U.S. flag

An official website of the United States government

NCBI Bookshelf. A service of the National Library of Medicine, National Institutes of Health.

Institute of Medicine (US) Committee on Military Nutrition Research. Caffeine for the Sustainment of Mental Task Performance: Formulations for Military Operations. Washington (DC): National Academies Press (US); 2001.

Cover of Caffeine for the Sustainment of Mental Task Performance

Caffeine for the Sustainment of Mental Task Performance: Formulations for Military Operations.

Show details

3Efficacy of Caffeine

Caffeine has been shown clinically to induce a variety of positive effects that have contributed to its extensive use worldwide. Caffeine use has been associated with increased alertness and enhanced physical performance, and as a countermeasure to the effects of sleep deprivation. Extensive research has been done on each of these effects of caffeine. A brief summary of research findings on the efficacy of caffeine is presented here.


Caffeine has been proposed as an ergogenic aid in physical performance. Its use is associated with a reproducible increase in endurance time in activities of moderate intensity and long duration. Caffeine consumed both at rest and during exercise increases a variety of physiological processes (heart rate, respiratory rate, blood pressure), probably through the secretion of epinephrine. Recent research indicates that caffeine may also act by altering pain perception since it has been reported to increase plasma β-endorphins during endurance exercise (Laurent et al., 2000). Typically, the magnitude of the exercise response far exceeds and masks the resting effects of caffeine intake. However, if the intensity of the exercise is low and the caffeine dose is high, the effect of the caffeine may be obvious even during exercise. Caffeine also shifts cellular metabolism, possibly through antagonism of adenosine receptors (Graham et al., 1994). Specifically, caffeine increases lipolysis via activation of hormone-sensitive lipase, decreases glycogenolysis via direct inhibition of glycogen phosphorylase, and increases blood glucose and oxygen consumption (Spriet, 1999). Earlier work indicated this increase in lipolysis may actually be stimulated by the caffeine metabolite, paraxanthine, rather than by caffeine itself (Hetzler et al., 1990). Energy derived from fat during exercise is increased with caffeine ingestion, while the energy derived from carbohydrate is somewhat reduced at the same intensity of exercise (Sasaki et al., 1987). Glycogen utilization is, at least initially, depressed (Erickson et al., 1987; Essig et al., 1980; Spriet et al., 1992). Blood lactate, which usually increases in exercise above 70–75 percent of VO2max, is not affected by caffeine at rest, and may (Flinn et al., 1990; McNaughton, 1986) or may not (Dodd et al., 1991; Gastin et al., 1990) be affected by caffeine during exercise, depending on the intensity of the exercise and the level of caffeine ingested.

Although in today's military there is an increasing reliance on sophisticated computer-controlled systems, special operations and infantry missions will always rely on the physical fitness of the soldier. These operations consist of either prolonged endurance or brief, high-intensity activity. The efficacy of caffeine in promoting physical performance is different for these two kinds of activity.

Four separate reviews (Dodd et al., 1993; Graham et al., 1994; Spriet, 1995; Tarnopolsky, 1994) have concluded consistently that caffeine enhances endurance performance in a variety of activities (i.e., running, cross-country skiing, cycling), with doses from 2 to 9 mg/kg, in naive and habituated, trained and untrained test subjects. The performance effects are seen at intakes that result in urinary caffeine levels below the legal limits stipulated by the International Olympic Committee and are more pronounced in well-trained athletes (Spriet, 1999).

These same reviews concluded that there was little effect of caffeine on activities requiring high power outputs over a short time, such as lifting, carrying, and sprinting. Such activities utilize primarily anaerobic generation of adenosine triphosphate, a process that is probably not affected by caffeine. In contrast, other studies have shown slightly increased power output due to caffeine intake (Anselme et al., 1992; Collomp et al., 1992; Wiles et al., 1992), or increased time to exhaustion in brief (2-minute) supramaximal exercise (Jackman et al., 1996). This suggests a possible direct effect of caffeine on muscle tissue (Green et al., 1990; Lopes et al., 1983; Tarnopolsky et al., 1992).

