Caffeine and the Andrew Health
common
cold
Smith, Marie Thomas, Kate Perry and Helen Whitney
Psychology Research Unit, Department of Psychology, University of Bristol, Bristol BS8 1TN, UK.
An experiment was carried out to determine whether caffeinated and decaffeinated coffee removed the malaise (reduced alertness, slower psychomotor performance) associated with having a common cold. One hundred volunteers were tested when healthy and 46 returned to the laboratory when they developed colds. Those subjects who remained healthy were then recalled as a control group. On the second visit subjects carried out two sessions, one pre-drink and another an hour after the drink. Subjects were randomly assigned to one of the following three conditions, caffeinated coffee (1.5 mg/kg caffeine/body weight), decaffeinated coffee or fruit juice. Subjects with colds reported decreased alertness and were slower at performing psychomotor tasks. Caffeine increased the alertness and performance of the colds subjects to the same level as the healthy group and decaffeinated coffee also led to an improvement. These results suggest that drugs which increase alertness can remove the malaise associated with the common cold, and that increased stimulation of the sensory afferent nerves
may also be
beneficial.
Key words: alertness; caffeine; coffee; psychomotor performance;
Introduction Minor illnesses such as the common cold are frequent, and a major cause of absence from work and education. In addition, there is now strong evidence showing that both experimentally-induced and naturally-occurring upper respiratory tract illnesses (URTIS) are associated with increased negative mood and impairments of mental functioning. These findings can be briefly summarized as follows (see Smith, 1990, 1992 for reviews of the earlier work).
widespread
Experimentally-induced URTIs Both colds and influenza have selective effects on performance, with only some functions, not all, being impaired. The profile of performance effects was different in studies of influenza to those observed in experiments on the effects of colds. Influenza illnesses were found to increase reaction times on tasks where there is uncertainty as to where or when the next target will appear (Smith et al., 1987a). Psychomotor tasks involving speed and accuracy of hand-eye co-ordination were not impaired during influenza, neither were cognitive tasks, such as logical reasoning. In contrast, cold-type viral infection impaired performance on psychomotor tasks such as a pursuittracking tasks, five-choice serial response task and pegboard task (see Smith et al., 1987a, b, 1988). Subjects with colds were not impaired on the detection tasks affected by influenza, and there was little evidence of subjects with colds showing changes in memory performance. The effects of the experimentallyinduced URTIs on mood also depended on the type of virus, with influenza leading to a general increase in negative affect and colds mainly reducing alertness (Smith et al., 1992). These effects may have reflected the severity of the illness rather than
upper
respiratory tract illness
the virus per se. The above results reflect differences between subjects who were symptomatic and those who were uninfected. Behavioural changes were also found in volunteers with subclinical infections and when subjects were tested during the incubation period of the illness and in convalescence. In other words, the performance changes were not dependent on the presence of symptoms (see Smith et al., 1988, 1989).
Naturally occurring URTIs The above findings have largely been replicated in studies of naturally-occurring illnesses (see Smith et al., 1993a, for details of the effects of naturally-occurring influenza and Hall and Smith, 1996 for effects of naturally-acquired colds). Other research has examined whether illness makes people more sensitive to the effects of other factors known to alter mood. Smith et al. (1993b) with colds were more sensitive to the effects of noise than were healthy volunteers. Main effects of both noise (on cognitive vigilance tasks) and colds (on psychomotor tasks) were found. However, there was also evidence to support the view that a combination of having a cold and being exposed to noise resulted in the worst performance (although this effect was restricted to a variable fore-period simple reaction-time task). Smith et al. (1995) investigated whether a low dose of alcohol ( 1.5 ml of vodka/kg body weight) produced different behavioural effects in healthy volunteers to those suffering from URTIS. Subjects with colds reported an increase in negative affect and were slower at performing the psychomotor tasks. Interactions between health status and alcohol conditions were also found. Alcohol improved the mood of healthy subjects but produced greater negative mood in those with colds. Similarly,
performance efficiency and investigated whether subjects
Downloaded from jop.sagepub.com at CAMBRIDGE UNIV LIBRARY on October 23, 2015
320
performance of selective and sustained attention tasks showed different effects of alcohol in healthy and ill subjects. Possible mechanisms underlying these effects The behavioural changes associated with URTIs have important practical implications and treatments for these illnesses should aim not only to reduce local and systemic symptoms but to remove the malaise. This is difficult as ,the mechanisms underlying the development of malaise are not clearly understood. Two approaches are outlined in this paper. The first involves increasing central arousal by administering caffeine. Previous research has shown that, even in low doses, caffeine can produce beneficial effects when alertness is low. This has been demonstrated by studying sleep deprivation (Lorist et al., 1994), nightwork (Smith et al., 1993c), and the post-lunch dip (Smith et al., 1990). The previous findings suggest that at least part of the malaise associated with a cold appears to involve reduced alertness. Because caffeine has been shown to increase alertness, it was predicted that caffeine should be able to return the cold-induced impairments to levels similar to those seen in healthy subjects. The second method considered here involved increasing sensory afferent stimulation by giving the subject a hot drink. Previous studies have shown that compounds which increase stimulation of the sensory afferent trigeminal nerve (e.g. zinc gluconate, Smith et al., 1991; nedocromil sodium, Barrow et al., 1990) can remove the malaise produced by experimentallyinduced URTIs. It was predicted that a hot drink would reduce the impairments associated with having a cold. The aims of the present study can, therefore, be summarized as follows. First, the study provided another opportunity to confirm that naturally occurring colds lead to decreased alertness and psychomotor impairments but have little effect on other cognitive functions (e.g. episodic, working or semantic memory). Secondly, the experiment examined whether increasing alertness would remove the impairments associated with having a cold, and the efficacy of two methods for doing this was compared.
