Psychological changes in athletes infected with Omicron after return to training: fatigue, sleep, and mood

Participants

A total of 267 athletes who resumed training in early January 2023 at the Chongming Sports Training Base in Shanghai participated in the first test. Among them, 31 athletes had not obtained a positive result in the nucleic acid or antigen test before the test (some athletes had suspected symptoms and had recovered), while six athletes reported having been infected before December 2022 due to personal and competition reasons and had already returned to regular training. After excluding these participants, a total of 230 athletes were included in the study, comprising 117 males and 113 females with a mean age of 18.62 years (SD = 4.16). These athletes participated in various sports, including fencing, modern pentathlon, boxing, judo, badminton, table tennis, gymnastics, martial arts, and handball.

One hundred and sixty-seven athletes participated in the second test after 1 week, with 86 males and 81 females (mean age 19.16 years, SD = 4.09). One hundred and twenty-four athletes participated in the third test after a further two-week interval, comprising 64 males and 60 females with a mean age of 18.74 years (SD = 3.87). Ninety-nine of these athletes completed the scale in all three tests, with 53 males and 46 females (mean age 18.99 years, SD = 3.93). Moreover, 117 athletes completed the exercise task in all three tests, consisting of 62 males and 55 females, with a mean age of 18.81 years (SD = 3.85).

The study protocol was approved by the Ethics Committee of the Shanghai Research Institute of Sports Science (Shanghai Anti-doping Agency) (Ethics Approval Number LLSC20230001), and complies with ethical principles stated in the latest version of the Declaration of Helsinki. All participants provided written consent before the first test.

Design

This study was structured into two main parts, utilizing quasi-experimental designs. The designs are described in Fig. 1 in association with the entire study procedure. The first part examines the impact of the duration of return to training on athletes’ fatigue, sleep, and mood, using self-report scales. Initially, athletes who returned to training in batches were divided into three groups based on the length of time they had been back in training. A 3 (duration: short, medium, and long) × 2 (test: pre-test, post-test) mixed design was employed, with time as a between-subjects variable and test point as a within-subjects variable. The pre-test refers to the athlete’s state when they first returned to training, and the post-test refers to the state at the second test, conducted after 1 week of training. By the time of the second test, the length of time the athlete had returned to training was classified into three categories: short (approximately 1 week), medium (1–2 weeks), and long (more than 2 weeks), as depicted in Fig. 1A. The dependent variables were the athletes’ mood, sleep, and fatigue status.

Additionally, the study analyzes the characteristics of psychological factors influenced by longer training periods. A one-factor two-level within-subjects design was employed to accomplish this, with the independent variable being the test (pre-test and post-test). The pre-test and post-test were the athletes’ mood, sleep, and fatigue status at the second and third tests, respectively, with approximately 2 weeks between tests, as shown in Fig. 1B.

The second part of the study analyzes the characteristics of exercise-induced subjective fatigue through an exercise task. A 3 (time: first test, second test, third test) × 2 (exercise: pre-exercise, post-exercise) within-subjects design was employed. The time refers to the three tests conducted, as depicted in Fig. 1C. The dependent variables were the level of physical and cognitive fatigue subjectively evaluated by the athletes.

Procedure

From late December 2022 to early January 2023, most athletes from sports teams at the Shanghai Chongming Sports Training Base returned to the base to resume training. The criteria for resuming training were twofold: (1) the antigen/nucleic acid test must show a negative result after a period of rest in the dormitory or home isolation following infection, and (2) the athlete’s team can resume training and calls the athlete back to the training base. To ensure training safety, sports scientists at the training base conducted three sequential tests of athletic ability after the athletes’ return to training in early January. This study was conducted with these three tests (Fig. 1), each completed within 1 week.

Each test consisted of two parts: a test of fatigue, sleep, and mood scales and a test of exercise-induced fatigue. The athletes completed the scales on-site or 1 day in advance on a dedicated webpage. For the first test, athletes were asked to recall their status during the week when they returned to training. For the second and third tests, athletes were asked to respond based on their status during the previous week from the day of testing. Tests for exercise-induced fatigue were conducted both before and after exercise.

Scales

(1) Symptoms questionnaire

The symptom questionnaire was developed based on the surveys used by medical institutions in Shanghai to investigate the symptoms (including fever, cough, sore throat, diarrhea, and other common symptoms) of the participants after infection and after they returned to training, as well as the dates of COVID-19-positive results in nucleic acid/antigen tests, their first negative results in nucleic acid/antigen tests after infection, and their return to training.

