Effect of Polydextrose on Subjective Feelings of Appetite during the Satiation and Satiety Periods: A Systematic Review and Meta-Analysis

Introduction: Subjective feelings of appetite are measured using visual analogue scales (VAS) in controlled trials. However, the methods used to analyze VAS during the Satiation (pre- to post-meal) and Satiety (post-meal to subsequent meal) periods vary broadly, making it difficult to compare results amongst independent studies testing the same product. This review proposes a methodology to analyze VAS during both the Satiation and Satiety periods, allowing us to compare results in a meta-analysis. Methods: A methodology to express VAS results as incremental areas under the curve (iAUC) for both the Satiation and Satiety periods is proposed using polydextrose as a case study. Further, a systematic review and meta-analysis on subjective feelings of appetite was conducted following the PRISMA methodology. Meta-analyses were expressed as Standardized Mean Difference (SMD). Results: Seven studies were included in the meta-analysis. There were important differences in the methods used to analyze appetite ratings amongst these studies. The separate subjective feelings of appetite reported were Hunger, Satisfaction, Fullness, Prospective Food Consumption, and the Desire to Eat. The method proposed here allowed the results of the different studies to be homogenized. The meta-analysis showed that Desire to Eat during the Satiation period favors polydextrose for the reduction of this subjective feeling of appetite (SMD = 0.24, I2 < 0.01, p = 0.018); this effect was also significant in the sub-analysis by sex for the male population (SMD = 0.35, I2 < 0.01, p = 0.015). There were no other significant results. Conclusion: It is possible to compare VAS results from separate studies. The assessment of iAUC for both the Satiation and Satiety periods generates results of homogeneous magnitudes. This case study demonstrates, for the first time, that polydextrose reduces the Desire to Eat during the Satiation period. This may explain, at least in part, the observed effects of polydextrose on the reduction of levels of energy intake at subsequent meals.


Introduction
The definition of appetite covers the whole field of food intake including selection, motivation, and preference. It also refers specifically to qualitative aspects of eating, sensory aspects, and responsiveness to environmental stimulation that can be contrasted with the homeostatic view based on eating in response to physiological stimuli and an energy deficit [1]. Appetite is measured by several means including the assessment of subjective feelings, food intake, gastrointestinal hormones, and the gastric emptying rate, to mention the most relevant ones. making it difficult to compare studies using different lengths of VAS scales (e.g., 100 mm paper and pencil vs. 64 mm electronic system). Further, the AUC usually reports both Satiation and Satiety as integrated periods in the analysis.
The incremental AUC (iAUC) has many characteristics of the AUC but has the advantage of adjusting the values at baseline [7]. Unfortunately, in many studies, this adjustment is done at the pre-meal timepoint, integrating both Satiation and Satiety periods in the analysis. It would be preferrable to analyze them separately. Therefore, the aim of this review is to find and propose a methodology based on iAUC during the individual Satiation and Satiety periods. This will generate homogenous magnitudes from the independent studies and allow us to compare the subjective feelings of appetite in a meta-analysis.
The use of iAUC/AUC in meta-analysis may seem challenging due to the differences in observation period length in the studies. However, since all the observation periods of the studies contain the phenomena of interest and the use of Standardized Mean Difference (SMD) in calculation of the effect size compensates for the differences in the experimental settings of the studies, commensurable effect sizes for the studies are obtained [8].
The studies conducted on appetite suppression using polydextrose are proposed as a case study. Recently, a meta-analysis showed that polydextrose effectively reduces voluntary energy intake at a subsequent meal, especially when it is administered as part of a mid-morning preload before an ad libitum lunch [9]. Polydextrose has also been associated with lower levels of ghrelin, and higher increases in glucagon-like peptide 1 (GLP-1) and peptide tyrosine-tyrosine (PYY) [10,11]. This evidence makes this food ingredient an ideal candidate to test new methodological concepts to assess appetite suppression.

