Eur J Appl Physiol DOI 10.1007/s00421-014-2973-z
Original Article
Combined caffeine and carbohydrate ingestion: effects on nocturnal sleep and exercise performance in athletes Ben Miller · Helen O’Connor · Rhonda Orr · Patricia Ruell · Hoi Lun Cheng · Chin Moi Chow
Received: 19 April 2014 / Accepted: 29 July 2014 © Springer-Verlag Berlin Heidelberg 2014
Abstract Purpose In athletes, caffeine use is common although its effects on sleep have not been widely studied. This randomised, double-blind, placebo-controlled crossover trial investigated the effects of late-afternoon caffeine and carbohydrate-electrolyte (CEB) co-ingestion on cycling performance and nocturnal sleep. Methods Six male cyclists/triathletes (age 27.5 ± 6.9 years) completed an afternoon training session (TS; cycling 80 min; 65 % VO2max) followed by a 5 kJ kg−1 cycling time trial (TT). Caffeine (split dose 2 × 3 mg kg−1) or placebo was administered 1 h prior and 40 min into the TS. A 7.4 % CEB (3 ml kg−1 every 15 min) was administered during the TS, followed 30 min after by a standardised evening meal. Participants retired at their usual bedtime and indices of sleep duration and quality were monitored via polysomnography. Data: mean ± SD. Results All participants performed better in the caffeine TT (caffeine 19.7 ± 3.3; placebo 20.5 ± 3.5 min; p = 0.006), while ratings of perceived exertion (caffeine 12.0 ± 0.6; placebo 12.9 ± 0.7; p = 0.004) and heart rate (caffeine 175 ± 6; placebo 167 ± 11 bpm; p = 0.085) were lower in the caffeine TS. Caffeine intake induced significant
Communicated by Michael Lindinger. B. Miller · H. O’Connor · R. Orr · P. Ruell · H. L. Cheng · C. M. Chow (*) Exercise, Health and Performance, Faculty of Health Sciences, The University of Sydney, P.O. Box 170, Lidcombe, NSW 1825, Australia e-mail: chin‑[email protected] C. M. Chow Delta Sleep Research Unit, Faculty of Health Sciences, The University of Sydney, Lidcombe, NSW, Australia
disruptions to a number of sleep indices including increased sleep onset latency (caffeine 51.1 ± 34.7; placebo 10.2 ± 4.2 min; p = 0.028) and decreased sleep efficiency (caffeine 76.1 ± 19.6; placebo 91.5 ± 4.2 %; p = 0.028), rapid eye movement sleep (caffeine 62.1 ± 19.6; placebo 85.8 ± 24.7 min; p = 0.028) and total sleep time (caffeine 391 ± 97; placebo 464 ± 49 min; p = 0.028). Conclusions This study supports a performance-enhancing effect of caffeine, although athletes (especially those using caffeine for late-afternoon/evening training and competition) should consider its deleterious effects on sleep. Keywords Caffeine · Performance-enhancing substances · Athletic performance · Physical endurance · Sleep · Polysomnography Abbreviations AI Arousal index ANOVA Analysis of variance BMI Body mass index CEB Carbohydrate-electrolyte beverage CI Confidence interval E Epinephrine EEG Electroencephalogram HPLC High-performance liquid chromatography NE Norepinephrine NREM Non-rapid eye movement PSG Polysomnography REM Rapid eye movement ROL Rapid eye movement sleep onset latency RPE Ratings of perceived exertion SD Standard deviation SE Sleep efficiency SOL Sleep onset latency SWS Slow wave sleep
13
TST Total sleep time VO2max Maximal oxygen uptake WASO Wake after sleep onset
Introduction Caffeine is a commonly consumed psychoactive substance found in a diverse range of foods and beverages, some medications and dietary supplements (Burke 2008). In elite athletes, caffeine is commonly consumed to enhance exercise performance, and this practice is particularly prevalent in sports such as cycling and triathlon (Burke 2008). Since the removal of caffeine from the World Anti-Doping Agency Prohibited Substances List in 2004, evidence suggests that its use in elite sporting competition has increased (Chester and Wojek 2008). However, this observation has not been consistent across all studies (Del Coso et al. 2011). Research investigating the ergogenic potential of caffeine is extensive, with studies in endurance (O’Rourke et al. 2008; Paluska 2003), intermittent and team sports supporting worthwhile performance-enhancing effects (Astorino and Roberson 2010). The influence of combined caffeine and carbohydrate-electrolyte beverage (CEB) consumption has also been explored, as CEBs provide fuel and sensory-related performance benefits for exercise bouts lasting as little as 60 min (Jeukendrup et al. 1997; Carter et al. 2004). Research demonstrates a beneficial effect of combined caffeine and CEB intake on improving endurance performance (Conger et al. 2011). Despite its ergogenic potential, caffeine consumption is associated with a range of adverse effects such as heart palpitations, locomotor agitation and sleep disruptions (Paluska 2003; Landolt et al. 2004). In the general population (including those who habitually consume one to three caffeinated beverages in the morning), low-dose caffeine use prior to bedtime can lead to greater difficulty in falling asleep (long sleep onset latency, SOL), reduced total sleep time (TST), decreased proportion of time spent asleep while in bed (sleep efficiency, SE) and reduced deep sleep (slow wave sleep, SWS) (Landolt et al. 1995). Not only are these adverse effects socially significant given the widespread use of caffeine, they are also important in the competitive sport setting as adequate sleep is critical for performance, post-exercise recovery and adaptation to training (Samuels 2008; Halson 2008). Moreover, factors such as strenuous exercise and performance anxiety may further exacerbate poor sleep in athletes (Driver and Taylor 2000; Savis et al. 1997). The sleep disruptive effects of caffeine are especially relevant to those who use it to compete over sequential days (e.g. many team sports, tennis, swimming and athletics finals) or for evening training/competition as its stimulating effect may continue well into the night.
