Does the arterial cannulation site for circulatory arrest influence stroke risk?

Background We investigated whether axillary/subclavian artery inflow with a side graft decreases the risk of stroke versus cannulation at other sites during hypothermic circulatory arrest. Methods Between January 1993 and May 2003, 1,352 operations with circulatory arrest were performed for complex adult cardiac problems. A single arterial inflow cannulation site was used in 1,336 operations, and these formed the basis for comparative analyses. Cannulation sites were axillary plus graft in 299 operations, direct cannulation of the aorta in 471, femoral in 375, innominate in 24, and axillary or subclavian without a side graft in 167. Retrograde brain perfusion was used in 933 (69%). A total of 272 (20%) were for emergencies, 432 (32%) were reoperations, and 439 (32%) were for dissections. A total of 617 (46%) had aortic valve replacement and 1,160 (87%) ascending, 415 arch (31%), and 248 descending (18%) aortic replacements. Indications also included arteriosclerosis (n = 301) and calcified aorta (n = 278). Primary comparisons were made by using propensity matching, and, secondarily, risk factors for stroke or hospital mortality were identified by multivariable logistic regression. Results Stroke occurred in 6.1% of patients (81/1,336): 4.0% (12/299) of those had axillary plus graft and 6.7% who had direct cannulation (69/1,037; p = 0.09; p = 0.05 among propensity-matched pairs). Operative variables associated with stroke included direct aortic cannulation, aortic arteriosclerosis, descending aorta repair, and mitral valve replacement. The risk of hospital mortality was higher (11%; 42/375) for patients who had femoral cannulation than axillary plus graft (7.0%; 21/299; p = 0.06; p = 0.02 among propensity-matched pairs). Conclusions Axillary inflow plus graft reduces stroke and is our method of choice for complex cardiac and cardioaortic operations that necessitate circulatory arrest. Retrograde or antegrade perfusion is used selectively.