Table 3 Binary logistic regression of ANOVA selected metabolite and existing markers in MPO immunised animals at day 210.

From: Changes in urinary metabolomic profile during relapsing renal vasculitis

Predictors B S.E. Wald df p OR Model Accuracy
MPO Control Total p
Haematuria 0.46 0.909 0.256 1 0.613 1.584 20 85.7 58.3 0.665
ACR −0.004 0.005 0.718 1 0.397 0.996
Haematuria 0.171 1.168 0.021 1 0.884 1.187 80 85.7 83.3 0.122
ACR −0.003 0.01 0.074 1 0.785 0.997
2-OG −13.29 8.447 2.475 1 0.116 0
2-OG −13.36 7.432 3.231 1 0.072 0 80 85.7 83.3 0.017
  1. A predictive model built with existing markers is poorly predictive (58.3% accuracy p = 0.665) of relapse status in EAV (top panel). Adding 2-oxoglutarate (2-OG) to the model improved prediction and p values slightly but not to significance (middle panel). Removal of the current biomarkers, leaving 2-oxoglutarate alone resulted in a statistically significant model predicting 83.3% of relapse status (lower panel). OR = Odds ratio.