13
Eur J Appl Physiol
Under non-exercise settings, caffeine has been shown to have a half-life of 4–6 h (Graham 2001). However, its pharmacokinetics during exercise has not been widely studied. While exercise is expected to prolong caffeine metabolism due to redirection of hepatic blood flow (where it is metabolised by cytochrome P450) to the muscles, one study observed a decrease in caffeine half-life to 2.3 h with moderate intensity cycling (Collomp et al. 1991). With limited research on the possible changes to caffeine half-life during exercise and the potential of this to influence sleep in the post-exercise period, it is surprising that the deleterious effects of caffeine on sleep have not been widely studied in sport. To address the lack of research on the impact of caffeine ingestion on sleep in athletes, the present pilot study investigated the combined effect of caffeine and CEB intake [as would be typical during a training session (TS) or endurance competition] in the late afternoon on nocturnal sleep. The influence of caffeine and CEB co-ingestion on endurance exercise performance was measured as a secondary outcome. It was hypothesised that caffeine and CEB coingestion would enhance performance during a late-afternoon exercise session, but elicit significant disruptions to indices of sleep quality and duration on the same night.
Methods Participants Well-trained male cyclists or triathletes (cycling ≥250 km week−1) were recruited. Heavy caffeine users (>300 mg day−1), those reporting a history of irregular sleep–wake cycles, sleep disorders or frequent use of sleep medication were excluded. Smokers and regular alcohol consumers (>60 g week−1) were also excluded due to possible influences on caffeine sensitivity (Nehlig et al. 1992). Written informed consent was obtained from all participants. The study was approved by the Human Research Ethics Committee of the University of Sydney and was performed in accordance with the ethical standards laid down in the 1964 Declaration of Helsinki. Study design and experimental procedures This was a randomised, double-blind, placebo-controlled crossover trial comparing combined caffeine and CEB versus combined placebo and CEB ingestion on exercise performance and subsequent nocturnal sleep. Participants visited the research laboratory on three occasions. Prior to the initial visit, they were required to keep a food/beverage and sleep diary for 48 h, and refrain from strenuous exercise, caffeine and alcohol consumption
Eur J Appl Physiol Fig. 1 A schematic diagram of: a the experimental day; b the laboratory training session and time trial. Times are given based on mean times of all participants. C caffeine ingestion, CEB carbohydrate-electrolyte beverage, RPE ratings of perceived exertion, HR heart rate, P placebo ingestion, PSG polysomnography, TS training session, TT time trial
for 24 h. At the initial visit, anthropometric characteristics and information on medical history and caffeine consumption were collected (Desbrow 2013). Participants then completed a late morning VO2max test on an electronically braked cycle ergometer (Lode Excalibur Sport, Groningen, The Netherlands), followed by a TS and time trial (TT) familiarisation later that day. In the evening, participants consumed a standardised meal (5,450 kJ; 61 % carbohydrate; 26 % protein; 18 % fat) and underwent an overnight sleep study (polysomnography, PSG). Stature and body mass (in cycling shorts) were recorded to the nearest 0.5 cm and 0.01 kg using a wall-mounted stadiometer (Seca, Hamburg, Germany) and an electric floor scale (Mettler-Toledo Ltd, Port Melbourne, Australia), respectively. The experimental phase commenced a week from the initial visit. Participants completed two experimental days that were separated by a minimum of two and a maximum of 14 days, during which they were instructed to maintain their habitual training and dietary practices. On the experimental days, participants consumed either caffeine or placebo in a randomised, counterbalanced order and completed a cycling TS followed by a performance TT and an overnight PSG. Placebo or caffeine at 6 mg kg−1 was administered as a split dose (Conway et al. 2003), to minimise detection of caffeine-related stimulatory/side effects and to preserve blinding. In competitive sports, similar caffeine doses of 200–400 mg day−1 are reported to be used frequently by athletes (Chester and Wojek 2008). The first dose (3 mg kg−1) was ingested 1 h before an afternoon TS, with the second dose (3 mg kg−1) taken 40 min into the TS (Fig. 1a). The effectiveness of the blinding was assessed at the end of each TT and on the morning after each experimental day by asking participants to report their perceived treatment status. Prior to each experimental
