Résumé : While polysomnography remains the current gold standard in sleep investigation, guidelines for single night versus consecutive recordings in a sleep laboratory have been disputed mainly because of two phenomena: the first-night effect and night-to-night variability. One hundred and twenty nine subjects, that underwent two consecutive nights of polysomnographic recording in a general University Hospital's sleep lab, were divided into four groups: sleep-related breathing disorders (SRBD), insomnia, movement and behavioral disorders and a healthy control (HC) group based on their complaints at admission and sleep study results. Sleep parameters of both consecutive two nights were compared and analyzed. All groups showed a significant first-night effect. However the latter seemed more pronounced in the insomnia group. Furthermore, a clinically significant intra-patient night-to-night variability was found for the apnea-hypopnea index in the SRBD-group. Due to the observed first-night effect among any subject group and the potential impact of night-to-night variability of the apnea-hypopnea index, we conclude that the clinical assessment of sleep disorders should be similar in every patient. Hence, the present study underlines the importance of two consecutive nights of polysomnographic recording as a potential reference standard for the execution of sleep investigations. © 2012 Elsevier Ireland Ltd.