To investigate snacking frequency in relation to energy intake and food choices, taking physical activity into account, in obese vs reference men and women.
Cross-sectional, descriptive study.
In total, 4259 obese, middle-aged subjects (1891 men and 2368 women) from the baseline examination of the XENDOS study and 1092 subjects (505 men and 587 women) from the SOS reference study were included.
A meal pattern questionnaire describing habitual intake occasions (main meals, light meals/breakfast, snacks, drink-only), a dietary questionnaire describing habitual energy and macronutrient intake and a questionnaire assessing physical activity at work and during leisure time were used.
The obese group consumed snacks more frequently compared to the reference group (P<0.001) and women more frequently than men (P<0.001). Energy intake increased with increasing snacking frequency, irrespective of physical activity. Statistically significant differences in trends were found for cakes/cookies, candies/chocolate and desserts for the relation between energy intake and snacking frequency, where energy intake increased more by snacking frequency in obese subjects than in reference subjects.
Obese subjects were more frequent snackers than reference subjects and women were more frequent snackers than men. Snacks were positively related to energy intake, irrespective of physical activity. Sweet, fatty food groups were associated with snacking and contributed considerably to energy intake. Snacking needs to be considered in obesity treatment, prevention and general dietary recommendations.
Obesity is increasing at alarming rates worldwide.1 The disease is due to an undesirable positive energy balance over prolonged periods of time. The influence of meal frequency and snacking on energy imbalance and body weight is not clear. Some studies have reported a positive relation between low eating frequency and obesity,2, 3, 4, 5 whereas other studies found no relation.6, 7, 8 Gender differences have also been reported, that is, BMI has been found to be negatively associated with eating frequency in men and positively associated in women.9, 10 Similar inconsistencies have been found in the relation between eating frequency and energy intake, where some studies have reported a positive relation,10, 11 whereas other studies found no relation at all12 or gender differences.9
It has been suggested that the inconsistent associations reported between BMI, energy intake and eating frequency/snacking are due to differences in physical activity.8, 9 Drummond et al13 hypothesized that the lack of a relationship between increased snacking and BMI may be explained by the fact that frequent snackers may have higher energy intakes due to higher physical activity levels. This implies that those who have a frequent snack intake are more physically active and also maintain normal weight.
Yet, the prevalence of obesity is accelerating and simultaneously reports from population studies show an increasing energy intake related to an increase in snack intake.14, 15 In particular, a more frequent snacking pattern is associated with overconsumption of energy in children and adolescents15, 16, 17 and also with higher body weight.16 A similar pattern is found in male and female adult snackers who had a higher energy intake than never snackers, although BMI did not differ by snacking frequency.8
Food choices related to snacking are also a cause of concern. Different food groups, such as baked goods, sweets and beverages, have been related to snacking in both adolescent and adult normal weight subjects, as well as normal populations.8, 15, 18, 19, 20 Comparisons of snacking patterns and snack food choices between obese and nonobese subjects are sparse. No differences in snacking pattern were found between Swedish obese and normal weight men,21 whereas higher snacking frequency was found in obese women compared to a normal population.11 Energy-dense food choices as well as frequent snacking may facilitate increased energy intake.
Hence, we hypothesised that obese subjects have eating patterns, especially snacking patterns, deviating from a normal population. We therefore wanted to compare such patterns in a large sample of obese subjects vs a normal population, and to test for possible gender differences. Thus, the aim of this study was to investigate snacking frequency in relation to energy intake and food choices, taking physical activity into account, in obese vs reference men and women.
Subjects and methods
The study comprised two groups of subjects: a group of obese men and women and a reference group of men and women representing a normal population.
The obese men and women volunteered to participate in the XENDOS (XENical in the prevention of Diabetes in Obese Subjects) study. The aim of the XENDOS study was to assess the effectiveness of a weight loss drug, Xenical®, plus lifestyle changes compared with lifestyle changes alone in the prevention of type 2 diabetes. The study was performed at 22 centres in Sweden during 1997–2001.22
Subjects were participants in the baseline examination preceding the intervention part of the XENDOS study. Inclusion criteria were age 30–60 y old and BMI ≥30 kg/m2. In total, 4470 subjects were examined at baseline and all men (n=1891) and women (n=2368) with complete information on dietary intake were included in this study.
