Stroke in surgery of the thoracic aorta: Incidence, impact, etiology, and prevention

Objectives: To determine the incidence, impact, etiology, and methods for p revention of stroke after surgery of the thoracic aorta. Methods: A total of 317 thoracic aortic operations on 303 patients (194 mal e, 109 female) aged 13 to 87 years (mean 61 years) were reviewed. There wer e 218 procedures on the ascending aorta and arch and 99 on the descending a orta. Of the 218 procedures on the ascending aorta and arch, 86 involved ca rdiopulmonary bypass, 122 involved deep hypothermic circulatory arrest, 2 i nvolved antegrade cerebral pet-fusion, and 8 involved "clamp and sew" or le ft heart bypass. Of the 99 procedures on the descending aorta, 20 involved "clamp and sew," 69 involved left heart or full bypass, and 10 involved dee p hypothermic circulatory arrest. A total of 206 cases were elective and 97 were emergency operations. Results: Twenty-three (7.3%) of 317 patients had a stroke. Fifteen strokes occurred in operations on the ascending aorta and 8 in operations on the de scending aorta (6.9% vs 8.1%; P =.703). Stroke occurred in 16 (16.5%) of 97 emergency operations and 7 (3.4%) of 206 elective operations (P =.001). In the 300 patients surviving the operation, stroke was a significant predict or of postoperative death (9/23 [39.1%] vs 23/277 [8.3%]; P =.001). Analysi s of operative reports, brain images, and neurologic consultations revealed 15 of the 23 strokes were embolic, 3 were ischemic, 3 hemorrhagic, and 2 i ndeterminate. Patients with stroke had longer intensive care unit stays (18 .4 vs 6.8 days; P =.0001), longer times to extubation (12.7 vs 3.8 days; P < .0012), longer postoperative stays (31.4 vs 14.3 days; P =.001), and decr eased age-adjusted survival (relative risk 2.775; P =.0013). After implemen tation of a rigorous antiembolic regimen, both strokes and mortality trende d downward. Conclusions: (1) Stroke complicates surgery of both the ascending and desce nding thoracic aorta and warrants consideration in decision making. (2) Str okes are largely embolic. (3) Antiembolic measures for particles and air ar e essential, including gentle aortic manipulation, thorough debridement, tr ansesophageal echocardiography to identify aortic atheromas, carbon dioxide flooding of the field, and (in descending cases) proximal clamp applicatio n before initiating femoral perfusion.