Pasquet and colleagues observed, in keeping with previous studies [1–5], an association between exposure to non-nucleoside reverse transcriptase inhibitor (NNRTI)-containing antiretroviral therapy and vitamin D deficiency in their cross-sectional cohort study [6]. Their study was insufficiently powered to look at individual NNRTI, and no significant association between vitamin D deficiency and either efavirenz or nevirapine was observed. However, comparable crude and adjusted coefficients for efavirenz and nevirapine led the authors to question whether the observed association between vitamin D deficiency and NNRTI reflects an NNRTI class effect rather than a phenomenon restricted to efavirenz as reported by ourselves [2]. An increasing body of evidence supports an association between efavirenz and 25-hydroxy-vitamin D [25(OH)D] deficiency [2,7,8]. Initiation of efavirenz appears to be consistently associated with reductions in 25(OH)D levels [5,6,9], and preliminary data suggest that etravirine and efavirenz may have similar effects on 25(OH)D homeostasis [10]. Although we found an association between efavirenz use and severe 25(OH)D deficiency [adjusted odds ratio (aOR) 2.0 (1.5, 2.7)] as well as raised alkaline phosphatase levels [aOR 1.6 (1.02, 2.4)], nevirapine use appeared to be protective against 25(OH)D deficiency [aOR 0.6 (0.3, 1.5)] and raised alkaline phosphatase [aOR 0.5 (0.3, 0.9)] in multivariate models adjusted for sex, ethnicity, season and CD4 cell count [2], and no association between nevirapine and 25(OH)D deficiency or insufficiency was observed in the SUN cohort [7]. Whilst comparable reductions in 25(OH)D have been observed in patients initiating efavirenz and nevirapine [5], the small sample size of this study precluded inclusion of individual NNRTI in multivariate analyses. Furthermore, no significant reductions in 25(OH)D from baseline were observed in 18 patients who initiated zidovudine, lamivudine and nevirapine [11] and in 27 patients who commenced nevirapine together with ritonavir-boosted lopinavir [12]. Whereas the effects of antiretroviral treatment-associated reductions in 25(OH)D on bone remain to be defined, a trend toward lower bone mineral density (BMD) with efavirenz and higher BMD with nevirapine has been reported in HIV infected children [13]. In clinical practice, efavirenz and nevirapine are preferentially used in some patients and avoided in others; associations with either NNRTI in observational cohort studies may thus be subject to channelling bias. For example, nevirapine may be the drug of choice for premenopausal African women with low CD4 cell counts and efavirenz for white men with preserved CD4 cell counts. As black ethnicity, female sex and low nadir CD4 cell count were all independently associated with severe vitamin D deficiency in our study [2], an inability to adjust for these factors may lead to a spurious association between nevirapine and vitamin D deficiency in cross-sectional studies. On the contrary, 61% of our patients were black, and it is possible that the effects of nevirapine on 25(OH)D homeostasis differ by ethnicity [14]. Of note, the two studies that reported a decline in 25(OH)D with nevirapine or an association between nevirapine and vitamin D deficiency predominantly included white men with preserved CD4 cell counts, did not adjust for sex, ethnicity and nadir CD4 cell count, and showed that 25(OH)D levels in black patients were minimally affected by antiretroviral treatment [5,6]. The consistent finding that vitamin D deficiency is very common in HIV-infected patients raises more important questions than which antiretroviral drug(s) may be implicated, namely what is the clinical significance of low 25(OH)D and is there any benefit from vitamin D supplementation? Vitamin D deficiency in the general population has been associated with numerous adverse health outcomes including cardiovascular disease, cancer, infection, osteopenia and fractures [15]. HIV infection is associated with an increased risk of (opportunistic) infections, cancer [16] and low BMD [17], and exposure to antiretroviral therapy with an increased risk of cardiovascular events [18] and further reductions in BMD [19]. As the greatest reductions in BMD are observed in patients who initiate antiretroviral therapy [12,20–21], the contribution of vitamin D deficiency and the potential of vitamin D supplementation to reduce this initial bone loss deserve further study. An effect of ethnicity and specific antiretrovirals, especially efavirenz and tenofovir, should be considered in the design of these studies. Acknowledgements None of the authors has any financial or personal relationships with people or organizations that could inappropriately influence this work, although all authors have, at some stage in the past, received funding from a variety of pharmaceutical companies for travel grants. In addition, F.A.P. has received research funding, speaking engagements or consultancy fees from GlaxoSmithKline, ViiV, Tibotec, Bristol-Meyers Squibb and Gilead Sciences.