Meal patterns, including intermittent fasting – a scoping review for Nordic Nutrition Recommendations 2023

Introduction ‘Meal patterns’ refers to eating frequency, snacking, regularity, and timing. Here also including intermittent fasting. The effect of meal patterns on health is inconsistent and when updating the Nordic Nutrition Recommendations 2023 (NNR2023), summarizing the evidence is important. Aims To describe the evidence for the role of meal patterns on bodyweight, body composition, and cardiovascular risk factors (i.e. blood pressure and lipid- and glukose metabolism) in healthy people living with normal weight, overweight, or obesity in all age groups. Methods An initial search in PubMed found 481 reviews, of which 104 were identified based on titles. Of these, 47 were excluded based on title and abstracts. Of the remaining 57 reviews, 16 were included reporting search terms and inclusion/exclusion criteria. In addition, 8 reviews from reference list or known by authors were included. In total, 24 reviews were relevant. Cochrane Library was searched with no results. Results All reviews were rated low or critically low (AMSTAR 2). No consistent findings on eating frequency and body weight or composition were found in children/adolescents or adults. In snacking, mixed results were found, although among adults, some consistent results showed positive associations between snacking and body weight. In regularity, breakfast skipping showed mixed results in children/adolescents on body weight and composition. Among adults, randomized controlled trials on breakfast skipping showed a minor impact on improved weight loss. In prospective studies on timing, lower energy intake during late afternoon/evening was related to less body weight. Intermittent fasting reduced body weight but was not superior to continuous energy restrictions. Cardiovascular risk factors were assessed in a minority of the reviews, and despite some beneficial effects, the evidence was limited. Conclusion Given the overall low to critically low quality of the reviews, the evidence is limited and inconclusive. No consistent results providing evidence for setting recommendations for meal patterns were shown. In this regard, meal patterns may vary within the context of an energy balanced and nutritionally adequate diet.


M
eal patterns refers to eating frequency or occasions, snacking, timing, and regularity.Regularity may include time between meals, day-to-day variation, or skipping meals.Meal patterns varies between individuals according to age, culture, and traditions.For instance, meal patterns reported from the European Popular scientific summary • The term meal patterns refers to eating frequency, snacking, regularity, and timing of meals.
• Meal patterns in the Nordic countries can be described as regular, and most people eat breakfast.
• Snacking may be associated with higher body weight in adults, while lower late afternoon/evening energy intake has been associated with lower body weight or weight loss.• Intermittent fasting has been shown to reduce body weight similar to continuous energy restriction.
• No causal relationship between meal patterns including intermittent fasting and health related outcomes has been established.Mette Svendsen and Heléne Bertéus Forslund Prospective Investigation into Cancer and Nutrition (EPIC) calibration study showed a north-south difference in frequencies of eating occasions and timing.The participants from the Nordic and Central Europe reported more eating occasions compared to southern Europe.Europeans in the north reported about 1/3 of total energy intake from snacking.Furthermore, northern Europe had a higher proportion of energy eaten later in the day compared to the Mediterranean participants, although the latter having later timing of meals (1,2).However, no clear associations between meal timing and body mass index (BMI), lipids, or type 2 diabetes mellitus were found (2).The research on meal patterns in relation to health outcomes has shown divergent results.One reason is the lack of consensus about the definitions of eating occasions.Numerous definitions have been proposed by investigators, although the terms 'meals' and 'snacks or snacking' are commonly used to describe eating occasions (3,4).'Meals' generally refers to main meals (breakfast, lunch, and dinner) and 'snacks or snacking' to in-between meal eating or drinking.Whether only drinks with or without energy are included or not in eating occasions or snacks varies among studies (3,4).In addition, the distinction between eating occasions can be set by the researcher or by the respondents themselves.This makes the interpretations related to outcomes even more difficult.
Another reason for divergent results in meal pattern research is methodological problems, such as various classification of eating occasions (i.e.cultural, time, preparation, energy, or nutrient content) and different time space when separating eating occasions, as well as underreporting.For example, Murakami and Livingstone showed that the association between the number of eating occasions and BMI and waist circumference changed by different definitions.By using a definition taking energy content of each eating occasion into account as well as under-reporting, inverse associations changed to positive associations (5).Hence, using different time space when separating eating occasions is also crucial.
Meal patterns may have an effect on energy balance, that is energy intake and energy expenditure.Having a high or a low eating frequency can alter energy intake as well as appetite or satiety.However, eating frequency has a minor impact on energy expenditure under iso-caloric conditions.In this regard, meal patterns including timing may have an impact on energy intake, homeostatic (6), and hedonic (7) body weight regulation as well as lipid and glucose metabolism (6,7) in the fasting and post-prandial state and by that increase risk for cardiovascular disease.In addition, internal factors such as circadian rhythms may also have an impact on eating patterns and health-related outcomes.
Intermittent fasting has been included in the present scoping review as a result of a public hearing process for the Nordic Nutrition Recommendations 2023 (NNR2023) project (8).Fasting has traditionally been used for religious purposes and for general health reasons (9).However, the increased popularity of intermittent fasting during later years is mostly related to the effect on weight loss.Different variants of intermittent fasting exist (Table 1), although the common strategy is to restrict eating in specific periods during the day or week.There is scarce knowledge about the effects of intermittent fasting compared to continuous energy restrictions when similar energy restrictions are compared.
The aim of this scoping review was to describe the evidence for the role of meal patterns (eating frequencies, snacking, breakfast skipping, timing, and intermittent fasting) on body weight, body composition, and cardiovascular risk factors (i.e.blood pressure and lipid-and glucose metabolism) in healthy people living with normal weight, overweight, or obesity in all age groups (Box 1).

