New Preoperative Spinal Cord Ischemia Risk Stratification Model for Patients Undergoing Thoracic Endovascular Aortic Repair.

Our objective was to determine significant predictors of spinal cord ischemia (SCI) following Thoracic Endovascular Aortic Repair (TEVAR) and to further develop a simple and clinically orientated risk score model. A retrospective review of data from the Society of Vascular Surgery/Vascular Quality Initiative national data set was performed for all patients undergoing TEVAR from January, 2014 to June 2018. Preoperative demographics, procedure-related variables, and clinical details related to SCI were examined. A SCI risk score was developed utilizing a multivariable logistic regression model. For the 7889 patients in the final analysis who underwent TEVAR during the study period, the mean age was 67.6 ± 13.9, range 18 to 90 years, and the majority was male (65%). Postoperative outcomes included stroke (3.0%), myocardial infarction (2.9%), inhospital mortality (5.4%), transient SCI (1.5%), and permanent SCI (2.1%). Nearly half of the overall cases were performed in high volume centers. Predictors of increased risk for SCI included age by decade (odds ratio [OR]: 1.2), celiac coverage (OR: 1.5), current smoker (OR: 1.6), dialysis (OR: 1.9), 3 or more aortic implanted devices (OR: 1.7), emergent or urgent surgery (OR: 1.5), adjunct aorta-related procedure (OR: 2.5), adjunct not related (OR: 2.6), total estimated length of aortic device (19-31 cm, OR: 1.9 and ≥32 cm, OR: 3.0), ASA class 4 or 5 (OR: 1.6), and procedure time ≥154 minutes (OR: 1.8). Two predictors decreased the risk of SCI, cases from high-volume centers (OR: 0.6) and eGFR ≥ 60 (OR: 0.6). To evaluate the risk score model, probabilities of SCI from the original regression, raw score, and raw score categories resulted in area under the curve statistics of 0.792, 0.786, and 0.738, respectively. Spinal cord ischemia remains one of the most feared complications of TEVAR. Incidence of SCI in this large series of patients with TEVAR was 3.6% with nearly 60% being permanent. The proposed model provides an assessment tool to guide clinical decisions, patient consent process, risk-assessment, and procedural strategy.