Current recommendations, by the National Sleep Foundation (NSF), suggest that school-aged children (6-13 years of age) and adolescents (14-17 years of age) should achieve 9-11 hours and 8-10 hours of sleep each night, respectively (Hirshkowitz et al., 2015). Junior (<18 years of age) athletes often receive less sleep than these recommendations (Riederer, 2020). Sleep quality has also been shown to be impaired in junior athletes (Suppiah et al., 2021). Sleep duration and sleep quality may be impaired by lifestyle factors, including school commencement times, homework, and potential social media and gaming (Hansen et al., 2017; Wahlstrom et al., 2017). Additionally, junior athletes' sporting commitments may impact their sleep opportunity, thus reducing their sleep duration and sleep quality (Dumortier et al., 2018; Kölling et al., 2016).
Studies in young adult tennis players have found that partial sleep deprivation (reduced sleep duration) negatively impacts serve, forehand and backhand accuracy (Reyner & Horne, 2013; Vitale et al., 2021). One week of sleep extension to nine hours per day has been shown to increase serve accuracy in college tennis players (Schwartz & Simon, 2015). These studies indicate a positive link between sleep duration and execution of tennis skills. However, increased sleep duration as part of a mixed-method recovery strategy (the use of multiple recovery strategies) did not positively impact tennis performance outcomes (Lever et al., 2021). Though, it did improve lower body power and reduce perceived muscle soreness in junior tennis players, indicating a link between increased sleep duration and improved physical performance (Duffield et al., 2014).
While there is a paucity of studies investigating the effects of sleep on physical performance in junior athletes (Riederer, 2020), numerous studies have been conducted in adult athletes from various sports, with equivocal findings (Watson, 2017). Specifically, reaction times have been shown to slow when partial (reduced sleep duration) or complete (maintain wakefulness) sleep deprivation occurs (Fullagar et al., 2015; Reilly & Edwards, 2007; Watson, 2017). Maximal strength appears to be unaffected by sleep duration (Reilly & Edwards, 2007; Sinnerton & Reilly, 1992; Watson, 2017). However, the effects of sleep duration on aerobic capacity and sprint performance are less clear, with some (Mah et al., 2011; Peacock et al., 2018; Watson, 2017), but not all (Reilly & Edwards, 2007; Sinnerton & Reilly, 1992), studies reporting decrements in performance following nights with reduced sleep duration.
Current studies have reported associations between physical factors and tennis performance (Fett et al., 2020; Ulbricht et al., 2016). Given these associations, aspiring junior tennis players must optimise their physical performance to ensure tennis success. Understanding the factors influencing physical performance, including sleep, is then of paramount importance. Therefore, we aimed to examine the extent to which sleep duration and sleep quality metrics influence physical performance metrics in Australian and German junior tennis players. We hypothesised that junior tennis players with sleep duration and sleep quality metrics that meet the NSF recommendations will have superior performance in physical tests.
Measures
Sleep-wake behaviour
The Consensus Sleep Diary (CSD) measured players' sleep duration and sleep quality (Carney et al., 2012). This measure was developed by the Pittsburgh Assessment Conference in 2012 and has since been validated (Carney et al., 2012; Maich et al., 2018). When completing the diary, players were required to report the following information for each day/night: (i) what time did they go to bed (bedtime), (ii) what time they tried to sleep, (iii) how long it took them to fall asleep, (iv) how many times they woke during the night and for how long, (v) what time they woke for the day, and (vi) what time they got out of bed. Additionally, the CSD included 5-point Likert scales for players to rate their sleep quality and restfulness. The information derived by the CSD enables calculation of time in bed (TIB), sleep onset latency (SOL), wake after sleep onset (WASO), sleep duration, sleep efficiency (SE), sleep quality and restfulness. Sleep duration, SOL, WASO and SE, measured by the Consensus sleep diary, have been validated against actigraphy (Maich et al., 2018). The CSD can be seen in Figure S1 of the supplementary file.
Each player's total sleep score was calculated based on age-appropriate (school-aged: 9 to 13 years or teenager: 14 to 17 years) sleep duration and sleep quality (sleep latency, wake after sleep onset and sleep efficiency) recommendations from the NSF (Hirshkowitz et al., 2015; Ohayon et al., 2017). The recommended ranges for each sleep variable can be seen in Figure 2. A detailed description of how the total sleep score was determined can be seen in Figure S2. The total sleep scores, presented as a percentage, represent the level of adherence (no adherence 0% to complete adherence 100%) each player had to sleep duration and sleep quality recommendations outlined by the National Sleep Foundation (Hirshkowitz et al., 2015; Ohayon et al., 2017).
Physical and Mental Fatigue
The Chalder Fatigue Scale (CFS) evaluated fatigue, including physical and mental fatigue (Chalder et al., 1993). The CFS is an eleven-item validated questionnaire designed to indicate the current physical and mental fatigue. Scores are then totalled and range from 0-33, with higher scores indicating greater fatigue levels. A score of 22 indicates a player was feeling fatigued at the time of completion.
Anthropometrics
Standing heights were measured and recorded to the nearest centimetre (0.01 m) using a tape measure. Players stood tall with their back against a wall, the distance from the ground to the top of their heads recorded. Body mass was measured using a set of electronic scales and was recorded to the nearest gram (0.01 kg).
