Study On Wind-related Syndromes Of Diabetic Peripheral Neuropathy And Construction Of Neural Network Prediction Mode

ObjectiveStudy 1:To analyze the distribution of the syndromes of diabetic peripheral neuropathy(DPN)patients by collecting their four-diagnosis information,and to explore the manifestation of wind-related syndrome in this disease.Study 2:To explore the differences between wind syndrome and non-wind syndrome patients by collecting general information and laboratory indicators of DPN patients,and to extract characteristic indicators of wind syndrome.Based on neural network analysis technology and characteristic indicators,a syndrome prediction model is constructed to assist in the clinical diagnosis of wind syndrome and explore the significance of wind syndrome in this disease.MethodsStudy 1:A total of 246 DPN patients who met the inclusion criteria and visited Dongzhimen Hospital of Beijing University of Chinese Medicine from October 2020 to January 2023 were included,and their general information and four-diagnosis information were collected.Statistical software was used for statistical description and differential analysis of patient data.The distribution of four-diagnosis information was analyzed by frequency and frequency analysis.Factor analysis was used to extract the common factors of information,and factor scores were calculated according to the formula,to summarize the distribution of syndromes.Study 2:A total of 246 DPN patients who met the inclusion criteria and visited Dongzhimen Hospital of Beijing University of Chinese Medicine from October 2020 to January 2023 were included,and their general information,four-diagnosis information and laboratory indicators were collected.Patients were divided into wind syndrome group and non-wind syndrome group according to whether they had wind syndrome.Statistical software was used for differential analysis of data between the two groups,and a neural network multilayered perceptron program was applied to construct a diabetic peripheral neuropathy wind syndrome prediction model.ResultsStudy 1:1 General data analysis results:A total of 246 DPN patients were analyzed in this study.Among them,135 were male,accounting for 54.9%of all patients,and 111 were female,accounting for 45.1%.There was no statistically significant difference in gender distribution(P>0.05).Compared patients’ Age,BMI,and diabetes duration,and analyzed the difference between the different gender groups.The age of the patients ranged from 26 to 80 years old,with a median of 63.0(13.5)years old.The median age of male patients was 61.0(13.0)years old,and that of female patients was 66.0(14.0)years old,with a significant statistical difference between the two genders(P<0.01).There was no statistically significant difference in BMI or disease duration between genders(P>0.05).The age distribution of DPN patients was highest in the 60-74 year old group,followed by the 45-59 year old group,and there was no statistically significant difference between the age groups of both genders(P>0.05).The number of patients who were overweight or obese(97 and 52,respectively)accounted for more than 60%of the total patient population.2 Distribution of four diagnostic features:Dry mouth and throat,fatigue,numbness of limbs,insomnia,backache and leg pain,obesity,spontaneous sweating,itchy skin,dry stool,restlessness,cold limbs,bitter taste,and chest tightness were the main symptoms among the 246 DPN patients.Greasy tongue coating,yellow tongue coating,dark red tongue,white tongue coating and red tongue were the main tongue appearances.String-like pulse,slippery pulse,thin pulse and sunken pulse were the main pulse appearances.3 Factor analysis results:Five common factors were extracted.The first factor,characterized by numbness of limbs,spasm of hands and feet,itching of skin,dryness of skin,abnormal sensation of limbs,restlessness,inconvenient movement of limbs,cold limbs,shortness of breath and laziness to speak,frequent and scanty urination,hot hands and feet,distension and fullness of the epigastric region,characterized by the wind.The second factor,characterized by shortness of breath,fatigue,dizziness,dark lips,dull complexion,palpitations,chest tightness,poor appetite,loose stools,vivid dreams,and insomnia,characterized by Qi deficiency and blood stasis.The third factor,characterized by heaviness in the head,backache and leg pain,obesity,sticky mouth,abdominal distension,dizziness,vivid dreams,infrequent urination,and restlessness,characterized by phlegm and dampness.The fourth factor,characterized by bitter taste,abdominal distension,soreness in limbs,sticky mouth,limited limb extension,and frequent urination at night,characterized by Qi stagnation.The fifth factor,characterized by excessive hunger,easy satiety,spontaneous sweating,and frequent urination,characterized by Yin deficiency and Qi deficiency.The 246 DPN patients were classified into the corresponding factor based on their factor scores,and the DPN syndrome elements were ultimately ranked from high to low in the following order:Qi deficiency and blood stasis(40.24%),wind(24.80%),Qi stagnation(14.63%),phlegm and dampness(14.23%),and Qi and Yin deficiency(6.10%).Study 2:1 Analysis of differences:In terms of general information,there was no statistically significant difference in gender,age,BMI,and course of disease between the group with wind syndrome and the group without wind syndrome(P>0.05).In terms of tongue appearance,there was a statistically significant difference between the two groups in comparing pale and dull tongue and teeth-marked tongue(P<0.05),with fewer patients in the group with wind syndrome having these tongue appearances.In terms of pulse appearance,there was a statistically significant difference between the two groups in comparing string-like pulse(P<0.05),with fewer patients in the group with wind syndrome having this pulse appearance.In terms of laboratory indicators,there was no statistically significant difference in FPG,2hPG,FCP,HbA1c,TC,TG,LDL-C,HDL-C,HCY,ALT,AST,Cr,UA,and TSH levels between the two groups(P>0.05).UACR showed a statistically significant difference between the group with wind syndrome and the group without wind syndrome(P<0.05),with U ACR levels in the group with wind syndrome ranging from 5.30-458.90 mg/g,with a median of 86.10(271.90)mg/g,while the group without wind syndrome had UACR levels ranging from 3.40-700.90 mg/g,with a median of 32.90(143.40)mg/g.The UACR level in the group with wind syndrome was significantly higher than that in the group without wind syndrome.2 Results of predictive model construction:The input layer of the model consisted of numbness of limbs,spasm of hands and feet,itching of skin,dryness of skin,abnormal sensation of limbs,restlessness,inconvenient movement of limbs,cold limbs,shortness of breath and laziness to speak,frequent and scanty urination,hot hands and feet,distension and fullness of the epigastric region,pale and dull tongue,teeth-marked tongue,and string-like pulse.The output layer was whether the patient had wind syndrome("1" for wind syndrome and "0" for non-wind syndrome).The neural network predictive model constructed in this study had a training sample accuracy of 84.4%,a test sample accuracy of 85.1%,and an area under the ROC curve(AUC)of 0.901,indicating good predictive efficacy of the model.The various factors were ranked according to their standardized importance,with abnormal sensation of limbs having the highest importance(100.0%),followed by hot hands and feet(88.5%),teethmarked tongue(71.9%),dryness of skin(53.6%),itching of skin(44.9%),distension and fullness of the epigastric region(40.8%),pale and dull tongue(35.3%),spasm of hands and feet(29.1%),cold limbs(23.5%),restlessness(21.9%),numbness of limbs(20.1%),frequent and scanty urination(19.9%),shortness of breath and laziness to speak(9.6%),string-like pulse(7.9%)and inconvenient movement of limbs(6.1%).ConclusionStudy 1:The main types of syndrome differentiation in diabetic peripheral neuropathy are Qi deficiency and blood stasis,wind,Qi stagnation,phlegm and dampness,and Qi and Yin deficiency.The wind syndrome of DPN is worth attention and research,and wind pathogen can reflect the clinical features of the disease.Study 2:Compared with non-wind syndrome patients,wind syndrome patients have higher levels of UACR,which reflects the pathogenesis of wind pathogen.In this study,the neural network prediction model has a good effect on predicting wind syndrome in diabetic peripheral neuropathy.