FGF2和BMP-2对Ⅲ、Ⅳ型慢性骨髓炎患者病灶清除联合封闭负压引流治疗预后的预测价值

Predictive value of FGF2 and BMP-2 in the treatment prognosis of type Ⅲ-Ⅳ chronic osteomyelitis patients treated with lesion clearance and closed negative pressure drainage

摘要:

目的 探讨成纤维细胞生长因子2（FGF2）和骨形态发生蛋白-2（BMP-2）对Ⅲ、Ⅳ型慢性骨髓炎患者病灶清除联合封闭负压引流治疗预后的预测价值。方法 前瞻性选取2020年1月—2021年12月在新疆医科大学第一附属医院住院治疗的105例Ⅲ、Ⅳ型慢性骨髓炎患者作为研究对象，均接受病灶清除联合封闭负压引流治疗，按不同治疗预后分为疗效好组75例（71.4%）和疗效差组30例（28.6%）。比较两组患者的临床资料、血清炎症因子、FGF2及BMP-2表达水平；采用多因素Logistic回归分析影响患者预后的独立危险因素，分析FGF2及BMP-2与预后的关系；构建相关列线图模型，绘制受试者工作特征（ROC）曲线和决策曲线，分析FGF2、BMP-2及联合预测模型的预测效能和净收益率。结果 疗效差组Ⅳ型Cierny-Mader分型及窦道形成患者占比高于疗效好组（P <0.05）。疗效差组患者术前红细胞沉降率（ESR）、C反应蛋白（CRP）及肿瘤坏死因子-α（TNF-α）水平均高于疗效好组（P <0.05），疗效差组患者术前FGF2及BMP-2水平均低于疗效好组（P <0.05）。多因素Logistic回归分析结果显示，Cierny-Mader分型［O^R =5.036（95% CI：1.369，9.894）］、窦道形成［O^R =2.987（95% CI：1.156，7.247）］、FGF2［O^R =0.446（95% CI：0.129，0.735）］和BMP-2［O^R =0.485（95% CI：0.212，0.738）］为影响Ⅲ、Ⅳ型慢性骨髓炎患者预后的危险因素（P <0.05）。基于FGF2、BMP-2构建预测预后的列线图模型，校准曲线显示，Ⅲ、Ⅳ型慢性骨髓炎患者治疗疗效的预测值与实际观测值十分接近；ROC曲线分析结果显示，Cierny-Mader分型、窦道形成、FGF2及BMP-2预测预后的曲线下面积分别为0.783（95% CI：0.754，0.875）、0.752（95% CI：0.761，0.893）、0.823（95% CI：0.789，0.885）及0.811（95% CI：0.797，0.875），FGF2及BMP-2的最佳截断值分别为18.9 ng/L和113.5 ng/L，4者联合预测的曲线下面积为0.952（95% CI：0.896，0.991）；决策曲线分析结果显示，Cierny-Mader分型、窦道形成、FGF2及BMP-2预测预后均具有良好的净收益率，并且联合预测的总体净收益率高于单一指标。结论 基于Cierny-Mader分型、窦道形成、FGF2及BMP-2 4个指标构建的列线图模型能准确预测Ⅲ、Ⅳ型慢性骨髓炎患者病灶清除联合封闭负压引流治疗预后。

Abstract:

Objective To investigate the predictive value of fibroblast growth factor 2 (FGF2) and bone morphogenetic protein-2 (BMP-2) for the prognosis of lesion clearance combined with closed negative pressure drainage treatment in patients with type Ⅲ and Ⅳ chronic osteomyelitis.Methods A total of 105 patients with type Ⅲ and Ⅳ chronic osteomyelitis treated at the First Affiliated Hospital of Xinjiang Medical University from January 2020 to December 2021 were prospectively included as research subjects. All patients received lesion clearance combined with closed negative pressure drainage treatment and were divided into a good efficacy group (75 cases, 71.4%) and a poor efficacy group (30 cases, 28.6%) according to different treatment outcomes. The clinical data, serum inflammatory factors, and FGF2 and BMP-2 expression levels were compared between the two groups. Multivariate logistic regression analysis was used to identify independent risk factors affecting patient prognosis and to analyze the relationship between FGF2, BMP-2, and prognosis. Receiver operating characteristic (ROC) curves and decision curves were constructed to analyze the predictive efficacy and net benefit of FGF2, BMP-2, and the combined prediction model.Results The proportion of type Ⅳ Cierny-Mader classification and sinus formation in the poor efficacy group was higher than that in the good efficacy group (P < 0.05). The levels of erythrocyte sedimentation rate (ESR), C-reactive protein (CRP), and tumor necrosis factor-alpha (TNF-α) in the poor efficacy group were higher than those in the good efficacy group (P < 0.05). The levels of FGF2 and BMP-2 in the poor efficacy group were lower than those in the good efficacy group (P < 0.05). Multivariate logistic regression analysis showed that Cierny-Mader classification [O^R = 5.036 (95% CI: 1.369, 9.894)], sinus formation [O^R = 2.987 (95% CI: 1.156, 7.247)], FGF2 [O^R = 0.446 (95% CI: 0.129, 0.735)], and BMP-2 [O^R = 0.485 (95% CI: 0.212, 0.738)] were independent risk factors affecting the prognosis of patients with type Ⅲ and Ⅳ chronic osteomyelitis (P < 0.05). A prediction model based on FGF2 and BMP-2 was constructed, and the calibration curve showed that the predicted values of treatment efficacy in patients with type Ⅲ and Ⅳ chronic osteomyelitis were close to the actual observed values. The ROC curve analysis showed that the areas under the curve for predicting the prognosis of Cierny-Mader classification, sinus formation, FGF2, and BMP-2 were 0.783 (95% CI: 0.754, 0.875), 0.752 (95% CI: 0.761, 0.893), 0.823 (95% CI: 0.789, 0.885), and 0.811 (95% CI: 0.797, 0.875), respectively. The optimal cutoff values for FGF2 and BMP-2 were 18.9 ng/L and 113.5 ng/L, respectively, and the area under the curve for the combined prediction model was 0.952 (95% CI: 0.896, 0.991). Decision curve analysis showed that Cierny-Mader classification, sinus formation, FGF2, and BMP-2 had good net benefit rates for predicting prognosis, and the overall net benefit rate of the combined prediction was higher than that of a single index.Conclusion The prediction model based on Cierny-Mader classification, sinus formation, FGF2, and BMP-2 can accurately predict the prognosis of patients with type Ⅲ and Ⅳ chronic osteomyelitis treated with lesion clearance combined with closed negative pressure drainage.