Response to caffeine ingestion may vary among studies as a consequence of the caffeine habits of participants. As mentioned elsewhere in this report, chronic use of caffeine results in habituation to some of its effects, possibly by up-regulation of adenosine receptors. The epinephrine response to circulating caffeine or methylxanthine by-products may be attenuated as a result (Tarnopolsky et al., 1989; van Soeren et al., 1993). If the epinephrine response is required for the performance-enhancing effects of caffeine to be realized, habitual users may require a higher dose of caffeine to garner the positive results (Spriet et al., 1992). The dose of caffeine required for significant improvements in physical performance ranges from 3 to 9 mg/kg (Graham and Spriet, 1995). It should be noted, as well, that exercise has been shown to counteract the anxiety that may accompany high doses of caffeine. Youngstedt et al. (1998) showed that after ingestion of 800 mg of caffeine, cycling for 60 minutes at 60 percent of VO2max significantly reduced anxiety compared with consumption of this amount of caffeine while at rest.

Carbohydrate-Caffeine Mixtures

The most important theoretical mechanism of action of caffeine in the context of physical performance of the whole organism is a shift in the primary fuel used for exercise. In adipocytes, caffeine promotes lipolysis by increasing cyclic adenosine monophosphate levels, which in turn increase stimulation of hormone-sensitive lipase. The resulting increase in circulating free fatty acids hypothetically spares muscle glycogen. An independent effect of caffeine on muscle glycogenolysis has also been postulated (as discussed in previous section). In addition, carbohydrate has been shown to enhance performance during continuous exercise lasting at least 50–60 minutes (Armstrong and Maresh, 1996). The hypothesis has been put forward that incorporating the lipolytic qualities of caffeine with the carbohydrate utilization-promoting qualities of carbohydrate ingestion might augment the performance effects of both, suggesting that caffeine delivered in a carbohydrate-containing medium may further enhance performance. The following three studies have tested the efficacy of such a mixture.

Wemple et al. (1997), using a carbohydrate and electrolyte drink (3 mL/kg) with and without caffeine (60 mg per dose), evaluated time to exhaustion at 85 percent VO2max after 3 hours of continuous cycling exercise in six trained subjects. Cycling performance was not affected by including caffeine in the carbohydrate-containing fluid. However, caffeine intake in this experiment was extremely low.

Kovacs et al. (1998) added different doses of caffeine (2–4.5 mg/kg) to a carbohydrate-electrolyte solution and examined the effects on substrate metabolism and endurance performance time in 15 trained subjects during a 1-hour time trial. The addition of caffeine to the carbohydrate-electrolyte drink resulted in a significant improvement in the performance times as compared to placebo or carbohydrate-electrolyte drink alone, with a maximum effect at an intake of about 3 mg of caffeine per kilogram. There was no apparent change in metabolic fuel used during the cycling exercise, thus ruling out fuel shifts as the mechanism by which caffeine augmented the carbohydrate effect. No caffeine-only treatment was included in this experiment, leaving the question open as to how much of the effect was due to caffeine alone and how much to the interaction of caffeine and carbohydrate.

Sasaki and colleagues (1987) looked at the effect of placebo, sucrose, caffeine (approximately 6 mg/kg of body weight), and a sucrose plus caffeine mixture on time to fatigue in five trained males running at 80 percent VO2max. There was no additive effect of caffeine on time to exhaustion when it was given with sucrose, although the mean distance covered was greater in the two trials where the subjects consumed sucrose compared to placebo. Caffeine alone resulted in a distance intermediate between the two sucrose trials, but it was not significantly different from either. Caffeine alone was associated with an increase in energy derived from fat, whereas sucrose alone was associated with an increased utilization of carbohydrate. Sucrose in combination with caffeine maintained the higher carbohydrate utilization equivalent to sucrose alone. The small number of subjects in this experiment makes it difficult to project these findings to all other populations, including military personnel.

