References Study Location/Cohort	Time frame	Population	Main findings	Possible limitations
C. BLOOD 25(OH)D MEASURED BEFORE INFECTION
Chodick et al. (224) Israel	From 1st Jan 2020	- n = 14,520, including: - n = 1,317 SARS-CoV-2+ - n = 13,203 SARS-CoV-2–	- 25(OH)D levels in those with (23.6 ± 8.6 ng/mL, mean ± SD) SARS-CoV-2 infection, similar to those without (24.1 ± 9.1 ng/mL)	- Lack of clarity on 25(OH)D assay
Hastie et al. (225, 226) UK Biobank (n = 502,624)	16th March−14th April 2020	n = 449 adults with confirmed COVID-19 and 25(OH)D test (2006–2010) of 348,598 eligible participants	- 25(OH)D not significantly associated with confirmed COVID-19 (OR 1.00, 95% CI 0.998–1.01, p = 0.208) after adjustment for ethnicity, age, sex, month of assessment, income, BMI, comorbidities	- 25(OH)D measured some time before COVID-19 pandemic
Hastie et al. (227)UK Biobank (n = 502,624)	5th March−25th April 2020	n = 656 adults with confirmed COVID-19 and 25(OH)D test (2006–2010) with n = 203 deaths of 341,484 eligible participants	- Severe infection and mortality (HR 0.98, 95% CI 0.91–1.06, p = 0.696) not significantly associated with 25(OH)D after adjustment ethnicity, age, sex, month of assessment, income, BMI, comorbidities	- Possible “over-adjustment” as BMI and ethnicity both may affect 25(OH)D [commentary from (228)]
Kaufman et al (229) USA	9th March−19th June 2020	From n = 218,372 tested for SARS-CoV-2	- Significant negative relationship between lower rates of SARS-CoV-2 positivity and higher blood 25(OH)D levels (OR 0.984 for every 1 ng/ml increase in 25(OH)D, 95% CI 0.983–0.986, p <0.001) following adjustment for latitude, ethnicity, gender and age (with 25(OH)D levels seasonally adjusted)	- No consideration of co-morbidities
Meltzer et al., (230) Chicago, USA	3rd March−10th April 2020	n = 489 patients tested for COVID-19 who had their 25(OH)D levels tested in the last year (prior to testing positive for SARS-CoV-2)	- Testing positive for COVID-19 (n = 71) was significantly associated with increased risk (RR 1.8 95% CI 1.1–2.1, p = 0.02) for being vitamin D deficient [25(OH)D <20 ng/ml] in multivariable analysis	- Uncertainty about 25(OH)D assay
Raisi-Estabragh et al., (231) UK Biobank (n = 497,996)	16th March−18th May 2020	n = 1,326 with positive COVID-19 test and n = 3,184 with negative COVID-19 test all with blood 25(OH)D test (2006–2010)	- No significant association (OR 1.00 95% CI 1.00–1.00, p = 0.72) between seasonally adjusted 25(OH)D levels and COVID-19 positivity in a model that also considered sex, age, and BAME ethnicity	- 25(OH)D measured some time before COVID-19 pandemic - Observed significant associations for BAME ethnicity OR 1.8, 95% CI 1.4–2.2, p = 9.27 ×10−7)

Identified in PubMed as of 18th November 2020.

^aMean 25(OH)D levels reported for each country, ranging in year of measurement.

Significance defined as p <0.05, 25(OHD = 25-hydroxyvitamin D.

CI, confidence interval; CRP, c-reactive protein; HR, hazard ratio; OR, odds ratio; RR, relative risk.

BAME, black, Asian and minority ethnic; 25(OH)D, 25-hydroxyvitamin D; CI, confidence interval; HR, hazard ratio; OR, odds ratio; RR, relative risk.