Identification Of Gestational Diabetes Mellitus Related Urinary Biomarkers Along Pregnancy(From Early Pregnancy To Postpartum)by Using Metabolomics Analysis

ObjectiveGestational diabetes mellitus(GDM)has many adverse effects on pregnant women and fetuses.At present,no prediction marker for GDM in early pregnancy is accepted.There is still a lack of recognized early predictors.This study was designed to identity valuable metabolic biomarkers for GDM.MethodsThis was a prospective observed cohort study.140 pregnant women were recruited in early pregnancy,and followed up to 6 weeks postpartum.Glucose challenge test and 75g oral glucose tolerance test were performed after 24 weeks of pregnancy,and GDM was diagnosed according to the latest ADA standard.Urinary samples were collected in the first(<12 weeks),second(24～28 weeks)and third(32-weeks)trimester of pregnancy.Urinary metabolomics were analyzed by ultra-performance liquid chromatography tandem mass spectrometry.15 cases of GDM women and 50 cases of control women were used for longitudinal analysis;15 cases of GDM women and 15 cases of age matched control women were used for difference analysis.Results1.Clinical parameters for GDM prediction.21 cases of pregnant women were diagnosed as GDM,the incidence was 15%.BMI in early pregnancy had the best predictive ability for GDM in all clinical parameters.Its area under the receiver operating characteristic(ROC)curve was 0.82,the sensitivity and specificity were 77.8%and 78.0%respectively when cut point was 22.9kg/m2.The area under the ROC curve could increase to 0.91 when combined with varied clinical parameters,then the positive likelihood ratio was 5.57,negative likelihood ratio was 0.21.2.Effects of normal pregnancy and GDM on urinary metabolites.The steroid hormones biosynthesis and tyrosine metabolic pathway was affected in normal pregnancy.Steroid hormone synthesis was increased and tyrosine metabolism was decreased.The tryptophan metabolic pathway was mainly affected in GDM pregnancy.Kynurenic acid、4,6-Dihydroxyquinoline、3-Hydroxyanthranilic acid and other major urinary metabolites showed a decreasing trend.3.Early prediction of GDM by urinary metabolites.In the first,second and third trimester of pregnancy,7 kinds,9 kinds and 7 kinds different urinary metabolites were identified between GDM and control group.In early pregnancy,the area under the ROC curve of D-1-Deoxy-erythro-hexo-2,3-diulose for prediction of GDM was the largest(0.84,p=0.002),the sensitivity and specificity were both 80%.The area under the ROC curve could increase to 0.94 when combined with three different urinary metabolites,the sensitivity and specificity were 93.3%and 86.7%respectively,the positive likelihood ratio was 14.1 and the negative likelihood ratio was 0.08.4.Differences in urinary metabolites of GDM women with different postpartum glucose tolerance.Glucose tolerance was confirmed postpartum in 9 cases of GDM women,6 cases were diagnosed as impaired glucose regulation(IGR)and 3 cases were diagnosed as normal glucose tolerance(NGT).9 kinds of urinary metabolites were found significantly different between the two groups.4 kinds of urinary metabolites were decreased in IGR group including Humulinic acid A,L-Octanoylcarnitine,Gingerol and Lactapiperanol C.5 kinds of urinary metabolites were increased in IGR group including Pipecolic acid,3-Hydroxyanthranilic,Sterol,2-Methylguanosine and 25-Acetylvulgaroside.5.Differences in urinary metabolites between GDM women with different treatment.4 cases of GDM women were treated with insulin.Compared with the lifestyle intervention group,the level of methoxamine was decreased in the first trimester of pregnancy in the insulin treatment group,the level of Lactapiperanol C was increased