The Association Between Pancreatic Fat Content And Glucose Metabolism Status,Islet Cell Function,And Insulin Resistance

Background and aims:Increased visceral fat content is associated with an increased risk of diabetes.The pancreas is one of the common depositing organs of visceral fat.Basic research showed that the increase of pancreatic fat content could cause pancreatic islet cell damage by increasing insulin resistance,impairing insulin secretion,and promoting pancreatic islet cell apoptosis.A number of previous clinical studies evaluated the association between pancreatic fat content and prevalence of diabetes,islet cell function,however,the conclusions were not consistent.The pancreas is a retroperitoneal organ with a deep location and uneven distribution of pancreatic fat,so it is not easy to measure the pancreatic fat content.The assessment methods of pancreatic fat content were also inconsistant among studies.In this study,we will use artificial intelligence method,deep learning,to analyze the pancreatic pixels of the abdominal CT images,which were collected during the epidemiological survey of the general population in Pinggu District,Beijing.We will evaluate the association between pancreatic fat content,pancreas-spleen ratio,pancreatic volume,and the prevelence diabetes,pancreatic ?-cell function,and insulin resistance.MethodThis study was a cross-sectional study and 4002 objects came from the Pinggu Metabolic Disease Cohort were enrolled.740 participants who had excessively alcoholic intake;had a history of pancreatic related diseases or surgery;were with estimated glomerular filtration rate<30 ml/kg/1.73m2 or with missing important variables were excluded.Finally,3262 participants were enrolled.The average age was 50 years old and 39.3%of this population were men.577 of them were current diabetes and newly-diagnosed diabetes,836 were pre-diabetes.Clinical data(including questionnaires,physical examination,laboratory test)and imaging data(abdominal CT images)of this cohort were obtained.Pancreatic fat content,pancreas-spleen ratio,and the volume of pancreas in abdominal CT were identified by artificial intelligence methods.The linear regression models were used to evaluate the association between pancreatic fat content,pancreas-spleen ratio,pancreatic volume and glycosylated hemoglobin,HOMA-IR and glucose disposal index.Logistics regression models were used to evaluate the association between pancreatic fat content,pancreas-spleen ratio,pancreatic volume and the presence of diabetes,pre-diabetes,and insulin resistance.The area under the receiver operating curve(AUC)was used to compare the accuracy for detecting diabetes,pre-daibetes,and insulin resistance by using pancreatic fat content,pancreas-spleen ratio,pancreas volume,abdominal visceral fat area,abdominal subcutaneous fat area,body mass index,waist circumference and liver-spleen ratio.ResultIn men and women,pancreatic fat content was positively associated with HbA1c(all P<0.01)and increased pancreatic fat content was independently associated with the risk of diabetes(all Pfor trend<0.01).In women,high level of pancreatic fat content was associated with an increased risk of pre-diabetes(Pfor trend<0.01).The pancreatic fat content was negatively associated with the glucose disposal index in men(P<0.01).In the men and women,the level of HOMA-IR gradually increased with the increase of pancreatic fat content and increased pancreatic fat content was positively associated with insulin resistance(all Pfor trend<0.01).The AUC of pancreatic fat content in men and women for detectiong diabetes were 0.70 and 0.75,respectively.Conclusion:In men and women,increased pancreatic fat content was independently associated with risk of diabetes and insulin-risistance.Increased pancreatic fat content was associated with decreased pancreatic ?-cell function only in men.Pancreatic fat content had moderate accuracy for detecting diabetes in men and women.