After several papers reported that Whole Body Cryotherapy (WBC) can improve objective and subjective markers of sleep, supported by anecdotal reports of post-exposure sleepiness from players at Southampton FC (SFC; PhD sponsor), the original aim of this thesis was to elucidate the effect of WBC on sleep in professional football players. However, after the UK COVID-19 lockdowns, WBC was not considered covid safe and, therefore, sleep became the central theme. Sleep plays an important role in the maintenance of both physiological and psychological homeostasis. During sleep, the release of human growth hormone and other anabolic hormones peak, inflammatory processes are modulated, and memories and skills are consolidated. Therefore, sleep is considered integral to athletic recovery and player well-being. Despite this, professional football players regularly present with sub-optimal sleep duration and/or quality. However, the factors associated with sleep variability are not fully understood, and there is no consensus on what the optimal level of sleep for athletes is. Therefore, this thesis conceptualised the following research questions: (1) What is known about the quality and duration of sleep amongst professional footballers? (2) What factors affect sleep in professional football players, specifically at SFC? (3) What are suitable and effective ways of improving sleep in professional football players? These questions were addressed across 2 systematic reviews (Chapters 2 & 4), an interventional study (Chapter 3), an observational cohort study (Chapter 5), a method agreement study (Chapter 6), and finally a case study (Chapter 7).
Chapter 3 presents a study that aimed to (1) investigate the effect of a WBC applied across an in-season microcycle on the objective and subjective sleep quality in under-18 (U18) professional footballers, and (2) determine the effect of WBC on game-day inflammation, testosterone, and cortisol. Unfortunately, this study was curtailed by the COVID lockdowns. Nevertheless, novel findings were reported. Specifically, whilst objective sleep data were not significantly different between groups, players who received WBC during the microcycle preceding a competitive fixture, reported a greater sense of alertness following wake, as determined by the Leeds Sleep Quality Index. Whilst these results are subjective, they could also be indicative of improved sleep architecture following WBC. However, considering objective sleep was determined from wrist-worn activity monitors without the capability to detect sleep stages, this cannot be known with certainty.
In Chapter 4, a scoping review of observational studies was performed that suggested that professional football players’ mean sleep duration, sleep latency, and wake after sleep onset (WASO), were all within recommended guidelines (these same reference limits were also used for Chapter 4). This conclusion was made on the basis that over 63% of the included studies reported means that were above the lower reference boundary for sleep duration. Despite this, several papers reported error bars that exceeded the reference limits, suggesting that suboptimal sleep remains common among individual players. In Chapter 5, an observational study was performed on under-18 professional SFC players, and the results matched what was observed from the scoping review in Chapter 4. Specifically, whilst sleep duration on matchday+1 (the day proceeding matchday) presented with a beta estimate (derived from linear mixed models) of 400mins, the remaining day types presented with sleep durations of above 420mins, the lower end of the reference limits. Nevertheless, in this study, confidence intervals breached the reference limits, therefore, further suggesting that suboptimal sleep occurs in this population. In tandem, results from Chapter 4 and Chapter 5 potentially indicate that group-level interventions are unnecessary. Rather, practitioners may find it more efficient to target support to players who report sleep disturbances.
The scoping review presented in Chapter 4 also suggested that professional football players' sleep was also more variable compared to age-matched controls and several factors (e.g. scheduling variables) were associated with disrupted sleep. Chapter 5 builds on these findings by demonstrating for the first time that scheduled start time (the time players were scheduled to arrive at training or for a fixture) was associated with the amount of sleep that U18 players attained. Specifically, for every hour increase in start time, player sleep duration increased by an estimated 19.1mins (CI:9.4–28.79; p<0.001). This occurred in tandem with an 18mins (CI:9.3–26.6; p<0.001) later wake time, per hour increase in scheduled start time. It is not clear to what magnitude start time would have to be extended to generate increases in player performance, secondary to increased sleep duration. However, considering the player's age from this study (age: 17.3 ± 0.7yrs), a later start time may befit their intrinsic chronotype and, therefore, support the players by reinforcing their natural sleep habits.
Whilst data from Chapter 5 support the notion that scheduling variables are associated with sleep in U18 professional footballers, they also suggest that sleep is not meaningfully associated with external workload. Global positioning and accelerometry data were collected and collated across 1-day, 7-day, and 28-day periods. For every 100m increase in high-speed running (>5.5 m·s−1), sleep onset and wake time were extended by 4.68min (CI:2.78—6.58mins) and 3.38mins (CI: 1.27—5.5mins), respectively. However, considering that workload had no significant effect on total sleep duration, the changes to wake time and sleep onset time should not concern practitioners.
In Chapters 3, 5, and 7, objective sleep monitoring was completed using ReadiBand wrist-worn activity monitors. Though, it was acknowledged that these devices cannot readily link objective sleep quality and performance, and players' data could be missing due to poor band adherence. Therefore, another approach was trialled where the effect that inadequate sleep has on cognitive variables that are sensitive to sleep loss was determined, rather than measuring sleep directly. Consequently, this thesis also assessed the use of a novel virtual reality eye-tracking device that could rapidly administer an oculomotor task which was reported to be sensitive to total sleep deprivation. However, to be efficacious in a footballing environment, the device would have to demonstrate sensitivity to the daily fluctuation of sleep. Target radial variation (a measure of spatial accuracy) was found to be significantly correlated with perceived daytime sleepiness (r=0.33, p=0.005), however, no further relationships were observed between oculomotor function, psychometric vigilance, daytime sleepiness, and sleep metrics. In a retrospective analysis on a second data set from military personnel (that was included to augment the original analysis), only psychomotor vigilance, and not oculomotor function, were associated with the total amount of sleep achieved. This suggested that this device would not be efficacious in a footballing environment as a replacement for sleep monitoring.
Following the research presented in Chapters 4 and 5, it was surmised that a bespoke approach to sleep intervention would be more efficacious than team-based interventions. To this end, a framework was conceptualised in collaboration with a multidisciplinary team from SFC (Chapter 7). Next, a player was referred to the scheme after reporting excessive night time awakenings. After consultation, the player completed several subjective questionnaires to assess sleep quality (Pittsburgh Sleep Quality Index), insomnia severity (Insomnia Severity Index), and daytime sleepiness (Epworth Sleepiness Scale) followed by a period of objective sleep monitoring. The sleep monitoring confirmed excessive nighttime awakenings and based on the responses from the initial consultation, a sleep hygiene intervention was applied tailored to the players' responses during the initial consultation. Results revealed improved subjective sleep quality, insomnia severity, and nighttime awakenings. Whilst a case study cannot establish causality, it does provide a potential framework for practitioners looking to provide targeted sleep interventions.
Conclusions:
In general, professional football players' sleep quantity, latency, and WASO is within available population-based reference limits.
Scheduling variables, and not workload variables, are associated with activity monitor-derived objective sleep metrics in professional football players.
Scheduled start time is associated with the amount of sleep that professional U18 football players receive.
An oculomotor task does not have the requisite sensitivity to detect acute sleep loss in professional football players.
A bespoke sleep intervention strategy can be efficacious in an applied footballing environment for players reporting sleep disruption.
ORCID: https://orcid.org/0000-0002-2806-9478, Hill, Jessica 