Childhood and adolescence as ‘critical periods’

The prevalence of cardiometabolic disease risk markers, including insulin resistance and dyslipidaemia, and even the manifestations of certain cardiometabolic diseases are increasing globally among young people^{(Reference Andes, Cheng and Rolka1–Reference Lawrence, Divers and Isom3)}. Given that cardiometabolic disease develops over time, an understanding of disease precursors rather than manifestation is more relevant to young people. Even so, in the absence of global data, recent data of 3⋅47 million 10–19-year-olds in the US reported a 95⋅3 % relative increase in estimated type 2 diabetes prevalence over 16 years between 2001 and 2017^{(Reference Lawrence, Divers and Isom3)}. Furthermore, US National Health and Nutrition Examination Survey data have shown that 18⋅0 % of 12–18-year-olds have prediabetes^{(Reference Andes, Cheng and Rolka1)} and 20⋅2 % of 8–17-year-olds have an adverse lipid concentration^{(Reference Kit, Kuklina and Carroll4)}. With rises in obesity among young people being a key driver, these data suggest cardiometabolic disease is set to manifest at an earlier age in future generations, prolonging the health and economic burden^{(Reference Andes, Cheng and Rolka1,Reference Weiss, Bremer and Lustig5)} . Importantly, targeting young people to aid disease prevention rather than focusing efforts towards treatment in later life may also benefit the planet through reduced greenhouse gas emissions associated with disease diagnosis and treatment within healthcare systems; for example, due to reduced manufacturing of blood analysis consumables and drugs, motorised transportation of goods and services, and clinical waste disposal (see Fig. 1)^{(Reference Eckelman and Sherman6)}.

Fig. 1. Potential avenues that support a focus on children and adolescents (young people) for cardiometabolic disease prevention through physical activity and/or diet with co-benefits to mitigating climate change. Note that examples of pathways linking physical activity and dietary manipulations in young people with climate change mitigation include, but are not limited to, active travel, human-nature connections and plant-based diets.

Moderating postprandial glycaemia, insulinaemia and lipaemia is at the cornerstone of cardiometabolic disease prevention^{(Reference Blaak, Antoine and Benton7,Reference Heine, Balkau and Ceriello8)} . Such risk markers share common physiological pathways, leading to their ‘clustering’ and placing young people at further risk^{(Reference Weiss, Bremer and Lustig5)}. As such, public health recommendations that address multiple cardiometabolic disease risk markers stand to create the largest impact. Most waking hours are spent inducing marked metabolic fluctuations, emphasising the importance of postprandial rather than fasting measures. Postprandial glycaemia can increase cardiometabolic disease risk through oxidative stress, inflammation and atherosclerosis^{(Reference Blaak, Antoine and Benton7,Reference Heine, Balkau and Ceriello8)} with repeated rises and falls indicative of high glycaemic variability appearing particularly harmful^{(Reference Ceriello, Esposito and Piconi9,Reference Sun, Zhao and Teng10)} . Even in ‘healthy’ populations, higher postprandial glucose concentrations increase type 2 diabetes, cardiovascular disease (CVD) and all-cause mortality risk^{(Reference Abdul-Ghani, Abdul-Ghani and Ali11–Reference Ning, Zhang and Dekker14)}. Likewise, regular exposure to elevated postprandial triacylglycerol (TAG) concentrations promotes a pro-atherogenic lipid phenotype, including low HDL concentrations and the accumulation of small, dense LDL^{(Reference Cohn15)} that increases susceptibility to future clinical cardiovascular events. Taken together, these findings may explain the lack of sensitivity of fasting measures alone for the diagnosis of cardiometabolic disease^{(Reference Cosson, Hamo-Tchatchouang and Banu16)}.