The reference group of men and women were recruited from the SOS (Swedish Obese Subjects) reference study, a population-based study conducted to obtain a reference group for the SOS study. The SOS study is described elsewhere.23 The SOS reference study was conducted in two Swedish cities (Mölndal and Örebro) during 1994–1999. Totally, 2037 subjects aged 37–60 y old were randomly selected and invited to participate in the study. The participation rate in the SOS reference study was 54% (n=524) among men and 58% (n=611) among women. All men (n=505) and women (n=587) with complete information on dietary intake were included in this study.
All XENDOS and SOS reference subjects underwent a health examination including anthropometric measures and completed questionnaires on meal patterns, dietary intake and socioeconomic aspects.
The studies were approved by ethical committees in Sweden and informed consent was obtained from all participants.
Meal pattern questionnaire
A self-administered meal pattern questionnaire describing habitual daily intake occasions was used. The subjects were asked to specify time for each intake occasion and choose one of four different types of intake occasions: main meal, light meal/breakfast, snacks or drink-only corresponding to their intake during ‘an ordinary day’.11 The total number of intake occasions was analysed as well as the total number of each type of intake occasion. Accordingly, total intake occasions were defined as total number of main meals, light meals/breakfast, snacks and drink-only, whereas eating occasions included all intake occasions except drink-only. Drink-only includes drink occasions both with and without energy. In subanalyses, snacking occasions were divided into four categories: 0, 1, 2 and 3 or more snacks per day.
A self-administered dietary questionnaire describing daily energy, macro- and micronutrient intake during the last 3 months was used. The questionnaire has been validated and judged to give valid estimates of energy intake in obese as well as in normal-weight subjects.24 Total energy intake, absolute and relative intake of macronutrients and fibre expressed as g/1000 kcal were used in the analysis. In addition, total energy intake was divided into 15 different food groups: cooked meals, cereals, sandwiches, toppings without bread, eggs, fast food (pizza/hamburgers/sausages), light meals (omelette/soup), desserts, fruit, nonalcoholic drinks, alcoholic drinks, milk, salty snacks (nuts/chips/popcorn), candies/chocolate and cakes/cookies.
Physical examination and laboratory values
The physical examination included anthropometric measures of weight and height. Weight was measured on calibrated scales in light indoor clothes to the nearest 0.1 kg and height to the nearest 0.01 m. BMI (kg/m2) was calculated as weight divided by height squared. Blood samples for cholesterol, triglycerides, blood-glucose and insulin were taken in the morning after a night's fast.
Education, occupational physical activity and leisure time physical activity
A questionnaire was used to assess education level and physical activity levels at work and during leisure time. Education level was reported on a six-category scale, but for our analyses this variable was dichotomised as: low education (elementary–nine-y compulsory school) and high education (upper secondary school–university). Occupational physical activity level was categorized in five levels: unemployed, sedentary work, moderately sedentary work, moderately heavy work and heavy work. Leisure time physical activity was categorized in four levels: sedentary lifestyle, moderate activity, moderate exercise and heavy exercise. The participants chose one of the alternatives corresponding to their usual activity pattern. In our analyses, sedentary leisure time and sedentary work were coded as a dichotomous variable, sedentary yes=1, no=0.
