Distribution Rule Of TCM Syndrome Of Glucose And Lipid Metabolism Disease And Objective Study And Application Of Eye Diagnosis

BackgroundGlucose and lipid metabolism,as the main source of energy supply in the human body,plays a key role in life activities,but glucose and lipid metabolism disorders will produce a series of diseases such as hyperlipidemia,diabetes,fatty liver,obesity,arteriosclerotic cardiovascular and cerebrovascular diseases.The latest survey shows that the overall prevalence of dyslipidemia in Chinese adults is as high as 40.40%,and the incidence of diabetes has exceeded 140 million,becoming the first in the world.Cardiovascular deaths caused by the above diseases account for the first of the total causes of death.Glycolipid metabolic diseases not only have a high incidence,but also often appear in combination or complications,that is,abnormal blood glucose is often accompanied by dyslipidemia or hypertension and other diseases.In 2008,the tutor team investigated dyslipidemia and found that 84.2%of hyperlipidemia patients had diabetes,hypertension and other diseases.In 2012,a survey by the Chinese Diabetes Society showed that 72%of T2DM patients had dyslipidemia and one or two of hypertension,and the comprehensive compliance rate was only 5.6%.Compared with patients with T2DM alone,T2DM patients with dyslipidemia and hypertension have a 6-fold higher risk of cardiovascular disease,indicating that glucose and lipid metabolism disorders also greatly increase the risk of cardiovascular disease.And according to a study published in Diabetes Care in 2016,intensive hypoglycemic therapy did not significantly reduce the incidence of cardiovascular events.Therefore,prevention of cardiovascular disease requires a thorough assessment and control of cardiovascular disease risk factors(hyperglycemia,hypertension,and dyslipidemia).However,at present,multiple risk control of cardiovascular diseases mostly uses the combination of multiple drugs(lipidlowering,antiplatelet,anticoagulant,thrombolytic,vasodilator,antioxidant and antiinflammatory drugs),which has the clinical problems of increased risk of side effects,patient intolerance,heavy economic burden,and poor compliance.Therefore,it is necessary to give full play to the advantages and characteristics of the overall view of TCM and syndrome differentiation and treatment,and carry out comprehensive understanding,diagnosis and treatment.On the basis of many years of clinical practice,literature traceability and scientific empirical evidence.the tutor put forward a new concept of "glucose and lipid metabolism disease · phlegm turbidity"."Glycolipid metabolism disease(GLMD)is a disease characterized by glucose and lipid metabolism disorders,with hyperglycemia,dyslipidemia,fatty liver and atherosclerosis as the main clinical manifestations,which requires comprehensive prevention and control from the whole.The name of TCM disease is called"phlegm".Objective"Glucose and lipid metabolism disease·phlegm turbidity" is a newly proposed concept,and it is necessary to explore its etiology and pathogenesis,syndrome distribution rules,diagnostic methods,etc.,in order to provide the basis and guidance for its clinical comprehensive integrated prevention and control.Therefore,this paper systematically studies the distribution rule of TCM syndrome of glucose and lipid metabolism disease,and provides the basis and basis for syndrome differentiation and treatment of glucose and lipid metabolism disease.Artificial intelligence is introduced into the objective study of the characteristics of "looking at the eyes" in TCM inspection,which lays a foundation for the objectification of TCM diagnosis of glucose and lipid metabolism diseases.The objective index of eye diagnosis is combined with the macroscopic four diagnostic information of TCM,and applied to the construction of the diagnostic model of "glucose and lipid metabolism disease·phlegm turbidity",which not only provides the basis for the application of macroscopic four diagnostic information and objective index of TCM in the diagnostic model,but also provides a new diagnostic method and model originating from the four diagnostic methods of TCM for the coexistence of glucose and lipid metabolism disorders.Methods1.Study on the distribution characteristics of TCM syndrome of glucose and lipid metabolism disease based on the information of four diagnostic methods of TCMBased on the standard "International Guidelines