Performance Characteristics of the 50 Gram Glucose Challenge Test Compared to the 75 Gram Oral Glucose Tolerance Test for Gestational Diabetes Screening at the Kenyatta National Hospital.

Abstract

Introduction: The International Diabetes Federation (IDF) in 2015 estimated that, 16.2% of livebirths were complicated with hyperglycaemia in pregnancy(1), 81.5% due to GDM(1). The prevalence of GDM at Kenyatta National Hospital (KNH) is estimated at 11.6-16.7%(2)(3).The International Association of Diabetes and Pregnancy Study Groups recommend a universal one-step 75-g oral glucose tolerance test (OGTT) as a screening strategy for GDM(4). The American College of Obstetricians and Gynaecology (ACOG) prefer a universal two-step screening strategy using a 50-g glucose challenge test (GCT) to determine candidates for an OGTT(5).There are insufficient studies on the various screening and diagnostic criteria and cut off for a GCT compared to OGTT especially in this setting. Objectives: to evaluate the performance characteristics (sensitivity, specificity, positive predictive value (PPV), negative predictive value (NPV) and likelihood ratios) and receiver- operator characteristic curve (ROC)) of the 50g GCT compared with 75g OGTT; assess the correlation between abnormal GCT and OGTT to Glycated hemoglobin (HbA1c) levels and determine the risk factors for GDM. Study setting: the study was conducted at Kenyatta National Hospital antenatal clinic and antenatal wards. Methodology: This was a hospital based cross-sectional study where 438 eligible antenatal clients, at 24-28 weeks of gestation without pre-gestational diabetes underwent a 50g GCT followed by a 75g OGTT within two weeks. Plasma glucose level of >7.8mmol/l on GCT was considered as positive test and used to estimate the performance characteristics of the GCT. The ROC curve was generated, area under the curve (AUC) calculated. The best threshold value of the GCT for detecting GDM was obtained from the ROC curve. The performance characteristics of the test was determined at different cut-offs. Positive samples on both the 50g GCT and 75g OGTT were also correlated to HbA1c levels. Risk factors for development of GDM were determined. P-value was significant at p<0.05. Results: Of the 438 patients enrolled, 100 (23%) had GDM based on a 75g OGTT. A total of 107 had a positive GCT at a cut off plasma glucose levels of >7.8mmol/l, out of whom 54 (50%) had abnormal OGTT. The sensitivity, specificity, PPV, NPV, PLR, NLR of the 50g GCT at a cut off of ≥7.8 mmol/l was 56%, 85%, 52%, 87%, 3.71 and 0.52 respectively. The ROC curve gave a GCT cut off of 7.5mmol/l and area under the curve of 0.7 hence sensitivity and specificity of 92% and 73% respectively. The HbA1c was positive (≥6.5%) in only 13 (13%) of the diabetic patients and 12 (11%) of GCT positive patients making it a poor screening tool for GDM. Glycosuria was associated with GDM in both univariate and multivariate analysis. Conclusion: The 50 gram GCT at a threshold of 7.8mmol/l is a good screening test in this setting due to its high specificity and high negative predictive value making it a useful test in excluding GDM. The sensitivity of the test at the 7.8mmol/l cut off was lower, however lowering the threshold to 7.5mmol/l raises the sensitivity but lowers the specificity of the test. Lowering the current recommended threshold of 7.8mmol/l will lead to unnecessarily performing an OGTT. HbA1C levels do not correlate with abnormal GCT and OGTT. Glycosuria was associated with increased risk of GDM. Recommendations: Universal screening for GDM with 50gram GCT is justified. 1-hour plasma glucose of >7.8mmol/l should be evaluated by 75g OGTT.