To Predict Nephrotoxicity Of Chinese Herbal Medicines Based On Traditional Chinese Herbal Medicine Theory

Objective: To study the correlation of nephrotoxicity of CHMs with the drug-property(four properties,five flavors,channel tropism)and efficacy for the establishment of prediction models of nephrotoxicity based on Traditional Chinese Herbal Medicine theory,and to evaluate the models.Methods: A comprehensive literature search was performed for CHMs with nephrotoxicity.And the CHMs without nephrotoxicity were collected from Chinese Materia Medica·Featured which removed CHMs with nephrotoxicity.The collected data of Part 1,from Chinese Materia Medica,covered the drug-property(containing the four properties,five flavors and channel tropism)of the included herbs,but excluding the CHMs without four properties or five flavors or channel tropism.And the data of Part 2,from Chinese Materia Medica,covered the efficacy of the included herbs,but excluding the CHMs without efficacy.The correlation of nephrotoxic or non-nephrotoxic CHMs with the drug-property(four properties,five flavors and channel tropism)and the efficacy was analyzed separately.Then the related variables were input to the database.The LRM,NNM and BNM were applied to screen the variables to be input.The ROC curves of the three models were drawn and the AUC were calculated,and the accuracy of the three prediction models was evaluated.Results: Part 1 CHMs with or without nephrotoxicity was correlated with the four properties and five flavors(P<0.05).LRM,NNM and BNM were set up.Five variables,which were hot and neutral of the properties;pungent,bitter and sweet of the flavors,were selected to establish the LRM.And the pungent,bitter,hot were risk factors(OR>1),but the sweet,neutral were protective factors(OR<1).The four variables,hot,pungent,warm and bitter were selected by NNM.And the sweet and neutral were selected by BNM.However,the top 1 of predictive variables in NNM was hot and but in BNM was sweet.The AUC of ROC curves of three models were 0.703(LRM),0.739(NNM),0.661(BNM).Part 2 CHMs with or without nephrotoxicity was correlated with the efficacy(P<0.05),with the coefficient being 0.338.One-factor Logistic regression analysis was performed to screen significant variables for models.The KW(Kill Worms),CBSB(Cool the Blood and Stanch Bleeding)and NHTM(Nourish the Heart and Tranquilize the Mind),as the important factors of nephrotoxicity,were all selected by LRM,NNM and BNM.And the KW was risk factor(OR>1),but the CBSB and NHTM were protective factors(OR<1).The AUC of three models showed the prediction ability of NNM(0.788)was better than the LRM(0.724)and BNM(0.742).Conclusions: 1.Combination of Traditional Chinese Herbal Medicine theory with modern mathematical statistics method,this study had been predicted and tested the important variables affecting the nephrotoxicity of CHMs from the four factors,five flavors and efficacy.2.Three models had been set up for predicting nephrotoxicity of CHMs.The prediction ability of NNM was better than the LRM and BNM,but the explanatory ability of LRM was the best.3.The modern mathematics method of data statistics might be a promising method for the studying nephrotoxicity and toxicology of CHMs.