Effects of gastrointestinal delivery of non-caloric tastants on energy intake: a systematic review and meta-analysis

This systematic review and meta-analysis was registered in the PROSPERO database on 26 February 2020 (ID: CRD42020171182). The present systematic review and meta-analysis were performed according to the Preferred Reporting Items for Systematic Reviews and Meta-Analyses (PRISMA) guidelines [46]. The description of the PICOS (participants, intervention, comparison, outcome, and setting) criteria used to define the research question are depicted in Table 1. A structured search in the Cochrane, PubMed, Embase, and Medline databases was performed up to 26 February 2020 with the following search strategy: ((((((((((((tastant) OR Taste) OR Taste receptor) OR bitter taste) OR quinine) OR denatonium benzoate) OR umami) OR sodium glutamate) OR monosodium glutamate) OR sweet) OR non-nutritive sweeteners)) AND ((((((((((((((((energy intake) OR intake) OR food intake) OR appetite sensation) OR satiation) OR satiety response) OR satiety) OR satiety hormones) OR glucagon-like peptide-1) OR peptide YY) OR Ghrelin) OR leptin) OR cholecystokinin) OR motilin) OR motility) OR gastric emptying).

Eligibility of each paper was assessed independently by two researchers (TK and DK) according to predefined criteria. Papers reporting the effects of post-oral delivery of non-caloric tastants on eating behavior (e.g., energy intake, GI peptides, appetite sensations, GI motility, GI symptoms, brain signaling and other effects) were included. To investigate the function of gastrointestinal taste receptors, papers needed to properly bypass oral taste effects. Therefore, studies investigating delivery methods using catheters, capsules or other methods resulting in adequate masking of oral taste were included. Moreover, papers were excluded if they were reviews, comments, replies on an original paper, or abstracts without available full text. No limitations on publication date were set. Paper inclusion was agreed upon by both reviewers. A third reviewer (AM) was consulted with regard to inclusion in case of disagreement between the two reviewers.

This systematic review looked at various aspects of eating behavior after post-oral delivery of non-caloric tastants. Outcome measures of interest were (1) energy intake, (2) GI symptoms and perceptions, and (3) mechanisms of effect.

Two reviewers (TK) and (DK) carried out the data extraction. Two authors were contacted to elaborate their data and they replied. Name of author, year of publication, country, sample size, age of subjects, BMI of subjects, tastants used, method of administration, energy intake, appetite sensations, GI symptoms, GI peptides, GI motility, and brain signaling in homeostatic and hedonic regions were abstracted and presented in tables. Principle summary measures are differences in means.

Two independent reviewers (TK and DK) used the revised Cochrane risk of bias tool to assess risk of bias in randomized trials (RoB 2) to assess the quality of included papers [47]. The quality of the paper was assessed only once when a paper described multiple studies. A third reviewer (AM) was consulted in case of discordance between the two reviewers. The RoB 2 tool, assessing the quality of randomized controlled trials, consists of five domains covering bias arising from the randomization process, bias due to deviations from intended interventions, bias due to missing outcome data, bias in the measurement of the outcome, and bias in selection of the reported result. The scoring system assesses the risk of bias on these domains (low risk of bias, some concerns, high risk of bias). When an individual domain received score of a particular level of bias, overall risk of bias was determined to be at least as severe.

A meta-analysis was performed if at least two studies described the use of non-caloric tastants of the same basic taste on energy intake and clinical heterogeneity between studies was acceptable. Clinical heterogeneity was discussed by two independent reviewers (TK and DK). In case of discordance between these reviewers, a third reviewer (AM) was consulted. Meta-analyses were performed using a random effect model by the metaphor package in R (version 3.6.3) [48]. Energy intake in Kcal after tastants and control were pooled using the data provided by included studies. Sensitivity analyses were performed when same studies employed different doses of tastants in order for those subjects to not influence the results to a greater extent than subjects form other studies. The I² was used to quantitatively measure statistical heterogeneity between studies (p value < 0.05).

Paper inclusion, exclusion, and reasons for exclusion are presented in a diagram (Fig. 1) according to the PRISMA statement for reporting a systematic review and meta-analysis.

