To compare surface and intramuscular laryngeal electromyography (EMG) with adductor pollicis muscle EMG after 0.1 mgxkg(-1) cisatracurium.This prospective study included ten patients undergoing surgery with risk of damage to the recurrent laryngeal nerve (RLN). The tracheas were intubated after fentanyl/propofol without the aid of muscle relaxation. A surface laryngeal electrode was attached to the tube and placed amidst the vocal cords; two straight needles were inserted endoscopically into the left lateral cricoarytenoid muscle. Single twitch stimulation of the left RLN (0.1 Hz) was performed transcutaneously; skin EMG of the left adductor pollicis muscle was performed at 0.1 Hz. After supramaximal stimulation for 10 min, 0.1 mgxkg(-1) cisatracurium was injected. Lag, onset time and peak effect were measured and compared.Good correlation (r = 0.9, 0.8, P < 0.005) and good comparability of the two methods of laryngeal EMG (mean difference and limits of agreement: 0 +/- 28 sec for lag time, -2 +/- 84 sec for onset time) was shown. The mean surface laryngeal lag and onset times were 67 +/- 22 sec and 198 +/- 72 sec, compared with the adductor pollicis muscle (98 +/- 30 sec and 242 +/- 59 sec) at P < 0.01. Peak effects at larynx (92 +/- 9%) and adductor pollicis muscle (95 +/- 3%) were similar.Surface laryngeal EMG is comparable to intramuscular laryngeal EMG to determine degree and onset of the neuromuscular blockade. Increasing muscle relaxation does not cause the surface electrode to lose contact with the vocal cords and therefore underestimate onset time and peak effect.