Pupillometry in clinically sleepy patients.

OBJECTIVES Investigators have suggested using pupillometry to assess alertness in hypersomnolent patients. In this study we assessed hypersomnolent patients and normal volunteers by using pupillometry and examined the usefulness of this technique for the diagnosis of pathologic sleepiness in individual patients. METHODS Forty-nine patients were examined by pupillometry and their sleepiness was assessed by using the multiple sleep latency test (MSLT). Thirty-three normal well-rested volunteers were also examined by pupillometry. The patients were classified as having 'mild', 'moderate', or 'severe' sleepiness, based on their mean MSLT sleep latency. Several dynamic variables of pupil diameter were calculated from the pupillograms and correlated with the mean MSLT sleep latency, and were compared between severity groups of patients and the well-rested normal subjects. RESULTS All but two pupillometric variables were significantly correlated with sleep latency. All except the same two pupillometric variables of the sleepiest group were significantly different from those of normal subjects. However, only 51% of patients with mean sleep latencies less than 10 min and 35% of patients with mean sleep latencies of less than 5 min could be correctly identified by pupillometry. CONCLUSIONS Pupillometry is clearly associated with differences in alertness between groups of patients. However, pupillometric assessment cannot substitute for the MSLT in most cases.