The Association Between Varying Degrees of Hyperglycemia During Pregnancy and Adverse Short- and Long-Term Outcomes

Abstract

Background: Women with gestational diabetes mellitus (GDM) are at increased risk of adverse perinatal outcomes and long-term outcomes, including type 2 diabetes mellitus (T2DM) and cardiovascular diseases (CVD). However, evidence on the association between GDM and stillbirth is inconclusive, and recent studies suggest that women with mild hyperglycemia not meeting a clinical diagnosis of GDM may also confer increased risks of short- and long-term outcomes. Given the lack of universal screening and diagnostic criteria for GDM, relying solely on a GDM diagnosis may underestimate true risks. This thesis aimed to: (1) assess the association of varying degrees of hyperglycemia and risk of stillbirth through a systematic review and meta-analysis; (2) explore the impact of early- and late-onset GDM on late stillbirth using a retrospective cohort study; (3) evaluate the association of varying degrees of hyperglycemia and risk of diabetes via a systematic review and meta-analysis; and (4) examine the association of hyperglycemia during pregnancy and risk of T2DM in a nested case-control study.

Methods: Objectives 1 and 3 involved comprehensive searches in MEDLINE, EMBASE, and CINAHL. Meta-analyses were conducted to generate pooled odds ratios (ORs) and 95% confidence intervals (CIs) if the included studies were homogeneous. For objectives 2 and 4, the Cerner Real-World Dataset (CRWD), which comprises the United States electronic health record (EHR) data from 2000 to 2016, was used as the data source. Objective 2 included a cohort of women with a liveborn or stillborn surviving ≥ 28 weeks. The study outcome of late stillbirth was defined as death of a fetus occurred 28 weeks or more completed weeks of gestation. Maternal and clinical characteristics among women with early GDM (0-23 weeks), late GDM (24-28 weeks) and non-GDM were described. Logistic regression was performed to determine the association of early GDM and late GDM with the risk of late stillbirth. Objective 4, a nested case-control study was conducted among women with glucose challenge test (GCT), or oral glucose tolerance test (OGTT) measured during pregnancy. A case was defined as having T2DM after index delivery until the end of follow-up time, while a control was defined as not being identified as having T2DM during the same follow-up period. Optimal variable matching was used for matching controls to cases (5:1) without replacement based on maternal age at delivery (±1 year), birth year (±1 year), and gestational age at delivery (±1 week) of the index pregnancy. Patients were divided into four groups: normal GCT, normal OGTT but abnormal GCT, one abnormal value on OGTT, and two or more abnormal values on OGTT. For these groups with one abnormal value and these with two or more abnormal values on OGTT, further analyses were conducted after stratifying by fasting glucose status. Maternal and clinical characteristics were compared between cases and controls. Conditional logistic regression was used to derive the best estimate of the association between glucose level and risk of T2DM.

Results: For the study on the association between maternal hyperglycemia and risk of stillbirth, the systematic review found that pooled ORs for risk of stillbirth among women with at least two abnormal values on OGTT and women with just one abnormal value or normal OGTT but abnormal GCT under the Carpenter & Coustan (C&C) criteria were 3.65 (95% CI, 1.74, 7.64), and 1.29 (95% CI, 0.56, 2.98), respectively, as compared to women with normal GCT. For women with at least one abnormal value diagnosed by the International Association of Diabetes and Pregnancy Study Groups (IADPSG) criteria, the pooled OR for the risk of stillbirth was 1.19 (95% CI, 0.94, 1.49), as compared to women with normal glucose level. Our cohort study found that the adjusted ORs (aORs) for the risk of late stillbirth among women with early GDM and late GDM were 3.50 (95% CI, 2.85, 4.29) and 3.20 (95% CI, 2.71, 3.76), respectively, as compared to women without GDM.

For the study on the association between maternal hyperglycemia and future risk of diabetes, the systematic review found that the pooled ORs for the risk of diabetes among women with at least two abnormal values and one abnormal value on 100-g OGTT were 12.29 (95% CI, 8.36, 18.07), and 4.44 (95% CI, 2.71, 7.28), respectively, as compared to women with normal GCT or OGTT. The nested case-control study found that aORs for the risk of T2DM among women with normal OGTT but abnormal GCT, one abnormal value on OGTT, and those with at least two abnormal values on OGTT diagnosed by the National Diabetes Data Group (NDDG) criteria were 1.66 (95% CI,1.35, 2.04), 5.69 (95% CI, 4.34, 7.47), and 13.20 (95% CI, 10.52, 16.57), respectively. Whereas the aORs for the risk of T2DM among women with normal OGTT but abnormal GCT, one abnormal value, and these with at least two abnormal values diagnosed by the Carpenter & Coustan (C&C) criteria were 1.21 (95% CI, 0.94, 1.55), 2.62 (95% CI, 1.95, 3.51), and 11.69 (95% CI, 9.61, 14.23), respectively. For women with at least two abnormal values on OGTT by either C&C or NDDG criteria, abnormal fasting glucose imposed additional risk of T2DM.

Conclusion and Implications: Women with borderline hyperglycemia during pregnancy that did not meet current diagnostic criteria for GDM (i.e., one abnormal value on OGTT) in some jurisdictions may be associated with increased risks of stillbirth and (especially) diabetes, although the associations were not as strong and as consistently observed in women with universally accepted diagnosis (i.e., two or more abnormal values on OGTT) of GDM. Although no immediate medical intervention is needed, attention should be paid to women with borderline hyperglycemia during pregnancy for recommendations to intensify lifestyle changes to reduce their risk of adverse short- and long-term outcomes. Large scale studies assessing the association of fasting glucose status, along with the number of abnormal values, with adverse short- and long-term outcomes, should be conducted to confirm the preliminary findings from this thesis. Since universal screening policies are now in place in most countries/jurisdictions, prenatal care offers many young adult women opportunities of first contact with health care providers for screening of glucose levels. Pregnancy is a good time to detect women at increased risk of adverse short- and long-term outcomes, with targeted interventions to reduce the burden of the diseases for the affected women and their offspring.

This thesis work focused on stillbirth and T2DM, two of the most prominent adverse outcomes associated with hyperglycemia during pregnancy. As the next step, I plan to conduct a retrospective cohort study using large population health databases in Ontario, Canada to assess the association between maternal hyperglycemia and CVDs.