Screening accuracy of the 50g-glucose challenge test in twin compared with singleton pregnancies

Context — The optimal 50g-glucose challenge test (GCT) cut-off for the diagnosis of gestational diabetes (GDM) in twin pregnancies is unknown.

Objective — To explore the screening accuracy of the 50g-GCT and its correlation with the risk of large for gestational age (LGA) newborn in twin compared to singleton pregnancies.

Design — Population-based retrospective cohort study (2007-2017).

Setting — Ontario, Canada

Population — Patients with a singleton (n=546,892[98.4%]) or twin (n=8,832[1.6%]) birth who underwent screening for GDM using the 50g-GCT.

Methods — We compared the screening accuracy, risk of GDM, and risk of LGA between twin and singleton pregnancies using various 50g-GCT cut-offs.

Results — For any given 50g-GCT result, the probability of GDM was higher (p=0.0067) while the probability of LGA was considerably lower in the twin compared with the singleton group, even when a twin-specific growth chart was used to diagnose LGA in the twin group (p<0.001). The estimated false positive rate (FPR) for GDM was higher in twin compared with singleton pregnancies irrespective of the 50g-GCT cut-off used. The cut-off of 8.2 mmol/L (148 mg/dL) in twin pregnancies was associated with an estimated FPR (10.7-11.1%) that was similar to the FPR associated with the cut-off of 7.8 mmol/L (140 mg/dL) in singleton pregnancies (10.8%).

Conclusions — The screening performance of the 50g-GCT for GDM and its correlation with LGA differ between twin and singleton pregnancies.