Method Individuals from the Health Psychology Research Unit subject panel took part in the study. Prior to the experiment subjects
completed questionnaires measuring personality, psychosocial factors and health-related behaviours. This background information was important to ensure that effects attributed to experimental manipulations did not merely reflect differences in the above factors. Subjects were practised at the tasks and completed a baseline session (when all subjects were healthy). If subjects then developed an URTI they returned to the Unit and completed another session of tests, followed by the drink manipulation and another test session. Subjects were randomly assigned to drinks conditions. The caffeine manipulation was double-blind. The juice condition was included for two main reasons.
First, it was a cold drink rather than a hot drink. Second, it represented a condition where subjects were aware that no caffeine had been administered. Previous research with this design has shown little difference between the decaffeinated coffee and juice conditions (Smith et al.,
1993c). However, it was unknown whether this would apply subjects with colds were tested. When about half of the original panel had returned with a URTI the remaining subjects were recalled and retested (forming the healthy group). The entire study lasted about 10 weeks and was carried out in the months of January-March. when
Subjects subjects (47 females, 53 males, mean age 21 years, age range 18-30 years) took part in the study. Thirtyfive subjects were assigned to the caffeine condition (15 colds, 20 healthy), 32 to the decaffeinated condition (16 colds, 16 healthy) and 33 to the juice condition (16 colds and 17 healthy). Prior to participating in the study the volunteers were given both written and verbal information regarding the study and signed a consent form. One hundred
Nature of the drinks Subjects were given a 150-ml drink (representative of a typical mug of coffee). One condition consisted of orange juice, another decaffeinated coffee and the third decaffeinated coffee with 1.5 mg/kg body weight of caffeine tablets added (average amount of caffeine =97.5 mg). Subjects were free to add milk and sugar to the coffee according to their normal practice.
Schedule of testing
subjects completed a familiarization session which involved practise at the tests and completion of questionnaires measuring personality traits (the Eysenck Personality Inventory and Spielberger Trait Anxiety Inventory), healthrelated behaviours (alcohol consumption, smoking and diet) and stress (life events, perceived stress and negative moodAll
Cohen et al., 1993). Baseline sessions were completed in the morning (between 09.00 and 11.00 hours) and the second session was at the same time of day. Subjects had abstained from consuming caffeinated drinks for 2 h prior to testing. At the end of the second session the subjects consumed their drink. Subjects were then allowed a rest before the third session, which commenced 1 h after the end of the second session. These timings were selected on the basis of previous studies showing effects of caffeine 1 h after ingestion. see
Criteria for a cold Subjects were told to return to the unit if they developed a cold. They were told to wait at least 24 h after the start of the illness before returning but to ensure that they were retested within 4 days of the start of the illness. All subjects who reported URTIs completed a standard symptom check list and
categorized as having a cold if they had two or more respiratory tract symptoms (runny or blocked nose, sneezing, sore throat) and produced greater than 0.2 g of nasal secretion (subjects blew their nose and then waited for 15 min before using a pre-weighed tissue to blow their nose again. Weight of nasal secretion was determined by a further weighing). Sublingual temperature was recorded to determine whether the subjects had a fever. were
upper
Downloaded from jop.sagepub.com at CAMBRIDGE UNIV LIBRARY on October 23, 2015
321
Measurement of mood Mood ratings were recorded using bipolar visual analogue scales presented on the computer screen. Eighteen scales were presented (e.g. drowsy-alert; tense-calm) and these were based on the scales used by Herbert et al. (1976). Subjects completed the mood ratings before and after the performance tasks.
Performance tests Two main types of tests were examined here. Psychomotor tasks which had been shown to be sensitive to cold effects (see Introduction) were included, as were memory tasks which were not impaired by having a cold. All of these tasks, except for the pegboard task, were presented on an IBM compatible PC.
Psychomotor tasks Simple reaction-time task.
In this task a box was displayed on the screen and this was followed after a period of 1-8 sec by a square (the target) being presented in the middle of the box. The subject had to press a key as soon as the square was detected and, following this, another box was presented. This task lasted for 3 min.
Five-choice serial response task. Five boxes were displayed on the screen and a square appeared in one of the boxes. The subject pressed the corresponding key and the square then appeared in another box and the subject had to press the next key. This task lasted for 3 min and the number of responses made, number of errors and number of gaps (occasional long responses > 1500 msec) were recorded.
and decide whether the sentence was a true description of the order of the letters. If it was, the subject pressed the T key on the keyboard, if it wasn’t, they pressed the F key. The sentences ranged in syntactic complexity from simple active to passive negative (e.g. A is not followed by B). Subjects carried out the task for 3 min.
Semantic-processing task. This test, developed by Baddeley (1981), measures speed of retrieval of information from general knowledge. Subjects were shown a sentence and had to decide whether it was true (e.g. canaries have wings). The number completed in 3 min was recorded, as was the accuracy of doing the test.
Results Analyses of variance distinguishing colds/healthies and drinks conditions showed that subjects in the various groups did not differ in terms of demographics, personality, stress levels or health-related behaviours. The subjects with URTIs had greater nasal symptoms and signs than the healthy subjects but did not report systemic symptoms. Indeed, there was no evidence of an increased temperature in the URTI group, which suggests that the illnesses were colds rather than influenza. Effects of a cold (pre-drink) of the data from Session 2 allows one to examine having a cold produced changes in mood and performance. Analyses of covariance were carried out with the baseline data as covariates and pre-drink scores as the dependent variables. This statistical technique adjusts the dependent variables to take account of unwanted variation at baseline. The between-subject factor was colds vs healthy
Analysis whether
Pegboard test. This test involved transferring pegs from a full pocket solitaire set to an empty one. Pegs were transferred one at a time from the full set to the equivalent hole in the empty set. The subjects were instructed to use their dominant hand. For right-handed subjects the full set was on the left and the empty set on the right. The first peg to be moved was the one at the extreme top right, and the volunteer went down each column, ending with the peg in the extreme bottom left of the set (left-handed subjects had the full set on the right and proceeded from left to right rather than right to left). Memory tasks A list of 20 words was presented on the PC of one every 2 sec. At the end of the list, the subject had 2 min to write down (in any order) as many of the words as possible. Subjects were shown a different list at each test session.
