Ding S, Ma D, Guo C. Division of Cardiovascular Diseases, Wuhan General Hospital, Wuhan, Hubei, P.R. China. Dinfsfmd@Yahoo.com. J Interv Card Electrophysiol 4(1):225-30, 2000. We have developed a noninvasive transesophageal signal averaging technique for direct recording of sinus node electrogram. In this study, sinus node electrograms were recorded from 106 of 138 patients (77%), comparable to that (46%) recorded by conventional transesophageal technique, 59 were male and 47 were female ranging in age from 10-74 years (mean 44.2+/-12.4 years). The signals from lead I, surface averaged lead and esophagus averaged lead were amplified (up to 100 microV/cm), filtered (0.1-50 Hz), AD converted to 16-bit accuracy at a sampling rate of 2 KHz and averaged by using the three channel low-noise amplifier. The signal averaged esophageal sinus node potentials are deflections of low-amplitude and low-frequency preceding the P wave. Two morphologies, the domed wave (64 of 106 patients, 60%) and the smooth upstroke slope (42 of 106 patients, 40%), can be seen. The directly recorded sinoatrial conduction time was 82.3 +/-18.6 msec (mean+/-2 SD), ranged from 23-112 msec, amplitude was 3.8-27.7 microV and dv/dt was 0.42-1.92 mV/sec. The sinoatrial conduction time recorded by the transesophageal catheter technique was comparable to that (80.4+/-18. 1 msec) recorded by the transvenous catheter method perfectly. We think that signal averaged sinus node electrogram could be recorded in sinus rhythm in most patients with normal sinus node function and proper filter settings, high amplification and anti-drift technique are important in recording signal averaged esophageal sinus node electrogram.