Humans appear to have innate energy regulation mechanisms that manifest in sensations of satiation during a meal and satiety post ingestion. Interactions between these mechanisms and the macronutrient profile of their contemporary food environment could be responsible for the dysregulation of this mechanism, resulting in a higher energy intake. The aim of this systematic review was to determine the impact of dietary fibre and fat both in isolation and combination on satiation and satiety.
A systematic review of the literature was undertaken, from inception until end December 2017, in accordance with the PRISMA guidelines, in: Scopus, Food Science and Tech, CINAHL, and Medline databases. The search strategy was limited to articles in English language, published in peer-reviewed journals and human studies. Studies were selected based on inclusion/exclusion criteria.
A total of 1490 studies were found initially using the selected search terms that were reduced to 12 studies suitable for inclusion. Following on from this, a meta-analysis was also conducted to determine any satiety effects from any potential interaction between dietary fat and fibre on satiety, no significant effects were found.
Owing to high energy density, fat (per kJ) had a weak effect on satiation as determined by the effect per gram for each unit of energy. The addition of fibre theoretically improves satiety by slowing the absorption of various nutrients including fat, although the meta-analysis as part of this review was unable to demonstrate an effect, perhaps reflecting a lack of sensitivity in research design. The potential to improve satiation and satiety responses by consuming fat together with carbohydrates containing fibre warrants further investigation.
The global prevalence of obesity has been increasing, this near uniform rise is despite diversity in the genetic, food, and living environment as well as economic factors. Globally, the proportion of adults with a body mass index of 25 kg/m² or greater increased from 28.8% to 36.9% in men, and from 29.8% to 38.0% in women between 1980 and 2013 . Increasing energy intake has been the conventional explanation for weight gain at the individual  and at the societal level. This simplistic model is no longer sufficient and there is a need for a re-examination of conventional explanations and an exploration into other possible factors driving the obesity crisis such as possible effects of biology on behaviour, and how these interact .
There is an obligate requirement to consume dietary fat in order to maintain good health [4, 5]. The diets of a number of hunter gatherer groups were heavily dependent on energy-dense, if at times scarcely available, animal-based foods . Historical and anthropological studies have shown hunter-gatherers generally to be healthy, fit, and largely free of the degenerative cardiovascular diseases common in modern societies . Thus, fat intake per se may not be of greatest relevance to the development of obesity but rather macronutrient intake in the context of other dietary components with which it is consumed, such as dietary fibre, as in the present review.
Fat is highly desirable to humans and this may be due to various factors beyond its nutritional properties, such as cultural and psychological influences . As economic development progresses, consumers consistently make dietary choices which incorporate increased levels of fat and this behaviour is most clearly evident today in the changing dietary choices of consumers in developing countries [9, 10]. Yet high dietary fat intake is considered to be a significant contributing factor to increasing rates of overweight and obesity over recent decades .
Fibre is a type of carbohydrate not digestible by humans . Dietary fibre is known to be associated with improved satiation, satiety, and reduced food intake . Fibre has been shown to improve the ability of those on a low energy diet to maintain compliance . Conversely, in observational studies, consumption of diets which are high in fat but low in fibre content have been shown to be associated with increased levels of overweight [15, 16].
Mechanisms for maintaining weight are thought to exist in all animals. In humans, food intake is controlled by appetite signalling through a complex system of hormones [17, 18]. Digestive peptide hormones such as ghrelin, gastrin, and cholecystokinin are responsible for control of food intake through a complex process of biochemical responses, interactions, and feedback mechanisms. By such means, excessive energy intake can be balanced by a compensatory reduction in energy intake through alterations of appetite .
Satiation and satiety are crucial concepts in the understanding of digestion and are associated with gut hormone signalling . Satiation refers to physiological responses to food intake during the consumption of food that leads to a cessation of eating, whereas satiety refers to physiological responses that delay the taking of the next meal .
Several reviews have examined the effect of fibre on satiation and satiety [22,23,24]. Other individual studies have looked at the addition of fibre and its effect on the rate of absorption of nutrients, such as fat, on satiety . This systematic review was conducted to evaluate the effect on satiation and satiety of the co-ingestion of fat and fibre and to identify any gaps in the current literature. A meta-analysis was also conducted on the satiety effects that result from the interaction between fat and fibre.
