Campus & Community

Regimen enhances caffeine’s ability to target key sleep system

4 min read

Small, frequent doses best strategy

Caffeine is the world’s most widely used stimulant, yet scientists still do not know exactly how it staves off sleep. Researchers at Brigham and Women’s Hospital, Harvard Medical School, and other institutions have now discovered that caffeine works by thwarting one of two interacting physiological systems that govern the human sleep-wake cycle. The researchers, who report their findings in the May issue of the journal Sleep, propose a novel regimen, consisting of frequent low doses of caffeine, to help shift workers, medical residents, truck drivers, and others who need to stay awake get a bigger boost from their tea or coffee.

“Most people take a huge jolt of coffee in the morning to jump-start their day – they get the super grande latte from Starbucks,” said Charles Czeisler, who was recently appointed the Frank Baldino Jr. Ph.D. Professor of Sleep Medicine at Harvard Medical School (HMS). “Their caffeine levels soar only to fall as the day progresses in the face of rising sleepiness. They might be better off taking much smaller more frequent doses of caffeine, equivalent to a quarter of a cup of coffee, as the day wears on.”

Though many studies have measured caffeine’s sleep-averting effects, most do not take into account that sleep is governed by two opposing but interacting processes. The circadian system promotes sleep rhythmically – an internal clock releases melatonin and other hormones in a cyclical fashion. In contrast, the homeostatic system drives sleep appetitively – it builds the longer one is awake. If the two drives worked together, the drive for sleep would be overwhelming. As it turns out, they oppose one another.

Czeisler, who also leads the Division of Sleep Medicine at HMS, and his colleagues had reason to suspect that caffeine might be working to blunt the homeostatic system. For one thing, caffeine is thought to block the receptor for adenosine, a critical chemical messenger involved in the homeostatic drive for sleep. If that were true, then caffeine would be most effective if it were administered in parallel with growing pressure from the sleep homeostatic system, and also with accumulating adenosine.

To test their hypothesis, the scientists sequestered 16 male subjects in private suites, free of time cues, for 29 days. Instead of keeping to a 24-hour day, researchers scheduled the subjects to live on a 42.85 hour day (28.57-hour wake episodes), simulating the duration of extended wakefulness commonly encountered by doctors, and military and emergency services personnel. The extended day was also designed to disrupt the subjects’ circadian system while maximizing the effects of the homeostatic push for sleep.

Following a randomized, double-blind protocol, subjects received either one caffeine pill, containing 0.3 mg per kilogram of body weight, roughly the equivalent of two ounces of coffee, or an identical-looking placebo. They took the pills upon waking and then once every hour. The goal of the steady dosing was to progressively build up caffeine levels in a way that would coincide with – and ultimately, counteract – the progressive push of the homeostatic system, which grows stronger the longer a subject stays awake.

The strategy worked. Subjects who took the low-dose caffeine performed better on cognitive tests. They also exhibited fewer accidental sleep onsets, or microsleeps. EEG tests showed that placebo subjects were unintentionally asleep 1.57 percent of the time during the scheduled wake episodes, compared with 0.32 percent for those receiving caffeine.

Czeisler and his colleagues suggest that shift workers, medical residents, truck drivers, and others who need to stay alert consider taking frequent low doses of caffeine. “While caffeine is no substitute for sleep, those who must stay awake for extended periods would benefit from this kind of dosing regimen to help maintain their performance and reduce the risk of lapses of attention,” said Czeisler.