Edited by: Kelly Costello Allison, Perelman School of Medicine of the University of Pennsylvania, USA
Reviewed by: Annika Petra Christine Lutz, University of Luxembourg, Luxembourg; Suzanne Higgs, University of Birmingham, UK
Specialty section: This article was submitted to Eating Behavior, a section of the journal Frontiers in Nutrition
This is an open-access article distributed under the terms of the Creative Commons Attribution License (CC BY). The use, distribution or reproduction in other forums is permitted, provided the original author(s) or licensor are credited and that the original publication in this journal is cited, in accordance with accepted academic practice. No use, distribution or reproduction is permitted which does not comply with these terms.
Eating frequently during the day, or “grazing,” has been proposed to assist with managing food intake and weight. This systematic review assessed the effect of greater eating frequency (EF) on intake and anthropometrics in human and animal experimental studies. Studies were identified through the PubMed electronic database. To be included, studies needed to be conducted in controlled settings or use methods that carefully monitored food intake, and measure food intake or anthropometrics. Studies using human or animal models of disease states (i.e., conditions influencing glucose or lipid metabolism), aside from being overweight or obese, were not included. The 25 reviewed studies (15 human and 10 animal studies) contained varying study designs, EF manipulations (1–24 eating occasions per day), lengths of experimentation (230 min to 28 weeks), and sample sizes (3–56 participants/animals per condition). Studies were organized into four categories for reporting results: (1) human studies conducted in laboratory/metabolic ward settings; (2) human studies conducted in field settings; (3) animal studies with experimental periods <1 month; and (4) animal studies with experimental periods >1 month. Out of the 13 studies reporting on consumption, 8 (61.5%) found no significant effect of EF. Seventeen studies reported on anthropometrics, with 11 studies (64.7%) finding no significant effect of EF. Future, adequately powered, studies should examine if other factors (i.e., disease states, physical activity, energy balance and weight status, long-term increased EF) influence the relationship between increased EF and intake and/or anthropometrics.
Approximately two out of every three adults in the U.S. are overweight or obese (
One key area in obesity treatment is reducing energy intake (
The relationship between EF and body weight in humans was first examined with a cross-sectional investigation that was published in 1964 (
In the 50 years since the relationship of EF and weight status was initially investigated in humans, numerous observational studies have been conducted examining this relationship in adults and outcomes are very mixed [for review, see Bellisle and colleagues (
While previous reviews have been published in the area of EF, intake, and/or weight, these reviews have included observational studies and/or failed to include animal experimental research (
Studies were identified by searching the PubMed electronic database from January to March 2014, and in November 2015. The search terms included were eating frequency, snacking, and feeding frequency AND appetite, satiety, satiation, energy intake, body weight, obesity, and metabolism. These terms were used as they are meaningful terms for both human and animal research, and thus could be used for both searches (human and animal). No constraints were set on date of publication or type of study; however, only articles published in English were reviewed. Identified abstracts were screened for duplicates. Abstracts identified were reviewed (Guoxun Chen, Matthew R. Goff, and Hollie A. Raynor) and full articles for abstracts meeting criteria were retrieved and further evaluated (Matthew R. Goff and Hollie A. Raynor). Discrepancies in decisions were discussed and resolved with Guoxun Chen and Seletha A. Poole. Additionally, bibliographies of identified reviewed studies (
For human studies, only original research studies reporting the influence of a manipulation of increasing EF on a measure of energy intake, as assessed via daily energy consumed or energy intake during an
Inclusion criteria for animal studies were similar to that of human studies, in that only original research studies reporting the influence of a manipulation of increasing EF on a measure of food intake or body weight were considered for inclusion. All experimental designs were eligible for inclusion. Studies that only included animal models representing glucose or lipid metabolism disease states were not included in this review. There was no restriction on length of feeding manipulation to be included in the review.
For the human studies, general study characteristics (author, year of publication, study design, and length of study), characteristics of study population (age, gender, BMI), experimental manipulation (EF and diet prescription), study outcomes (energy intake and anthropometrics), and potential mechanisms (energy expenditure, self-reported appetite, and cardiometabolic/hormonal measures) were extracted from included studies. Hollie A. Raynor and Seletha A. Poole independently extracted data from each study. Discrepancies regarding data extraction were resolved by discussion.