in the second trimester of pregnancy and the level of QH2 was increased in the third trimester of pregnancy.6.Prediction of adverse pregnancy outcomes by urinary metabolites.Adverse pregnancy outcomes were found in 19 cases of pregnant women.Compared with the control group,the maternal age,gravidity and fasting blood glucose level of women with adverse pregnancy outcomes were significantly higher(p<0.05).Among the identified urinary differential metabolites,the highest area under the ROC curve for independently predicting adverse pregnancy outcomes was found in indolylacryloylglycine(0.84,p<0.001).Combined with L-histidine could improve the prediction ability,the area under the ROC curve increased to 0.92,the positive likelihood ratio was 15.9,and the negative likelihood ratio was 0.17.ConclusionsThe normal pregnancy process mainly affects steroid hormone biosynthesis and tyrosine metabolism.When GDM occurs,the tryptophan metabolism is affected.Urinary differential metabolites in early pregnancy have certain value for prediction of GDM,and combined with various metabolites could improve the accuracy of GDM prediction.Urinary metabolites in early pregnancy may have early predictive value for adverse pregnancy outcomes.Objective:To analyze the characteristics of dynamic blood glucose in hypoglycemia patients with different causes and explore the application value of dynamic glucose monitoring system(DGMS)in the diagnosis and differential diagnosis of hypoglycemia.Methods:The blood glucose levels of 20 hypoglycemia patients were monitored by DGMS for 3-6 days,including 6 cases of insulinoma,7 cases of reactive hypoglycemia,2 cases of autoimmune hypoglycemia,3 cases of islet cell hyperplasia,1 case of drug-induced hypoglycemia,1 case of congenital hyperinsulinism.To analyze the differences of blood glucose profiles in hypoglycemia patients with different causes.Results:The average monitoring duration of each patient was 88.9±10.8 hours,and 38-936 points of hypoglycemia were detected in the patients with hypoglycemia.The average blood glucose of each group from low to high were:congenital hyperinsulinism(2.97±0.59 mmol/L),islet cell hyperplasia(4.83± 1.32 mmol/L),insulinoma(4.87± 1.37 mmol/L),drug-induced hypoglycemia(5.98±0.8 mmol/L),reactive hypoglycemia(6.38±1.99 mmol/L),autoimmune hypoglycemia(7.63±2.67 mmol/L).The constituent ratio of hypoglycemia from low to high were:autoimmune hypoglycemia and drug-induced hypoglycemia(0%),reactive hypoglycemia[(0.5%(0.0%,1.6%)],insulinoma[9.0%(5.2%,16.3%)]，islet cell hyperplasia[9.5%(1.1%,14.2%),congenital hyperinsulinism(58.5%).To analyze the blood glucose profiles in day and night,the night blood glucose levels were lower than that of daytime in insulinoma(5,16± 1.37 mmol/L vs 4.44± 1.24 mmol/L),reactive hypoglycemia(6.93 ± 2.19 mmol/L vs 5.57± 1.25 mmol/L),autoimmune hypoglycemia(8.08±2.79 mmol/L vs 6.95±2.31 mmol/L),islet cell hyperplasia(5.23±1.31 mmol/L vs 4.11 ±1.00 mmol/L)and congenital hyperinsulinism(3.06±0.57 mmol/L vs 2.83±0.59 mmol/L),all P<0.05;the daytime blood glucose levels were lower than that of night in drug-induced hypoglycemia(5.90±0.81 mmol/L vs 6.11±0.77 mmol/L,P<0.05).62.2%and 78.9%of the hypoglycemia happened at night in patients with insulinoma and islet cell hyperplasia,and 43.2%of the hypoglycemia happened at night in patients with congenital hyperinsulinism.The highest standard deviation of blood glucose(2.67 mmol/L),largest amplitude of glycemic excursions(9.2 mmol/L)and coefficient of variation(35%)were found in in patients with autoimmune hypoglycemia.The highest range of blood glucose(20.4 mmol/L)was found in patients with reactive hypoglycemia.The coefficient of variation(31.6%vs 22.4%)and range of blood glucose(20.4mmol/L vs 