Childhood and adolescence are vulnerable times of life and have been associated with entering the initial stages of cardiometabolic disease development. Among girls and boys, pubertal insulin resistance reduces the ability of insulin to increase glucose uptake and utilisation, resulting in a compensatory increase in insulin secretion and plasma insulin concentrations to maintain normal blood glucose concentrations^{(Reference Goran and Gower17,Reference Moran, Jacobs and Steinberger18)} . Pubertal insulin resistance is transient, typically reported to begin during early puberty, peak during mid-puberty and recover by the end of puberty^{(Reference Goran and Gower17)}. Yet, more recent longitudinal data have demonstrated that the first stages are detectable from about 7 years of age^{(Reference Jeffery, Metcalf and Hosking19)}. Unfortunately, the diets of many young people are not conducive to the puberty-related reduction in insulin sensitivity; for example, free sugar intakes exceeded the recommended level in 87 % of 4–10-year-olds and 92 % of 11–18-year-olds in the UK in 2020⁽²⁰⁾. The factors contributing to pubertal insulin resistance remain elusive, but appear to include rises in insulin-like growth factor-1 and adiposity^{(Reference Jeffery, Metcalf and Hosking19,Reference Kelsey and Zeitler21,Reference Roemmich, Clark and Lusk22)} , and affect girls more than boys^{(Reference Moran, Jacobs and Steinberger18,Reference Jeffery, Metcalf and Hosking19,Reference Cooper, Dring and Morris23)} . Due to the additive effect of adiposity, unfortunately pubertal insulin resistance in young people with obesity does not fully resolve, resulting in an increased risk of β-cell exhaustion and cardiometabolic disease at a younger age^{(Reference Kelsey and Zeitler21)}. Likewise, the antecedents of atherosclerotic disease originate in the formative years^{(Reference McGill, McMahan and Herderick24)}, where aortic and coronary atherosclerotic lesions can develop in the first two decades of life and progress in prevalence and severity over time^{(Reference Holman, McGill and Strong25,Reference Strong and McGill26)} . Further, CVD risk markers that manifest in youth are predictive of subclinical atherosclerosis in adulthood^{(Reference Davis, Dawson and Riley27,Reference Raitakari, Juonala and Kähönen28)} and may increase the risk of premature death in adults^{(Reference Franks, Hanson and Knowler29)}. These data highlight the importance of protecting cardiometabolic health from an early age.

For effective cardiometabolic disease prevention, childhood and adolescence have been identified as a critical time to establish healthy, sustainable physical activity and dietary behaviours for life-long impact^{(Reference Eccles, Midgley and Wigfield30,Reference Todd, Street and Ziviani31)} . Yet, young people remain under-represented in the field of exercise and nutrient metabolism. Drawing on data from adult samples, acute exercise is a potent stimulus for moderating postprandial glycaemia, insulinaemia and lipaemia for up to 2–3 d^{(Reference Henriksen32–Reference Maraki and Sidossis34)}. Critically, such findings may not apply to the distinct hormonal, metabolic and behavioural profiles of children and adolescents, including pubertal insulin resistance^{(Reference Goran and Gower17,Reference Moran, Jacobs and Steinberger18)} , fuel utilisation^{(Reference Riddell35,Reference Timmons, Bar-Or and Riddell36)} and physical activity patterns^{(Reference Armstrong and Welsman37)}. The lack of paediatric-specific data is concerning given the low physical activity levels of many young people; for example, about 71 % of European 2–18-year-olds do not meet the physical activity guidelines when measured objectively^{(Reference Steene-Johannessen, Hansen and Dalene38)}, and about 81 % of 11–17-year-olds were classified as insufficiently physically active globally based on population survey data of 1⋅6 million participants^{(Reference Guthold, Stevens and Riley39)}. Attention to adolescents is of particular relevance due the adolescent decline in physical activity, which is more pronounced among girls than boys and supported by a large body of longitudinal research employing objective physical activity measures^{(Reference Farooq, Martin and Janssen40)}. During this time, adolescents often begin to skip breakfast^{(Reference Monzani, Ricotti and Caputo41)}, which may interact with physical activity and glycaemic control^{(Reference Betts, Richardson and Chowdhury42–Reference Gonzalez48)}. However, a comprehensive summary of the available paediatric evidence on the acute glycaemic, insulinaemic and lipaemic responses to physical activity and the role of breakfast–physical activity interactions have not yet been undertaken.

