Kaufman · PloS one 2020 · retrospective cohort study · n=191,779

SARS-CoV-2 positivity rates associated with circulating 25-hydroxyvitamin D levels.

Cited 396 times in the scientific literature.

Level 3 - non-randomized controlled study

Retrospective observational cohort study using clinical laboratory records.

PubMed 32941512 · doi:10.1371/journal.pone.0239252 · record verified 2026-08-30

What was done

A retrospective observational study evaluated deidentified laboratory data from 191,779 patients across all 50 US states tested for SARS-CoV-2 between mid-March and mid-June 2020 who also had a 25-hydroxyvitamin D (25(OH)D) test result within the prior 12 months. Residential zip codes were linked to US Census data to adjust for race/ethnicity proportions and latitude. Weighted polynomial and multivariable logistic regressions were used to assess the association between seasonally adjusted 25(OH)D levels and SARS-CoV-2 positivity.

What was found

The overall SARS-CoV-2 positivity rate was 9.3% (95% CI, 9.2–9.5%) and mean seasonally adjusted 25(OH)D was 31.7 ng/mL (SD 11.7). Positivity was 12.5% (95% CI, 12.2–12.8%) in 39,190 patients with deficient 25(OH)D (<20 ng/mL), 8.1% (95% CI, 7.8–8.4%) in 27,870 patients with adequate levels (30–34 ng/mL), and 5.9% (95% CI, 5.5–6.4%) in 12,321 patients with levels ≥55 ng/mL. After adjusting for demographic factors, higher 25(OH)D was associated with lower positivity (adjusted OR 0.984 per ng/mL increase, 95% CI, 0.983–0.986; p < 0.001).

Why it matters

This study demonstrates a consistent, dose-dependent inverse correlation between pre-existing serum vitamin D levels and SARS-CoV-2 test positivity in a large national cohort, reinforcing the need for prospective interventional trials.

Limits

The retrospective observational design cannot establish causality. Vitamin D levels were measured up to 12 months prior to SARS-CoV-2 testing and may not reflect status at infection. Clinical confounders like BMI, comorbidities, individual socioeconomic status, and behavioral exposure risks were not captured.

Cited by