6.8mmol/L)were both most different between daytime and night in patients with reactive hypoglycemia.Conclusions:The application of DGMS and the analysis of the parameters such as average blood glucose,blood glucose profiles and variability of blood glucose are helpful for the diagnosis and differential diagnosis of hypoglycemia.Objective:To evaluate the clinical and biochemical characteristics,the secretion characteristics of insulin,glucagon and glucagon like peptide 1 after oral glucose load in pregnant women with different glucose tolerance status.Methods:We retrospectively analyzed 74 cases pregnant women with positive results of 50g glucose challenge test who received regular follow-up over the duration of pregnancy in Peking Union Medical College Hospital from January 2009 to June 2012.A further 100 g oral glucose tolerance test(OGTT)was performed,subjects were divided into three groups as gestational diabetes mellitus(GDM)group(25 cases),impaired glucose tolerance(IGT)group(25 cases)and normal glucose tolerance(NGT)group(24 cases).The general clinical data and biochemical indexes of the three groups were compared,and the related index of insulin resistance and the function of pancreatic islet beta cell was calculated.Glucose,insulin,glucagon and glucagon like peptide 1(GLP-1)were measured in OGTT.The secretion characteristics of each hormone and their correlation with other indicators were evaluated.Results:Compared with the NGT group,the age(33.9±4.1 vs 31.8±3.0years,P＝0.039),diastolic blood pressure(75.0± 10.8 vs 68.5±6.7mmHg,P＝0.035),GCT(9.21 ±0.75 vs 8.52±0.50mmol/L)and glycosylated hemoglobin Alc(5.39±0.34 vs 5.18±0.20%,P＝0.008)were higher in subjects of the GDM group.The body heights of the pregnant women in the GDM group were significantly lower than that of the IGT group and NGT group(161.2±4.3 vs 163.9±4.8 vs 163.6±4.0,P1 ＝P2＝0.038).In OGTT,the area under curve(AUC)of glucose of the GDM group was significantly higher than that of the IGT group and NGT group(26.58±2.02 vs 23.2±1.51 vs 19.84± 1.95[mmol/(L·h)],P1<0.001,P2<0.001).The peak values of insulin secretion in the GDM group and IGT group were delayed to 2-hour in OGTT.The 3-hour insulin levels(124.23±87.16 vs 75.26±49.14μIU/mL,P＝0.016)and the AUC of insulin(2.49±1.36 vs 1.84±0.91[×103pmol/(L·h)],P = 0.047)in OGTT of the GDM group were significantly higher than that of the NGT group.Compared with the NGT group,the glucagon levels in each time point of OGTT and the AUC of glucagon levels were reduced in the GDM group and IGT group,but no significant differences were found.The peak value of glucagon levels were emerged at 3-hour of OGTT in all groups.The GLP-1 levels in each time point of OGTT were gradually increased from NGT group to IGT group to GDM group,but no significant differences were found.The peak value of GLP-1 level was presented at 1-hour of OGTT in NGT group and IGT group and at 2-hour of OGTT in the GDM group.The valley value of GLP-1 level was appeared at 3-hour of OGTT in all groups.The further statistical analysis demonstrated that the AUC of glucagon levels in OGTT were negatively correlated with 0-hour(r=-0.287,P＝0.013)and 1-hour glucose levels(r=-0.266,P＝0.022)in OGTT and positively correlated with ISSI(r＝0.297,P＝0.010)and HOMA-p(r＝0.236,P＝0.043).Moreover,the AUC of GLP-1 levels in OGTT were negatively correlated with the levels of C reactive protein(r=-0.264,P =0.035).Conclusion:Our study suggests that insulin resistance and the dysfunction of pancreatic islet β cell were observed in pregnant women with GDM and IGT in the period of midtrimester pregnancy.Potential GLP-1 resistance and inadequate secretion may exist in GDM patients.Glucagon levels may not start to change remarkably before 28 gestational weeks.