Impact of Smoking Status on Long-Term Postoperative Wound Dehiscence and Infection Following Robotic-Assisted Spine Surgery

Introduction: Smoking is a well-documented risk factor for wound infection and dehiscence after spine surgery. However, the effect of smoking remains to be evaluated among robotic-assisted cases. This study aims to characterize the effect of smoking on long-term postoperative wound complications for robot-assisted spine surgeries.

Methods: All patients undergoing robotic-assisted spine surgery between May 2018 and March 2022 were retrospectively identified and categorized according to smoking status, defined as current, former, and non-smoker. Primary outcomes were 6-month long-term postoperative wound dehiscence and infection, with secondary analyses of intraoperative complications, comorbidities, and demographic factors. Risk factors for wound dehiscence and infection were identified using two predictive algorithms, XGBoost classifier and logistic regression. These models were evaluated through Receiver Operating Characteristic (ROC) analysis with the Area Under the Curve (AUC) calculated. Shapley additive explanations (SHAP) were applied to the best-performing model for feature importance.

Results: Of the 503 patients identified, 47 (9%) were current smokers, 198 (39%) were former smokers, and 258 (51%) were non-smokers. The non-smoker cohort was predominantly female (61%, p=0.002), while former and current smokers were more evenly split (51% and 44%). Former smokers are the oldest cohort, with a mean age of 65 years (p < 0.001), while current and non-smokers are in the 50–60-year range. Current smokers have higher comorbidity burden, with 43% having 2+ CCI index scores (p=0.049). Extremity edema is more common among current and former smokers than non-smokers (6% vs 3% vs 1%, p=0.031). Both predictive models had an AUC of 0.97 with a favorable ROC curve profile. After SHAP was applied, smoking status was classified as the fifth most significant predictor of wound dehiscence after sex, age at surgery, number of screws placed, and preoperative diabetes. Smoking status was also classified as the third most important predictor of wound infection after number of screws placed and age at surgery.

Conclusion : While smoking status can significantly increase the risk of long-term wound complications after robot-assisted spine surgery, other factors like age and number of screws placed must also be considered. These results emphasize the necessity for targeted preoperative counseling and risk stratification to improve surgical outcomes in this patient population.