investigations were done as service evaluation or standard of care.
Serum 25(OH)D was measured by enzyme immunoassay (Immunodiagnostics Systems Limited, Boldon, Tyne and Wear, UK), with intraassay and interassay coefficients of variation of less than 10% for the duration of the study. Vitamin D status was defined according to 25(OH)D level as severely deficient [<10 μg/l (<25 nmol/l)], deficient [10–20 μg/l (25–50 nmol/l)], insufficient [20–30 μg/l (50–75 nmol/l)], or optimal [>30 μg/l (>75 nmol/l)] . As patients were sampled over a 7-month period, season of sampling was included in the analysis as a dichotomous variable and considered to have occurred in summer (from June to September) or winter (from October to December). Demographic, clinical, and laboratory parameters were abstracted from the electronic patient record (EPR) and described for patients with or without severe VDD [25(OH)D <10 μg/l].
Factors associated with severe VDD and upper quartile ALP levels (compared to all other patients) were examined in multivariate logistic regression models using Stata 10.0 (Stata Corporation, College Station, Texas, USA). In the absence of bone-specific ALP, these analyses were restricted to patients with normal aspartate transaminase (AST<50 IU/l) to better reflect increased bone turnover and included gamma-glutamyl transferase (GGT) as a continuous variable . All factors significant to P less than 0.1 in univariate analyses were tested in multivariate models. All reported P values are two-sided.
Vitamin D levels were measured in 1077 patients. The median age was 41 [interquartile range (IQR) 36–47] years, 41% of patients were women and 61% of black ethnicity. The median current CD4 cell count was 456 [IQR 328–616] cells/μl, and 9 and 55% of patients had current and nadir CD4 cell counts less than 200 cells/μl, respectively. At the time of sampling, 78% were taking cART, which was ritonavir-boosted protease inhibitor-based in 31% and NNRTI-based in 69%; 12% were naive to antiretroviral therapy. Thirty-three percent of samples were obtained during the winter months.
Prevalence of and risk factors for severe vitamin D deficiency
One-third (34.8%) of patients were severely vitamin D deficient, 38.7% were deficient, and 17.7% had suboptimal vitamin D levels. Only 8.7% had optimal vitamin D levels of at least 30 μg/l. In a multivariate model that included all 1077 patients, sampling in winter, black ethnicity, nadir CD4 cell count less than 200 cells/μl, and current cART use were significantly associated with severe VDD (Table 1a).
When the analysis was restricted to 843 patients on cART, winter season, black ethnicity, and nadir CD4 cell count less than 200 cells/μl remained associated with severe VDD. In addition, exposure to efavirenz (EFV)-based cART, but not nevirapine-based (NVP-based) or protease inhibitor-based cART, was associated with severe VDD (Table 1b). Age, sex, current CD4 cell count, current HIV-1 RNA load, estimated glomerular filtration rate (eGFR) less than 60 ml/min, serum albumin, and use of tenofovir (TDF), abacavir, or other nucleoside reverse transcriptase inhibitors (NRTIs) were not independently associated with severe VDD in either of the above models (data not shown). In supplementary analyses of patients on cART, cumulative exposure to EFV [adjusted odds ratio (aOR) 1.0003 (1.0001–1.0004) per day of EFV exposure, P = 0.001] was significantly associated with severe VDD, correcting for season, nadir CD4 cell count, and ethnicity. Patients exposed to EFV for 30–90 days had significantly lower median (IQR) 25(OH)D levels [8.6 (5.3–9.6) μg/l] than those exposed for less than 30 days [13.9 (9–21.1) μg/l] (P = 0.04; Kruskal–Wallis test).
Factors associated with raised alkaline phosphatase (with aspartate transaminase <50 IU/l)
The median (IQR) ALP was 71.5 (56–89) IU/l in patients not on cART and 85 (69–108) IU/l in those on cART (P ≤ 0.0001; Kruskal–Wallis test). In multivariate analysis including all patients, current cART use and increased GGT were significantly associated with upper quartile ALP, whereas black ethnicity was protective (Table 2a).
In analyses restricted to patients on cART, current TDF and EFV use and GGT were independent risk factors for upper quartile ALP, whereas current NVP use was protective (Table 2b). We found no association between upper quartile ALP and severe VDD among all patients [univariate OR 1.2 (0.9–1.6)] or among those on cART [univariate OR 1.1 (0.8–1.5)].
In supplementary analyses, we examined the potential effect of EFV/TDF coadministration on ALP. Compared to patients on regimens containing neither EFV nor TDF, those on regimens containing TDF and EFV [aOR 5.4 (2.8–10.1)] or TDF without EFV [aOR 3.5 (2.0–6.1)], but not those on EFV without TDF [aOR 1.6 (0.8–3.2)] were more likely to have upper quartile ALP levels. These analyses were then stratified by vitamin D level (above or below the median, 13 μg/l). Among patients with 25(OH)D levels less than 13 μg/l, use of EFV without TDF was significantly associated with upper quartile ALP [aOR 2.6 (1.2–5.8)] (Table 2c). This association was not seen among patients with vitamin D levels of at least 13 μg/l (Table 2d). In addition, there appeared to be an additive effect of TDF and EFV in combination on upper quartile ALP at 25(OH)D levels above and below 13 μg/l (Table 2c and 2d).