After removal of duplicates, a total of 11,271 abstracts were assessed, and 45 full texts were screened for eligibility (Fig. 1). A total of 19 papers, describing 25 studies met the inclusion criteria. Twenty-six papers were excluded for various reasons: abstract for a conference (13), review (5), study protocol (4), animal study (2), no adequate bypassing of oral taste receptors (1), no full text available (1). The flow-chart for screening and inclusion of papers is depicted in Fig. 1.

Supplementary Table 1 provides an overview of the results of the included papers.

The results of the quality assessment are summarized in Fig. 2. Overall, papers scored decently on the risk of bias assessment. Bias arising from the randomization process was determined to raise some concerns by the majority of papers due to lacking information on the method of randomization. Most papers described randomization of subjects but provided no information on the method of randomization.

An overview of the studies describing effects of post-oral delivery of non-caloric tastants on GI symptoms and perceptions is provided in Table 4.

An overview of the studies describing effects of post-oral delivery of non-caloric tastants on the mechanisms of effect involved in regulating eating behavior is provided in Table 5. The mechanisms of interest were GI peptide release, GI motility, and brain signaling.

Most studies described the effects of post-oral delivery of sweet [35, 36, 38, 40‐44] or bitter [27‐33, 35, 37, 39, 42, 45] stimuli. Only one study used an umami stimulus [42] and two studies described the effects of a tastant mixture [34, 42]. It has been hypothesized that taste can predict the type of food that is being ingested (i.e., bitter for potential toxic substances, umami for glutamate and sweet for saccharides) [8], although this theory is probably an oversimplification and does not sufficiently reflect the complexity of the underlying biological responses. According to traditional beliefs, carbohydrates are considered as fuel for the human body [54], whereas one wants to avoid toxic substances [55]. Therefore, it is conceivable that most researchers hitherto have generally focused on sweet and bitter substances. However, umami taste is a predictor of amino acids. Up to now, little is known about solely post-oral effects of umami taste, although it has been widely accepted as a basic taste since the discovery of umami taste receptors in 2002 [56, 57]. However, more data are available on oral delivery of umami stimuli. Oral delivery of umami stimuli has been shown to elicit a GLP-1 response [58, 59]. Moreover, adding MSG to a novel flavor is able to condition liking for that flavor [60]. Furthermore, Dermiki et al. showed that adding MSG to novel flavored soups resulted in increased food intake in elderly subjects without eliciting increased liking [61]. Given the effects of oral MSG on food liking, food intake, and GI peptides, the effect of post-oral delivery of umami tastants should be further elucidated.

This review and meta-analysis clearly shows that the most potent stimuli to influence eating behavior and reduce food intake are the bitter substances. The obvious explanation for this is the innate aversion for bitter taste [62]. That is, delivery of solely bitter tastants, in the absence of other flavors, would result in a warning signal with the intention to stop the intake of that particular substance. However, it should be noted that several studies show that this negative affective response to bitter can be decoupled by, for instance, the positive response to caffeine [9, 10]. This process is called flavor-consequence learning [63]. Moreover, Higgins et al. showed an increase in pale ale intake in individuals with increased bitter perception [64]. This indicates that, over the years, humans have learned to appreciate bitter tastes, mainly when combined with other flavors (i.e., liking for black coffee and beers).

The question, however, arises whether any such mechanism would also hold true for post-oral taste receptor stimulation.

Data on post-oral delivery of sweet tastants are limited to three studies showing contradictory results [40, 42, 44]. Interestingly, more data are available on the effect of oral consumption of non-caloric sweeteners on energy intake. A review and meta-analysis by Rogers et al. that also included studies using oral stimulation showed a reduction of energy intake after consumption of low caloric sweeteners when compared with sugar consumption but not when compared to consumption of water in a short-term setting [65]. Since there is a preference for sweet foods and beverages in humans, further research to investigate the effects of post-oral delivery of sweet tastants on energy intake is warranted.

