Upper airway obstructions during sleep cause recurrent brief awakenings or microarousals. Standard criteria exist for sleep and respiratory event scoring, however, there are different definitions currently used to score microarousals. We therefore compared three definitions of microarousal (ranging from 1.5-3 s in duration) and one of awakening (>15 s). We examined their occurrence at the termination of apnoeas and hypopnoeas and their correlation with daytime sleepiness in patients with sleep apnoea/hypopnoea syndrome (SAHS). Sixty-three patients (aged 49, SD 10) had overnight polysomnography, multiple sleep latency tests (MSLT) and Epworth Sleepiness Scales (ESS). There were significantly more microarousals by any definition than there were awakenings (P<0.001) and there were more 1.5 s than 3 s microarousals (P<0.001). Significantly more apnoeas and hypopnoeas were terminated by 1.5 s microarousals (83% and 81%) than by 3 s microarousals (75%) (all P<0.001). Apnoea/ hypopnoea index (AHI) correlated significantly with objective daytime sleepiness (rho=-0.30, P<0.01). There were weakly significant relationships between all three microarousal definitions (-0.24<rho<-0.22, 0.03<P<0.04) and objective daytime sleepiness. None of the arousal definitions correlated with Epworth Sleepiness Scales scores. These results suggest that although 1.5 s microarousals are found at the end of more respiratory events, relationships between 3 and 1.5 s definitions, and daytime sleepiness are similar. This indicates that any of the three microarousal definitions can be used for visual assessment of sleep fragmentation.