Evaluation of Physiologic Sleepiness in Anesthesia Residents

Using the MSLT, Howard and colleagues^[349] evaluated physiologic daytime sleepiness of anesthesia residents under three different conditions: (1) "baseline" (daytime shift, no on-call duty period in previous 48 hours), (2) "post-call" (immediately after a 24-hour work and in-house on-call period); and (3) "sleep extended." In the sleep-extended condition, residents were told to maximize sleep and were allowed to arrive for work at 10 AM (3 to 4 hours later than normal) for 4 consecutive days prior to testing. They were not on call during this time. The sleep-extended condition was included to provide a true control state of maximal rest and optimum alertness.

In this study, the anesthesia residents had an MSLT score of 6.7 + 5.3 minutes in the "baseline" condition and 4.9 + 4.7 minutes in the "post-call" condition, both scores revealing the nearly pathologic levels of daytime sleepiness seen in patients with narcolepsy or sleep apnea. The "baseline" group slept an average of 7.1 + 1.5 hours per night, whereas the "post-call" group reported an average of 6.3 + 1.9 hours of sleep during their night on-call. Ironically, although the on-call periods were during rotations that often have very busy call nights, only a few subjects were, in fact, awake most of the night. In the "sleep extended" condition, the subjects extended their sleep to an average of over 9 hours per night, and MSLT scores were in the normal range (12.0 + 6.4 minutes). These results clearly demonstrate that medical personnel who have not been on-call cannot be assumed to be "rested" when compared with "fatigued" post-call residents. These data also indicate that under "normal" working conditions the resident physicians studied were physiologically sleepy to near-pathologic levels. This finding documents a previously unknown level of chronic sleep deprivation in this population. Notably, these data cast substantial doubt on previous studies of the performance of medical personnel that have relied on the assumption that individuals working under "normal" conditions are truly rested.

Evaluation of Subjective Sleepiness

In the previously discussed study, Howard and colleagues also investigated the degree of discrepancy between the residents' subjective sleepiness (how sleepy they felt) and their physiologic sleepiness (how easily they fell asleep). Subjective sleepiness was measured using a validated numeric scale (Stanford Sleepiness Scale)^{[366] [367]} ; physiologic sleepiness was measured using the MSLT, as described earlier. Subjects' self-reported sleepiness immediately prior to each sleep opportunity during the MSLT did not, in general, correlate with their MSLT score. As in previous studies, subjective sleepiness correlated better with physiologic sleepiness when subjects were extremely alert or extremely sleepy.

The authors also found that subjects demonstrated little ability to determine whether they had actually fallen asleep during the MSLT sleep opportunities. For example, in 51% of trials in which the electroencephalographic/electrooculographic (EEG/EOG) measurements showed that the subject had fallen asleep, the subjects thought they had remained awake throughout the test. These results support the contention that medical personnel are physiologically vulnerable to degraded alertness, yet are unable to perceive this decrement. Thus, an anesthetist could in fact fall asleep during a case, awaken, and be totally unaware of the lapse in vigilance.