2.1. Zwiększenie oksydacji tłuszczów prowadzące do redukcji tkanki tłuszczowej (ID 735, 1484)

The claimed effect is “fat metabolism/energy expenditure”. The Panel assumes that the target population is the general population.
In the context of the proposed wordings, the Panel assumes that the claimed effect refers to an increased fat oxidation leading to a reduction in body fat mass.
The Panel considers that increased fat oxidation leading to a reduction in body fat mass might be a beneficial physiological effect.

2.2. Zwiększenie wydatków energetycznych prowadzące do redukcji masy ciała (ID 1487)

The claimed effect is “support of resting metabolic rate and thermogenesis”. The Panel assumes that the target population is the general population.
Enhancement of resting metabolic rate and thermogenesis is interpreted as an increase in total substrate oxidation (energy expenditure), which is a measurable outcome. An increase in energy expenditure may lead to a decrease in body weight, which is considered a beneficial physiological effect.
The Panel considers that increased energy expenditure leading to a reduction in body weight might be a beneficial physiological effect.

2.3. Zwiększenie czujności (ID 736, 1101, 1187, 1485, 1491, 2063, 2103)

The claimed effects are “cognitive and mental performance”, “mental and physical stimulant effect”, “mental state and performance”, “mental performance (where mental performance stands for those aspects of brain and nerve functions which determine aspects like concentration, learning, memory and reasoning, as well as resistance to stress)”, “mental performance and cognitive function (enhances mental alertness during intense muscular activity)”, and “mental performance”. The Panel assumes that the target population is the general population.
In the context of the proposed wordings and the clarifications provided by Member States, the Panel assumes that the claimed effects refer to alertness. Alertness may relate to either a cognitive (i.e. behavioural) or an affective (i.e. subjective self-rating) construct. Cognitive alertness refers to a state of enhanced arousal and readiness to receive and process information and respond. Alertness is a well defined construct and can be measured by validated psychometric cognitive tests.
The Panel considers that increased alertness might be a beneficial physiological effect.

2.4. Zwiększenie uwagi (ID 736, 1485, 1491, 2375)

The claimed effects are “cognitive and mental performance”, “mental performance (where mental performance stands for those aspects of brain and nerve functions which determine aspects like concentration, learning, memory and reasoning, as well as resistance to stress)”, “mental performance and cognitive function (enhances mental alertness during intense muscular activity)”, and “invigoration of the body”. The Panel assumes that the target population is the general population.
In the context of the proposed wordings, the Panel assumes that the claimed effects refer to attention, which means the ability to concentrate while processing information. There are two broad categories
of attention. Selective attention is the ability to concentrate on one task or source of information to the exclusion of others. Sustained attention (vigilance) is the ability to concentrate over a long period of time. Attention is a well defined construct which can be measured by validated psychometric tests.
The Panel considers that increased attention is a beneficial physiological effect.

3.1. Zwiększenie oksydacji tłuszczów prowadzące do redukcji tkanki tłuszczowej (ID 735, 1484)

Among the references provided for the scientific substantiation of the claim were narrative reviews and human studies which were unrelated to the claimed effect (e.g. performance during prolonged physical exercise, cognitive function and mood) and eight references, which reported on acute, single dose human intervention studies on the effects of caffeine on fat oxidation and/or blood lipids which did not assess changes in body fat mass (Acheson et al., 1980; Arciero et al., 1995; 2000; Dulloo et al., 1999; Hadjicharalambous et al., 2006; McNaughton, 1986; Rumpler et al., 2001; Ryu et al., 2001). The Panel considers that no conclusions can be drawn from these references for the scientific substantiation of the claim.
One reference described a longer-term human intervention study on the effects of caffeine plus epigallocatechin on body weight and body composition (Westerterp-Plantenga et al., 2005), and another reference reported the effects of a caffeine-containing vitamin and plant-based supplement administered for eight weeks on body composition compared to placebo (microcrystalline cellulose) (Malek et al., 2006). The Panel considers that no conclusions can be drawn from studies using caffeine in combination with other substances for the scientific substantiation of a claim on caffeine alone.
The Panel concludes that a cause and effect relationship has not been established between the consumption of caffeine and increased fat oxidation leading to a reduction in body fat mass.

