What this is
- This research investigates the relationship between and , as well as body mass index (BMI), in Swedish adults and adolescents.
- It utilizes data from two national dietary surveys, Riksmaten Adults 2010-11 and Riksmaten Adolescents 2016-17, to analyze meal frequency, breakfast skipping, and timing of energy intake.
- Findings indicate that vary by weight status, with implications for public health strategies aimed at improving dietary habits.
Essence
- Higher meal frequency and breakfast consumption are linked to better and lower obesity risk in Swedish adolescents and adults. Skipping breakfast and low eating frequency correlate with poorer .
Key takeaways
- Adults and adolescents with obesity reported lower meal frequency and higher rates of breakfast skipping. This trend suggests that are influenced by weight status.
- A high meal frequency (OR 0.44) and late energy distribution (OR 0.70) are associated with a decreased risk for overweight or obesity in adolescents. This indicates that meal timing and frequency can impact weight management.
- A low eating frequency and breakfast skipping are inversely related to . This finding emphasizes the importance of regular meals and breakfast for maintaining a healthy diet.
Caveats
- The study's cross-sectional design limits causal inferences between and or obesity. Longitudinal studies are needed to clarify these relationships.
- Self-reported data on weight and dietary intake may introduce misclassification and bias, particularly among individuals with obesity who tend to under-report their intake.
Definitions
- meal patterns: The distribution of food intake across the day, including timing, frequency, and regularity.
- diet quality: An assessment of food intake adherence to dietary guidelines, reflecting nutritional adequacy and variety.
AI simplified
Introduction
The relationship between temporal meal patterns, body weight, and cardiometabolic health has gained attention in recent years [1, 2]. Temporal meal patterns describe how food intake is distributed across the day in terms of timing, frequency, and regularity [3] (hereafter referred to as meal patterns). This area of research falls within the concept of chrononutrition, which focuses on how the timing of food intake interacts with the body’s circadian rhythms and thus influence metabolic processes [4, 5]. Evidence suggests that the associations between meal patterns and metabolic health may be caused by the interaction between ingested foods, circadian rhythms, and metabolism, which in turn affects how the body utilises nutrients at different time points of the day and how they influence cardiometabolic functions [5].These potential mechanisms have been described elsewhere [4]. Obesity and diet quality are two other factors that influence cardiometabolic health [6–8], and meal patterns have been associated with both [2, 9].
Given that meal patterns may influence metabolic processes and risk of disease via circadian disruption it is important from a public health perspective to understand their impact on diet quality and obesity. These insights may be used for developing preventive interventions that promote healthier eating habits across diverse populations. Meal patterns, such as meal frequency and lunch- and snack consumption, differ between countries, but few countries provide general recommendations regarding meal patterns, which might be relevant if they are associated with diet- and health-related outcomes [10, 11]. Also, about 15% of the Swedish population between 10 and 80 years skip breakfast regularly, with the highest prevalence in adolescents and younger adults where almost one-third skip breakfast [12]. This has also been shown in other national surveys, such as the US where 20–30% of young adults regularly skip breakfast [13] as well as in observational research where daily breakfast consumption in adolescents ranged between 38 to 73% across Europe and North America [14].
Despite the growing interest, previous research on meal patterns and health-related outcomes has yielded conflicting results, highlighting the need for further investigation in larger, representative populations [15]. Although variables such as meal frequency, breakfast consumption, and late-day food intake have been extensively studied, findings remain inconsistent and methodological heterogeneity and varying definitions complicates direct comparisons across studies [3]. In a previous study, we explored meal patterns in the current study populations and their associations with nutrient intakes, where a high eating frequency and eating breakfast were associated with a higher absolute intake of some key nutrients such as whole grains, vitamin D and folate [12]. Therefore, it is of interest to investigate whether meal patterns are associated with diet quality and risk factors, such as obesity. Moreover, nationally representative data from Sweden on these associations are currently lacking, limiting our understanding of how meal patterns relate to public health in a Swedish context. This study aimed to evaluate the meal patterns variables: meal frequency, breakfast skipping and late versus early energy distribution and their respective association with diet quality, mirrored as the adherence to the Swedish food-based dietary guidelines, and overweight and obesity in two representative samples of the Swedish population. To better understand these relationships, the present study can provide valuable insights for developing public health strategies aimed at improving meal patterns in the population and thus reducing the burden of obesity-related diseases.
