Vascular Surgical Emergencies during the COVID Era: Endovascular Aortic Repair for Ruptured Abdominal Aortic Aneurysm as a Model
Emily A Grimsley, Haroon M Janjua, Mark Asirwatham, Meagan Read, Paul C Kuo, Dean J Arnaoutakis, Christopher A Latz
University of South Florida Morsani College of Medicine, Tampa, FL
INTRODUCTION: The COVID-19 pandemic caused significant stress on healthcare systems resulting in changes in processes of care. The effect of COVID and altered care processes on surgical outcomes has not been characterized. In particular, vascular surgical emergency outcomes have not been reviewed. Using COVID prevalence as a surrogate for processes of care, our aim is to determine outcomes of endovascular aortic repair (EVAR) in ruptured abdominal aortic aneurysm (rAAA).
METHODS: Using the Center for Disease Control-COVID19 data, COVID mortality per three-month quarter was calculated in Florida. The quarters with the three highest mortality rates and three lowest mortality rates were used to establish the COVID-heavy (CH) and COVID-light (CL) timeframes, respectively. Three quarters of 2019 were used for the pre-COVID (PC) timeframe. The Florida AHCA database was queried using ICD-10 codes to identify patients diagnosed with rAAA who underwent EVAR during the three timeframes. COVID+ patients were excluded. Primary outcomes were mortality, morbidity, and length of stay (LOS). Morbidity was a composite score variable of all complications per patient. Stepwise linear and logistic regression and gradient boosting machine (GBM) models with 10-fold cross-validation determined what factors most impacted primary outcomes. Timeframe was forced into the final regression models. Secondary outcomes included univariate analysis of individual complication rates and discharge to hospice.
RESULTS: There were 316 patients included. There was no difference in surgical volume during CH (n=106) or CL (n=94). There were no significant differences in patient sex, age groups, ethnicity, race, Charlson Comorbidity Index category (CCI), or comorbidities except for significantly more patients with peripheral vascular disease in EVAR CH and CL, compared to PC (p=.03), though this did not significantly influence primary outcomes in regression. Univariate analysis of mortality, LOS, and complications were not significant per timeframe. Zero of the 23 variables included in the regression model were significant predictors of mortality. Stepwise regression showed that timeframe was not a significant influencer for any primary outcome. Linear regression demonstrated the top three variables that increased LOS were metastatic cancer (B 13.48, p=.004), rheumatic disease (B 12.14, p=.002), and CCI >2 (B 3.8, p=.013) with a model R2 =.174. Significant variables that increased morbidity were CHF (OR 3.3, p=.003) and complicated diabetes (DM; OR 2.6, p=.02), while being African American decreased morbidity (OR 0.42, p=.02). Overall regression accuracy for morbidity was 0.62 (95% CI 0.57-0.68; F1 0.62). GBM showed CHF, DM, and African American race make up 60% of the relative influence on morbidity (Figure 1). Overall GBM accuracy for morbidity was 0.61 (95% CI 0.60-0.63; F1 0.62).
CONCLUSIONS: Despite increased stress on the healthcare system during the COVID-19 pandemic, outcomes following EVAR for rAAA were unchanged in Florida. These results imply that despite periods of COVID-heavy stress on the healthcare system, emergency surgery can still be accomplished with similar results. 
Back to 2023 Abstracts