Eight studies described effects on energy intake after post-oral delivery of bitter tastants. These studies show a modest decrease in energy intake following post-oral delivery of bitter tastants [27, 29, 30, 32, 33, 37, 39, 42]. These decreases in energy intake varied from 11 to 88 kcal in the acute setting to 340 kcal on a daily basis. Pooled effects of post-oral delivery of bitter tastants show a significant decrease of 54 kcal compared with placebo in the acute setting. In itself, a reduction of 54 kcal in a single meal is rather small. However, in case this reduction in energy intake can be replicated several times daily over a longer period of time, this may indeed lead to a daily caloric deficit and subsequent weight loss. It should be noted, however, that these modest effects on caloric intake point towards a role in weight control rather than weight loss.

Based on the included papers, no clear effects of post-oral delivery of non-caloric tastants on appetite sensations was found. However, it must be noted that appetite sensations are rarely measured as a primary outcome. Consequently, most studies might not have been adequately powered to detect differences in appetite sensations. Therefore, interpretation of these results should be done with care.

Interesting to point out is the finding of Deloose et al. showing a longer sustained satiation in response to bitter after a standardized test meal [32]. This observation indicates that adding a bitter tastant to a caloric carrier could result in a delay until the next meal. Such a combination could result in a decrease in snacking in between meals. In line with this, Mennella et al. found in their study a reduction in caloric consumption during the day, after intake of a breakfast containing an encapsulated bitter mixture leading to prolonged satiation [37].

Only one study investigated the effect of post-oral delivery of an umami stimulus. A decrease in desire to eat and hunger was found [42].

According to the studies described in this review, noncaloric tastant administration is considered to be safe. One study reported side effects after administration of high doses of xylitol and erythritol, which was not surprising given the doses employed [43]. Other studies showed no GI-symptoms or other side effects after post-oral delivery of non-caloric tastants [27‐34, 38, 42, 45]. It should be taken into account that, almost all studies focused on acute effects of post-oral delivery of non-caloric tastants and not on prolonged, daily administration.

The United States Food and Drug Administration (U.S. FDA) issues an acceptable daily intake (ADI) for various food components, including tastants. For example, the U.S. FDA issued a code of regulations stating soft drinks are allowed to contain 83 parts per million quinine [66]. Moreover, a systematic review describing the use of quinine to treat muscle cramps showed an increase of gastrointestinal complaints, headache and tinnitus after daily intake of 200–500 mg of quinine for 3 days up to several weeks [67]. Even more, the Medicines and Healthcare products Regulatory Agency (MHRA) issued a reminder on the dose-dependent effects on the QT interval [68]. Therefore, prolonged intake of high doses or combination with other drugs that prolong the QT interval should be avoided. Interestingly, in studies included in this review, the dose of the bitter tastant quinine ranged from 18 to 600 mg [27, 29‐31, 33‐35, 42]. This did not result in side effects in the studies investigating the acute effects of these compounds. However, when applying quinine daily to prevent obesity, the maximum dose should be carefully considered. This should be considered in future study protocols.

Gastric emptying was delayed by high doses of xylitol and erythritol [43]. However, other studies showed no effect of post-oral delivery of non-caloric tastants on gastric emptying [27, 29, 32, 35]. Therefore, gastric emptying does not seem to be attenuated by post-oral delivery of non-caloric tastants.

GI motility appears to be attenuated as one research group has consistently found gastric motility and the GI peptides motilin and ghrelin to be affected by post-oral delivery of non-caloric tastants [28, 31‐33]. Moreover, this research group recently published a review that elaborated the role of motilin as a regulator of hunger and food intake in humans [69], which indicates motilin as a possible target in combating the obesity epidemic. These interesting data are awaiting replication by other research groups.

Four studies showed increased CCK and/or GLP-1 levels after post-oral delivery of non-caloric tastants [27, 29, 37, 43]. Other studies that investigated these traditional satiety peptides did not find any effect of post-oral delivery of non-caloric tastants [30, 36, 38, 42]. This raises the question whether the focus should shift from more traditional satiety peptides towards motilin and/or ghrelin. It must be noted that all studies reported on plasma levels of GI peptides, pointing to systemic effects. Up to now, it is still unclear what the effect of taste receptor activation is on local secretion of GI peptides. It is conceivable that GI peptides are elevated at a splanchnic level or that they exert a more local or paracrine effect.