This systematic review was conducted according to The Preferred Reporting Items for Systematic Reviews and Meta Analyses (PRISMA) Statement , it was registered with PROSPERO (CRD42017064244).
An electronic literature search of peer-reviewed journal articles was conducted across four databases—Scopus, Food Science and Tech, CINAHL, and Medline. Medical subject headings were considered in the development of the search terms (Section 2.2). The search was limited to English language and human studies. Studies from inception to December 2017 were included. Titles, abstracts, and methods were screened by the lead author for relevance based on selection criteria (Section 2.3) and any duplicate papers arising from separate searches were removed. The articles deemed relevant were selected for further consideration (Fig. 1). The search was further limited to peer-reviewed original research articles with full text available. Review, meta-analysis and other types of papers (e.g., author manuscript, letters to editors) were excluded. In addition, a hand search was undertaken of the reference lists of relevant articles (including reviews), and those deemed eligible were included with the intent to ensure that all articles relevant to the research had been captured by the search strategy. The abstracts of articles deemed relevant by the lead author were then independently reviewed by a further investigator (D.M.) for relevance. If consensus was not reached, the article was moved on to the next stage for a review of the full text. The full-texts of the remaining eligible studies were independently reviewed by the investigators (A.W. and D.M.) against the inclusion and exclusion criteria. Any final discrepancies were resolved by referral to a third researcher (K.P.). Disagreements were discussed until consensus was reached in all cases.
The search terms included: (“low fat” OR “high carbohydrate” OR “high fat” OR “low carbohydrate”) AND (“high fiber” OR “increased fiber” OR “high fibre” OR “increased fibre”) AND diet AND (satiation OR satiety). Search engines were utilised for the search which was limited to randomised controlled trials, randomised crossover studies and experimental studies published since inception until the end of 2017.
Selection criteria and data extraction
Studies were eligible for inclusion if they: (1) examined the effect on satiation or satiety of diets containing fat and carbohydrates with varying levels of dietary fibre; (2) interventions were acute—each test or control meal was administered within the same day; (3) utilise healthy individuals of either normal weight or overweight/obese with no age constraints. Studies where subjects suffered from any metabolic disorder such as diabetes, studies on bariatric patients and patients taking a weight loss drug were excluded. Also excluded were: animal studies; studies where the primary focus is psychological; studies where the primary focus is exercise; studies with participants taking a psychiatric medication; and studies where participants are taking any drug which may decrease hunger. Any studies that had a washout period of < 6 days were also excluded.
Data extraction of included studies was completed by two independent reviewers (A.W. and D.M.). Information such as study title, study type, number of subjects, randomisation, blinding, nutrients assessed, washout period, outcomes, and other sources of potential bias were extracted from each study.
The primary means of data synthesis was the collection of data from visual analogue scales (VAS) used in each study to assess appetite. VAS are a standardised system of appetite assessment that requires a response to a question relating to the degree to which a subject is experiencing a certain aspect of appetite sensation .
VAS satiety ratings were examined as part of the meta-analysis. Results were assessed by measuring fullness/hunger at a particular time point. For the purposes of this review, a time point closer to the subsequent meal was anticipated to have greater relevance. Three hours post baseline was selected as the maximum time point measured, which was common to all studies and so this was the time point used to assess satiety in the meta-analysis.
Where numerical data were not available, authors were contacted. If authors did not respond and supply data, then estimates of the results were made using graphs published in each study. If a sufficient quality of data was not able to be synthesised in this manner the study was excluded.
Additional measures of satiety such as the effect on blood plasma hormone levels were also assessed where possible.
The review software Revman (Version 5.3. Copenhagen: The Nordic Cochrane Centre, The Cochrane Collaboration) was used for the meta-analysis. The outcome to be assessed in the meta-analysis was the change in mean between those treatments including fat and those including fat with higher levels of dietary fibre. Changes in satiety were measured by subtracting the VAS score for fullness/hunger rating at baseline from the VAS score for fullness/hunger at 3 hours postprandial.