For the animal studies, general study characteristics (author, year of publication, comparison groups, and testing duration), characteristics of study population (species, strain), experimental manipulation (EF and diet prescription), study outcomes (food intake and anthropometrics), and potential mechanisms (cardiometabolic/hormonal measures) were extracted from the included studies. Guoxun Chen and Matthew R. Goff independently extracted data from each study. Discrepancies regarding data extraction in both human and animal studies were resolved by discussion.
For human studies, extracted studies were divided into two categories based on the setting of the study: research or field. All identified studies reported on at least one of the primary outcomes (energy intake, anthropometrics). Energy expenditure (dietary-induced thermogenesis or total energy expenditure), self-reported appetite (hunger, fullness, and satiety), and/or cardiometabolic/hormonal measures (glucose, insulin, cholesterol, high-density lipoprotein, low-density lipoprotein, triglyceride, triacylglycerol, gastric inhibitory polypeptide, ghrelin, leptin, glucagon-like peptide-1, peptide YY, and free fatty acids) were reported as potential mechanisms or indicators of potential mechanisms in the relationship between EF, energy intake, and anthropometrics.
For animal studies, extracted studies were divided into two categories based on testing duration: EF studies conducted <1 month and EF studies conducted >1 month (month was defined as a single 30-day period). All identified studies reported at least one of the previously mentioned outcomes (food intake or body weight). Potential mechanisms regarding the relationship between EF and intake or weight, such as cardiometabolic/hormonal measures (glucose-6-phosphate dehydrogenase, glycerol-3-phosphate dehydrogenase, insulin, insulin-like growth factor-1, protein, hematocrit, glucose, urea, glucagon, leptin, lipogenesis, and malic enzyme), were also reported.
The initial search for human studies yielded 972 records. After removing duplicate records and including additional relevant articles identified through bibliographies of included records, 69 articles were assessed for eligibility and 15 were included in this review. The flow of included studies is outlined in Figure
Citation | Participants | Study design | EF prescription | Diet prescription | Intervention length and assessments | Measures taken | Results | ||||
---|---|---|---|---|---|---|---|---|---|---|---|
Anthropometrics | Energy expenditure | Self-reported appetite regulation | Cardiometabolic/hormonal | ||||||||
Allirot et al. ( |
RC: |
674.8 kcal of conventional foods; required to consume all provided foods | 4, 240-min laboratory sessions with 7-d between sessions; |
≠ | F4 ↓ F1 | ||||||
Allirot et al. ( |
RC: |
674.8 kcal of conventional foods; required to consume all provided foods | 4, Laboratory sessions of varying length with 7-d between sessions; EE measured during basal period (T-30 to 0 min) and over 430 min; self-reported appetite regulation measured in 2 sessions, 9× over 390 min with an |
≠ | NR | ||||||
Antoine et al. ( |
Mean ± SD of pts age and BMI for entire sample NR |
RC: |
1200 kcal; not specified if required to consume all provided foods | 2, 14-d interventions in metabolic ward without washout period; assessments occurred at 1, 15, and 29 d | ≠ | ||||||
Bortz et al. ( |
Mean ± SD of pts age and BMI for entire sample NR |
NRC: |
600 kcal/day liquid meals; unspecified if required to consume all provided foods | Intervention in metabolic ward of unspecified length with washout period unspecified; BW and cardiometabolic/hormonal measures every 6 d | No difference with significance NR | ||||||
Chapelot et al. ( |
Mean ± SD of pts age and BMI for entire sample NR |
NRC: |
No prescribed diet | 2, 24-h laboratory sessions with 28-d between sessions following prescription; all meals and snack |
≠ | ≠ | NR | ||||
Dallosso et al. ( |
Mean ± SD of pts age and BMI for entire sample NR |
RC: |
