Background:  Hypovitaminosis D (hypD) is a frequent condition in industrialized countries and is associated with several pathological states. A high prevalence of hypD in HIV⁺ patients has been reported, but the reasons are unclear. The aim of this study is to estimate the prevalence of hypD in a cohort of HIV⁺ patients before and after start of combination antiretroviral therapy (cART), and its association with anthropometric, environmental, and clinical data.

Methods:  Patients of the ICONA cohort, for whom a stored plasma sample was available before and after starting cART, were used for dosing 25(OH)-vitaminD (vitD) concentration; all data used were recorded at blood collection. VitD insufficiency defined as 25(OH)-vitD <75 nmol/L, values <30 nmol/L considered vitD deficiency. Differences in the proportion of patients with insufficiency were tested by chi-square test and logistic regression analysis accounting for non independence of tests. The change in absolute levels of vitD pre/post cART was modelled by linear regression controlling for confounders and seasonality (only pairs of tests measured in the same season included).

Results: We studied 856 patients contributing 1,505 tests: 262 before and 1,243 after a median of 14 months (IQR: 8 to 19) of ART (369 non-nucleoside reverse transcriptase inhibitor-based and 874 protease inhibitor-based regimens). Median age 36 years (IQR: 20 to 69); 800 (93%) were from Italy, 11 (1%) from Europe, 23 (3%) from Africa, 18 (2%) from Central/South America, 4 (0.5%) from Asia. VitD insufficiency and deficiency were found in 807 (54%) and 98 (7%) of the tests.

In multivariable analysis, only age (OR = 1.35 per 10 years older, 95%CI 1.2 to 1.6, P =0.0001), Western-world nationality (OR = 0.35 vs African, Asian, or Centre/South American, 95%CI 0.2 to 0.6, P =0.0001) and duration of ART (OR = 1.33 per year longer, 95%CI 1.1 to 1.5, P =0.0001) were independently associated with vitD insufficiency risk . There was no evidence from the linear regression model that change in vitD levels from pre-ART to post ART varied according to whether patients were receiving protease inhibitors vs non-nucleoside reverse transcriptase inhibitors (mean difference/year = +15 nmol/L, 95% CI -6 to +36, P =0.17).

Conclusions: This is the first large study confirming a very high prevalence of hypD in HIV⁺ patients, much more frequent than in the general population. These data should carefully considered in the clinical management of HIV⁺ patients, cause our patients were very much younger that HIV^‑ case series reported in the literature, and high negative impact of hypD on risk of morbidity in this population should be carefully taken into account.