day, participants were prescribed a standardised carbohydrate diet (7 g kg−1 day−1) for consumption in the preceding 24 h. They were also required to refrain from strenuous exercise, caffeine and alcohol for 24 h and maintain a regular sleep–wake pattern (assessed by a sleep log). Participants also fasted for 4 h prior to commencement of each experimental visit (from ~1300 hours). Measurement of VO2max During the VO2max test at the initial visit, participants performed four submaximal workload stages at 100, 150, 200, 250 W for four min each, followed by a four min recovery at 75 W. The workload was then increased at a rate of 30 W min−1 until volitional fatigue. Oxygen consumption was measured using Douglas bags with expired air collected for a minimum of 30 s prior to exhaustion. Expired air was analysed for percent of oxygen and carbon dioxide using Servomex Pm1111E and Ir1507 sensors (Servomex, Crowborough, UK). Gas volume was measured with a dry gas metre (Harvard Apparatus Ltd, Edenbridge, UK) with oxygen calculated using indirect calorimetry equations. Participant VO2max was attained upon volitional fatigue and a plateau in VO2. They were also monitored for other VO2max indicators including achievement of a heart rate within 10 beats min−1 of age-predicted maximum (220 − age), and a respiratory exchange ratio >1. Linear regression equations generated from power output and oxygen consumption were used to determine individual workload for the experimental sessions. Training session and performance time trial The experimental visit involved an 80 min TS at 65 % of VO2max, followed by a 5 kJ kg−1 TT (approximately
13
20 min) on a cycle ergometer. The chosen workload was based on a previous study reporting coefficients of variation (duration 0.7–4.6 %; power 1.5–4.6 %) for similar cycling TTs (Hopkins et al. 2001). During the TS, a commercially available CEB (carbohydrate 7.4 g 100 ml−1; Powerade, Coca-Cola, Australia) was consumed every 15 min at a volume of 3 ml kg−1 (equivalent to ~1 g kg−1 h−1 carbohydrate). Heart rate was measured every five min using a heart rate monitor (Polar Electro Inc., Lake Success, USA). Ratings of perceived exertion (RPE) were measured at 15 min intervals using the Borg scale (Fig. 1b) (Borg 1982). Body mass was recorded immediately pre- and post-exercise to monitor sweat loss (adjusting for any faecal or urine output, but not metabolic water). The TS and TT were performed in the fed state and under thermoneutral conditions (20 °C, 40 % relative humidity) inside an environmental chamber. Participants were asked to report any adverse effects such as tremors, heart palpitations, anxiety or gastrointestinal discomfort. Polysomnographic (PSG) recording Full PSG with international 10–20 electrode placement (C3/A2, O2/A1, F3/A2), respiratory recordings (to exclude sleep breathing disorders) and leg electromyogram (to exclude limb movement disorders) were applied on the familiarisation night to preclude the “first night effect” resulting from attachment of electroencephalograph (EEG) electrodes to the scalp and body surface in an unfamiliar environment (Agnew et al. 1966; Lorenzo and Barbanoj 2002). Sleep was recorded using the Compumedics Sleep system (Compumedics Ltd, Abbotsford, Australia). On subsequent experimental nights, only sleep EEG, electrooculogram and electromyogram were recorded. Sleep was ad libitum: participants went to bed at their usual bedtime and woke spontaneously in the morning in synchrony with the circadian control of the sleep–wake cycle. Participants slept in a darkened, air-conditioned room at a constant temperature of 21 °C. Sleep studies were scored by an experienced sleep physiologist blinded to participant treatment status according to standard scoring criteria (Rechtschaffen and Kales 1968). Sleep recordings were evaluated for variables of TST (the total amount of sleep time during the sleep period); SE (the percentage of time spent asleep while in bed); SOL (the time from lights out to the first epoch of sleep); arousal index per hour (AI, defined by EEG desynchronisation of more than 3 s); wake after sleep onset (WASO, the total amount of time awake during the sleep period); non-rapid eye movement (NREM) sleep stages 1 and 2 (light sleep) and stage 3 (deep sleep or SWS); rapid eye movement (REM) sleep and REM sleep onset latency (ROL, the time from sleep onset to the first REM sleep epoch).
