To Set Up a Logistic Regression Prediction Model for Hepatotoxicity of Chinese Herbal Medicines Based on Traditional Chinese Medicine Theory

• Main Authors: Hongjie Liu, Tianhao Li, Lingxiu Chen, Sha Zhan, Meilan Pan, Zhiguo Ma, Chenghua Li, Zhe Zhang
• Format: Article
• Language: English
• Published: Hindawi Limited 2016-01-01
• Series: Evidence-Based Complementary and Alternative Medicine
• Online Access: http://dx.doi.org/10.1155/2016/7273940
• ISSN: 1741-427X; 1741-4288

Aims. To establish a logistic regression (LR) prediction model for hepatotoxicity of Chinese herbal medicines (HMs) based on traditional Chinese medicine (TCM) theory and to provide a statistical basis for predicting hepatotoxicity of HMs. Methods. The correlations of hepatotoxic and nonhepatotoxic Chinese HMs with four properties, five flavors, and channel tropism were analyzed with chi-square test for two-way unordered categorical data. LR prediction model was established and the accuracy of the prediction by this model was evaluated. Results. The hepatotoxic and nonhepatotoxic Chinese HMs were related with four properties (p<0.05), and the coefficient was 0.178 (p<0.05); also they were related with five flavors (p<0.05), and the coefficient was 0.145 (p<0.05); they were not related with channel tropism (p>0.05). There were totally 12 variables from four properties and five flavors for the LR. Four variables, warm and neutral of the four properties and pungent and salty of five flavors, were selected to establish the LR prediction model, with the cutoff value being 0.204. Conclusions. Warm and neutral of the four properties and pungent and salty of five flavors were the variables to affect the hepatotoxicity. Based on such results, the established LR prediction model had some predictive power for hepatotoxicity of Chinese HMs.