Frequency of Reoperative Coronary Artery Bypass Grafting in Medicare Beneficiaries: 20-Year Tends and Outcomes

J. Eisenga¹, J. DiMaio², J. Banwait³, S. Hale³, J. Schaffer^4
1Baylor University Medical Center, Dallas, Texas ²Baylor Scott & White Health, Dallas, Texas ³Baylor Scott and White The Heart Hospital- Plano, Plano, Texas ⁴Baylor Scott and White The Heart Hospital, Plano, Texas

Purpose: Improvements in medical care and percutaneous interventions have reduced the incidence of redo coronary artery bypass grafting (CABG). Nevertheless, redo-CABG remains an important therapy in patients with severe, progressive coronary artery disease. We sought to analyze trends and long-term outcomes in Medicare beneficiaries undergoing redo-CABG versus primary isolated CABG.

Methods: Medicare administrative claims data were queried to identify beneficiaries undergoing isolated CABG from 1999-2019. 2.44 million isolated CABG surgeries were identified using ICD9/10 codes. 1.7 million records with useable data were identified (≥1yr of Medicare coverage, a traditional (non-advantage) Medicare plan, complete demographic, hospital, and surgeon data, location demographics to assess neighborhood disadvantage, CPT codes confirming isolated CABG). Outcomes between first time CABG (N=1,612,939) and redo-CABG (N=89,496) were compared. Cochran-Armitage test was used to assess the incidence of redo-CABG over this period. Overlap propensity score weighting was performed to account for demographic, operative, surgeon and hospital differences. The number of arterial and venous grafts was not included in propensity weighting as these differences were assumed to be inherent in redo-CABG. Kaplan-Meier survival analysis was performed to analyze the primary outcome of all-cause mortality. In the weighted sample, survival was compared using a Cox regression-based test for equality of survival curves.

Results: 1,702,435 beneficiaries with usable data underwent isolated coronary artery bypass grafting over a 20-year period from 1999-2019. A total of 89,496 patients (5.3%) underwent redo coronary bypass surgery. Median follow up was 14 years. The incidence of repeat surgery decreased dramatically over this period (Cochrane-Armitage z=-101.475, prob>z < 0.001). Patients who underwent repeat CABG were more likely to be < 65 years old (0.12,0.10; SMD=0.08), white (0.77,0.67; SMD=0.22), and had a higher incidence of cardiovascular co-morbidities (CHF, prior MI, and atrial fibrillation). However, these patients had a lower incidence of renal disease and fewer required dialysis at the time of surgery. Patients undergoing redo CABG were less likely to undergo concomitant MAZE procedure, less likely to have their operation done off pump, and more likely to have all venous grafting compared to first time operations (0.48,.01;0.932). In our weighted analysis, median survival of the first-time surgical group was 9.5 years, compared to 8.6 years in the redo CABG group; p< 0.001 by Cox regression-based test for equality of survival curves. Post operative survival in the first-time surgical group and the redo group was: 96.5% vs 93.4% at 30-days, 94.5% vs 91.2% at 90 days, and 91% vs 87.5% 1 year post-operatively.

Conclusion: The incidence of redo CABG has decreased dramatically in Medicare beneficiaries over the two decades. Nevertheless, it remains an important aspect of cardiac surgery. Although redo CABG presents a higher perioperative risk, the long-term results of this surgery are quite reasonable compared to first time revascularization.

Identify the source of the funding for this research project: Baylor Scott & White Research Institute