Free-recall task.
subjects. Mood A global increase in negative affect was reported by subjects with colds (both pre- and post-performance mood, all p < 0.05). This effect is illustrated by the drowsy-alert (postperformance rating) in Fig. I [main effect of colds:
MSerror = 68.4].
F( 1,97) =13.4, p < 0.001,
screen at a rate
Psychomotor tasks Both the simple reaction-time task and five-choice serial response task were performed more slowly [simple reactiontime task: F(1,97) 7.35, p < 0.01, MSerror = 7018; five-choice task: F( 1,97) = 7.64, p < 0.01, MSerror = 326] when the subject had a cold (see Fig. 1 ). There was no significant effect of having a cold on the pegboard task, which confirms the negative finding of Hall and Smith (1996). =
Delayed-recognition memory task. At the end of the test session, subjects were shown a list of 40 words which consisted of the 20 words shown at the start of the session, plus 20 distracters. The subjects had to decide as quickly as possible whether each word had been shown in the original list or not. Both speed and accuracy of responding were recorded.
Memory tasks There
was no
effect of
Logical-reasoning task. This test was developed by Baddeley (1968), and the subjects were shown statements about the
memory tasks (all p > In summary, the
order of the letters A and B followed by the letters AB or BA (e.g. A follows B: BA). The subjects had to read the statement
confirmed that
psychomotor
having 0.05).
a
cold
on
performance of
the
results from the pre-drink session having a cold reduces alertness, leads to slowing but has no effect on memory.
Downloaded from jop.sagepub.com at CAMBRIDGE UNIV LIBRARY on October 23, 2015
322
1 Effects of the common cold (pre-drink). (a) Alertness, scores are the adjusted mean ratings (range = to 50), high scores indicate greater alertness. (b) Simple reaction-time task, scores are the adjusted mean reaction time (msec). (c) Five-choice serial response task, scores are the adjusted mean number of responses in 3 min
Figure
Post-drink ef fects In these analyses baseline measurements were still used covariates and post-drink scores as the dependent variable. Main effects of colds. All of the effects found pre-drink still significant in the post-drink session. No new effects
were were
significant. Main effects of drink. The only main effect of drink conditions was found in the analysis of the simple reactiontime data [F(2,93) = 4.89, p < 0.01], with those given caffeine having faster reaction times than the other groups.
Colds x drinks interactions. Interactions between colds and drinks were found for the post-performance alertness ratings
[e.g. drowsy-alert: F(2,93)=4.40, p < 0.05, MSerror = 72.0], simple reaction-time task [F(2,93) 4.24, p < 0.05, MSerror=8686] and five-choice serial response speed [F(2,93)=5.35, p < 0.05, MSerror = 468]. These data are =
shown in Table l. In these analyses those with colds given caffeine returned to a similar level to that of the healthy subjects, and there
Table 1
Effects of different drinks
on
alertness and psychomotor
indication of beneficial effects (especially on those with colds of even the decaffeinated coffee. These effects are shown in Fig. 2. Tukey tests showed that the juice/colds group were significantly impaired compared to the juice no/cold group, whereas the colds and healthy groups were not significantly different in the coffee conditions. No interactions between colds/drinks were significant in the other tasks. was
as
also
some
speed) for
Discussion The results obtained in this study confirm that minor illnesses, such as the common cold, can alter mood and psychomotor performance. The major findings, however, are those showing interactions between drinks and cold status. Caffeinated coffee largely eliminated the effects of having a cold. This confirms results from studies of other low arousal states which show that even a small dose of caffeine can be beneficial when alertness is reduced. Caffeine produces its behavioural effects in a number of ways. The primary action of caffeine is competitive antagonism
performance of subjects with colds and healthy subjects
-~---
Scores
are
the
adjusted
means
from the
analyses of covariance. Downloaded from jop.sagepub.com at CAMBRIDGE UNIV LIBRARY on October 23, 2015
&dquo;
323
2 Differences (post-drink) between subjects with colds and healthy subjects in the different drink conditions. (a) Alertness, scores are the differences between the means of the cold and healthy groups (healthy-colds), high scores indicated greater cold-induced reduction in alertness. (b) Simple reaction-time task, scores are the difference between the mean reaction times of the colds and healthy groups (colds-healthy), (msec), high scores indicate greater cold-induced slowing of reaction time. (c) Five-choice serial response task, scores are the difference between the mean number done for the colds and healthy groups (healthy-colds), high scores indicate greater cold-induced slowing of responses
Figure
of adenosine receptors. However, there is also evidence that the noradrenergic and dopaminergic neurotransmitter systems can be altered by caffeine. Indeed, recent research provides a plausible mechanism which could account for the effects of caffeine in low-alertness states. Smith and Nutt (1996) have shown that clonidine, a drug which reduces the turnover of central noradrenaline, can induce a low-arousal state. This can be removed either by factors which increase alertness (e.g. white noise) or by ingestion of idazoxan, a drug which increases central noradrenaline. There are a number of pieces of evidence which suggests that caffeine increases central noradrenaline (see Fredholm, 1993). First, caffeine influences adenosine, which in turn inhibits noradrenaline release. Second, animal research has shown increased turnover of noradrenaline and increased locus coeruleus timing following caffeine. Finally, there is evidence of down-regulation of receptors following chronic administration of caffeine. It is plausible, therefore, that the effect of caffeine seen here reflects changes in the noradrenergic neurotransmitter system. If this is the case then compounds such as idazoxan should produce similar benefits. The noradrenaline model can also account for the larger effects of alcohol and noise in subjects with colds, in that both of these factors influence noradrenaline. It is, of course, possible that other neurotransmitter systems are involved as well (e.g. dopamine) and further research will be required to determine if this is the case. The present result also has important implications for therapy. Certain medications already contain caffeine and it will be of interest to investigate whether caffeine presented in this form and dose also has beneficial effects. Another result found here also has important implications for treatment of the illness. Decaffeinated coffee produced improvements in those with colds which suggests that hot drinks are beneficial in this type of illness. This could reflect the local relief produced by the drink or could be due to increased stimulation
of the sensory nerves. If the former is true then other forms of medication (e.g. decongestants) may also produce similar effects. On the other hand, if increased trigeminal stimulation is important then compounds such as menthol should also reduce the malaise associated with the common cold. However, it is clear that the improvements associated with a hot drink are much smaller than when the hot drink includes caffeine. It is also possible that the effects observed in the decaffeinated coffee condition reflected possible expectancy effects associated with the double-blind procedure. Again, further research is needed to resolve this issue. The present results show that the common cold leads to reduced alertness and slower psychomotor speed. These impairments can be removed by caffeinated coffee and, to some extent, by decaffeinated coffee. These results have important practical implications and suggest further studies to determine the underlying mechanisms. The findings also suggest that medication should not only provide symptomatic relief but reduce the malaise associated with the cold. Addition of caffeine may be one way to achieve this, and there is clearly some support for the practice of taking hot drinks when suffering from colds. Further studies must not only consider the topic of malaise induced by colds but also examine whether similar effects are apparent in other diseases. The results of such studies will not only be relevant to the treatment of malaise but also to our understanding of the mechanisms underlying effects of reduced arousal on behaviour.