Means and standard deviations were used to describe age, anthropometric variables, metabolic variables, energy and macronutrient and number of intake occasions. Dichotomous variables, for example, education and physical activity, were presented as percentages. A logistic regression model was used to describe association between obesity (dependent variable, obese group=1, reference group=0) and three different types of intake occasions controlling for energy intake (kcal/1000) and age. In Table 2 statistical significance of differences between the obese and reference group and between men and women was tested using the likelihood ratio test. The null model for the test of gender assumes no difference in means between men and women. The null model for the test of obese and reference groups assumes no difference in means between the two study groups. In both tests, the alternative model assumes differences in means for all four group/gender combinations. In Table 3, Figures 2 and 3, general linear models, adjusted for age and physical activity, were used to analyse linear trends in energy and nutrient intake related to snacking categories. The test in trend is similar for those of test of means above. The trends (slopes) were compared instead of means. All P-values presented were initially unadjusted for multiple comparisons. When analysing snack categories in relation to dietary intake and metabolic variables, the Bonferroni correction for multiple comparisons was performed. The SAS 8.02 statistical package was used for all analyses (SAS Institute Inc., Cary, NC 27513, USA).
Characteristics of the obese and reference groups are presented in Table 1. A larger proportion of the obese men reported low education compared to reference men, whereas no difference in education level was found between the two groups of women. As expected, the obese men and women had higher levels of b-glucose, s-insulin, s-cholesterol and s-triglycerides and lower s-HDL than the reference men and women.
Number of intake occasions and energy intake
Gender differences in intake occasions between obese and reference subjects were found. The obese men and women most frequently reported six intake occasions per day compared to five intake occasions in the reference men and women (Figure 1a and b). Table 2 shows the mean number of intake occasions and mean number of eating occasions, as well as the intake occasions by type of intake: main meals, light meals/breakfast, snacks and drink-only. The mean number of intake occasions was higher in the obese group than in the reference group (P<0.001). Number of intake occasions also tended to be higher in women compared to men (P=0.055). When excluding drink-only, that is, number of eating occasions, the gender difference was significant (P<0.001). Snacks were more frequently consumed in the obese group compared to the reference group (P<0.001), as well as in women compared to men (P<0.001).
In contrast to what is commonly seen in dietary studies, the obese men and women reported a significantly higher energy intake than reference men and women (P<0.001) (Table 2). Number of intake occasions was independently related to obesity in an energy- and age-adjusted logistic regression model. Odds ratio for number intake occasions was 1.21, CI=1.15–1.27, P<0.001 and for energy (kcal/1000) 1.49, CI=1.37–1.62, P<0.001. When using number of snacks instead of intake occasions in the model OR for number of snacks was 1.27 CI=1.19–1.35, P<0.001 and for energy (kcal/1000) 1.48 CI=1.37–1.61, P<0.001.
Leisure time physical activity also differed between groups. In the obese group, a larger proportion of men and women reported sedentary leisure time physical activity than reference men and women (P<0.001), but there was no gender difference. Conversely, a gender difference, but no group difference, was noted for occupational physical activity. A larger proportion of the men than the women reported a low physical activity level at work (P<0.008) (Table 2).
Intake frequency in relation to energy intake
The relation between energy intake and reported number of intake occasions is shown in Figure 2. The intake occasions were divided into three different meal types: (Figure 2a) principal meals (main meal+light meal/breakfast), (Figure 2b) drink-only and (Figure 2c) snacks. The energy intake related to reported intake occasions were analyzed taking age and physical activity into account. Energy intake increased significantly by increasing number of principal meals in both obese men and women, whereas no such trends were found in the reference group. There was a negative trend in energy intake by drink-only category in all groups but the reference women and neither group nor gender trend differences were found. On the other hand, increased snacking frequency was associated with higher energy intake in all four groups; however, the trends were steeper in the obese group (P<0.01 for difference in trend between obese and reference groups).
Snacking frequency in relation to dietary intake
Table 3 presents the distribution of BMI and dietary variables, adjusted for age and physical activity, by each snack category; 0, 1, 2 and ≥3 snacks per day. BMI was not associated with number of snacks per day in either group after correction for multiple comparisons. However, energy intake was related to number of snacks as reported above. Increases in energy intake per day between the lowest and highest snack category were 830 kcal (29.2%) for the obese men, 640 kcal (25.5%) for the reference men, 584 kcal (25.1%) for the obese women and 312 kcal/day (15.5%) for the reference women.