(2) Patient global impression of change (PGIC)

The PGIC was used to evaluate the subjective perception of symptom change after an intervention. This scale has only one item, a 7-point scale (ranging from one for much worse to seven for much improved) that subjects selected based on their physical condition compared to the pre-intervention state. The scale has been used in a study on the effects of exercise on facilitating recovery after infection with COVID-19 (Araújo et al., 2023). In the present study, during the first test, subjects were required to compare their current status with that before the infection and their current status with that when they had obtained their first negative result in the nucleic acid or antigen test after infection. During the second and third tests, participants were required to compare their current status with the previous test.

(3) Visual analog scale—training and daily fatigue (VAS-TDF)

Perceived training and daily fatigue were evaluated using the visual analog scale. Four questions were asked about physical fatigue in training, cognitive fatigue in training, overall physical fatigue in life, and overall cognitive fatigue in life. Participants responded using either their mobile phones or paper-based questionnaires. On the mobile phone webpage, subjects had to drag the slider on a line that fits the length of the screen of the phone to the position that matched their feelings, and the webpage converted the score automatically according to the dragging distance (0–100; 0 means no fatigue at all, 100 represents the extreme point of fatigue). This method has been used in studies of both physical and cognitive fatigue in different sports and is effective in evaluating fatigue (Coimbra et al., 2021; Gündoğdu et al., 2021). This method has also been used in studies to measure the mental state of Chinese people during the COVID-19 pandemic (Yan et al., 2021). Athletes used VAS-TDF to assess fatigue in training and daily life in the current study.

(4) Athens Insomnia Scale for non-clinical application (AIS-NCA)

The AIS-NCA was revised from the Athens Insomnia Scale and consisted of seven items, each scored on a scale of 1–5, with each option indicated by the text adapted to that item. All items are divided into two dimensions, sleep problems, and daytime functioning, and can be calculated for both subscales and the total scale score, with higher scores indicating more severe sleep quality problems (Sattler, Seddig & Zerbini, 2021). The scale has been used in studies of sleep related to COVID-19 and has demonstrated good utilization (Zerbini et al., 2022). The Chinese version of the questionnaire used by the participants consists of six items, constituting the two dimensions. The internal consistency coefficients were good in this study, with Cronbach’s α = 0.688 for the whole scale, Cronbach’s α = 0.636 for sleep problems, and Cronbach’s α = 0.761 for daytime functioning (based on the data of the first test).

(5) Depression, Anxiety, and Stress Scale (DASS-21)

The DASS-21 consists of twenty-one items scored from 0 to 3, divided into three dimensions of depression, anxiety, and stress, with higher scores, indicating more severe symptoms (Gong et al., 2010; Lovibond & Lovibond, 1995). Researchers have tested the reliability of this scale in athletes during the epidemic, demonstrating its validity in evaluating depression, anxiety, and stress in this population (Vaughan, Edwards & MacIntyre, 2020). The scale has good cross-cultural validity among Chinese adolescents (Mellor et al., 2014). The internal consistency coefficients were good in this study, with Cronbach’s α = 0.928 for the whole scale, Cronbach’s α = 0.575 for depression, Cronbach’s α = 0.767 for anxiety, and Cronbach’s α = 0.838 for stress (based on the data of first test).

(6) Visual analog scale—exercise-induced fatigue (VAS-EIF)

A visual analog scale (VAS) with 10 cm lines was employed to assess exercise-induced fatigue in this study. This method is relatively efficient in evaluating acute fatigue (Schlichta et al., 2022). The questionnaire comprised five questions related to physical fatigue, cognitive fatigue, sleepiness, decreased thinking clarity, and attention, all of which were completed on a paper-based form. For each question, a 10-cm line segment was provided, and participants indicated their level of fatigue by marking a vertical line on the line segment. The longer the length of the marked line, the higher the level of fatigue, except for the question related to attention, where the opposite was true. Sleepiness, inability to think clearly, and attention were complementary to central fatigue. Sleepiness, an experience where sleep is the primary pathway to recovery, was used to complement feelings related to fatigue but with differences in recovery (Mairesse et al., 2017). The ability to think and to pay attention were modified from the NASA-task load index interpretation of mental demand (Hart & Staveland, 1988). VAS-EEF was used to evaluate athletes’ perceived fatigue before and after the exercise task.

(7) Karolinska Sleepiness Scale (KSS)

The KSS was employed in the present study to evaluate the current subjective level of sleepiness or alertness. It is a widely used scale in driving, sleep deprivation, and shift work studies. The KSS consists of only one item scored from 1 to 9 (one indicating extreme alertness, and nine indicating excessive sleepiness). A corresponding text label accompanies each option, and subjects should select the one that best matches their feelings (Baulk, Reyner & Horne, 2001; Miley, Kecklund & Åkerstedt, 2016). KSS complemented the VAS-EEF to measure fatigue levels before and after the exercise task.