42 kcal/kg body weight daily for a 65-kg reference man with light activity pattern; unspecified if required to consume all provided foods | 2, 14-d interventions in metabolic ward with washout period unspecified; BW and EE measured at 0, 7, 14, 21, and 28 d | Gorging ↑ Nibbling at 7 d | ≠ | |||||
Dougkas et al. ( |
RC: |
Standard breakfast of 348 kcal and in snack sessions 201 kcal of dairy snack; unspecified if required to consume all provided standard breakfast; snack required to be consumed in 5 min | 4, 230-min laboratory sessions with 1 week between sessions; |
All snacks ↓ water (no snack) | |||||||
Speechly and Buffenstein ( |
RC: |
33% EER; required to consume all provided foods | 2, 405-min laboratory sessions with unspecified washout period; |
Single ↑ Multi | ≠ | ||||||
Speechly et al. ( |
RC: |
33% EER; required to consume all provided foods | 2, 405-min laboratory sessions with unspecified washout period; |
Single ↑ Multi | Single ↑ Multi at 315 min | ||||||
Swindells et al. ( |
Mean ± SD of pts age and BMI for entire sample NR |
NRC: |
100% of EER; unspecified if required to consume all provided foods | 27-d intervention in metabolic ward (6 d of 3 meals, 6 d of 2 meals, 6 d of 3 meals, 6 d of 9 meals, 3 d of 3 meals) with no washout period; BW measures every day | No difference with significance NR | ||||||
Verboeket-van de Venne et al. ( |
Mean ± SD of pts age and BMI of entire sample NR |
RC: |
Average daily energy requirement based on 7 d food record; instructed to consume all provided foods and asked to return any foods not consumed | 2, 1-week interventions (6 d free-living in which food was provided) and 1 d in respiration chamber without washout period | ≠ |
Citation | Participants | Study design | EF prescription | Diet prescription | Intervention length and assessments | Measures taken | Results | ||||
---|---|---|---|---|---|---|---|---|---|---|---|
Diet | Anthropometrics | Energy expenditure | Self-reported appetite regulation | Cardiometabolic/hormonal | |||||||
Finkelstein and Fryer ( |
Mean ± SD of pts age and BMI for the entire sample NR |
RCT | Low-kcal (1400 kcal), low-fat (40 g fat) with all foods provided | 60-d intervention; BW measured every 7th d; all cardiometabolic/hormonal measures except glucose measured at fasting on d 0 and 60; fasting glucose measured every 7th d | ≠ | ||||||
Iwao et al. ( |
Mean ± SD of pts age and BMI for entire sample NR; gender of entire sample not reported |
RCT | Low-kcal (1200 kcal), low-fat (19.8% energy from fat) with all food provided as a liquid supplement; unspecified if required to consume all provided foods | 14-d intervention; assessments occurring at 0 and 14 d, cardiometabolic/hormonal measures measured at fasting | ≠ | ≠ | ≠ | ||||
Murphy et al. ( |
RC: |
2000 kcal; unspecified if required to consume all provided foods | 2, 2-week interventions with 3 weeks between interventions; assessments occurred at 0, 2, and 7 weeks with cholesterol, HDL, and LDL cholesterol measures taken at fasting and all other cardiometabolic/hormonal measures taken over 8 h | ||||||||
Stote et al. ( |
RC: |
BEE x 1.3-1.5; required to consume all provided foods | 2, 8-week interventions with 11 weeks between interventions; BW measured daily before evening meal; self-reported appetite regulation measures taken daily before evening meal; cardiometabolic/hormonal measures taken at fasting at 0, 4, and 8 weeks of each intervention | One meal ↓ Three meal | |||||||
Verboeket-van de Venne et al. ( |
Mean ± SD of pts age and BMI for entire sample NR |
RC: |
Based on average daily energy requirement based on a 7-d food record; instructed to consume all provided foods and asked to return any foods not consumed | 2, 1-week interventions (6-d free-living in which food was provided and 1 d in respiration chamber) without washout period; BW measured at beginning and end of each 1-week period; average daily metabolic rate measured over 6-d in free-living conditions | ≠ | ≠ |
Out of the 11 studies conducted in laboratory/metabolic ward settings, 4 manipulated EF within one eating occasion (breakfast), while the remaining 7 studies manipulated EF occurring over a longer period (