Method
Population
This study was completed as a secondary analysis of data collected in two cross-sectional, national dietary surveys by the Swedish Food Agency (SFA), Riksmaten Adults 2010–11 (n = 1796) and Riksmaten Adolescents 2016–17(n = 2967). In total, this provided an age span of 10 to 80 years, representative of the Swedish population. The data from the two surveys were analysed separately.
Riksmaten adolescents
Riksmaten Adolescents 2016–17 was a school-based dietary survey completed in grades 5, 8, and 11 [16]. Statistics Sweden provided a sample of 619 schools to obtain a representative sample. Out of these, 131 schools (22%) agreed to participate, and in total 5145 participants were recruited and invited from one or two chosen classes at each school. Eligible to participate were all students in each class. A total of 3477 (68%) participated in some part of the study, and 2967 (57%) completed all three days of food registrations, which was an inclusion criterion for this study. Participants registered their food for three days using the web-based, validated RiksmatenFlex method, previously described elsewhere [17]. Staff from the SFA visited the schools and guided the participants in how to complete the dietary assessment and measured their height and weight. The first day of the dietary assessment was performed as a 24 h recall and was always the day before the school visit. The second day was a prospective food record of the day of the school visit. The third day was randomly assigned to take place 3–9 days after day 2. All days of the week were represented proportionally. The web-based method is linked to a food composition database specifically adapted for the survey (The Swedish food composition database, version Riksmaten Adolescents 2016–17), which enabled automatic estimation of energy and nutrient intakes. The participants and their parents answered web questionnaires on background characteristics and food intake.
Riksmaten adults
The Riksmaten Adults 2010–11 survey was a national dietary survey on Swedish adults between 18 and 80 years. Participants were invited to register their food for 4 consecutive days using the web-based, validated Riksmaten method (The Swedish food composition database, version Riksmaten Adults 2010–11) [18, 19], and fill out questionnaires [20]. Further information on the study procedure has been described elsewhere [20, 21]. A representative sample based on age, sex, and living area was drawn by Statistics Sweden. A total of 5003 people were invited, of whom 1796(36%) completed the food diaries. In order to participate, participants needed to speak Swedish. The recruitment was evenly distributed throughout the year. Almost all, 98% of the participants completed all four days, but all participants with approved registrations for at least two days were included in this study. All days of the week were equally distributed in the dataset, as the starting day for each participant was randomly selected.
The food composition database used in the adult survey did not include free sugars. For the present study, intake of free sugars was estimated for adults by utilising a recent version of the Swedish Food composition database (Version 20,240,529). Further, some manual estimation of the free sugars content of some foods was needed since they had been removed in the food database used for the estimation. In total 1908 foods were included in the survey (1699 used) of which 212 foods needed manual estimation of free sugars. Estimation of free sugars was completed with the same method as in the rest of the food composition database [22].
Background characteristics
Data on age and sex were collected from the class lists for adolescents and for adults, data on age and sex were collected from Statistics Sweden. Age was categorized into the following age groups 18–30, 31–44, 45–64 and 65–80 years for adults, and school grades 5(11–12 years), 8 (14–15 years) and 11(17–18 years) were used for the age categories of adolescents. Physical activity levels were assessed using self-reported questionnaire data. Adolescent participants answered one question about leisure-time physical activity during the previous week. Physical activity level (PAL) for adult participants was computed by combining responses from two questions—one assessing physical activity during work hours and the other evaluating non-work hour related physical activity [23]. Smoking status was collected from the questionnaires (excluding the youngest age group). Education was self-reported in the questionnaires and defined as the participants’ highest education level for adults, and for adolescents the highest education of the parents in the household. In both surveys, some participants provided blood and urine samples, but these were not used in this study.
Anthropometric measures
Weight and height for adolescents were measured by field staff from the SFA using standardized methods [16]. For adults, weight and height were self-reported in the questionnaires. For adults underweight was defined as body mass index (BMI) < 18.5 kg/m2, Normal weight BMI 18.5–24.9 kg/m2, Overweight BMI 25.0–29.9 kg/m2, and obesity BMI > 30.0 kg/m2. ISO-BMI cut-offs weight status were used for adolescents (IOTF) [24]. Associations between meal patterns and obesity in the adult population were explored. Among adolescents, where the obesity prevalence was notably low (4%), relationships between meal patterns and both overweight and obesity were assessed to enhance statistical power.