Methods
After an initial search conducted by the NNR2023 Committee, we expanded the search terms and otherwise followed the methodology published previously (10).
PubMed was searched on 07 May 2021 with the following search terms: "meal pattern*"[Title/Abstract] OR "meal frequenc*"[Title/Abstract] OR "eating frequenc*"[Title/ Abstract] OR "meal tim*"[Title] OR "snack*"[ Title/ Abstract] OR "meal skipping"[ Title/Abstract] OR The initial search identified 481 publications.Both authors read titles, of which 104 reviews were identified.Of these, 47 were excluded based on title and abstracts.Of the remaining 57 reviews, 16 were included reporting search terms and inclusion/exclusion criteria.In all, 41 reviews were excluded as this information was not provided, or outcomes were irrelevant.Publications conducted with normal weight, overweight, and obesity were included, whereas studies in special groups (i.e.patients, religious groups, and shift workers) were excluded.In the updated search on 01 October 2021, six publications were found, but no further studies were included because lack of information about search methods or irrelevant outcomes.In addition, four publications from reference lists and four publications known by authors were included.In total, 24 reviews were relevant for setting guidelines.These were extracted according to data on first author, number of publications, study design, population, intervention/exposure, comparator, dietary method (if relevant), outcomes, and AMSTAR 2 grading in summary tables.When differences were discovered, we discussed and agreed upon the final tables.
Cochrane Library was searched on 31 December 2021 with no results.No qualified systematic reviews, as defined by the NNR2023 project, were available.
In Finnish preschool children aged 3-6 years (N = 557), 88-100% of the days recorded (1-4 days) included breakfast, lunch, afternoon snack, dinner, and evening snack; of which, the lunch and dinner meals contributed most to the total energy intake (11).Regular breakfast eating was also reported among 6-year-old Icelandic children (N = 4,360) (12).UNGKOST 2015 in Norwegian children and adolescence aged 9 and 13 years showed that ≥90% of the 9 years old children ate breakfast, lunch, and dinner every day.Among the 13-year-old adolescents, daily eating of breakfast, lunch, and dinner was reported by 81, 71, and 92%, respectively.A majority of all participants did also have an evening eating occasion.Among the teenagers, girls were skipping meals more often than boys (13).Among Swedish adolescents (N = 3,099, aged 11-18 years), it was shown that about 22% of energy intake was in-between meals (14).
Among Norwegian adults in Norkost 3, meal patterns were assessed by two 24-h recalls and showed that on both days, 92% of participants consumed breakfast, 59% of participants consumed lunch, 89% of participants consumed dinner, and 33% of participants consumed an evening meal.The mean number of snacks including drinks (all eating or drinking occasions ≥50 kJ) was 1.6 per day for men and 1.9 per day for women.Moreover, 93% of men and 97% of women consumed at least one snack.In the meal pattern with three main meals, the snacks contributed to 29% of energy intake in men and 31% in women (15).In the FinDiet 2017 study, it was shown that the number of eating or drinking occasions was, on average, seven times per day, and almost all participants ate breakfast.The main meals contributed to 60% of the daily total energy intake, and most of the energy was eaten at lunch time 11-12 o´clock and at dinner about 17 o´clock (16).In addition to the national dietary surveys cited above, meal patterns were assessed in a cross-sectional, questionnaire-based internet survey with stratified random samples of Danish, Finnish, Norwegian, and Swedish adults in the age of 15-80 years (N = 7,531, completion rate 9-13%).This survey showed that most of the • This paper is one of many scoping reviews commissioned as part of the Nordic Nutrition Recommendations 2023 (NNR2023) project (8) • The papers are included in the extended NNR2023 report, but, for transparency, these scoping reviews are also published in Food & Nutrition Research • The scoping reviews have been peer reviewed by independent experts in the research field according to the standard procedures of the journal • The scoping reviews have also been subjected to public consultations (see report to be published by the NNR2023 project) • The NNR2023 committee has served as the editorial board • While these papers are a main fundament, the NNR2023 committee has the sole responsibility for setting dietary reference values in the NNR2023 project Mette Svendsen and Heléne Bertéus Forslund respondents in all the Nordic countries reported an eating frequency of 4-5 or more meals (not defined) per day.A regular eating pattern was seen with the majority eating breakfast, lunch, and dinner (17).