This review paper will critically summarise the extant literature on the cardiometabolic responses up to 48 h after completing physical activity bouts among children and adolescents. The possible interactions of breakfast with physical activity on cardiometabolic disease risk are then discussed. Reflecting traditional cardiometabolic disease risk markers and the available paediatric literature, this review paper will pay specific attention to glycaemia, insulinaemia and lipaemia, notwithstanding that an array of other markers exists. Given the seriousness and timeliness of global climate change with responsibility to act across disciplines, this review paper offers new insights into the inherent interactions between child–adolescent behaviour and cardiometabolic health in the context of environmental sustainability to aid climate change mitigation efforts, including exploring future research avenues. Based on these themes, Fig. 1 provides a platform for the concepts integral to this review paper, offering potential avenues that support a focus on young people for cardiometabolic disease prevention with co-benefits to mitigating climate change through direct and indirect pathways. From the outset, it should be highlighted that the term ‘physical activity’ is used to describe bodily movements with an energy expenditure >1⋅5 metabolic equivalents (METs), whereas the term ‘exercise’ is used to describe the subset of physical activity that has pre-planned, structured characteristics. The collective term ‘young people’ is used where findings apply to mixed samples of children and adolescents and where it is not essential to distinguish between chronological age (referred to as ‘age’ for simplicity) and pubertal stage.

Cardiometabolic responses to acute physical activity

As illustrated in Fig. 2a, experimental investigations on acute physical activity have typically employed crossover designs using a 1 or 2 d study paradigm, whereby the physical activity stimulus is either applied before measuring fasting and/or postprandial cardiometabolic responses or is accumulated throughout the measurement period. These experimental designs over multiple days reflect the expected lag between exposure to the physical activity stimulus and cardiometabolic benefit. The test day protocols for the measurement of outcome variables are designed in line with the differing pattern, duration and magnitude of postprandial glycaemic, insulinaemic and lipaemic responses to a single or multiple meals consumed across the day, as shown in Fig. 2b.

Fig. 2. (a) Schematic of typical protocols adopted to assess the postprandial cardiometabolic responses to acute exercise or physical activity bouts. Typically, a 1 or 2 d protocol is used, with exercise/physical activity manipulations and rest (control) conditions performed on day 1 and postprandial cardiometabolic risk marker responses assessed on day 1 and/or day 2 and in some cases day 3. Protocols differ according to whether the primary outcome is plasma glucose and insulin or TAG. Repeated capillary or venous blood samples are collected at pre-determined time-points in the fasted and postprandial state. OGTT, oral glucose tolerance test; HCM, high-carbohydrate meal; HFM, high-fat meal. (b) Illustration of the typical circulating plasma glucose and TAG response to consumption of a glucose load or meal at 0 h (indicated by the arrow). Postprandial glycaemic and insulinaemic responses are typically immediate, peaking at about 15–30 min and returning to baseline within about 2 h^{(20,Reference Zakrzewski and Tolfrey49,Reference Dring, Cooper and Williams56,Reference Zakrzewski, Stevenson and Tolfrey114)} , whereas postprandial TAG rises slower to reach a peak at about 3–4 h before returning to baseline after 6–8 h in the absence of a second-meal challenge^{(Reference Tolfrey, Doggett and Boyd65–Reference Barrett, Morris and Stensel68)}. As such, glucose and insulin concentrations require more frequent sampling over a shorter duration often in response to a carbohydrate-based test meal/drink, whereas less frequent blood sampling over a longer duration is required to capture postprandial TAG profiles, often in response to a high-fat meal. The macronutrient content of the test meal is typically designed to induce a sufficient metabolic challenge specific to the primary cardiometabolic risk marker under investigation to ensure ample scope for intervention effects. Mixed macronutrient meals have also been employed for the measurement of both glycaemia and lipaemia or to enhance ecological validity.