Out of the five studies conducted in field settings, three used a randomized crossover design (
The initial search for animal studies yielded 76 results. After removing duplicate records and including additional relevant articles identified through the bibliographies of included records, three additional articles were assessed for eligibility. Of the 79 articles chosen for the screening process, 10 were selected to be included in this review. The flow of included studies is outlined in Figure
Citation | Animal model | Comparison groups | EF manipulation | Diets | Testing duration | Measures taken | Results | ||
---|---|---|---|---|---|---|---|---|---|
Food intake | Anthropometrics | Cardiometabolic/hormonal | |||||||
Anderson et al. ( |
Rat pups (no specific model system noted) | Gastrostomy coupled with HFF vs. LFF | Gastrostomy tubes inserted 24 h of age; standard formula provided with daily amount = 0.5 kcal/g body weight | 5–7 d | |||||
Atalayer and Rowland ( |
Albino mice | Feeding opportunities available during 12 h dark period and 1st 4 h of light period: 4×/d = access to food for 40 min at beginning of every 4th h; 8×/d = access to food for 20 min at beginning of every 2nd h; 16×/d = access to food for 10 min at beginning of every h; all groups had 160 min/d access to food; size of feeding opportunity depend on food cost (an additional manipulation) | 20 mg Purina chow pellets (10.4% kcal from fat + 24.1% kcal from protein) | 3–4 d | ≠ | ≠ | |||
DeVries et al. ( |
Holstein cows | Experiment 1: |
1×/d at 5:30 a.m.;
2×/d at 5:30 a.m. and 3:15 p.m.; 4×/d at 5:30 a.m.,
11:00 a.m., 3:15 p.m., and 10:30 p.m. |
Experiment 1: total mixed ration 1 (51.2% concentrate and 48.8% forage) |
3 d adjustment period followed by 7 d observation period | ≠ (In both exp. 1 and 2) | |||
Nussbaum et al. ( |
Simmental-red Holstein calves, Braunvieh-Brown Swiss calves, and Holstein Friesian calves | LFF at 8:00 a.m. and 5:00 p.m. by bucket feeding; HFF ranged from 6 to 14×/d by automated computer feeding |
Colostrum (d 1–3) followed by mature milk powder (d 4–14) and finally mature milk (d 15–28); bucket feeding of colostrum and milk powder contained water in concentration of 4 g/100 g and 5 g/100 g, respectively; automated feeding ranged from 0.5 to 1.5 L per portion | 28 d | ≠ | ||||
Vicari et al. ( |
Holstein-Friesian Calves | 1×/d at 12:00 p.m.;
2×/d at 12:00 p.m. and 12:00 a.m.; 4×/d at 6:00 a.m.,
12:00 p.m.; 6:00 p.m., and 12:00 a.m. |
Experimental milk replacer diet | 14 d for period 1; 28 d washout period; 14 d for period 2 | ≠ | ≠ |
Citation | Animal model | Comparison groups | EF manipulation | Diets | Testing duration | Measures taken | Results | ||
---|---|---|---|---|---|---|---|---|---|
Food intake | Anthropometrics | Cardiometabolic/hormonal | |||||||
Mantysaari et al. ( |
Finnish Ayrshire Cows | Fed 1×/d vs. 5×/d |
Diet: total mixed ration (grass silage and concentrate mix); concentrate mix contained 60.6% barley, 27% rapeseed meal, 10% molasses sugar beet pulp, and 2.4% vitamin and mineral mix | From calving to 28 weeks lactation | HFF ↓ LFF | ≠ | |||
Muiruri and Leveille ( |
Sprague-Dawley rats | Purified diet (70% glucose, 19% casein, and 12% fat) | 6 weeks (3 weeks adaptation; 3 weeks experiment) | Group 4 ↑ |
|||||
Robles et al. ( |
Holstein Heifers | 1×/d at 8:00 a.m.;
2×/d at 8:00 a.m. and 8:00 p.m.; 3×/d at 8:00 a.m., 2:00 p.m., and 8:00 p.m.; 4×/d at 8:00 a.m., 12:00 p.m., 4:00 p.m., and 8:00 p.m. |
Concentrate diet and barley straw | 4, 2-week periods | ≠ | ||||
Steelman et al. ( |
Quarter Horse Yearlings | 2×/d at 7:00 a.m. and
7:00 p.m.; 3×/d at 7:00 a.m., 3:00 p.m., and 11:00 p.m.; 4×/d at 1:00 a.m., 7:00 a.m., 1:00 p.m., and 7:00 p.m. |
Concentrate diet (pellets) and Bermuda grass hay | 33 d | ≠ | ||||
Wu et al. ( |
Rare Minnow | Factorial experiment of temperature × feeding frequency |
NR | 8 weeks | 2×/d > 3×/d > 1×/d at ambient temp |
Table
The five studies with an experimental period of <1 month included an assortment of animal models, such as rat pups, mice, calves, and cows, with 5–56 animals per condition (
