Sukanda Metheetrairut¹, Chontipha Gleebkaew¹

Objective: To establish the random plasma glucose (RPG) threshold for predicting gestational diabetes mellitus (GDM) in pregnant individuals with a gestational age of less than 24 weeks.

Materials and Methods: A prospective cohort study was conducted at Rajavithi Hospital between September 2022 and August 2023. Pregnant women with gestational ages below 24 weeks were recruited during their initial antenatal visit, and RPG and hemoglobin A1c (HbA1c) levels were measured. Participants with pregestational diabetes mellitus were excluded. All participants underwent screening for GDM using a 50-g glucose challenge test (GCT) followed by diagnosis via a 100-g oral glucose tolerance test (OGTT). Test accuracy was assessed using the area under the receiver operating characteristic curve (aROC).

Results: Eight hundred three pregnant women were analyzed, with a GDM prevalence of 8.5%. An RPG threshold of 88.5 mg/dL or greater demonstrated the highest clinical relevance, with a sensitivity of 45.6% and a specificity of 82.5% (AUC 0.64, 95% CI 0.56 to 0.72). A “combined GDM predictive score” was developed using multivariate logistic regression analysis, incorporating RPG, HbA1c, age, glucosuria, and body mass index (BMI). This composite score yielded a sensitivity of 60.3% and a specificity of 85.3% (AUC 0.79, 95% CI 0.72 to 0.85). Furthermore, the specificity of the “combined GDM predictive score” in a non-high-risk subgroup was notably high at 92.4% (AUC 0.73).

Conclusion: Although RPG alone was not found to be a good predictor of GDM, the “combined GDM predictive score” is a useful tool for ruling out GDM in non-high-risk pregnant women before 24 weeks of gestation.

Received 9 September 2024 | Revised 4 February 2025 | Accepted 7 February 2025

DOI: 10.35755/jmedassocthai.2025.3.191-196-01449