13
Eur J Appl Physiol
Caffeine preparation and biochemical analysis Caffeine and placebo (lactose) capsules were pre-prepared from 100 % pure powder and inserted into gelatine capsules (Bronson & Jacobs Pty Ltd, Villawood, Australia) by a pharmacist (RO) blinded to participant treatment status. Blood (4 ml) was collected from an indwelling cannula inserted into the antecubital vein for analysis of plasma caffeine concentration immediately before exercise, mid-way through the TS (immediately prior to the 2nd treatment dose), immediately post-TT, and 30 min prior to bedtime for both the caffeine and placebo trials. All samples were analysed for plasma caffeine from the caffeine trial and only the pre-exercise sample from the placebo trial was analysed. Caffeine was analysed by high-performance liquid chromatography (HPLC) as described previously (Holland et al. 1998). Briefly, 150 µl of 0.8 M perchloric acid was added to 150 µl of serum from each sample. After vortex mixing, proteins were removed by centrifugation at 4,000g for 10 min at 4 °C. A 125 µl aliquot of supernatant was neutralised with 9.8 µl of 4 M sodium hydroxide. Samples were centrifuged for 5 min at 4,000g and then 100 µl was injected onto a 15 cm Nova-Pak column (Waters Corporation, Milford, USA), with all peaks eluted by 20 min. The mobile phase was KH2PO4 (1.735 g) and 150 ml methanol per litre, the flow rate was 1 ml min−1 and peaks were detected at 274 nm. The Shimadzu HPLC system consisted of two LC-10ADVP pumps, a SPD10AVP UV–Vis detector and Class VP software for data acquisition and analysis. Standards ranging from 0 to 5 µg ml−1 were included in every assay, with unknowns analysed in duplicate. Peak area was used for determining the concentration of the unknown, and corrected for the dilution factor. Statistics Statistical analysis was performed using SPSS 17.0 (IBM Corporation, Somers, NY, USA). Data were checked for normality and equal variance between the trials. Differences between the caffeine versus placebo treatments were assessed using paired t tests for all physiological (except HR), performance and sleep variables that did not violate normality and equal variance assumptions, with the Wilcoxon signed-rank tests used for the non-normal variables. Between-treatment differences in RPE and HR over time were analysed using two-factor (treatment and time) repeated measures ANOVA. Descriptive data for RPE and HR are presented as a mean ± SD of all time points measured. Plasma caffeine was analysed at three time points (pre-exercise, post-exercise and bedtime)
Eur J Appl Physiol Table 1 Sleep indices after caffeine or placebo ingestion
Sleep variable
Caffeine
Placebo
Z valuea
P valueb
Mean ± SD
SOL (min) Stage 1 (min) Stage 1 (%) Stage 2 (min) Stage 2 (%) Stage 3/SWS (min) Stage 3/SWS (%) REM sleep (min) REM sleep (%) REM latency (min) Total sleep time (min) WASO (min) Arousal index (per min) Number of awakenings
51.1 ± 34.7 27.8 ± 12.6 7.0 ± 2.2 250 ± 68 63.4 ± 4.7 52.8 ± 19.3 13.9 ± 4.8 62.1 ± 19.6 15.7 ± 2.4 164.3 ± 112.3 391.4 ± 96.9 75.1 ± 86.6 12.9 ± 3.2 31.5 ± 12.3
10.2 ± 4.6 18.5 ± 7.6 4.2 ± 2.2 289 ± 37 62.0 ± 2.5 75.5 ± 9.1 15.6 ± 3.2 85.8 ± 24.7 18.2 ± 3.7 106.0 ± 44.0 464.3 ± 48.9 31.9 ± 17.0 11.8 ± 4.5 28.3 ± 14.6
0.028 0.080 0.116 0.075 0.461 0.225 0.463 0.028 0.116 0.116 0.028 0.046 0.600 0.893
76.1 ± 19.6
91.5 ± 4.2
−2.201 −1.753 −1.572 −1.782 −0.738 −1.214 −0.734 −2.201 −1.572 −1.572 −2.201 −1.992 −0.524 −0.135
REM rapid eye movement, SOL sleep onset latency, SWS slow wave sleep, WASO wake after sleep onset, % percent of total sleep episode a
Z values based on negative ranks (Wilcoxon signed-rank test) b
P values are two-tailed values
Sleep efficiency (%)
using repeated measures ANOVA and Bonferroni post hoc tests. Data are presented as mean ± SD with significance set at p 300 mg day−1. Ten participants were subsequently recruited but four discontinued after the familiarisation session secondary to work constraints and injury unrelated to this study. The remaining six participants had a mean age of 27.5 ± 6.9 years (range 18–40 years), BMI of 23.1 ± 0.3 kg m−2, and VO2max of 63.3 ± 2.7 ml kg−1 min−1. Self-reported caffeine intake indicated all but one participant (who reported a habitual intake of approximately 300 mg day−1) had habitual caffeine intakes of
Original Article
Combined caffeine and carbohydrate ingestion: effects on nocturnal sleep and exercise performance in athletes Ben Miller · Helen O’Connor · Rhonda Orr · Patricia Ruell · Hoi Lun Cheng · Chin Moi Chow
Received: 19 April 2014 / Accepted: 29 July 2014 © Springer-Verlag Berlin Heidelberg 2014