Protocol Registration
This review was conducted according to the methodology described by the Preferred Reporting Items for Systematic Reviews and Meta-Analyses: PRISMA Statement [4]. The Protocol was registered at the International Prospective Register of Systematic Reviews (PROSPERO) with number CRD42013005261 on 9 August 2013. The methodology was used to analyze the available data on the effects of polydextrose on subjective feelings of appetite and levels of energy intake. This report communicates the results on the subjective feelings of appetite. Results on different levels of energy intake are communicated in a separate report [9].

Eligibility Criteria and Information Sources
Eligible study designs were acute or chronic, randomized, and used placebo-controlled nutritional interventions where polydextrose was administered alone or in combination with other foods or food ingredients including supplements. Participants were males and females rated as normal, overweight, or obese, but otherwise healthy. The chosen interventions were those intended to assess the effects of polydextrose on appetite ratings and energy intake when available. Subjective feelings of appetite included, but were not limited to Hunger, Satisfaction, Fullness, Prospective Food Consumption, and the Desire to Eat.
Eligible reports included papers from scientific journals, conference abstracts, and theses reported in English-language literature before 31 July 2013, except for a full report provided by Nerys Astbury originally published as an abstract [12], and a manuscript from Kaisa Olli, later published as Olli et al. [11]. Searches were conducted on the following databases: BIOSIS Previews, CAB Abstracts, Foodline: Science, FSTA, Medline, SciSearch, Science Direct, Wiley Online Library, and www.ClinicalTrials.gov. Further information on recently completed trials, unpublished research, and research reported in grey literature was identified by searches for relevant documents hosted on Google Scholar.

Study Selection and Quality Assessment
One researcher (Alvin Ibarra) screened and selected the records. The authors of the selected original articles were requested to provide any missing information and full data sets on anthropometric measurements, subjective feelings of appetite, and the levels of energy intake. A second independent researcher (Kaisa Olli), checked the assessment, and any discrepancies were resolved by consulting a third researcher (Kirsti Tiihonen). The reviewed articles that were considered not relevant for this study were noted along with the reason for their exclusion.
A similar system was followed to assess the risk bias of each included study. The assessment followed the procedure described in the Cochrane Handbook for Systematic Reviews of Interventions [8].

Strategy for Data Synthesis
The data analysis was divided into two main sections. First, a narrative synthesis focusing on the subjective feelings of appetite was conducted to compare the methodologies used. For example, the description of the way that subjective feelings of appetite were measured was scrutinized, the kinds of foods that were provided during the studies, and the manner in which polydextrose was administered were considered. The second section focused on the analysis of the subjective feelings of appetite using the methodologies proposed in this review.

Definition of Satiation and Satiety Periods
The periods are defined as follows: Satiation: VAS scores immediately before to immediately after the meal consumption. Satiety: VAS scores immediately after the meal consumption until immediately before the next meal.
For practical reasons, all points of measurement before or after both periods (in case they exist) were discarded or treated as separate cases.

iAUC Estimation and Interpretation
Once the periods were defined, all the VAS scores were transformed to the same scale (0-100) and the post-meal time point was shifted to time point 0. Then, the VAS score at time point 0 was subtracted from all the VAS scores in the Satiation and Satiety periods. In other words, the curve consisting of VAS scores was shifted to 0 in timepoint 0. After that, the iAUC values were calculated separately for the Satiation (iAUC Satiation ) and Satiety (iAUC Satiety ) periods using the trapezoidal rule. iAUC above the zero reference was assigned a positive value and below was assigned negative one, and both were expressed as minutes times millimeter (min.mm). Figure S1 of the Supplementary Materials provides an example for Hunger, Prospective Food Consumption, and the Desire to Eat; Figure S2 provides an example for Satisfaction and Fullness. These figures exemplify the behavior of an ideal appetite suppressive verum as compared to a placebo control. Notice that the magnitudes for the same subjective feeling of appetite during the Satiation and Satiety periods are opposites. This is an important point to take into consideration for the correct interpretation of iAUC Satiation and iAUC Satiety values of a specific subjective feeling over the appetite suppressive effect of a given food product.
Using the following rule, it is possible to correct the opposite values of iAUC Satiation and iAUC Satiety for each subjective feeling of appetite in order to ensure that the potential effect of a food product as an appetite suppressing agent is interpreted correctly: Multiply iAUC Satiation by´1: to indicate that negative values for Hunger "reduce hunger"; for Prospective Food Consumption, "lower the amount expected to eat"; and for the Desire to Eat, "reduce the desire to eat".
Multiply iAUC Satiety by´1: to indicate that positive values for Satisfaction "increase satisfaction"; and for Fullness, "increase fullness".
Note that the adjustments made to the methodology do not affect the significance of the meta-analysis.

