Discussion

The present study aimed to establish if different types of BC were associated with differing portion sizes. Our study has demonstrated that our study population overserved the majority of BC types tested, with denser BC being overserved to a greater degree than less dense varieties.

Previous studies had shown that individuals had a tendency to pour larger portions than the RSS^{(20,Reference Schwartz and Byrd-Bredbenner23)} . However, limitations to these studies were that they had only compared portions of flaked BC and therefore the findings could not be applied to other types of BC. SACN⁽³⁰⁾ published that there was a tendency for high-fibre BC to be overconsumed; however, only two categories were compared. In our analysis, the majority of BC appeared to be overserved regardless of the type of BC. The exception to this was the pre-portioned BF category, indicating that they were easier to portion. However, the degree of over-portioning tended to be associated with the physical characteristics of the BC, consistent with previous literature showing the influence of physical properties of food on portion size^{(Reference Osterholt, Roe and Rolls27,Reference Rolls, Meengs and Roe28)} . To improve portion size awareness, the size of a portion should become more prominent on packaging and there should be standardisation across types of BC. Alternatively, CEEREAL should consider updating the suggested RSS to reflect a realistic portion size, which would subsequently change the traffic light nutritional information often found on the front of the pack. Unrealistic portion sizes may mislead the consumer to perceive a product as nutritious, leading to larger portion sizes^{(Reference Labbe, Rytz and Godinot31)} and although Public Health England ⁽¹⁰⁾ included BC in their sugar reduction programme, government policy should place more pressure on manufacturers of high-energy dense BC, such as granola, to reformulate products to improve their nutritional value.

Differing densities of BC may explain some of the findings in our study. CEEREAL⁽²²⁾ considers product densities when proposing the RSS. Figure 3 showed that the smallest median DFR in weight after BF originated from BC including OP, OPC, GP and EGP, as well as F1, all of which were the less dense types of BC. It appeared that the participants’ tendency to overserve had less of an impact on serving weight with these types of cereal. In contrast, denser cereals such as granola, SW and oats had a significantly larger DFR indicating a greater degree of overserving and larger portion sizes. Incorporation of air into products and hence a reduction in density produces a decrease in energy consumed^{(Reference Osterholt, Roe and Rolls27)}. Two other studies also demonstrated how differing volumes of isoenergetic preloads influenced satiety in men as a result of overestimating energy content^{(Reference Rolls, Bell and Waugh32,Reference Rolls, Castellanos and Halford33)} . These studies showed that there was a dissociation between volume, weight and energy content, therefore altering the individual’s perception of portion size. The BC with a RSS of 30 g and a greater volume of air (GP, EGP, OPC and OP) had a significantly smaller median DFR suggesting a smaller difference between the manufacturers RSS and what the participants served themselves. Collectively, this may suggest that reducing the density of cereal may be an effective way to decrease energy intake.

Portion size is an important determinant of energy intake and given that all but one types of BC tested lead to a degree of overserving, we examined the impact this had on energy and macronutrient intake. Previous literature has demonstrated that when portion size increases, energy intake increases significantly and is sustained over subsequent days, that is, no compensation^{(Reference Rolls, Roe and Meengs34)}. In our analysis, granola had the largest freepour measure and was also the most energy dense. Consequently, this led to a large DFR of energy, supporting previous research showing that high-energy dense foods have the largest effect on energy intake^{(Reference Rolls, Roe and Meengs34)}. SW and oats are less energy dense and had a significantly smaller DFR for energy than granola. The significance of these observations can be seen when comparing our findings to the government recommendations from the One You campaign, specifically the advice that breakfast should provide approximately 400 kcal⁽³⁵⁾. Our analysis suggests that individuals regularly consuming the median portion size for granola would be consuming approximately 490 kcal at breakfast per day, exceeding these recommendations. This does not include other food items consumed at breakfast, such as milk on the cereal, which would increase the energy intake further and contribute to weight gain if this was not compensated for at other mealtimes. Therefore, government policy should encourage companies manufacturing high-energy dense BC to reformulate products to reduce energy density, as a result of lack of portion size awareness. Research has shown that energy density is not strongly associated with food choice; therefore, reformulation may not influence market sales, but increase health benefits^{(Reference Brunstrom, Drake and Forde36)}.

In our study, the types of cereal F1, OP and BF were low in fat, saturated fat and total sugar, but contained modest amounts of protein and fibre. However, this same cannot be said about categories such as F2, GP, EGP and OPC, especially considering sugar. As a single meal, breakfast contributes more to the daily intake of carbohydrate, total sugars and less to total daily intake of protein, total fat, saturated fat and fibre^{(Reference Gaal, Kerr and Ward37)}. Breakfast remains an important source of micronutrients, especially in young children and less so in adolescents^{(Reference Coulthard, Palla and Pot38)}. However, our finding that freepour measures were larger than the RSS is significant, and the importance of breakfast combined with the potential for a significant source of sugar suggests that BC reformulation should be taken seriously. For context, the median serving for granola in our study would see an individual consumes almost 14 g of sugar, with a similar observation seen in the F2 group (approximately 18 g/serving). If the median portion size of granola was routinely consumed, it would likely have negative implications on nutritional health as it may contribute to excess energy and sugar intake^{(Reference Livingstone and McNaughton39)}. There needs to be substantial drive to reformulate breakfast products to improve their nutritional composition and contribution to overall nutrient intake.