Other Effects on Physical Performance

It has been postulated that caffeine might impinge on physical performance via changes in body temperature and fluid balance. Caffeine apparently has no effect on rectal temperature, plasma volume change, or sweat rate during endurance exercise in warm (25–29°C) environments (Falk et al., 1990; Gordon et al., 1982). No similar studies have been conducted in hotter conditions; however, if an effect is not seen at 25–29°C, it is unlikely that there would be a differential response due to caffeine at temperatures greater than 29°C. Further, a study by Cohen et al. (1996) on performance in a hot and humid environment showed no effect of consuming 5 or 9 mg of caffeine per kilogram on time to exhaustion, body temperature, or blood levels of glucose and lactate during multiple 21-km runs in trained men and women.

High-altitude exposure may augment the positive effects of caffeine on endurance performance. Exercise performance is dramatically reduced by altitude exposure, and maximal effort may be diminished by as much as 25 percent. Submaximal performance may be improved with acclimatization, but maximal effort does not normally recover (IOM, 1996). However, Fulco et al. (1994) showed that ingestion of caffeine (4 mg/kg) could increase the time to exhaustion in eight trained men riding a cycle ergometer at 80 percent of high-altitude VO2max (65 percent of sea-level VO2max) at 4,300 m, but not at sea level. This positive effect was present after 1 hour of altitude exposure (54 percent increase in time to exhaustion with caffeine ingestion 1 hour before exercise) and tended to remain after 2 weeks of acclimatization (24 percent increase). Because Fulco et al. did not find any differences in substrate metabolism between the two conditions, they hypothesized that the mechanism of improvement involved an increase in residual lung capacity (tidal volume) or an improvement in muscle strength. Similarly, Berglund and Hemmingsson (1982) showed that caffeine significantly decreased the race time (by 101 seconds after one lap, 152 seconds after two) of trained cross-country skiers in a 21-km race at 2,900 m. No change in race time occurred in a test at an altitude of 300 m.

A combination of caffeine and ephedrine enhances running performance (Bell and Jacobs, 1999), but also raises metabolic heat production and thus poses a theoretical risk of hyperthermia during exercise-heat stress. However, during 2 hours of brisk treadmill walking in a 40°C hot, dry environment, Bell et al. (1999) observed that this increased metabolic heat production was offset by increased heat dissipation and that the internal body temperature change was no greater than during a control trial. However, recent information on adverse cardiovascular and central nervous system events resulting from the use of ephedra-containing supplements (Haller and Benowitz, 2000) makes the use of a caffeine-ephedra combination less than desirable. Although hyperthermia is more likely when prolonged, strenuous exercise and intense environmental stress are concurrent, the effects of caffeine in this situation have not been examined.


Both common experience and the results of scientific investigations support the belief that caffeine enhances performance on a variety of cognitive tasks. However, a review of the experimental literature reveals inconsistencies in the amount of caffeine that is required to produce positive effects on cognitive behavior. These discrepant findings can be explained by differences among experiments in a number of variables including whether or not subjects were tested following a period in which they had abstained from using caffeine, the tasks used to assess cognitive behavior, the age and gender of the subjects, the subjects' history of caffeine use, and whether the subjects were rested or sleep deprived.

There has been some debate whether caffeine enhances cognitive performance or simply restores degraded performance following caffeine withdrawal in rested individuals. James (1994, 1995, 1998) argued that the majority of studies reporting the effects of caffeine in rested subjects studied moderate caffeine consumers (200–300 mg/d) who were required to abstain from caffeine for some period of time prior to cognitive testing (2–24 hr). Abstinence for regular caffeine users could have resulted in symptoms of withdrawal which include headaches, fatigue, and irritability (Griffiths and Mumford, 1995; Griffiths et al., 1990). James (1994, 1995, 1998), hypothesized that comparisons between caffeine and placebo conditions in experiments assessing the effects of caffeine on cognitive behavior could represent a reversal of deteriorated performance. This may be due to caffeine withdrawal in the placebo condition compared to baseline performance in the presence of caffeine.