Evaluation of energy intake
Under- and overreporting of energy intake were evaluated using the ratio of energy intake (EI) and resting energy expenditure (REE) with the method by Black [25]. REE was calculated based on weight and age [26]. The PAL used in the calculations were the group averages, 1.4 for adolescents and 1.67 for adults. The EI:REE cut-offs was calculated to 0.93–3.01 for adults and 0.83–2.35 for adolescents. Participants outside these cut-offs were identified as misreporters for sensitivity analysis purposes, but were not excluded from the samples.
Meal patterns
In the web-based dietary assessment, the participants reported their meals using predefined meal types (breakfast, lunch, dinner/evening meal, snack, other eating, drinks only) together with clock time. This information was used to define the participants’ meal patterns. In the present study, we examined meal frequency, breakfast skipping and late energy peak as meal pattern variables. Meal frequency was defined as the number of reported eating occasions per 24 h, with a minimum energy contribution of 50 kcal [3]. Meals only containing drinks were included in this definition. The time between two meals was at least 15 min for adults and one hour for adolescents due to different designs and functionalities of the web-based methods. The mean number of EO was calculated for each participant. Participants were also grouped based on their meal frequency into “3 or fewer meals”, “4 to 5 meals” and “6 or more meals”. Breakfast skipping was identified as participants who did not register the meal-type breakfast on at least one of the days, regardless of clock-time, provided that the breakfast tracked contributed at least 50 kcal. Both irregular breakfast skippers (skipped breakfast on at least one of the registration days) and complete breakfast skippers (no breakfast reported) were grouped into “Breakfast skippers” due to the limited number of cases in each subgroup, which precluded separate analyses.
The distribution of energy during the day was defined as early energy peak when having reported a greater amount of energy between 06.00–14.59 while having reported a greater amount of energy between 15.00–23.59 was called a late energy peak. These cut-offs were used to align with previous studies [9, 12, 27]. The energy intake in each time slot was calculated for all participants and this was calculated to a ratio[12]. A ratio < 1 indicates a late peak and > 1 an early peak. The mean ratio over the reported days per individual was used in the analyses.
Diet quality index
The dietary data was collected using a method that captures the current intake of individuals. Fish intake is an example of a less frequently consumed food group that can show significant day-to-day variation within individuals [32]. To align with nutritional recommendations, which focus on long-term needs (usual intake), current intake can be converted into usual intake using a statistical method. Multiple methods are available, and for this study the multiple source method (MSM) was used, which has been used in other large studies as well as by SFA [28, 33]. Therefore, before generating the score for fish, the intake of fish and shellfish was converted to the usual intake using the MSM method stratified by sex together with data from the questionnaire on fish consumption [33, 34].
| Dietary component | Calculation |
|---|---|
| Fruit and vegetables – at least 500 g per day (potatoes and legumes not included)* | Intake/500 |
| Red and processed meat – max 350 g per week* | 1 − ((intake − 350)/350) |
| Fiber—at least 2.5 g/MJ per day | Intake per MJ /2.5 |
| Whole grain—at least 75 g per 10 MJ | Intake per 10 MJ /75 |
| Fish and shellfish—2 to 3 times per week, average intake 45 g/d | Intake/45 |
| Polyunsaturated fat—at least 7.5 E% | E%/7.5 |
| Monounsaturated fat—at least 15 E% | E%/15 |
| Saturated fat—max 10E% | 1 − ((E% − 10)/10) |
| Added and free sugar—max 10E% * | 1 − ((E% − 10)/10) |
Statistics
Statistical analyses were performed using STATA version 18 (StataCorp, College Station, Texas, USA). The statistics were performed separately for the two surveys, adolescents and adults, because of differences in data collection procedures and sample size. The significance level was set at p < 0.05. Variables were tested for normality with Shapiro–Francia W′ test for normality. Differences in SHEIA25 scores, individual SHEIA score components and meal patterns between weight statuses were evaluated using t-test, Mann–Whitney U-test or Pearson’s x2 test.
A linear regression analysis was performed to explore the association between meal patterns (meal frequency, breakfast skipping and late energy peak) and diet quality. Separate models were performed for each meal pattern variable. A basic model was adjusted for age, and a multivariable model was adjusted for sex, age, physical activity, education level and smoking status (adults). The models were stratified by sex because both meal patterns and diet quality differed by sex. Since the schools could be considered as clusters for the adolescents, a mixed-effect regression model was used with school as a random intercept.
