Study On The Alternative Mechanism Of Surrogate Indicators For Clinical Efficacy Evaluation Of Chronic Heart Failure In Traditional Chinese Medicine

Objective:Based on the Logistic regression model and Cox regression model to select alternative indicators for evaluating the clinical efficacy of chronic heart failure in traditional Chinese medicine,a Bayesian network model was introduced to explore the causal relationship between the surrogate indicators and endpoint events,and the logistic regression model was introduced to analyze the interaction between the surrogate indicators.And further clarify the alternative mechanism of surrogate indicators for clinical efficacy evaluation of chronic heart failure in traditional Chinese medicine,and provide a scientific basis for constructing the replacement index of traditional Chinese medicine clinical efficacy evaluation of chronic heart failure.Method:Relying on the medical record information management system of 7 hospitals affiliated to Shandong University of Traditional Chinese Medicine,Rizhao Traditional Chinese Medicine Hospital,Weifang Traditional Chinese Medicine Hospital,Jinan Traditional Chinese Medicine Hospital,Qingdao Haici Medical Group,Zibo Traditional Chinese Medicine Hospital,Weihai Traditional Chinese Medicine Hospital,Patients were admitted from January 1,2012 to December 31,2017,and patients with a diagnosis of chronic heart failure were discharged as a study cohort.According to the "Cohort Study Questionnaire for Patients with Chronic Heart Failure in Traditional Chinese Medicine Hospital",the relevant information of patients during hospitalization and follow-up period was collected,and a database was created using Epidate 3.1 software and the data was entered,and the data was exported to Excel software.According to previous studies,cardiac death was used as the primary endpoint,NYHA cardiac function grading was used as the primary surrogate indicator,amino-terminal B-type natriuretic peptide precursor,left ventricular ejection fraction,and hemoglobin were used as secondary replacement indicators.The Bayesian network model was used to calculate the strength of the correlation between the primary and secondary surrogate indicators and the endpoint event,and to determine the surrogate ability of the surrogate index to the endpoint event.The logistic regression model was used to analyze the interaction of NYHA cardiac function grading with amino-terminal B-type natriuretic peptide precursors,left ventricular ejection fraction,and hemoglobin,and further clarifies the alternative mechanism of surrogate indicators for evaluating the clinical efficacy of chronic heart failure in Chinese medicine,and constructs surrogate indicators for evaluating the clinical efficacy of chronic heart failure in Chinese medicine.Result:1.NYHA cardiac function grading had the strongest correlation with end-point cardiac death(P= 0.44),followed by amino-terminal B-type natriuretic peptide precursor(P = 0.22),and left ventricular ejection fraction(P = 0.20),and hemoglobin(P = 0.14).The parent node of NYHA’s cardiac function grading is cardiogenic death,so the conditional probability is affected by cardiogenic death.When cardiac death occurs,the probability of NYHA cardiac function grading to grade II is 0.07,the probability of grade III is 0.57,and the probability of grade IV is 0.36;when cardiac death has not occurred,the probability of NYHA cardiac function grading to grade II is 0.24,the probability of grade III is 0.62,and the probability of grade IV is 0.14.The parent node of the amino-terminal B-type natriuretic peptide precursor is cardiac death and NYHA cardiac function grading,so the conditional probability is affected by cardiac death and NYHA cardiac function grading.When cardiac death occurs and NYHA cardiac function is grade Ⅱ,the probability of amino terminal B-type natriuretic peptide precursors being less than 450 pg / ml is 0.59,and the probability is greater than or equal to450 pg / ml is 0.41;when cardiac death has not occurred and NYHA cardiac function is grade Ⅱ,the probability of amino terminal B-type natriuretic peptide precursor being less than 450 pg / ml is 0.79,and the probability of 450 pg / ml or more is 0.21;when cardiac death occurs and NYHA cardiac function is grade Ⅲ,the probability of amino-terminal B-type natriuretic peptide precursor being less than 450 pg / ml is 0.53,and the probability of 450 pg / ml or more is 0.47.When cardiac death has not occur and NYHA cardiac function is grade III,the probability of amino-terminal B-type natriuretic peptide precursor being less than 450 pg / ml is 0.61,and the probability of 450 pg / ml or more is 0.39;when cardiogenic death occurs and NYHA cardiac function is grade IV,the probability of amino terminal B type natriuretic peptide precursor being less than 450 pg / ml is 0.44,and the probability of 450 pg / ml or more is 0.56.When cardiac death has not occurred and the NYHA cardiac function is grade IV,the probability of the amino terminal B-type natriuretic peptide precursor being less than 450 pg / ml is 0.49,and the probability of 450 pg / ml or more is 0.51.The parent node of the left ventricular ejection fraction is cardiac death and NYHA cardiac function grading,so