A clearer picture of caffeine's effects on cognitive function and behavior has begun to emerge, however. Caffeine can enhance performance on some types of cognitive tasks, and some aspects of mood in rested individuals independent of its ability to reverse symptoms of withdrawal and regardless of the background consumption of caffeine. Warburton (1995) demonstrated that caffeine administered in doses of 0, 75, and 150 mg to adult male, nonsmoking, regular caffeine users, without abstinence from caffeine prior to treatment, improved attention, problem solving, and delayed recall and significantly improved mood ratings. Rogers et al. (1995), using caffeine doses of 0, 70, and 250 mg/day in caffeine users (>200 mg/d) and nonusers (<15 mg/d), demonstrated that although caffeine withdrawal had a negative effect on mood, it did not appear to affect psychomotor performance. Jarvis (1993) reported results of a large survey study on coffee and tea consumption showing a highly significant dose-response relationship between habitual caffeine intake and psychomotor performance (simple reaction time, choice reaction time, incidental verbal memory, and visuo-spatial reasoning). This report also clearly demonstrates that tolerance to the performance-enhancing effects of caffeine, if it occurs at all, is incomplete with the result that higher daily caffeine consumers tend to perform better than do low consumers (Jarvis, 1993).

Using objective measures of alertness (multiple sleep latency test, visual and auditory vigilance tasks), Zwyghuizen-Doorenbos et al. (1990) demonstrated in rested, moderate (<250 mg/d) caffeine users that caffeine administered in 250-mg doses twice a day compared to placebo improved daytime alertness and reaction time on auditory vigilance tasks. Kenemans and Lorist (1995), using male and female undergraduate students with an average coffee consumption of 5.9 cups/day, demonstrated that caffeine given in a single dose of 3 mg/kg body weight (≈250 mg/day) increased cortical activation, increased sensitivity (rate at which information on stimuli is accumulated), and increased both speed and accuracy of target selection.

Amendola et al. (1998) reported caffeine at doses of 0, 64, 128, and 256 mg/day enhanced accuracy and reduced reaction time on auditory and visual vigilance tasks in a dose-related manner. Moreover, caffeine significantly increased self-reports of vigor and decreased reports of fatigue, depression, and hostility on the Profile of Moods Scale (POMS). Self-assessments of energy levels were also improved by caffeine (Lieberman et al., 1987; Sicard et al., 1996). However, caffeine did not improve long-term memory (list learning), false alarms in an auditory vigilance task, commission of errors in a four-choice reaction time, or motor coordination. In a simulated military situation involving a tedious task that required sustained attention for proficient performance (i.e., sentry duty), caffeine eliminated the vigilance decrement that occurred with increasing time on duty, reduced subjective reports of tiredness, and did not impair rifle firing accuracy (Johnson, 1999). Additionally, in this situation, caffeine increased the number of correct target identifications in both males and females. However, the reason for this differed with gender. With prolonged sentry duty and no caffeine, men were more likely to fire at friendly targets and women were less likely to fire at foes. Caffeine returned both of these deficits to baseline levels (Johnson, 1999).

Thus, caffeine's effects on cognitive function and mood can be detected in rested individuals, both users and nonusers of caffeine, using a variety of standardized tests. Only certain behavioral functions appear to be susceptible to the influence of moderate doses of caffeine (32–256 mg). In particular, it appears that in well-rested individuals, low and moderate doses of caffeine preferentially affect functions related to vigilance (i.e., the ability of the individual to maintain alertness and appropriate responsiveness to the external environment for sustained periods of time), but have limited effects on memory and problem-solving abilities. At high doses caffeine can interfere with performance of tasks requiring fine motor control (Durlach, 1998; Rogers and Dernoncourt, 1998).

The effects of caffeine on cognitive behavior vary according to dose, the subject's experience with caffeine, and gender. In general, low to intermediate doses (100–600 mg) of caffeine are associated with increased alertness, energy, and concentration, while higher doses can lead to anxiety, restlessness, insomnia, and tachycardia (Heishman and Henningfield, 1992, 1994). Individuals who do not consume caffeine on a regular basis appear to be more susceptible to the negative consequences of caffeine than regular consumers. With respect to gender, because of their smaller lean body mass, women may be more affected by a given dose of caffeine than men.

