We undertook this study to evaluate the dependence of Doppler microembolic signal (MES) counts, detected during cardiopulmonary bypass, on the type of oxygenator used. A total of 90 patients, 71 men and 19 women, aged 60 +/- 10 years (mean +/- SD), undergoing elective cardiac surgery for coronary artery bypass grafting (one vessel, n = 6; two vessels, n = 24; three vessels, n = 33; four vessels, n = 9) or valve replacement (mitral valve, n = 2, aortic valve, n = 15, both valves, n = 1) were monitored with transcranial Doppler sonography during the complete surgical procedure. The surgical and anesthetic techniques were standardized in all patients, except for the type of membrane oxygenator used (COBE CML Duo, n = 55 or DIDECO D 703, n = 35). MES count was expressed as total number of MES detected in both middle cerebral arteries during cardiopulmonary bypass (CPB) and also as total MES number divided by the CPB duration in minutes (MES min(-1)). No significant differences in patients' age or sex and type and duration of operation were noted between the two groups. MES incidence while the patients were on cardiopulmonary bypass was 100%. MES counts during CPB were 309 (236-502) and 143 (86-233) for DIDECO and COBE oxygenators, respectively (p < 0.00001). MES min(-1) were also significantly higher in patients operated with DIDECO, as compared to COBE oxygenators (3.7 (2.4-5.6) versus 1.5 (1-2.4), respectively, p < 0.0001). Inter-observer variability was satisfactory (k = 0.72). Use of a DIDECO D 703 oxygenator resulted in significantly higher MES counts as compared to the COBE CML Duo oxygenator. The clinical relevance of this finding remains to be determined.