Heterogeneity, risk of bias, and quality assessment
Heterogeneity between studies in the meta-analysis was determined in two ways: first, by visual inspection of a forest plot generated by the meta-analysis and, second, by means of a calculation of I². I² provides an estimate of the proportion of variability in a meta-analysis that is explained by differences between the included trials rather than by sampling error .
Risk of bias was assessed by considering the level of randomisation of each study, the degree to which the study was blinded and other potential sources of bias. The Physiotherapy Evidence Database (PEDro) [29, 30] scale was used to assess the risk of bias in the studies.
Statistical significance was set at a Confidence Interval of 95%. Excel was used to generate the spreadsheets and to calculate means, difference in means, standard error of mean, and standard deviations; converting between measures of spread where necessary.
A total of 1490 studies were found initially using the selected search terms. By the removal of duplicates and excluding studies based on a reading of the study abstracts, this number was reduced to 141 studies. There were 10 studies identified as meeting the criteria for inclusion in the review. The two additional studies dealing only with the effect of increases in dietary fibre were acute randomised crossover trials obtained from the reference lists of reviews and were included after a consensus decision by all researchers that they fully met the inclusion criteria [31, 32].
All studies included in the review assessed satiety by means of a VAS. Four of the studies examined the effect on blood plasma hormonal excretions. The characteristics of each of the studies used in the review is summarised in Table 1.
Heterogeneity, risk of bias, and quality assessment
All of the studies included in this review were randomised controlled trials using a crossover design. Aside from randomisation, efforts to reduce bias in many of the studies were minimal. Only four of the studies were blinded [33,34,35,36], two were double-blinded [33, 35]. Assessment of risk of bias is reported in Table 2.
Characteristics of studies included in the meta-analysis
Two separate meta-analyses were conducted: one on fullness ratings and the other on ratings of hunger.
Three studies were excluded from both analyses for not having a comparable measure of VAS results—VAS results shown as area under the curve rather than at specific time points [37,38,39]. Two studies were excluded due to a lack of inclusion of error measurements with the published data [40, 41]. Attempts were made to contact authors in each of these cases but were unsuccessful.
For the analysis of fullness, out of the remaining seven studies, five studies were included in the meta-analysis [42,43,44,45,46] and two were excluded [47, 48] owing to the lack of a rating for fullness as these two studies were concerned with hunger and measured this aspect accordingly. Two of these [43, 46] were broken down into subgroups by age or gender to form a total of thirteen separate cohorts included in the analysis.
For the hunger analysis, out of the remaining seven studies, five of the studies [42,43,44, 47, 48] were included in the meta-analysis and two were excluded [45, 46] owing to the lack of a rating for hunger [45, 46] as these two studies were concerned with the maintenance of fullness.
Studies used in the meta-analysis were concerned with the effect of changes in dietary composition on fullness and hunger. All the studies used a crossover type design with the subjects acting as their own controls. Participants would be exposed to either the test treatment or control according to into which group they had been randomised. Subjects would return at a time at least 1 week after the initial experiment in which the treatment order was reversed.
Interventions would typically be composed of a test meal containing fat with some additional amount of fibre incorporated into the meal, or given as an additive, and a control. Test meals included high fat [45, 48], low fat , and mixed [42,43,44, 46] treatments. One study tested the effect of two different fibres  and the results of both these treatments were included in the meta-analysis on hunger effects.
Studies tested isoenergenic treatments almost exclusively [42, 43, 45, 47, 48] although some studies tested treatments with different energy levels for within-group comparisons [44, 46]. Changes in appetite would be measured at various time points until the conclusion of the experiment. Three studies also measured the effect on the secretion of various digestive hormones [43, 47, 48] and four studies examined subsequent energy intake [42,43,44, 47].