Meta-Analysis
For this meta-analysis, data sets were investigated using a random effects model which considers our chosen studies to be a sample of a larger universe of studies. The model was chosen because there were minor differences in both study design and the participants' characteristics. Therefore, a common effect size could not be assumed for all the studies. The treatment-effect size was analyzed using SMD with a 95% confidence interval. The effect size was estimated using Hedges' g measure. The between-study variation was estimated using a restricted maximum likelihood approach. The results of the meta-analysis was visualized using a Forest plot, which illustrates the results of the individual studies as well as the summary random effect. The numbers of effect sizes, K, were reported; and the total heterogeneity in the dataset was tested using the Q and Higgins I 2 statistics. The publication bias was analyzed visually using Funnel plots and assessed using the Egger's test.
All statistical analyses were performed using software "R" version 3.2.2 (R Core Team, R Foundation for Statistical Computing, Vienna, Austria) [13], and metafor package version 1.9-8 (Wolfgang Viechtbauer, Maastricht, The Netherlands) [14]. The "R" code is available online at https://github.com/avoltusfi/pdx_vas_meta. Table S2 of the Supplementary Materials shows the number of hits obtained in each database during the screening. Twenty-two complete studies on the effects of polydextrose on energy intakes and subjective feelings of appetite were assessed for eligibility for inclusion in the systematic review and meta-analysis using the same criteria. The meta-analysis on energy intakes is published [9]. In the present report on subjective feelings of appetite, fifteen studies were excluded with reasons as provided by Ibarra et al. [9], the lack of original data being the main cause for exclusion. Most of the data used in this meta-analysis was delivered by authors of the original publications under confidentiality agreements.

Included Studies
The flow diagram on Figure 1 demonstrates the screening process used to determine how the included studies were selected.
Seven studies were included for the assessment of subjective feelings of appetite: Olli et al.  [19]. In the previous meta-analysis on energy intakes [9], two of these studies were not included [11,18] because they did not measure food intake. The original data on subjective feelings of appetite of one of the studies [10] was not accesible.
The studies included in this analysis represent a universe of 135 participants, of which 66 were males and 59 were females. The doses of polydextrose tested in these studies ranged from 6.25 to 25.0 grams. Table 1 summarizes the design, procedures, and main outcomes of these studies with respect to subjective feelings of appetite.

Risk of Bias
The only double-blind studies included in this analysis were conducted by Olli et al. [11] and Schwab et al. [18]. All other studies were single-blinded, meaning that whilst the volunteers were unaware of the treatments they were given, the investigators were aware. Single-blinded studies are susceptible to a high risk of bias. Further, King et al. [19] communicated that sixteen volunteers were enrolled in the study; but a later revision of the clinical report revealed that results on subjective feelings of appetite were calculated using only fourteen participants (7 male and 7 female), and results on energy intake on fifteen participants ( Table 1). The reasons for these discrepancies were not clarified in the original report leading to a high risk of bias for this study.
All studies included in this review used commercial polydextrose, Litesse Ultra ® , or Litesse Two ® , manufactured by DuPont. Studies conducted by Astbury et al. [12,15], were sponsored by the Biotechnology and Biological Sciences Research Council (BBSRC) of the United Kingdom and Mars UK acting as a private partner. All other studies were sponsored, at least in part, by DuPont.