Single exercise bouts

Short, frequent physical activity bouts

Short, frequent physical activity bouts that break up sedentary behaviour have received growing interest as evidence of the high prevalence of sedentary behaviour across the lifespan continues to accumulate^{(Reference Bauman, Petersen, Blond, Leitzmann, Jochem and Schmid80)}. Sedentary behaviours are any waking behaviours characterised by an energy expenditure ≤1⋅5 METs, while in a sitting, reclining or lying posture^{(Reference Tremblay, Aubert and Barnes81)}, which can be replaced by physical activity of a light (1⋅6–2⋅9 METs), moderate (3⋅0–5⋅9 METs) or vigorous (≥6⋅0 METs) intensity during waking hours. This avenue of research is timely as sedentary behaviour has emerged as a strong determinant of mortality and cardiometabolic disease risk in adults^{(Reference Ekelund, Tarp and Steene-Johannessen82,Reference Patterson, McNamara and Tainio83)} and is associated with adverse cardiometabolic health outcomes in young people^{(Reference Carson, Hunter and Kuzik84)}. Efforts to mitigate the deleterious effects of prolonged sitting (a common proxy for sedentary behaviour) through short, frequent physical activity bouts are likely to have greater appeal for those who are not motivated to perform continuous bouts of structured exercise, where paediatric populations typically perceive prolonged exercise as more demanding than adults^{(Reference Timmons and Bar-Or85)}. This pattern of activity may also more closely reflect the sporadic physical activity patterns of children and adolescents^{(Reference Armstrong and Welsman37)} and be more conducive to integrate into the school day.

A meta-analysis of studies across the lifespan showed that interrupting sitting with short, frequent light- to moderate-intensity physical activity bouts reduces postprandial glucose and insulin, but not TAG, concentrations^{(Reference Saunders, Atkinson and Burr86)}. Although the impact of disrupting sitting with regular activity bouts on postprandial cardiometabolic risk markers was not modified by age, a closer examination of the paediatric-specific studies presented here suggests relatively modest and inconsistent effects. In this regard, breaking up prolonged sitting with 3 min moderate-intensity walking bouts performed every 30 min during a 3 h OGTT reduced glucose and insulin concentrations but did not alter postprandial lipaemia in 7–11-year-old children with healthy weight^{(Reference Belcher, Berrigan and Papachristopoulou87)} and reduced insulin concentrations in 7–11-year-old children who were overweight or obese^{(Reference Broadney, Belcher and Berrigan88)}. Using less frequent physical activity bouts in 13–14-year-old boys, accumulating four bouts of work-matched high-intensity physical activity (2 × 1 min intervals at 90 % peak power) reduced the postprandial glucose response on the intervention day when compared with continuous moderate-intensity (about 6 min at 90 % gas exchange threshold) exercise and a rest control trial, although no effect was observed for postprandial lipaemia^{(Reference Bond, Williams and Jackman89)}. In contrast, postprandial glucose, insulin and TAG concentrations were all unaffected in response to interrupting sitting with 2 min light-intensity walking bouts every 20 min for 8 h compared to 8 h of uninterrupted sitting in 10–14-year-olds; this lack of effect was maintained even when the 2 min light walking breaks were supplemented with two, 20 min moderate-intensity treadmill bouts^{(Reference Saunders, Chaput and Goldfield90)}. Similarly, breaking up sitting time with bouts of self-paced walking and standing accumulated during two consecutive simulated school days did not alter postprandial glucose, insulin or TAG concentrations assessed on the second simulated school day or the next day in 11–13-year-old physically inactive girls^{(Reference Smallcombe, Biddle and Slater91)}. Thus, the immediate and delayed effect of an ecologically valid standing and walking intervention appears to be negligible.

Discrepancies between studies are most likely explained by differences in participant characteristics and study design features, including variations in the physical activity or exercise stimulus, meal content or glycaemic load, timing of activity bouts and frequency of outcome assessments. Specifically, the relatively low frequency of blood sampling of ≥90 min intervals in some studies may have missed important glycaemic fluctuations in close proximity to the meals and activity/standing bouts^{(Reference Saunders, Chaput and Goldfield90,Reference Smallcombe, Biddle and Slater91)} . Using continuous glucose monitoring devices that measure interstitial glucose concentrations at about 5 min intervals, interrupting prolonged sitting with 2 min resistance activity bouts every 18 min reduced the incremental, but not total, glucose area under the curve for 3 h after both breakfast and lunch meals in adolescents aged 14–17 years^{(Reference Fletcher, Salmon and McNaughton92)}. Yet, these changes were relatively subtle in magnitude with detection likely facilitated by the improved sensitivity of continuous glucose monitors to capture glucose dynamics.