We observed a small but statistically significant difference in mean serum calcium [corrected for albumin (CCA)] levels between patients with 25(OH)D less than 10 μg/l and more than 10 μg/l [2.17 (SD 0.09) and 2.18 (SD 0.08) nmol/l, respectively, P = 0.04], though mean CCA levels remained within the normal range. There was no significant difference in serum phosphate levels between patients with and without severe VDD.
Low 25(OH)D levels were almost universal in this HIV cohort, and one-third of patients had severe VDD. Traditional risk factors such as black ethnicity and winter season, as well as current EFV use and CD4 cell nadir less than 200 cells/μl were independent risk factors for severe VDD. In addition, we identified exposure to EFV and TDF, but not severe VDD, as an independent risk factor for increased bone turnover (as assessed by upper quartile ALP with normal AST) among patients on cART. Although several studies have documented increased bone turnover, raised ALP, and/or reduced bone mineral density in patients receiving TDF [7–10], the associations between EFV and severe VDD, EFV and upper quartile ALP, and the apparently additive effects of EFV and TDF on ALP levels have not been reported previously.
Several mechanisms can be postulated by which EFV may affect vitamin D homeostasis: EFV induces CYP3A4 and CYP24. Induction of CYP3A4, a 25(OH)D hydroxylase, which converts vitamin D to 25(OH)D, may reduce the available amount of vitamin D substrate, whereas induction of CYP24, which catalyzes 25(OH)D and 1,25(OH)2D, may result in reduced vitamin D levels. In addition, by reducing transcription of CYP2R1, another 25(OH)D hydroxylase , EFV may also reduce 25(OH)D production. Of note, phenobarbital, a recognized cause of osteomalacia, also induces CYP3A4 while suppressing CYP2R1 [11,12].
Although not measured in our patients, the raised ALP in patients receiving EFV is likely to reflect increased parathyroid hormone (PTH) production. By inducing CYP24, EFV may increase the breakdown of 25(OH)D and 1,25(OH)2D to their inactive metabolites. As 1,25(OH)2D inhibits PTH production (both directly, through reduced PTH gene transcription, and indirectly, via increased calcium absorption), reduced 1,25(OH)2D levels may result in elevated PTH, particularly in patients on TDF in whom elevated PTH is more common [7–10]. Of interest, the effect of EFV exposure on ALP was restricted to patients with 25(OH)D less than 13 μg/l, suggesting a threshold below which the effects of EFV on vitamin D homeostasis become severe enough to result in disinhibition of PTH secretion. Interestingly, a similar threshold effect has been observed for the association between VDD [25(OH)D <15 μg/l] and insulin resistance . By contrast, we found the effect of TDF on ALP, which may be mediated by increased PTH production or increased renal phosphate losses , to be independent of 25(OH)D levels.
The lack of association between severe VDD and upper quartile ALP in this study may reflect the fact that only 8.7% of patients had optimal 25(OH)D levels, resulting in comparisons between patients with VDD of varying severity. The relationship between 25(OH)D, PTH and ALP is known to be nonlinear and dependent on factors such as ethnicity and dietary calcium intake [15–17]. For example, there is no threshold 25(OH)D level at which secondary hyperparathyroidism develops and black patients have lower ALP for given levels of PTH [15–17].
This study has several limitations, including the cross-sectional nature of the analyses, the lack of PTH measurements and specific markers of bone turnover, and the very small number of patients with optimal vitamin D levels that could have served as a better control group. Nonetheless, the large sample size, the diversity of the patient population and their low propensity to seek healthcare, including vitamin supplementation, from other sources have provided a valuable resource to assess the associations between specific antiretrovirals, vitamin D levels and bone turnover. The observed associations will, however, need to be confirmed in longitudinal studies and their clinical significance remains to be defined.
In summary, we found that current EFV use was associated with VDD, a condition associated with multiple adverse health outcomes and a well known cause of hyperparathyroidism. Numerous studies have found osteopenia and osteoporosis, and possibly bone fractures, to be more common among HIV-infected patients [18–21]. The implications of VDD for bone health, cardiovascular status, and immune function in HIV-infected patients deserve further study. The apparent additive effect of EFV and TDF on markers of increased bone turnover in this cohort of patients with low vitamin D levels is novel and potentially important, given the high prevalence of VDD and the widespread use of these drugs in first-line cART.
Serum 25(OH)D quantification was performed as part of a National Health Service (NHS) audit by C.F.M. T.W., M.P., and F.A.P. designed the study. T.W. and F.I. performed the analyses, and K.C., C.F.M., M.P., C.B.T., and F.A.P. assisted with the interpretation of the results. T.W., K.C., and F.A.P. wrote the article with input from all authors. The final version of the article was approved by all authors.
No funding was received to conduct this work.
None of the authors has any financial or personal relationships with people or organizations that could inappropriately influence this work, though several authors have, at some stage in the past, received funding from a variety of pharmaceutical companies for research, travel grants, speaking engagements, or consultancy fees.
