Department of Orthopaedic Surgery, Columbia University Irving Medical Center, New York, NY.
Department of Epidemiology, Mailman School of Public Health, Columbia University, New York, NY.
J Bone Joint Surg Am. 2022 Feb 16;104(4):364-375. doi: 10.2106/JBJS.21.00751.
Despite tremendous efforts, the incidence of surgical site infection (SSI) following the surgical treatment of pediatric spinal deformity remains a concern. Although previous studies have reported some risk factors for SSI, these studies have been limited by not being able to investigate multiple risk factors at the same time. The aim of the present study was to evaluate a wide range of preoperative and intraoperative factors in predicting SSI and to develop and validate a prediction model that quantifies the risk of SSI for individual pediatric spinal deformity patients.
Pediatric patients with spinal deformity who underwent primary, revision, or definitive spinal fusion at 1 of 7 institutions were included. Candidate predictors were known preoperatively and were not modifiable in most cases; these included 31 patient, 12 surgical, and 4 hospital factors. The Centers for Disease Control and Prevention definition of SSI within 90 days of surgery was utilized. Following multiple imputation and multicollinearity testing, predictor selection was conducted with use of logistic regression to develop multiple models. The data set was randomly split into training and testing sets, and fivefold cross-validation was performed to compare discrimination, calibration, and overfitting of each model and to determine the final model. A risk probability calculator and a mobile device application were developed from the model in order to calculate the probability of SSI in individual patients.
A total of 3,092 spinal deformity surgeries were included, in which there were 132 cases of SSI (4.3%). The final model achieved adequate discrimination (area under the receiver operating characteristic curve: 0.76), as well as calibration and no overfitting. Predictors included in the model were nonambulatory status, neuromuscular etiology, pelvic instrumentation, procedure time ≥7 hours, American Society of Anesthesiologists grade >2, revision procedure, hospital spine surgical cases <100/year, abnormal hemoglobin level, and overweight or obese body mass index.
The risk probability calculator encompassing patient, surgical, and hospital factors developed in the present study predicts the probability of 90-day SSI in pediatric spinal deformity surgery. This validated calculator can be utilized to improve informed consent and shared decision-making and may allow the deployment of additional resources and strategies selectively in high-risk patients.
Prognostic Level III. See Instructions for Authors for a complete description of levels of evidence.
尽管已经付出了巨大的努力,但在小儿脊柱畸形的外科治疗后,手术部位感染(SSI)的发生率仍然令人担忧。尽管先前的研究已经报道了一些 SSI 的风险因素,但这些研究受到了无法同时调查多个风险因素的限制。本研究的目的是评估广泛的术前和术中因素对 SSI 的预测作用,并开发和验证一种量化个体小儿脊柱畸形患者 SSI 风险的预测模型。
纳入在 7 家机构中的 1 家机构接受初次、翻修或确定性脊柱融合术的小儿脊柱畸形患者。候选预测因素在术前已知且在大多数情况下不可改变;这些因素包括 31 个患者、12 个手术和 4 个医院因素。采用美国疾病控制与预防中心(CDC)定义的术后 90 天内 SSI 标准。经过多次插补和多重共线性检验后,使用逻辑回归选择预测因素以建立多个模型。数据集被随机分为训练集和测试集,进行 5 重交叉验证以比较每个模型的判别能力、校准能力和过拟合程度,并确定最终模型。通过模型开发了风险概率计算器和移动设备应用程序,以便计算个体患者发生 SSI 的概率。
共纳入 3092 例脊柱畸形手术,其中 132 例(4.3%)发生 SSI。最终模型的判别能力(受试者工作特征曲线下面积:0.76)、校准能力和不过拟合均良好。纳入模型的预测因素包括非步行状态、神经肌肉病因、骨盆内固定物、手术时间≥7 小时、美国麻醉医师协会(ASA)分级>2、翻修手术、医院每年脊柱手术例数<100 例、血红蛋白水平异常、超重或肥胖体质指数。
本研究中建立的包含患者、手术和医院因素的风险概率计算器可预测小儿脊柱畸形手术 90 天 SSI 的概率。经验证的计算器可用于提高知情同意和共同决策,并可选择性地在高危患者中部署额外的资源和策略。
预后 III 级。欲了解完整的证据水平说明,请参见作者须知。