摘要

摘要：
目的：探究鲁拉西酮诱发药源性锥体外系反应的影响因素并建立预测模型。

方法：回顾性分析我院使用过鲁拉西酮治疗的住院患者，根据服药后是否发生锥体外系反应分为观察组和对照组，比较两组一般资料、是否合并使用多巴胺受体拮抗剂、5-羟色胺调节剂、锂盐、丙戊酸盐、日最大服用剂量、尿酸水平、肌酐水平、总胆红素、谷草转氨酶、谷丙转氨酶以及联合使用的抗精神病药数；使用R语言双向逐步回归筛选预测变量并构建预测模型；采用ROC曲线分析预测模型的预测效能。

结果：共纳入388例患者，其中有104例（26.8%）患者发生锥体外系反应，双向逐步回归纳入日最大服用剂量、总胆红素、肌酐水平、尿酸水平、BMI、合用锂盐、合用抗精神病药数、疾病诊断类型、病程9个预测变量构成最优模型，其中病程、总胆红素、尿酸水平是鲁拉西酮致锥体外系反应保护因素；肌酐水平、疾病诊断类型、日最大服药剂量是其风险因素; ROC曲线下面积（AUC）为0.811（95%CI0.759~0.862）。

结论：病程、总胆红素、尿酸水平是鲁拉西酮致锥体外系反应保护因素；肌酐水平、疾病诊断类型、日最大服药剂量是其风险因素；根据预测变量构建的预测模型具有良好的预测效能。

关键词：关键词：鲁拉西酮；锥体外系反应；影响因素；预测模型

ABSTRACT

Abstarct:
Objective To investigate the influencing factors of lurasidone-induced pharmacogenic extrapyramidal reactions.

Methods The inpatients treated with lurasidone in our hospital were retrospectively analyzed and divided into an observation group and a control group according to the occurrence of extrapyramidal reactions after taking lurasidone, comparing the two groups' general information, the combined use of dopamine receptor antagonists, 5-hydroxytryptamine, lithium, valproate, maximum daily dose, uric acid level, creatinine level, total bilirubin, aspartate transaminase, alanine transaminase and the number of antipsychotic used in combination; the influence of factors and predictive models were analyzed by two-way step-by-step regression using the R language; the predictive effectiveness of predictive models was analyzed using ROC curves.

Results A total of 388 patients were included, of whom 104 (26.8%) had extrapyramidal reactions. Bidirectional stepwise regression incorporating nine predictor variables of maximum daily dose, total bilirubin, creatinine level, uric acid level, BMI, coadministration of lithium, coadministration of number of antipsychotics, type of disease diagnosis, and duration of illness constituted the optimal model, in which duration of illness, total bilirubin, and uric acid level were the extrapyramidal reactions protective factors of lurasidone; creatinine level, type of disease diagnosis, and maximum daily dose of the drug were risk factors；and the area under the ROC curve（AUC）was 0.811（95%CI0.759~0.862）.

Conclusion  s Disease duration, total bilirubin, and uric acid levels were protective factors for extrapyramidal reactions to lurasidone; creatinine level, type of disease diagnosis, and maximum daily dose of the drug were risk factors; and prediction models constructed on the basis of the predictor variables had good predictive efficacy.

Key words: Keywords lurasidone; Extrapyramidal Symptoms; Influencing factors; Predictive model