A total of 592 participants were examined in the meta-analysis on fullness and 116 participants in the hunger analysis. One study in the meta-analyses assessed only males  with the remainder being mixed gender studies. One study broke down the participants by gender to examine effects on males and females as subgroups . Although child studies did not form part of the exclusion criteria of the review, all participants in the studies were adults. Ages of participants ranged from 18 to 68 years. One study that broke down participants in different age groups  was included in the analysis on fullness. All subjects were healthy although studies which included overweight participants were also utilised. The hunger analysis included one study which included obese patients .
Main findings of meta-analysis
Six of the cohorts in the meta-analysis on fullness showed an increased effect at the 3-hour post-prandial time point. According to our analysis the Zhou study  produced conflicting findings between the various age subgroups with the 20–30 years age group showing an increase in fullness, whereas all other age groups showed a decrease. Three studies in the hunger meta-analysis reported a decrease in hunger at the 3-hour post-prandial time point.
No effect was found for an increase in fullness (CI 95%: − 2.39, 2.26, p = 0.96) and a non-significant 5% reduction (CI 95%: − 12.76, 1.82, p = 0.14) was detected for hunger, (Fig. 2).
The value for I2 in the meta-analysis on fullness was 81 and 84% in the hunger meta-analysis indicating a high degree of heterogeneity among the studies.
Analysis of studies not included in the meta-analysis
Test meals included high fat , low fat , and mixed [35, 36, 49] treatments. Studies tested isoenergenic treatments , another a combination of isoenergenic and non-isoenergenic treatments  and others used a preload design [33, 35, 36]. Changes in appetite were measured at multiple time points until the conclusion of the experiment. One study also measured the effect on the secretion of various digestive hormones  and four studies examined subsequent energy intake [33, 35, 36, 49].
Similar to the studies included in the meta-analysis, subjects of the studies not included in the meta-analysis were of normal weight or overweight. Two studies in this systematic review that were not included as part of the meta-analysis utilised obese subjects [39, 48]. Ages of subjects ranged from 18 to 50 years.
Four of these papers found some influence on appetite in diets containing fat with increased levels of dietary fibre [38,39,40,41]. All studies detecting an increased satiety effect attributed this effect mainly to the actions of the dietary fibre. One study noted that lower energy levels of the high fibre treatment resulted in lower carbohydrate uptake increasing levels of non-esterified fatty and greater fat oxidation . Whether this higher level of fat oxidation had any effect on satiety levels was not explored in the study. Improved satiety through slowing the absorption of nutrients in the small intestine [37, 40] or slowed gastric emptying  through the actions of increased viscosity of soluble fibres  was cited in several of these papers.
Despite a focus on the weaker satiety-inducing properties of fat-rich foods, one study , which tested the effect on satiety of four breakfasts, detected an increased level of satiety and lower subsequent energy intake of an isocaloric egg and bacon (EB) breakfast containing fat and protein but no carbohydrate compared to a croissant breakfast containing carbohydrate and a similar level of fat but low levels of protein and fibre. The degree to which this effect was due to the higher levels of protein in the EB breakfast, as opposed to any potential increase in satiety resulting from slower absorption of the meal in the small intestine, remains unclear.
Only one paper found no increase in satiety levels or any change in energy intake from a test meal containing increased levels of dietary fibre  but this study was concerned with a hypothesised increase in satiety from the fermentation of fibre in the colon. It should be noted that as an acute study there was a limitation on this study’s ability to detect a chronic fermentation. Furthermore, the number of types of fibres examined was limited to four particular types of fermentable fibres and may not be applicable to dietary fibres in general.
No significant effects on hunger were found when a diet containing fat had an increased fibre content (five percent reduction; p = 0.14) with no corresponding increase in fullness. The results indicate that any effect from an intervention of this kind, should one exist, may be associated with a delay in the return of hunger as opposed to maintenance of fullness. This is in agreement with the data from other studies examining potential enhancement of satiety using other macronutrient compositions such as high protein diets  and ketogenic diets .
An inability to detect a clear increase in satiety in this study may be due to the relatively short post-prandial time point of three hours at which the satiety measurements were made. The impact of changes in satiety may be more meaningful if they were made closer to the commencement of the subsequent meal. This highlights potential limitations in interpreting data from these types of studies. It may be more applicable to public health if future investigations into the nature of hunger by assessment of this aspect of appetite for a longer period—even up to 5–6 h—post meal ingestion and ideally longer, although perhaps logistically less practical.