The association between meal patterns and weight was explored using multivariable logistic regression analysis with adjustments for age and sex as model 1 and model 2 further included diet quality, physical activity, smoking (adults) and education level. A mixed-effect logistic model with schools as random intercept were used for adolescents.
Sensitivity analyses were performed by excluding mis-reporters in the multivariable regression analyses on both diet quality and weight status.
Results
Logistic regression models for the association between meal patterns and obesity (adults) or overweight or obesity (adolescents) are presented as odds ratios. Model 1 was adjusted for age and sex. Model 2 was adjusted for age, sex, diet quality, physical activity, smoking (adults) and education level/parental education. Model 2 was for adolescents completed as a logistic mixed model with schools as random intercept
| Adolescents | Adults | |||||||||||||
|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|
| School year | Sex | All | Age group | Sex | All | |||||||||
| 5 | 8 | 11 | Girls | Boys | 18–30 | 31–44 | 45–64 | 65–80 | Women | Men | ||||
| n | 989 | 1012 | 966 | 1656 | 1311 | 2967 | 334 | 430 | 665 | 367 | 1005 | 791 | 1796 | |
| ± SD | 11.5 ± 0.39 | 14.5 ± 0.40 | 17.7 ± 0.64 | 14.6 ± 2.6 | 14.5 ± 2.6 | 14.6 ± 2.6 | 23.7 ± 3.6 | 37.9 ± 4.0 | 54.5 ± 5.8 | 70.4 ± 4.2 | 47.1 ± 16.7 | 49.0 ± 16.4 | 48.0 ± 16.6 | |
| Sex n (%) | Men | 455(46) | 455(45) | 401(42) | N/A | N/A | 1311(44) | 131(39) | 184(43) | 308(46) | 169(46) | N/A | N/A | 791(44) |
| Women | 534(54) | 557(55) | 565(58) | N/A | N/A | 1656(56) | 203(61) | 247(57) | 357(54) | 198(54) | N/A | N/A | 1005(56) | |
| Weight * status n(%) | Normal weight | 676(69) | 764(76) | 679(71) | 1187(72) | 932(72) | 2119(72) | 214(64) | 226(53) | 269(40) | 156(43) | 548(55) | 317(40) | 865(48) |
| Overweight | 182(19) | 141(14) | 173(18) | 280(17) | 216(17) | 496(17) | 55(17) | 115(27) | 255(38) | 147(40) | 255(25) | 317(40) | 572(32) | |
| Obesity | 35(4) | 30(3) | 54(6) | 58(4) | 61(5) | 119(4) | 55(17) | 80(19) | 138(21) | 61(17) | 177(18) | 157(20) | 334(19) | |
| Underweight | 80(8) | 74(7) | 54(6) | 118(7) | 90(7) | 208(7) | 9(3) | 7(2) | 3(0) | 3(1) | 22(2) | 0 | 22(1) | |
| Physical activity** n(%) | 1 | 173(18) | 226(22) | 303(32) | 408(25) | 294(23) | 702(24) | 1.69 | 1.66 | 1.66 | 1.68 | 1.67 | 1.67 | 1.67 |
| 2 | 222(23) | 166(17) | 208(22) | 387(23) | 209(16) | 596(20) | N/A | N/A | N/A | N/A | N/A | N/A | N/A | |
| 3 | 410(42) | 365(36) | 228(24) | 566(34) | 437(34) | 1003(34) | N/A | N/A | N/A | N/A | N/A | N/A | N/A | |
| 4 | 173(18) | 251(25) | 220(23) | 288(17) | 356(27) | 644(21.8) | N/A | N/A | N/A | N/A | N/A | N/A | N/A | |
| Misreporting | Under | 69 (7) | 100 (10) | 92 (10) | 145(9) | 116(9) | 261(9) | 74 (22) | 48 (11) | 106 (16) | 45 (12) | 146(15) | 127(16) | 273(15) |
| Over | 38 (4) | 42 (4) | 26 (3) | 42(3) | 63(5) | 106(4) | 2 (1) | 5 (1) | 2 (0) | 2 (1) | 4(0) | 7(1) | 11(1) | |