A number of studies have reported on the effect of age on physiological and cognitive responses to caffeine. Arciero et al. (1995) reported that caffeine ingestion (5 mg/kg fat-free mass) increased free fatty acids and tended to increase rate of appearance of fatty acids in younger men (19–26 years old), but not in older men (65–80 years old); while norepinephrine kinetics and fat oxidation were not affected by caffeine in either age group. Arciero et al. (1998) reported on effects of caffeine ingestion (5 mg/kg fat-free mass) on blood pressure, heart rate, norepinephrine kinetics, and behavioral mood in younger and older men. Resting baseline blood pressure was significantly lower for younger men than for older men. Following caffeine ingestion, blood pressure increased significantly above baseline for older men whereas it remained statistically unchanged in younger men. Heart rates in both groups were unaffected by caffeine ingestion. Norepinephrine kinetics (appearance and clearance rates) were not affected by caffeine in either group, although older men had higher norepinephrine concentrations with caffeine. Older men reported declines in feelings of tension and anger following caffeine ingestion, while younger men reported increased feelings of anger.

Rees et al. (1999) examined the interaction of caffeine and age and found that 250 mg of caffeine significantly decreased reaction times in both 20- to 25-year-olds and 50- to 65-year-olds with no effect on word recall. In contrast, Hogervorst et al. (1998) evaluated the effects of 225 mg of caffeine on memory and memory-related processes in three age groups: young (20–34 y), middle-aged (46–54 y), and older (66–74 y). Short-term memory was negatively affected by caffeine in the young group, positively affected in the middle-aged group, and had no effect in the older group. Jarvis (1993), in a large survey study on coffee and tea consumption, found that when results for reaction time tests were categorized by age group (16–34 y, 35–54 y, 55+y), caffeine intake had a greater performance-enhancing effect for older people (35–54 y, 55+y) than younger people (16–34 y). The author hypothesized that this greater sensitivity to caffeine in older adults might be due to the fact that older people tend to operate further below their ceiling than do the young. Alternatively, since the survey only measured coffee and tea consumption, the caffeine intake in the young group was more likely to be underestimated due to much heavier cola and soft drink use in this age group (Jarvis, 1993). Amendola et al. (1998), using subjects in two age groups (18–30 y and >60 y), tested oral caffeine doses of 0, 64, 128, and 256 mg and found a dose-dependent improvement in mood and performance on the modified Wilkinson Auditory Vigilance Task that was not affected by age.

Thus, it would appear that caffeine effects on performance of vigilance types of tasks is independent of age, while caffeine effects on memory-related tasks may be age-dependent.


Effects of Sleep Deprivation on Cognitive Behavior

Military personnel face many situations in which extended wakefulness may be required, including sentry duty, deployment-related activities, air transportation during emergencies, submarine duty, and combat. As part of their duties in these situations, individuals may have to perform complex cognitive tasks. The performance of these tasks is compromised during periods of extended wakefulness. Sleep deprivation leads to a sequence of impairments in cognitive functioning. These impairments include decreases in alertness, decrements in mental performance, reductions in self-reports of vigor, increases in sleepiness and fatigue, and increases in response reaction time (Kautz, 1999; Newhouse et al., 1989; Penetar et al., 1993, 1994; Wyatt, 1999).

