Associations between paediatric fatigue and eating behaviours

Summary Background In adults, poor sleep quality is associated with increased obesogenic eating behaviours; less is known about this relationship in youth. The objectives of this study were to assess the strength of association between fatigue‐related quality of life (QoL) and eating behaviours among youth and to describe the associations in participants with percent body fat (%BF) above and below the 90th percentile for sex and age. Methods Caregiver‐reported measures of fatigue (Pediatric QoL Multidimensional Fatigue Scale) and eating behaviours (Child Eating Behaviour Questionnaire) were obtained from participants aged 8–17 years. %BF was measured by iDXA and grouped by sex‐ and age‐specific percentiles. Multiple linear regression adjusting for age, sex and race/ethnicity was used. Results Of the 352 participants (49% male), 44.6% had %BF >90th percentile. General, sleep/rest and cognitive fatigue QoL was inversely associated with food approach behaviours: food responsiveness, enjoyment of food, emotional overeating and desire to drink. For participants with %BF >90th percentile, higher general fatigue QoL was associated with higher satiety responsiveness (0.13; 95% confidence interval [CI 0.03, 0.24]). For participants with %BF ≤90th percentile, higher general fatigue QoL was associated with less satiety responsiveness (−0.16; 95% CI [−0.31, −0.01]). Conclusion Less fatigue symptoms were associated with less behaviours associated with food approach among paediatric participants. For participants with %BF >90th percentile, less symptoms of general fatigues corresponded with more satiety. Though causation has yet to be established, youth with elevated %BF should be screened for fatigue symptoms and offered counselling on sleep hygiene or a sleep medicine referral to help mitigate weight gain.

Possible factors contributing to the association between poor sleep quality and obesity include the effect of sleep quality on motivation/cognitive fatigue, obesogenic eating behaviours and caloric intake in children. 31,32 Both cognitive fatigue (subjective feeling of exhaustion that follows sustained cognitive demands) and decreased motivation/anhedonia were associated with elevated BMI, independent of sex and age, in youth with obesity. 33 Reduced motivation and cognitive fatigue may decrease the likelihood of engaging in certain behaviours such as routine physical activity and choosing healthy foods. For example, adolescents with disrupted sleep/wake schedules with late bedtime/late rise patterns of sleep were more likely to engage in additional screen time and less physical activity than youth with early bedtime/early rise patterns. 34 Additionally, poor sleep quality and decreased sleep duration have been associated with obesogenic eating behaviours, such as emotional eating and food responsiveness in both youth and adults. 31,35,36 Sleep restriction has also been associated with increased snacking, increased consumption of carbohydrates, fat, salt and sugar, and larger portions in both youth and adults. [37][38][39][40][41][42][43][44] Fatigue symptoms are not only associated with sleep quality or duration (sleep-/rest-related fatigue), but also can be classified as cognitive fatigue or general fatigue (i.e., feeling physically weak, feeling too tired to do things, trouble starting things). 45 Clarifying the association between fatigue-related quality of life (QoL) and specific eating behaviours may help elucidate the fatigue-obesity relationship as well as inform targeted behavioural modification strategies in youth with overweight, obesity and/or a rapidly increasing body fat percentage (%BF).
The objectives of this study were to (1) assess the strength of association between three subscales of fatigue-related QoL (general, sleep/rest and cognitive fatigue) and eating behaviours among youth and (2) describe the associations between fatigue QoL and eating behaviours in children with %BF above and below the age-and sexspecific 90th percentile. The a priori hypothesis was that general, sleep/rest and cognitive fatigue QoL subscales would be inversely associated with food approach behaviours (food responsiveness, emotional overeating. enjoyment of food and desire to drink). That is, higher fatigue QoL score (less fatigue symptoms) would be associated with less endorsement of food approach behaviours. Participants with %BF >90th percentile were hypothesized to have more food approach behaviours significantly associated with lower general, sleep/rest and cognitive fatigue QoL than participants with % BF ≤ 90th percentile.

| Study design and participants
This report was a secondary analysis of data collected as part of a cross-sectional study of vascular health among children and adolescents with a range of BMI and %BF values. 46 Children and adolescents (age range: 8-17 years old) were enrolled in the study.