Six-minute stair climb test

The exercise task employed in this study was a six-minute stair climb test. This test was based on the test utilized in a previous study (Da Costa et al., 2017). As a submaximal test, the intensity in this study was adjusted to correspond with the participating athletes’ physical form and physical function. The test was divided into two phases, each lasting 3 min. In the first phase, athletes climbed up and down a stair of dimensions 15 cm in height and 30 cm in width at a frequency of 120 beats per minute. Three minutes later, the athletes immediately proceeded to the second phase, where they climbed up and down a higher step stair (40 cm for male and 30 cm for female athletes) at the same rate.

This task was a component of the athletic ability monitoring program implemented following athletes’ return to training. Before and after the exercise task, athletes were required to wear sports physiology data collection equipment and complete several tests. The pre-test was administered using the scales before the exercise. At the same time, the post-test was carried out using the scale after the post-exercise sports physiology data collection (approximately 3 min later).

Statistical analysis

The data were processed and analyzed using Excel 2016 and SPSS 26.0. Athletes who did not complete the required scales or exercise tasks were excluded from the statistical analysis. Depending on the design, this study employed repeated measures ANOVA, one-sample t-tests, and paired-sample t-tests to analyze the data. These statistical tests have been shown to have relatively good tolerance and robustness to the data, and the use of correction values can also contribute to the results in a beneficial way (Games, Keselman & Clinch, 1979; Poncet et al., 2016; Sullivan & D’Agostino, 1992; Zimmerman & Zumbo, 1992). To ensure the reliability of the results, appropriate correction values were used in the ANOVA based on the results of tests such as the sphericity test. Moreover, G*power 3.1.9.2 was used to calculate the minimum sample size required for reliable results. Specifically, the athletes’ infection and symptom characteristics were evaluated using descriptive statistical results. At the same time, their subjective physical condition changes (PGIC) were tested using one-sample t-tests (PGIC = 4 for no significant difference). The minimum sample size was calculated to be 54 (effect size d = 0.50, α = 0.05, 1−β = 0.95). The data from the scales in the first and second tests were analyzed using repeated measures ANOVA with Bonferroni correction for multiple comparisons (Fig. 1A). For those that did not satisfy the sphericity test hypothesis, the Greenhouse-Geisser correction was utilized. The minimum sample size was calculated to be 66 (effect size f = 0.25, α = 0.05, 1−β = 0.95, r = 0.50). The data from the scales in the second and third tests were analyzed using paired-samples t-tests (Fig. 1B). It was confirmed that there were significant correlations between the two tests before the statistics (ps < 0.001). The minimum sample size was calculated to be 54 (effect size dz = 0.50, α = 0.05, 1−β = 0.95). The data from the exercise-induced fatigue in the three tests were analyzed using repeated measures ANOVA (Fig. 1C). For those that did not satisfy the sphericity test hypothesis, the Greenhouse-Geisser correction was utilized. The presence of a non-integer number of degrees of freedom in the statistical results indicated that the pvalues were corrected. The minimum sample size was calculated to be 28 (effect size f = 0.25, α = 0.05, 1−β = 0.95, r = 0.50).

Symptoms and recovery

Among the 230 athletes, 87.82% reported fever after infection. Of those with fever, 96.53% remembered and reported their maximum body temperature, with a mean value of 39.26 °C (SD = 0.84). Additionally, 90.43% of athletes experienced dry cough, 89.13% experienced fatigue, 73.04% experienced sore throat, 56.09% experienced decreased smell/taste, 43.48% experienced diarrhea, 80.87% experienced muscle aches, and 79.57% experienced sleepiness. The mean duration from the initial positive result of nucleic acid or antigen test to a negative retest was 6.77 days (SD = 3.29; n = 229). It took 6.91 days (SD = 4.62; n = 229) for athletes to resume training after regaining a negative result.

Most athletes experienced dry cough (65.65%) at the first test and fatigue (64.78%). Additionally, 24.78% experienced sore throat, 24.35% experienced reduced smell/taste, 21.30% experienced diarrhea, 52.61% experienced muscle soreness, and 53.91% experienced sleepiness.

