P39. Predictive factors of immediate postoperative adverse events in anterior cervical discectomy and fusion: a study of 509 patients

Abstract

<h3>BACKGROUND CONTEXT</h3> Postoperative adverse events following anterior cervical discectomy and fusion (ACDF) can have a significant impact on patient quality of life and health care resource utilization. <h3>PURPOSE</h3> To identify predictive factors associated with the incidence of adverse events within 1-2 weeks post-ACDF. <h3>STUDY DESIGN/SETTING</h3> Retrospective cohort study. <h3>PATIENT SAMPLE</h3> A total of 509 patients who (1) received ACDF between January 2013 and January 2020, (2) had a follow-up visit with their surgeon 1-2 weeks postsurgery. <h3>OUTCOME MEASURES</h3> Incidence of immediate postoperative adverse events (AE) was determined via retrospective chart review. Adverse events included the following: dysphagia, hematoma, wound dehiscence, fever, vocal cord paralysis and dysphonia. <h3>METHODS</h3> Demographic, clinical and surgical data was retrospectively collected via chart review. Recorded data variables included the following: time of follow-up, age, gender, insurance type, BMI, hospital, presence of preoperative (preop) radiculopathy, presence of preop myelopathy, presence of preop sensory deficit, presence of preop motor deficit, preop smoking status (current/former smoker or nonsmoker), diabetes diagnosis, osteoporosis diagnosis, sleep apnea diagnosis, presence of cardiac comorbidity, preop anxiety (yes/no), preop depression (yes/no), preop exercise (none or any exercise), preop prescription pain medication use (yes/no), surgical history (primary or revision), number of levels (single or multilevel), level type (C2, C3, etc.) and allograft use (yes/no). Univariate analyses were first performed to evaluate the correlation between each factor and incidence of postoperative complications. Variables that were found to be significant at the 0.1 alpha level were deemed to be candidates for multivariable analysis. A binary logistic regression model was developed to determine if any of the candidate variables were independent predictors (alpha level of 0.05) of immediate postoperative AE incidence in ACDF. <h3>RESULTS</h3> Fifty patients (9.8%) had at least one AE at 1-2 weeks post-ACDF. A total of 54 adverse events occurred and included the following: dysphagia (n=37), hematoma (n=7), wound dehiscence (n=6), vocal cord paralysis (n=2), fever (n=1), and dysphonia (n=1). Mean (± standard deviation) patient age was 54 ± 11 years. Average patient follow-up was 12 days. The following patient factors were discovered via univariate analyses to be significantly correlated with AE incidence: presence of preop motor deficit (n=296, p=0.009), BMI 35 (n=69, p=0.019), use of allograft (n=311, p < 0.001), revision surgery (n=15, p < 0.001), treatment level C4 (n=240, p=0.012) and treatment level C5 (n=406, p=0.008). These factors were included in the binary logistic regression model and controlled for as covariates. Multivariable regression analysis identified presence of preoperative motor deficit (OR 2.25, 95% CI 1.14-4.44, p=0.019), BMI 35 (OR 2.37, 95% CI 1.12-5.0, p=0.024), use of allograft (n=311, p < 0.001), revision surgery (n=15, p < 0.001), and treatment level C5 (n=406, p=0.008) as significant independent predictors of AE incidence at 1-2 weeks post-ACDF. The model was statistically significant (p < 0.001) and a good fit for the data (Hosmer-Lemeshow p-value of 0.429). <h3>CONCLUSIONS</h3> Preop motor deficit, BMI of 35, use of allograft, revision surgery and treatment level C5 were found to be significant independent predictors of immediate postoperative AE incidence in ACDF. <h3>FDA DEVICE/DRUG STATUS</h3> This abstract does not discuss or include any applicable devices or drugs.