| Highest education/parent’s highest education | Primary | 41 (4) | 48 (5) | 35 (4) | 67(4) | 57(4) | 124(6) | 17 (5) | 15 (4) | 91 (14) | 136 (37) | 124(12) | 135(17) | 259(14) |
| High school | 303 (31) | 288 (29) | 351 (36) | 521(31) | 421(32) | 942(32) | 159 (48) | 160 (37) | 272 (41) | 100 (27) | 383(38) | 308(39) | 691(38) | |
| University | 609 (62) | 598 (59) | 515 (53) | 988(60) | 734(56) | 1722(58) | 126 (38) | 222 (52) | 259 (39) | 110 (30) | 441(44) | 276(35) | 717(40) | |
| None | 0 | 0 | 0 | 0 | 0 | 0 | 0 | 1 (0) | 0 | 2 (1) | 2(0) | 1(0) | 3(0) | |
| Missing | 36 (4) | 78 (8) | 65 (7) | 80(5) | 99(8) | 179(6) | 32 (10) | 32 (7) | 43 (7) | 18 (5) | 54(5) | 71(9) | 125(7) | |
| Smoking | Never | N/A | 941 (93) | 669 (70) | 914(82) | 697(82) | 1611(82) | 199 (66) | 266 (67) | 286 (46) | 155 (44) | 508(53) | 398(55) | 906(54) |
| Daily | N/A | 0 (0) | 50 (5) | 33(3) | 17(2) | 50(3) | 26 (9) | 26 (7) | 66 (11) | 26 (7) | 88(9) | 56(8) | 144(9) | |
| Former | N/A | 14 (1) | 50 (5) | 36(3) | 28(3) | 64(3) | 20 (7) | 75 (19) | 251 (40) | 157 (45) | 292(31) | 211(29) | 503(30) | |
| Rarely | N/A | 29 (3) | 168 (18) | 115(10) | 82(10) | 197(10) | 56 (19) | 32 (8) | 20 (3) | 15 (4) | 67(7) | 56(8) | 123(7) | |
| I don’t want to answer | N/A | 25 (3) | 23 (2) | 22(2) | 26(3) | 48(2) | N/A | N/A | N/A | N/A | N/A | N/A | N/A | |
| Adolescents | ||||||
|---|---|---|---|---|---|---|
| Underweighta | Normal weight | Overweight | Obesity | All adolescents | valueNormal weight vs. overweight or obesityP | |
| n | 208 | 2119 | 496 | 119 | 2967 | |
| Meal frequency mean ± sd | 4.4 ± 0.9 | 4.3 ± 0.9 | 4.0 ± 0.9 | 3.8 ± 0.9 | 4.2 ± 0.9 | < 0.001 |
| 3 or fewer n (%) | 40 (19) | 459(22) | 160 (32) | 58(49) | 725(24) | < 0.001 |
| 4 to 5 n (%) | 146 (70) | 1526(72) | 320 (65) | 59(50) | 2067(70) | < 0.001 |
| 6 or more n (%) | 22 (11) | 134(6) | 16 (3) | 2(2) | 175(6) | < 0.001 |
| Breakfast skipping n (%) | 30(14) | 409(19) | 112(23) | 39(33) | 591(20) | < 0.001 |
| Late energy peak n (%) | 84(49) | 823(39) | 156(31) | 35(29) | 1102(37) | < 0.001 |
| Total SHEIA25mean ± sdb | 5.56 ± 1.1 | 5.53 ± 1.0 | 5.50 ± 1.0 | 5.36 ± 1.0 | 5.52 ± 1.0 | 0.28 |
| Individual component score (0 to 1) | ||||||
| Fruit and vegetables | 0.46 ± 0.27 | 0.46 ± 0.27 | 0.42 ± 0.25 | 0.38 ± 0.22 | 0.45 ± 0.26 | 0.014 |
| Meat and processed meat | 0.48 ± 0.43 | 0.46 ± 0.44 | 0.46 ± 0.43 | 0.40 ± 0.42 | 0.46 ± 0.43 | 0.57 |
| Fiber | 0.79 ± 0.19 | 0.81 ± 0.17 | 0.79 ± 0.18 | 0.78 ± 0.18 | 0.80 ± 0.17 | 0.38 |
| Whole grains | 0.44 ± 0.33 | 0.44 ± 0.32 | 0.43 ± 0.33 | 0.37 ± 0.30 | 0.44 ± 0.32 | 0.15 |
| Fish | 0.53 ± 0.27 | 0.51 ± 0.29 | 0.50 ± 0.30 | 0.52 ± 0.30 | 0.51 ± 0.29 | 0.98 |
| Polyunsaturated fat | 0.61 ± 0.15 | 0.63 ± 0.16 | 0.64 ± 0.17 | 0.64 ± 0.17 | 0.63 ± 0.16 | 0.75 |
| Monounsaturated fat | 0.86 ± 0.12 | 0.86 ± 0.14 | 0.87 ± 0.14 | 0.89 ± 0.13 | 0.86 ± 0.14 | 0.07 |
