(IP065) A novel preoperative risk prediction model of non-home discharge after open thoracoabdominal aortic aneurysm repair

Objectives: To build a preoperative risk prediction model of non-home discharge (NHD) after open thoracoabdominal aortic aneurysm (TAAA) repair.

Methods: Retrospective single-institutional chart review of patients who underwent open TAAA repair between 2013 and 2024. Patients who had urgent or emergent surgery, were admitted from non-home setting, or discharged to unknown location were excluded from the study. Univariate followed by multivariate logistic regression models were used to identify preoperative predictors of NHD after TAAA repair. Receiver operating curve (ROC) and concordance statistic (c-statistic) were used to evaluate the strength of the prediction model. Descriptive analysis was used to illustrate the outcomes of patients depending on their discharge status. Kaplan Meier curves were used to analyze survival time according to the discharge destination.

Results: Among 135 patients who underwent open TAAA repair during the study period, 117 patients met the inclusion criteria and were included in the study. Average age was 54 ± 13 years (range: 21-78), and 68% (n=80) were male. Seventy-six percent of the repairs (n=89) were done for extent II TAAAs. Early mortality rate was 8.5% (n=10), and NHD occurred for 28% of patients (n=33). On multivariate logistic regression analysis, older age (OR=1.08, 95%CI 1.03-1.13, p< 0.01), type 2 diabetes (OR=6.70, 95%CI 1.15-39.07, p=0.03), peripheral artery disease (OR=5.36, 95%CI 1.67-17.25), and not using aspirin preoperatively (OR=0.286, 95%CI 0.10-0.82, p=0.02) were associated with increased likelihood of NHD and were included in the final risk score model (Table1). The risk score model accurately predicted NHD with a c-statistic of 0.83 (Fig.1). On descriptive analysis, the rates of major adverse events were higher in the NHD group, including postoperative pulmonary complications (82% vs 22%, p< 0.01), stroke (24% vs 1%, p< 0.01), spinal cord ischemia (14% vs 1%, p< 0.01), myocardial infarction (7% vs 0%, p=0.01), and renal failure requiring dialysis (55% vs 5%, p< 0.01). Mean follow-up was 56 ± 40 months. On Kaplan Meier analysis, long-term survival was significantly lower in NHD patients (p=0.013). (Fig.2).

Conclusions: Older patients with type 2 diabetes and peripheral artery disease, and who do not use aspirin preoperatively are at highest risk of NHD after TAAA repair. In addition, NHD patients were more likely to have post-operative morbidity and long-term mortality. Using preoperative factors only, our risk score model can accurately predict patients who will have a NHD after open TAAA repair. This may enhance preoperative counseling and shared decision-making, as well as expedite hospital discharge.