A Subjective Assessment of the Prevalence and Factors Associated with Poor Sleep Quality Amongst Elite Japanese Athletes

A total of 891 candidates for the 17th Asian games Incheon 2014, who were over 20 years old, participated in this study. The survey started on February 20 and ended on July 15, 2014. (The 17th Asian games Incheon 2014 was held between September 19 and October 4). Written informed consent was obtained after explaining the purpose, procedure, and possible risks of this study. This study was approved by the Japan Institute of Sports Sciences Ethics Committee.

Data from 817 (449 males and 368 females) of the 891 athletes (91.7%) with no missing data were analyzed.

The athletes were asked to complete the questionnaire after measuring their heights and weights. The questionnaire consisted of questions to obtain general information, the PSQI [30], the Epworth Sleepiness Scale (ESS) [31], two-question case-finding instruments [32], and a checklist for sleep hygiene. The general information included sex, age, sport, training-practice schedules, napping habits, and frequency of overseas trips in the past 6 months. Training-practice schedules were ascertained on each day of the week. Training-practice times per week were calculated from the training-practice times on each day.

The PSQI consists of 19 self-rated questions. The athletes were asked about their bed time, sleep latency, and getting-up time, and the time in bed and sleep efficiency were calculated as part of the PSQI. Subjective sleep quality was evaluated based on the global score of the PSQI. In the Japanese version, a PSQI global score higher than 5.5 is considered to indicate poor sleep quality [33]. The PSQI contains five other questions that are not reflected in the global score and are rated by a bed partner or roommate [26, 30]. Since the athletes were not accompanied by a bed partner or roommate at the time of completing the questionnaire, the latter five questions were modified to be answered by the athletes themselves in this study as follows:

For each of the five items, athletes selected an appropriate answer from the following: “Not at all,” “Yes, sometimes,” and “Yes, frequently.” For the logistic regression analyses, “Yes, sometimes” and “Yes, frequently” were defined as “the athlete has the symptom.” For the logistic regression analyses, “the athlete has the symptom” was assigned the dummy variable “1,” and the absence of the symptom was assigned “0.”

Napping habits were clarified by asking three questions. Firstly, the athletes stated their frequency of napping by selecting one of five choices: no napping habits, less than once per week, 1–2 times per week, 3–5 times per week, and almost every day. Athletes with napping habits were requested to answer the second and third questions. In the second question, athletes stated their timing of starting to nap by selecting one of nine choices: before noon, eight time bands of every 1 hour from 12:00 to 19:00, and depends on case. Thirdly, athletes stated their duration of napping by selecting one of five choices: 30 min or less, 31 to 60 min, 61 to 90 min, more than 90 min, and depends on case. For the logistic regression analyses, “no napping habit” and “within 30 min” were assigned the dummy variable “0,” and the others were assigned “1.”

Daytime sleepiness was evaluated by the ESS score (Japanese version). An ESS score higher than 10.5 is considered to indicate excessive daytime sleepiness [31].

Two-question case-finding instruments for depression were used to evaluate each athlete’s mood [32]. The instruments consisted of two questions: (1) “During the past month, have you often been bothered by feeling down, depressed, or hopeless?” and (2) “During the past month, have you often been bothered by little interest or pleasure in doing things?” Athletes were requested to answer “yes” or “no” to each question. For Japanese adults, a sensitivity of 99% and a specificity of 60.5% have been reported if the respondent answers “yes” to either of the two questions. If the respondent answers “yes” to both questions, a sensitivity of 87.9% and a specificity of 81.4% have been reported [34]. For the logistic regression analyses, the answer “yes” to both questions was considered “the athlete feels a depressive mood.” For the logistic regression analyses, “the athlete feels a depressive mood” was assigned the dummy variable “1,” and its absence was assigned “0.”

Each question had three possible replies: “I already do so,” “I may be able to, but do not already do so,” and “I think it is difficult to do so.” For the logistic regression analyses, “I may be able to, but do not already do so” and “I think it is difficult to do so” were defined as “the athlete does not already do so.” For the logistic regression analyses, “the athlete does not already do so” was assigned the dummy variable “1,” and the other was assigned “0.”

The athletes were asked about the frequency of overseas trips in the past 6 months.

The questionnaire took approximately 10 min to complete.

In addition, the results of training-practice schedules, sleep schedule, the PSQI, the ESS, and clinical information were categorized within each variable, and then the number and percentages of athletes were described in the categories.

To explore factors associated with poor sleep quality (PSQI global score > 5.5), logistic regression analyses were performed using the following independent variables: age, sex, BMI, frequency of overseas expeditions, training-practice schedules, sleep schedules, clinical information, depressive mood, and sleep hygiene. Initially, we examined these variables using a univariate analysis. Then, a multiple logistic regression analysis was performed using the variables showing significant correlations in the univariate analysis (p < 0.05). For the continuous variables, the results were used for the logistic regression analyses. For the variables regarding clinical information, depressive mood, and sleep hygiene, dummy variables were used for the logistic regression analysis. Wald statistics were used to examine the significant odds ratios generated by the regression analyses. All analyses were performed using IBM SPSS 22 for windows. Significance was set at p < 0.05.