| Saturated fat | 0.61 ± 0.25 | 0.62 ± 0.25 | 0.62 ± 0.26 | 0.59 ± 0.26 | 0.62 ± 0.25 | 0.82 |
| Added and free sugar | 0.77 ± 0.34 | 0.74 ± 0.34 | 0.77 ± 0.33 | 0.78 ± 0.34 | 0.75 ± 0.33 | 0.74 |
| Adults | ||||||
|---|---|---|---|---|---|---|
| Underweighta | Normal weight | Overweight | Obesity | All adults | valueNormal weight vs. obesityP | |
| n | 22 | 865 | 572 | 334 | 1796 | |
| Meal frequency mean ± sd | 4.8 ± 1.2 | 4.7 ± 1.1 | 4.6 ± 1.0 | 4.4 ± 1.2 | 4.6 ± 1.1 | < 0.001 |
| 3 or fewer n (%) | 3 (14) | 93 (11) | 74 (13) | 66 (20) | 237 (13) | < 0.001 |
| 4 to 5 n (%) | 13 (59) | 604 (70) | 391 (68) | 216 (65) | 1226 (68) | 0.09 |
| 6 or more n (%) | 6 (27) | 168 (19) | 107 (19) | 52 (16) | 333 (19) | 0.04 |
| Breakfast skipping n (%) | 3 (14) | 82 (9) | 45 (8) | 48 (14) | 179 (10) | 0.02 |
| Late energy peak n (%) | 9 (41) | 300 (35) | 188 (33) | 123 (37) | 622 (35) | 0.49 |
| Total SHEIA25mean ± sdb | 6.19 ± 1.2 | 6.37 ± 1.1 | 6.38 ± 1.1 | 6.02 ± 1.1 | 6.30 ± 1.1 | < 0.001 |
| Individual component score (0 to 1) | ||||||
| Fruit and vegetables | 0.57 ± 0.29 | 0.58 ± 0.28 | 0.56 ± 0.26 | 0.49 ± 0.27 | 0.56 ± 0.27 | < 0.001 |
| Meat and processed meat | 0.64 ± 0.40 | 0.52 ± 0.43 | 0.46 ± 0.43 | 0.39 ± 0.43 | 0.48 ± 0.43 | < 0.001 |
| Fiber | 0.87 ± 0.16 | 0.88 ± 0.15 | 0.89 ± 0.15 | 0.85 ± 0.17 | 0.88 ± 0.16 | < 0.001 |
| Whole grains | 0.48 ± 0.34 | 0.62 ± 0.33 | 0.62 ± 0.32 | 0.56 ± 0.32 | 0.61 ± 0.32 | 0.006 |
| Fish | 0.69 ± 0.30 | 0.72 ± 0.29 | 0.72 ± 0.30 | 0.65 ± 0.31 | 0.71 ± 0.29 | < 0.001 |
| Polyunsaturated fat | 0.74 ± 0.17 | 0.71 ± 0.19 | 0.72 ± 0.19 | 0.73 ± 0.19 | 0.72 ± 0.19 | 0.12 |
| Monounsaturated fat | 0.87 ± 0.14 | 0.83 ± 0.14 | 0.83 ± 0.15 | 0.85 ± 0.15 | 0.84 ± 0.14 | 0.002 |
| Saturated fat | 0.53 ± 0.34 | 0.66 ± 0.27 | 0.69 ± 0.26 | 0.63 ± 0.27 | 0.66 ± 0.27 | 0.11 |
| Added and free sugar | 0.79 ± 0.31 | 0.85 ± 0.26 | 0.89 ± 0.23 | 0.86 ± 0.26 | 0.86 ± 0.25 | 0.78 |
| Adolescents | Adults | ||||||||
|---|---|---|---|---|---|---|---|---|---|
| Coefficient | 95% CI | Coefficient | 95% CI | ||||||
| 3 or fewer ref. more than 3 EO | |||||||||
| Girls | Age adjusted | −0.16 | −0.28 | −0.03 | Women | Age adjusted | −0.52 | −0.76 | −0.29 |
| Multivariable adjusted | −0.10 | −0.22 | 0.04 | Multivariable adjusted | −0.32 | −0.56 | −0.07 | ||
| Boys | Age adjusted | −0.14 | −0.26 | −0.03 | Men | Age adjusted | −0.23 | −0.42 | −0.03 |
| Multivariable adjusted | −0.10 | −0.22 | 0.02 | Multivariable adjusted | −0.14 | −0.35 | 0.06 | ||
| 6 or more ref. fewer than 6 EO | |||||||||
| Girls | Age adjusted | 0.18 | 0.02 | 0.33 | Women | Age adjusted | 0.19 | 0.04 | 0.34 |
| Multivariable adjusted | 0.17 | 0.01 | 0.32 | Multivariable adjusted | 0.14 | −0.01 | 0.29 | ||
| Boys | Age adjusted | 0.06 | −0.16 | 0.27 | Men | Age adjusted | 0.2 | 0.02 | 0.38 |
| Multivariable adjusted | 0.01 | −0.20 | 0.23 | Multivariable adjusted | 0.15 | −0.03 | 0.33 | ||
| Breakfast skipping ref. breakfast eaters | |||||||||