A variety of instruments have been used to quantify the effects of sleep deprivation on behavior in controlled-experimental as well as simulated real-world situations. Alertness has been assessed using objective measures such as ambulatory vigilance monitors, visual and auditory vigilance tasks, and subjective measures such as self-reports and questionnaires. Studies using these measures have found that sleep deprivation impairs performance on vigilance tasks and decreases self-reports of alertness (Bonnet and Arand, 1994a,b; Bonnet et al., 1995; Caldwell et al., 1995; Penetar et al., 1993). A number of mental tasks, such as a serial add-subtract test, logical reasoning, mental rotation, perceptual cueing, and memory tests have been used to assess the effects of sleep deprivation on higher cognitive processes. Using these tasks, mental performance deteriorates as a function of sleep deprivation (Bonnet, 1999; Caldwell et al., 1995; Kautz, 1999; Newhouse et al., 1989; Penetar et al., 1993; Smith, 1999; Stickgold, 1999). Of particular significance, sleep deprivation leads to impairments in performance on cognitive tasks that would be encountered in military situations, such as piloting helicopters, fixed-winged aircraft, submarines, or advance warning aircraft; monitoring sonar or radar screens; and sentry duty. Sleep deprivation also affects mood as measured by standard scales such as the POMS and visual analogue scales. More specifically, as subjects become incrasingly sleep-deprived, increases in fatigue, tension, and depression and decreases in vigor are reported (Bonnet, 1999; Caldwell et al., 1995; Kautz, 1999; Newhouse et al., 1989; Penetar et al., 1993; Smith, 1999; Stickgold, 1999). Sleepiness, as assessed by objective measures including latency to sleep, eyelid movements, electroencephalograms, and muscle tone, and subjective measures such as self-report sleepiness scales, increases directly as a function of the amount of sleep deprivation incurred.

Recent advances in the understanding of sleep mechanisms have identified adenosine as a moderator of the sleep-inducing effects of prolonged wakefulness. Studies have shown that extracellular concentrations of adenosine in the cholinergic regions of the basal forebrain increased progressively during prolonged wakefulness and declined slowly during recovery sleep (Porkka-Heiskanen, 1999; Porkka-Heiskanen et al., 1997). Caffeine, as a known antagonist of adenosine, could thus be expected to promote wakefulness by preventing neuronal uptake of the sleep-promoting adenosine.

Two recently identified neuropeptides (orexins A and B, or hypocretins) are produced exclusively by a well-defined group of neurons in the lateral hypothalamus. These unique orexin peptides act directly at axon terminals to stimulate the release of the major inhibitory neurotransmitter, gamma-amino benzoic acid, and the major excitatory neurotransmitter, glutamate. Together, these two neurotransmitters are responsible for almost all fast synaptic activity in the hypothalamus.

Chemelli and colleagues (1999) reported the development of a strain of orexin knockout mice that developed symptoms virtually identical to narcolepsy in humans. To further evaluate the role of orexin in stimulating wakefulness, the antinarcoleptic drug, modafinil (see Chapter 6) or placebo was administered to normal mice. Modafinil strongly activated the orexin neurons in the lateral hypothalamus. No research has yet been reported that examines the effect of caffeine or paraxanthine on orexin neurons.

Restoration of Sleep Deprivation-Induced Cognitive Deficits with Sleep

All of the above-listed decrements in cognitive behavior can best be reversed by reconstituting sleep. There is a dose effect for the restorative effects of sleep duration on cognitive performance (Bonnet, 1999; Bonnet and Arand, 1994b; Bonnet et al., 1995). Any amount of sleep from as little as a 15-minute nap can restore some degree of function, although the longer the sleep episode, the greater the amount of cognitive function restored (Bonnet et al., 1995). Since the drive for sleep is governed by both a homeostatic and a circadian drive, which are interactive (Wyatt, 1999), these factors must be taken into consideration in determining the timing of naps and their effectiveness in reconstituting mental functioning. Naps are effective both prior to (prophylactic naps) and during (restorative naps) a period of sleep deprivation (Bonnet, 1999; Bonnet and Arand, 1994a; Bonnet et al., 1995).

