SummaryThis study aimed to determine whether bispectral index (BIS) can be used as an indicator of sedation and recovery with intravenous midazolam. In Part A, 30 healthy patients undergoing third molar extraction under local anaesthesia were recruited. They were sedated with intravenous midazolam titrated to clinical endpoints. BIS values were recorded when adequately sedated (BIS S ) and when clinical recovery criteria were met (BIS R ). In Part B, another 30 patients were sedated to the range of BIS S obtained in Part A. Recovery was assessed postoperatively when the range of BIS R from Part A was reached. BIS titrated patients required less midazolam (p < 0.001). Seventy percent of Part B patients required increments of midazolam during surgery, compared to 16.7% in Part A (p < 0.001). Total dose of midazolam given was lower in Part B (p = 0.025). BIS is not effective as a sole indicator of endpoint in sedation with intravenous midazolam.