We recently reported in Brain (Lugaresi et al., 1998) 20 cases of idiopathic recurrent stupor due to endozepine-4 accumulation in the blood.Samples obtained from nine of these patients were analysed by gas chromatography-mass spectrometry to rule out contaminating synthetic benzodiazepines (Rothstein et al., 1992).All these samples contained endozepine-4 concentrations high enough to account for the stuporous state of the patients.The other 11 cases of idiopathic recurring stupor were diagnosed on the basis of clinical criteria alone since they were identical to the patients with documented endozepine-4 accumulation.All presented the same clinical picture and EEG pattern (low amplitude, unreactive, background activity) during stupor and a reversal of the stuporous state after flumazenil administration (awakening and EEG normalization).Again, toxicological immunoenzymatic tests failed to detect even traces of synthetic benzodiazepines.We subsequently investigated a cluster of nine patients presenting recurrent stuporous attacks with almost simultaneous onset which had occurred in a restricted rural area near Lucca in Tuscany.Except for the extraordinary clustering in time and space, these patients were in all respects similar to the sporadic idiopathic recurring stupor patients previously encountered by us.The stuporous episodes in fact lasted 1-2 days and were followed by confusion and amnesia; ictal EEG was characterized by the typical low voltage, 13-14 Hz background activity, and flumazenil administration led to transient awakening and EEG normalization.Routine toxicological immunoenzymatic assay had ruled out the presence of benzodiazepines.However, in the meantime, a newer more specific toxicological assay, liquid chromatography-mass spectrometry, had become available to us.Because of the unusual epidemiology, we used this technique to re-analyse blood samples from the Tuscan patients.This time we detected in the blood of all of these patients the benzodiazepine lorazepam which had not