As a consequence of increasing energy intake, the absolute intake of protein, fat and carbohydrate also increased by snacking frequency in all four groups. The relative macronutrient intake is presented in Table 3. In obese men and women, the proportion of fat intake increased (P<0.05, after Bonferroni correction) and the proportion of protein intake decreased (P<0.05, after Bonferroni correction) by increasing snacking category. However, no significant trends were found in the reference group. The proportion of total carbohydrate intake did not differ by snacking category either in the obese or in the reference men and women, although carbohydrates from mono-disaccharides increased in men. The difference in trend between men and women was significant (P<0.05).
The absolute intake of alcohol decreased by increased snacking in all four groups. For relative alcohol intake, there was both a gender (P<0.05) and a group (P<0.01) difference in trend. The relative decrease in alcohol intake by snacking category was similar and significant in obese men and women. However, in reference men the trend was steeper, whereas in reference women the trend was not significant after Bonferroni correction. No significant trends in fibre intake were found in the two groups of men and women.
Snacking frequency and energy intake from the 15 food groups
A statistically significant difference in trend between the obese and reference groups was found for the relation between energy intake and snacking frequency for three food groups: cakes/cookies (P<0.01), candies/chocolate (P<0.05) and desserts (P<0.01) Figures 3a–c. For cakes/cookies energy intake increased by increasing snacking frequency in both obese and reference men and women; however, the increase was steeper in the obese men and women. The same pattern was seen for candies/chocolate, although the trend was not significant in reference women after Bonferroni correction. Energy intake from desserts increased significantly by snacking frequency only in the obese men and women.
Furthermore, a significant difference in trend between men and women was found for the relation between energy intake from milk and snacking frequency, Figure 3d. Both obese and reference men reported increased milk consumption with increasing snack category; however, no significant relation was found for women.
Sandwich intake increased by increasing snacking frequency (Figure 3e). The trends were similar in all four groups, but the intake level was substantially higher in men than in women. However, the differences between obese and reference men and between obese and reference women were small and non significant.
High-energy intake from snacks may be compensated by lower energy intakes in meals. In order to investigate possible compensation in certain food groups corresponding to main meals, the energy intake from two food groups, cooked meals and light meals, were combined (Figure 3f). A significant difference in trend between obese and reference groups (P<0.05) was found for energy intake in the combined food group by increasing snacking frequency. The obese women reported an increase in energy intake from the combined food groups, whereas no statistical trend was found in the obese men. There was no change in trend for energy intake in the combined food groups by increasing snacking frequency in the reference men and women, although the reference group reported a lower energy intake than the obese group.
For the remaining food groups (toppings without bread, cereals, egg, fast food, nonalcoholic drinks, alcoholic drinks, fruit and salty snacks), neither gender nor group differences were found in energy intake by snacking category. The intake trends were inconsistent and the contribution to total energy intake was generally small.
Snacking frequency in relation to metabolic variables
The metabolic variables blood pressure, b-glucose, s-insulin, s-cholesterol, s-triglycerides and s-HDL were not significantly related to snacking frequency in any of the four study groups after Bonferroni correction.
Our results showed that Swedish obese men and women were more frequent snackers than Swedish reference men and women and also that women consumed more snacks than men. Energy intake increased by higher snacking frequency, irrespective of physical activity. Furthermore, the obese group's food intake pattern differed from that of the normal population in showing a more pronounced energy intake, especially from sweet, fatty food choices among obese frequent snackers.
Our finding of a positive relationship between intake occasions, especially regarding snacks, and high BMI in both men and women is at variance with previous studies in adults. In a review of epidemiological studies on the relation between meal frequency and body weight, Bellisle et al25 concluded that although many studies failed to find any significant relation, those that did consistently reported an inverse relation. Recent studies on the relation between intake frequency and BMI in nonobese individuals or normal populations also report a negative relation in men,10, 12 but positive in females.10 Furthermore, no association between snacking and BMI was found in either males or females.8 In contrast, we compared intake patterns of a group of obese subjects with subjects from a normal population and found group differences as well as gender differences.
