Association Of Plasma Carnitine Metabolites,Renal Function In Early Pregnancy With Gestational Diabetes Mellitus

Background and ObjectiveGestational diabetes mellitus(GDM)is one of the most common complications during pregnancy,defined as impaired glucose tolerance to varying degrees first discovered or diagnosed during pregnancy.Data from the IDF Diabetes Atlas 10 th Edition published by the International Diabetes Federation showed that 16.7% of females aged 20 to 49 years globally suffered from hyperglycemia,80.3% of which was GDM.GDM not only negatively influences the physical and mental health of pregnant women and fetuses,as well as the long-term health of mothers and offspring,but also brings a heavy economic burden of maternal and infant care.In recent years,the development of metabolomics technology has promoted the study of metabolomics in the pathogenesis of GDM.Carnitine metabolites,important small molecule metabolites in mitochondrial oxidative energy supply,have been reported to be closely related to insulin resistance and the development of GDM.However,due to factors such as study design and sample size,the results are controversial.For multiple adverse effects of GDM on maternal and infant health,it is vital to identify the highrisk population of GDM as early as possible to reduce the risk of GDM.Many risk prediction models based on traditional risk factors were used to predict the risk of GDM.However,the predictive ability of these models was limited,and little external validation was conducted.Few prediction models could be used for Chinses pregnant women clinically.Therefore,it is necessary to explore new predictors among Chinses pregnant women and establish an effective prediction model for GDM.Besides,GDM is affected by both genetic and environmental factors.Previous studies reported that the polygenic risk score(PRS)based on type 2 diabetes was related to the GDM risk,and there was an interaction between genetic variation and small molecule metabolites on the risk of GDM.However,genetic background varies across populations.Studies examining the association of PRS with GDM risk among Chinese women are less,and it is unclear whether the associations between plasma carnitine metabolites and GDM risk are influenced by the individual inherent genetic background.Previous studies have found that abnormal renal function might be associated with multiple risk factors of GDM,such as obesity,hypertension,and insulin resistance,and renal function indicators like serum creatinine,uric acid,and estimated glomerular filtration rate(e GFR)have been evidenced to significantly related to the risk of GDM or type 2 diabetes.However,the number of prospective studies is limited,and the sample sizes are smaller,and the potential biological mechanism is still confused.In addition,the kidney is an important organ for carnitine and acylcarnitine metabolism,and previous studies have suggested that renal function was associated with plasma carnitine metabolites.However,whether renal function might cause a disorder of glucose metabolism by influencing plasma carnitine metabolites has not been studied.In view of this,based on the Tongji-Shuangliu Birth Cohort(TSBC),we designed a nested case-control study among pregnant women recruited from 2017 to 2019.Targeted metabolomics detection technology was used to detect the level of plasma carnitine metabolites in early pregnancy.This study aimed to 1)examine the association between plasma carnitine metabolites in early pregnancy and the risk of GDM and to explore the predictive ability of carnitine metabolites on GDM risk;2)examine the association of genetic risk related to type 2 diabetes and plasma carnitine metabolites in early pregnancy with the risk of GDM;3)estimate the associations between early-pregnancy renal function indicators and the risk of GDM,select carnitine metabolites related to renal function indicators,and explore the mediating roles of plasma carnitine metabolites in the associations between renal function indicators and GDM risk.Methods Part 1 Plasma carnitine metabolites in early pregnancy and risk of gestational diabetes mellitusA nested case-control study of 328 GDM cases and 656 controls was established based on the TSBC.The GDM cases and controls were matched at 1:2 on age(±3years)and gestational age(±4 weeks).GDM was diagnosed according to the criteria recommended by International Association of the Diabetes and Pregnancy Study Groups.The concentration of 39 plasma carnitine metabolites in early pregnancy was detected by targeted metabonomic detection technology based on the platform of ultrahigh performance liquid chromatography-mass spectrometry.Two-sample t-test and multivariable-adjusted conditional logistic regression were used to identify differential carnitine metabolites between GDM cases and controls.The least absolute shrinkage and selection operator(LASSO)regression and 10-fold cross-validation were applied to select carnitine biomarkers related to GDM.The receiver operating characteristic curve and area under the curve(AUC)were used to estimate the predictive ability of carnitine metabolites on GDM.Part 2 Genetic risk of type 2 diabetes mellitus,plasma carnitine metabolites in early pregnancy,and risk of gestational diabetes mellitusThis study was conducted among 855 pregnant women(285 GDM cases and 570controls)with genome-wide typing data from a nested case-control study in Part 1.Genomic DNA of peripheral blood was extracted by commercial DNA extraction kits,and Illumina Human Omni Zhong Hua-8 Bead Chips were used for genome-wide genotyping.Type 2 diabetes polygenic risk score(PRS)was estimated using a weighted method based on 174 single nucleotide polymorphisms significantly related to type 2 diabetes among Eastern Asians.Multiple linear regression was applied to estimate the associations of type 2 diabetes PRS with plasma