The lack of consistent reductions in postprandial lipaemic responses may be a consequence of the 1 d study protocol often employed in these studies^{(Reference Belcher, Berrigan and Papachristopoulou87–Reference Saunders, Chaput and Goldfield90)}. In this regard, accumulating 6 × 10 min bouts of running at 70 % of peak oxygen uptake did not alter next day postprandial glucose and insulin concentrations, but did reduce total and incremental area under the TAG curve by 11 and 16 %, respectively, in adolescent boys aged 12–14 years^{(Reference Sedgwick, Morris and Nevill72)}. These differences did not reach statistical significance, but are similar in magnitude to other exercise postprandial studies in children and adolescents. In support of the delayed effect of exercise on postprandial lipaemia, an in-school protocol in 11–13-year-old boys comparing the effect of 40 × 30 m sprints in one uninterrupted after school bout and accumulated in four bouts of 10 × 30 m sprints strategically placed during natural breaks in the school day reported reductions in postprandial TAG concentrations the next day, despite no effect on the day of exercise^{(Reference Smallcombe, Barrett and Sherar78)}. This may provide a viable model for translating the metabolic health benefits reported with laboratory-based exercise protocols into free-living settings providing the exercise stimulus is of sufficient duration and/or intensity. Importantly, since favourable effects among adults are most consistently seen in clinical, overweight/obese and physically inactive populations^{(Reference Saunders, Atkinson and Burr86)}, further research in such paediatric populations is warranted and has high public health relevance.

Acute breakfast consumption–physical activity interactions for cardiometabolic health

Cross-sectional and prospective observational studies show that frequent breakfast consumption is associated with reduced cardiometabolic disease risk in young people^{(Reference Monzani, Ricotti and Caputo41,Reference Souza, Neves and Gorgulho93)} and manifestation in adults^{(Reference Ballon, Neuenschwander and Schlesinger94,Reference Chen, Zhang and Ge95)} . Such findings may in part be explained by higher free-living physical activity and enhanced glucose control in response to breakfast consumption v. breakfast omission, as reported in randomised controlled trials and crossover trials in adults^{(Reference Betts, Richardson and Chowdhury42–Reference Jovanovic, Leverton and Solanky47)}. An eating pattern that includes breakfast to aid the provision of a continuous carbohydrate supply may be of greater importance to young people v. adults due to their elevated resting and physical activity energy expenditures^{(Reference Harrell, McMurray and Baggett96,Reference Rogol, Clark and Roemmich97)} and higher reliance on exogenous carbohydrate as a fuel^{(Reference Timmons, Bar-Or and Riddell36)}. Indeed, increased exogenous glucose availability (i.e. a physiological response) as well as perceived energy levels (i.e. a psychological response) are potential mediators of the breakfast–physical activity relationship. An understanding of such responses is particularly relevant to adolescent girls due to the adolescent decline in breakfast consumption frequency and physical activity levels as well as pubertal insulin resistance disproportionately affecting girls^{(Reference Moran, Jacobs and Steinberger18,Reference Jeffery, Metcalf and Hosking19,Reference Cooper, Dring and Morris23,Reference Farooq, Martin and Janssen40,Reference Monzani, Ricotti and Caputo41)} . Depending on the country and definition of breakfast employed, the prevalence of skipping breakfast can range from 10 to 30 % among young people, and is higher among adolescents (v. children) and girls (v. boys)^{(Reference Monzani, Ricotti and Caputo41)}.