| Anthropometrics and %BF
Height and weight measurements were obtained using a wallmounted stadiometer and an electronic scale, respectively. The mean of three separate measurements of height and weight was used. BMI was calculated as the body weight in kilogrammes divided by the height in meters squared. %BF and fat-free mass were determined by dual energy x-ray absorptiometry (iDXA GE Healthcare), using standard positioning techniques, and were conducted and analysed by trained staff. 47 Participants were separated into two groups based on %BF measured by iDXA: (1) participants ≤90th percentile for %BF and (2) participants >90th percentile for %BF. %BF percentile cut-points were determined using previously published data on nationally representative %BF percentiles for youth aged 8-20 years old. 48 In this previous analysis, 81% of male and 86% of female youth with Class 1 obesity (BMI >95th to <120th of the 95th percentile) were above the 90th percentile for %BF. 48 There was a decrease in the proportion of youth with Class I obesity as the %BF threshold decreased from the 85th percentile (53% male, 60% female) and 75th percentile (33% male, 37% female). 48 Therefore, the 90th percentile for %BF was used to in our study to describe a larger proportion of participants with both excess adiposity and Class I obesity.

| Pediatric Quality of Life Fatigue Scale
The

| Statistical analysis
Data from all participants with PedsQL MDFS™, CEBQ and nonmissing %BF measurements were analysed (n = 352). Descriptive statistics using mean and standard deviation or N and percent summarized the cohort. Multiple linear regression using robust standard errors was performed to evaluate associations between PedsQL MDFS™ scale score and CEBQ eating domain score for the entire cohort adjusting for age, sex, race and ethnicity. Multiple linear regression using robust standard errors was performed to evaluate associations between PedsQL MDFS™ scale score and CEBQ eating domain score in both %BF groups by use of an interaction term and adjusting for age, sex, race and ethnicity. %BF groups were defined as those above tor those at or below the age-and sex-specific 90th %BF percentile. 46 Because the analysis by %BF cut-point was a secondary objective and because we sought to limit the use of multiple comparisons, we intentionally did not directly compare associations between %BF groups. The PedsQL MDFS™ scale scores were scaled by a factor of 25 for more easily interpretable results. All statistical calculations were performed on R 3.5.0. Statistical significance was defined as p value <0.05.

| Participant characteristics
Three hundred fifty-two participants (48.9% male, 78.0% White, 11.4% Latino/Hispanic) were included in this analysis (   associations between cognitive fatigue QOL score and food avoid behaviours for participants with %BF >90th percentile.

| DISCUSSION
This cross-sectional study evaluated the association between three subscales of fatigue-related QoL (general, sleep/rest and cognitive fatigue) and eating behaviours and characterized the association in among youth with %BF above and below the age-and sex-specific 90th percentile. Less general, sleep-/rest-related and cognitive fatigue were associated with lower endorsement of all four food approach behaviours (food responsiveness, emotional overeating, enjoyment of food and desire to drink) in the overall cohort. Additionally, associations between general fatigue and food avoid behaviours (satiety responsiveness, slowness in eating, emotional undereating and food fussiness) differed in directionality for participants with %BF above and below the age-and sex-specific 90th percentile. Improvement in general, sleep-/rest-related and cognitive fatigue may be a lifestyle management focus for children with rapidly increasing BMI (even with a %BF <90th percentile) as addressing food approach behaviours may taper rapid weight gain. For example, specific food approach behaviours, such as desire to drink and emotional overeating, may serve as targets for higher yield behaviour interventions for patients with endorsement of general, sleep-/rest-related or cognitive fatigue symptoms regardless of BMI or %BF. Endorsement of desire to drink behaviour has been associated with a higher preferences for soda, fruit juice and milk and higher consumption of these beverages in children. 52 Excess fruit juice and sugar-sweetened beverage (SSB) consumption is an important contributor in the development of obesity among children and adolescents, and limiting access to these beverages is an important dietary intervention for weight management. 53   overeating, is also recommended when fatigue is identified in youth with overweight or obesity. In conclusion, high fatigue symptoms were significantly associated with increased food approach behaviours in youth. Future efforts aimed at characterizing the potential causal pathways between fatigue and poor sleep quality leading to unhealthy eating behaviours may inform more targeted and effective prevention and treatment strategies for obesity in youth.

ACKNOWLEDGEMENTS
We would like to thank the children and adolescents who participated in this study. We would also like to thank Ms Annie Sheldon and Ms Erin Hurley for their excellent coordination of this study and Ms Cameron Naughton for her programme management expertise.

INFORMED CONSENT AND ETHICAL CONSIDERATIONS
This study was conducted according to the guidelines laid down in the Declaration of Helsinki, and the University of Minnesota Institutional Review Board approved all procedures involving human subjects.