Response to Letter to the Editor From Göbl and Tura: “Oral Glucose Tolerance Test-based Measures of Insulin Secretory Response in Pregnancy”

We thank Drs. Gobl and Tura for their interest in our recent publication (1), in which we compared oral glucose tolerance test (OGTT)-based measures of insulin secretory response to first-phase insulin response measured by an intravenous glucose tolerance test (IVGTT) in the same pregnant participants. The motivation for this analysis was the observation that several past investigations have used OGTT-based measures of insulin secretory response in pregnancy studies (2-5), without data on how these measures correspond to gold-standard measures from intravenous glucose administration or hyperglycemic clamps. Our intent was to describe the degree to which more easily administered oral glucose challenges approximate the more laborious, “gold-standard” measures. Our study, to our knowledge, is the first to provide such data in pregnancy.

In their letter, Drs. Gobl and Tura suggest that the Stumvoll first-phase estimate and the ratio of the area under the insulin curve to the area under the glucose curve (AUC_ins/AUC_glu) may not be good approximations of the first-phase insulin response in pregnancy. They suggest that the R² between an OGTT measure of insulin secretory response and the corresponding IVGTT measure must be >0.50 for that measure to be considered valid. We respectfully contend that the threshold they suggest is arbitrary. Whereas the P < 0.05 criterion is accepted as the standard for statistical significance, there is no conventionally accepted standard for clinical significance, scientific importance, or effect size, which need to be addressed on a study-by-study basis, weighing pertinent substantive issues. Our data provide useful quantification of the degree to which OGTT-based measures correlate with the response to intravenous glucose; we reported the correlations, their corresponding P-values, and provided the associated scatterplots, leaving it to the reader to judge what correlations they consider to be substantial. Many of the correlations we report are r > 0.5 with P ≤ 0.005, which do not seem insubstantial. For investigators concerned about the threshold for variance explained in pregnancy, we suggest utilization of the AUC_ins/AUC_glu measure, which captures 62% of the variance in IVGTT-measured insulin secretory response in early gestation and 48% of the variance in late gestation, the highest of the measures we examined. While the writers’ assertion that OGTT-based insulin secretory response measures can never be validated directly in pregnancy is true in the purest sense, there is valuable methodologic information to be gained by rigorously comparing different measures of insulin secretory response to those that have been used in past canonical studies (6, 7). With regard to the clinical significance of these measures, previous studies have suggested that OGTT-based measures of insulin secretory response have the expected relationship with insulin resistance and gestational diabetes (2, 3). We are currently conducting prospective longitudinal studies that investigate the changes in insulin secretory response to an oral glucose load in pregnant women with and without gestational diabetes; these data may further demonstrate the clinical relevance of these measures.

Drs. Gobl and Tura correctly identify the incretin effect as an important determinate of insulin secretory response to oral glucose. There are limited data on the incretin effect in pregnancy, although some evidence suggests that it is diminished (8). We agree that it is likely that the incretin effect accounts, at least partially, for variation in the response between oral and intravenous glucose apparent in our analysis. We might also note that most observational study aims can be categorized into 1 of 2 objectives: “explanation” or “prediction.” The fact that the incretin effect may provide a mechanistic explanation for our observed imperfect correlations does not necessarily obviate the predictive validity of our findings. Furthermore, if the incretin effect uniformly affects the insulin response to oral glucose across gestation, the correlations we describe would be minimally affected by it. A longitudinal study of the incretin effect in pregnancy would fill a major gap in the literature. Based on the comments of Drs. Gobl and Tura, we hope to address the incretin effect in our ongoing prospective longitudinal studies of glycemic physiology that span preconception, pregnancy, and postpartum. We encourage others with relevant data to conduct studies of the incretin effect in this population. In the interim, our recent study serves a practical purpose: equipping investigators with the tools to study insulin secretory response in pregnancy using the simple and widely deployed OGTT.

Funding

This work was supported in part by NICHD HD22965-19 (P.M.C.) and NCRR CTSA Ul1 RR 024989. C.E.P. is supported by career development awards from the National Institute of Diabetes and Digestive and Kidney Diseases (NIDDK) (K23DK113218) and the Robert Wood Johnson Foundation’s Harold Amos Medical Faculty Development Program. The data analysis was supported by Harvard Catalyst/Harvard University’s Clinical and Translational Science Center (National Center for Advancing Translational Sciences, National Institutes of Health Award UL 1TR002541) and financial contributions from Harvard University and its affiliated academic healthcare centers.

Disclosures

The authors disclose no conflicts of interest.