Acknowledgement The research described in this paper was supported by a grant from the Institute for Scientific Information on Coffee.
Downloaded from jop.sagepub.com at CAMBRIDGE UNIV LIBRARY on October 23, 2015
324
Smith A P, Nutt D J (1996) Noradrenaline and attention lapses. Nature 380: 291 Smith A P, Tyrrell D A J, Coyle K B, Willman J S (1987a) Selective effects of minor illnesses on human performance. Br J Psychol :183-188 78 Smith A P, Tyrrell DAJ, Al-Nakib W, Coyle KB, Donovan C B, Higgins PG, Willman JS (1987b) Effects of experimentally induced respiratory virus infections and illness on psychomotor
Address for correspondence Professor Andrew Smith
Department of Psychology University of Bristol 8 Woodland Road
Bristol BS81TN UK
Email: [email protected]
:144-148 performance. Neuropsychobiology 18 Tyrrell DAJ, Al-Nakib W, Coyle KB, Donovan C B, Higgins PG, Willman JS (1988) The effects of experimentallyinduced respiratory virus infections on performance. Psychol
Smith A P,
References reasoning test based on gramBaddeleyAD matical transformation. Psychonom Sci 10: 341-342 Baddeley A D (1981) The cognitive psychology of everyday life. Br J : 257-269 Psychol 72 Barrow G I, Higgins P G, Al-Nakib W, Smith A P, Wenham R B M, Tyrrell DAJ (1990) The effect of intranasal nedocromil sodium on viral upper respiratory tract infections in human volunteers. Clin Exp Allergy 20: 45-51 Cohen S, Tyrrell D A J, Smith AP (1993) Negative life events, perceived stress, negative affect and susceptibility to the :131-140 common cold. J Pers Soc Psychol 64 Fredholm B (1993) Central pathways affected by xanthines. In: Proceedings of seventh international caffeine workshop. International Life Sciences Institute, Washington DC Hall S R, Smith A P (1996) An investigation of the effects and aftereffects of naturally- occurring upper respiratory tract illnesses on mood and performance. Physiol Behav 59: 569-577 Herbert M, Johns M W, Dore C (1976) Factor analysis of analogue scales measuring subjective feelings before sleep and after sleep. Br J Med Psychol 49: 373-379 Lorist M, Snel J, Kok A (1994) Influence of caffeine on information processing stages in well rested and fatigued subjects. Psycho: 411-421 pharmacology 113 Smith A P (1990) Respiratory virus infections and performance. (1968) A three-minute
Phil Trans R Soc Lond B327: 519-528 Smith A P (1992) Colds, influenza and performance. In Smith A P, Jones D M (eds), Handbook of human performance, vol 2. Academic Press, London, pp. 197-218
Med 18: 65-71 Smith A P, Tyrrell DAJ, Al-Nakib W, Barrow GI, Higgins PG, Leekam S, Trickett S (1989) Effects and after-effects of the common cold and influenza on human performance. Neuropsy: 90-93 chobiology 21 Smith A P, Rusted J M, Eaton-Williams P, Savory M, Leathwood P (1990) Effects of caffeine given before and after lunch on :160-163 sustained attention. Neuropsychobiology 23 Smith A P, Tyrrell DAJ, Al-Nakib W, Barrow GI, Higgins P G, Wenham R (1991) The effects of zinc gluconate and nedocromil sodium on performance deficits produced by the common cold. J : 251-254 Psychopharmac 5 Smith A P, Tyrrell D A J, Barrow G I, Higgins P G, Willman J S, Bull S, Coyle K B, Trickett S (1992) Mood and experimentally-induced respiratory virus infections and illnesses. Psychol Hlth 6: 205212
Smith A P, Thomas M, Brockman P, Kent J, Nicholson KG (1993a) Effects of influenza B virus infection on human performance. BMJ 306: 760-761 Smith A P, Thomas M, Brockman P (1993b) Noise, respiratory virus infections and performance. In Vallet M (ed.) Proceedings of the sixth international congress on noise as a public health problem. Actes Inrets 34, pp. 311-314 Smith A P, Brockman P, Flynn R, Maben A, Thomas M (1993c) An investigation of the effects of coffee on alertness and performance during the day and night. Neuropsychobiology 27: 217-233 Smith A P, Whitney H, Thomas M, Brockman P, Perry K (1995) A comparison of the acute effects of a low dose of alcohol on mood and performance of healthy volunteers and subjects with upper respiratory tract illnesses. J Psychopharmac 9: 225-230
Downloaded from jop.sagepub.com at CAMBRIDGE UNIV LIBRARY on October 23, 2015
common
cold
Smith, Marie Thomas, Kate Perry and Helen Whitney
Psychology Research Unit, Department of Psychology, University of Bristol, Bristol BS8 1TN, UK.
