Study design: A retrospective database study.
Objective: The aim of this study was to develop and validate a sliding scale for predicting perioperative complications associated with adult spinal deformity (ASD) surgery.
Summary of background data: ASD surgery can have high perioperative complication rates, which is associated with increased morbidity and mortality.
Methods: Data on consecutive ASD patients undergoing posterior corrective surgery over a 6-year interval were collected from a prospective database. The patients' preoperative general condition, surgical status, and perioperative complications occurring within 30 days of surgery were analyzed and independent predictors were determined using multivariable logistic regression analysis. We made the sliding scale using cut-off values from receiver operating curve analysis and validated the performance of this model.
Results: Three hundred four patients were included with a mean age of 62.9 years. One hundred eight patients (35.5%) were affected by at least one perioperative complication with a total of 195 perioperative complications, including neurological (12.8%), excessive blood loss (11.2%), delirium (11.2%), and infection (3.6%). Total independent predictors were age [odds ratio (OR): 1.042], operation time (OPT) (OR: 2.015), and estimated blood loss (EBL) (OR: 4.885) with cut-off values of approximately 70 years, 6 hours, and 2000 mL, respectively. Fusion of ≥10 segments (OR: 2.262), three-column osteotomy (OR: 1.860), current use of antihypertensive (OR: 2.595) and anticoagulant (OR: 7.013), and body mass index (OR: 1.160) were risk factors for neurological complications, infection, and deep vein thrombosis/pulmonary thrombosis. Our proposed sliding scale had a sensitivity of 91%, specificity of 58.1%, and the incidence of perioperative complication in the validation dataset was smaller than that without this scale (P ≤ 0.05).
Conclusion: Patients' age, current medication, and degenerative pathology might be independent preoperative as well as operative predictors. An age and comorbidities based sliding scale with classifications of OPT and EBL may be useful for risk prediction in ASD surgery.
Level of evidence: 3.