25-hydroxyvitamin D is a predictor of COVID-19 severity of hospitalized patients

Abstract

Objectives: Studies investigating the association between vitamin D and severity of COVID-19 have mixed results perhaps due to immunoassay assessment of total 25-hydroxyvitamin D (tD) (the sum of 25-hydroxyvitamin-D2 [25-OH-D2] and 25-hydroxyvitamin-D3 [25-OH-D3]). Liquid chromatography tandem mass spectrometry (LC-MS/MS) has high analytical specificity and sensitivity for 25-OH-D2 and 25-OH-D3, and thus enables a more accurate assessment of impact on COVID-19 outcomes.

Methods: We established reference intervals for 25-OH-D3 and tD using LC-MS/MS. 25-OH-D2, 25-OH-D3 and tD were quantitated for 88 COVID-19 positive and 122 COVID-19 negative specimens. Chi-square or Fisher's exact tests were used to test associations in binary variables. T-Tests or Wilcoxon rank sum tests were used for continuous variables. Cox proportional hazards were used to test associations between 25-OH-D3 or tD levels and length of stay (LOS). For mortality and ventilation, logistic regression models were used.

Results: COVID-19 patients with deficient (<20 ng/mL) levels of 25-OH-D3 had significantly longer LOS by 15.3 days. COVID-19 P patients with deficient (<20 ng/mL) and insufficient (<30 ng/mL) of tD had significantly longer LOS by 12.1 and 8.2 days, respectively. Patients with insufficient levels of tD had significantly longer LOS by 13.7 days. COVID-19 patients with deficient serum 25-OH-D3 levels had significantly increased risk-adjusted odds of in-hospital mortality (OR [95% CI]: 5.29 [1.53-18.24]); those with insufficient 25-OH-D3 had significantly increased risk for requiring ventilation during hospitalization was found at LCMS insufficient cutoff (OR [95% CI]: 2.75 [1.10-6.90]).

Conclusions: There is an inverse relationship of 25-hydroxyvitamin D levels and hospital LOS for COVID-19 patients. Vitamin D status is a predictor for severity of outcomes. LCMS results are useful for assessing the odds of mortality and the need for ventilation during hospitalization.

Fig 1. Linear plots and regression parameters.

Correlation of Cascadion tD results to A) VDSCP tD B) Architect tD of all patient samples C) Architect tD of samples contain detectable 25-OH-D2 and D) Architect tD of samples with undetectable 25-OH-D2.

Fig 2. Kaplan-Meier curves for hospital discharge.

Unadjusted Kaplan-Meier curves for hospital discharge by 25-OH-D3 and tD levels above and below cutoffs. For 25-OH-D3, there were 11 patients censored at the time of death with deficient levels and 10 with non-deficient levels. For tD, 8 and 13 patients were censored at time of death for deficient and non-deficient levels, respectively.