Qualitative Results of Appetite Ratings
Five subjective feelings of appetite were identified and assessed using visual analogue scales (VAS). All studies measured Hunger, six studies looked at Fullness and the Desire to Eat, three studies evaluated Satisfaction, and two studies examined Prospective Food Consumption. Table 2 summarizes the methodology used to measure each of these subjective feelings of appetite. Some studies reported other parameters as measured by VAS, which were not necessarily related to appetite, such as assessment of participants' well-being or sensory characteristics of investigational products. These extraneous parameters were, therefore, excluded from this review.

Risk of Bias
The only double-blind studies included in this analysis were conducted by Olli et al. [11] and Schwab et al. [18]. All other studies were single-blinded, meaning that whilst the volunteers were unaware of the treatments they were given, the investigators were aware. Single-blinded studies are susceptible to a high risk of bias. Further, King et al. [19] communicated that sixteen volunteers were enrolled in the study; but a later revision of the clinical report revealed that results on subjective feelings of appetite were calculated using only fourteen participants (7 male and 7 female), and results on energy intake on fifteen participants ( Table 1). The reasons for these discrepancies were not clarified in the original report leading to a high risk of bias for this study.
All studies included in this review used commercial polydextrose, Litesse Ultra ® , or Litesse Two ® , manufactured by DuPont. Studies conducted by Astbury et al. [12,15], were sponsored by the Biotechnology and Biological Sciences Research Council (BBSRC) of the United Kingdom and Mars UK acting as a private partner. All other studies were sponsored, at least in part, by DuPont.

Qualitative Results of Appetite Ratings
Five subjective feelings of appetite were identified and assessed using visual analogue scales (VAS). All studies measured Hunger, six studies looked at Fullness and the Desire to Eat, three studies evaluated Satisfaction, and two studies examined Prospective Food Consumption. Table 2 summarizes the methodology used to measure each of these subjective feelings of appetite. Some studies reported other parameters as measured by VAS, which were not necessarily related to appetite, such as assessment of participants' well-being or sensory characteristics of investigational products. These extraneous parameters were, therefore, excluded from this review. Participants were asked to consume a standardized dinner at 20:00 the day before the test and to refrain from consuming alcohol and undertaking vigorous exercise. The day of the test, they had a standardized breakfast at home (10% daily energy expenditure), arrived at the laboratory at 10:45 and were served a preload at 11:00. Appetite ratings were collected at baseline, 0 min (immediately after preload) and 30, 60, 90 min later, and 0 (immediately after), 30 and 60 min after the test meal. The test meal was ad libitum (657 kJ/100 g). Then, participants were instructed to complete food diaries for the rest of the day.
There were no significant differences on subjective feelings of appetite between PDX groups and control.
Ranawana et al. [16] 400 g preload (0.0 g PDX) 400 g preload ( Participants were asked to refrain from consuming alcohol and undertaking vigorous exercise. The day of the test, they arrived at 8:00 at the laboratory, after 10 h fast. Appetite ratings were determined before the breakfast, after the breakfast, 1 h after the breakfast, before preload and 15, 30, 45 and 60 (just before ad libitum lunch) min after the preload, and after the lunch. The size of the first breakfast meal was measured and the same size was used for the next session.
There were no significant differences on subjective feelings of appetite between the PDX group and control. Participants were asked to consume a standardized dinner at 20:00 the day before the test and to refrain from consuming alcohol and undertaking vigorous exercise. The day of the test, they had a standardized breakfast at home (10% daily energy expenditure), arrived to the laboratory at 10:45 and were served a preload at 11:00. Appetite ratings were collected at baseline, 0 min (immediately after preload) and 30, 60, 90 min later, and 0 (immediately after), 30 and 60 min after the test meal. The test meal was ad libitum (657 kJ/100 g). Then, participants were instructed to complete food diaries for the rest of the day.
There were no significant differences on subjective feelings of appetite between the PDX group and control.
Schwab et al. [18] 400 mL drink (0.0 g PDX or SPB) 400 mL drink (     Changes from baseline (between preload to the next meal) using two-way ANOVA

iAUCSatiation and iAUCSatiety
The iAUC Satiation and iAUC Satiety values for the polydextrose and the control groups are reported in the Supplementary Materials for Hunger (Table S3), Satisfaction (Table S4), Fullness (Table S5), Prospective Food Consumption (Table S6), and the Desire to Eat (Table S7).

