The associations of maternal liver biomarkers in early pregnancy with the risk of gestational diabetes mellitus: a prospective cohort study and Mendelian randomization analysis

Abstract

Background: Associations of liver function with the risk of gestational diabetes mellitus (GDM) remain unclear. This study aimed to examine the relationship and the potential causality between maternal liver biomarkers and the risk of subsequent GDM, as well as to evaluate the interaction between liver biomarkers and lipids on GDM risk.

Methods: In an ongoing Zhoushan Pregnant Women Cohort, pregnant women who finished the first prenatal follow-up record, underwent liver function tests in early pregnancy, and completed the GDM screening were included in this study. Logistic regression models were used to investigate the association, and the inverse-variance weighted method supplemented with other methods of two-sample Mendelian randomization (MR) analysis was applied to deduce the causality.

Results: Among 9,148 pregnant women, 1,668 (18.2%) developed GDM. In general, the highest quartile of liver function index (LFI), including ALT, AST, GGT, ALP, and hepatic steatosis index, was significantly associated with an increased risk of GDM (OR ranging from 1.29 to 3.15), especially an elevated risk of abnormal postprandial blood glucose level. Moreover, the causal link between ALT and GDM was confirmed by the MR analysis (OR=1.28, 95%CI:1.05-1.54). A significant interaction between AST/ALT and TG on GDM risk was observed (P _interaction = 0.026).

Conclusion: Elevated levels of LFI in early pregnancy were remarkably associated with an increased risk of GDM in our prospective cohort. Besides, a positive causal link between ALT and GDM was suggested.

Keywords: Mendelian randomization; gestational diabetes mellitus; interaction; liver biomarkers; prospective cohort study.

Figure 1

The associations of maternal liver biomarkers in early pregnancy with risk of GDM. The model was adjusted for age, maternal BMI in the first trimester, educational level, gravidity, parity, weight gain before OGTT, gestational age of measurement, smoking, and alcohol consumption. GDM, gestational diabetes mellitus; ALT, alanine aminotransferase; AST, aspartate aminotransferase; GGT, gamma-glutamyl transferase; ALP, alkaline phosphatase; HSI, hepatic steatosis index.

Figure 2

Causal association of genetically predicted liver biomarkers with the risk of GDM. MR, Mendelian randomization; SNP, single nucleotide polymorphism; MR-PRESSO, MR-pleiotropy residual sum and outlier; MVMR, multivariable MR; GDM, gestational diabetes mellitus; ALT, alanine aminotransferase; AST, aspartate aminotransferase; GGT, gamma-glutamyl transferase; ALP, alkaline phosphatase.

Figure 3

The associations of maternal liver biomarkers in early pregnancy with risk of the GDM subtypes. The model was adjusted for age, maternal BMI in the first trimester, educational level, gravidity, parity, weight gain before OGTT, gestational age of measurement, smoking, and alcohol consumption. GDM, gestational diabetes mellitus; ALT, alanine aminotransferase; AST, aspartate aminotransferase; GGT, gamma-glutamyl transferase; ALP, alkaline phosphatase; HSI, hepatic steatosis index; i-IFG, isolated impaired fasting glucose; i-IGT, isolated impaired glucose tolerance; b-GDM, both impaired fasting glucose and impaired glucose tolerance.

Figure 4

Associations of maternal liver biomarkers in early pregnancy with GDM risk stratified by overweight. (A) Associations of ALT in early pregnancy with GDM risk stratified by overweight; (B) Associations of AST in early pregnancy with GDM risk stratified by overweight; (C) Associations of GGT in early pregnancy with GDM risk stratified by overweight; (D) Associations of ALP in early pregnancy with GDM risk stratified by overweight; (E) Associations of AST/ALT ratio in early pregnancy with GDM risk stratified by overweight. The model was adjusted for age, maternal BMI in the first trimester, educational level, gravidity, parity, weight gain before OGTT, gestational age of measurement, smoking, and alcohol consumption. Based on their quartiles, live enzymes were divided into four categories (Q1-Q4). ALT was divided into 1~10,11~13,14~19, and 20~87 U/L. AST was divided into 3~14,15~17,18~20, and 21~63 U/L. GGT was divided into 2~9,10~11,12~15, and 16~88 U/L. ALP was divided into 15~43,44~50,51~57, and 58~139 U/L. AST/ALT ratio was divided into 0.32~0.93,0.94~1.20,1.21~1.50, and 1.51~17.00. The points in the figure represent the odds ratios (OR), and the error line represents the 95% confidence interval (CI). The numbers in the figure represent the corresponding OR value. GDM, gestational diabetes mellitus; ALT, alanine aminotransferase; AST, aspartate aminotransferase; GGT, gamma-glutamyl transferase; ALP, alkaline phosphatase.