The decision to choose the 3-hour postprandial time point, at which to measure satiety, was made to try and obtain comparable data and minimise heterogeneity of that data from a range of studies with differing methodologies and outcome measures. This may have not been optimal, for the hypothesised effect of fibre from fruit and vegetables may fall prior to this, and any effects of fat or other nutrients, e.g., via cholecystokinin may occur at a later point post prandially. Findings from Zhou et al.  may reflect these dynamics. In addition, it is important to note that Zhou et al.  assessed fullness and satisfaction but not hunger.
The challenge in interpretation of the data is increased by the variability of satiety study designs, including duration, the use of free-living or laboratory studies, isoenergenic, or ad libitum studies, the statistical power of the study, prior diet of the subjects, and the intrinsic risk of subjectivity associated with the assessment satiety, appetite, and hunger . This aspect is further complicated by a lack of consistency in terminology used in describing the various aspects of appetite sensation. Although there is a fixed convention in use, some researchers use terms with differing meanings interchangeably.
This review utilised studies that assessed appetite using a VAS scale. However, a limitation could be how different researchers used differing VAS statements and scales, which could have had a bearing on the results. The development of a consensus on a uniform scale of VAS for appetite studies, would be of benefit for future research in this field. In addition, VAS scales tend to produce a polarity in results, there for manipulation of data derived from VAS scales such as calculation of sums and differences may not be appropriate and this is another of a multiplicity of statistical issues relating to the use of data derived from VAS scales, again making cross-study comparisons challenging .
Fibre and satiation
Six of the studies in this review assessed the effects of a limited fibre [37,38,39,40, 47, 48]. However, different fibres have different physical and chemical properties and these in turn have varying effects on human digestion and thus appetite. Further, a distinction must be made between fibre naturally occurring within a food matrix and that which is added to a food. Other systematic reviews have assessed a wide variety of dietary fibre on satiety [22, 24] with varying results but generally reporting a small effect. The volume of ingested food has been shown to be primarily responsible for satiation independent of energy content . As such, the greater volume taken up by fibre containing carbohydrates displaces food content with greater energy density, delivering satiation, and assisting in regulating energy intake at levels not deleterious to the human metabolism . However, these reviews have assessed the effect on satiety of diet in general and not considered how this effect would be exaggerated by the coupling of fat with fibre. As both satiation and satiety are involved, the potential for utilising such a change in dietary composition would be best demonstrated over the long term (months to years). Based on this, there is a need for more longitudinal studies examining the chronic effects of dietary composition on satiation and satiety.
Fat and satiation
Most of the studies in this systematic review found a very weak effect of fat on satiation [41, 43, 44, 46] as determined by the effect on satiation per unit of energy, which is consistent with other findings in this area [31, 50,51,52]. This effect has also been seen with increasing fat and energy intakes in populations and associated increases in rates of obesity [9, 10, 32, 53].
It is important to note, the view of single nutrients is perhaps reductionist, as macronutrients, including fat, are almost never consumed on their own. Interactions with other nutrients and their effects on physiological, biochemical, and other factors are important to consider . As in the present study, fat’s weak effect on satiation may be influenced by the other nutrients with which it is consumed, such as bulky high fibre carbohydrates.
Fibre and satiety
One mechanism through which fibre has been associated with an increase in satiety is via delaying gastric emptying. Three studies in this review examined the effect of a delay in gastric emptying on satiety [39, 40, 44] and found a correlation between an increase in gastric emptying time and satiety. Of these, one  found that the addition of a gel-forming liquid fibre that caused a delay in gastric emptying tended to decrease hunger, increase fullness, and reduce the amount of food subjects wanted to eat but had a lesser effect on the desire to eat. The discrepancy between sensations recorded before the test meal and the actual amount of food consumed subsequently may be explained by the possibility that the delayed gastric emptying prevented nutrients entering the small intestine and hence the inhibitory effects on food ingestion owing to hormonal factors were absent.
