Skip to main content

Table 3 Predictive value for GDM using traditional and novel biomarkers identified in our cohort

From: Exploring preconception signatures of metabolites in mothers with gestational diabetes mellitus using a non-targeted approach

Models

R2

AUC

P value

Ref (model 1)

Ref (model 2)

Ref (model 3)

Metabolites prediction model

 Model 1, adjusting for 36:4, 38:6

0.127

0.694

N/A

N/A

N/A

 Model 2, adjusting for eight metabolites (34:1, 34:2, 36:2, 36:4, 38:4, 38:5, 38:6, and 40:6)

0.194

0.754

0.554

N/A

N/A

Traditional risks prediction model

 Model 3, adjusting for maternal age, ppBMI, family history of GDM, prior history of GDM, preconception prediabetes, parity

0.048

0.620

0.382

0.612

N/A

Metabolites and traditional risks combined prediction model

 Model 4, model 1 + model 3

0.236

0.773

0.360

0.738

0.480

 Model 5, model 2 + model 3

0.377

0.843

0.433

0.257

0.300

  1. Abbreviations: GDM Gestational diabetes mellitus, ppBMI pre-pregnancy body mass index, AUC Area under the curve