Preoperative Risk Factors for 1 year Mortality in Patients Undergoing Fenestrated Endovascular Aortic Aneurysm Repair in the US Aortic Research Consortium
Charles Adam Banks1, Emily L Spangler1, Zdenek Novak1, Andres Schanzer2, Mark Farber3, Matthew Sweet4, Anthony Lee5, Matthew Eagleton6, Gustavo Oderich7, Adam W Beck1
1University of Alabama at Birmingham, Birmingham, AL;2University of Massachusetts Memorial Hospital Division of Vascular and Endovascular Surgery, Worcester, MA;3University of North Carolina Division of Vascular Surgery, Chapel Hill, NC;4University of Washington Division of Vascular and Endovascular Surgery, Seattle, WA;5Boca Raton Regional Hospital Division of Vascular Surgery, Boca Raton, FL;6Massachussetts General Hospital Division of Vascular and Endovascular Surgery, Boston, MA;7University of Texas Health Science Center at Houston Division of Cardiothoracic and Vascular Surgery, Houston, TX
BACKGROUND:Over the last three decades, the technological revolution in endovascular aortic aneurysm repair (EVAR) has allowed for expansion of endovascular techniques to a wide variety of aortic pathologies. Fenestrated and branched devices (F/BEVAR) present an endovascular alternative to repair more extensive aortic aneurysms, including complex abdominal aortic aneurysms (AAA) and thoracoabdominal aortic aneurysms (TAAA). There are few fenestrated devices approved for off-the-shelf use by the US Food and Drug Administration (FDA) while most remain in clinical trials. Custom-made F/BEVAR devices within Investigative Device Exemption (IDE) studies are an important alternative for cases beyond the instructions for use (IFU) for currently available devices. The US-Aortic Research Consortium (US-ARC) was developed in 2018 as a collaboration between 10 academic centers utilizing IDEs for the study of custom-made F/BEVAR devices. This collaborative, prospectively maintained database allows for the ongoing clinical appraisal of these endografts. The decision to repair AAA or TAAA is primarily based upon previously established size criteria and the associated rupture risk. This procedure is performed prophylactically as a preventative measure against aneurysm rupture, which has a high associated mortality. Early survival (1-year) following elective aneurysm repair can be utilized as an indicator of successful repair and provides a reasonable and easy to interpret countermeasure to the annual rupture risk based on diameter that is often provided as an indication for repair. One could argue that mortality within 1-year after elective repair represents a failure of therapy and ultimately unnecessary intervention. Patients undergoing elective complex AAA or TAAA have a variety of operative risk profiles secondary to high comorbid status. Mortality at 1-year following endovascular repair may differ between individual patients due to the presence of specific risk factors. The purpose of this study was to identify preoperative factors associated with 1-year mortality following F/BEVAR and develop a descriptive model for 1-year mortality based on the presence of specific risk factors. This model may allow for pre-operative risk stratification of individual patients and assist in the objective selection of patients to undergo elective complex AAA or TAAA repair.
METHODS:The US-ARC database was queried for all patients undergoing F/BEVAR for complex AAA or TAAA from 2005-2022. Utilizing this dataset, a multi-institutional retrospective cohort study was performed. Only elective repairs were included in the analysis. Patients presenting with aneurysm-related symptoms or rupture were excluded. Patients with unknown mortality status or under 1-year of postoperative follow-up were also excluded. Demographics were analyzed overall and based on extent of aortic aneurysmal pathology. The cohort was divided into three groups based on aneurysmal location or Crawford Extent (Group 1: complex AAA, Group 2: Type 1-3 TAAA, and Group 3: Type 4-5 TAAA). Patients were categorized into these three groups to preserve statistical power.The primary outcome in this study was 1-year overall survival. This outcome was modeled and stratified by aneurysm extent, to identify variables associated with 1-year mortality. Univariable regression analysis was utilized to identify unadjusted preoperative factors associated with 1-year mortality. Variables that met a univariable screening criteria of p<.20 as well as variables of clinical interest were introduced into a multivariable Cox regression model to determine risk-adjusted preoperative variables associated with 1-year mortality. Variables with a high degree of missingness defined as >40% could not be evaluated in the multivariable model (e.g. functional status and ambulatory status). Composite variables consisting of varying combinations of the variables associated with 1-year mortality identified in multivariable Cox regression analysis were created. Kaplan-Meier analysis was utilized for each composite variable to determine 1-year survival based on number and type of risk factors present. Lastly, logistic regression was performed to build a predictive model incorporating the variables utilized in multivariable Cox regression.
