​Marc L. Schermerhorn¹, Philip P. Goodney², Ellen D. Dillavou³, Jeffrey Jim⁴, Christopher T. Kenwood⁵, Flora S. Siami^5
1Beth Israel Deaconess Medical Center, Boston, MA; ²Dartmouth-Hitchcock Medical Center, Lebanon, NH; ³University of Pittsburgh Medical Center, Pittsburgh, PA; ⁴Washington University, St. Louis, MO; ⁵New England Research Institutes, Inc., Watertown, MA.

OBJECTIVES: CMS requires high risk criteria for CAS reimbursement. The impact of these criteria on outcomes remains uncertain, potentially biasing comparative effectiveness analysis. We evaluate this using data from the SVS Vascular Registry™.

METHODS: We analyzed 10,107 patients undergoing CEA (6,370) and CAS (3,737). Patients were stratified by CMS high risk (HR) criteria. The primary endpoint was composite death, stroke and MI (MACE) at 30 days.

RESULTS: CAS patients were more likely to have preoperative stroke (26 vs. 21%) or TIA (23 vs. 19%) than CEA. While age ≥80years was similar, CAS had higher prevalence of all other HR criteria. For CEA, HR patients had higher MACE than normal risk: SX (7.3 v 4.6%, p<.01) and ASX (5 v 2.2%, p<.0001). For CAS there was no difference: SX (9.1 vs. 6.2%, p=.24) or ASX (5.4 vs. 4.2%, p=.61) All CAS patients had MACE rates similar to HR CEA. After risk adjustment, CAS had higher rates than CEA for MACE (OR 1.2, 95% CI 1.0-1.5), death (1.5, 1.0-2.2) and stroke (1.3, 1.0-1.7), while there was no difference in MI (OR 0.8, 0.6-1.3). Multivariate analyses for predictors of MACE are reported in the Table. Radiation and restenosis were protective for MACE from CAS.

CONCLUSIONS: CEA is safer in the majority of patients with carotid disease. Normal risk CAS may be confined to restenosis and radiation.

AUTHOR DISCLOSURES: E. D. Dillavou, Nothing to disclose; P. P. Goodney, Nothing to disclose; J. Jim, Nothing to disclose; C. T. Kenwood, Nothing to disclose; M. L. Schermerhorn, Nothing to disclose; F. S. Siami, Nothing to disclose.

Table: Multivariate logistic models for 30-day MACE

Posted April 2012