You'd get very concerned if, while waiting to board an airplane, you saw the pilot stumble out of the airport bar. New Stanford research suggests that you should be just as worried if you see the pilot yawning and rubbing his eyes.

In a test of reaction times, people who were tired because of disrupted sleep performed about as poorly as subjects who were legally drunk, the researchers report. The study is the first to show severe impairment in people who have only mild to moderate sleep disturbances.

Nelson B. Powell, DDS, MD, leader of the research team, said he hopes that the results will stimulate a discussion about the need for safety guidelines to cover sleepiness -- rules that might resemble those already in place for blood alcohol levels.

Alcohol's slowing effect on reaction times is well documented, said Powell, who is co-director of the Stanford Sleep Disorders Clinic and Research Center. That is one reason why society demands that people responsible for the safety of others -- truck drivers, train engineers, airline pilots -- limit their alcohol consumption before working.

Yet, Powell noted, society does not seem to be as concerned about the harmful impact of sleeplessness, even though fatigue contributed to almost every high-profile accident in recent memory, from the grounding of the Exxon Valdez to the Chernobyl and Three Mile Island nuclear accidents.

Powell wants to see that change, so he and five colleagues set out to measure how the disorder known as sleep apnea affects reaction times. People with sleep apnea -- about 24 percent of middle-aged men and 9 percent of middle-aged women -- stop breathing multiple times during the night. These interruptions come during sleep, so the patient may not be aware of them, but they prevent the body from settling into a deep, relaxing sleep and result in daytime drowsiness.

Those with extreme apnea, who are often so tired that they may struggle to remain awake during a conversation or while driving to the corner store, clearly represent a danger. However, Powell's group wanted to learn if less severe forms of the disorder might also compromise safety.

So they recruited 113 patients with mild to moderate sleep apnea and compared their reaction times with those of 80 normal volunteers who had slept well the three previous nights. Members of the latter group took the reaction-time test sober to provide baseline data. Then they gradually got drunk and performed the test three additional times: once at a blood alcohol level of 0.057 percent, again at 0.08 percent, and finally at 0.083 percent.

A blood alcohol level of 0.057 percent exceeds the legal limit for driving a commercial vehicle, which is 0.04 percent, but falls short of the limit to drive a car in California and many states, which is 0.08 percent.

Comparing the two groups on seven measures of reaction time -- including average time, maximum time, and average of the ten fastest times -- showed a surprising degree of impairment in the apnea patients. On all seven measures, their results were worse than those of the drinking group at a blood alcohol level of 0.057 percent (see graph on page 7). And on three measures, the apnea patients scored as badly or worse than the drinkers who were legally drunk. "That really stunned us," Powell said.

Taking one example, the average reaction time for the drinkers with a blood alcohol level of 0.057 percent was 263 milliseconds (just over a quarter of a second), which increased to 276 milliseconds by the time their alcohol level rose to 0.08 percent. The average time for the apnea patients was 266 milliseconds.

One of Powell's co-authors, David F. Dinges, PhD, a psychiatry professor at the University of Pennsylvania, developed the computerized reaction-time test for NASA as a way to evaluate pilots and astronauts. The 10-minute test is straightforward: the subjects push a button whenever a red light flashes. The length of time between flash and push is the reaction time.

Powell cautioned that he isn't proposing specific safety guidelines. But he hopes this study will start a national discussion and debate about whether such rules are necessary. For example, he suggested that it might be prudent to require airline pilots to pass a reaction-time test before they are allowed to fly. Alternatively, we might want to consider monitoring their sleep to ensure they are well rested when they sit down in the cockpit.

"Being arrested for sleepiness -- that isn't going to happen, but maybe it should," Powell said. "It might make people in sensitive positions take responsibility."

Further research could focus on other ways in which mental or physical abilities suffer in the sleep deprived, Powell suggested. "I am showing you just one measure of abnormality in performance," he said.

The group's results appear in the October issue of the journal Laryngoscope.

Powell's Stanford colleagues on the paper are both from the Sleep Disorders Clinic and Research Center: Robert W. Riley, DDS, MD, associate clinical professor of psychiatry and behavioral sciences, and Christian Guilleminault, MD, professor of psychiatry and behavior sciences. The remaining authors are Marc B. Blumen, MD, an assistant professor of otolaryngology at Hopital Foch in Suresnes, France; and Kenneth B. Schechtman, PhD, an associate professor of biostatistics at Washington University School of Medicine in St. Louis, Mo. SR