The Association of Low Vitamin K Status with Mortality in a Cohort of 138 Hospitalized Patients with COVID-19

Linneberg, Allan; Kampmann, Freja Bach; Israelsen, Simone Bastrup; Andersen, Liv Rabøl; Jørgensen, Henrik Løvendahl; Sandholt, Håkon; Jørgensen, Niklas Rye; Thysen, Sanne Marie; Benfield, Thomas

It has recently been hypothesized that vitamin K could play a role in COVID-19. We aimed to test the hypotheses that low vitamin K status is a common characteristic of patients hospitalized with COVID-19 compared to population controls and that low vitamin K status predicts mortality in COVID-19 patients. In a cohort of 138 COVID-19 patients and 138 population controls, we measured plasma dephosphorylated-uncarboxylated Matrix Gla Protein (dp-ucMGP), which reflects the functional vitamin K status in peripheral tissue. Forty-three patients died within 90 days from admission. In patients, levels of dp-ucMGP differed significantly between survivors (mean 877; 95% CI: 778; 995) and non-survivors (mean 1445; 95% CI: 1148; 1820). Furthermore, levels of dp-ucMGP (pmol/L) were considerably higher in patients (mean 1022; 95% CI: 912; 1151) compared to controls (mean 509; 95% CI: 485; 540). Cox regression survival analysis showed that increasing levels of dp-ucMGP (reflecting low vitamin K status) were associated with higher mortality risk (sex- and age-adjusted hazard ratio per doubling of dp-ucMGP was 1.49, 95% CI: 1.03; 2.24). The association attenuated and became statistically insignificant after adjustment for co-morbidities (sex, age, CVD, diabetes, BMI, and eGFR adjusted hazard ratio per doubling of dp-ucMGP was 1.22, 95% CI: 0.82; 1.80). In conclusion, we found that low vitamin K status was associated with mortality in patients with COVID-19 in sex- and age-adjusted analyses, but not in analyses additionally adjusted for co-morbidities. Randomized clinical trials would be needed to clarify a potential role, if any, of vitamin K in the course of COVID-19.