Meta-Analyses of Subjective Feelings of Appetite
Four indicators for the subjective feelings of appetite were included in the meta-analysis: Hunger, Satisfaction, Fullness, and the Desire to Eat. There were two studies that assessed Prospective Food Consumption (Hull et al. [17] and Ranawana et al. [16]), but their reporting of results was not consistent; therefore, they could not be analyzed together.
Results of the meta-analysis are summarized in Table 3. The only significant result (p = 0.018) in the whole group was observed with the Desire to Eat category with small effect size (0.24) [25]. Here, the results of the random effects model indicate that the meta-analysis significantly favors polydextrose for the "reduction of the desire to eat" over the placebo during the Satiation period ( Figure 2). The Higgins I 2 statistic for this variable was zero, evidencing the high consistency of the data. In addition, the indicator of Egger's test was not significant, confirming a low level of bias. The results of the other indicators of subjective feelings of appetite, including the Funnel plots, are presented in the Supplementary Materials ( Figures S3-S9).

iAUCSatiation and iAUCSatiety
The iAUCSatiation and iAUCSatiety values for the polydextrose and the control groups are reported in the Supplementary Materials for Hunger (Table S3), Satisfaction (Table S4), Fullness (Table S5), Prospective Food Consumption (Table S6), and the Desire to Eat (Table S7).

Meta-Analyses of Subjective Feelings of Appetite
Four indicators for the subjective feelings of appetite were included in the meta-analysis: Hunger, Satisfaction, Fullness, and the Desire to Eat. There were two studies that assessed Prospective Food Consumption (Hull et al. [17] and Ranawana et al. [16]), but their reporting of results was not consistent; therefore, they could not be analyzed together.
Results of the meta-analysis are summarized in Table 3. The only significant result (p = 0.018) in the whole group was observed with the Desire to Eat category with small effect size (0.24) [25]. Here, the results of the random effects model indicate that the meta-analysis significantly favors polydextrose for the "reduction of the desire to eat" over the placebo during the Satiation period ( Figure 2). The Higgins I 2 statistic for this variable was zero, evidencing the high consistency of the data. In addition, the indicator of Egger's test was not significant, confirming a low level of bias. The results of the other indicators of subjective feelings of appetite, including the Funnel plots, are presented in the Supplementary Materials ( Figures S3-S9).   When analyzing results by sex for the category Desire to Eat, the meta-analysis favors polydextrose but only in the male population. Table 3 shows that there were no other effects on subjective feelings of appetite in the whole groups or in the sex-specific sub-groups.

