Table 3 Binomial regression models output for detecting the best predictors of the binomial (two categories: Resp vs. TRD) study group variable.

From: Whole-blood expression of inflammasome- and glucocorticoid-related mRNAs correctly separates treatment-resistant depressed patients from drug-free and responsive patients in the BIODEP study

Logistic models Explanatory variables Likelihood ratio test Negelkerke’s pseudo-R2
Chi2 (degree of freedom) P value
Mod. (i) Trait-anxiety 23.9 (1) <0.001 0.53
State-anxiety 0.4 (1) 0.533
CRP 0.2 (1) 0.961
Neutrophils absolute 5.9 (1) 0.015
Total white cells 0.3 (1) 0.601
Total score CTQ 0.2 (1) 0.727
Mod. (ii) CXCL12 4.0 (1) 0.038 0.89
CCL2 4.9 (1) 0.023
IL-1beta 3.8 (1) 0.048
IL-6 3.6 (1) 0.037
GR 18.4 (1) <0.001
P2RX7 11.5 (1) 0.003
SGK1 2.2 (1) 0.125
TNF-alpha 3.7 (1) 0.042
FKBP5 4.5 (1) 0.004
A2M 2.1 (1) 0.076
MIF 6.1 (1) 0.018
STAT1 5.6 (1) 0.009
CRP 2.8 (1) 0.086
Mod. (iii) # GR 5.7 (1) 0.017 0.90
P2RX7 14.0 (1) <0.001
TNF-alpha 4.1 (1) 0.040
Trait-anxiety 3.9 (1) 0.051
IL-6 4.2 (1) 0.042
CCL2 3.8 (1) 0.053
IL-1beta 6.6 (1) 0.010
CXCL12 5.7 (1) 0.031
Neutrophils absolute 1.2 (1) 0.277
FKBP5 2.4 (1) 0.124
MIF 2.5 (1) 0.113
STAT1 1.4 (1) 0.235
  1. #Explanatory variables of the model (iii) were standardised in order to take into account the different variable ranges.
  2. Mod. (i) considering only (significantly different between group) clinical and blood immune variables; mod. (ii) considering only (significantly different between group) genes variables and mod. (iii) considering both genes and clinical-blood immune variables resulted remained significant in mod. (i) and (ii).