| Girls | Age adjusted | −0.30 | −0.43 | −0.18 | Women | Age adjusted | −0.63 | −0.85 | −0.41 |
| Multivariable adjusted | −0.25 | −0.37 | −0.13 | Multivariable adjusted | −0.50 | −0.72 | −0.27 | ||
| Boys | Age adjusted | −0.35 | −0.48 | −0.22 | Men | Age adjusted | −0.33 | −0.57 | −0.09 |
| Multivariable adjusted | −0.33 | −0.47 | −0.20 | Multivariable adjusted | −0.26 | −0.51 | 0 | ||
| Late energy peak ref. early energy peak | |||||||||
| Girls | Age adjusted | −0.16 | −0.25 | −0.06 | Women | Age adjusted | −0.28 | −0.42 | −0.15 |
| Multivariable adjusted | −0.14 | −0.24 | −0.04 | Multivariable adjusted | −0.27 | −0.41 | −0.14 | ||
| Boys | Age adjusted | −0.15 | −0.25 | −0.04 | Men | Age adjusted | −0.18 | −0.33 | −0.03 |
| Multivariable adjusted | −0.15 | −0.25 | −0.04 | Multivariable adjusted | −0.13 | −0.28 | 0.02 | ||
Discussion
This study aimed to examine the associations between meal patterns and diet quality as well as meal patterns and overweight and obesity in two representative samples of the Swedish population. The results suggest that temporal meal patterns differ according to weight status in both adolescents and adults. Participants with normal weight had a higher eating frequency and a lower proportion of breakfast skippers, compared to adults with obesity or adolescents with overweight or obesity. The diet quality also differed according to weight status in adults, with those of normal weight having a higher diet quality compared to those with obesity. A low eating frequency, skipping breakfast and having a late energy peak were all inversely associated with diet quality, while a high eating frequency was positively associated in both adolescents and adults. However, the present study found fewer associations between meal patterns and diet quality in the stratified analyses, especially for adult men. Most previous research on adults has also shown positive associations between meal frequency and diet quality [35]. In children and adolescents, cross-sectional studies have shown both positive and negative associations between diet quality and eating frequency as well as with snack frequency [35–37]. Breakfast consumption has, in line with our results, been consistently associated with improved diet quality, and skipping breakfast has been associated with lower diet quality [3, 38–40].
Our results indicate that a higher meal frequency is associated with higher adherence to Swedish dietary guidelines in both adults and adolescent females but not in males. This indicates that the additional meals in males do not improve the diet quality. We did not examine snack frequency per se but meal frequency in this study, and it is likely that these additional meals are snacks. Whether snacks are healthy or not probably differs between countries and individuals, and would affect these relationships. In line with our findings, snacks/between-meal EO has been shown to contribute to the largest proportion of free sugars in this sample of adolescents [41].