Upon initial return to training, athletes perceived their condition to be worse than before infection (M = 3.38, SD = 1.46), t(207) = −6.128, p < 0.001, Cohen’s d = −0.425. However, they perceived their physical condition to be better than when the nucleic acid or antigen test had just turned negative (M = 4.52, SD = 1.29), t(207) = 5.864, p < 0.001, Cohen’s d = 0.407. At the second test, athletes perceived their physical condition to be better than at the previous test (M = 4.751, SD = 1.00), t(168) = 9.782, p < 0.001, Cohen’s d = 0.752. Similarly, at the third test, athletes perceived their physical condition to be better than at the previous test (M = 5.00, SD = 1.09), t(107) = 9.577, p < 0.001, Cohen’s d = 0.922.

Changes after different durations of training

The results of the descriptive statistics are shown in Table 1. There was only a main effect of test on overall experienced physical fatigue, F(1,164) = 11.197, p = 0.001, η_p² = 0.063; fatigue was lower on the post-test. There was no significant main effect of duration, F(2,164) = 2.599, p = 0.077, η_p² = 0.031, and no significant interaction, F(2,164) = 0.552, p = 0.577, η_p² = 0.007. Additionally, there was no significant main effect of the test on overall experienced cognitive fatigue, F(1,164) = 0.806, p = 0.371, η_p² = 0.005. There was also no significant main effect of duration, F(2,164) = 0.321, p = 0.726, η_p² = 0.004, and no significant interaction, F(2,164) = 2.485, p = 0.086, η_p² = 0.029.

Regarding physical fatigue experienced during training, there was only a main effect of test, F(1,164) = 16.266, p < 0.001, η_p² = 0.090, with perceived fatigue being lower on the post-test. However, there was no significant main effect of duration, F(2,164) = 1.575, p = 0.271, η_p² = 0.016, and no significant interaction, F(2,164) = 1.317, p = 0.271, η_p² = 0.016. Furthermore, there was no significant main effect of test on cognitive fatigue experienced during training, F(1,164) = 0.402, p = 0.527, η_p² = 0.002. There was no significant main effect for duration, F(2,164) = 0.247, p = 0.781, η_p² = 0.003, and no significant interaction, F(2,164) = 1.482, p = 0.230, η_p² = 0.018.

Overall sleep quality showed a significant main effect of test, F(1,164) = 20.012, p < 0.001, η_p² = 0.109, indicating better sleep quality on the post-test. However, no significant main effect of duration was found, F(2,164) = 0.164, p = 0.849, η_p² = 0.002, and no significant interaction was observed, F(2,164) = 0.587, p = 0.557, η_p² = 0.007. Regarding the sleep problem dimension, a significant main effect of test was observed, F(1,164) = 11.560, p = 0.001, η_p² = 0.066. Sleep disturbance was lower in the post-test. However, no significant main effect of duration was observed, F(2,164) = 0.864, p = 0.423, η_p² = 0.010, and no significant interaction was found, F(2,164) = 0.677, p = 0.509, η_p² = 0.008. A significant main effect of test was observed for the daytime function dimension, F(1,164) = 16.190, p < 0.001, η_p² = 0.090. Specifically, daytime dysfunction was lower in the post-test. Nevertheless, no significant main effect of duration was found, F(2,164) = 2.159, p = 0.119, η_p² = 0.026, and no significant interaction was observed, F(2,164) = 0.145, p = 0.865, η_p² = 0.002.

Regarding depression, a significant main effect of test was observed, F(1,164) = 8.879, p = 0.003, η_p² = 0.051. Scores on the post-test were lower. However, no significant main effect of duration was observed, F(2,164) = 0.978, p = 0.378, η_p² = 0.007, and no significant interaction was observed, F(2,164) = 0.590, p = 0.556, η_p² = 0.007. On anxiety, a significant main effect of test was observed, F(1,164) = 27.981, p < 0.001, η_p² = 0.146. Scores on the post-test were lower. However, no significant main effect of duration was found, F(2,164) = 1.957, p = 0.145, η_p² = 0.023, and no significant interaction was observed, F(2,164) = 0.434, p = 0.649, η_p² = 0.005. Regarding stress, there was no significant main effect of test, F(1,164) = 2.187, p = 0.141, η_p² = 0.013. Additionally, no significant main effect of duration was found, F(2,164) = 0.538, p = 0.585, η_p² = 0.007, and no significant interaction was observed, F(2,164) = 0.167, p = 0.846, η_p² = 0.002.

