Within a double-blind, comparative study on the effects of the long-half-life benzodiazepine (BDZ), quazepam, and the short-half-life BDZ, triazolam, on clinical symptomatology, sleep and anxiety of 45 patients with insomnia based on a mild to moderate generalized anxiety disorder (GAD) (ICD-9 code: 307.42-1, 300,0; ASDC-APSS-Code: A.2.a), we compared, in a first step at baseline, drug-free polysomnographic and psychometric data of 22 patients recorded in the laboratory (L-group) and 21 patients recorded by the Oxford Medilog 9000 system at home (H-group) with those of normal controls. Sleep efficiency, total sleep time, wake within total sleep period (middle insomnia) and wake before buzzer (late insomnia) were significantly deteriorated in both patient groups as compared with controls, while sleep induction time only differed significantly in home recordings. Regarding sleep architecture, stage (S)2 was reduced, S3 and S4 increased in the H-group only, while no intergroup differences were seen in S1, SREM and REM latency. Subjective sleep quality was reduced in both patient groups, but not awakening quality. Psychometric tests in the morning demonstrated for the noopsyche, only a significantly deteriorated psychomotor activity in both patient groups. In the thymopsyche, evening well-being and mood in the morning were reduced in both the L- and H-group, affectivity and morning well-being only in the H-group. The psychopharmacological part of the study was completed by 40 patients (there were 4 drop-outs in the triazolam, 1 in the quazepam group). They were treated after 1 week placebo with either 15-30 mg (median 15 mg) quazepam or 0.25-0.5 mg (median 0.25 mg) triazolam for 4 weeks, and thereafter for 2 weeks with placebo. Anxiety (rated by HAMA and SAS) improved significantly with both drugs and remained improved throughout 2 weeks post-drug placebo, with quazepam being slightly superior to triazolam. Polysomnography demonstrated a shortened sleep onset only after quazepam. Sleep efficiency improved after acute administration of both drugs, but the improvement was maintained by quazepam only (tolerance development with triazolam). Rebound insomnia was observed only in the 1st post-triazolam placebo night (significant intergroup difference based on confirmatory testing). S2 increased, S3 + S4 decreased under and after quazepam, which represents a normalization in home-recorded GAD patients. S1 decreased with both drugs, SREM only under quazepam. Subjective sleep quality behaved very similarly to objective sleep efficiency. Awakening quality improved after acute therapy with both drugs, somatic complaints only with quazepam.(ABSTRACT TRUNCATED AT 400 WORDS)