for Clinical Practice of Traditional Chinese Medicine Glucose and Lipid Metabolism Disease" issued by the World Federation of Chinese Medicine Societies,according to the strict diagnosis,inclusion and exclusion criteria,eligible patients with glucose and lipid metabolism disease were screened,and the general data,TCM four diagnostic information,and biochemical indicators of glucose and lipid metabolism of the patients were collected by means of the "Four Diagnostic Information Collection Form for Glucose and Lipid Metabolism Disease in Traditional Chinese Medicine" independently developed by the tutor team,and the syndrome and syndrome element distribution rules were explored by two attending and above TCM physicians for clinical syndrome differentiation and the early development of syndrome element score calculation methods by the research group,respectively,and the correlation between syndrome and syndrome elements and the objective characteristics of "looking eyes" in inspection was explored on this basis.2.Objectification of eye diagnosis in TCM Inspection of Glucose and Lipid Metabolism DiseaseBased on the characteristics of syndrome distribution,it is found that "liver syndrome"is one of the main syndromes of glucose and lipid metabolism disease,and the liver opens in the eyes.Therefore,artificial intelligence white eye shadowless imaging system is further used to collect the pictures of eye diagnosis,and digitally extract the shape,color,and location characteristics of the eye collaterals.1.Collect the eye diagnosis characteristics of simple abnormal glucose metabolism or lipid metabolism(referred to as"glycolipid monopathy")such as type 2 diabetes,hyperlipidemia,and nonalcoholic fatty liver disease,and analyze the commonness and difference characteristics between them;2.Collect the eye diagnosis characteristics of glycolipid metabolism disease(abnormal glucose and lipid metabolism,referred to as "glycolipid comorbidity")and analyze the commonness and difference between "glycolipid comorbidity" and "glycolipid monopathy";3.Analyze the correlation between the eye diagnosis characteristics of "glycolipid comorbidity" and TCM syndromes and clinical test indicators,and elucidate the value and significance of "looking eyes" in the diagnosis of glycolipid metabolism disease.3.Application of objective index and macroscopic four diagnostic information of eye diagnosis in the construction of diagnostic model of glucose and lipid metabolism diseaseOn the basis of elucidating the distribution rule of TCM syndrome of glucose and lipid metabolism disease and the objectification of TCM eye diagnosis characteristics,a crosssectional study was used to collect the general physical examination,macroscopic information of TCM four diagnoses,objectification characteristics of eye diagnosis,clinical biochemistry and other information of the study subjects,and the included study subjects were randomly divided into training set and internal validation set according to a 7:3 ratio.Integrate multi-modal information such as physical examination(numerical information),macroscopic information of four diagnoses of TCM(text information),and characteristics of eye diagnosis(image information),first screen variables through extreme gradient elevation algorithm(XGBoost)to determine the important factors affecting the occurrence of glucose and lipid metabolism disease;then construct a diagnostic model of glucose and lipid metabolism disease using logistic regression and various machine learning algorithms such as K nearest neighbor method,artificial neural network,and XGBoost algorithm;finally comprehensively evaluate the diagnostic efficacy of the model constructed by different methods and optimize the optimal model to provide the basis for the application of macroscopic four diagnoses of TCM and objective indicators in the diagnostic model.Results(一)Syndrome distribution characteristics of glucose and lipid metabolism diseaseA total of 335 patients with glucose and lipid metabolism disease were included in this study,and their syndrome distribution characteristics were as follows:(1)The total syndrome frequency distribution rule of glucose and lipid metabolism disease from high to low was phlegm-dampness block syndrome(50.1%),liver-kidney yin deficiency syndrome(42.7%),liver-qi stagnation and spleen-yang deficiency syndrome(31.9%),spleen-kidney yang deficiency syndrome(21.8%).qi stagnation and blood stasis syndrome(8.4%).and dampness-heat internal accumulation syndrome(8.1%);the compound syndrome of glucose and lipid metabolism disease(47%)accounted for more,and most patients also clamped two syndromes.In addition,phlegm-dampness block syndrome+liver-kidney yin deficiency syndrome accounted for the most proportion(8.36%),and in combination with the three syndromes,phlegm-dampness block syndrome+ liver-kidney yin deficiency syndrome+liver-qi stagnation and spleen deficiency syndrome accounted for the most proportion(6.1%).