RESULTS:DemographicsA total of 1,927 patients met the inclusion criteria for this study (complex AAA: N=655 (34%); Type 1-3 TAAA: N=716 (37.2%); Type 4-5 TAAA: N=541 (28.1%)). Overall, the cohort was predominantly Caucasian (87.5%) and Male (72.1%) with an average age of 73.3±8.0 years. The mean preoperative aneurysm size was 64.4±11.2 mm with 9.9% of patients with prior aortic dissection. When stratified by extent of aortic pathology, there were fewer male patients within the Type 1-3 TAAA group (57.4% vs 80.2% Type 4-5 TAAA and 82.4% complex AAA; p<0.001), and Type 1-3 TAAA patients were considerably younger (71.6±8.7 Type 1-3 TAAA; 74.5±7.4 complex AAA; 74.0±7.42 Type 4-5 TAAA; p<0.001). A lower percentage of Type 1-3 TAAA patients were diagnosed with CAD (39.0% vs 52.5% complex AAA and 52.2% Type 4-5 TAAA; p<0.001) or had a prior MI (15.9% vs 24.3% complex AAA and 25.8% Type 4-5 TAAA; p<0.001). Lastly patients with Type 1-3 TAAA had a larger mean aortic diameter at time of repair (66.1±10.5 mm vs 62.7±12.5 mm complex AAA and 64.2±10.2 Type 4-5 TAAA; p<0.001) and more Type 1-3 TAAA patients had undergone prior TEVAR (32.0% vs 1.50% complex AAA and 5.40% Type 4-5 TAAA; p<0.001). Overall Survival and Risk Factors Associated with MortalityUnivariable Cox regression analysis of preoperative variables revealed multiple factors associated with 1-year mortality including chronic obstructive pulmonary disease (COPD), COPD requiring home oxygen, cardiac arrythmia, congestive heart failure (CHF), preoperative creatinine > 1.7 mg/dL, end stage renal disease (ESRD), Female sex, Age > 75, Type 1-3 TAAA, and maximum aortic diameter. Multivariable Cox regression identified the following significant risk factors associated with 1-year mortality: current smoking (HR 1.56: CI 1.16, 2.08), COPD (HR 1.34: CI 1.01, 1.72), CHF (HR 2.23: CI 1.60, 3.11), Aortic diameter>7cm (HR 1.99: CI 1.35, 2.93), Aortic diameter 6-7cm (HR 1.46: CI 1.05, 2.04), Age>75 (HR 1.96: CI 1.45, 2.58), and Extent 1-3 (HR 1.70: CI 1.20, 2.42).Kaplan-Meier analysis revealed overall 1-year survival was 85.3%. Separate Kaplan-Meier analyses were performed using different combinations of the significant risk factors identified in the multivariable Cox regression model (Figure 1A). An increase in the number of risk factors was associated with significantly lower 1-year survival with CHF demonstrating the greatest effect (Figure 1A). Patients lacking any of the risk factors experienced optimal observed 1-year survival at 93% compared to 65% in patients with all three risk factors (Log-rank<0.001) (Figure 1A). To evaluate the effect of significant anatomical factors identified in the multivariable Cox regression model, this process was repeated incorporating Age>75, Max aortic diameter>7cm, and Type 1-3 TAAA (Figure 1B). Patients with all three risk factors have a 67% 1-year survival compared to 92% in patients with none of the risk factors (Log-rank<0.001) (Figure 1B). Stratification by Extent of Aortic PathologyWhen examined by extent of aortic pathology, observed 1-year survival was highest among complex AAA repairs at 89.2%, 86.1% in Type 4-5 TAAA, and 80.9% in Type 1-3 TAAA (p<0.001). Multivariable cox regression revealed factors associated with 1-year mortality. In complex AAA these included: CHF (HR 3.04 CI: 1.58, 5.85) and max aortic diameter>7cm (HR 3.35: CI 1.70, 6.61). In Type 1-3 TAAA factors associated with 1-year mortality included: COPD (HR 1.59: CI 1.10, 2.30), CHF (HR 2.65: CI 1.63, 4.28), and Age>75 (HR 2.29: CI 1.57, 3.34). Finally, in Type 4-5 TAAA only age>75 (HR 1.86: CI 1.06, 3.28) and creatinine>1.7 (HR 2.10: CI 1.08, 4.06) were associated with 1-year mortality. Multiple combinations of these risk factors were incorporated into composite variables to determine the effects on 1-year mortality. Overall, the increasing number of risk factors was proportional to an increase in 1-year mortality (Figure 2). The combination of all significant risk factors in each group demonstrated an observed 1-year mortality of 38% in complex AAA, 56% in Type 1-3 TAAA, and 45% in Type 4-5 TAAA (Figure 2). In Type 1-3 TAAA, CHF alone carried a significant mortality burden of 50% at 1-year. Predictive Modeling for 1-year Mortality Logistic regression incorporating the variables from the Cox regression model yielded similar findings identifying current smoking, COPD, CHF, Max aortic diameter>7cm, Max aortic diameter 6-7cm, Age>75, Type 1-3 TAAA, and creatinine>1.7, as variables associated with 1-year mortality. A predictive model incorporating these variables was generated. The model included the entire cohort without stratification to preserve sample size and extent of aortic pathology as a risk factor. Appraisal of the model revealed an area under the curve of 0.70 with p-value<0.001. The model requires further validation from independent cohort. CONCLUSIONS:Conclusions: This study describes multiple risk factors that are associated with an increase in 1-year mortality following F/BEVAR. Despite the relatively low perioperative risk imposed by F/BEVAR, our study indicates that preoperative cardiopulmonary risk stratification and optimization may play a larger role especially given the impact of CHF on 1-year mortality. Additionally, this study indicates that different risk factors are associated with 1-year mortality depending on extent of aortic pathology.Elective repair of complex AAA or TAAA is typically offered to patients if future rupture risk outweighs operative risk. These findings demonstrate that highly comorbid patients with smaller aneurysms (i.e. lower rupture risk) may not benefit from repair. Descriptive and predictive models for 1-year mortality based on individual patient risk factors can serve as an adjunct in clinical decision-making when offering patients F/BEVAR.
Back to 2024 Abstracts