Abstract Purpose In athletes, caffeine use is common although its effects on sleep have not been widely studied. This randomised, double-blind, placebo-controlled crossover trial investigated the effects of late-afternoon caffeine and carbohydrate-electrolyte (CEB) co-ingestion on cycling performance and nocturnal sleep. Methods Six male cyclists/triathletes (age 27.5 ± 6.9 years) completed an afternoon training session (TS; cycling 80 min; 65 % VO2max) followed by a 5 kJ kg−1 cycling time trial (TT). Caffeine (split dose 2 × 3 mg kg−1) or placebo was administered 1 h prior and 40 min into the TS. A 7.4 % CEB (3 ml kg−1 every 15 min) was administered during the TS, followed 30 min after by a standardised evening meal. Participants retired at their usual bedtime and indices of sleep duration and quality were monitored via polysomnography. Data: mean ± SD. Results All participants performed better in the caffeine TT (caffeine 19.7 ± 3.3; placebo 20.5 ± 3.5 min; p = 0.006), while ratings of perceived exertion (caffeine 12.0 ± 0.6; placebo 12.9 ± 0.7; p = 0.004) and heart rate (caffeine 175 ± 6; placebo 167 ± 11 bpm; p = 0.085) were lower in the caffeine TS. Caffeine intake induced significant
Communicated by Michael Lindinger. B. Miller · H. O’Connor · R. Orr · P. Ruell · H. L. Cheng · C. M. Chow (*) Exercise, Health and Performance, Faculty of Health Sciences, The University of Sydney, P.O. Box 170, Lidcombe, NSW 1825, Australia e-mail: chin‑[email protected] C. M. Chow Delta Sleep Research Unit, Faculty of Health Sciences, The University of Sydney, Lidcombe, NSW, Australia
disruptions to a number of sleep indices including increased sleep onset latency (caffeine 51.1 ± 34.7; placebo 10.2 ± 4.2 min; p = 0.028) and decreased sleep efficiency (caffeine 76.1 ± 19.6; placebo 91.5 ± 4.2 %; p = 0.028), rapid eye movement sleep (caffeine 62.1 ± 19.6; placebo 85.8 ± 24.7 min; p = 0.028) and total sleep time (caffeine 391 ± 97; placebo 464 ± 49 min; p = 0.028). Conclusions This study supports a performance-enhancing effect of caffeine, although athletes (especially those using caffeine for late-afternoon/evening training and competition) should consider its deleterious effects on sleep. Keywords Caffeine · Performance-enhancing substances · Athletic performance · Physical endurance · Sleep · Polysomnography Abbreviations AI Arousal index ANOVA Analysis of variance BMI Body mass index CEB Carbohydrate-electrolyte beverage CI Confidence interval E Epinephrine EEG Electroencephalogram HPLC High-performance liquid chromatography NE Norepinephrine NREM Non-rapid eye movement PSG Polysomnography REM Rapid eye movement ROL Rapid eye movement sleep onset latency RPE Ratings of perceived exertion SD Standard deviation SE Sleep efficiency SOL Sleep onset latency SWS Slow wave sleep
13
TST Total sleep time VO2max Maximal oxygen uptake WASO Wake after sleep onset
Introduction Caffeine is a commonly consumed psychoactive substance found in a diverse range of foods and beverages, some medications and dietary supplements (Burke 2008). In elite athletes, caffeine is commonly consumed to enhance exercise performance, and this practice is particularly prevalent in sports such as cycling and triathlon (Burke 2008). Since the removal of caffeine from the World Anti-Doping Agency Prohibited Substances List in 2004, evidence suggests that its use in elite sporting competition has increased (Chester and Wojek 2008). However, this observation has not been consistent across all studies (Del Coso et al. 2011). Research investigating the ergogenic potential of caffeine is extensive, with studies in endurance (O’Rourke et al. 2008; Paluska 2003), intermittent and team sports supporting worthwhile performance-enhancing effects (Astorino and Roberson 2010). The influence of combined caffeine and carbohydrate-electrolyte beverage (CEB) consumption has also been explored, as CEBs provide fuel and sensory-related performance benefits for exercise bouts lasting as little as 60 min (Jeukendrup et al. 1997; Carter et al. 2004). Research demonstrates a beneficial effect of combined caffeine and CEB intake on improving endurance performance (Conger et al. 2011). Despite its ergogenic potential, caffeine consumption is associated with a range of adverse effects such as heart palpitations, locomotor agitation and sleep disruptions (Paluska 2003; Landolt et al. 2004). In the general population (including those who habitually consume one to three caffeinated beverages in the morning), low-dose caffeine use prior to bedtime can lead to greater difficulty in falling asleep (long sleep onset latency, SOL), reduced total sleep time (TST), decreased proportion of time spent asleep while in bed (sleep efficiency, SE) and reduced deep sleep (slow wave sleep, SWS) (Landolt et al. 1995). Not only are these adverse effects socially significant given the widespread use of caffeine, they are also important in the competitive sport setting as adequate sleep is critical for performance, post-exercise recovery and adaptation to training (Samuels 2008; Halson 2008). Moreover, factors such as strenuous exercise and performance anxiety may further exacerbate poor sleep in athletes (Driver and Taylor 2000; Savis et al. 1997). The sleep disruptive effects of caffeine are especially relevant to those who use it to compete over sequential days (e.g. many team sports, tennis, swimming and athletics finals) or for evening training/competition as its stimulating effect may continue well into the night.