The lack of consistency between studies may be due to different definitions of intake occasions26 and/or different assessment methods.4, 9, 12, 20, 21 Inconsistencies in the definition of meals and snacks make comparisons between studies hard to interpret. For example, no generally accepted definition exists of what constitutes high or low meal frequency26 and, furthermore, distinctions between a meal, a snack and a drink are ambiguous. This ambiguity may be due to cultural aspects,27 as well as the use of different time periods9, 20 or varying energy levels to distinguish between meals and snacks.28 We defined snacking as an intake occasion that was not considered a main meal, light meal/breakfast or drink-only. Moreover, intake occasions may also be predefined and ‘force’ the respondent into certain patterns,20 or self-defined, leaving the definition to the respondent.29 The latter has been advocated by Booth30 and was used in this study.
Secondly, common methods in dietary assessment, such as food records9, 12, 19 and 24-h recalls,5, 8, 21 are often used to assess habitual intake frequency, but may not be appropriate for this purpose. For example, energy intake is often under-reported when using these methods.21, 31 This bias may be specifically related to snacking,32 since if energy intake is under-reported, presumably intake frequency is also under-reported, particularly in relation to snacking. In a study of Swedish men using the 24-h recall method, the obese group reported fewer intake occasions than did the normal weight group, but the difference disappeared after excluding under-reporters.21 Although intake occasions may also be under-reported in our study, our results still show a higher frequency of intake occasions in the obese group than in the normal population. Assuming that snacking frequency is under-reported, we can speculate that results from studies indicating a negative relation between eating frequency and body weight might be due to the use of methods unable to adequately assess snack intake.
We combined two independent methods, ‘the meal pattern questionnaire’, assessing habitual intake occasions and ‘the dietary questionnaire’, assessing habitual energy intake. In ‘the meal pattern questionnaire’ the number of intake occasions was assessed but not energy content or specific food groups. However, the energy intake reported in the ‘the dietary questionnaire’ was related to reported number of snacks in the meal pattern questionnaire. Although we could not assess energy intake in each intake occasion, we found that energy intake was higherwith increased snacking frequency, which is consistent with results from large population surveys using other methods.8, 14, 15 In a previous study using these assessment methods, we found that obese women had a more frequent intake pattern than women in a normal population and, furthermore, that snacking was positively related to energy intake.11 This finding is in accordance with the present study, but is inconsistent with other studies that have reported that obese subjects have a lower intake frequency than nonobese. However, it should be noted that the frequency of intake occasions found in the reference subjects concurs with that found in other studies in normal weight subjects and normal populations.19, 21, 33, 34 For example, in a Nordic study, eating frequency (‘just a drink’ excluded) was reported to be 4.1 in a Swedish normal population.34 Bellisle et al19 reported an average intake of 2.7 meals and 1.3 snacks per day in French normal weight adults and in Swedish students.35 In comparison, our obese group reported a similar meal frequency (main meals+light meals/breakfast), but more frequent snacks. The meal frequency was related to increased energy intake only in the obese group whereas snack frequency was related to increased energy intake in both the obese and the reference group. The role of snacking as a contributor to total energy intake should not be neglected.
Certain food groups are frequently consumed as snacks. In young adults, the top contributors to energy from snacking were desserts, beverages, milk and salty snack foods.15 In adult nonobese and normal populations, a large proportion of energy intake from snacks derives from baked goods, sweets, fruit and dairy products.8, 18, 19, 20 Although our data do not permit comparable analyses, we did find that the most frequent snackers had a considerable daily energy intake from candies/chocolate; the percentage of daily energy was 8 and 11% in obese men and women, respectively, vs 6 and 7.5% in reference men and women, respectively (data not shown). Studies have reported that snacks contribute to 20–25% of daily energy intake.20, 36 Corresponding data are not available in this study for methodological reasons, but food groups that have been related to snacking, such as sweets, soft drinks and baked goods,20 contributed considerably to energy intake, especially in the frequent snackers. These food items easily lead to energy surplus and do not facilitate weight control.