the conditional probability is affected by cardiac death and NYHA cardiac function grading.When cardiac death occurs and the NYHA cardiac function is grade Ⅱ,the probability of the left ventricular ejection fraction value being less than 40% is 0.50,and the probability of being 40% or greater is 0.50;when cardiac death has not occurred and the cardiac function of NYHA is grade Ⅱ,the probability of the left ventricular ejection fraction value being less than 40% is 0.28,and the probability of being40% or greater is 0.72;when cardiac death occurs and NYHA cardiac function is grade Ⅲ,the probability of left ventricular ejection fraction being less than 40% is 0.46,and the probability of being 40% or greater is 0.54;when cardiac death has not occurred and the NYHA cardiac function is grade III,the probability of the left ventricular ejection fraction being less than 40% is 0.33,and the probability of being 40% or greater is 0.67;when cardiac death occurs and NYHA cardiac function is grade IV,the probability of left ventricular ejection fraction being less than 40% is 0.55,and the probability of being 40%or greater is 0.45;when no cardiac death occurs and NYHA cardiac function is grade IV,the probability of left ventricular ejection fraction less than 40% is 0.56,and the probability of being 40% or greater is 0.44.The parent node of hemoglobin is cardiac death and NYHA cardiac function grading,so the conditional probability is affected by cardiac death and NYHA cardiac function grading.When cardiac death occurs and NYHA cardiac function is grade Ⅱ,the probability of hemoglobin less than 110 g / L is 0.06,and the probability of 110 g / L or greater is 0.94;when cardiac death has not occurred and NYHA cardiac function is gradeⅡ,The probability of hemoglobin less than 110 g / L is 0.17,and the probability of 110 g /L or more is 0.83;when cardiac death occurs and NYHA cardiac function is grade III,the probability of hemoglobin less than 110 g / L is 0.34,the probability of 110 g / L or more is0.66;when cardiac death has not occurs and NYHA cardiac function is grade III,the probability of hemoglobin less than 110 g / L is 0.24,and the probability of 110 g / L or more is 0.76;when cardiac death occurs and NYHA cardiac function is grade IV,the probability of hemoglobin less than 110 g / L is 0.35,and the probability of 110 g / L or more is 0.65;when cardiac death has not occurs and the cardiac function of NYHA is grade IV,the probability of hemoglobin is less than 110 g / L is 0.29,the probability of110 g / L or more is 0.71.2.NYHA cardiac function grading and amino-terminal B-type natriuretic peptide precursor,NYHA cardiac function grading and left ventricular ejection fraction,NYHA cardiac function grading and hemoglobin all have a multiplicative interaction.Before adjusting for confounding factors,the effect value of NYHA cardiac function grading and amino terminal B-type natriuretic peptide precursor multiplication interaction term was 1.010(1.004 to 1.016),and it was statistically significant(P = 0.001).Logistic regression analysis was used to adjust the left ventricular ejection fraction(LVEF),left ventricular end-diastolic diameter(LVEDD),low-density lipoprotein(LDL-C),and hemoglobin(HB).the effect value of NYHA cardiac function grading and amino terminal B-type natriuretic peptide precursor multiplication interaction term was 1.374(1.178 to1.601),and was statistically significant(P <0.001).Before adjusting for confounding factors,the effect value of the multiplicative interaction term between NYHA cardiac function grading and left ventricular ejection fraction was 1.215(1.106 to 1.335),and it was statistically significant(P <0.001).Logistic regression analysis was used to adjust the amino terminal B-type natriuretic peptide precursor,left ventricular end-diastolic diameter(LVEDD),low-density lipoprotein(LDL-C),and hemoglobin(HB)confounding factors.The effect value of the multiplicative interaction term between NYHA cardiac function grading and left ventricular ejection fraction was 1.006(1.002 to 1.011),and was statistically significant(P = 0.009).Before adjusting for confounding factors,the effect value of NYHA cardiac function grading and hemoglobin multiplication interaction terms was 1.002(1.001 to 1.003),and it was statistically significant(P <0.001).Logistic regression analysis was used to adjust the amino-terminal B-type natriuretic peptide precursor,left ventricular end-diastolic diameter(LVEDD),low-density lipoprotein(LDL-C),and left ventricular ejection fraction(LVEF).the effect value of NYHA cardiac function grading and hemoglobin multiplication interaction terms was 1.003(1.001-1.004),and it was statistically significant(P = 0.009).Conclusion:1.Evaluation of the clinical efficacy of traditional Chinese medicine for chronic heart failure,the surrogate indicators’ ability to substitute for endpoint events in order of NYHA cardiac function grading,amino-terminal B-type natriuretic peptide precursor,left ventricular ejection fraction,and hemoglobin.2.NYHA cardiac function grading and amino-terminal B-type natriuretic peptide precursor,NYHA cardiac function grading and left ventricular ejection fraction,NYHA cardiac function grading and hemoglobin all have a multiplicative interaction.