Vitamin D is best known for its effects on bone health, but a growing appreciation of its many immunomodulatory actions has prompted significant recent interest in its potential to prevent and treat diverse viral and bacterial infections [1–4]. Much of this research has focused on the question of whether vitamin D supplementation might slow progression of disease caused by human immunodeficiency virus (HIV) and prevent tuberculosis and other opportunistic infections in HIV-infected people [5]. Physiological concentrations of calcitriol, the active vitamin D metabolite, inhibit HIV replication in human macrophages in vitro and restrict growth of Mycobacterium tuberculosis through the induction of autophagy [6, 7]. Observational studies conducted in HIVinfected people who are not taking antiretroviral therapy (ART) have reported associations between vitamin D deficiency and susceptibility to tuberculosis, upper respiratory tract infection, and oral candidiasis [8, 9], raising the possibility that supplementation might reduce susceptibility to infections in this patient group. Surely, though, once a potent pharmacological therapy like ART is initiated, isn’t deficiency in a mere micronutrient unlikely to influence clinical outcomes in HIV-infected people? Not so, according to a cohort study published in this issue of the Journal [10]. Sudfeld and colleagues studied 1103 HIV-infected adults in Tanzania who were taking ART and participating in a randomized controlled trial of a multivitamin supplement that did not contain vitamin D. Clinical assessments were performed at monthly intervals, and the median duration of follow-up was 20.6 months. After multivariate adjustment, vitamin D deficiency (defined as a serum concentration of 25-hydroxyvitamin D of 10% weight loss (HR, 2.10; P = .02) but not with risk of malaria, pneumonia, or anemia. A milder degree of deficiency, termed “vitamin D insufficiency” and defined as a serum 25-hydroxyvitamin D concentration of 20–30 ng/mL, did not associate with any of these outcomes. Is the reported association between baseline vitamin D deficiency and increased risk of tuberculosis likely to be causal? In favor of this conclusion, the association survived statistical adjustment for a number of potential confounders of the relationship between vitamin D status and risk of opportunistic infection, including sex, age, season of sampling, body mass index, measures of HIV disease stage at baseline, and ART regimen. The authors also went to considerable lengths to investigate whether this association arose as a result of reverse causality, by excluding from their analysis individuals who developed tuberculosis within a month of enrollment. Notably, the association between baseline vitamin D deficiency and increased risk of subsequent tuberculosis was still observed even when diagnoses of pulmonary tuberculosis arising in the first 2 months of the study were excluded in a sensitivity analysis. This finding is significant, as tuberculosis itself might contribute to vitamin D deficiency, by reducing patients’ sun exposure or increasing consumption of 25-hydroxyvitamin D by activated macrophages. The observation that vitamin D deficiency preceded the onset of tuberculosis in the present study is consistent with findings from a cohort study conducted in HIVuninfected patients [11], and it effectively excludes reverse causality as an explanation for the association between vitamin D deficiency and tuberculosis risk reported here. A final point in favor of a causal interpretation is the biological plausibility of the hypothesis: in Received and accepted 31 October 2012; electronically published 16 November 2012. Correspondence: Adrian R. Martineau, MRCP, PhD, Centre for Primary Care and Public Health, Barts and The London School of Medicine, Queen Mary University of London, 58 Turner St, London E1 2AB, United Kingdom (a.martineau@ qmul.ac.uk). The Journal of Infectious Diseases 2013;207:373–5 © The Author 2012. Published by Oxford University Press on behalf of the Infectious Diseases Society of America. All rights reserved. For Permissions, please e-mail: journals. permissions@oup.com. DOI: 10.1093/infdis/jis697