Physical assessments
Flexibility
Flexibility was measured using a Sit and Reach test; players are required to sit on the ground with their legs extended and the soles of their feet against a sit and reach box (Flex tester, Novel Products Incorporated, Rockton, USA). Players had their hands on top of one another with palms facing down and reached forward as far as possible without their knees bending, holding this position for 2-3 s (Roetert et al., 1992). Players performed three trials, one after the other, with the best trial achieved recorded and analysed.
Linear speed
Linear speed was assessed over 5, 10 and 20-m distances using timing light sensors (Speed light, Swift Performance, Queensland, Australia). The testing procedure was previously used with junior tennis players; it involves players beginning their linear sprint from a standing start, positioned 50 cm behind the first timing light sensor (Ulbricht et al., 2016). Players were allowed practice runs if required after the initial instructions were provided. Each player then performed two maximal 20-m sprints with a two-minute passive recovery between sprints. Sprint times over 5-, 10- and 20-m distances were recorded to the nearest ms (0.01 s).
Tennis Agility Test
Agility was assessed using the Tennis Agility Test (TAT), which required players to start in a standing 'ready position', straddling the centre mark of the baseline. The tester initiated the test who performed a forehand or backhand swing, indicating that the player runs to their right or left, respectively. The player was required to run to the doubles sideline, perform a forehand or backswing, then run to the opposing doubles sideline, perform a forehand or backhand swing before running back to the centre mark to complete the test. The players were provided with a clear explanation of the test and allowed a practice run before three trials, interspersed with two-minute passive recovery periods. The three trials' best reaction time (the time between tester and player movements) and total time (time to complete the test) were included for analysis. Both times were recorded to the nearest 0.3 ms (0.03 s) using Kinovea (open-access video analysis software; https://www.kinovea.org ).
Counter movement jump
Players jump height during a counter movement jump (CMJ) was used to determine lower body power. To reduce the involvement of arm-swing during the jump, players held a pole across their shoulders (Legg et al., 2017). A linear positional transducer (GymAware, Kinetic Performance Technology, Canberra, Australia) attached to the end of the pole was used to measure the jump height of players. Players were instructed to jump as high as possible whilst keeping the pole level; the trial was repeated if the player tilted the pole. Three maximal CMJs were performed with two-minute passive recovery periods between jumps. The maximal height of each jump was recorded to the nearest centimetre (0.01 m).
Overhead medicine ball throw
The upper body power of players was assessed using an overhead medicine ball throw; this test has previously been used with junior tennis players (Ulbricht et al., 2016). This test requires players to stand with their feet side by side and, using both hands, throw a two-kilogram medicine ball overhead as far as possible. Players were instructed not to step forward when throwing as the measurement was taken from their feet to the point where the ball landed; if a player stepped forward, the throw was retaken. A total of three throws were performed, with a two-minute passive recovery period between throws. The furthest horizontal distance from the thrower to the landing position of the medicine ball was recorded and used for analysis. All throws were recorded to the nearest five centimetres (0.05 m).
Grip strength
Upper body strength was assessed using a grip strength test, commonly used to measure strength in junior tennis players (Fett et al., 2020; Girard & Millet, 2009; Ulbricht et al., 2016). A hand dynamometer (Advanced Hand Dynamometer, TTM, Japan) was gripped by the player and positioned by the side of their body; it was then squeezed maximally for three seconds. The player's grip strength was measured in kilograms (kg), and the best of two trials on each hand, separated by two minutes of passive recovery, was used for further analysis.
Repeat Sprint Ability
The anaerobic capacity of players was assessed using the repeat sprint ability test; this test has recently been performed with junior tennis players (Vitale et al., 2021). The test required players to sprint 20 metres as fast as possible every 20 s for ten repetitions. The sprint times of players were measured using timing gates (Speed light, Swift Performance, Queensland, Australia) placed one metre above ground level. The anaerobic capacity was determined by the fatigue decrement score, calculated using the following formula (Chapman & Sheppard, 2011; Vitale et al., 2021).
Fatigue decrement (%) = ((total sprint time - ideal time) / total sprint time) x 100
Hit and Turn Tennis Test
The aerobic capacity of players was assessed using a tennis-specific endurance test called the Hit and Turn Tennis Test; this test is reliable and valid (Ferrauti et al., 2011). The test was delivered by a standardised audio file that dictates the direction and speed players move. Players were required to run and sidestep in the indicated direction and perform forehand and backhand swings until they could not make the time or voluntarily withdrew.
Statistical analysis
As determined by the Shapiro-Wilk test, data were normally distributed and presented as mean ± standard deviation (range). A T-test was conducted to identify if the sleep latency, wake after sleep onset, sleep duration or sleep efficiency differed between school-aged children and teenagers. General linear modelling was also used to determine the influence of sleep duration and sleep quality on each of the physical performance metrics. All models were adjusted for sex, age and nationality. False discovery rate (FDR) correction was applied to account for multiple models and to mitigate false positive results. The effect sizes were reported as Cohen's d or r (partial correlations). Cohen’s d thresholds were identified as small = 0.2, medium = 0.5 and large = 0.8 (Cohen, 2013). While Cohen's r thresholds were identified as small = 0.1, medium = 0.3 and large = 0.5 (Cohen, 2013). Analysis was conducted using R Studio software package, Version 1.1 (RStudio Team, 2020). A p-value <0.05 was considered statistically significant.