Inconsistent cross-sectional findings on the association between breakfast consumption and objectively measured physical activity among young people^{(Reference Corder, van Sluijs and Steele98–Reference Vissers, Jones and Corder100)} highlight the need for experimental trials to provide more definitive causative interpretations of the findings with reduced interference from confounders. Using a 3 d crossover trial to compare standardised breakfast consumption v. breakfast omission in adolescent girls, physical activity was unaffected despite incomplete energy intake compensation of the breakfast meal^{(Reference Zakrzewski-Fruer, Plekhanova and Mandila101)}. If repeated habitually, this lack of energy compensation would be expected to result in weight gain in response to breakfast consumption v. omission, contradicting the higher adiposity reported among habitual breakfast skippers^{(Reference Monzani, Ricotti and Caputo41,Reference Souza, Neves and Gorgulho93)} . Yet, it is possible that wrist-worn accelerometry used in this study was not sensitive enough to detect differences in physical activity in response to breakfast manipulation. Using combined heart rate accelerometry, a 7 d crossover trial reported that adolescent girls spent more time in light-intensity physical activity before 10:30 and after school, and less time sedentary after school during daily v. intermittent (i.e. alternate breakfast omission and consumption) standardised breakfast consumption^{(Reference Zakrzewski-Fruer, Wells and Crawford102)}. The 19⋅8 min/d higher light-intensity physical activity during daily v. intermittent breakfast consumption complemented previous findings among adults that breakfast-induced changes in physical activity may be characterised by the replacement of sedentary behaviours with spontaneous light-intensity physical activity, whereas moderate- to vigorous-intensity physical activity is typically planned and structured^{(Reference Betts, Richardson and Chowdhury42)}. Interestingly, no differences were reported during the 10:30–15:00 time segment, possibly because the school timetable restricts physical activity opportunities, limiting the ‘window of opportunity’ among school-aged young people. Thus, previous reports that morning physical activity typically before around 12:00 has been shown to be particularly sensitive to breakfast manipulation in adults^{(Reference Betts, Richardson and Chowdhury42–Reference Yoshimura, Hatamoto and Yonekura44)} may not apply to adolescents during school days. Although it is plausible that contrasting the extremes of daily breakfast omission v. consumption across 7 d may confer more pronounced effects, no difference in physical activity quantified via combined heart rate accelerometry was found when comparing these extremes in adolescent girls who habitually skipped breakfast^{(Reference Zakrzewski-Fruer, Seall and Tolfrey103)}. Nevertheless, the girls were able to accurately compensate after 10:30 for the missed energy at breakfast through increased energy intake, primarily in the form of carbohydrate, during breakfast omission v. breakfast consumption. The lack of physical activity response may have been because the girls were already accustomed metabolically and behaviourally to skipping breakfast, suggesting that habituation to breakfast omission could alter responses to breakfast manipulation. However, ethical restrictions pose challenges in asking girls who habitually consume breakfast to engage in extended periods of breakfast omission.

Inconsistent extant findings may be due to environmental constraints inherent to the lives of adolescent girls, including the school timetable, extra-curricular activities, family routines and commitments. As such, further research that includes opportunities to be physically active would be useful in extending current knowledge. Adding a further degree of complexity, the nature of the causal relationship between breakfast and physical activity may also depend on physical activity intentions. Indeed, breakfast may promote physical activity or, alternatively, intentions to engage in physical activity may prompt breakfast consumption in preparation, and early morning exercise may result in ‘breakfast consumption’ (i.e. within 2–3 h of waking)^{(Reference O'Neil, Byrd-Bredbenner and Hayes104)}. Studies in this area have used carbohydrate-based breakfasts to examine the potential impact on free-living physical activity as this is the primary fuel for physical activity^{(Reference Zakrzewski-Fruer, Plekhanova and Mandila101–Reference Zakrzewski-Fruer, Seall and Tolfrey103)}; nevertheless, the effect of breakfast quantity and quality requires investigation. Additionally, data among boys, younger children and ‘at risk’ young people are required to inform population-specific recommendations.