Similar to sugar, salt content of cereal is also a cause for concern and has been recently reported as such⁽²⁹⁾. Salt intake remains an important public health nutrition target, and as a nutrient has strong links to cardiovascular health, specifically blood pressure^{(Reference Huang, Trieu and Yoshimura40)}. Indeed, the gradual reduction of salt intake in the UK diet is speculated to be one of the contributing factors to lower stroke and CHD mortality^{(Reference He, Pombo-Rodrigues and MacGregor41)}, with salt content of BC decreasing as part of the governments incremental targets^{(Reference Pombo-Rodrigues, Hashem and He42)}. The salt reduction targets published in 2020⁽⁴³⁾ continue to include BC as a category, with new targets that are lower than the previous recommendations published in 2017. Thus, it is essential that cereal products be reformulated to meet these new goals. In our study, the cereals used had a wide range of salt contents, ranging from <0·01 to 1·12 g/100 g. For children, reducing salt intake can be a challenge, especially as previous studies have suggested that while a low-salt BC can be consumed for up to 8 weeks, there is still no significant change in salt preference^{(Reference Bobowski and Mennella44)}. In contrast, low-salt interventions in adults of 5 and 12 months duration do lead to changes in salt preference when trialled using various foodstuffs^{(Reference Blais, Pangborn and Borhani45,Reference Bertino, Beauchamp and Engelman46)} . For salt reductions to be effective in children, it would seem that improving educational awareness of the consequences of excessive salt is more effective than simply reducing the salt content of a product^{(Reference He, Wu and Feng47)}, and a combination of the two would have the greatest impact on health. Combined with result from Action on Salt⁽²⁹⁾, our finding that all cereal categories were overserved suggests there is considerable potential for consumers to be unknowingly consuming more salt at breakfast than they are perhaps aware of.

Despite the negative consequences of overserving highlighted above, there is genuine potential for the overserving finding in our study to be exploited for public health benefit. Overserving BC may contribute positively to an individual’s health by increasing the consumption of carbohydrates, fibre and protein, in addition to vitamins and minerals that the cereal product was fortified with. For example, the larger portions seen in SW, granola and oat categories – while leading to higher amounts of energy and sugar in the present study – could be used to increase intake of fibre, protein and micronutrients if such products were reformulated correctly. Considering fibre as an example, current UK intakes are woefully inadequate with the majority of every age group not meeting the revised target of 30 g/d⁽⁹⁾. Indeed, recent analyses have shown the importance of higher cereal fibre intake on markers of cardiovascular health such as lower waist-hip ratio^{(Reference Barrett, Amoutzopoulos and Batterham48)}, in addition to being inversely associated with the prevalence of diverticular disease^{(Reference Aune, Sen and Norat49)} and inflammation^{(Reference Sang, Idehen and Zhao50)}. Thus, there is scope to utilise the overserving effect seen in our study to improve the population’s intake of nutrients such as fibre; however, this requires input from manufacturers to reformulate products accordingly, reducing energy density of BC and simultaneously lowering salt and sugar.

There are important limitations to our study that warrant discussion. Our study population was well-educated with 30 % of participants having achieved an undergraduate degree or equivalent. A previous study has demonstrated that educational level influences nutritional knowledge^{(Reference Hendrie, Coveney and Cox51)}, with additional work suggesting increased nutritional knowledge positively influences dietary intake, usually of fruits and vegetables^{(Reference Spronk, Kullen and Burdon52)}, and is associated with a higher frequency of breakfast consumption^{(Reference Matsumoto, Ishige and Sakamoto53)}. Thus, further studies should be performed in more diverse socio-economic groups. Moreover, 68 % of participants were classified as students. Studies have acknowledged that some students have a less healthful lifestyle than the general population^{(Reference Sprake, Russell and Cecil54)}. The time of day that participants completed the study was not controlled, due to the project’s time constraints. This meant that the researchers could not control the satiety and hunger of the participants, which may have influenced portion size. Future studies are required to determine how the portion size awareness of BC differs in a fasted state, which is likely when individuals would serve BC. It is also important to note that while our participants were familiar with all cereals used in the study, they were not required to habitually consume each type of cereal. Stipulating this as a requirement for the study would have negatively impacted on recruitment and also introduced more variation by removing the repeated measures aspect of the study design. Lastly, it must be recognised that the cereals analysed in our study do not fully represent the variety seen within each type of BC category.

In conclusion, the present study contributes to existing literature that consumers have a tendency to overconsume BC. However, we show that this is not uniform across different cereal types and the degree of overserving varies with the type of BC used. Granola, the most energy dense BC, was associated with the least portion size awareness and hence greatest degree of overserving. This lack of portion size awareness may impact on nutritional health if routinely overserved, due to the high energy, fat and sugar content. The present study supports previous work showing that denser foods are associated with reduced portion size awareness as a result of the dissociation between volume, weight and energy content. Government policy should encourage manufacturers to update the RSS and standardise portion size across cereals, as well as acknowledge differences for adults and children. Product reformulation should be encouraged to positively exploit the overserving seen in our study.