3.2. Zwiększenie wydatków energetycznych prowadzące do redukcji masy ciała (ID 1487)

Among the references provided for the scientific substantiation of the claim were textbooks which did not contain any primary data which could be used for the scientific substantiation of the claim. Eleven human studies investigated acute effects of caffeine intake on energy expenditure (Arciero et al., 1995; 2000; Astrup et al., 1990; Collins et al., 1994; Dulloo et al., 1989; 1999; Engels et al., 1999; Koot and Deurenberg, 1995; MacNaughton et al., 1990; Poehlman et al., 1989; Tagliabue et al., 1994). None of the studies provided assessed the effects of caffeine consumption on body weight or body composition. The Panel considers that no conclusions can be drawn from these references for the scientific substantiation of the claim.
The Panel concludes that a cause and effect relationship has not been established between the consumption of caffeine and increased energy expenditure leading to a reduction in body weight.

3.3. Zwiększenie czujności (ID 736, 1101, 1187, 1485, 1491, 2063, 2103)

A number of the references provided addressed endpoints other than the claimed effect, including mood, lipid or carbohydrate metabolism, metabolic rate, physical performance and cardiovascular endpoints, or investigated substances other than caffeine, or caffeine in combination with other substances, or the effect of caffeine withdrawal rather than the effect of caffeine on subjects under normal conditions of use. A meta-analysis (Riby, 2004) addressed the effects of glucose, and a systematic review (Hoyland et al., 2008) reported on macronutrients but not caffeine, which is the
subject of the claim. The Panel considers that no conclusions can be drawn from these references for the scientific substantiation of the claim.
Two double-blind, placebo-controlled intervention trials assessed the effect of a commercial standardised guarana extract (containing approximately 12 % caffeine) on measures of alertness (Haskell et al., 2007; Kennedy et al., 2004). The Panel notes that the placebo (capsule containing no guarana extract) used in these two studies did not control for substances other than caffeine. The Panel considers that no conclusions can be drawn from these studies for the scientific substantiation of the claim.
A total of 19 randomised, double-blind, placebo-controlled intervention studies assessed the effects of caffeine on reaction time (RT), a measure of alertness as a cognitive construct. The cognitive tests used to measure RTs can be classified into simple reaction time tasks (nine studies), choice reaction time tasks (six studies) and other vigilance tasks measuring speed of reactions (e.g. rapid information processing tasks, visual or auditory vigilance tasks; 10 studies). These studies differed with respect to their design (cross-over, parallel), their sample size (11 to 120 subjects), the baseline characteristics of participants (age range 18-57 years; usual coffee consumption from 0 to 7 cups/day), the doses of caffeine administered (range 12.5 mg to 500 mg in drinks/capsules) and the time between caffeine administration and RT testing (range 20 min to 8 hours). The majority of studies used cross-over designs with relatively small sample sizes (15-25 subjects), involved young regular caffeine consumers (males and females, 18-30 years) who were on caffeine withdrawal for at least 12 hours, and administered caffeine doses ranging from 100 to 300 mg in a single dose, 45-90 min before RT testing.
Of the nine trials which assessed the effect of caffeine on simple reaction time tasks, six reported a significant reduction in RTs following caffeine consumption (dose range 12.5 mg to 320 mg) (Brice and Smith, 2002; Haskell et al., 2005; Robelin and Rogers, 1998; Smit and Rogers, 2000; Smith et al., 1992; 1993), while three studies found no effect (dose range 32 mg to 320 mg) (Hewlett and Smith, 2007; Hogervorst et al., 1999; Lieberman et al., 1987). A significant effect of caffeine on simple reaction time tasks was observed regardless of whether subjects were on caffeine withdrawal (4 trials) (Haskell et al., 2005; Robelin and Rogers, 1998; Smit and Rogers, 2000; Smith et al., 1993) or not (2 trials) (Brice and Smith, 2002; Smith et al., 1992). Most of these studies were conducted in rested subjects. One study which investigated the effect of caffeine on simple reaction time tasks during the day and at night (rested or sleep deprived subjects) reported a significant effect of caffeine in both conditions (Smith et al., 1993). The Panel notes that six of the nine studies which assessed the effects of caffeine consumption on simple RT tests showed a significant reduction in RT following caffeine consumption.
