Correction

This article refers to:

Vitamin D as a prognostic biomarker in COVID-19: single-center study and meta-analyses

Article title: Vitamin D as a prognostic biomarker in COVID-19: single-center study and meta-analyses

Authors: Vasilije Novakovic, Thomas Benfield, Henrik Løvendahl Jørgensen and Nikki Have Mitchell

Journal: Scandinavian Journal of Clinical and Laboratory Investigation

Bibliometrics: Volume 83, Number 3, pages 173–182

DOI: https://doi.org/10.1080/00365513.2023.2191333

When the article was first published online, Table 2 and Supplementary Table B were published with incorrect values. Now, these are updated as below and republished online.

Table 2. Odds of death within 30 days of hospital admission for COVID-19, based on logistic regression analysis.

	OR (95% CI) for death at 30 days given vitamin D sufficiency vs. vitamin D deficiency	OR (95% CI) for death at 30 days given age ≥70 years vs. <70 years	OR (95% CI) for death at 30 days given male sex vs. female sex	OR (95% CI) for death at 30 days given present comorbidity vs. no comorbidity	OR (95% CI) for death at 30 days given BMI ≥30 vs. <30
Unadjusted	1.84 (1.10–3.07)	–	–	–	–
Adjusted for age	1.35 (0.79–2.30)	3.59 (2.25–5.74)	–	–	–
Adjusted for age and sex	1.41 (0.82–2.42)	3.68 (2.30–5.89)	1.46 (0.98–2.17)	–	–
Adjusted for age, sex, comorbidity and BMI^a	1.51 (0.85–2.66)	3.02 (1.85–4.94)	1.31 (0.86–1.99)	1.85 (0.23–14.79)	1.17 (0.74–1.87)

OR: odds ratio; CI: confidence interval.

The logistic regression model was adjusted in a stepwise manner, adding an extra covariate to the analysis in each adjustment step.

This was done in the following order:

Vitamin D sufficiency (≥50 nmol/L) compared to vitamin D deficiency (<50 nmol/L), age group (≥70 years compared to <70 years), sex (male compared to female), comorbidity (present compared to absent) and BMI (≥30 compared to <30).

Adjustment for comorbidity and BMI were done in the same step.

The OR of the association of each covariate is added to the table as the covariate is introduced in the analysis.

^a One hundred and two individuals had missing data on weight and/or height, so n = 583.

Supplementary Table B. Odds of death within 30 days of hospital admission for COVID-19, based on logistic regression analysis.

	OR (95% CI) for death at 30 days given vitamin D level 50–74 nmol/L vs. vitamin D level <50 nmol/L	OR (95% CI) for death at 30 days given vitamin D level 75–99 nmol/L vs. vitamin D level <50 nmol/L	OR (95% CI) for death at 30 days given vitamin D level ≥100 nmol/L vs. vitamin D level <50 nmol/L	OR (95% CI) for death at 30 days given age ≥70 years vs. <70 years	OR (95% CI) for death at 30 days given male sex vs. female sex	OR (95% CI) for death at 30 days given present comorbidity vs. no comorbidity	OR (95% CI) for death at 30 days given BMI ≥30 vs. <30
Unadjusted	1.53 (0.84–2.76)	1.73 (0.95–3.15)	2.42 (1.34–4.37)	–	–	–	–
Adjusted for age	1.25 (0.68–2.30)	1.30 (0.70–2.42)	1.52 (0.82–2.83)	3.49 (2.17–5.62)	–	–	–
Adjusted for age and sex	1.27 (0.69–2.33)	1.38 (0.74–2.58)	1.68 (0.90–3.16)	3.55 (2.20–5.71)	1.51 (1.01–2.26)	–	–
Adjusted for age, sex, comorbidity and BMI^a	1.45 (0.76–2.74)	1.41 (0.72–2.74)	1.71 (0.88–3.34)	2.94 (1.79–4.85)	1.33 (0.87–2.04)	1.92 (0.24–15.34)	1.17 (0.74–1.87)

OR: odds ratio; CI: confidence interval.

The logistic regression model was adjusted in a stepwise manner, adding an extra covariate to the analysis in each adjustment step.

This was done in the following order:

Vitamin D level of either 50–74 nmol/L, 75–99 nmol/L or ≥100 nmol/L compared to vitamin D deficiency (<50 nmol/L), age group (≥70 years compared to <70 years), sex (male compared to female), comorbidity (present compared to absent) and BMI (≥30 compared to <30).

Adjustment for comorbidity and BMI were done in the same step.

The OR of the association of each covariate is added to the table as the covariate is introduced in the analysis.

^a One hundred and two individuals had missing data on weight and/or height, so n = 583.

In the article, it is stated that all measurements of vitamin D were taken before hospital admission of the COVID-19-patients. However, the vitamin D value used for each included patient was calculated as the mean of all available measurements for the given patient, regardless of whether the measurements were taken before or after admission. Thus, stating that only vitamin D measurements from before hospital admission were used is incorrect.

Further, in Table 2 and Supplementary Table B, the OR estimates relating to the covariates ‘BMI’ and ‘Comorbidity’ have swapped places. Corrected tables have been provided.