Gender differences were less pronounced than obesity differences in the relation between snacking and energy intake from the defined food groups. Energy intake from candies/chocolate, cakes/cookies and desserts increased more steeply by increasing number of snacks in the obese compared to reference men and women. A nibbling pattern has been considered of advantage for body weight control and for preventing obesity.37, 38 Even if we assume that the most frequent snackers in our study were nibblers, we still have to consider what people in fact eat when they nibble. To recommend a nibbling pattern in obesity prevention and treatment may be counter effective if total energy intake and food choices are neglected. Nibbling might be recommended if it can be assumed that energy intake from snacks is compensated for in subsequent meals, that is, that the meals contain less energy. However, our findings revealed no such compensation, which is in accordance with experimental studies.39, 40, 41 Nibblers, as opposed to gorgers, have also been reported to compensate energy intake after an energy-reduced lunch.42 In addition to weight control, a further argument against the nibbling theory is that no consistent pattern was found between snacking/intake frequency and metabolic variables, which is in contrast to previous studies.2, 10, 43
A limitation in our study is that our obese subjects were volunteers for an intervention study and may thus not be representative for all obese subjects, but rather for men and women seeking help for obesity. Moreover, like most other studies, our study is cross-sectional and thus we cannot conclude causality. On the other hand, our results are consistent with the results from the Bogalusa heart study in children where snacking was associated to higher BMI longitudinally.16 Another limitation is that the meal pattern questionnaire used here has not been validated. Appropriate criteria against which to validate this questionnaire, such as reliable dietary assessment methods or biomarkers, are currently unavailable and thus validation of habitual intake frequency is impossible to accomplish. On the other hand, the dietary questionnaire assessing energy and macronutrient intake is validated against energy expenditure, urinary nitrogen and food records.24
By combining two independent methods, we found a link between frequent snacking, high-energy intake and obesity. In conclusion, our study showed that men and women with obesity have more frequent intake occasions, higher energy intake and more sedentary leisure time than a reference population. Furthermore, we found that obese subjects were more frequent snackers than reference subjects and women were more frequent snackers than men. High physical activity could not explain high-energy intake and snacking. Certain food groups were associated with snacking and contributed considerably to energy intake. Food choices and snacking need to be considered in obesity treatment, prevention and general dietary recommendations.
World Health Organization. Obesity: preventing and managing the global epidemic: report of a WHO consultation. Geneva; 2000.
Fábry P, Hejl Z, Fodor J, Braun T, Zvolánková K . The frequency of meals. Its relation to overweight, hypercholesterolemia and decreased glucose-tolerance. Lancet 1964; 2: 614–615.
Metzner HL, Lamphiear DE, Wheeler NC, Larkin FA . The relationship between frequency of eating and adiposity in adult men and women in the Tecumseh Community Health Study. Am J Clin Nutr 1977; 30: 712–715.
Kant AK, Schatzkin A, Graubard BI, Ballard-Barbash R . Frequency of eating occasions and weight change in the NHANES I Epidemiologic Follow-up Study. Int J Obes Relat Metab Disord 1995; 19: 468–474.
Ma Y, Bertone ER, Stanek III EJ, Reed GW, Hebert JR, Cohen NL, Merriam PA, Ockene IS . Association between eating patterns and obesity in a free-living US adult population. Am J Epidemiol 2003; 158: 85–92.
Dreon DM, Frey-Hewitt B, Ellsworth N, Williams PT, Terry RB, Wood PD . Dietary fat: carbohydrate ratio and obesity in middle-aged men. Am J Clin Nutr 1988; 47: 995–1000.
Edelstein SL, Barrett-Connor EL, Wingard DL, Cohn BA . Increased meal frequency associated with decreased cholesterol concentrations; Rancho Bernardo, CA, 1984–1987. Am J Clin Nutr 1992; 55: 664–669.
Hampl JS, Heaton CL, Taylor CA . Snacking patterns influence energy and nutrient intakes but not body mass index. J Hum Nutr Diet 2003; 16: 3–11.