Three studies showed a significant decrease in RTs on choice-reaction time tasks after caffeine consumption (dose range 32 mg to 300 mg) (Lieberman et al., 1987; Lorist et al., 1994a; van Duinen et al., 2005). Lieberman et al. (2002) reported significantly decreased RTs on one choice-reaction time task after caffeine consumption at 200 mg and 300 mg, but not at 100 mg. One study observed a significant effect of caffeine after exercise, but not before (dose range 150 mg to 320 mg) (Hogervorst et al., 1999). One study found no effect (dose of 1 mg caffeine per kg body weight) (Hewlett and Smith, 2007). The Panel notes that five of the six studies which assessed the effects of caffeine consumption on choice-reaction time tasks showed a significant reduction in RT following caffeine consumption.
Of the 10 trials which reported the effect of caffeine on speed of reactions using other vigilance tasks, eight showed a significant decrease in RTs following caffeine consumption (dose range 75 mg to 500 mg), whereas two studies found no effect (dose range 12.5 mg to 250 mg) (Ruijter et al., 2000c; Smit and Rogers, 2000). Again, a significant effect of caffeine on RTs was observed regardless of whether subjects were caffeine deprived (4 trials) (Fine et al., 1994; Hasenfratz and Bättig, 1994;
Rosenthal et al., 1991; Smith et al., 1990) or not (4 trials) (Frewer and Lader, 1991; Lieberman et al., 2002; Smith et al., 1992; Warburton, 1995). Most of these studies were conducted in rested subjects. Two studies tested the effect of caffeine under stress conditions (sleep deprivation and/or exercise), and found a significant effect of caffeine on RTs for visual or auditory vigilance tasks (Lieberman et al., 2002; Rosenthal et al., 1991). The Panel notes that 8 of the 10 studies which assessed the effects of caffeine on other vigilance tasks showed a significant reduction in RT following caffeine consumption.
The Panel notes that the vast majority of the studies which investigated the effects of caffeine consumption on RTs using simple and choice RT tests, and other vigilance tasks, showed a reduction in RTs following caffeine consumption. The Panel also notes that the evidence provided by consensus opinions/reports (ANZFA, 2000; IoM, 2001) shows that there is good consensus on the role of caffeine in increasing speed of reaction time, and in the maintenance of speed of reactions/increased alertness, particularly in low arousal situations (e.g. sleep deprivation and fatigue).
Overall, 12 of the studies provided assessed the effects of different caffeine doses on RTs (Frewer and Lader, 1991; Hasenfratz and Bättig, 1994; Haskell et al., 2005; Hewlett and Smith, 2007; Hogervorst et al., 1999; Lieberman et al., 1987; 2002; Robelin and Rogers, 1998; Rosenthal et al., 1991; Smit and Rogers, 2000; Smith et al., 1993; Warburton, 1995). Two studies investigated dose-response effects. Using doses of 100 mg, 200 mg and 300 mg, Lieberman et al. (2002) found a dose-related effect of caffeine on RTs in a visual vigilance test (no effect on RTs in a 4-choice RT test). Warburton (1995) reported a dose-related effect of 75 mg and 150 mg caffeine in RTs on a rapid visual information processing task. The results from the other studies are heterogeneous (in some studies the effect appeared to increase with increasing caffeine doses, whereas in others the decrease in RTs appeared unrelated to the caffeine dose used). This heterogeneity suggests that the effective dose of caffeine may vary from individual to individual, and may depend on many factors (IoM, 2001). Although some effects have been observed at very low caffeine doses (<75 mg), results were inconsistent (Hewlett and Smith, 2007; Lieberman et al., 1987; Smit and Rogers, 2000). The majority of significant results were obtained when subjects consumed caffeine doses in the range of 75-500 mg. This finding is consistent with consensus opinions reporting on caffeine doses in the range of 100-400 mg (IoM, 2001) and of 60-600 mg (ANZFA, 2000) which have consistently demonstrated reductions in reaction time and enhanced speed of reaction in vigilance tests.
In weighing the evidence, the Panel took into account that the evidence provided by consensus opinions/reports, and by the majority of the studies submitted for the scientific substantiation of the claim, showed that there was good consensus on the role of caffeine in increasing alertness measured as speed of reaction times in healthy individuals of both sexes, at doses of at least 75 mg.
The Panel concludes that a cause and effect relationship has been established between the consumption of caffeine and increased alertness.

