Risk Factors for and Prediction of Early Thromboembolic Disease Following Adult Spinal Deformity Surgery: An Analysis of >7,400 Patients with Spinal Deformity

Abstract

Background: The aim of this study was to determine the risk factors associated with deep vein thrombosis (DVT) or pulmonary embolism (PE) within 30 days after multilevel adult spinal deformity (ASD) surgery and to develop risk prediction models.

Methods: A retrospective observational study was conducted using the American College of Surgeons National Surgical Quality Improvement Program database from 2010 to 2019. Current Procedural Terminology (CPT) codes 22843 and 22844 were used to query the database and to identify patients who underwent surgical correction of ASD with ≥7 levels of posterior instrumentation. The primary outcomes were the incidences of, and risk factors for, postoperative DVT and PE. Multiple logistic regression was utilized to identify variables associated with an elevated risk of DVT or PE within 30 days after surgery and to develop prediction models for assessing risk.

Results: A total of 7,445 patients (56% female; 73% Caucasian; mean age, 61 years) met the inclusion criteria. Postoperatively, the rate of any venous thromboembolism (VTE; i.e., DVT or PE) was 3.4% (254 patients), the rate of DVT was 2.0% (151 patients), and the rate of PE was 1.7% (127 patients). The following independent predictors of any VTE were identified: weight (odds ratio [OR], 1.054; 95% confidence interval [CI]: 1.027 to 1.081), age per decade of life (OR, 1.106; 95% CI: 1.012 to 1.209), body mass index (BMI; OR, 1.032; 95% CI: 1.015 to 1.049), medicated hypertension (OR, 1.523; 95% CI: 1.168 to 1.987), chronic corticosteroid use (OR, 2.654; 95% CI: 1.848 to 3.812), American Society of Anesthesiologists (ASA) class (OR, 1.768; 95% CI: 1.426 to 2.192), and total operative time (OR, 1.002; 95% CI: 1.002 to 1.003) (p < 0.05 for all). When incorporated into a single model, total operative time, BMI, ASA class, and chronic corticosteroid use were associated with VTE risk.

Conclusions: Four major risk factors were identified as being associated with postoperative VTE risk in patients undergoing surgery for ASD. Corticosteroid use for a chronic medical condition was the strongest predictor of VTE risk, followed by ASA class, BMI, and operative time. Knowledge of these risk factors can aid in preoperative risk assessment, informed consent, and medical decision-making, such as in determining the clinical thresholds for VTE testing and chemoprophylaxis.

Level of evidence: Prognostic Level III . See Instructions for Authors for a complete description of levels of evidence.

Fig. 1

Final model for combined DVT and PE probability. Data are presented as regression coefficients with standard error (SE) values for variables that were significantly predictive of overall VTE risk (p < 0.05). Sig. = significance, ASACLAS = ASA class, OPTIME = operative time, STEROIDCODED = chronic corticosteroid use. Fig. 1-A An example prediction calculation utilizing the regression coefficients to calculate risk for a sample patient. Fig. 1-B Individual predictor cutoff values for significant increases in VTE risk.