An experiment was carried out to determine whether caffeinated and decaffeinated coffee removed the malaise (reduced alertness, slower psychomotor performance) associated with having a common cold. One hundred volunteers were tested when healthy and 46 returned to the laboratory when they developed colds. Those subjects who remained healthy were then recalled as a control group. On the second visit subjects carried out two sessions, one pre-drink and another an hour after the drink. Subjects were randomly assigned to one of the following three conditions, caffeinated coffee (1.5 mg/kg caffeine/body weight), decaffeinated coffee or fruit juice. Subjects with colds reported decreased alertness and were slower at performing psychomotor tasks. Caffeine increased the alertness and performance of the colds subjects to the same level as the healthy group and decaffeinated coffee also led to an improvement. These results suggest that drugs which increase alertness can remove the malaise associated with the common cold, and that increased stimulation of the sensory afferent nerves
may also be
beneficial.
Key words: alertness; caffeine; coffee; psychomotor performance;
Introduction Minor illnesses such as the common cold are frequent, and a major cause of absence from work and education. In addition, there is now strong evidence showing that both experimentally-induced and naturally-occurring upper respiratory tract illnesses (URTIS) are associated with increased negative mood and impairments of mental functioning. These findings can be briefly summarized as follows (see Smith, 1990, 1992 for reviews of the earlier work).
widespread
Experimentally-induced URTIs Both colds and influenza have selective effects on performance, with only some functions, not all, being impaired. The profile of performance effects was different in studies of influenza to those observed in experiments on the effects of colds. Influenza illnesses were found to increase reaction times on tasks where there is uncertainty as to where or when the next target will appear (Smith et al., 1987a). Psychomotor tasks involving speed and accuracy of hand-eye co-ordination were not impaired during influenza, neither were cognitive tasks, such as logical reasoning. In contrast, cold-type viral infection impaired performance on psychomotor tasks such as a pursuittracking tasks, five-choice serial response task and pegboard task (see Smith et al., 1987a, b, 1988). Subjects with colds were not impaired on the detection tasks affected by influenza, and there was little evidence of subjects with colds showing changes in memory performance. The effects of the experimentallyinduced URTIs on mood also depended on the type of virus, with influenza leading to a general increase in negative affect and colds mainly reducing alertness (Smith et al., 1992). These effects may have reflected the severity of the illness rather than
upper
respiratory tract illness
the virus per se. The above results reflect differences between subjects who were symptomatic and those who were uninfected. Behavioural changes were also found in volunteers with subclinical infections and when subjects were tested during the incubation period of the illness and in convalescence. In other words, the performance changes were not dependent on the presence of symptoms (see Smith et al., 1988, 1989).
Naturally occurring URTIs The above findings have largely been replicated in studies of naturally-occurring illnesses (see Smith et al., 1993a, for details of the effects of naturally-occurring influenza and Hall and Smith, 1996 for effects of naturally-acquired colds). Other research has examined whether illness makes people more sensitive to the effects of other factors known to alter mood. Smith et al. (1993b) with colds were more sensitive to the effects of noise than were healthy volunteers. Main effects of both noise (on cognitive vigilance tasks) and colds (on psychomotor tasks) were found. However, there was also evidence to support the view that a combination of having a cold and being exposed to noise resulted in the worst performance (although this effect was restricted to a variable fore-period simple reaction-time task). Smith et al. (1995) investigated whether a low dose of alcohol ( 1.5 ml of vodka/kg body weight) produced different behavioural effects in healthy volunteers to those suffering from URTIS. Subjects with colds reported an increase in negative affect and were slower at performing the psychomotor tasks. Interactions between health status and alcohol conditions were also found. Alcohol improved the mood of healthy subjects but produced greater negative mood in those with colds. Similarly,
performance efficiency and investigated whether subjects
Downloaded from jop.sagepub.com at CAMBRIDGE UNIV LIBRARY on October 23, 2015
320
performance of selective and sustained attention tasks showed different effects of alcohol in healthy and ill subjects. Possible mechanisms underlying these effects The behavioural changes associated with URTIs have important practical implications and treatments for these illnesses should aim not only to reduce local and systemic symptoms but to remove the malaise. This is difficult as ,the mechanisms underlying the development of malaise are not clearly understood. Two approaches are outlined in this paper. The first involves increasing central arousal by administering caffeine. Previous research has shown that, even in low doses, caffeine can produce beneficial effects when alertness is low. This has been demonstrated by studying sleep deprivation (Lorist et al., 1994), nightwork (Smith et al., 1993c), and the post-lunch dip (Smith et al., 1990). The previous findings suggest that at least part of the malaise associated with a cold appears to involve reduced alertness. Because caffeine has been shown to increase alertness, it was predicted that caffeine should be able to return the cold-induced impairments to levels similar to those seen in healthy subjects. The second method considered here involved increasing sensory afferent stimulation by giving the subject a hot drink. Previous studies have shown that compounds which increase stimulation of the sensory afferent trigeminal nerve (e.g. zinc gluconate, Smith et al., 1991; nedocromil sodium, Barrow et al., 1990) can remove the malaise produced by experimentallyinduced URTIs. It was predicted that a hot drink would reduce the impairments associated with having a cold. The aims of the present study can, therefore, be summarized as follows. First, the study provided another opportunity to confirm that naturally occurring colds lead to decreased alertness and psychomotor impairments but have little effect on other cognitive functions (e.g. episodic, working or semantic memory). Secondly, the experiment examined whether increasing alertness would remove the impairments associated with having a cold, and the efficacy of two methods for doing this was compared.