3.4. Zwiększenie uwagi (ID 736, 1485, 1491, 2375)

A number of the references provided addressed endpoints other than the claimed effects, including mood, cognitive domains other than attention, lipids or carbohydrate metabolism, metabolic rate, physical performance and cardiovascular endpoints, or investigated substances other than caffeine, caffeine in combination with other substances, the effect of caffeine withdrawal rather than the effect of caffeine on subjects under normal conditions of use, or the effect of caffeine in patients with impaired adrenaline responses. A meta-analysis (Riby, 2004) addressed to glucose, and a systematic review (Hoyland et al., 2008) reported on macronutrients but not caffeine, which is the subject of the claim. The Panel considers that no conclusions can be drawn from these references for the scientific substantiation of the claim.
Two double-blind, placebo-controlled intervention trials assessed the effect of a commercial standardised guarana extract (containing approximately 12 % caffeine) on measures of attention (Haskell et al., 2007; Kennedy et al., 2004). The Panel notes that the placebo (capsule containing no guarana extract) used in these two studies did not control for substances other than caffeine. The Panel considers that no conclusions can be drawn from these references for the scientific substantiation of the claim.
One study (Brice and Smith, 2002) investigated the effects of caffeine on performance in three attention tests (focused attention task, categoric search task, 5-choice serial RT). Results for the focused attention task were not reported, and although it was stated that measures of RT and accuracy were obtained for the categoric search task, these results were also not reported. The 5-choice serial RT test was administered for too short a period (3 minutes) to adequately assess sustained attention. Another study (Loke, 1990) examined the effects of caffeine on performance in a symbol cancellation task, but did not report the main effects of the intervention. Haskell et al. (2005) observed the effects of caffeine on performance in a rapid visual information processing test and a digit vigilance reaction time test, in both habitual caffeine consumers and non-consumers. Primary analyses of the effects of caffeine on rapid visual information processing were not reported. No description of the digit vigilance reaction time was provided, and whether this is a validated test of sustained attention cannot be determined. The Panel considers that no conclusions can be drawn from these studies for the scientific substantiation of the claim.
A total of 22 randomised, double-blind, placebo-controlled intervention studies assessed the effects of caffeine on attention. The studies differed with respect to their design (cross-over, parallel), their sample size (11 to 120 subjects), the baseline characteristics of participants (age range 18-72 years; usual coffee consumption from 0 to 7 cups/day), the doses of caffeine administered (range 12.5 mg to 500 mg in drinks/capsules), and the time between caffeine administration and attention testing (range 20 min to 8 hours). The majority of studies used cross-over designs with relatively small sample sizes (15-25 subjects), involved young regular caffeine consumers (males and females, 18-30 years) who were on caffeine withdrawal for at least 12 hours, and administered caffeine doses ranging from 100 mg to 300 mg in a single dose, 45-90 min before attention testing.
Ten studies examined the effect of caffeine on selective attention by means of a range of psychometric tasks (e.g. visual selective search task, colour selection task, spatial-selection task, Stroop test and categoric search attention task). Six of these studies also examined the effect of caffeine on event-related potentials (ERPs). The ERP is an electrophysiological measure of cortical activity which is time-locked to a specific stimulus event. The ERP is analysed into waves which vary in latency, amplitude and direction (negative and positive polarity). Different components of the ERP are associated with different mental operations. Of relevance to selective attention is the occurrence of the N2b negative wave 200 ms after stimulus onset. This N2b component is related to covert (internal) orienting of attention towards relevant information, and increased selective attention is identified by a greater amplitude in the N2b component.
