BACKGROUND: Experimentally, isoflurane, a commonly used volatile anesthetic agent, mimics the cardioprotective effects of ischemic preconditioning via a mechanism that could involve the activation of protein kinase C. The present study was designed to assess the clinical relevance of this observation in patients undergoing elective CABG. METHODS AND RESULTS: Twenty patients were included in the study. In 10 of them, preconditioning was elicited after the onset of cardiopulmonary bypass via a 5-minute exposure to isoflurane (2.5 minimum alveolar concentration), followed by a 10-minute washout before aortic cross-clamping and cardioplegic arrest. Ten case-matched control patients underwent an equivalent period (15 minutes) of prearrest isoflurane-free bypass. Outcome measurements included troponin I and creatine kinase-MB isoenzyme (until the third postoperative day) levels and the activity of ecto-5'-nucleotidase, which contributes to adenosine production and is considered to be a reporter of protein kinase C activation, as assessed in right atrial biopsy samples taken before bypass and at the end of the preconditioning protocol (or after 15 minutes of bypass in control patients). Aortic cross-clamping times did not differ between the 2 groups: 52+/-14 and 48+/-14 minutes (mean+/-SD) in control and isoflurane-preconditioned patients, respectively. Likewise, prebypass values of ecto-5'-nucleotidase were similar in control (3.54+/-0.86 nmol x mg protein(-1) x min(-1)) and isoflurane-treated (2.98+/-1.08 nmol x mg protein(-1) x min(-1)) patients. The values subsequently remained unchanged in control patients (3.62+/-0.94 nmol x mg protein(-1) x min(-1)), whereas they significantly increased after isoflurane preconditioning (4.74+/-0. 50 nmol x mg protein(-1) x min(-1); P<0.002 versus baseline values, P<0.004 versus time-matched values in control patients). This was paralleled by a consistently smaller release of troponin I, which yielded an area under the curve and a peak value of 204+/-147 ng x mL(-1) x min(-1) and 3.98+/-2.83 ng/mL, respectively, versus 284+/-136 ng x mL(-1) x min(-1) and 5.88+/-3.64 ng/mL, respectively, in control patients. The release of creatine kinase-MB featured a similar pattern. There were no adverse effects related to isoflurane. CONCLUSIONS: These data support a cardioprotective effect of isoflurane and, more generally, demonstrate the feasibility of pharmacologically preconditioning the human heart during cardiac surgery.