(2)According to the statistical method of syndrome element score developed by the research group in the early stage,the empirical syndrome of patients with glucose and lipid metabolism disease is mainly dampness syndrome,qi stagnation syndrome and phlegm syndrome,the deficiency syndrome is mainly yin deficiency,yang deficiency and blood deficiency,and the lesion sites are mainly in the liver.spleen and kidney.(二)Characteristics of glucose and lipid metabolism disease and its correlation with clinical parameters1.Glucose and lipid metabolic disease eye diagnosis characteristicsA total of 889 subjects were included in this study,including 213 subjects in the nonglycolipid metabolism disease group,20 subjects in the type 2 diabetes alone group,95 subjects in the hyperlipidemia group,50 subjects in the non-alcoholic fatty liver disease group,and 511 subjects in the glycolipid metabolism disease group.(1)The ocular characteristics of the simple type 2 diabetes group were as follows:compared with the non-glucose and lipid metabolism group,the score of "colliculus" feature in the morphological characteristics of the simple type 2 diabetes group was 2(0.75,3.25),and there was a statistically significant difference,P<0.01;the score of "yellowish" feature in the color characteristics was 1(0,1.25),the score of "red" was 1(0,2),and the score of"yellow" feature was 1(0,2),all of which were statistically significant,P<0.01;(2)The important ocular characteristics of the hyperlipidemia group were as follows:compared with the non-glucose and lipid metabolism group,the score of "colliculus" feature in the morphological characteristics of the hyperlipidemia group was(0,(0,1.5),and the score of "haze diffuse" feature was 3.85 ± 1.185,and the difference was statistically significant,P<0.05);the regional characteristics were "A","C","G","I",and "K" zone feature scores were statistically significant,P<0.05;There were significant differences in the scores of "dark color","light color","gray" and "cyan" features among the color features.(3)The important ocular characteristics of the non-alcoholic fatty liver disease group were:compared with the non-glucose and lipid metabolism disease group,the "colliculus"characteristic score in the morphological characteristics of the non-alcoholic fatty liver disease group was(0.2),and the "fog diffuse" characteristic score was 4.14±1.05,and the difference was statistically significant,P<0.01;the important ocular color characteristics were "yellow" characteristic score 1(0,2),and the difference was statistically significant,P<0.01;the "A" characteristic score in the ocular collateral location characteristics was 2.5(2,5),and the difference was statistically significant,P<0.05.(4)the important ocular characteristics of the glycolipid metabolism disease group were:compared with the non-glycolipid metabolism disease group,the "colliculus" feature score in the morphological characteristics of the glycolipid metabolism disease group was 1(0.2),and the "spot" feature score was 3(1.5),and the difference was statistically significant,P<0.05;the location characteristics were "A","C","E","G","I" and"K" regions,and the feature scores were significantly different;the "dark red","yellowish","pink","red",and "yellow"feature scores were significantly different between the two groups.