Background: A large number of observational studies have reported harmful effects of low 25-hydroxyvitamin D (25OHD) levels on non-skeletal outcomes. We performed a systematic quantitative review on characteristics of randomized clinical trials (RCTs) included in meta-analyses (MAs) on non-skeletal effects of vitamin D supplementation.
Methods and findings: We identified systematic reviews (SR) reporting summary data in terms of MAs of RCTs on selected non-skeletal outcomes. For each outcome, we summarized the results from available SRs and scrutinized included RCTs for a number of predefined characteristics. We identified 54 SRs including data from 210 RCTs. Most MAs as well as the individual RCTs reported null-findings on risk of cardiovascular diseases, type 2 diabetes, weight-loss, and malignant diseases. Beneficial effects of vitamin D supplementation was reported in 1 of 4 MAs on depression, 2 of 9 MAs on blood pressure, 3 of 7 MAs on respiratory tract infections, and 8 of 12 MAs on mortality. Most RCTs have primarily been performed to determine skeletal outcomes, whereas non-skeletal effects have been assessed as secondary outcomes. Only one-third of the RCTs had low level of 25OHD as a criterion for inclusion and a mean baseline 25OHD level below 50 nmol/L was only present in less than half of the analyses.
Conclusions: Published RCTs have mostly been performed in populations without low 25OHD levels. The fact that most MAs on results from RCTs did not show a beneficial effect does not disprove the hypothesis suggested by observational findings on adverse health outcomes of low 25OHD levels.