Introduction/Aims Prolonged distal median motor latency (DML) may occur in carpal tunnel syndrome (CTS), potentially causing an electrodiagnostic dilemma in acquired demyelinating polyneuropathies. We aimed to demonstrate that parameter values obtained from conventional median nerve conduction studies can distinguish immune-mediated demyelination from compression-induced damage. Methods We retrospectively reviewed the median nerve records of 73 control individuals, 125 patients with pure CTS, 31 with CTS and diabetic distal symmetric polyneuropathy, 36 with acute inflammatory demyelinating polyneuropathy, and 23 with chronic inflammatory demyelinating polyneuropathy. All patients exhibited DML prolonged by >= 50%. Optimal cut-off values were calculated with receiver operating characteristic curves. Results The sensitivity and specificity of the optimal cut-off values forming the modified electrodiagnostic criteria were as follows: (a) distal compound muscle action potential duration of >= 8 ms (55.4% and 99.5%, respectively), (b) reduction of forearm motor conduction velocity by >= 30% (30.4% and 99.5%, respectively), (c) prolongation of minimum F-wave latency by >= 30% (58.9% and 99%, respectively), or (d) abnormal temporal dispersion of > 30% (14.1% and 100%, respectively). Using the modified electrodiagnostic criteria and the 2021 European Academy of Neurology/Peripheral Nerve Society electrodiagnostic criteria to define immune-mediated demyelination yielded sensitivities of 79.3% and 89.1% and specificities of 98.5% and 92.3%, respectively. Increased proximal conduction time and the presence of a mixed nerve-sparing pattern in the forearm did not offer additional diagnostic accuracy. Discussion The proposed electrodiagnostic criteria for the median nerve are reliable for differentiating between immune-mediated demyelination and compression-induced damage, even in patients with extremely prolonged DML.