Our results using the PSQI global score indicated that sleep quality is poor in 28% of elite Japanese athletes. A previous study [36] that examined general populations in Japan reported that the PSQI global scores were 4.51 ± 2.14 for males in their 20s and 5.30 ± 2.48 for females in their 20s. This study also reported that 30.1% of males in their 20s and 36.4% for females in their 20s scored above the cutoff for poor sleep quality on the PSQI. Our results of the mean PSQI global scores and prevalence of poor sleep quality for both males and females were slightly better than those of age-matched general populations in Japan. The differences between sexes observed in the sleep schedules, PSQI global, and ESS scores in this study were in line with those in previous studies [36, 37].

The multiple logistic analysis revealed that athletes’ sleep quality was significantly associated with five factors: “time in bed,” “depressive mood,” “eating breakfast every morning,” “avoiding the use of electronic devices (PC, smartphone) just before bedtime,” and “not thinking about troubles in bed.” Regarding excessive daytime sleepiness, Doi and Minowa examined 4722 Japanese and reported that insufficient nocturnal sleep, an irregular sleep-wake schedule, and depression were associated risk factors [37]. Although the evaluated parameters differed between the latter research and ours (PSQI global and ESS scores), common background-related factors may exist.

Our study shows that the mean time in bed was 7 h and 29 min in total, corresponding to 7 h and 37 min for males and 7 h and 20 min for females. As shown in Table 2, males from five sports and females from nine sports showed a mean time in bed of less than 7 h. The 2015 NHK Japanese Time Use Survey revealed that the sleep duration was 7 h and 27 min for males in their 20s (n = 424) and 7 h and 18 min for females in their 20s (n = 437) [38]. Since sleep duration is generally shorter than time in bed, our results may suggest that time in bed of Japanese athletes was not as long as those of age-matched general populations. At present, it is unclear and debatable whether a time in bed that is the same as the general population is sufficient for athletes. However, some researchers recommend a longer nocturnal sleep, i.e., 8–10 h [39‐41] for athletes. Samuels and Alexander [40] recommend 8–10 h of nocturnal sleep plus a 30-min nap between 2 and 4 pm for athletes. Bompa and Haff [41] reported that athletes require 9 to 10 h of sleep, 80–90% of it during the night. Rountree [42] recommends sleep extension with a longer training-practice time. Our results suggest that the time in bed of elite Japanese athletes is inconsistent with their recommendations.

Our study revealed that 34 (4.2%) athletes were considered short sleepers (time in bed < 6 h). Seventeen of the 34 short sleepers get up earlier than 6:00. However, only 4 of the 17 short sleepers reported that their training-practice starts before 8:00. Sargent et al. [4] reported that early-morning training markedly restricts the time in bed of swimmers. Our results differ from theirs. For an example, all of our kabaddi players reported that their training-practice starts around 19:00 and ends at 21:00. Twelve of 20 kabaddi players get up at 7:00 or earlier because of their jobs. For them, the schedules of both training-practice and their job may be the reasons for their short time in bed.

In this study, more than 80% of the athletes replied that they use electronic devices (PC, smartphone) just before bedtime. The Ministry of Posts and Telecommunications (Japan) reported that 53.7% of Internet users in Japan had delayed bedtimes, and 45.4% of them had a shortened sleep duration as a result of internet use [43]. It is also known that lights from electronic devices reduce melatonin levels, increase alertness at night, delay the circadian clock, and decrease alertness the next morning [22]. Thus, using electronic devices just before bedtime may lead to a shortened sleep duration and poor sleep quality [23‐25].

It is already known that skipping breakfast is associated with poor sleep quality [44]. Previous researchers have stated that busy lifestyles cause some people to skip breakfast, but some also suggested that skipping breakfast affects circadian rhythms. For example, in the liver, breakfast promotes food-induced entrainment of the circadian clock [45]. In addition, some researchers suggest that dietary components influence the synthesis of serotonin and melatonin, and this contributes to a good sleep quality [46]. Skipping breakfast may lead to the loss of these effects.

Hammond et al. reported that 26% of athletes showed self-reported mild to moderate symptoms of depression [47]. In their study, they demonstrated that in elite athletes, negative changes in their performances were correlated with depression. Guilliver et al. examined 224 elite Australian athletes and reported that 27.2% of them displayed depressive symptoms [48]. Yang et al. examined 257 collegiate athletes and reported that 21% of their athletes showed depression, and anxiety and pain were related to that depression [49]. Other factors such as sports injuries, career termination, and performance outcomes that are below expectations are also risk factors of depression [50]. It is also well-known that a depressive mood is one of the symptoms of overtraining or overreaching [51]. Although it was unclear whether our athletes were suffering from overtraining or overreaching, such risks may arise during periods of high training loads [52].