Drummond SE, Crombie NE, Cursiter MC, Kirk TR . Evidence that eating frequency is inversely related to body weight status in male, but not female, non-obese adults reporting valid dietary intakes. Int J Obes Relat Metab Disord 1998; 22: 105–112.
Titan SM, Bingham S, Welch A, Luben R, Oakes S, Day N, Khaw KT . Frequency of eating and concentrations of serum cholesterol in the Norfolk population of the European prospective investigation into cancer (EPIC-Norfolk): cross sectional study. BMJ 2001; 323: 1286–1288.
Bertéus Forslund H, Lindroos A, Sjöström L, Lissner L . Meal patterns and obesity in Swedish women – a simple instrument describing usual meal types, frequency and temporal distribution. Eur J Clin Nutr 2002; 56: 740–747.
Ruidavets JB, Bongard V, Bataille V, Gourdy P, Ferrieres J . Eating frequency and body fatness in middle-aged men. Int J Obes Relat Metab Disord 2002; 26: 1476–1483.
Drummond S, Crombie N, Kirk T . A critique of the effects of snacking on body weight status. Eur J Clin Nutr 1996; 50: 779–783.
Nielsen SJ, Siega-Riz AM, Popkin BM . Trends in energy intake in U.S. between 1977 and 1996: similar shifts seen across age groups. Obes Res 2002; 10: 370–378.
Zizza C, Siega-Riz A-M, Popkin BM . Significant increase in young adults snacking between 1977–1978 and 1994–1996 represent a cause for concern!. Prev Med 2001; 32: 303–310.
Nicklas TA, Yang SJ, Baranowski T, Zakeri I, Berenson G . Eating patterns and obesity in children. The Bogalusa Heart Study. Am J Prev Med 2003; 25: 9–16.
Dwyer JT, Evans M, Stone EJ, Feldman HA, Lytle L, Hoelscher D, Johnson C, Zive M, Yang M . Adolescents’ eating patterns influence their nutrient intakes. J Am Diet Assoc 2001; 101: 798–802.
Cross AT, Babicz D, Cushman LF . Snacking patterns among 1,800 adults and children. J Am Diet Assoc 1994; 94: 1398–1403.
Bellisle F, Dalix AM, Mennen L, Galan P, Hercberg S, de Castro JM, Gausseres N . Contribution of snacks and meals in the diet of French adults: a diet-diary study. Physiol Behav 2003; 79: 183–189.
Summerbell C, Moody R, Shanks J, Stock M, Geissler C . Sources of energy from meals vs snacks in 220 people in four age groups. Eur J Clin Nutr 1995; 49: 33–41.
Andersson I, Rössner S . Meal patterns in obese and nonobese men: The ‘Gustaf’ study. Eur J Clin Nutr 1996; 50: 639–646.
Torgerson JS, Arlinger K, Kappi M, Sjostrom L . Principles for enhanced recruitment of subjects in a large clinical trial. The XENDOS (XENical in the prevention of Diabetes in Obese Subjects) study experience. Control Clin Trials 2001; 22: 515–525.
Sjöström L, Larsson B, Backman L, Bengtsson C, Bouchard C, Dahlgren S, Hallgren P, Jonsson E, Karlsson J, Lapidus L, Lindroos AK, Lindstedt S, Lissner L, Narbro K, Näslund I, Olbe L, Sullivan M, Sylvan A, Wedel H, Ågren G . Swedish obese subjects (SOS). Recruitment for an intervention study and a selected description of the obese state. Int J Obes Relat Metab Disord 1992; 16: 465–479.
Lindroos AK, Sjöström L, Lissner L . Validity and reproducibility of a self-administered dietary questionnaire in obese and nonobese subjects. Eur J Clin Nutr 1993; 47: 461–481.
Bellisle F, McDevitt R, Prentice AM . Meal frequency and energy balance. Br J Nutr 1997; 77: S57–S70.
Gatenby SJ . Eating frequency: methodological and dietary aspects. Br J Nutr 1997; 77 (Suppl 1): S7–S20.