Lorist et al. (1994b) reported that caffeine (250 mg, n=30 subjects per intervention arm, 18-25 years) significantly reduced RT in a selective attention task (p<0.02), but had no effect on performance accuracy. The Panel notes that faster response time, indicated by reduced RT, and with no associated increase in error rate, is evidence for improved attention and efficiency at processing information. In the same study, analysis of the ERP showed that the N2b amplitude was greater in the caffeine than in the placebo condition (p<0.03). Lorist et al. (1995) reproduced these findings in a sample of 15 young (18-23 years) and 15 older (60-72 years) subjects, where caffeine (250 mg, n=30 per intervention arm) significantly reduced RT on the same selective attention task (p=0.001) with no effects on accuracy. Additionally, the N2b amplitude was significantly increased in the caffeine compared to the placebo condition (p<0.03). Lorist et al. (1996) also observed a significant effect of caffeine (3 mg/kg body weight, n=16 subjects per intervention arm, 19-29 years) in the focused attention condition of a visual
selective search task. RT was significantly reduced (p=0.004), and the N2b amplitude was significantly increased (p<0.025). Lorist et al. (1994a) reported similar results in a study on the effects of caffeine on performance in three different tests of selective attention (stimulus degradation, stimulus-response compatibility and time uncertainty tasks). Significant effects of caffeine (250 mg, n=30 per intervention arm) were found in all tasks, with reduced RTs (p<0.004) and fewer errors (p<0.05). ERP measures were also obtained, but data on the N2b component were not reported. Ruijter et al. (2000a) found a significant effect of caffeine (250 mg, n=11 per intervention arm) on RT in a colour selection task (p=0.032) with no associated effects on errors. Ruijter et al. (2000a) also obtained ERP measures, and reported a significantly greater N2b amplitude in the caffeine group (p=0.023). In the same experiment (Ruijter et al., 2000b), no effect of caffeine was found on RT in a spatial selection task, albeit accuracy was significantly improved (p=0.029). The effect of caffeine on the N2b amplitude was not significant. Hogervost et al. (1999) reported a significant effect of caffeine at each of three different doses (150 mg, 225 mg or 320 mg, n=15 per intervention arm) on a Stroop test after strenuous exercise (p<0.05), but not before exercise. Frewer and Lader (1991) studied the effects of caffeine at two doses (250 mg or 500 mg, n=12 per intervention arm) on one- and two-target letter cancellation tasks. Caffeine at 250 mg produced significantly faster completion times compared to placebo on the one-target task (p<0.05), and at 500 mg on the two-target task (p<0.05). Hewlett and Smith (2007) observed no effects of caffeine (1 or 2 mg/kg body weight, n=30 or 60 subjects per intervention arm) on performance in a focused attention choice reaction time task and a categoric search task. Lorist and Snel (1997) failed to observe any significant effect of caffeine (3 mg/kg body weight, n=16 subjects per intervention arm) on performance in a visual selective attention task.
In summary, all 10 of these studies included participants who were habitual caffeine users who had avoided caffeine consumption for a period ranging from 7-24 hours. One study included groups of high- and low-caffeine users (Hewlett and Smith, 2007), but no effects of caffeine were observed in either group. Three of the 10 studies assessed the effects of different caffeine doses on measures of selective attention (Frewer and Lader, 1991; Hewlett and Smith, 2007; Hogervorst et al., 1999). Only Frewer and Lader (1991) reported any dose-related effects. Caffeine at 250 mg produced significantly faster completion times compared to placebo on a one-target letter cancellation task (p<0.05), but the effect of caffeine at 500 mg was not significant. These effects were reversed for a two-target task (250 mg: not significant; 500 mg: p<0.05).
The Panel notes that although there were notable differences in the duration of avoidance of caffeine prior to testing, and in the measures used to assess selective attention, overall a consistent effect of caffeine was observed on a variety of measures of selective attention (15 positive outcomes vs. 6 negative outcomes, from 10 trials).
A total of 14 studies examined the effect of caffeine on sustained attention (vigilance). Sustained attention was measured by a variety of psychometric tasks which included digit detection, rapid visual information processing, visual vigilance, continuous performance, 5-choice serial RT, Wilkinson auditory vigilance task and the Bakan vigilance task.
