Comment on: Goldberg et al. Circadian Variation in the Response to the Glucose Challenge Test in Pregnancy: Implications for Screening for Gestational Diabetes Mellitus. Diabetes Care 2012;35:1578–1584

W e read with great interest the article by Goldberg et al. (1), in which the authors sought to further characterize the metabolic phenotype of the population of pregnant women with the positive glucose challenge test (GCT) test in relation to the time of day when the test was done. They found that women who had GCT done in the afternoon have a better metabolic function and a lower risk of subsequent gestational diabetes mellitus (GDM) on the consecutive oral glucose tolerance test (OGTT). In other words, the authors found a causal relationship between the way the GCT test is performed (time of day of the test) and the properties of the studied subjects (metabolic phenotype of pregnant women), which necessarily gives rise to doubts and calls for caution. One of the knownweaknesses of GCT screening is that its implementation does not require clear relationship to food intake (fasting, pre-, and postprandial phases, or postabsorption period), which creates very different metabolic conditions under which the test is done during the day. Then, applying a single threshold GCT $7.8 mmol/L would lead to the lower positive predictive value of the test in the afternoon, as has been previously described (2,3). In practice, this manifests so that during the daytime there will be an increasing proportion of healthy pregnant women, either with normal glucose tolerance (NGT) or gestational impaired glucose tolerance (GIGT), having the GCT test positive. Normal OGTT results of these women will be the reason for a “better”metabolic phenotype of the study group, which had GCT test in the afternoon. Diurnally rising proportions of NGT plus GIGT women within the study groups is shown in Fig. 2A (1). Thus, we may assume this would also result in a higher mean insulin sensitivity and an improved mean b-cell function in the group, which had GCT test in the afternoon. Finally, we think that although some of the ideas and facts presented in this study are of interest, its results cannot be fully accepted without performing an additional statistical adjustment for daytime dependent proportion of NGT plus GIGT women in the study groups. Most likely, following such statistical correction it would not be possible to confirm any special metabolic phenotype of pregnant women dependent upon day-timing of GCT.