Chamontin A, Pretzer G, Booth DA . Ambiguity of ‘snack’ in British usage. Appetite 2003; 41: 21–29.
Longnecker MP, Harper JM, Kim S . Eating frequency in the Nationwide Food Consumption Survey (USA), 1987–1988. Appetite 1997; 29: 55–59.
Roos E, Prättäla R . Meal pattern and nutrient intake among adult finns. Appetite 1997; 29: 11–24.
Booth DA . Evidence-based reduction of obesity: identification of a subculture's least fattening eating patterns. Appetite 1999; 32: 80–85.
Lissner L, Habicht J-P, Strupp B, Haas J, Levitsky D, Roe D . Body composition and energy intake: do overweight women overeat and underreport? Am J Clin Nutr 1989; 49: 320–325.
Poppitt SD, Swann D, Black AE, Prentice A . Assessment of selective under-reporting of food intake by both obese and non-obese women in a metabolic facility. Int J Obes Relat Metab Disord 1998; 22: 303–311.
Becker W . Befolkningens kostvanor och näringsintag(HULKEN). Vår Föda 1992; 44: 349–362.
Makela J, Kjaernes U, Pipping Ekstrom M, L’Orange Furst E, Gronow J, Holm L . Nordic meals: methodological notes on a comparative survey. Appetite 1999; 32: 73–79.
Bellisle F, Monneuse M, Steptoe A, Wardle J . Weight concerns and eating patterns: a survey of university students in Europe. Int J Relat Metab Disord 1995; 19: 723–730.
Gatenby AO, Anderson AD, Southon S, Mela DJ . ‘Meals’ and ‘snacks’. implications for eating patterns in adults. Appetite 1995; 24: 292.
Livsmedelsverket. Svenska Näringsrekommendationer. Uppsala, Sweden 1997.
Kirk TR . Role of dietary carbohydrate and frequent eating in body-weight control. Proc Nutr Soc 2000; 59: 349–358.
Rolls BJ, Kim S, McNelis AL, Fischman MW, Foltin RW, Moran TH . Time course of effects of preloads high in fat or carbohydrate on food intake and hunger ratings in humans. Am J Physiol 1991; 260: R756–R763.
Marmonier C, Chapelot D, Louis-Sylvestre J . Metabolic and behavioral consequences of a snack consumed in a satiety state. Am J Clin Nutr 1999; 70: 854–866.
Marmonier C, Chapelot D, Louis-Sylvestre J . Effects of macronutrient content and energy density of snacks consumed in a satiety state on the onset of the next meal. Appetite 2000; 34: 161–168.
Westerterp-Plantenga M, Wijckmans-Duysens N, Ten Hoor F . Food intake in the daily environment after energy-reduced lunch, related to habitual meal frequency. Appetite 1994; 22: 173–182.
Jenkins DJA, Ocana A, Jenkins AL, Wolever TMS, Vuksan V, Katzman L, Hollands M, Greenberg G, Corey P, Patten R, Wong G, Josse RG . Metabolic advantages of spreading the nutrient load: effects of increased meal frequency in non-insulin-dependent diabetes. Am J Clin Nutr 1992; 55: 461–467.
This study was supported by the Swedish Research Council (grant 05239) and F Hoffmann-La Roche.
About this article
Patterns of added sugars intake by eating occasion among a nationally representative sample of Australians
European Journal of Nutrition (2018)
Is snack consumption associated with meal skipping in children and adolescents? The CASPIAN-IV study
Eating and Weight Disorders - Studies on Anorexia, Bulimia and Obesity (2017)
The contribution of snacks to dietary intake and their association with eating location among Norwegian adults – results from a cross-sectional dietary survey
BMC Public Health (2015)
It’s the power of food: individual differences in food cue responsiveness and snacking in everyday life
International Journal of Behavioral Nutrition and Physical Activity (2015)
A mid-morning snack of almonds generates satiety and appropriate adjustment of subsequent food intake in healthy women
European Journal of Nutrition (2015)