Meal patterns and health outcomes relevant for Nordic and Baltic countries
In Nordic and Baltic countries, about 17-44% of total deaths are caused by cardiovascular diseases among all ages.High blood pressure, low density lipoprotein (LDL) cholesterol, blood glucose, and high BMI are important risk factors for cardiovascular diseases in men and women (18).Furthermore, metabolic syndrome, that is abdominal obesity, increased blood pressure, impaired glucose metabolism, and/or dyslipidemia, increases risk for cardiovascular diseases (19).

Evidence for setting guidelines for meal patterns on health-related outcomes
We categorized meal patterns into eating frequency, snacking, skipping breakfast, timing, and intermittent fasting.Most of the reviews included cross-sectional studies, not relevant for setting guidelines.Therefore, we focused on the prospective studies and randomized clinical trials (RCTs) included in the reviews when referring the evidence below.

Eating frequency
At total of seven reviews (20)(21)(22)(23)(24)(25)(26) were identified (AMSTAR 2: critically low for all), and no consistent findings with regard to eating frequency on body weight and body composition were found (Table 2).Among children and adolescents, mixed results were seen in the prospective studies included in the reviews (20,24,25).However, a meta-analysis of 11 studies (N = 18,849 boys and girls, aged 2-19 years) showed that the highest category of eating frequency compared with the lowest was associated with a beneficial effect on body weight, especially in boys (not significant in girls when analyses were divided by gender), but the included studies were mostly cross-sectional, not relevant for setting guidelines (26).In adults, the reviews that included prospective studies (20,21,24,25) and RCTs (20,22,23) showed no consistent findings, with almost equal associations in different directions.For instance, in the review by Kant, it was reported from a prospective study (n > 7,000 men and women) followed for 10 years that eating frequency had no detrimental effect on body weight when controlling for energy intake, but in another prospective study (n > 20,000 men), also followed for 10 years, a higher risk of weight gain with higher eating frequency (P = 0.001) was seen when energy intake was not adjusted for (24).An important notion with regard to some of the RCTs included in the reviews is about the eating frequencies that were compared and the duration of the studies.For instance, in the review by Garcidueñas-Fimbres (20), two of the six RCTs compared two low eating frequencies (one meal eaten during early evening versus three main meals per day), and the duration time of the RCTs ranged from 12 h to 18 weeks (20).Cardiovascular risk factors were reported in three reviews (20,22,25).Results from RCTs showed beneficial effects on glucose (20), and single arm CTs showed some beneficial associations in total-and LDL-cholesterol (22) with higher eating frequency (20).No association between eating frequency and metabolic syndrome was reported either in children (20) or adults (25).However, due to the overall critically low quality of the reviews, there is no evidence for setting a general guideline for eating frequency according to health-related outcomes.

