]目的•建立预测骨折后接受手术治疗患者发生异位骨化(heterotopicossification, HO)风险的nomogram临床评分 系统。方法•选取2018年8月至2020年10月在浙江省舟山医院骨科或脑外科住院的骨折并接受手术的患者共124例, 其中 男性81例, 女性43例, 年龄25〜91岁[平均年龄(56.48±15.45)岁]。收集患者入院临床资料, 包括性别、年龄、是否合 并脑外伤、麻醉方式、手术持续时间、术中出血量等。采集患者骨折后1d、7d、15d共3个时间点的外周血, ELISA法检测 血清骨成形蛋白 2 (bone morphogenetic protein 2, BMP-2)、BMP-4、BMP-5、BMP-7及细胞因子白介素-4 (interleukin 4, IL-4)、IL-6、IL-10、丫干扰素(interferon y, IFN-丫)、转化生长因子-卩(transforming growth factor-p, TGF-p)的浓度。按 随访结果将患者分为可见HO组和未见HO组, 以0=0.15从临床和实验室指征中挑选2组间差异具有统计学意义的变量纳入 单变量Cox回归模型;再以0=0.05从单变量Cox回归模型中挑选变量纳入多变量Cox回归模型来筛选患者发生异位骨化的 独立风险因子, 并计算风险率(HR);最后通过R语言中的rms包进行nomogram可视化输出。结果•随访共发现13例异位 骨化患者, 随访发现时间中位数为71 (38, 292) d。通过组间比较、单变量Cox回归分析、多变量Cox回归分析得到合并 脑外伤(HR=2.932, P=0.038),手术持续时间(HR=1.005, P=0.007),术中出血量(HR=1.004, P=0.022), 15 d时的 BMP-2 (HR=1.009, P=0.044)、BMP-4 (HR=1.004, P=0.011)、TGF-p (HR=1.011, P=0.046),以及7d时的BMP-7 (HR= 1.004, P=0.008)共7个对异位骨化结局有显著影响的独立因子并输出得到骨折后90d、180 d. 360 d患者异位骨化发生概 率预测的nomogram评分系统。结论•预测骨折后手术患者异位骨化风险的nomogram评分系统的7个评分指标分别为是否 合并脑外伤, 手术持续时间, 术中出血量, 骨折后15d外周血BMP-2、BMP-4、TGF-p水平, 以及骨折后7d外周血BMP-7 水平. Objective* To establish a nomogram scoring system to predict morbidity of heterotopic ossification (HO) in the patients undergoing surgery after fracture. Methods • From August 2018 to October 2020, 124 patients with fractures from department of orthopaedics or brain surgery, including 81 males and 43 females with an average age of (56.48±15.45) years old (ranging from 25-91 years old), were enrolled in the study. Clinical features including gender, age, with brain trauma or not, anesthetic mode, operation duration time, and blood loss during operation were collected. Peripheral blood samples were collected at three time points, i.e. 1 d, 7 d and 15 d after fracture, and the sera levels of bone morphogenetic protein 2 (BMP-2), BMP-4, BMP-5, BMP-7, interleukin 4 (IL-4), IL-6, IL-10, interferon y (IFN-y), and transforming growth factor-p (TGF-p) were detected by ELISA. All the patients were divided into HO group and non-HO group according to follow-up outcomes. The variables with statistically significant differences (a =0.15) between the two groups selected from clinical and laboratory indications were included in the univariate Cox regression model. Then the variables selected from univariate Cox regression model (a =0.05) were incorporated into multivariate Cox regression model to screen the independent risk factors of HO and calculate the risk ratios (HR). Finally, the nomogram scoring system was output through rms package in R language. Results • The follow-up observation found 13 HO cases after a median time of 71 (38, 292) d. Seven independent risk factors for HO, i. e. brain trauma (HR=2.932, P=0.038), operation duration time (HR= 1.005, P=0.007), blood loss during operation (HR=1.004, P=0.022), BMP-2 (15d) (HR=1.009, P=0.044), BMP-4 (15 d) (HR=1.004, P=0.011), TGF-卩(15 d) (HR=1.011, P=0.046), and BMP-7 (7 d) (HR=1.004, P=0.008), were selected by comparison between groups, univariate COX regression and multivariate COX regression sequentially. The nomogram to predict morbidity of HO after 90 d, 180 d, and 360 d was generated. Conclusion • The seven scoring indexes of nomogram scoring system for predicting the risk of HO in the patients undergoing surgery after fracture are with brain trauma or not, operation duration time, blood loss during operation, levels of BMP-2, BMP-4 and TGF-卩 in peripheral blood 15 d after fracture and BMP-7 7 d after fracture. [ABSTRACT FROM AUTHOR]