Skip to main content
. 2020 Sep 14;24:557. doi: 10.1186/s13054-020-03254-1

Table 2.

Variables associating with individual GDF15 concentrations in multivariable linear regression analysis

Model with variables explaining serum GDF15 concentrations upon ICU admission Scaled estimate (95%CI) p value
ICU-independent variables
 Age 353 (− 528; 1234) 0.43
 BMI − 674 (− 2057; 709) 0.33
 History of diabetes mellitus 731 (314; 1147) 0.0006
 History of malignancy 485 (60; 909) 0.02
 NRS score of ≥ 5 vs < 5 424 (− 24; 872) 0.06
ICU-related variables
 Admission diagnosis
  Cardiac surgery 16 (− 761; 793) 0.96
  Elective other surgery 1566 (445; 2687) 0.006
  Emergency other surgery − 1256 (− 2013; − 499) 0.001
  Medical disease − 326 (− 1477; 824) 0.57
 APACHE-II score 4454 (3211; 5697) < 0.0001
 Sepsis upon admission 2232 (594; 3869) 0.007
 Infection upon admission − 1412 (− 3059; 234) 0.09
 Mechanical ventilation on ICU day 1 296 (− 452; 1044) 0.43
 Steroids on ICU day 1 703 (212; 1194) 0.005
Laboratory parameters
 Blood glucose upon ICU admission − 1975 (− 4405; 455) 0.11
 Plasma creatinine on ICU day 1 22,102 (19,679; 24,526) < 0.0001
 Plasma total bilirubin on ICU day 1 13,569 (10,097; 17,041) < 0.0001
 Plasma CRP on ICU day 1 − 1605 (− 3132; − 78) 0.03

Scaled estimates are shown for all variables included in the linear regression model built to explain individual GDF15 concentrations upon ICU admission. Age, BMI, APACHE-II score, and the laboratory parameters were added as continuous variables; all other variables were added as categorical variables. The R2 of the model was 0.23. A sensitivity analysis using BMI (< 18.5 kg/m2, 18.5–25 kg/m2, > 25 kg/m2) and age (18– < 30 years, 30– < 50 years, 50– < 70 years, 70– < 90 years, > 90 years) as categorical variables was performed to rule out non-significant associations in case of non-linear relationships, which also yielded insignificant associations for all categories (data not shown)

Abbreviations: GDF15 growth-differentiation factor-15, ICU intensive care unit, BMI body mass index, NRS nutritional risk screening, APACHE acute physiology and chronic health evaluation, CRP C-reactive protein