(5)Through stratified analysis of glucose and lipid metabolism disease,divided into glucose metabolism disorder combined with hyperlipidemia group,glucose metabolism disorder combined with non-alcoholic fatty liver disease group,glucose metabolism disorder combined with hyperlipidemia&non-alcoholic fatty liver disease group,glucose metabolism disorder combined with hyperlipidemia&non-alcoholic fatty liver disease group&atherosclerosis group,each layer group was compared with non-glucose and lipid metabolism disease group,respectively,and it was found that the typical morphological characteristics of white eyes of glucose and lipid metabolism disease were "colliculus",and the colors involved "dark red","yellowish","pink","pink","yellow","dark","pale",and"gray",with "A" area,"C" area,"E" area,"G" area,"I" area,"L" area,"K" area,"M" area,and "N" area compared with non-glucose and lipid metabolism disease group.2.Association between eye diagnosis characteristics of glucose and lipidmetabolism diseases and clinical parameters(1)The correlation between the characteristics of eye diagnosis and clinical evidence and evidence elementsIn this study.335 patients with glucose and lipid metabolism disease who were included in the valid TCM syndrome questionnaire had a relatively large proportion of "phlegmdampness internal accumulation" in their syndrome frequencies,a relatively large proportion of "liver" syndrome in the viscera in the syndrome element frequencies,and a relatively large proportion of "yin deficiency" in the eight principles.The syndrome elements and syndrome differentiation were compared with the characteristics of ophthalmogram,and the results were as follows:①In terms of morphological characteristics,it was found that there was no significant difference in the correlation coefficient between each syndrome in morphological characteristics;there was significant difference in the correlation coefficient between spleen syndrome and stomach syndrome and the score of "colliculus" characteristics,which was negatively correlated;there was a negative correlation between dampness syndrome and the score of "spot" characteristics,and there was a positive correlation between phlegm syndrome and the score of "blood vessel" characteristics.②In terms of location characteristics,the reflection points on location were different between different syndromes,and there were some common reflection points between syndrome elements.Phlegm syndrome,blood stasis syndrome,qi deficiency syndrome,yang deficiency syndrome and kidney syndrome were positively correlated with the characteristic scores of "D" area.Dampness syndrome,yin deficiency syndrome and liver syndrome were negatively correlated with the characteristic score of"G" area.③In terms of color characteristics,the color of white eye collaterals was different in each different syndrome,and there were crosses between different syndrome elements.The correlation coefficients of liver-kidney yin deficiency syndrome,qi stagnation syndrome,qi deficiency syndrome,heat syndrome and "dark pink" characteristic scores were statistically significant.Liver depression and spleen deficiency syndrome,qi deficiency syndrome,yang excess syndrome,spleen syndrome,kidney syndrome and "dark brown" characteristic correlation coefficient score was statistically significant.Heat syndrome was positively correlated with the score of "yellowish" feature.Phlegm syndrome and yang excess syndrome were positively correlated with the characteristic score of "powder dark color".Qi stagnation syndrome and heat syndrome were negatively correlated with the score of "pink"feature.Qi stagnation syndrome,phlegm syndrome and yang deficiency syndrome were positively correlated with the score of "light color" feature.Qi stagnation syndrome,yang excess syndrome and yang deficiency syndrome were negatively correlated with the score of"pink" feature,and heat syndrome was positively correlated with the score of "pink" feature.Dampness-heat syndrome and heat syndrome were positively correlated with the score of"red" feature,and qi stagnation syndrome,qi deficiency syndrome and yang deficiency syndrome were negatively correlated with the score of "red" feature.Dampness syndrome and dampness-heat syndrome were positively correlated with the score of "purplish color"characteristics.(2)Correlation between clinical biochemical and demographic parameters and "eyesight" characteristicsCorrelation analysis was performed between clinical biochemical indicators and visual characteristics in the total population of 889 subjects included,and it was found that BMI,waist circumference,waist-to-hip ratio,fasting blood glucose,triglyceride,and high-density lipoprotein cholesterol were positively correlated with "colliculus" in terms of morphological characteristics,and the correlation coefficient was statistically significant;BMI,waist circumference,waist-to-hip ratio,fasting blood glucose,cholesterol,triglyceride,and lowdensity lipoprotein cholesterol were positively correlated with the characteristic score of "C"area,and the correlation coefficient was statistically significant;BMI,waist circumference,and waist-to-hip ratio were significantly different from the characteristic score of "dark pink","dark gray"."yellowish","red","dark","light","pink","red","yellow","gray",and "cyan"in terms of color characteristics.