Fine et al. (1994) found that caffeine (200 mg, n=20 subjects per intervention arm) significantly increased the number of correct responses (p<0.0005) and decreased response times (p=0.002) compared to placebo in a visual vigilance task. Hasenfratz and Bättig (1994) reported that caffeine at three different doses (1.5 mg, 3 mg and 6 mg/kg body weight, n=20 per intervention arm) significantly decreased RT (p<0.01), but had no effects on processing rate compared to placebo in the rapid information processing task. No dose-response effect was observed in this study. Rosenthal et al. (1991) reported similar effects for two doses of caffeine (75 mg and 150 mg, n=12 subjects per intervention arm), both of which significantly reduced RT in an auditory vigilance task (p<0.001) with no effects on error rate. No dose-response effect was observed. Smith et al. (1990) showed that caffeine (3 mg/kg body weight, n=32 subjects per intervention arm) significantly improved performance on the Bakan vigilance task with reduced RT and better accuracy compared to placebo
(p<0.05). The effects were observed both before and after lunch. Smith et al. (1992) replicated these results in a study on the effects of caffeine (4 mg/kg body weight, n=24 subjects per intervention arm) on the Repeated Digits Vigilance Task. Caffeine significantly reduced RT (p<0.0001) and increased accuracy (p<0.0001), both before and after lunch. Smith et al. (1993) extended these findings to show that caffeine (1.5 or 3 mg/kg body weight, n=24 per intervention arm) significantly increased the number of correct responses on the 5-Choice Serial Response Task (p<0.005) at two different doses. These effects were found both during the day and at night. Frewer and Lader (1991) investigated the effects of caffeine (250 mg or 500 mg, n=12 subjects per intervention arm) on two different tests of sustained attention (rapid information processing task and continuous attention task). Accuracy in the rapid information processing task was significantly increased at both caffeine doses (p<0.02), but RT was significantly reduced only at the higher 500 mg dose (p<0.01). Caffeine at both doses significantly improved accuracy in the continuous attention task (p<0.05). Lieberman et al. (1987) investigated the effects of caffeine (32 mg, 64 mg, 128 mg and 256 mg, n=20 subjects per intervention arm) on three different tests of sustained attention (Wilkinson auditory vigilance task, continuous performance task, 4-choice visual RT). Caffeine at all doses significantly improved accuracy on the Wilkinson auditory vigilance task (p<0.0016), and significantly reduced RT on the 4-choice visual RT test (p=0.02), with no accompanying effects on accuracy. No effects of caffeine were observed on the continuous performance task. Warburton (1995) examined the effects of caffeine (75 mg and 150 mg, n=18 subjects per intervention arm) on the rapid information processing test in participants with minimal caffeine deprivation. There was a significant effect of caffeine on performance accuracy (p<0.01), with greater accuracy occurring in the 150 mg condition compared to placebo. A significant reduction in RT was also observed in the caffeine compared to placebo conditions (p<0.05), which was also dose related (p<0.05). Smit and Rogers (2000) studied the effects of caffeine (12.5 mg, 25 mg, 50 mg and 100 mg; n=23 subjects per intervention arm) on the rapid visual processing task in habitual high- and low-caffeine users. Caffeine at all doses significantly improved performance of high-caffeine users (p<0.004), but had no effect on the performance of low-caffeine users, and there were no effects on RT in either group.
Two further studies examined the effects of caffeine on sustained attention in sleep-deprived individuals. Bonnet et al. (1995) studied the effects of either a single dose (400 mg) or repeated doses (150-300 mg every 6 hours for 42 hours) of caffeine on performance in a visual vigilance test in participants with varying durations of sleep deprivation. There were 27 participants in the placebo group, 17 in the 150-300 mg group, and 12 in the 400 mg group. Visual vigilance performance was significantly improved in the caffeine groups, but only up to 24 hours sleep deprivation (group x time interaction p<0.001). Lieberman et al. (2002) studied the effects of caffeine (100 mg, 200 mg and 300 mg, n=17 subjects per intervention arm) in participants who underwent 72 hours sleep deprivation as well as physical and environmental stress challenges prior to receiving their caffeine intervention. Performance was measured in two tests of sustained attention (4-choice visual RT test and scanning visual vigilance test). Caffeine produced significant beneficial, dose-related effects including an increase in the number of correct responses (p=0.0464) and a decrease in RT (p=0.0418) on the scanning visual vigilance test. Significantly more correct responses in the 4-choice visual RT test were shown by both the 200 mg and 300 mg caffeine groups compared to placebo, at one hour after caffeine intake (p<0.05).