It is generally considered that an increase in fibre will result in additional delay in gastric emptying  but the satiety-enhancing effects of fibre are now thought to be due in larger part to slower absorption of nutrients rather than to a delay in gastric emptying . Slower absorption of nutrients is considered to occur in two ways: through physical obstruction of the nutrients in the digestive tract by insoluble fibre  and also by an increase in the viscosity of the small intestinal digesta caused by the presence of soluble fibre. Although solubility of fibres was previously thought to be the most important variable having an influence on satiety, more recently the change in viscosity of fluids present in the digestive tract is now thought to be of greater relevance . An increase in short chain fatty acids through fermentation via microflora in the colon is another means by which satiety may be delivered by insoluble fibre and this was examined by one of the papers in this review , although this study failed to find an effect.
Fat and satiety
Through its effects in promoting the secretion of the digestive hormone cholecystokinin, which is known to mediate a slowing in the release of nutrients from the stomach , fat will cause a delay in gastric emptying. Unlike with dietary fibre, this phenomenon may be explained by a slowing in gastric emptying commensurate with fat’s greater energy density. Specifically, the rate of gastric motility has been shown to be regulated by the energy density of nutrients consumed . Therefore, any possible increase in satiety caused by a delay in gastric emptying would reflect fat’s higher energy density.
Studies in this review have demonstrated differing levels of satiety in meals containing fat. A high-fat breakfast reduced energy intake at a subsequent lunch but resulted in higher energy intake at dinner and a higher energy intake overall for the day . Another test breakfast with a higher fat and protein content provided greater satiety . It appears that the satiety derived from fat may be variable, dependent on the other nutrients with which it is consumed and this is also true when fat is consumed with dietary fibre.
Fibre and fat interaction and satiety
The results of this review suggest that an increased level of fibre results in an increase in satiety. Other studies in this review have identified an increase in satiety to levels similar to that gained from high-fat treatments through a combination of low levels of fat with fibre  and a significantly improved satiety response and secretion of higher levels of digestive hormones such as glucagon-like peptide 1 when coupling a high-fat meal with a high fibre content .
As the benefits from a slower absorption effectively amount to a more efficient return in terms of sensory feedback from the ingestion of nutrients, it logically follows that greater utility would be derived by combining this effect with more energy-dense nutrients. Therefore, it is entirely possible that the significantly reduced satiety as a result of the more rapid digestion of fat may accelerate the weight gain associated with a diet both high in fat and low in fibre content.
The findings from this review suggest that future research in this area would be best directed at investigating the deferment of hunger rather than maintaining sensations of fullness. This is in agreement with other studies on satiety working with other macronutrients such as protein  and also on ketogenic diets . It is plausible that delaying the onset of hunger has greater relevance and utility than maintaining sensations of fullness in the context of a potential reduction in energy intake. Further, more may be gained by assessment of hunger for a longer period of time post-meal ingestion.
The main limitations of this systematic review were the general low quality of studies and the heterogeneity in the study protocols. Studies used a small number of participants and although studies were randomised, details on how randomisation was conducted were not reported. Owing to the nature of food-based studies, few of the studies were blinded and only two studies stated that they were double-blinded.
Issues also were identified relating to assessment of subjective ratings of appetite. VAS scales have a variety of statistical and other issues limiting their ability to be usefully manipulated as data. The lack of a uniform scale for VAS would also have an impact on the quality of data.
Owing to high energy density, fat has weak effect on satiation as determined by the effect per kJ of energy. Carbohydrates that contain high levels of dietary fibre are effective at producing satiation at relatively low levels of energy intake. This is owing to the greater volume occupied by carbohydrates containing fibre relative to those that do not. The fibre content slows the absorption of fat consumed with it, thereby allowing an optimal level of satiety to be derived from the fat relative to its energy content. A delay in hunger rather than an increased fullness is more likely, however, this review was unable to conclusively find such an effect. The potential to improve satiation and satiety by consuming carbohydrates containing fibre together with fat, thereby preventing or at least minimising weight gain warrants further investigation.
We acknowledge the support provided by the University of Canberra in the creation of this review.