Although sleep disturbance is known as one of the symptoms of depression, sleep deprivation lowers the psychological threshold for the perception of stress [53]. In addition, there is a report that individuals with short (< 7 h/night) and long (≧ 9 h/night) sleep durations show an increased genetic influence for depression [54]. Thus, sleep quantity and quality are associated with depression and vice versa.

Symptoms of chronic stress, burnout, anxiety, and nervousness are prevalent amongst athletes [55]. Erlacher et al. studied 632 German athletes and reported that 65.8% of them experienced poor sleep during nights before important competitions [13]. Romyn et al. demonstrated strong negative associations between state anxiety and sleep quality in athletes [56]. Kashani et al. reported a significant correlation between perceived stress and the PSQI global score [57]. In non-pharmacologic treatment of insomnia, Lande and Gragnani recommend trying to resolve problems prior to bedtime or to make resolution a priority the following day [58]. Siebern et al. recommends relaxation techniques including progressive muscle relaxation, deep breathing techniques, body scanning, and autogenic training [59].

Although the logistic regression analysis did not reveal a significant association with sleep quality, 40% of the athletes had been informed by someone about “snoring loudly” and/or “leg twitching or jerking during sleep” (Table 5). Similarly, Swinbourne et al. examined 175 athletes in New Zealand and demonstrated that 38% of them were defined as snorers and 8% reported having apnoeic episodes [18]. Tuomilehto et al. examined 107 professional ice-hockey players and found that 14 of them had obstructive sleep apnoea [19]. In our studies, loud snoring and a long pause between breaths were widely distributed in many BMI categories, and higher percentages of athletes were found in larger BMI categories. Emsellem and Murtagh said that a BMI above 28.0 kg/m² and a neck circumference greater than 40.0 cm are risk factors for obstructive sleep apnoea syndrome [16]. However, this BMI criterion might be too high for the Japanese. Itasaka et al. divided 257 Japanese subjects into three groups: normal weight (BMI under 24.0 kg/m²), mildly obese (BMI 24.0–26.4 kg/m²), and obese (BMI 26.4 kg/m² or higher), and showed that the apnoea-hypopnea index, intraoesophageal pressure, and lowest oxygen saturation became significantly worse according to the degree of obesity [60]. In this study, the athletes with a BMI at or above 25 kg/m² showed a higher frequency of being informed about long pauses between breaths. However, some athletes with a BMI below 25 kg/m² were also similarly informed. Other factors, such as microgenia, retrognathia, rhinostenosis, and enlarged tonsils, should also be examined to identify causes of apnoea/hypopnea. Regarding leg twitching or jerking, there are both physiological and pathological causes. If the leg twitching or jerking is physiological and does not fragment sleep, treatment is rarely necessary [61]. For pathological forms, sleep-related movement disorders, such as periodic limb movement disorder and restless legs syndrome, may be related. A low ferritin level (> 50 μg/L) is one of the risk factors of restless leg syndrome [62]. Koehler et al. reported that 31% of male athletes and 57% of female athletes showed serum ferritin levels below 35 μg/L [63]. Checking blood profiles might be recommended if sleep is disturbed by frequent limb movements. In a study on 107 professional ice hockey players, Tuomilehto et al. found 14 with obstructive sleep apnoea, 13 with insomnia, 4 with restless legs syndrome and periodic leg movements, 1 with parasomnia, and 1 with delayed sleep-wake syndrome [19]. We should keep in mind that factors not showing an association in our logistic analysis may also affect individual athlete’s sleep quality.

For both males and females, depressive mood and the hygiene-factor “not thinking about troubles in bed” were correlated with poor sleep quality (Tables 6 and 7). Time in bed for males and getting up time and the hygiene-factor “preventing an irregular getting up time (within 2 h)” for females were also correlated with poor sleep quality. These results suggest that psychological factors and sleep schedules are common factors that affect poor sleep quality.

Firstly, we investigated Japanese athletes’ sleeping habits using a questionnaire and not by actual measurements. If applicable, it would be desirable to evaluate sleep schedules with equipment such as actigraphy [2‐4, 8, 9, 11, 12]. Secondly, Samuels et al. developed a new athlete-specific sleep-screening questionnaire because they felt that the PSQI classified a higher than expected number of athletes as poor sleepers [20]. For example, 24.7% of males and 32.1% of females showed a PSQI global score above 5.5, whereas our PSQI subcategory scores indicated that the number of athletes with insomnia seems to be small. It is debatable whether the PSQI is the most suitable sleep-screening questionnaire for athletes. Thirdly, athletes younger than 20 years old were excluded from this study. Since many of them spend the daytime at high school or junior high school, their training schedules and sleeping habits may be different from those shown by our results. Fourthly, we could not evaluate athletes’ performance in this research. Further studies are necessary to verify that an improvement in sleep leads to enhanced sport performance.