Due to the lack of clarity on the potential association and/or causal link between breakfast consumption and physical activity among adolescents, alternative mediating factors may play a more prominent role in explaining the favourable cardiometabolic health profiles of habitual breakfast consumers. Among adults, breakfast consumption attenuates glycaemic and insulinaemic responses to subsequent meals when compared to breakfast omission, as termed the ‘second-meal effect’^{(Reference Chowdhury, Richardson and Tsintzas45–Reference Jovanovic, Leverton and Solanky47)}. If repeated habitually, a series of moderated glycaemic responses induced by breakfast consumption could reduce cardiometabolic disease risk through a reduced magnitude of glycaemic excursions and glycaemic variability^{(Reference Blaak, Antoine and Benton7,Reference Heine, Balkau and Ceriello8)} . A key mechanism for the ‘second-meal effect’ in adults is increased glucose conversion into muscle glycogen^{(Reference Jovanovic, Leverton and Solanky47)}. Such responses may differ in adolescents due to their lower reliance on endogenous glucose and reduced capacity for muscle glycogen storage, particularly in younger, less mature individuals^{(Reference Riddell35,Reference Timmons, Bar-Or and Riddell36,Reference Eriksson and Saltin105)} , and with an estimated 32 % reduction in insulin sensitivity occurring from pre- to mid-puberty^{(Reference Goran and Gower17)}. The slower rate of gastric emptying that has also been proposed to contribute to attenuations in glycaemia in response to breakfast consumption v. omission in adults^{(Reference Gonzalez48)} may, again, not apply equally to young people. Indeed, puberty, age and sex may affect gastric emptying in healthy children and adolescents, with slower rates reported in girls v. boys and during mid-late puberty v. pre-puberty^{(Reference Kovacic, Zhang and Nugent Liegl106)}. Unfortunately, current knowledge on breakfast consumption and glycaemia among adolescents relies on a few novel studies that have pooled data from adolescents and adults aged 13–20-year-olds with little consideration of pubertal status, age and sex^{(Reference Alwattar, Thyfault and Leidy107,Reference Bauer, Reynolds and Douglas108)} ; hence, the responses of children and adolescents specifically remain unknown.

The available research based on pooled adolescent–adult data suggests that the ‘second-meal’ effect depends on factors such as habitual breakfast patterns and breakfast meal composition. In a novel study of girls classified as overweight/obese (mean age 19 years), 3 d of skipping breakfast v. consuming a normal-protein (12 g protein) breakfast v. consuming a high-protein (32 g protein) breakfast did not affect post-lunch glycaemic and insulinaemic responses when assessed on day 4 among habitual breakfast skippers^{(Reference Alwattar, Thyfault and Leidy107)}. However, the habitual breakfast consumers, who did not complete the breakfast skipping trial, had lower afternoon and daily glucose responses following the higher-protein breakfast v. the normal-protein breakfast^{(Reference Alwattar, Thyfault and Leidy107)}. Extending a similar protocol over 12 weeks with a similar yet mixed-sex sample, high protein (35 g) breakfast consumption appeared to be more beneficial for free-living glycaemia than normal protein (13 g) breakfast consumption^{(Reference Bauer, Reynolds and Douglas108)}; yet, this was not an acute, controlled study and reductions in energy intake, unhealthy evening snacking of high-fat/high-sugar foods may have contributed to the findings^{(Reference Leidy, Hoertel and Douglas109)}. Other than age and/or puberty, a potential reason for the unclear findings to date may be that the repeated ingestion of glucose appears to be key to inducing the ‘second-meal effect’ among adults, with low glycaemic index breakfasts being particularly beneficial^{(Reference Bonuccelli, Muscelli and Gastaldelli110,Reference Liljeberg and Björck111)} . Further, the energy content of the breakfast meals^{(Reference Alwattar, Thyfault and Leidy107,Reference Bauer, Reynolds and Douglas108)} was modest when considering data in adults^{(Reference Chowdhury, Richardson and Tsintzas45–Reference Jovanovic, Leverton and Solanky47)} and recommendations that breakfast should provide at least 20 % of daily energy needs^{(Reference Gibney, Barr and Bellisle112,Reference Monteagudo, Palacin-Arce and Mdel113)} . Considering daily glucose profiles, low glycaemic index breakfasts also induce a lower immediate response compared with high glycaemic index breakfast among adolescents^{(Reference Cooper, Dring and Morris23,Reference Zakrzewski, Stevenson and Tolfrey114)} . Using a standardised carbohydrate-rich low glycaemic index breakfast, our recent work with adolescent girls who consumed breakfast habitually demonstrated significant and large reductions in peak, post-lunch (2 h) and total trial (5 h) glycaemia and insulinaemia in response to breakfast consumption v. omission followed by a standardised high glycaemic lunch provided at 3 h, with minimal inter-individual variability^{(Reference Zakrzewski-Fruer, Morari and Champion115)}. Further, enjoyment during exercise performed in the afternoon was enhanced with breakfast consumption v. omission^{(Reference Zakrzewski-Fruer, Morari and Champion115)}. Although this work did not target adolescents ‘at risk’ of poor glycaemic control, it suggests that moderations in glycaemia with breakfast consumption may be important in driving a favourable cardiometabolic health profile. Further research on the possible interactive effect of breakfast consumption and physical activity requires investigation due to the potential link between these two variables^{(Reference Corder, van Sluijs and Steele98–Reference Vissers, Jones and Corder100,Reference Zakrzewski-Fruer, Wells and Crawford102)} . Such experimental research would help move closer to developing an evidence-based standardised definition of breakfast consumption for health that can be employed by researchers and practitioners.

