Study | Design/setting | Population | Exposure/intervention | Outcomes | Results | Limitations |
Serum vitamin D | ||||||
D’Avolio et al73 | Cross-sectional study Canton of Tessin, Switzerland | 107 patients with data on SARS-CoV-2 and 25(OH)D measurement | Vitamin D analysis, conducted within 7 weeks of the SARS-CoV-2 PCR result. Control patients with 25(OH)D data during the same period. | SARS-CoV-2 infection | Group 1 comprised 27 patients with positive PCR test results for SARS-CoV-2, while group 2 comprised 80 patients with a negative PCR result for SARS-CoV-2. Significantly lower 25(OH)D levels (p=0.004) in SARS-CoV-2 patients even after stratifying patients according to age >70 years. | Few patients from a single hospital. No available clinical information about the severity of COVID-19 symptoms. No data on other potential confounding variable SARS-CoV-2 and the 25(OH)D status were performed on different days. |
Hastie et al70 | Retrospective cohort study UK Biobank Cohort including England, Scotland and Wales | 502 624 participants aged 37–73 years between 2006 and 2010 | Biochemical assay of 25(OH)D, a measure of vitamin D status. Vitamin D was imputed if it was below or above the limit of detection. | Confirmed COVID-19 (at least one positive test result) | Complete data on 348 598 UK Biobank participants 449 had confirmed COVID-19. Of these, 385 (85.8%) were white compared with 64 (14.2%) non-white (black, South Asian and others). Vitamin D was associated with COVID-19 univariably but not after adjustment for confounders. Ethnicity was associated with COVID-19. | UK Biobank is not representative of the general population. Baseline measurements, including 25(OH)D concentration and health status, were obtained a decade prior to conduct of the study. |
Ilie et al71 | Ecological study 20 European countries | Population of 20 included European countries | Mean levels of vitamin D in each country | Cases of COVID-19 per 1 million population in each country. Deaths from COVID-19 per 1 million population. | Negative correlations between mean levels of vitamin D and the number of COVID-19 cases per 1 million, and mortality per 1 million. | The number of cases per country is affected by the number of tests performed and by the different measures taken by each country to prevent the spread of infection. |
Vitamin D supplementation | ||||||
Fasano et al74 | Case–control survey. A single tertiary centre in Lombardy, Italy | 1486 Parkinson’s disease (PD) patients were included in the survey 1207 family members (controls) | Vitamin D | ‘Confirmed’ or ‘probable’ diagnosis of COVID-19 | 12.4% of PD patients with confirmed or probable COVID-19 had been taking vitamin D. 22.9% of PD patients without COVID-19 had been taking vitamin D. | Well-known limitation of a telephone survey. Community-dwelling PD patients. Some patients could not be reached possibly due to death from COVID-19. COVID-19 diagnosis could not be confirmed in many cases. Younger age of non-PD COVID-19 cases. |