The five studies with an experimental period of >1 month included cow, rat, horse, and rare minnow models, with 3–25 animals per condition (
The purpose of this systematic review was to provide a comprehensive review of experimental research conducted in both humans and animals in the areas of greater EF, food intake, and body weight. Twenty-five studies, using varying study designs, EF manipulations, and lengths of experimentation, were identified and included in the review (
When potential mechanisms were examined, four studies reported on the effect of greater EF on energy expenditure (one study reported on two different measures of energy expenditure), with two studies (50.0%) finding no influence of EF, one study reporting reduced energy expenditure with greater EF, and one study not reporting significance (
The hypothesis that increased EF may influence energy intake and/or anthropometrics continues to be sustained in the literature. For example, evidence cited in previous reviews on EF have suggested that greater EF may not have a strong impact on energy intake and/or anthropometrics, yet these reviews have still concluded that there may be an effect of EF on energy intake or anthropometrics (
As the initial research which suggested that greater EF may influence intake and anthropometrics was observational, and issues about accuracy of self-reported dietary intake have been suggested as a factor in producing results that indicate that greater EF is related to a lower weight status (
As this review did not include animal models of diseased states or human participants with health conditions other than overweight or obesity, greater EF may have an influence on energy intake, anthropometrics, and/or cardiometabolic/hormonal outcomes in human or animal models demonstrating impaired glucose or lipid metabolism (i.e., diabetes and cardiovascular disease). Additionally, it has been suggested that physical activity may be an important variable in the relationship between greater EF and energy intake and anthropometrics (
There are several limitations to this systematic review. Within the studies, there is great diversity of the EF manipulations, the types of measures collected, and when during the EF manipulation the measures were collected. This makes it challenging to pool effect sizes across the investigations. Additionally, the animal studies used several animal models, and the models may have differing “typical” eating patterns (i.e., animals that typically graze vs. animals that do not), producing a potential differential response to EF manipulations. These differing models again make it challenging to pool outcomes. Finally, while many studies used experimental designs that capitalized on using within-subject factors, the sample sizes in the investigations as a whole were small, which could indicate that the studies may be underpowered to detect differences in outcomes between EF conditions. These limitations reduce ability to draw firm conclusions regarding the effect of EF on energy intake and anthropometrics.
In summary, the human and animal experimental studies included in this review suggest that greater EF may not necessarily influence energy intake or anthropometrics. This indicates that contrary to what is commonly proposed in the lay literature, eating more frequently during the day (i.e., “grazing”) may not assist with reducing energy intake or improving weight status.
HR and GC developed study premise. HR, MG, and GC reviewed all identified abstracts, and HR and MG reviewed all identified articles. HR, MG, SP, and GC abstracted data from articles. HR, MG, and SP developed initial draft of paper, and all authors contributed to and approved the final version of the manuscript.
The authors declare that the research was conducted in the absence of any commercial or financial relationships that could be construed as a potential conflict of interest.
AUC, area under the curve; BMI, body mass index; EF, eating frequency.