13
Eur J Appl Physiol
Under non-exercise settings, caffeine has been shown to have a half-life of 4–6 h (Graham 2001). However, its pharmacokinetics during exercise has not been widely studied. While exercise is expected to prolong caffeine metabolism due to redirection of hepatic blood flow (where it is metabolised by cytochrome P450) to the muscles, one study observed a decrease in caffeine half-life to 2.3 h with moderate intensity cycling (Collomp et al. 1991). With limited research on the possible changes to caffeine half-life during exercise and the potential of this to influence sleep in the post-exercise period, it is surprising that the deleterious effects of caffeine on sleep have not been widely studied in sport. To address the lack of research on the impact of caffeine ingestion on sleep in athletes, the present pilot study investigated the combined effect of caffeine and CEB intake [as would be typical during a training session (TS) or endurance competition] in the late afternoon on nocturnal sleep. The influence of caffeine and CEB co-ingestion on endurance exercise performance was measured as a secondary outcome. It was hypothesised that caffeine and CEB coingestion would enhance performance during a late-afternoon exercise session, but elicit significant disruptions to indices of sleep quality and duration on the same night.
Methods Participants Well-trained male cyclists or triathletes (cycling ≥250 km week−1) were recruited. Heavy caffeine users (>300 mg day−1), those reporting a history of irregular sleep–wake cycles, sleep disorders or frequent use of sleep medication were excluded. Smokers and regular alcohol consumers (>60 g week−1) were also excluded due to possible influences on caffeine sensitivity (Nehlig et al. 1992). Written informed consent was obtained from all participants. The study was approved by the Human Research Ethics Committee of the University of Sydney and was performed in accordance with the ethical standards laid down in the 1964 Declaration of Helsinki. Study design and experimental procedures This was a randomised, double-blind, placebo-controlled crossover trial comparing combined caffeine and CEB versus combined placebo and CEB ingestion on exercise performance and subsequent nocturnal sleep. Participants visited the research laboratory on three occasions. Prior to the initial visit, they were required to keep a food/beverage and sleep diary for 48 h, and refrain from strenuous exercise, caffeine and alcohol consumption
Eur J Appl Physiol Fig. 1 A schematic diagram of: a the experimental day; b the laboratory training session and time trial. Times are given based on mean times of all participants. C caffeine ingestion, CEB carbohydrate-electrolyte beverage, RPE ratings of perceived exertion, HR heart rate, P placebo ingestion, PSG polysomnography, TS training session, TT time trial
for 24 h. At the initial visit, anthropometric characteristics and information on medical history and caffeine consumption were collected (Desbrow 2013). Participants then completed a late morning VO2max test on an electronically braked cycle ergometer (Lode Excalibur Sport, Groningen, The Netherlands), followed by a TS and time trial (TT) familiarisation later that day. In the evening, participants consumed a standardised meal (5,450 kJ; 61 % carbohydrate; 26 % protein; 18 % fat) and underwent an overnight sleep study (polysomnography, PSG). Stature and body mass (in cycling shorts) were recorded to the nearest 0.5 cm and 0.01 kg using a wall-mounted stadiometer (Seca, Hamburg, Germany) and an electric floor scale (Mettler-Toledo Ltd, Port Melbourne, Australia), respectively. The experimental phase commenced a week from the initial visit. Participants completed two experimental days that were separated by a minimum of two and a maximum of 14 days, during which they were instructed to maintain their habitual training and dietary practices. On the experimental days, participants consumed either caffeine or placebo in a randomised, counterbalanced order and completed a cycling TS followed by a performance TT and an overnight PSG. Placebo or caffeine at 6 mg kg−1 was administered as a split dose (Conway et al. 2003), to minimise detection of caffeine-related stimulatory/side effects and to preserve blinding. In competitive sports, similar caffeine doses of 200–400 mg day−1 are reported to be used frequently by athletes (Chester and Wojek 2008). The first dose (3 mg kg−1) was ingested 1 h before an afternoon TS, with the second dose (3 mg kg−1) taken 40 min into the TS (Fig. 1a). The effectiveness of the blinding was assessed at the end of each TT and on the morning after each experimental day by asking participants to report their perceived treatment status. Prior to each experimental
