Background: A recent study demonstrated that almost 75% of strokes after coronary artery revascularization surgery occur in patients classified preoperatively as low to medium risk. Thus, despite the use of risk classification, most strokes can occur when not expected. We hypothesized that optimization of cerebral oxygen delivery variables by using noninvasive cerebral oximetry could reduce the incidence of stroke.
Methods: Cerebral oximetry was used by all surgeons to monitor cerebral oxygen saturation in all cardiac surgery patients from January 1, 2002, until June 30, 2003 (n = 1034; 18 months, treatment group). Cerebral oxygen delivery was optimized during surgery by modifying oxygen delivery and consumption variables to maintain oximetry values at or near the patient's preinduction baseline. Stroke was defined according to guidelines of the Society of Thoracic Surgeons. The incidence of stroke in the treatment group was compared with that for patients who underwent cardiac surgery between July 1, 2000, and December 31, 2001, (n = 1245; 18 months, control group) before cerebral oximetry was incorporated.
Results: Age and sex distribution were similar in the 2 groups. The study group had significantly more patients in New York Heart Association (NYHA) classes III and IV than the control group, and patients in the study group were sicker overall. Despite this difference, the study group overall had fewer permanent strokes (10 [0.97%] versus 25 [2.5%]; P < .044). This difference remained significant when the results were controlled for NYHA class and on-pump or off-pump surgery. When the patients were examined by NYHA class, the proportion of patients requiring prolonged ventilation was significantly smaller in the study group (6.8% versus 10.6%; P < .0014), as was the length of hospital stay (P < .046).
Conclusions: The treatment group, which underwent all cardiac surgeries with optimized cerebral oxygen delivery using cerebral oximetry monitoring, demonstrated a significantly lower incidence of permanent stroke. Because our study is retrospective, a prospective randomized trial is warranted.