carnitine metabolites in early pregnancy,as well as biomarkers of glucose metabolism in early and secondtrimester pregnancy.Multivariable-adjusted conditional logistic regression was used to examine the relationship between type 2 diabetes PRS and GDM risk,as well as the interaction of type 2 diabetes PRS with carnitine metabolites on GDM risk.Part 3 Association of renal function in early pregnancy with gestational diabetes mellitus and the mediating roles of the carnitine metabolitesBased on TSBC,6770 pregnant women with complete renal function examination data were enrolled in this study.Multivariable-adjusted logistic regression and multiple linear regression were separately used to assess the associations of renal function indicators in early pregnancy with the risk of GDM and blood glucose during the oral glucose tolerance test in the second-trimester pregnancy.The mediation analyses of carnitine metabolites were conducted among pregnant women from the nested case-control study in Part 1.First,multiple linear regression was applied to identify carnitine metabolites related to renal function indicators in early pregnancy;second,LASSO regression was conducted to select carnitine metabolites related to renal function indicators and establish renal function indicators-related carnitine score;third,multivariable-adjusted conditional logistic regression was performed to examine the associations between renal function indicators-related carnitine score and GDM risk;finally,mediation analyses were used to test the mediation effects of renal function indicators on GDM risk through the carnitine score.Results Part 1 Plasma carnitine metabolites in early pregnancy and risk of gestational diabetes mellitusTwo-sample t-test and multivariable-adjusted conditional logistic regression identified 10 differential carnitine metabolites between GDM cases and controls.Seven carnitine biomarkers were further selected by LASSO regression,and 6 of them were significantly associated with GDM risk in multivariable-adjusted conditional logistic regression.C5-DC,C13:0,C16:1,C16:2,and C18:2 were negatively associated with the risk of GDM,and the odds ratio(OR)of per 1 log-standard deviation(SD)increment on the concentration of these carnitine metabolites ranged from 0.76 to 0.85,while C4-OH was positively associated with GDM,and the OR was1.60 [95% confidence interval(CI): 1.04,2.46;Q4 vs Q1].When the 7 carnitine biomarkers were included in the predictive models based on traditional predictive factors and clinical glucose indicators,the AUC was improved from 0.704(0.668,0.740)to 0.724(0.690,0.758).Part 2 Genetic risk of type 2 diabetes mellitus,plasma carnitine metabolites in early pregnancy,and risk of gestational diabetes mellitusType 2 diabetes PRS was positively associated with the risk of GDM,and pregnant women with higher PRS had an increment of 57% in GDM risk(1.57;1.02,2.42;Q4 vs Q1).Results of multiple linear regression showed that type 2 diabetes PRS was positively associated with baseline fasting blood glucose and glycosylated hemoglobin.The associations between carnitine metabolites and GDM risk were stronger among pregnant women with higher PRS.Except for C4-OH,those with higher PRS and lower carnitine metabolites had a much higher risk of GDM,and the ORs of GDM ranged from 1.45 to 1.80.Certain carnitine metabolites(including C16:1and C16:2)showed significant interaction with PRS on the risk of GDM(P for interaction ≤ 0.006),and each unit increment on these two carnitine metabolites and PRS at the same time contributed to 42% and 35% reduction of GDM risk respectively.Part 3 Association of renal function in early pregnancy with gestational diabetes mellitus and the mediating roles of the carnitine metabolitesSerum uric acid,uric acid to creatinine ratio,and e GFR were positively associated with GDM risk after adjusting for confounders,and per 1 log-SD increment of concentrations or value associated with 23%(1.23;1.11,1.36),31%(1.31;1.19,1.45),and 15%(1.15;1.04,1.27)higher risk of GDM separately.Increased serum creatinine,creatinine to weight ratio,and cystatin C were associated with decreased risk of GDM with ORs of 0.88(0.80,0.97),0.88(0.78,0.99),and 0.89(0.80,0.98)for 1 log-SD increment.A total of 5,12,8,13,and 12 carnitine metabolites were selected to establish uric acid,creatinine,cystatin C,e GFR,and creatinine to weight ratio-related carnitine score among 933 pregnant women from the nested case-control study.Increasement of 1 SD in creatinine and creatinine to weight ratio-related carnitine scores were separately associated with 22%(0.78;0.66,0.91)and 25%(0.75;0.64,0.88)reduction of GDM risk,while the OR of the e GFR-related score was 1.25(1.07,1.46).Mediation analyses showed that serum creatinine,creatinine to weight ratio,and e GFR-related carnitine score positively mediated the associations between serum creatinine,creatinine to weight ratio,and e GFR with GDM risk,and the proportions were 43.1%,81.9%,and 56.7%,respectively.ConclusionA set of plasma carnitine metabolites in early pregnancy were found to be associated with the risk of GDM in this study and could improve the predictive ability of the traditional GDM prediction model to some extent.Type 2 diabetes PRS was also positively associated with GDM risk,and the associations between certain carnitine metabolites and GDM risk were stronger among those with higher PRS.This study showed significant interactions between type 2 diabetes PRS and certain carnitine metabolites on the risk of GDM.In addition,this study found that serum uric acid,uric acid to creatinine ratio,and e GFR were positively associated with GDM risk,while creatinine,creatinine to weight ratio,and cystatin C were negatively associated with GDM risk.Certain carnitine metabolites positively mediated the associations of serum creatinine,creatinine to weight ratio,and e GFR with the risk of GDM.