Skip to main content

Table 3 Interaction between serum GDF15 concentrations upon ICU admission and randomization to early PN versus late PN for impact on clinical outcome

From: The clinical potential of GDF15 as a “ready-to-feed indicator” for critically ill adults

Outcome

Likelihood of earlier live discharge from ICU

HR (95% CI)

p

1A. GDF15 (per ng/ml added)

0.957 (0.947–0.967)

< 0.0001

1B. Randomization to early PN vs late PN

0.939 (0.883–0.998)

0.04

2. GDF15, randomization and interaction

 GDF15 (per ng/ml added)

 

< 0.0001

 Randomization to early PN vs late PN

 

0.02

 Interaction GDF15-randomization

 

0.21

Risk of new infection

OR (95% CI)

p

1A. GDF15 (per ng/ml added)

1.014 (1.007–1.021)

0.0001

1B. Randomization to early PN vs late PN

1.245 (1.071–1.447)

0.004

2. GDF15, randomization and interaction

 GDF15 (per ng/ml added)

 

< 0.0001

 Randomization to early PN vs late PN

 

0.003

 Interaction GDF15-randomization

 

0.46

Risk of ICU acquired weakness

OR (95% CI)

p

1A. GDF15 (per ng/ml added)

1.026 (1.006–1.045)

0.006

1B. Randomization to early PN vs late PN

1.457 (1.001–2.122)

0.04

2. GDF15, randomization and interaction

 GDF15 (per ng/ml added)

 

0.002

 Randomization to early PN vs late PN

 

0.03

 Interaction GDF15-randomization

 

0.19

  1. Stepwise multivariable models were built including all patients with upon ICU admission serum sample available for GDF15 measurement (N = 4393). Step 1: Models investigating the independent effect of GDF15 concentrations upon ICU admission (step 1A) and randomization to early PN vs late PN (step 1B) on outcome. All models were adjusted for the baseline risk factors age, body mass index, severity of illness, diagnostic category (cardiac surgery, emergency surgical, elective surgical and medical ICU admission), and risk of malnutrition. Step 2: Both GDF15 concentrations upon ICU admission and randomization to early PN versus late PN, as well as the interaction between randomization and GDF15 were included in the model, adjusting for the same baseline risk factors. This allowed to investigate whether upon-admission GDF15 was determining benefit or harm from early PN versus late PN
  2. Abbreviations: ICU intensive care unit, OR odds ratio, CI confidence interval, GDF15 growth-differentiation factor-15, PN parenteral nutrition, HR hazard ratio