Skip to main content
. 2021 Nov 26;11:161. doi: 10.1186/s13613-021-00952-z

Table 1.

Multivariate linear regression analysis output for variables independently associated with IgA, IgG and IgM on day 1

Variable B 95% CI p
IgA
 Female sex − 34.1 − 47.5: − 20.7 7.0 × 10–7
 BMI 2.6 1.4: 3.8 1.5 × 10–5
 White blood cell count at baseline 1.4 0.7: 2.1 6.0 × 10–5
 CV comorbidities 28.6 11.2: 46.1 0.001
 Randomized to crystalloids 19.4 6.3: 32.5 0.004
 Onco-haematologic disease − 48.0 − 87.0: − 9.1 0.016
 Creatinine − 3.5 − 7.6: 0.6 0.090
 Immunocompromised state − 18.3 − 40.0: 3.3 0.097
IgG
 Female sex − 88.5 − 135.5: − 41.4 2.4 × 10–4
 White blood cell count at baseline 4.2 1.7: 6.6 0.001
 Randomized to crystalloids 73.7 27.5: 120.0 0.002
 BMI 5.8 1.7: 10.0 0.006
 Shock − 55.1 − 101.8: − 8.4 0.021
 Liver disease 188.5 − 3.5: 380.4 0.054
 Heart rate − 0.9 − 2.0: 0.3 0.127
 Immunocompromised state − 9.3 − 79.1: 60.6 0.794
IgM
 Age − 0.6 − 0.9: − 0.2 0.002
 Shock − 14.4 − 25.3: − 3.6 0.009
 Onco-haematologic disease − 30.1 − 58.2: − 2.0 0.036
 White blood cell count at baseline 0.4 − 0.1: 1.0 0.127
 Mean arterial pressure 0.2 − 0.2: 0.5 0.329

Output for multivariable linear regression analysis with continuous Ig concentration as dependent variable

Included in the regression analyses are the variables which were significant in Additional file 1: Tables S1, S2 and S3. For IgG, AIDS was not included due to confounding by immuno-compromised state. Location of infection was not included in this analysis