骨科手术输血后不良反应的影响因素及预测模型构建与分析

Factors affecting the adverse reactions to blood transfusion during orthopedic surgery and prediction model development

摘要:

目的 探讨骨科手术输血后不良反应的影响因素，并在其基础上初步构建预测模型，为骨科手术输血后不良反应风险评估提供参考依据。方法 前瞻性选取2019年1月—2022年6月四川省人民医院行骨折手术治疗、符合输血指征且骨科手术输血后出现不良反应的60例患者作为观察组，按照1∶1随机纳入同期该院行骨折手术治疗且需要骨折手术输血的60例患者作为对照组。比较两组患者性别、年龄、吸烟史、骨折类型、血型分布、输注血制品类型、术中出血量、术中输血量、身体质量指数、病史、麻醉分级、输血史、过敏史、发血至输血时间、血清嗜酸性粒细胞趋化因子、C反应蛋白（CRP）、肿瘤坏死因子α（TNF-α）、细胞间黏附分子1（ICAM-1）、基质金属蛋白酶9（MMP-9）、血小板计数（PLT）、血红蛋白（Hb）水平等，采用单因素分析及非条件多因素Logistic逐步回归分析筛选骨科手术输血后不良反应的危险因素。根据非条件多因素Logistic逐步回归分析结果，初步构建骨科手术输血后不良反应预测模型，并以受试者工作特征（ROC）曲线分析预测模型的诊断价值。结果 两组患者性别、年龄、吸烟史、骨折类型、血型分布、输注血制品类型、术中出血量、术中输血量、体质量指数、病史、麻醉分级、Hb、PLT比较，差异无统计学意义（P >0.05）。观察组输血史、过敏史、发血至输血> 0.5 h占比高于对照组（P <0.05）。观察组嗜酸性粒细胞趋化因子、CRP、TNF-α、ICAM-1、MMP-9水平高于对照组（P <0.05）。ROC曲线分析显示，TNF-α的曲线下面积（AUC）、敏感性、特异性均最高，分别为0.900（95% CI：0.883，0.915）、92.7%（95% CI：0.685，0.749）、90.9%（95% CI：0.629，0.741）。多因素Logistic回归分析结果显示：输血史［O^R=2.856（95% CI：1.185，6.883）］、过敏史［O^R=4.334（95% CI：1.423，13.200）］、发血至输血时间> 0.5 h［O^R=3.020（95% CI：1.329，6.863）］、CRP ≥ 12.980 mg/L［O^R=4.699（95% CI：1.478，14.940）］、嗜酸性粒细胞趋化因子≥ 93.512 pg/mL［OR=3.124（95% CI：1.171，8.334）］、TNF-α≥ 103.434 ng/mL［O^R=2.983（95% CI：1.501，5.928）］、ICAM-1≥ 225.901 ng/mL［O^R=2.275（95% CI：1.210，4.277）］和MMP-9 ≥ 401.915 ng/mL［O^R=3.962（95% CI：1.521，10.32）］是骨科手术输血后不良反应的危险因素（P <0.05）。根据多因素Logistic回归分析结果，将临床相关指标CRP、嗜酸性粒细胞趋化因子、TNF-α、ICAM-1、MMP-9纳入预测模型，Logit （P）=-32.180+1.547×CRP+1.139×嗜酸性粒细胞趋化因子+1.093 ×TNF-α+ 0.822×ICAM-1+1.377×MMP-9。结果表明预测模型预测骨科手术输血后不良反应的临界值为31.764，AUC为0.937（95% CI：0.908，0.967），敏感性为94.2%（95% CI：0.769，0.901），特异性为96.4%（95% CI：0.776，0.925）。预测模型敏感性优于各项指标单独预测。结论 CRP、嗜酸性粒细胞趋化因子、TNF-α、ICAM-1、MMP-9与骨科手术输血后不良反应有关，因此临床进行骨科手术时可参考影响因素及其预测模型的风险预警作用，给予针对性干预，以降低骨科手术输血后不良反应风险。