However, in an earlier, well-designed study, Dinges et al. (1987) examined the effects of temporal placement of naps for alertness during a 56-hour period of sleep deprivation. A 2-hour nap was preceded by either 6, 18, 30, 42, or 54 hours of wakefulness. Naps were placed 12 hours apart near the circadian peak or circadian trough. Performance was measured by a visual reaction time test, and mood was assessed using the Stanford Sleepiness Scale (SSS). Results indicated that a nap at any time during the period of sleep deprivation improved reaction time performance but not SSS ratings. The earlier naps (6 and 18 hours into the wakefulness period) yielded better, and longer-lasting reaction time performance improvements which could be detected more than 24 hours after the nap, despite the fact that these naps were comprised of lighter sleep than later naps. Bonnet (1999) also found that quality of sleep differs between prophylactic naps and naps taken during sleep deprivation. Prophylactic naps are associated with longer sleep latencies and less deep sleep than post-deprivation recovery sleep. Dinges et al. (1987) also found circadian placement of naps had no effect on any parameter measured, and concluded that napping prior to a night of sleep loss is more important in meeting subsequent performance demands than is circadian placement of the nap. Napping appears to prevent sleepiness more readily than it permits recovery from sleepiness. In addition, a negative side effect of sleep during a period of sleep deprivation (restorative sleep) is sleep inertia, a short period of mental confusion upon awakening from such naps that can last as long as 30 minutes (Dinges, 1989; Stamph, 1989).

Restoration of Sleep Deprivation-Induced Cognitive Deficits with Caffeine

When sleep is not an option, caffeine can help to alleviate decrements in cognitive functioning resulting from shift work (Walsh et al., 1990, 1995), performance during circadian troughs (Gander et al., 1998; Reyner and Horne, 2000), restricted or disrupted sleep (Belland and Bissell, 1994; Rosenthal et al., 1991), and complete sleep deprivation (Bonnet, 1999; Jarvis, 1993; Johnson, 1999; Kautz, 1999; Lieberman, 1999; Lorist et al., 1994a,b; Smith and Rubin, 1999). The effectiveness of caffeine in reversing sleep deprivation-induced decrements in performance varies among subjects, and its ability to restore mental performance is influenced by a number of factors. These include prior caffeine exposure, dosage schedule, formulation of caffeine, metabolic factors, concurrent drug use, degree of sleep deprivation, and time of day of dose administration (Kaplan et al., 1997; Kuznicki and Turner, 1986; Linde, 1995; Lorist et al., 1994a,b). From the limited data available, gender does not appear to play a role in the effects of caffeine on mental abilities. However, this variable and other potential factors, such as P450 enzyme polymorphism, age, body weight, stress hormonal and other endocrine responses, concurrent illness, and drug interactions (Kamimori et al., 1999), which might potentially contribute to intra- or intersubject variability to the effects of caffeine, should be assessed further.

In sleep-deprived subjects, judicious use of caffeine can restore alertness, performance on mental tasks, and positive mood states. For example, Smith and Rubin (1999) found that caffeine had a similar profile to amphetamine and phentermine in that it reversed the sleep deprivation-induced increased response time and number of errors on a visual vigilance task, as well as the sleep deprivation-induced decrements in a running memory test. Similarly, Bonnet and Arand (1994b) observed that caffeine increased alertness and performance on a visual vigilance task, mental arithmetic tests, and logical reasoning in sleep-deprived subjects. A number of researchers have shown that caffeine is also effective in delaying sleep onset in sleep-deprived subjects (Bonnet, 1999; Kautz, 1999; Penetar, 1999; Smith, 1999). With respect to mood, caffeine administration in sleep-deprived subjects decreased reports of confusion and fatigue and increased reports of vigor, but had no effect on reports of tension, anger, and depression using the POMS (Kautz, 1999). Using visual analog scales, caffeine intake led to reports of decreased sleepiness and increased alertness, ability to concentrate, confidence, talkativeness, energy levels, anxiety, jitteriness, and nervousness (Kautz, 1999). One study suggested that some of the effects of caffeine were associated with increased measures of hypothalamic-pituitary-adrenal axis activity (plasma cortisol levels). However, further studies utilizing more extensive sampling are needed to confirm this effect.

