A Study On The Factors Influencing Of The Tangential Point Of Glucose Metabolism Diagnosis Among The Tibetan Population In Lhasa Based On The Method Of Noninvasive Blood Glucose Detection By Metabolic Heat Conformation

Objective（1）To establish and modify noninvasive blood glucose calculation models based on metabolic heat integration method of Tibetan population of Lhasa area;（2）The accuracy of the noninvasive blood glucose detection model was verified by paired analysis of blood glucose values detected by noninvasive blood glucose detection technology,automatic biochemical analyzer and fingertip blood glucose meter;（3）To analyze the factors influencing the accuracy of blood glucose detection model based on metabolic thermal integration technology among Tibetan in Lhasa.Methods A total of 200 normal people and 600 patients with type 2 diabetes mellitus（T2DM）recruited as subjects in the experiment in Lhasa,Tibet received blood glucose monitoring by noninvasive blood glucose meter（NIGA02）,automatic biochemical analyzer（hexokinase method）and finger capillary blood glucose meter（Roche blood glucose meter）at fasting and 2 hours after meal,respectively.At the same time,the subjects’upper arm venous blood was collected to measure fasting insulin and C-peptide,glycated hemoglobin（Hb A1c）,blood routine,thyroid stimulating hormone（TSH）and other indicators.The cross-validation method was used,with venous plasma glucose as the reference frame and Parkes error grid analysis in ISO 19157:2013 as the evaluation index.80%of the noninvasive detection data were used to correct the model to meet the detection accuracy of Tibetan in Lhasa,and the remaining 20%were used to verify the modified model.The blood glucose values measured by the noninvasive blood glucose meter were compared with the venous plasma blood glucose values.The consensus error grid analysis（Parkes error）was used as the main evaluation index,and the mean relative error（MARD）,Bland-Altman analysis and regression analysis were used as the secondary indicators to evaluate the accuracy of the noninvasive blood glucose meter in among the Tibetan in Lhasa.Double ordinate plots were used to evaluate the factors affecting the accuracy and determine the optimal detection range.Results In the normal population,Parkes error grid analysis showed that 94.66%of the blood glucose values was located in the A and B regions,and there was no blood glucose value of the E region.Bland-Altman consistency test showed that 298（93.71%）of 318 paired blood glucose values between noninvasive glucose meter and venous plasma glucose were within the consistency limit.The Pearson correlation coefficient was 0.43,the p value of regression model was<0.0001.MARD was 25.60%.In T2DM patients,the Parkes error grid analysis of the blood glucose values of the noninvasive glucose meter and venous plasma showed that 96.39%of the blood glucose values were located in the A and B regions,and there was no blood glucose value of the E region.Bland-Altman consistency test showed that 810（94.19%）of the 860 paired blood glucose values between noninvasive glucose meter and venous plasma glucose were within the consistency limit.The Pearson correlation coefficient was 0.76,the p value of regression model was<0.0001.The MARD was 24.90%.The factors influencing the accuracy of glucose metabolism include oxygen saturation（Sp O₂）,blood flow velocity,environmental temperature,environmental humidity,fasting insulin（Fins）and C-peptide（C-p）.Conclusion The noninvasive blood glucose detection technology based on metabolic thermal integration has high accuracy in blood glucose monitoring of the diabetic in Lhasa.The noninvasive glucose monitor is the most suitable for ambient temperatures between 17℃and 23℃,ambient humidity between 25%and 55%,blood flow between 240Bpu and 280Bpu,and Sp O₂ over 83%,with significant effects of insulin and C-peptide concentrations on accuracy.