Abstract:

Objective To determine the factors affecting the adverse reactions to blood transfusion during orthopedic surgery, on the basis of which to develop a prediction model, so as to provide a reference for the risk assessment of adverse reactions to blood transfusion during orthopedic surgery.Methods From January 2019 to June 2022, sixty patients who underwent surgical treatment for fractures in our hospital with indications for blood transfusion and developed adverse reactions to blood transfusion during orthopedic surgery were prospectively included as the observation group. Another 60 patients undergoing surgery for fractures in need of blood transfusion in our hospital during the same period were included as the control group.Results The sex composition, age, the proportion of patients with a history of smoking, fracture patterns, the distribution of blood types, the type of blood products, the intraoperative blood volume, the intraoperative blood transfusion volume, body mass index, the medical history, the American Society of Anesthesiology score, and the levels of hemoglobin (Hb) and platelet count (PLT) were not different between the two groups (P >0.05). The proportions of patients with a history of blood transfusion and the allergic history, and the frequency of the time from bleeding to transfusion longer than 0.5 h in the observation group were higher than those in the control group (P <0.05). The levels of eotaxin, C-reactive protein (CRP), tumor necrosis factor-α (TNF-α), and intercellular adhesion molecule-1 (ICAM-1), and matrix metalloproteinase-9 (MMP-9) in the observation group were higher than those in the control group (P <0.05). The receiver operating characteristic (ROC) curve analysis revealed that the area under the ROC curve (AUC), sensitivity, and specificity of TNF-α were the highest, with them being 0.900 (95% CI: 0.883, 0.915), 0.927 (95% CI: 0.685, 0.749), and 0.909 (95% CI: 0.629, 0.741), respectively. The multivariable Logistic regression analysis showed that the history of blood transfusion [O^R = 2.856 (95% CI: 1.185, 6.883) ], allergic history [O^R = 4.334 (95% CI: 1.423, 13.200) ], the time from bleeding to transfusion longer than 0.5 h [O^R = 3.020 (95% CI: 1.329, 6.863) ], CRP ≥ 12.980 mg/L [O^R = 4.699 (95% CI: 1.478, 14.94) ], eotaxin ≥ 93.512 pg/mL [O^R = 3.124 (95% CI: 1.171, 8.334) ], TNF-α ≥ 103.434 ng/mL [O^R = 2.983 (95% CI: 1.501, 5.928) ], ICAM-1 ≥ 225.901 ng/mL [O^R = 2.275 (95% CI: 1.210, 4.277) ] and MMP-9 ≥ 401.915 ng/mL [O^R = 3.962 (95% CI: 1.521, 10.32) ] were risk factors for adverse reactions to blood transfusion during orthopedic surgery (P < 0.05). According to the results of multivariable Logistic regression analysis, CRP, eotaxin, TNF-α, ICAM-1, and MMP-9 were included as variables in the prediction model Logit (P) = -32.180 + 1.547 × CRP + 1.139 × Eotaxin + 1.093 × TNF-α + 0.822 × ICAM-1 + 1.377 × MMP-9. The cutoff value of the prediction model was 31.764 for determining the adverse reactions to blood transfusion during orthopedic surgery, with the AUC being 0.937 (95% CI: 0.908, 0.967), the sensitivity being 0.942 (95% CI: 0.769, 0.901), and the specificity being 0.964 (95% CI: 0.776, 0.925), and the sensitivity of the prediction model was better than that of these indicators detected alone.Conclusions CRP, eotaxin, TNF-α, ICAM-1 and MMP-9 are associated with adverse reactions to blood transfusion during orthopedic surgery. Therefore, these factors could be established as red flags for the orthopedic surgery and be targeted to reduce the risk of adverse reactions to blood transfusion during orthopedic surgery.