Benedetto Umberto, Raja Shahzad G
Department of Cardiac Surgery, Harefield Hospital, London, United Kingdom.
Department of Cardiac Surgery, Harefield Hospital, London, United Kingdom.
J Thorac Cardiovasc Surg. 2014 Nov;148(5):2390-2396.e1. doi: 10.1016/j.jtcvs.2014.05.017. Epub 2014 May 15.
The effectiveness of the routine retrosternal placement of a gentamicin-impregnated collagen sponge (GICS) implant before sternotomy closure is currently a matter of some controversy. We aimed to develop a scoring system to guide decision making for the use of GICS to prevent deep sternal wound infection.
Fast backward elimination on predictors, including GICS, was performed using the Lawless and Singhal method. The scoring system was reported as a partial nomogram that can be used to manually obtain predicted individual risk of deep sternal wound infection from the regression model. Bootstrapping validation of the regression models was performed.
The final populations consisted of 8750 adult patients undergoing cardiac surgery through full sternotomy during the study period. A total of 329 patients (3.8%) received GICS implant. The overall incidence of deep sternal wound infection was lower among patients who received GICS implant (0.6%) than patients who did not (2.01%) (P=.02). A nomogram to predict the individual risk for deep sternal wound infection was developed that included the use of GICS. Bootstrapping validation confirmed a good discriminative power of the models.
The scoring system provides an impartial assessment of the decision-making process for clinicians to establish if GICS implant is effective in reducing the risk for deep sternal wound infection in individual patients undergoing cardiac surgery through full sternotomy.
在胸骨切开术关闭前常规在胸骨后放置庆大霉素浸渍胶原海绵（GICS）植入物的有效性目前存在一些争议。我们旨在开发一种评分系统，以指导使用GICS预防深部胸骨伤口感染的决策。
使用Lawless和Singhal方法对包括GICS在内的预测因素进行快速向后消除。该评分系统以部分列线图的形式报告，可用于从回归模型中手动获取深部胸骨伤口感染的个体预测风险。对回归模型进行了自举验证。
最终研究对象包括研究期间通过全胸骨切开术进行心脏手术的8750例成年患者。共有329例患者（3.8%）接受了GICS植入。接受GICS植入的患者深部胸骨伤口感染的总体发生率（0.6%）低于未接受植入的患者（2.01%）（P=0.02）。开发了一个预测深部胸骨伤口感染个体风险的列线图，其中包括GICS的使用情况。自举验证证实了模型具有良好的判别能力。
该评分系统为临床医生评估GICS植入物在降低通过全胸骨切开术进行心脏手术的个体患者深部胸骨伤口感染风险方面是否有效提供了公正的决策过程评估。
