Assessing the adequacy of anesthesia in the paralyzed patient is usually based on sympathetic and hemodynamic responses to noxious stimulation. Absence of such responses does not guarantee adequate anesthesia. A device monitoring the amplitude of provoked lower esophageal, contractility (PLEC) and the rate of spontaneous lower esophageal contractility (SLEC) has been developed as a potential monitor of the adequacy of anesthesia. This study determined the reliability of this device for monitoring anesthetic depth in 20 patients receiving fentanyl infusions who were undergoing coronary artery surgery and who were hemodynamically stable in the preoperative period. Premedication included midazolam 0.05 mg/kg i.m. and ranitidine 2 mg/kg p.o. Anesthesia was induced with fentanyl 50 μg/kg administered over 10 min and maintained by a fentanyl infusion 0.2 μgkg-1min-1. Following endotracheal intubation, a disposable esophageal monitoring probe, equipped with provoking and measuring balloons, was inserted and both the amplitude of provoked (PLEC) and the rate of spontaneous lower esophageal contractions (SLEC) were displayed and recorded. Precisely defined clinical signs of inadequate anesthesia included both somatic and hemodynamic responses to noxious stimulations. The presence of these responses was correlated with PLEC and SLEC and with fentanyl concentrations in plasma at specific times of noxious stimulation during the period preceding initiation of cardiopulmonary bypass. A total of 208 episodes of noxious stimulation were recorded at insertion of the nasal temperature probe (n = 8), at penetration of the skin by towel clips (n = 25), at skin incision (n = 20), at sternotomy (n = 20) and during multiple episodes of electrocauterization (n = 135). These provoked 52 clinical responses. The rate of the spontaneous contractions at times of response (4.5 4.6 contractions per 3 min, mean s.d.) was significantly greater than that at times of no response (2.5 2.7 contractions per 3 min; P<0.01), whereas there was no significant difference in the amplitude of provoked contractions (17 7 mmHg vs 15 4 mmHg (2.3 0.9 vs 2.0 0.5 kPa); P>0.05). The most favorable cut-off point was determined to be a rate of 3 contractions per 3-min period. This produced a false positive rate of 41.7% and a sensitivity (true positive rate) of 59.6%. There was no clear relationship between plasma concentrations of fentanyl and the rate of spontaneous contractions, or hemodynamic or somatic response to stimulation. We conclude that monitoring lower esophageal contractility is not reliable for detection of inadequate opioid anesthesia.