Snacking
We identified three reviews (25,27,28) with AMSTAR 2 critically low for all (Table 3).Generally, in the reviews, snacking was defined as eating or drinking between meals.
In children and adolescents, most of the publications included in the reviews were cross-sectional, with contradicting or null results.Mixed results were also seen in the prospective publications (25,27,28).Among adults, some consistency was seen as both the cross-sectional and prospective publications showed positive associations between snacking and body weight (25).None of the reviews reported cardiovascular outcomes.In most of the individual publications reported in the reviews, snacking was assessed by a simple question, 'yes or no', that did not indicate the number of snacks or the energy and nutrient content of the snack (Table 3).This makes interpretation difficult but indicates that among adults snacking may increase body weight.However, data on snacking quality are mostly lacking.In summary, there is no evidence for setting a general guideline about snacking according to health-related outcomes.

Breakfast skipping
A total of seven reviews were identified on breakfast eating or skipping (25,(29)(30)(31)(32)(33)(34) with AMSTAR 2 -low (29,30,33) and AMSTAR 2 -critically low (25,31,32,34).Body composition and cardiovascular risk factors were assessed in a minority of publications included in the reviews (Table 4).In children and adolescents, the findings were mixed in the prospective publications reported, but the reviews showed some consistency in cross-sectional publications in that skipping breakfast was associated with increased prevalence of overweight and obesity.However, as noted earlier, cross-sectional studies are not relevant for setting guidelines.In adults, meta-analyses (including prospective studies) showed skipping breakfast to be associated with about 20% increased risk for heart disease (31)     mellitus (34).No definition of the breakfast according to time of the day or nutrient content was reported.Hence, given the overall critically low quality of the reviews, these findings have to be interpreted with caution.It is difficult to know from the included publications how often breakfast has to be consumed to be preventive, and information of breakfast quality or timing is lacking.With regard to body weight, in the meta-analyses by Bonnet et al. (30) and Sievert et al. (33) including only RCTs in adults (N~450 in each of the meta-analyses), the effect of eating or skipping breakfast was investigated.Both reviews, with overlapping studies included, showed skipping breakfast to be associated with about 0.5 kg weight loss in participants with normal weight, overweight, and obesity during 2-16 weeks (30,33).The breakfasts varied among studies, but mostly participants prolonged their overnight fast and had lunch about noon in the skipping breakfast groups.Interpreting these results may indicate that delaying the first meal of the day seems not detrimental for control of body weight in adults.
Cardiovascular risk factors were reported in three reviews (25,30,32).Among children, cross-sectional data showed that skipping breakfast was associated with impaired effects on lipid profile, blood pressure, and insulin (32).In adults, increased LDL-cholesterol was reported (30) in RCTs, but no associations between breakfast skipping and metabolic syndrome were seen in one cross-sectional study (25).
Overall, the evidence with regard to breakfast skipping on health-related outcomes is limited, and no general guideline about skipping or eating breakfast can be set as long as energy and nutrient needs are met.

Timing
We identified four reviews on meal timing (35)(36)(37)(38) with AMSTAR 2 low (35,36) and critically low (35,36).Overall, the evidence is limited (Table 5).However, the reviews showed consistency in that lower energy intake during late afternoon or evening was related to lower body weight or weight loss (in prospective studies) (35)(36)(37).However, a meta-analysis of RCTs did not show significant associations between late eating and body weight (36).To be noted is that the RCTs included in the meta-analysis were mostly of short duration and hypo-caloric, with low generalizability for free-living conditions.Cardiovascular risk factors were reported in two reviews (35,38), showing reduction in markers for glucose metabolism with lower energy intake later in the day (35,38).This may indicate that a lower energy intake during the evening may be beneficial for body weight and glucose metabolism.However, the evidence of meal timing on health-related outcomes is limited, and no guideline can be set.

