No abstract

No abstract

a 北里大学東洋医学総合研究所(〒108 8642 東京都港区白金 5 9 1)， b 北里大学薬学部(〒108 8642 東京都港区白金 5 9 1)， c 東京理科大学薬学部(〒278 8510 千葉県野田市山崎 2641) ， d 株式会社常磐植物化学研究所(〒285 0801 千葉県 佐倉市木野子 158 番地) ， e 松山大学薬学部(〒790 8578 愛媛県松山市文京町 4 2) ， f 国立医薬品食品衛生研究所(〒210 9501 川崎市川崎区殿町 3 25 26) 現所属： † 東邦大学薬学部(〒274 8510 千葉県船橋市三山 2 2 1)  e-mail: hyuga-s＠insti.kitasato-u.ac.jp

Ephedra Herb is an important crude drug; it is used in various Traditional Japanese Medicine (Kampo) formulations. Its significant pharmacological effects have been believed to be attributed to ephedrine and pseudoephedrine, which sometimes induce adverse effects. On the other hand, it has been reported that some of these pharmacological effects are not dependent on ephedrine or pseudoephedrine. Ephedrine alkaloids-free Ephedra Herb extract has been newly developed. It has been reported to have analgesic, anti-influenza, and antimetastatic effects. This clinical trial was aimed at verifying the noninferiority of EFE's safety compared to that of Ephedra Herb extract (EHE) in humans. This was a single-institution, double-blinded, randomized, two-drug, two-stage, crossover comparative study. Twelve healthy male subjects were equally and randomly allocated into two groups: prior administration of EFE (EFE-P) and prior administration of EHE (EHE-P). In Stage 1, EFE and EHE were orally administered to the EFE-P and EHE-P groups, respectively, for six days. After a 4-week washout period, Stage 2 was initiated wherein the subjects were given a study drug different from Stage 1 study drug for six days. Eleven adverse events with a causal relationship to the study drugs (EHE: 8; EFE: 3) were noted; all events were mild in severity. With regard to the incidence of adverse events, EHE and EFE administration, respectively, accounted for 4 cases (out of 12 subjects, similarly below) and 1 case of increased pulse rate (p=0.32) and 3 cases and 1 case of insomnia (p=0.59). Further, there was one case of hot flashes (p=1.00) due to EFE administration and one case of dysuria (p=1.00) due to EHE administration. There were no significant differences in the incidences of adverse events between EHE administration and EFE administration. Therefore, we concluded that EFE is not inferior to EHE in terms of safety.

This paper focuses on the attempt to formulate the prescription prediction logic based on the medical data analysis towards the future computer-assisted-diagnosis for Kampo medicine. We constructed and evaluated prediction models for some frequently-used prescriptions using six kinds of machine learning algorithms including artificial neural network, multinomial logit, random forest, support vector machine, knearest neighbor, and decision tree. The possibility of prescription prediction and the necessary amount of data required for robust prediction are clarified.