BACKGROUND Many studies had been performed in the last years to prove the usefulness of ultrasonographic measurements of the median nerve in the diagnosis of carpal tunnel syndrome (CTS). We wanted to determine its reliability and to compare this technology with electromyography (EMG) in ordinary diagnostic conditions. METHODS The study involved 90 wrists with suspected CTS, 35 controlateral wrists and 52 control wrists. The diagnosis of CTS was confirmed in 81 cases by the hand symptom diagram and the Tinnel and Phalen sign. The EMG examination evaluated medianulnar sensory latency difference to the ring finger and wrist-to-palm sensory conduction velocity. For the ultrasound diagnosis, the cross sectional area of the median nerve at the level of the pisiform bone, was considered. The sensitivity and specificity of the two techniques was calculated. RESULTS Sensitive electroneurographic parameters showed a sensibility and specificity respectively of 79 and 80%. The cut-off point for ultrasound sensibility and specificity using ROC analysis was 11mm(2) for mean cross-sectional area. Sensitivity and specificity found in this way were 72% and 56%. Reliability was good with intra- and inter-reader intraclass correlation coefficients of 0.99, and interobserver coefficient of 0.88. Sonography found seven CTS among the 17 clinical CTS with normal electrophysiological findings. A statistically correlation was found between the cross-sectional section and the sensitive electrophysiologic parameters (r=0.43, p<0.001). CONCLUSIONS In our study, ultrasonographic diagnostic value are not as good as electrophysiological value, like found in recent literature, probably because of the composition of our group of patients which is including many causes of acroparesthesias. This can mean that in clinical practice, sonography is a complementary tool instead, for example in cases of equivocal EMG.