Discussion
Dietary fibers such as polydextrose are thought to impact Satiation and Satiety, which are processes that are involved in the body's appetite control system [26]. The classical definition of Satiation indicates that it leads to the termination of eating, accompanied by the satisfaction of appetite. Satiety is explained as the feeling of fullness, which hinders hunger and further consumption of food. Satiation and Satiety are both involved in limiting energy intake and are important in determining the total energy intake [2].
All the studies included in this analysis used methodologies which assess subjective feelings of appetite as validated in previous research [2,3,7,[20][21][22][23][24]. Although there were some minor variations, in general, the questions for each indicator for the feeling of appetite can be considered similar. This also applies to the two studies that used inventories in Finnish language (Olli et al. [11] and Schwab et al. [18]). However, the wordings used for questions at the upper and lower ends of the response scales were less consistent among studies. The biggest variation was found between the methodologies used to record changes on subjective feelings of appetite and the ways in which results were analyzed. Three studies asked participants to use a pencil to mark their subjective feelings of appetite on a 100 mm paper scale (Olli et al. [11], Ranawana et al. [16], and Schwab et al. [18]), while the other four studies used electronic systems with differing scale lengths (Astbury et al. [12], Astbury et al. [15], Hull et al. [17], and King et al. [19]). Olli et al. [11] was the only study which reported results as iAUC for the Satiation and Satiety periods separately, while Ranawana et al. [16] reported results as iAUC for the full assessment of appetite feelings. Both studies compared groups using paired Student's t-test. Schwab et al. [18] directly compared specific time points also using Student's paired t-test. King et al. [19] compared changes from the pre-to post-load using two-way repeated measures ANOVA. All other studies analyzed changes from the subjective appetite responses baseline to the preload (between test product to the next meal) using two-way repeated measures ANOVA (Astbury et al. [12], Astbury et al. [15] and Hull et al. [17]).
Two included studies measured subjective feelings of appetite acutely before and after a chronic administration of polydextrose: King et al. [19] in a crossover design and Schwab et al. [18] in a parallel design. In these interventions the same volunteers participated in the pre-and post-measurements. The results on appetite ratings between these two time points varied to a large extent, most likely due to the effect of the chronic administration of polydextrose. In addition, the number of participants in these two studies is small and the variance is large. Hence, the total impact of the two repeated studies on the total effect was very small. Therefore, in the current report, the same criteria was followed as per the meta-analysis on different levels of energy intakes [9], where we considered each acute measurement as an independent study.
Hunger was the only indicator of subjective feelings of appetite that was assessed in all the studies that were included. The meta-analysis of the Satiation period clearly did not favor polydextrose for reducing this parameter. The meta-analysis of the Satiety period also did not show a statistical difference. However, some of the included studies did report significant changes for Hunger during the Satiety period when different analysis methodologies were used (Hull et al. [17] and Olli et al. [11]). Future studies may help to confirm the tendency of polydextrose to reduce Hunger during the Satiety period.
Only three studies measured Satisfaction (Hull et al. [17], Olli et al. [11] and Schwab et al. [18]). The meta-analysis did not favor polydextrose during the Satiation or the Satiety periods for the increase of this parameter. However, two of the included studies reported that polydextrose increased Satisfaction significantly (Hull et al. [17]), or marginally (Olli et al. [11]), during the Satiety period. It is important to mention that the wording used to evaluate Satisfaction varies amongst the reported methodologies making it difficult to properly assess this subjective feeling of appetite. For example, some authors recommend using the term Satiety instead of Satisfaction (e.g., "How Satiated are you?" instead of "How Satisfied are you?") [3]. Nevertheless, future studies may help to confirm the tendency of polydextrose to increase the feeling of Satisfaction during the Satiety period.
Surprisingly, the meta-analysis did not favor polydextrose or the placebo for an increase of Fullness during the Satiation or Satiety periods. Polydextrose is a well-recognized dietary fiber [27]. In general, dietary fibers are well known to increase the feeling of Fullness due to their properties of adding bulk and producing viscosity in foods [26]. However, polydextrose in the study dose range of 6.25-25.0 g/day is highly soluble and has little impact on the viscosity of beverages [28]. This may explain, at least in part, why polydextrose did not show any effect on Fullness in the meta-analysis.
The only individual study that reported a reduction in Fullness was Hull et al. [17], but only after the ad libitum meal during the post-Satiety period.
