Table 3

Mendelian randomisation (MR) estimates of effect of vitamin D on COVID-19 outcomes

MR methodCOVID versus populationHospitalised COVID versus populationSevere respiratory COVID versus populationCOVID positive versus
COVID negative
Hospitalised COVID versus
non-hospitalised COVID
OR95% CIOR95% CIOR95% CIOR95% CIOR95% CI
Instrument A
IVW1.040.92 to to 1.310.960.64 to 1.431.150.99 to 1.351.440.75 to 2.78
MR-Egger intercept testP=0.57P=0.19P=0.14P=0.88P=0.14
Instrument B
IVW1.040.92 to to 1.441.010.71 to 1.431.150.98 to 1.341.420.86 to 2.35
MR-Egger intercept testP=0.62P=0.15P=0.06P=0.61P=0.14
Instrument C
IVW1.020.90 to to 1.371.050.82 to 1.341.050.92 to to 1.77
MR-Egger intercept testP=0.88P=0.85P=0.23P=0.12P=0.14
Instrument of risk of vitamin D deficiency
IVW1.000.99 to to to 1.050.990.97 to 1.000.970.90 to 1.05
MR-Egger intercept testP=0.95P=0.61P=0.92P=0.46P=0.21
Instrument of risk of vitamin D insufficiency
IVW1.000.99 to to to 1.050.990.98 to 1.000.970.90 to 1.05
MR-Egger intercept testP=0.87P=0.64P=0.76P=0.59P=0.22
  • OR representing change in risk of outcome per SD increase in log-transformed serum vitamin D (instruments A, B and C) or difference in risk of outcome for individuals at risk of vitamin D deficiency/insufficiency versus individuals with sufficient vitamin D.

  • IVW, inverse variance weighted.