Method Individuals from the Health Psychology Research Unit subject panel took part in the study. Prior to the experiment subjects
completed questionnaires measuring personality, psychosocial factors and health-related behaviours. This background information was important to ensure that effects attributed to experimental manipulations did not merely reflect differences in the above factors. Subjects were practised at the tasks and completed a baseline session (when all subjects were healthy). If subjects then developed an URTI they returned to the Unit and completed another session of tests, followed by the drink manipulation and another test session. Subjects were randomly assigned to drinks conditions. The caffeine manipulation was double-blind. The juice condition was included for two main reasons.
First, it was a cold drink rather than a hot drink. Second, it represented a condition where subjects were aware that no caffeine had been administered. Previous research with this design has shown little difference between the decaffeinated coffee and juice conditions (Smith et al.,
1993c). However, it was unknown whether this would apply subjects with colds were tested. When about half of the original panel had returned with a URTI the remaining subjects were recalled and retested (forming the healthy group). The entire study lasted about 10 weeks and was carried out in the months of January-March. when
Subjects subjects (47 females, 53 males, mean age 21 years, age range 18-30 years) took part in the study. Thirtyfive subjects were assigned to the caffeine condition (15 colds, 20 healthy), 32 to the decaffeinated condition (16 colds, 16 healthy) and 33 to the juice condition (16 colds and 17 healthy). Prior to participating in the study the volunteers were given both written and verbal information regarding the study and signed a consent form. One hundred
Nature of the drinks Subjects were given a 150-ml drink (representative of a typical mug of coffee). One condition consisted of orange juice, another decaffeinated coffee and the third decaffeinated coffee with 1.5 mg/kg body weight of caffeine tablets added (average amount of caffeine =97.5 mg). Subjects were free to add milk and sugar to the coffee according to their normal practice.
Schedule of testing
subjects completed a familiarization session which involved practise at the tests and completion of questionnaires measuring personality traits (the Eysenck Personality Inventory and Spielberger Trait Anxiety Inventory), healthrelated behaviours (alcohol consumption, smoking and diet) and stress (life events, perceived stress and negative moodAll
Cohen et al., 1993). Baseline sessions were completed in the morning (between 09.00 and 11.00 hours) and the second session was at the same time of day. Subjects had abstained from consuming caffeinated drinks for 2 h prior to testing. At the end of the second session the subjects consumed their drink. Subjects were then allowed a rest before the third session, which commenced 1 h after the end of the second session. These timings were selected on the basis of previous studies showing effects of caffeine 1 h after ingestion. see
Criteria for a cold Subjects were told to return to the unit if they developed a cold. They were told to wait at least 24 h after the start of the illness before returning but to ensure that they were retested within 4 days of the start of the illness. All subjects who reported URTIs completed a standard symptom check list and
categorized as having a cold if they had two or more respiratory tract symptoms (runny or blocked nose, sneezing, sore throat) and produced greater than 0.2 g of nasal secretion (subjects blew their nose and then waited for 15 min before using a pre-weighed tissue to blow their nose again. Weight of nasal secretion was determined by a further weighing). Sublingual temperature was recorded to determine whether the subjects had a fever. were
upper
Downloaded from jop.sagepub.com at CAMBRIDGE UNIV LIBRARY on October 23, 2015
321
Measurement of mood Mood ratings were recorded using bipolar visual analogue scales presented on the computer screen. Eighteen scales were presented (e.g. drowsy-alert; tense-calm) and these were based on the scales used by Herbert et al. (1976). Subjects completed the mood ratings before and after the performance tasks.
Performance tests Two main types of tests were examined here. Psychomotor tasks which had been shown to be sensitive to cold effects (see Introduction) were included, as were memory tasks which were not impaired by having a cold. All of these tasks, except for the pegboard task, were presented on an IBM compatible PC.
Psychomotor tasks Simple reaction-time task.
In this task a box was displayed on the screen and this was followed after a period of 1-8 sec by a square (the target) being presented in the middle of the box. The subject had to press a key as soon as the square was detected and, following this, another box was presented. This task lasted for 3 min.
Five-choice serial response task. Five boxes were displayed on the screen and a square appeared in one of the boxes. The subject pressed the corresponding key and the square then appeared in another box and the subject had to press the next key. This task lasted for 3 min and the number of responses made, number of errors and number of gaps (occasional long responses > 1500 msec) were recorded.
and decide whether the sentence was a true description of the order of the letters. If it was, the subject pressed the T key on the keyboard, if it wasn’t, they pressed the F key. The sentences ranged in syntactic complexity from simple active to passive negative (e.g. A is not followed by B). Subjects carried out the task for 3 min.
Semantic-processing task. This test, developed by Baddeley (1981), measures speed of retrieval of information from general knowledge. Subjects were shown a sentence and had to decide whether it was true (e.g. canaries have wings). The number completed in 3 min was recorded, as was the accuracy of doing the test.
Results Analyses of variance distinguishing colds/healthies and drinks conditions showed that subjects in the various groups did not differ in terms of demographics, personality, stress levels or health-related behaviours. The subjects with URTIs had greater nasal symptoms and signs than the healthy subjects but did not report systemic symptoms. Indeed, there was no evidence of an increased temperature in the URTI group, which suggests that the illnesses were colds rather than influenza. Effects of a cold (pre-drink) of the data from Session 2 allows one to examine having a cold produced changes in mood and performance. Analyses of covariance were carried out with the baseline data as covariates and pre-drink scores as the dependent variables. This statistical technique adjusts the dependent variables to take account of unwanted variation at baseline. The between-subject factor was colds vs healthy
Analysis whether
Pegboard test. This test involved transferring pegs from a full pocket solitaire set to an empty one. Pegs were transferred one at a time from the full set to the equivalent hole in the empty set. The subjects were instructed to use their dominant hand. For right-handed subjects the full set was on the left and the empty set on the right. The first peg to be moved was the one at the extreme top right, and the volunteer went down each column, ending with the peg in the extreme bottom left of the set (left-handed subjects had the full set on the right and proceeded from left to right rather than right to left). Memory tasks A list of 20 words was presented on the PC of one every 2 sec. At the end of the list, the subject had 2 min to write down (in any order) as many of the words as possible. Subjects were shown a different list at each test session.