Future directions: opportunities to impact human and planetary health using active travel to school as an example

With preventable cardiometabolic diseases and climate change progressively escalating over recent decades due to a complex interplay of modern human practices, multi-disciplinary approaches that target certain human behaviours may contribute to addressing these global challenges simultaneously. Results on the co-benefits of physical activity and nutrition in young populations for human and planetary health are of high interest, particularly due to cardiometabolic disease risk reaching concerning levels among young people and global climate change placing uncertainty on their futures (see Fig. 1)^{(Reference Gasparri, Eissa and Imbago-Jácome116)}. Active travel to school offers a physical activity opportunity for school-aged young people that can benefit both (their) human and planetary health. Further, morning active travel is in close proximity to the breakfast meal, providing a pathway for potential breakfast–physical activity interactions, as discussed previously^{(Reference Corder, van Sluijs and Steele98–Reference Vissers, Jones and Corder100,Reference Zakrzewski-Fruer, Wells and Crawford102)} . Concurrently, active travel reduces the greenhouse gas emissions associated with motorised vehicle transport, thus assisting climate change mitigation. In contrast to many physical activity opportunities, transport is most often a stipulation of everyday life with almost all school-aged young people travelling to school across the school year. Further, active travel may offer a convenient route to build physical activity into the day with a minimal time burden and interference with daily tasks; this is important given that lack of time or competing activities for time are common barriers to physical activity^{(Reference Martins, Costa and Sarmento117)}. Thus, encouraging school-based active travel may be a feasible approach to increase population-level physical activity levels.

Accumulating evidence among children and adolescents outlined in this review paper suggests that the likelihood of achieving cardiometabolic health benefits through active travel will depend on the duration, intensity (affected by the nature of the route, inclines and terrain), mode (e.g. walking, cycling, scooting) and frequency of the physical activity characterising the active travel^{(Reference Zakrzewski and Tolfrey49–Reference Dring, Cooper and Williams56,Reference Tolfrey, Doggett and Boyd65–Reference Sedgwick, Morris and Nevill79)} . Thus, it is important that such factors are amenable to change. The weight of the evidence from intervention and prospective studies among young people and adults combined supports the positive effects of active travel over longer periods and distances on cardiometabolic health outcomes^{(Reference Saunders, Green and Petticrew118)}. Yet, further research utilising study designs that can infer causality, including acute studies, with measures of cardiometabolic disease risk other than obesity is required to draw stronger conclusions among children and adolescents specifically^{(Reference Saunders, Green and Petticrew118)}. Determining the active travel characteristics that can confer health benefits would inform a health-based definition of ‘active travel’ that could be used in research and practice, which is currently lacking^{(Reference Saunders, Green and Petticrew118)}.

Considering real-life application, it may be most feasible for active travel to contribute to light- or moderate-intensity physical activity; although, active commutes have potential to reach vigorous intensities in certain instances, such as up-hill cycling. It is also crucial that the duration of the commute is sufficient to contribute to daily physical activity, which may be a challenge in young children who often live in close proximity to school^{(Reference Metcalf, Voss and Jeffery119)}. However, high levels of air pollution on some commutes may negate some of the positive health benefits, especially for children who are more susceptible than adults due to greater physiological sensitivity to air pollutants and higher respiration rates^{(Reference Al-Kindi, Brook and Biswal120)}. Additionally, active travel may interact with daily diet and physical activity behaviours, and such potential compensatory responses might either negate or enhance any effects of active travel. With future research, an understanding of the acute and chronic cardiometabolic risk marker responses to active travel across different intensities, modes, durations and frequencies has potential to inform messages that could help promote active travel for both health and climate change mitigation. Additionally, research focused on adaptation to climate change, such as strategies to keep cool during physical activity in extreme heat, will be important to promote active travel in our changing climates.