Estimating the degree of sleep fragmentation is an important part of a respiratory sleep study and is conventionally measured using EEG micro arousals or is inferred indirectly from respiratory abnormalities such as apnoeas and desaturations. There is a need for less labour-intensive measures of sleep fragmentation, and transient rises in blood pressure and heart rate may fulfil this role. Forty unselected sleep clinic referrals undergoing investigation for possible obstructive sleep apnoea (OSA) were studied with one night of polysomnography. Three conventional indices of sleep fragmentation (EEG micro arousals, apnoea/hypopnoea index (AHI) and oxygen saturation dip rate (SaO2 dips)) and two autonomic indices (heart rate and blood pressure rises) have been compared. Correlations between these five indices ranged from r=0.38 to r=0.73. Of the two autonomic indices, the correlations for blood pressure rises with SaO2 dips and EEG micro arousals were stronger (r=0.71 and r=0.65, respectively) than those for heart rate rises (0.55 and 0.51). All indices of sleep fragmentation, apart from heart rate rises, were similar in their correlation with subjective sleepiness (r-values 0.21-0.36). Arousals implied from blood pressure rises (using pulse transit time) can be measured easily, are objective, and appear no worse at predicting subjective sleepiness than either EEG micro arousals or AHI. They may therefore provide a useful alternative to manual scoring of micro arousals from the EEG as an index of sleep fragmentation in sleep clinic patients undergoing investigation for possible OSA.