Although two studies assessed Prospective Food Consumption (Hull et al. [17] and Ranawana et al. [16]), it was not possible to conduct a meta-analysis because Ranawana et al. [16] reported results in an non-aligned way (e.g., values for both polydextrose and control decay along the Satiety period). Hull et al. [17] reported that polydextrose reduced Prospective Food Consumption as compared to the placebo in the Satiety period. More studies measuring this parameter are needed in order to confirm this observation.
The most striking result was that the meta-analysis favors polydextrose for the reduction of the Desire to Eat during the Satiation period. There were no statistical differences during the Satiety period for this parameter. Six studies measured the Desire to Eat. However, only Hull et al. [17] reported a reduction of the Desire to Eat after the intake of polydextrose, both between the preload challenge until the subsequent meal at lunch time and after the lunch until dinner.
It has been proposed that the category of sex may play a role on the subjective feelings of appetite during the Satiety period, where generally females felt more satisfied than males [29]. However, little is known about this influence during the Satiation period. In this review, the meta-analysis shows that results by sex are comparable to the results of the entire mixed group during both the Satiation and Satiety periods, except for the Desire to Eat category during the Satiation period. Here the meta-analysis favors polydextrose only in the male population. The Desire to Eat category was assessed in six studies involving 69 males and 52 females, which represents an important part of the total universe of this review. To our knowledge, this is the first report on the effects of polydextrose on the Desire to Eat by sex. Future studies designed to assess differences on subjective feelings of appetite by sex (e.g., including equal sample sizes of males versus females) may help us to better understand this observation.
Results of this review indicate that some studies demonstrate that polydextrose affects several subjective feelings of appetite at doses which are in the commercial application range for foods and dietary supplements, i.e., between 6.25 g and 25.0 g in a single dose per day. When high doses of polydextrose have been tested, such as 56.7 g over the duration of a day [30], its effects on subjective feelings of appetite are conclusive: it reduces Hunger and the Desire to Eat, while increasing the feeling of Fullness and Satiety (Satisfaction).
In this review, the incremental area under the curve for the Satiation (iAUC Satiation ) and Satiety (iAUC Satiety ) periods of subjective feelings of appetite were analyzed to investigate the appetite-suppressing effect of polydextrose. Although other meta-analyses on VAS have been conducted using iAUC [31], to our knowledge, this is the first review that investigates this effect in both periods independently. The duration of the Satiation periods were almost the same in all studies, approximately 15 min, except in Olli et al. [11] where the period was 40 min. The lengths during the Satiety period varied more, between 60 (Ranawana et al. [16]) and 280 min (Olli et al. [11]) among studies. The differences in duration of these periods are perhaps the biggest limitation to the methodology used in this review. The lack of consistency among participants in the same study can also be considered a problem when validating this methodology. Once all individual iAUCs are evaluated in order to generate an average value, the effects of each period are clearly visualized in both iAUC Satiation and iAUC Satiety . The study by Olli et al. [11] is a good example of the methodology used for this purpose. However, inconsistent results amongst the participants may cause large deviations in the iAUC values and may produce negative values, which in turn may cause abnormal relative changes. The study by Schwab et al. [18] is a clear example of where this has occurred, and the methodology cannot easily be applied (see iAUC values in the Supplementary Materials). Nevertheless, inconsistent results are the exception with most of the studies, and overall, the methodology is very reliable. The estimation of iAUC allowed for the comparison of several studies using the same magnitude in this meta-analysis. Future studies may further estimate the reproducibility and validity of this methodology. Further, due to the fact that many of the included studies assessed energy intakes, the estimation of iAUC in the Satiation and Satiety periods could be complemented with the estimation of satiety quotients [32], with the aim of extending our understanding of the effects of a consumed food on appetite.

Conclusions
The methodology used to homogenize results of iAUC on the subjective feelings of appetite during the Satiation and Satiety periods allowed for a meaningful comparison of the results of the studies included in this meta-analysis. This review demonstrates, for the first time, that polydextrose reduces the Desire to Eat during the Satiation period. This effect was also significant in the sub-analysis by sex for the male population. The results on this subjective feeling of appetite may explain, at least in part, the observed effects of polydextrose on the reduction of energy intake at a subsequent meal. There were no differences on Hunger, Satisfaction, or Fullness during the Satiation or Satiety periods. A future meta-analysis incorporating more studies may confirm the tendency of polydextrose to reduce Hunger and increase Satisfaction during the Satiety period. This meta-analysis shows that polydextrose does not affect Fullness, possibly due to its high solubility and low viscosity.