Intermittent fasting
We identified six reviews that included participants with normal weight, overweight, and obesity (22,(39)(40)(41)(42)(43), AMSTAR 2 -critically low for all.Generally, the reviews included studies with different variants of intermittent fasting (Table 1) and showed that intermittent fasting reduced body weight (Table 6).The reduction in body weight and fat mass was about 3-14% during 3-14 weeks (22,43) but not superior to continuous energy restrictions.Importantly, the reviews included publications not comparing similar energy restrictions between intermittent fasting and continuous energy restrictions.Limited and mixed results were seen with regard to effects on fat free mass.Cardiovascular risk factors were reported in four reviews (22,39,40,43).Some beneficial effects were seen in triglycerides, high density lipoprotein (HDL) cholesterol, and glucose (22,43), as is expected due to weight reduction.Recent reviews (not from search) including only participants with overweight or obesity (44,45), with AMSTAR 2 critically low (44) and low (45), respectively, showed the intermittent fasting diet to improve body weight and lipid profile during 1 year of follow-up (44,45).However, the meta-analysis by He et al. (45), including 11 RCTs (N = 850 participants) with similar energy restrictions between the intermittent or continuous energy restricted groups, showed a marginally greater weight loss (weight mean difference: −0.95 kg [95% CI: −1.63, −0.27], P = 0.01) in the intermittent groups.Subgroup analysis showed this to be driven by short time studies (n = 300) with duration of 2-3 months (weight mean difference: −1.66 kg [−2.44, −0.27], P = 0.0001), with no difference between groups in longer term studies (n = 550) of 6-12 months (weight mean difference: −0.09 kg [−1.05, 0.87] P = 0.85).No significant findings in waist circumference, blood lipids, blood pressure, glucose, or HbA1c were seen (45).A review by Guerrero et al. including 17 RCTs (N = 1,091 participants) showed good adherence that may decline over time (reported in about 60% of the included studies) (44).
To summarize, the evidence is limited, and no guideline about intermittent fasting on health-related outcomes can be set for the general population.However, among people living with overweight and obesity, intermittent energy restriction seems to be equal to continuous energy restriction for medical treatment.

Mechanisms
The human night and day cycle is of importance for timing of meals.Lately, the impact of circadian rhythm and chrono-nutrition on health has been discussed.Circadian rhythms are biological rhythms that follow a 24-h cycle, including nutritional and metabolic processes.This suggests that we may consider not only what we eat but also (page number not for citation purpose) Meal patterns, including intermittent fasting    when we eat, as the timing of energy intake could affect chronic disease risk (46,47).Meal patterns may influence homeostatic regulation for body weight, glucose, and lipid metabolism through how eating frequency, regularity, and timing are synchronized according to circadian rhythms.Chrono-nutrition is an emerging field of research involving meal patterns, circadian rhythms, body weight regulation, and metabolic health (7).Circadian rhythms promote synchronicity of biological processes during approximately 24 h, and, among others, these rhythms signal sleep and awakening, feeding, and fasting (48).
The endogenous circadian system is primarily controlled by an autonomous master clock in the suprachiasmatic nucleus (SCN) of the hypothalamus, which is synchronized by light and entrains secondary clocks in the brain and most peripheral tissues of the body.These secondary clocks are also entrained by environmental cues and behaviors, such as eating and sleeping (6).When these behaviors fail to align with the master clock (SCN), mismatch may occur.This is most markedly seen among shift-workers (49).However, variability in sleep and eating pattern during the week, called social and eating jetlag (i.e.differences in time of sleeping and eating on free days and workdays) (50), may also have an influence on the synchrony of the biological processes related to body weight control and metabolism (7,49,51).Research on circadian rhythms in lipid and glucose metabolism and insulin sensitivity may support eating more of the total energy intake earlier in the day (7).
A factor that also may influence the effects of late eating is chrono-type (52), a measure of the individual preferences for morning or evening as can be influenced by genetic and environmental factors (50).Cross sectional studies have shown that chrono-types vary by age, with more morning types in young childhood (52), gradually more evening type during adolescence (53), and back to more morning type during older ages (54).However, this is not confirmed in prospective research but may indicate that differences in chrono-types (i.e.morning or evening) could be of importance to consider regarding the effects of meal patterns on health.

