Should all patients with socially unacceptable snoring (SUS) undergo polysomnography, or is history-taking sufficient to identify the presence of obstructive sleep apnoea syndrome (OSAS)? Three hundred and eighty consecutive patients with SUS who underwent sleep registration were evaluated retrospectively to determine the predictive value of a history of apnoea or excessive daytime somnolence (hypersomnia). Of the patients analysed, 54% had OSAS, defined as an apnoea/hypopnoea index (AHI) of >15. This is higher than previously reported in the literature (46.7%). (Incidental) apnoea was reported by 337 (89%) of the patients, with a sensitivity of 0.92 and a specificity of 0.13. The predictive value--as related to the gold standard, i.e. sleep registration, of a negative test and of a positive test for OSAS is low (0.56 and 0.59, respectively). Hypersomnia was reported by 280 (74%) of the patients, with a sensitivity of 0.29 and a specificity of 0.72. The predictive value of a negative test and a positive test for OSAS is also low, 0.45 and 0.56, respectively. These data confirm that apnoea and hypersomnia in the history do not have a reliable predictive value of an obstructive sleep apnoea syndrome. We conclude that sleep registration is indicated in all patients with SUS, to rule out or confirm the presence of OSAS.