Research suggests that doses of caffeine between 150 and 600 mg are effective in alleviating sleep deprivation-induced decrements in cognitive performance (Kelley et al., 1996; Penetar et al., 1993). Immediately following administration, doses in the range of 150 mg were just as effective as 300 or 600 mg in improving mental function in sleep-deprived subjects. However, the lower dose (150 mg) did not sustain performance on complex mental operations for as long as the higher doses (300 or 600 mg) (Kautz, 1999). Penetar et al. (1993) administered caffeine at levels of 0, 150, 300, and 600 mg following 49 hours of sleep deprivation and found a dose-related improvement in both subjective and objective measures of alertness and improvements in mood. Kelley et al. (1996) evaluated repeated doses of caffeine during 64 hours of sleep deprivation and measured effects on recovery sleep. Treatments were placebo, 300 mg of caffeine every 6 hours, or 400 mg of caffeine every 24 hours starting the evening of the first day of sleep deprivation. Subjects given the 300 mg every 6 hours developed a steady-state concentration of salivary caffeine by the third dose, while those receiving the 400 mg every 24 hours had salivary caffeine concentrations that peaked and then declined to near placebo level by 18 hours after administration. Caffeine had no effect on recovery sleep with respect to sleep latency, total sleep time, or rapid eye movement sleep. There was actually a nonsignificant increase in slow wave sleep with caffeine compared to placebo.

In comparison to 20 mg of amphetamine however, caffeine's effects are modest. Newhouse et al. (1989) found that 20 mg of amphetamine effectively restored alertness to almost 100 percent of rested values for 2 hours and remained significantly better than placebo for 7 hours after administration. In the Penetar et al. (1993) study caffeine restored alertness to approximately 50 percent of that seen in the rested condition with effects declining after 4.5 hours, although subjective measures of sleepiness following caffeine administration were restored to rested levels for 2 to 12 hours.

Restoration of Sleep Deprivation-Induced Cognitive Deficits with a Combination of Caffeine and Naps

Bonnet and Arand (1994a) compared the effectiveness of a 4-hour prophylactic nap alone to a 4-hour prophylactic nap followed by 200 mg of caffeine during the sleep deprivation period. Results showed that subjects given a combination of a 4-hour prophylactic nap prior to 24 hours of sleep deprivation and 200 mg of caffeine administered at 0130 and 0730 (normal circadian trough) during the sleep deprivation period maintained alertness and performance at levels equal to or better than those demonstrated prior to sleep deprivation, and was significantly better than the 4-hour prophylactic nap alone. In a subsequent study, Bonnet et al. (1995) evaluated differing lengths of prophylactic naps and differing doses of caffeine on performance during sustained operations and found that an 8-hour nap prior to the period of sleep deprivation was most effective in maintaining performance during the first 24 hours without sleep, and that repeated doses of caffeine at 150 or 300 mg every 6 hours were more effective than a single dose of 400 mg. However, neither nap nor caffeine conditions could maintain performance near rested levels beyond 24 hours.


Caffeine can significantly improve physical performance of an endurance nature. It is unclear at this time whether this is a result of increased production of free fatty acids to spare glycogen or an increase in release of endorphins that permits athletes to exercise longer by altering pain perception. Caffeine may be particularly beneficial in enhancing performance at high altitudes, with or without acclimation. The role of caffeine-carbohydrate combinations in enhancing physical performance still needs to be clarified.

Evidence is presented that caffeine can enhance certain types of cognitive performance, most notably vigilance and reaction times, in rested individuals regardless of whether or not they are regular caffeine users. The response to caffeine in caffeine users has been shown to be over and above any alleviation of withdrawal symptoms.

Sleep is the most effective means of reconstituting the decrements in cognitive functioning brought on by sleep deprivation. Thus, in situations where it is feasible, sleep should be promoted. When naps are not an option, caffeine alone could be used to partially alleviate sleep deprivation-induced impairments in cognitive behavior. Combining naps with judicious caffeine use may be the best remedy for sleep deprivation-induced decrements in cognitive function in military situations where adequate sleep cannot be obtained.

The doses of caffeine most likely to be effective without causing undesirable mood effects are within the range of 100 to 600 mg.

Copyright 2001 by the National Academy of Sciences. All rights reserved.
Bookshelf ID: NBK223791


  • PubReader
  • Print View
  • Cite this Page
  • PDF version of this title (1.1M)

Recent Activity

Your browsing activity is empty.

Activity recording is turned off.

Turn recording back on

See more...