day, participants were prescribed a standardised carbohydrate diet (7 g kg−1 day−1) for consumption in the preceding 24 h. They were also required to refrain from strenuous exercise, caffeine and alcohol for 24 h and maintain a regular sleep–wake pattern (assessed by a sleep log). Participants also fasted for 4 h prior to commencement of each experimental visit (from ~1300 hours). Measurement of VO2max During the VO2max test at the initial visit, participants performed four submaximal workload stages at 100, 150, 200, 250 W for four min each, followed by a four min recovery at 75 W. The workload was then increased at a rate of 30 W min−1 until volitional fatigue. Oxygen consumption was measured using Douglas bags with expired air collected for a minimum of 30 s prior to exhaustion. Expired air was analysed for percent of oxygen and carbon dioxide using Servomex Pm1111E and Ir1507 sensors (Servomex, Crowborough, UK). Gas volume was measured with a dry gas metre (Harvard Apparatus Ltd, Edenbridge, UK) with oxygen calculated using indirect calorimetry equations. Participant VO2max was attained upon volitional fatigue and a plateau in VO2. They were also monitored for other VO2max indicators including achievement of a heart rate within 10 beats min−1 of age-predicted maximum (220 − age), and a respiratory exchange ratio >1. Linear regression equations generated from power output and oxygen consumption were used to determine individual workload for the experimental sessions. Training session and performance time trial The experimental visit involved an 80 min TS at 65 % of VO2max, followed by a 5 kJ kg−1 TT (approximately
13
20 min) on a cycle ergometer. The chosen workload was based on a previous study reporting coefficients of variation (duration 0.7–4.6 %; power 1.5–4.6 %) for similar cycling TTs (Hopkins et al. 2001). During the TS, a commercially available CEB (carbohydrate 7.4 g 100 ml−1; Powerade, Coca-Cola, Australia) was consumed every 15 min at a volume of 3 ml kg−1 (equivalent to ~1 g kg−1 h−1 carbohydrate). Heart rate was measured every five min using a heart rate monitor (Polar Electro Inc., Lake Success, USA). Ratings of perceived exertion (RPE) were measured at 15 min intervals using the Borg scale (Fig. 1b) (Borg 1982). Body mass was recorded immediately pre- and post-exercise to monitor sweat loss (adjusting for any faecal or urine output, but not metabolic water). The TS and TT were performed in the fed state and under thermoneutral conditions (20 °C, 40 % relative humidity) inside an environmental chamber. Participants were asked to report any adverse effects such as tremors, heart palpitations, anxiety or gastrointestinal discomfort. Polysomnographic (PSG) recording Full PSG with international 10–20 electrode placement (C3/A2, O2/A1, F3/A2), respiratory recordings (to exclude sleep breathing disorders) and leg electromyogram (to exclude limb movement disorders) were applied on the familiarisation night to preclude the “first night effect” resulting from attachment of electroencephalograph (EEG) electrodes to the scalp and body surface in an unfamiliar environment (Agnew et al. 1966; Lorenzo and Barbanoj 2002). Sleep was recorded using the Compumedics Sleep system (Compumedics Ltd, Abbotsford, Australia). On subsequent experimental nights, only sleep EEG, electrooculogram and electromyogram were recorded. Sleep was ad libitum: participants went to bed at their usual bedtime and woke spontaneously in the morning in synchrony with the circadian control of the sleep–wake cycle. Participants slept in a darkened, air-conditioned room at a constant temperature of 21 °C. Sleep studies were scored by an experienced sleep physiologist blinded to participant treatment status according to standard scoring criteria (Rechtschaffen and Kales 1968). Sleep recordings were evaluated for variables of TST (the total amount of sleep time during the sleep period); SE (the percentage of time spent asleep while in bed); SOL (the time from lights out to the first epoch of sleep); arousal index per hour (AI, defined by EEG desynchronisation of more than 3 s); wake after sleep onset (WASO, the total amount of time awake during the sleep period); non-rapid eye movement (NREM) sleep stages 1 and 2 (light sleep) and stage 3 (deep sleep or SWS); rapid eye movement (REM) sleep and REM sleep onset latency (ROL, the time from sleep onset to the first REM sleep epoch).