Free-recall task.
subjects. Mood A global increase in negative affect was reported by subjects with colds (both pre- and post-performance mood, all p < 0.05). This effect is illustrated by the drowsy-alert (postperformance rating) in Fig. I [main effect of colds:
MSerror = 68.4].
F( 1,97) =13.4, p < 0.001,
screen at a rate
Psychomotor tasks Both the simple reaction-time task and five-choice serial response task were performed more slowly [simple reactiontime task: F(1,97) 7.35, p < 0.01, MSerror = 7018; five-choice task: F( 1,97) = 7.64, p < 0.01, MSerror = 326] when the subject had a cold (see Fig. 1 ). There was no significant effect of having a cold on the pegboard task, which confirms the negative finding of Hall and Smith (1996). =
Delayed-recognition memory task. At the end of the test session, subjects were shown a list of 40 words which consisted of the 20 words shown at the start of the session, plus 20 distracters. The subjects had to decide as quickly as possible whether each word had been shown in the original list or not. Both speed and accuracy of responding were recorded.
Memory tasks There
was no
effect of
Logical-reasoning task. This test was developed by Baddeley (1968), and the subjects were shown statements about the
memory tasks (all p > In summary, the
order of the letters A and B followed by the letters AB or BA (e.g. A follows B: BA). The subjects had to read the statement
confirmed that
psychomotor
having 0.05).
a
cold
on
performance of
the
results from the pre-drink session having a cold reduces alertness, leads to slowing but has no effect on memory.
Downloaded from jop.sagepub.com at CAMBRIDGE UNIV LIBRARY on October 23, 2015
322
1 Effects of the common cold (pre-drink). (a) Alertness, scores are the adjusted mean ratings (range = to 50), high scores indicate greater alertness. (b) Simple reaction-time task, scores are the adjusted mean reaction time (msec). (c) Five-choice serial response task, scores are the adjusted mean number of responses in 3 min
Figure
Post-drink ef fects In these analyses baseline measurements were still used covariates and post-drink scores as the dependent variable. Main effects of colds. All of the effects found pre-drink still significant in the post-drink session. No new effects
were were
significant. Main effects of drink. The only main effect of drink conditions was found in the analysis of the simple reactiontime data [F(2,93) = 4.89, p < 0.01], with those given caffeine having faster reaction times than the other groups.
Colds x drinks interactions. Interactions between colds and drinks were found for the post-performance alertness ratings
[e.g. drowsy-alert: F(2,93)=4.40, p < 0.05, MSerror = 72.0], simple reaction-time task [F(2,93) 4.24, p < 0.05, MSerror=8686] and five-choice serial response speed [F(2,93)=5.35, p < 0.05, MSerror = 468]. These data are =
shown in Table l. In these analyses those with colds given caffeine returned to a similar level to that of the healthy subjects, and there
Table 1
Effects of different drinks
on
alertness and psychomotor
indication of beneficial effects (especially on those with colds of even the decaffeinated coffee. These effects are shown in Fig. 2. Tukey tests showed that the juice/colds group were significantly impaired compared to the juice no/cold group, whereas the colds and healthy groups were not significantly different in the coffee conditions. No interactions between colds/drinks were significant in the other tasks. was
as
also
some
speed) for
Discussion The results obtained in this study confirm that minor illnesses, such as the common cold, can alter mood and psychomotor performance. The major findings, however, are those showing interactions between drinks and cold status. Caffeinated coffee largely eliminated the effects of having a cold. This confirms results from studies of other low arousal states which show that even a small dose of caffeine can be beneficial when alertness is reduced. Caffeine produces its behavioural effects in a number of ways. The primary action of caffeine is competitive antagonism
performance of subjects with colds and healthy subjects
-~---
Scores
are
the
adjusted
means
from the
analyses of covariance. Downloaded from jop.sagepub.com at CAMBRIDGE UNIV LIBRARY on October 23, 2015
&dquo;
323
2 Differences (post-drink) between subjects with colds and healthy subjects in the different drink conditions. (a) Alertness, scores are the differences between the means of the cold and healthy groups (healthy-colds), high scores indicated greater cold-induced reduction in alertness. (b) Simple reaction-time task, scores are the difference between the mean reaction times of the colds and healthy groups (colds-healthy), (msec), high scores indicate greater cold-induced slowing of reaction time. (c) Five-choice serial response task, scores are the difference between the mean number done for the colds and healthy groups (healthy-colds), high scores indicate greater cold-induced slowing of responses
Figure
of adenosine receptors. However, there is also evidence that the noradrenergic and dopaminergic neurotransmitter systems can be altered by caffeine. Indeed, recent research provides a plausible mechanism which could account for the effects of caffeine in low-alertness states. Smith and Nutt (1996) have shown that clonidine, a drug which reduces the turnover of central noradrenaline, can induce a low-arousal state. This can be removed either by factors which increase alertness (e.g. white noise) or by ingestion of idazoxan, a drug which increases central noradrenaline. There are a number of pieces of evidence which suggests that caffeine increases central noradrenaline (see Fredholm, 1993). First, caffeine influences adenosine, which in turn inhibits noradrenaline release. Second, animal research has shown increased turnover of noradrenaline and increased locus coeruleus timing following caffeine. Finally, there is evidence of down-regulation of receptors following chronic administration of caffeine. It is plausible, therefore, that the effect of caffeine seen here reflects changes in the noradrenergic neurotransmitter system. If this is the case then compounds such as idazoxan should produce similar benefits. The noradrenaline model can also account for the larger effects of alcohol and noise in subjects with colds, in that both of these factors influence noradrenaline. It is, of course, possible that other neurotransmitter systems are involved as well (e.g. dopamine) and further research will be required to determine if this is the case. The present result also has important implications for therapy. Certain medications already contain caffeine and it will be of interest to investigate whether caffeine presented in this form and dose also has beneficial effects. Another result found here also has important implications for treatment of the illness. Decaffeinated coffee produced improvements in those with colds which suggests that hot drinks are beneficial in this type of illness. This could reflect the local relief produced by the drink or could be due to increased stimulation
of the sensory nerves. If the former is true then other forms of medication (e.g. decongestants) may also produce similar effects. On the other hand, if increased trigeminal stimulation is important then compounds such as menthol should also reduce the malaise associated with the common cold. However, it is clear that the improvements associated with a hot drink are much smaller than when the hot drink includes caffeine. It is also possible that the effects observed in the decaffeinated coffee condition reflected possible expectancy effects associated with the double-blind procedure. Again, further research is needed to resolve this issue. The present results show that the common cold leads to reduced alertness and slower psychomotor speed. These impairments can be removed by caffeinated coffee and, to some extent, by decaffeinated coffee. These results have important practical implications and suggest further studies to determine the underlying mechanisms. The findings also suggest that medication should not only provide symptomatic relief but reduce the malaise associated with the cold. Addition of caffeine may be one way to achieve this, and there is clearly some support for the practice of taking hot drinks when suffering from colds. Further studies must not only consider the topic of malaise induced by colds but also examine whether similar effects are apparent in other diseases. The results of such studies will not only be relevant to the treatment of malaise but also to our understanding of the mechanisms underlying effects of reduced arousal on behaviour.
