Since the magnitude of the first-night effect has been shown to be a function of medical conditions and of settings in which polysomnographies are performed, it is essential to evaluate the habituation phenomenon in each case in order to determine the optimal recording methodology. A first-night effect was evidenced in certain cases of chronic fatigue syndrome, but not in others. To clarify this issue, a large group of patients with chronic fatigue syndrome who had no primary sleep disorders were selected and recorded for two consecutive nights in a hospital sleep unit. Several parameters, frequently associated with the first-night effect, were found to be influenced by the recording methodology: Total Sleep Time, Sleep Efficiency, Sleep Efficiency minus Sleep Onset, Sleep Onset Latency, Wake Time, Slow Wave Sleep, Rapid Eye Movement Sleep, Rapid Eye Movement Sleep Latency and Number of Sleep Cycles. Bland and Altman plots determined that the difference scores between the nights included a systematic bias linked to the order of recordings (first-night effect). Factorial analysis grouped the difference scores into three factors. No significant difference was observed between patients with generalized anxiety comorbidity and those with no psychiatric comorbidity, or between those with and without psychiatric comorbidity. Chronic fatigue syndrome must thus be added on the list of conditions where a clinically significant habituation effect takes place.