Skip to main content
. 2022 Jul 12;17(7):566–576. doi: 10.1089/bfm.2021.0292

Table 2.

Receiver Operator Curve Results for Select Metabolic Health Variables in Parallel with Gestational Diabetes Mellitus in Identifying Mothers with Persistent Severely Low Milk Output (<300 mL/24 Hours)

Metabolic health parameter AUC (95% CI) for model that includes GDM Se, true positive fraction at optimal cutoff Sp, true negative fraction at optimal cutoff Optimal Se/Sp cutoff in parallel with GDM Metabolic syndrome cutoffa
Metabolic syndrome severity z-scoreb 0.92 (0.83–1.00) 0.89 (16/18) 0.92 (22/24) ≥0.30 n/a
BMI 0.88 (0.77–0.99) 0.94 (17/18) 0.83 (20/24) ≥33.3 kg/m2 n/a
Waist circumference 0.89 (0.79–0.99) 0.89 (16/18) 0.79 (19/24) ≥104 cm ≥89 cm
Diastolic blood pressure 0.85 (0.74–0.97) 0.89 (16/18) 0.71 (17/24) ≥75 mm Hg ≥85 mm Hg
Systolic blood pressure 0.83 (0.70–0.96) 0.94 (17/18) 0.75 (18/24) ≥110 mm Hg ≥130 mm Hg
Plasma triglycerides 0.83 (0.71–0.95) 0.72 (13/18) 0.92 (22/24) ≥130 mg/dL ≥150 mg/dL
Plasma c-peptide 0.83 (0.70–0.95) 0.83 (15/18) 0.75 (18/24) ≥2.0 ng/mL n/a
HOMA-IRC-peptide 0.82 (0.72–0.95) 0.78 (14/18) 0.83 (20/24) ≥0.5 n/a
Plasma glucose 0.77 (0.62–0.92) 0.83 (15/18) 0.75 (18/24) ≥90 mg/dL ≥100 mg/dL
GDM 0.67 (0.55–0.80) 0.39 (7/18) 0.96 (23/24) Positive diagnosis n/a

Notes: Data are combined for the external control group and the moderate/normal nested control group (n = 24 total) and contrasted against the severely low milk output cases (n = 18). Optimal Se and Sp cutoffs were derived from models that combined each metabolic health parameter with GDM, where test positive is defined as either the cutoff or GDM being present (i.e., applying two diagnostic criteria in parallel). Results are ordered from highest AUC to lowest, based on the trapezoid rule and optimal Se/Sp is based on the shortest distance formula (SAS v9.4).

a

Metabolic syndrome cutoffs are based on the criteria used at NIHLB (https://www.nhlbi.nih.gov/health-topics/metabolic-syndrome).

b

Metabolic syndrome severity z-score, which is an algorithm where 0, >0, and <0 z-scores signify average, worse than average, and better than average metabolic health profiles, respectively, as compared with all U.S. adults aged 20–65 years.

AUC, area under the curve; CI, confidence interval; GDM, gestational diabetes mellitus; n/a, not applicable; Se, sensitivity; Sp, specificity.