Temporal meal patterns were associated with overweight and obesity, especially in adolescents. Taking lifestyle factors such as physical activity, smoking, and education level into account affected these relationships. Skipping breakfast was associated with obesity in adult men (OR 2.03). In adolescents, a high eating frequency (OR 0.44) and a late energy peak (OR 0.70) were associated with a lower risk for overweight or obesity, while a low eating frequency was associated with a higher risk (OR 1.77). In a previous study, we found a positive correlation between the number of meals and energy intake; however, this does not translate into a higher likelihood of overweight and obesity [12]. In both children and adults, most previous cross-sectional studies on meal frequency and body weight have shown a significant inverse association with body mass index(BMI), BMI z-score, and body weight; however, longitudinal research has shown mixed results [35]. Two meta-analyses showed that breakfast skipping was associated with a higher risk of having overweight and obesity, based on cross sectional and cohort studies globally [42, 43]. To gain a deeper understanding of these relationships, study designs beyond cross-sectional approaches are needed.
Misreporting of the energy intake, such as under- and overreporting, also affected the associations between meal patterns and weight status in the present study. This has been reported in previous research as well, and is an important factor to take into account when studying meal patterns [37, 44]. However, excluding these participants may also lead to bias since individuals with obesity tend to under-report their energy intake more [45–47].
A surprising finding in the present study was that in adolescents, a late energy peak was inversely associated with overweight and obesity (OR 0.70). Ingesting more energy later in the day is hypothesized to be less favorable for health due to the metabolic responses of food, such as poorer insulin sensitivity, compared to early in the day [48–50]. A previous meta-analysis has found a positive but small relationship between later eating rhythm and adiposity in children and adolescents, but with low certainty as there were many conflicting results and different definitions of later eating [51]. Previous observational research has shown a higher prevalence of obesity as well as a higher risk for cardiovascular disease risk in adults consuming their last meal later during the day, compared to earlier in the day [1, 52]. However, a meta-analysis of observational studies reported no association between BMI and evening energy intake [53]. A recent study has also evaluated these associations in relation to the polygenic risk of obesity, and found a significant interaction between meal timing and BMI among individuals with high polygenic risk, but not among those with low polygenic risk, suggesting that genetic predisposition should be considered in future research [54].
Adolescents also tend to have a later chronotype, and individuals with evening chronotypes have previously been shown to have higher body weight and being less adherent to a healthy diet [55–57]. The fact that we found an inverse relationship may be explained by that many adolescents who consume more energy later in the day might participate in active leisure activities and therefore eat when they arrive home later in the evening. Still, even though there was a lower risk for obesity, a late energy distribution was associated with a lower diet quality, which also needs to be considered. Foods consumed in the evening are probably of lower nutritional quality.
The large national representative samples and the validated methods used for dietary assessment are main strengths of this study. The combination of the two samples provided an opportunity to evaluate meal patterns in people aged 10–80 years, which is unique. The samples are quite generalizable to the Swedish population, especially for adolescents [16]. While the sampling was representative for adults; the participants were more often highly educated and included a lower proportion of people born abroad [20]. The dietary assessment offered detailed information regarding clock time, meal types and food intake combined.
Several limitations of this study need to be acknowledged. This study is cross-sectional and we cannot conclude any causal relationships between temporal meal patterns, diet quality and overweight and obesity. While we acknowledge that the inclusion of additional covariates such as genetic predispositions, chronic diseases, sleep, or other unmeasured factors could have provided further context, these data were not available in the present study. This study combines two surveys that differ slightly in dietary assessment methods, for example the time between two meals (15 min and 1 h), thus, analyses were carried out separately. Self-reported data can be a weakness, but also minimises external coding errors. Weight and height were self-reported for adults, which can lead to misclassification of weight status.
Knowledge of meal patterns of different population groups, such as people with different weight status, and how meal patterns are associated with diet quality is important information from a public health perspective. Future research is needed to evaluate causal relationships and to evaluate whether there are an association with metabolic health outcomes, especially longitudinal studies. Future research also needs to take misreporting into account when evaluating these relationships based on our findings and other similar research [44]. In addition, future studies could explore how meal patterns interact with circadian systems, energy metabolism, and hormonal factors to better understand the physiological mechanisms underlying these associations.
In conclusion, a low eating frequency, skipping breakfast or having a late energy peak are associated with a lower diet quality. Our study suggests that there may be a benefit in having a higher eating frequency and to consume breakfast, to achieve better adherence to the Swedish dietary guidelines and for weight status.
Supplementary Information
Below is the link to the electronic supplementary material. Supplementary file1 (DOCX 19 KB)