Two studies found no effects of caffeine on measures of sustained attention. Ruijter et al. (2000c) found no effects of caffeine (250 mg; n=12 subjects per intervention arm) on performance in a sustained attention task (a modified Bourdon test). The authors noted that the attention test used was more cognitively demanding than other sustained attention tests. Hewlett and Smith (2007) observed no effects of caffeine (1 or 2 mg/kg body weight, n=30 or 60 per intervention arm) on performance in a repeated digit detection task administered for five minutes. The Panel notes that the duration of this task may have been insufficient for an adequate assessment of sustained attention.
In summary, nine out of the 14 studies included participants who were habitual caffeine users and who had avoided caffeine consumption for a period ranging from three hours to one week. Four studies compared the effects of caffeine in groups of high- and low-caffeine users (Fine et al., 1994; Hewlett and Smith, 2007; Lieberman et al., 1987; Smit and Rogers, 2000), who had avoided caffeine consumption for 7-12 hours. Two of these studies reported significant effects of caffeine on sustained attention performance and which were the same in both high- and low-caffeine users (Fine et al., 1994; Lieberman et al., 1987). One study (Smit and Rogers, 2000) reported significant effects of caffeine in high-caffeine users, but no effects in low-caffeine users. Hewlett and Smith (2007) found no effects of caffeine in either high- or low-caffeine users, but the duration of the test period may have been insufficient for adequate assessment of sustained attention.
Ten of the 14 studies assessed the effects of different caffeine doses on measures of selective attention (Bonnet et al., 1995; Frewer and Lader, 1991; Hasenfratz and Bättig, 1994; Hewlett and Smith, 2007; Lieberman et al., 1987; 2002; Rosenthal et al., 1991; Smit and Rogers, 2000; Smith et al., 1993; Warburton, 1995). Two studies reported significant dose-related effects. Warburton (1995) found greater rapid information processing test accuracy in both 75 mg and 150 mg caffeine groups compared to placebo, and the effect was greater in the 150 mg group (p<0.01). RT was significantly reduced in both caffeine groups (p<0.05), and this effect was also dose related (p<0.05). Lieberman et al. (2002) examined the effects of three doses of caffeine (100 mg, 200 mg and 300 mg) on visual vigilance test performance, and reported a significant linear increase in accuracy with increasing caffeine intake (p<0.05).
The Panel notes that although there were notable differences in the duration of avoidance of caffeine prior to testing and in the measures used to assess sustained attention, overall a consistent effect of caffeine was observed on a variety of measures of sustained attention (35 positive outcomes vs. 12 negative outcomes, from 14 trials).
Overall, the Panel notes the consistency of the results of studies on both selective and sustained attention, with a majority showing positive effects of caffeine. Although some effects have been observed at very low caffeine doses (<75 mg), results were inconsistent (Hewlett and Smith, 2007; Lieberman et al., 1987; Smit and Rogers, 2000). The majority of significant results were obtained when subjects consumed caffeine doses in the range of 75-500 mg. This finding is consistent with consensus opinions reporting on caffeine doses in the range of 100-400 mg (IoM, 2001) and of 60-600 mg (ANZFA, 2000) which have consistently demonstrated an effect of caffeine on tests of attention.
In weighing the evidence, the Panel took into account that the evidence provided by consensus opinions/reports and by the majority of the studies submitted for the scientific substantiation of the claim showed that there was good consensus on the role of caffeine in increasing attention, measured by a range of psychometric tasks, in healthy individuals of both sexes, at doses of at least 75 mg.
The Panel concludes that a cause and effect relationship has been established between the consumption of caffeine and increased attention.

4.1. Zwiększenie czujności (ID 736, 1101, 1187, 1485, 1491, 2063, 2103)

The Panel considers that the following wording reflects the scientific evidence: “Caffeine helps to increase alertness”.

4.2. Zwiększenie uwagi (ID 736, 1485, 1491, 2375)

The Panel considers that the following wording reflects the scientific evidence: “Caffeine helps to improve concentration”.