Is Mid-morning Breakfast as Healthy as Early-morning Breakfast for Blood Sugar Control in Adolescent Girls?

Dietary manipulations that can moderate postprandial glycaemia and insulinaemia are at the cornerstone of type 2 diabetes (T2D) prevention. Indeed, repeated glycaemic excursions cause oxidative stress, inflammation, and atherosclerosis, increasing T2D risk. Adolescence is a critical time to promote dietary behaviours that can be sustained across the lifespan. Among the dietary factors influencing postprandial glycaemia, the decline in breakfast consumption during adolescence in girls is of concern, with only ~20% of UK adolescent girls consuming breakfast daily. Cross-sectional and prospective observational studies show that infrequent breakfast consumption is associated with T2D risk markers (e.g., glucose, insulin) in adolescents and T2D manifestation in adults. Randomised controlled trials in adults suggest that improvements in glycaemic control and insulin sensitivity may be key to explaining the reduced T2D risk when contrasting breakfast consumption with breakfast omission. Acutely, breakfast omission exaggerates glycaemic and insulinaemic responses to subsequent meals when compared to breakfast consumption, which has been termed the 'second-meal effect'. Thus, two moderate glycaemic responses (with breakfast and lunch) would be potentially better for reducing T2D risk than one very large glycaemic excursion (after lunch). Yet, findings based on adults may not apply to the distinct hormonal and metabolic profiles of adolescent girls. Specifically, it is proposed that the 'second-meal effect' occurs due to increased glucose conversion into muscle glycogen in adults. Such responses may differ in adolescents due to their higher reliance on exogenous glucose and fat as fuels with a lower reliance on endogenous glucose and reduced capacity for muscle glycogen storage, in addition to the 32% reduction in insulin sensitivity occurring from pre- to mid-puberty. Unfortunately, current knowledge on breakfast and glycaemia in adolescents relies entirely on a few novel studies that have pooled data from 13 to 20 year olds with little consideration of pubertal status and sex. Hence, the glycaemic responses to breakfast manipulation in adolescent girls are not still well understood. Common barriers to consuming breakfast among adolescent girls include not being hungry and a lack of time in the morning. As such, the option to consume breakfast later in the morning may be an attractive alternative for this population. In terms of timing, definitions have proposed that breakfast is consumed within 2-3 hours of waking, typically no later than 10:00. Yet, inconsistent definitions have been employed across the literature, mainly because evidence to inform a health-based (e.g., focused on glycaemic improvements) breakfast timing 'cut-off' is lacking. Thus, the unique contribution of this proposal is the use of an experimental cross-over design to directly contrast the effects of consuming an early-morning versus mid-morning standardised breakfast with breakfast omission on postprandial glycaemia and insulinaemia in adolescent girls classified as habitual breakfast skippers. By focusing on adolescent girls, breakfast timing and postprandial glycaemia, this proposal is unique, topical and has potential impact for T2D prevention.

Recruitment:

Girls will be recruited from local schools after gaining parental informed consent and child assent. These schools have expressed an interest in facilitating recruitment for the proposed research by providing a platform to invite girls to participate and permission for time off school to complete the measures. The participants will be invited to attend an assembly at their school and those that wish to discuss their participation with a parent/primary caregiver will take an information pack home that includes written details of the study. On return of a contact form, students and their primary caregiver will be asked to complete the consent, assent and pre-participation health screen questionnaire, before registration on the study.

Experimental design:

This study will employ a cross-over design. Participants will complete three conditions assigned according to the Latin square method: breakfast omission (BO), early-morning breakfast consumption (EM-BC) and mid-morning breakfast consumption (MM-BC). The conditions will be conducted ~7 days apart in girls who have not started their menses or in the early-follicular phase (~28 days apart) in girls with regular menses to minimise the potential confounding influence of menstrual cycle phase. In the 48 hours before the conditions, dietary intakes, bed time and wake/out of bed times will be replicated, and vigorous physical activity will be minimised (as confirmed via accelerometry).

After an overnight fast, resting metabolic rate (RMR) and substrate oxidation will be estimated via expired air analysis, and a capillary blood sample for the measurement of plasma glucose and insulin will be taken (~08:15 to 08:30). These measures will be collected at regular intervals, with a standardised lunch consumed at 4 h (~12:30) and followed by a 2-h lunch postprandial period. The standardised breakfast will be provided immediately after fasting measures at ~08:30 for EM-BC and 2 hours later at ~10:30 for MM-BC. Participants will remain sedentary throughout. Water intake will be replicated between the conditions.

Test meals:

A carbohydrate-rich, low glycaemic index (GI) breakfast providing ~70% carbohydrate, ~17% fat and ~13% protein will be used. The breakfast will be provided in quantities containing 0.04 g carbohydrate per kcal of individualised daily RMR. The standardised lunch providing 0.05 g of carbohydrate per kcal of daily RMR will be based on high GI carbohydrate. Meal consumption time will be limited to 15 minutes and replicated between the conditions.

Data and statistical analyses:

Pre-lunch (4 hours) and post-lunch (2 hours) incremental (iAUC) and total (tAUC) area under the curve will be calculated using the trapezium rule for the primary outcomes, plasma glucose and insulin. and the secondary outcomes, resting energy expenditure and substrate oxidation rates. In addition, peak plasma glucose concentration after lunch will be determined.

Linear mixed models will be used to compare all outcome variables between the conditions. Models will include fixed effects for condition (and time for the condition by time analyses) and a random intercept for participants, and will be adjusted for the order effect. The Holm-Bonferroni correction for multiple comparisons will be applied. Normality will be checked using Shapiro Wilk tests. Statistical significance will be accepted at p≤0.05. Cohen's effect sizes will be used to gauge the magnitude of differences.

Sample size estimation:

Based on 80% power at an alpha level of p=0.05, it is estimated that 21 breakfast skipping adolescent girls are required to detect a meaningful between-condition difference (Cohen's f=0.36) in post-lunch plasma glucose area under the curve. Recruitment will target 27 participants across the two study sites to account for an expected 20% attrition rate.