Part of this study was presented at the 5th IAS Conference on HIV Pathogenesis, Treatment and Prevention, held in Cape Town, South Africa from 19 to 22 July 2009 (abstract TUPEB186).
1. Holick MF. Vitamin D deficiency. N Engl J Med 2007; 357:266–281.
2. Haug C, Muller F, Aukrust P, Froland SS. Subnormal serum concentration of 1,25-vitamin D in human immunodeficiency virus infection: correlation with degree of immune deficiency and survival. J Infect Dis 1994; 169:889–893.
3. Rodriguez M, Daniels B, Gunawardene S, Robbins GK. High frequency of vitamin D deficiency in ambulatory HIV-positive patients. AIDS Res Hum Retroviruses 2009; 25:9–14.
4. Van Den Bout-Van Den Beukel CJ, Fievez L, Michels M, Sweep FC, Hermus AR, Bosch ME, et al
. Vitamin D deficiency among HIV type 1-infected individuals in the Netherlands: effects of antiretroviral therapy. AIDS Res Hum Retroviruses 2008; 24:1375–1382.
5. Rosenvinge MMM, Gedela KM, Copas AJP, Wilkinson AM, Sheehy CA, Bano GM, et al.Tenofovir-linked hyperparathyroidism is independently associated with the presence of vitamin D deficiency
. J Acquir Immune Defic Syndr
2010 (in press).
6. Pratt DS, Kaplan MM. Chapter 296. Evaluation of liver function.
In: Fauci AS, Braunwald E, Kasper D, Hauser SL, Longo DL, Jameson LJ, Loscalzo J, editors. Harrison's principles of internal medicine
(online version). AccessMedicine from McGraw-Hill. http://www.accessmedicine.com/content.aspx?aid=2873571
(Accessed 7 June 2010).
7. Martin A, Bloch M, Amin J, Baker D, Cooper DA, Emery S, et al
. Simplification of antiretroviral therapy with tenofovir-emtricitabine or abacavir-lamivudine: a randomized, 96-week trial. Clin Infect Dis 2009; 49:1591–1601.
8. Stellbrink HJ, Moyle G, Orkin C, Arribas J, Pearce H, Zucchi P, for the ASSERT Team. Assessment of safety and efficacy of abacavir/lamivudine and tenofovir/emtricitabine in treatment-naive HIV-1 infected subjects. ASSERT: 48-Week Result.12th European AIDS Conference
. Cologne, Germany. November 11–14, 2009.
9. Fux CA, Rauch A, Simcock M, Bucher HC, Hirschel B, Opravil M, et al
. Tenofovir use is associated with an increase in serum alkaline phosphatase in the Swiss HIV Cohort study. Antivir Ther 2008; 13:1077–1082.
10. 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. JAMA 2004; 292:191–201.
11. Ellfolk M, Norlin M, Gyllensten K, Wikvall K. Regulation of human vitamin D(3) 25-hydroxylases in dermal fibroblasts and prostate cancer LNCaP cells. Mol Pharmacol 2009; 75:1392–1399.
12. Pascussi JM, Robert A, Nguyen M, Walrant-Debray O, Garabedian M, Martin P, et al
. Possible involvement of pregnane X receptor-enhanced CYP24 expression in drug-induced osteomalacia. J Clin Invest 2005; 115:177–186.
13. Ashraf A, Alvarez J, Saenz K, Gower B, McCormick K, Franklin F. Threshold for effects of vitamin D deficiency on glucose metabolism in obese female African-American adolescents. J Clin Endocrinol Metab 2009; 94:3200–3206.
15. Sahota O, Mundey MK, San P, Godber IM, Lawson N, Hosking DJ. The relationship between vitamin D and parathyroid hormone: calcium homeostasis, bone turnover, and bone mineral density in postmenopausal women with established osteoporosis. Bone 2004; 35:312–319.
16. Aloia JF, Talwar SA, Pollack S, Feuerman M, Yeh JK. Optimal vitamin D status and serum parathyroid hormone concentrations in African American women. Am J Clin Nutr 2006; 84:602–609.
17. Sawaya BP, Monier-Faugere MC, Ratanapanichkich P, Butros R, Wedlund PJ, Fanti P. Racial differences in parathyroid hormone levels in patients with secondary hyperparathyroidism. Clin Nephrol 2002; 57:51–55.
18. Garcia Aparicio AM, Munoz Fernandez S, Gonzalez J, Arribas JR, Pena JM, Vazquez JJ, et al
. Abnormalities in the bone mineral metabolism in HIV-infected patients. Clin Rheumatol 2006; 25:537–539.
19. 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 2008; 93:3499–3504.
20. Brown TT, McComsey GA. Osteopenia and osteoporosis in patients with HIV: a review of current concepts. Curr Infect Dis Rep 2006; 8:162–170.
21. Brown TT, Qaqish RB. Antiretroviral therapy and the prevalence of osteopenia and osteoporosis: a meta-analytic review. AIDS 2006; 20:2165–2174.