One-hour glucose tolerance test outperforms two-hour test for predicting progression to diabetes
Machine learning was used to combine different variables, including clinical risk factors, plasma glucose and insulin levels, HbA1c level, and six metabolic markers, in 1,527 models for predicting progression risk.
A one-hour oral glucose tolerance test, alone or in combination with other metabolic variables, is better at predicting risk for progression to type 2 diabetes than the two-hour version, according to a recent study.
Researchers used data from a study in Finland to evaluate the performance of a variety of variables for helping to predict type 2 diabetes risk during a 75-g oral glucose tolerance test, where plasma glucose and insulin levels were measured at 0, 30, 60, and 120 minutes. The authors used machine learning to combine different variables, including clinical risk factors, plasma glucose and insulin levels, HbA1c level, and six metabolic markers, and then test 1,527 models for predicting progression risk. The results were published Nov. 15 by the Journal of Clinical Endocrinology & Metabolism.
Five hundred forty-three patients who took an oral glucose tolerance test and had serum data available were included in the study. Of this group, 146 progressed to type 2 diabetes over the 10-year follow-up period. A randomly chosen control sample was made up of 397 patients from the Finland study who did not progress to diabetes. In the predictive models, one-hour plasma glucose level outperformed all other individual markers, including fasting plasma glucose level and two-hour plasma glucose level. The 30-minute plasma glucose level, HbA1c level, and mannose were statistically comparable to one-hour plasma glucose level but had lower area under the curve and lower sensitivity. While HbA1c level was inferior to the one-hour plasma glucose level when used alone, the predictive models improved significantly when metabolic markers were combined with plasma glucose levels and HbA1c levels. Of the metabolic markers, mannose was found to be a robust predictor of progression.
The authors noted that impaired fasting glucose and impaired glucose tolerance status at baseline may have affected their results and that their findings should be evaluated in an independent study. Despite these limitations, they concluded that the one-hour plasma glucose level is effective for predicting future progression to type 2 diabetes, alone or in combination with other metabolic variables. “Shortening the standard 75-gram [oral glucose tolerance test] to one hour improves the predictive value as well as clinical usability,” the authors wrote. “International diabetes organizations should consider this recommendation as the 1-hr [plasma glucose] more accurately identifies individuals with dysglycaemia, critical in context of the burgeoning global population with obesity and glucose disorders.”