Acknowledgement The research described in this paper was supported by a grant from the Institute for Scientific Information on Coffee.
Downloaded from jop.sagepub.com at CAMBRIDGE UNIV LIBRARY on October 23, 2015
324
Smith A P, Nutt D J (1996) Noradrenaline and attention lapses. Nature 380: 291 Smith A P, Tyrrell D A J, Coyle K B, Willman J S (1987a) Selective effects of minor illnesses on human performance. Br J Psychol :183-188 78 Smith A P, Tyrrell DAJ, Al-Nakib W, Coyle KB, Donovan C B, Higgins PG, Willman JS (1987b) Effects of experimentally induced respiratory virus infections and illness on psychomotor
Address for correspondence Professor Andrew Smith
Department of Psychology University of Bristol 8 Woodland Road
Bristol BS81TN UK
Email: [email protected]
:144-148 performance. Neuropsychobiology 18 Tyrrell DAJ, Al-Nakib W, Coyle KB, Donovan C B, Higgins PG, Willman JS (1988) The effects of experimentallyinduced respiratory virus infections on performance. Psychol
Smith A P,
References reasoning test based on gramBaddeleyAD matical transformation. Psychonom Sci 10: 341-342 Baddeley A D (1981) The cognitive psychology of everyday life. Br J : 257-269 Psychol 72 Barrow G I, Higgins P G, Al-Nakib W, Smith A P, Wenham R B M, Tyrrell DAJ (1990) The effect of intranasal nedocromil sodium on viral upper respiratory tract infections in human volunteers. Clin Exp Allergy 20: 45-51 Cohen S, Tyrrell D A J, Smith AP (1993) Negative life events, perceived stress, negative affect and susceptibility to the :131-140 common cold. J Pers Soc Psychol 64 Fredholm B (1993) Central pathways affected by xanthines. In: Proceedings of seventh international caffeine workshop. International Life Sciences Institute, Washington DC Hall S R, Smith A P (1996) An investigation of the effects and aftereffects of naturally- occurring upper respiratory tract illnesses on mood and performance. Physiol Behav 59: 569-577 Herbert M, Johns M W, Dore C (1976) Factor analysis of analogue scales measuring subjective feelings before sleep and after sleep. Br J Med Psychol 49: 373-379 Lorist M, Snel J, Kok A (1994) Influence of caffeine on information processing stages in well rested and fatigued subjects. Psycho: 411-421 pharmacology 113 Smith A P (1990) Respiratory virus infections and performance. (1968) A three-minute
Phil Trans R Soc Lond B327: 519-528 Smith A P (1992) Colds, influenza and performance. In Smith A P, Jones D M (eds), Handbook of human performance, vol 2. Academic Press, London, pp. 197-218
Med 18: 65-71 Smith A P, Tyrrell DAJ, Al-Nakib W, Barrow GI, Higgins PG, Leekam S, Trickett S (1989) Effects and after-effects of the common cold and influenza on human performance. Neuropsy: 90-93 chobiology 21 Smith A P, Rusted J M, Eaton-Williams P, Savory M, Leathwood P (1990) Effects of caffeine given before and after lunch on :160-163 sustained attention. Neuropsychobiology 23 Smith A P, Tyrrell DAJ, Al-Nakib W, Barrow GI, Higgins P G, Wenham R (1991) The effects of zinc gluconate and nedocromil sodium on performance deficits produced by the common cold. J : 251-254 Psychopharmac 5 Smith A P, Tyrrell D A J, Barrow G I, Higgins P G, Willman J S, Bull S, Coyle K B, Trickett S (1992) Mood and experimentally-induced respiratory virus infections and illnesses. Psychol Hlth 6: 205212
Smith A P, Thomas M, Brockman P, Kent J, Nicholson KG (1993a) Effects of influenza B virus infection on human performance. BMJ 306: 760-761 Smith A P, Thomas M, Brockman P (1993b) Noise, respiratory virus infections and performance. In Vallet M (ed.) Proceedings of the sixth international congress on noise as a public health problem. Actes Inrets 34, pp. 311-314 Smith A P, Brockman P, Flynn R, Maben A, Thomas M (1993c) An investigation of the effects of coffee on alertness and performance during the day and night. Neuropsychobiology 27: 217-233 Smith A P, Whitney H, Thomas M, Brockman P, Perry K (1995) A comparison of the acute effects of a low dose of alcohol on mood and performance of healthy volunteers and subjects with upper respiratory tract illnesses. J Psychopharmac 9: 225-230
Downloaded from jop.sagepub.com at CAMBRIDGE UNIV LIBRARY on October 23, 2015