(三)Construction of diagnostic model for glucose and lipid metabolism disease1140 subjects were enrolled in this study,and their eye diagnosis characteristics,macroscopic information of four diagnoses of TCM,physical examination and other information were collected.Logistic regression,KNN nearest neighbor method,neural network,and gradient decision tree algorithm(XGBoost)were used to construct five models:model Ⅰ(eye diagnosis characteristics),model Ⅱ(macroscopic information of four diagnoses of TCM),model Ⅲ("eyesight" characteristics&physical examination),model Ⅳ(macroscopic information of four diagnoses of TCM&physical examination),and model Ⅴ(eye diagnosis¯oscopic information of four diagnoses of TCM&physical examination).Through AUC,precision recall(P-R)curve,accuracy,etc.,five models and algorithms were comprehensively compared,and it was found that the diagnostic efficacy of model five was the best,and the performance of XGBoost algorithm was the best,with an AUC of 0.91,an accuracy of 0.82,and an accuracy of 0.80 after parameter adjustment,indicating that the model was stable.The above results show that adding objective index of"looking at the eyes" feature increases the accuracy and AUC of the model.Based on the characteristics of "eyesight",macroscopic information of four diagnoses in TCM,and physical examination combination of a large sample population,logistic regression,KNN nearest neighbor method,neural network,and gradient decision tree algorithm(XGBoost)were used to construct model Ⅰ(simple eye diagnosis),model Ⅱ(macroscopic information of four diagnoses in TCM),model Ⅲ(eye diagnosis+physical examination),model Ⅳ(macroscopic information of four diagnoses in TCM+physical examination),and model Ⅴ(eye diagnosis+macroscopic information of four diagnoses in TCM+physical examination)according to data feature selection,and it was found that adding eye examination could increase the accuracy and AUC of the model after comprehensive comparison,and the performance of the XGBoost algorithm was superior,with an AUC of 0.91,an accuracy of 0.82,and an accuracy of 0.80 after parameter adjustment,indicating that the model was stable.Conclusion(1)Through cross-sectional investigation and research,it is found that the syndrome distribution of patients with glucose and lipid metabolism disease is mainly phlegmdampness block syndrome,liver-kidney yin deficiency syndrome,liver depression and spleen deficiency syndrome;the disease location mainly involves the liver,spleen,kidney and other viscera,providing the basis and basis for the syndrome differentiation and treatment of glucose and lipid metabolism disease.(2)Artificial intelligence was introduced into the objective study of eye diagnosis in TCM,and beneficial exploration of the objectification of TCM diagnosis was carried out.It was found that there were both common points and differences in the characteristics of eye diagnosis between glycolipid metabolism disease and simple type 2 diabetes,simple hyperlipidemia,and simple nonalcoholic fatty liver disease,and their common morphological characteristics were "colliculus",and their differences suggested that glycolipid metabolism disease was not a simple mutual superposition between single diseases.(3)On the basis of the above,a disease diagnosis model integrating physical examination,macroscopic information of four diagnoses of traditional Chinese medicine,and objective indicators of eye diagnosis is constructed using a variety of machine learning algorithms,and the best diagnostic model is optimized.It provides methodological reference and basis for the application of macroscopic four diagnostic information and objective indicators in diagnostic models of TCM.(4)The diagnostic model integrates the multi-modal information of TCM and provides a new diagnostic method integrating the macroscopic four diagnostic information of TCM and the objective index of eye diagnosis for the clinical diagnosis of glucose and lipid metabolism diseases,which can be used for large-scale screening of glucose and lipid metabolism diseases and has the characteristics of non-invasiveness and simplicity,and has certain clinical applicable value.