368 - Association between beta cell function measures and risk of medically indicated preterm birth: mediation effects by hypertensive disorders of pregnancy in Japan (2015-2021)

Publication Number: 368.3788

Richard Liang, Stanford University School of Medicine, Stanford, CA, United States; David Rehkopf, Stanford University School of Medicine, Palo Alto, CA, United States; Gary Shaw, Stanford University School of Medicine, stanford, CA, United States; Akiko Sankoda, National Center for Child Health and Development, Setagaya, Tokyo, Japan; Asako Mito, National Center for Child Health and Development, TOKYO, Tokyo, Japan; Naoko Arata, National Center for Child Health and Development, Tokyo, Tokyo, Japan

Background: Preterm birth is one of the leading causes of infant morbidity and mortality, yet the etiology is still poorly understood. Pancreatic beta cell function, or the body’s insulin response at a given insulin resistance, has recently been linked with increased risk of preterm birth. However, it is unknown whether hypertensive disorders of pregnancy (HDP) contribute to the association between beta cell function and medically indicated preterm birth (mPTB).

Objective: To examine the mediating role of HDP in the association between maternal beta cell function and risk of mPTB.

Design/Methods: We analyzed 2,176 singleton live births from the Diabetes and Pregnancy Outcome for Mother and Baby (DREAMBee) Study, a prospective, multicenter cohort recruited from 2015-2021 across tertiary hospitals in Japan. Pregnant women with a positive non-fasting 50-gram glucose challenge test (≥140 mg/dL) were enrolled and received a fasting 75-gram oral glucose tolerance test. Blood glucose and insulin levels were measured at the fasting, 1-hour, and 2-hour time points. These levels were used to calculate disposition index (insulin secretion x resistance) as the measure of beta cell function.

The exposure variable was a Z-standardized disposition index. The outcome was mPTB, defined as induced labor or cesarean delivery before 37 weeks of gestation. We conducted mediation analysis using Poisson regression and marginal structural models to parse out the total effect rate ratio (RR) into the (1) direct effect of disposition index on risk of mPTB and (2) indirect effect mediated through HDP (dichotomous variable). The models accounted for maternal age, parity, body mass index, education, and income. As sensitivity analyses, we examined Z-standardized measures of insulin secretion (Stumvoll First-Phase Estimate) and resistance (-1 x Matsuda index).

Results: Poor beta cell function, as measured by the disposition index, was associated with an increased risk of mPTB (total effect RR: 1.18 [95%CI: 0.96-1.45]). The direct effect RR was 1.17 (95%CI: 0.97-1.42), while the indirect effect RR mediated through HDP was 1.01 (95%CI: 1.00-1.01). The resulting proportion mediated through HDP was estimated to be 6%. The RR effect sizes for insulin secretion and resistance were not substantive.

Conclusion(s): HDP only partially mediates the association between poor beta cell function and risk of mPTB. Other factors may contribute more substantially to the etiology between beta cell function and mPTB.

Mediation analysis on risk of medically indicated preterm birth (mPTB) through hypertensive disorders of pregnancy (HDP)
mPTB = medically indicated preterm birth, HDP = hypertensive disorders of pregnancy, RR = rate ratio, CI = confidence interval; (a) Simplified proposed pathway; (b) Annotated pathway showing difference between direct effect and indirect effect mediated through HDP; (c) Mediation analysis for each type of insulin measurement. Each RR is for a 1-standard deviation unit increase from the mean. Models accounted for maternal age, parity, body mass index, education, and income.