Only one study investigated brain signaling after post-oral delivery of non-caloric tastants. These authors found an increase in activity in homeostatic and hedonic brain regions and a decrease of activity in the brain stem and medulla after intragastric delivery of quinine [33]. These are interesting findings, since these data suggest additional mechanisms of effect of post-oral delivery of non-caloric tastants on eating behavior. It remains to be further elucidated whether these changes are mediated by orexigenic GI peptides.

In this review and meta-analysis, we have summarized the current knowledge on the effects of non-caloric tastants on energy intake. Using non-caloric tastants to reduce energy intake could provide a useful tool to combat the obesity epidemic.

Specifically, the use of bitter tastants appears to be promising. It is important to note that most studies described in this review have been performed in healthy adults with a normal BMI or in slightly overweight healthy adults. In several studies blunted postprandial levels of PYY [70, 71] and GLP-1 [72‐74] were observed in obese subjects, and reductions in plasma CCK after weight loss [75, 76]. This points to alterations in the sensitivity of various receptors to GI peptides and in the magnitude of peptide secretion in obesity. Non-caloric tastants, because of their modest reduction in caloric intake, appear to be more suitable as a weight control intervention in a population that is worried about but has not yet gained excess weight.

This review and meta-analysis clearly illustrates the potential of bitter tastants in reducing caloric intake. However, several weaknesses of the currently available data should be elucidated. First, data are lacking uniformity. Most studies were underpowered: only small numbers of subjects have been included. Second, different study designs have been employed, making it difficult to compare results between various publications. An important variable with respect to energy intake reduction is the varying time interval between post-oral delivery of non-caloric tastants and meal intake. These intervals varied among studies between 30 min and 5 h but usually an interval around 60 min was chosen. The optimal interval between intervention and ad libitum meal is currently unknown, but standardization is necessary.

Third, most of the studies included in this review described the acute effects of post-oral delivery of non-caloric tastants on eating behavior. Only two studies described energy intake during the day [37, 39] and only one of those described an intervention period of 2 weeks [39]. Therefore, data on the long-term effects of post-oral delivery of non-caloric tastants are lacking. Consequently, it is unknown whether adaptation to the effects occurs. A fourth limitation is the lack of knowledge on the effects of post-oral delivery of umami tastants on eating behavior, as this was described in only one paper [42]. Fifth, only one study investigated several target locations in the GI tract [34]. Sixth, based on the present data it is unknown which bitter stimulus elicits the greatest effect on energy intake. Most studies described the use of either QHCl or DB.

To evaluate the effects of gastrointestinal delivery of non-caloric tastants on eating behavior, future studies should preferably standardize study design and doses of tastants. First, standardization could be achieved by creating consensus from the lead experts in the field. Second, mechanisms of effect should be more thoroughly investigated. For this, more research on the effects of post-oral delivery of non-caloric tastants on GI motility and systemic and local GI peptides secretion is needed. Based on the current data, we propose to focus on motilin and gastric motility. Third, compared with other tastants, knowledge on the effects of umami tastants is lacking. Therefore, the effects of post-oral delivery of umami tastants should be further elucidated. Fourth, taste receptors are expressed throughout the entire GI-tract [19‐22]. The most appropriate location(s) for tastant delivery to modulate eating behavior is unknown. More research investigating delivery of tastants in different locations in the GI-tract is needed. Fifth, the optimal dose of tastants to exert an effect on eating behavior is unknown. Studies should focus on finding the balance between the maximum possible effect without occurrence of side effects. Sixth, it should be investigated which bitter stimulus elicits the largest effect on energy intake. For this, studies should compare different bitter stimuli. We propose to focus on QHCl and DB, as most data are available on these bitter stimuli and both stimuli activate a wide range of bitter receptors. Lastly, when all the foundations are laid out, the field should move towards implementation of the successful interventions in battling the obesity epidemic. For this, the longer-term effects of post-oral delivery of non-caloric tastants on energy intake and ultimately body weight control should be investigated.