The lower the vitamin D, the higher the COVID death rate

Throughout the pandemic, researchers repeatedly noticed that people with low blood vitamin D seemed to fare worse with COVID-19. That observation left an unresolved chicken-and-egg problem: does vitamin D deficiency genuinely weaken immune defence against the virus, or does severe infection simply drag vitamin D levels down as a side effect? The distinction matters enormously, because only the first version would suggest that maintaining good vitamin D status could protect anyone. This review set out to gather further evidence on which way the arrow points.

The authors ran a systematic literature search for retrospective cohort and clinical studies that reported COVID-19 mortality rates alongside blood vitamin D3 levels. Crucially, they restricted themselves to studies where vitamin D was measured either before infection or on the day of hospital admission, which helps separate cause from consequence. Mortality rates from the clinical studies were corrected for age, sex and diabetes. They identified one population study and seven clinical studies, then analysed the data using correlation and linear regression.

Both independent datasets showed a negative relationship between vitamin D levels and mortality risk, though neither quite reached conventional statistical significance on its own. When the data were combined, the negative correlation became significant. The median vitamin D level across the combined data was 23.2 ng/mL, which is well below the level the authors highlight. Extending the regression line suggested a theoretical point of zero mortality at approximately 50 ng/mL, though that figure sits beyond the range where most of the actual data lie.

The authors argue their findings support vitamin D deficiency being a predictor of poor COVID-19 outcomes rather than merely a side effect of infection, partly because preinfection levels also tracked mortality. That said, no randomised trial was included here, so the causal question is not settled. The recommendation to push blood levels above 50 ng/mL goes further than correlational data can strictly support, and the zero-mortality point is a theoretical extrapolation rather than an observed result. Trials of supplementation remain the open question this paper cannot answer.

This is a reason to think vitamin D status is more than a passive marker of illness, since levels measured before people ever caught the virus still tracked their risk of dying. It is not, however, evidence that taking vitamin D would protect you: that claim would need randomised trials, and none appear here. The most defensible reading is that low vitamin D identifies people at higher risk, and that knowing your own level is a piece of information worth having when weighing your overall health picture.