(0704) Predictive Factors of Post-Parotidectomy Complications: A Retrospective Analysis From Our Experience

Introduction: Parotid surgery involving the dissection of the facial nerve branches is technically challenging. Even when performed by experienced surgeons, parotidectomies can result in well-documented postoperative complications. This study aims to identify the predictive factors of total parotidectomy complications.

Methods: We conducted a retrospective review of 875 patients who underwent parotidectomy between 2004 and 2022. Regression analysis was employed to identify unadjusted risk factors. This was followed by forward and backward stepwise regression to ascertain predictive factors. The ROC curve was calculated to determine the correctly classified patients based on our logistic model.

Results: Out of the 875 patients who underwent parotidectomies, 360 (41.1%) experienced at least one postoperative complication, including facial nerve paralysis (139 patients, 15.8%), hematoma (14 patients, 1.6%), seroma (13 patients, 1.5%), transitory numbness (203 patients, 23.2%), Frey’s syndrome (19 patients, 2.2%), first bite syndrome (9 patients, 1.0%), surgical site infection (10 patients, 1.1%), abscess (6 patients, 0.7%), sialocele (16 patients, 1.8%), salivary fistula (2 patients, 0.2%), flap necrosis (4 patients, 0.5%), recurrence of the tumor (14 patients, 1.6%), and scar issues (15 patients, 1.7%). Regression analysis revealed that anemia and diabetes mellitus were significantly associated with increased postoperative complications with odds ratios (OR) of 2.35 (95% CI 1.30 – 4.26, p = 0.005) and 1.54 (95% CI 1.07 – 2.23, p = 0.018), respectively. Another unidentified predictive factor had an OR of 1.38 (95% CI 1.03 – 1.85, p = 0.031). The model correctly predicted post-operative complications in 60% of cases with a specificity of 89.9%.

Conclusions: Comorbidities, such as anemia and diabetes mellitus, emerged as significant predictive factors for postoperative complications following parotidectomy. There is a need to further evaluate additional factors to enhance the predictive model. This would enable clinicians to stratify and adjust treatment protocols based on these predictive factors.