13
Eur J Appl Physiol
Caffeine preparation and biochemical analysis Caffeine and placebo (lactose) capsules were pre-prepared from 100 % pure powder and inserted into gelatine capsules (Bronson & Jacobs Pty Ltd, Villawood, Australia) by a pharmacist (RO) blinded to participant treatment status. Blood (4 ml) was collected from an indwelling cannula inserted into the antecubital vein for analysis of plasma caffeine concentration immediately before exercise, mid-way through the TS (immediately prior to the 2nd treatment dose), immediately post-TT, and 30 min prior to bedtime for both the caffeine and placebo trials. All samples were analysed for plasma caffeine from the caffeine trial and only the pre-exercise sample from the placebo trial was analysed. Caffeine was analysed by high-performance liquid chromatography (HPLC) as described previously (Holland et al. 1998). Briefly, 150 µl of 0.8 M perchloric acid was added to 150 µl of serum from each sample. After vortex mixing, proteins were removed by centrifugation at 4,000g for 10 min at 4 °C. A 125 µl aliquot of supernatant was neutralised with 9.8 µl of 4 M sodium hydroxide. Samples were centrifuged for 5 min at 4,000g and then 100 µl was injected onto a 15 cm Nova-Pak column (Waters Corporation, Milford, USA), with all peaks eluted by 20 min. The mobile phase was KH2PO4 (1.735 g) and 150 ml methanol per litre, the flow rate was 1 ml min−1 and peaks were detected at 274 nm. The Shimadzu HPLC system consisted of two LC-10ADVP pumps, a SPD10AVP UV–Vis detector and Class VP software for data acquisition and analysis. Standards ranging from 0 to 5 µg ml−1 were included in every assay, with unknowns analysed in duplicate. Peak area was used for determining the concentration of the unknown, and corrected for the dilution factor. Statistics Statistical analysis was performed using SPSS 17.0 (IBM Corporation, Somers, NY, USA). Data were checked for normality and equal variance between the trials. Differences between the caffeine versus placebo treatments were assessed using paired t tests for all physiological (except HR), performance and sleep variables that did not violate normality and equal variance assumptions, with the Wilcoxon signed-rank tests used for the non-normal variables. Between-treatment differences in RPE and HR over time were analysed using two-factor (treatment and time) repeated measures ANOVA. Descriptive data for RPE and HR are presented as a mean ± SD of all time points measured. Plasma caffeine was analysed at three time points (pre-exercise, post-exercise and bedtime)
Eur J Appl Physiol Table 1 Sleep indices after caffeine or placebo ingestion
Sleep variable
Caffeine
Placebo
Z valuea
P valueb
Mean ± SD
SOL (min) Stage 1 (min) Stage 1 (%) Stage 2 (min) Stage 2 (%) Stage 3/SWS (min) Stage 3/SWS (%) REM sleep (min) REM sleep (%) REM latency (min) Total sleep time (min) WASO (min) Arousal index (per min) Number of awakenings
51.1 ± 34.7 27.8 ± 12.6 7.0 ± 2.2 250 ± 68 63.4 ± 4.7 52.8 ± 19.3 13.9 ± 4.8 62.1 ± 19.6 15.7 ± 2.4 164.3 ± 112.3 391.4 ± 96.9 75.1 ± 86.6 12.9 ± 3.2 31.5 ± 12.3
10.2 ± 4.6 18.5 ± 7.6 4.2 ± 2.2 289 ± 37 62.0 ± 2.5 75.5 ± 9.1 15.6 ± 3.2 85.8 ± 24.7 18.2 ± 3.7 106.0 ± 44.0 464.3 ± 48.9 31.9 ± 17.0 11.8 ± 4.5 28.3 ± 14.6
0.028 0.080 0.116 0.075 0.461 0.225 0.463 0.028 0.116 0.116 0.028 0.046 0.600 0.893
76.1 ± 19.6
91.5 ± 4.2
−2.201 −1.753 −1.572 −1.782 −0.738 −1.214 −0.734 −2.201 −1.572 −1.572 −2.201 −1.992 −0.524 −0.135
REM rapid eye movement, SOL sleep onset latency, SWS slow wave sleep, WASO wake after sleep onset, % percent of total sleep episode a
Z values based on negative ranks (Wilcoxon signed-rank test) b
P values are two-tailed values
Sleep efficiency (%)
using repeated measures ANOVA and Bonferroni post hoc tests. Data are presented as mean ± SD with significance set at p 300 mg day−1. Ten participants were subsequently recruited but four discontinued after the familiarisation session secondary to work constraints and injury unrelated to this study. The remaining six participants had a mean age of 27.5 ± 6.9 years (range 18–40 years), BMI of 23.1 ± 0.3 kg m−2, and VO2max of 63.3 ± 2.7 ml kg−1 min−1. Self-reported caffeine intake indicated all but one participant (who reported a habitual intake of approximately 300 mg day−1) had habitual caffeine intakes of
