J Vasc Surg 2019 03 6;69(3):728-737. Epub 2018 Oct 6.
Division of Vascular and Endovascular Surgery, Beth Israel Deaconess Medical Center, Boston, Mass. Electronic address:
Background: Endovascular aneurysm repair (EVAR) has decreased the perioperative mortality for patients undergoing abdominal aortic aneurysm repair and has increased the rates of elective aneurysm repair in the elderly. However, Medicare will not cover abdominal aortic aneurysm screening for beneficiaries over 75 years of age. Consequently, abdominal aortic aneurysm treatment in this population depends on incidental detection. Targeted coverage for screening in this population, however, might be beneficial for a subgroup of patients.
Methods: To identify a subset of elderly patients who would potentially benefit from an expanded screening policy, we reviewed all patients greater than 75 years old undergoing elective EVAR in the Vascular Quality Initiative between 2003 and 2016. We used Cox regression with multivariable fractional polynomials to construct a risk model for 5-year survival in elderly patients to identify a subpopulation who might benefit the most from screening and performed internal validation using the bootstrapping technique.
Results: We identified 10,676 patients greater than 75 years old undergoing elective EVAR. Although perioperative mortality varied with age, it was only 2.1% in the oldest group of patients (>85 years). Significant predictors included in our final risk model for 5-year survival in the elderly included age, aortic diameter, hemoglobin, current smoking, white race, body mass index, renal function, congestive heart failure, statin use, chronic obstructive pulmonary disease, and ejection fraction. The risk model produced risk scores ranging from a possible -2 to 33. The mean and median risk score were 6.9 and 6.0, respectively, with a right skew. We categorized the risk scores into four groups: -2 to 4 points, 5-8 points, 9-13 points, and more than 13 points, with associated 5-year survivals of 88%, 79%, 68%, and 49%, respectively. The model showed adequate discrimination and calibration, with a C-statistic of 0.69 and a calibration score of 0.99 (predicted 5-year survival of 0.78 compared with an observed 5-year survival of 0.77) and a Brier score of 0.15. Internal validation demonstrated an optimism-corrected C-statistic of 0.69 and a calibration slope of 1.0.
Conclusions: Elective EVAR in elderly patients chosen to undergo repair is associated with acceptable perioperative mortality. Our risk score can be used to define optimal patients for expanded screening into all but the highest risk group based on expected postoperative 5-year survival to justify removing this Medicare coverage restriction.