Data gaps for future research
Different meal patterns may be associated with diet quality.In the present scoping review, diet quality was assessed in three reviews (20,29,32).In children and adolescence, some beneficial effect of intervention programs for eating breakfast was seen (29), but otherwise, the evidence was divergent.Among adults, meal frequency was positively associated with diet quality.This was not seen in a review by Leech et al. (55), who found a consistent inverse association between skipping breakfast and diet quality, but no associations between other meal patterns.Thus, the evidence with regard to the effect of meal patterns on diet quality is inconclusive and needs further investigation.
Research is needed to investigate the influence of meal patterns, especially with regard to meal timing, including chrono-nutrition, on body weight and composition, metabolism, and cardiovascular disease risk factors in all age groups.In this regard, more research is needed on the influence of eating breakfast on cardiovascular health.Results from the National Health and Nutrition Examination Survey III (56) confirmed the findings from the meta-analysis by Takagi et al. (31), but causality has not yet been proven.Thus, eating breakfast may be an epiphenomenon linked to dietary intake and other lifestyle factors (i.e.sleep and physical activity) (57).Moreover, research on how intake at one eating occasion influences the next should also be considered.Well-designed prospective and RCTs are needed, also including statistical analyses taking the meal patterns of the whole day into consideration with regard to energy and nutrient intake.To better understand the impact of meal patterns on health, especially body weight and metabolism in the light of meal timing, these issues need to be addressed in national surveys.Focus on how eating frequency, snacking, and timing contribute to energy intake and body weight in all ages, from children to elderly, during pregnancy and lactation, as well as metabolic risk factors should be emphasized in other well-designed trials.In addition, more research is needed on how culture and socioeconomic factors, gender, age, and different kind of households are associated with meal patterns and health.In order to reduce methodological problems regarding meal patterns, researchers ought to describe their definitions and classifications adequately.Different definitions also exist in the Nordic dietary surveys.A starting point could be to standardize this in the Nordic and Baltic countries.Consensus on the definitions of various eating occasions is crucial to make reliable comparisons and conclusions.
With regard to intermittent fasting, a higher loss of lean mass during intermittent fasting compared to continuous energy restrictions has been reported (45,58).A lot of variants for intermittent fasting exists, and less is known about the dietary quality during fasting days.However, lower dietary quality during intermittent fasting compared to continuous energy restriction has been reported (59).Further research is needed on intermittent fasting to clarify the effect of diet quality and if lean mass/muscle mass can be preserved by optimal intake of protein including plant-based protein and physical activity.

Limitations
All the reviews got AMSTAR 2 in the lowest categories.This was mainly due to inadequacy of the literature search (~2/3 of the reviews), risk of bias in the individual studies included (~½ of the reviews) and not considering bias when interpreting the results (~2/3 of the reviews).In addition, none of the reviews presented a list of excluded studies.Therefore, the reviews cannot provide an accurate and comprehensive summary of the available evidence.Furthermore, we selected reviews including healthy participant with normal weight, overweight, and obesity in all ages.However, meal patterns in the elderly were only reported in a minority of the reviews, and it was not possible to draw any conclusions.In addition, our search did not identify reviews of regularity other than skipping breakfast (i.e.no studies on time between meals or day-to-day variation).Moreover, due to the methods of using reviews, we may have missed newer studies not included in reviews.

Conclusion
No causal relationship between meal patterns including intermittent fasting and health-related outcomes was found.The evidence is limited and inconclusive due to low to critically low quality of the reviews, including mostly cross-sectional studies.The research on meal patterns lacks consistency mainly because of no common definitions of meals and snacks or not taking energy intake into consideration.Thus, with the evidence at this time point, we cannot give any advice on meal patterns based on health outcomes.In this regard, meal patterns may vary within the context of an energy balanced and nutritionally adequate diet.