Changes after a longer fixed training period

The results of the descriptive statistics are shown in Table 2. After an additional 2 weeks of training, a statistically significant increase in overall physical fatigue was observed in athletes, t(98) = −2.822, p = 0.006, Cohen’s d = −0.284, while no significant change was detected in overall cognitive fatigue, t(98) = −1.211, p = 0.229, Cohen’s d = −0.122. As for training-related fatigue, a significant increase was found in physical fatigue, t(98) = −2.123, p = 0.036, Cohen’s d = −0.213, while cognitive fatigue remained unchanged, t(98) = −1.463, p = 0.147, Cohen’s d = −0.147.

No significant changes were observed in overall sleep quality, t(98) = 1.059, p = 0.292, Cohen’s d = 0.106, sleep problems, t(98) = 0.380, p = 0.705, Cohen’s d = 0.038, and daytime function, t(98) = 1.572, p = 0.119, Cohen’s d = 0.158.

A significant decrease was found in the level of depression, t(98) = 3.256, p = 0.002, Cohen’s d = 0.327, while the level of anxiety remained unchanged, t(98) = 1.689, p = 0.094, Cohen’s d = 0.170. Moreover, the stress decreased significantly, t(98) = 2.265, p = 0.026, Cohen’s d = 0.228.

Changes in exercise task-induced fatigue

The results of the descriptive statistics are shown in Table 3. There was a significant interaction on subjectively experienced physical fatigue, F(2,228) = 4.830, p = 0.009, η_p² = 0.041. Simple effects analysis showed no significant difference between the pre-test and post-test in the first test (p = 0.122) and second test (p = 0.113), with only a higher level of physical fatigue in the post-test of the third test than in the pre-test (p = 0.021). A significant main effect of time was found, F(1.834,209.121) = 12.732, p < 0.001, η_p² = 0.100. Physical fatigue was higher on the first test than on the second test (p < 0.001) and third test (p < 0.001), but there was no difference between the second test and third test (p > 0.999). There was no significant main effect of exercise, F(1,114) = 1.606, p = 0.208, η_p² = 0.014.

Regarding the subjective experience of cognitive fatigue, there was a significant main effect of time, F(1.882,214.511) = 8.064, p = 0.001, η_p² = 0.066. Cognitive fatigue was higher for the first test than for the second test (p = 0.001) and third test (p = 0.005), but there was no significant difference between the second test and third test (p > 0.999). There was no significant main effect of exercise, F(1,114) = 0.202, p = 0.654, η_p² = 0.002; nor a significant interaction, F(2,228) = 1.296, p = 0.276, η_p² = 0.011.

There was a significant main effect of exercise on subjectively experienced sleepiness, F(1,114) = 20.027, p < 0.001, η_p² = 0.149. Sleepiness was lower on the post-test than on the pre-test. There was no significant main effect of time, F(1,114) = 1.482, p = 0.229, η_p² = 0.013. Similarly, no significant interaction was observed, F(2,228) = 2.464, p = 0.087, η_p² = 0.021.

A significant main effect of time was found on the subjective experience of being unable to think clearly, F(1.810,206.296) = 8.505, p < 0.001, η_p² = 0.069. The thinking was less clear on the first test than on the second test (p < 0.001) and third test (p = 0.017), but there was no significant difference between the second test and the third test (p > 0.999). There was no significant main effect for exercise, F(1,114) = 2.388, p = 0.125, η_p² = 0.021. There was no significant interaction, F(1.810,206.340) = 0.544, p = 0.564, η_p² = 0.005.

In addition, there was no significant main effect of time on the subjective experience of attention, F(2,226) = 0.789, p = 0.455, η_p² = 0.007. Similarly, there was no significant main effect of exercise, F(1,113) = 1.003, p = 0.319, η_p² = 0.009. Furthermore, no significant interaction was observed, F(2,226) = 2.529, p = 0.082, η_p² = 0.022.

Regarding sleepiness or lack of alertness measured by the KSS, an interaction between time and exercise was observed, F(1.789,202.194) = 4.121, p = 0.021, η_p² = 0.035. Simple effects analysis showed that, for the pre-test, the level of sleepiness was significantly higher on the first test compared to the second test (p < 0.001) and third test (p < 0.001). However, there was no significant difference between the second and third tests (p = 0.588). For the post-test, there was no significant difference between all three tests (p_1–2 = 0.169, p_1–3 = 0.188, p_2-3 > 0.999). Furthermore, there was a significant main effect of time, F(1.870,211.277) = 10.370, p < 0.001, indicating that sleepiness was higher during the first test compared to the second test (p = 0.001) and the third test (p = 0.001). However, there was no significant difference between the second and third tests (p = 0.984). Additionally, there was a significant main effect of exercise, F(1,113) = 25.963, p < 0.001, η_p² = 0.187, indicating that sleepiness levels were higher during the pre-test than the post-test.