| Study ID | |
| RefID (DistillerSR) | 4298 |
| Reference (authors, year, title, other info) | Biofortis, 2010. Study of Gastrointestinal Tolerability of Erythritol (Polyol) in Children. Report Ref. CER_TDEOH05. |
| Source (published/unpublished) | Unpublished |
| Study design | |
| Study type | HCT – multi‐centre, randomised, placebo‐controlled, cross‐over (with placebo only) study |
| Type of blinding | Double‐blinded |
| Duration of the study and length of follow‐up | 1 day per test material |
| Subjects | |
| Number of participants in the study | 185 participants (172 participants per protocol population) |
| Number of exposed/non‐exposed subjects or number of cases/controls (if applicable) | 184 exposed |
| Sex (male/female) | 95 males and 77 females (per protocol population) |
| Age (mean or range as reported) | 4–6 years old |
| Geography (country) | France |
| Ethnicity | Not reported |
| Confounders and other variables as reported | Not reported |
| Special health condition of participants | Healthy |
| Inclusion and exclusion criteria in the study |
Inclusion criteria: healthy children, 4–6 years old on day 1 of study (i.e. day of consumption of first beverage), BMI ≥ 13 kg/m2, accustomed to having breakfast, having a regular stool frequency ≤ 2 per day, able to drink 250 mL within 15 min, toilet‐trained/able to use a potty (day and night), informed consent of both parents/guardians, affiliated to the French social security Non‐inclusion criteria: participation in any clinical trial that included blood sampling and/or administration of substances up to 90 days before day 1 of the study, participation in any non‐invasive clinical trial up to 30 days before day 1 of this study (including blood sampling and/or, intravenous, inhalatory administration of substances), having a history of medical or surgical events that may significantly affect study outcome (e.g. gastric and digestive diseases), any current metabolic or endocrine diseases (including diabetes mellitus), use of medication (including antibiotics, laxatives and steroids), having regular GI complaints (e.g. stomach upsets, diarrhoea, constipation, flatulence, abdominal colic) Temporary exclusion (non‐inclusion) criteria: no consumption of breakfast before consumption of the test product, failure to consume at least one solid food during breakfast, vaccination within 7 days before day 1 of study or during study, regular stomach aches within 7 days before day 1 of the study, period of more than two faeces per day within 7 days before day 1 of study, light solid products consumption (e.g. confectionery, chocolate, chewing gum, biscuits) within 24 h of test product consumption |
| Other information | |
| Intervention/exposure | |
| Test material | Erythritol (Cargill) |
| Description of the intervention and estimated dietary exposure |
Children were required to consume breakfast approximately 2 h before attending the test centre, where they were provided the test drink for consumption within 15 min. Test drinks contained 5 g (n = 14), 15 g (n = 57), 20 g (n = 58) or 25 g (n = 55) erythritol in 250 mL of non‐carbonated fruit‐flavoured drink (Cargill). Placebo groups consumed 250 mL of non‐carbonated fruit‐flavoured drink, sweetened with saccharose and maltodextrin (Cargill) to a sweetening power equivalent to the corresponding erythritol beverages Children were not allowed to eat anything (and only drink water) for at least 1 h after consuming the test drink. For approx. 2 h after consumption, children remained at the test centre under observation of the investigator (collecting the eventual urine). At home, until 48 h after consumption, the legal guardians were asked to keep a defecation‐frequency diary for their child, collect urine for 24 h following consumption of the test substance, and complete a questionnaire on GI symptoms at 6 h and 24 h after consumption of the test substance The second test day was a repeat of the first test day except subject history was not taken. Children were asked to eat the same breakfast as before There was a washout period of ≥ 5 days between test days |
| Co‐exposure description (if applicable) | Not applicable |
| Endpoint measured, measurement time points and methods |
Primary outcome variable: GI tolerability as measured by the incidence of diarrhoea and/or significant GI symptoms following consumption of test item Secondary outcome variables: stool frequency and consistency, GI symptoms, total symptom intensity score, urinary erythritol excretion For each group, erythritol treatment effects were compared to the effects of treatment with placebo for: (i) diarrhoea (= single watery stool (i.e. Bristol Stool Scale (BSS) = 7) and/or > three faeces in 24 h), (ii) stools frequency and BSS ratings (for BSS, if no stool was passed for a subject, subject was not included in the analysis), (iii) GI symptoms (occurrence, frequency and intensity during 0–6 h and 6–24 h after consumption). A composite score of symptoms' intensity was computed for abdominal pain, nausea, borborygmi, bloating and gas. For each individual symptom's intensity and frequency scores and for the composite score, the maximum was taken into account, for analysis on overall period (0–24 h) and (iv) erythritol concentration in urine To measure group effect of erythritol: Y = product + sub‐group + product × sub‐group. A sensitivity analysis was performed on primary criteria where all missing data on GI symptoms or BSS were replaced by ‘strong intensity’ and ‘BSS = 7’ regardless of the product consumed. Additionally, all GI adverse events (AEs) reported in the case‐report form and for which no data on GI symptoms were available in the questionnaire were considered as clinically relevant diarrhoea/GI symptoms |
| Were sub‐groups analyses predefined? (yes/no, including justification) | Not applicable |
| Results | |
| Findings reported by the study author/s | Following consumption of 20 g erythritol, there was a significant difference between the groups for the incidence of diarrhoea and/or significant GI symptoms with the erythritol group having significantly more events than the placebo group (p = 0.0286), that was also significant for the 25 g dose (p < 0.0001) but not for the 5 g and 15 g doses. Therefore, 15 g erythritol in a single drinking occasion was the maximum tolerated dose in children |
| Statistical analysis | |
| Statistical methods (including power analyses, multiple comparison, potential sources of bias, adjustment for confounders, test for interactions) |
In the first instance, the number of children required was calculated based on the following: first‐order risk α = 10%, statistical power = 80%, PE‐PP ≥ 30%, (PE = % of subjects having the AE in question under erythritol, and PP = % of subjects having the AE under placebo). On this basis, the number of children required to be randomised in a given sub‐study was 13. In order to get a higher power, the number of 14 children per group was chosen Then, sample size was recalculated based on a first‐order risk α = 10%, the two‐sided confidence interval of the difference (PE‐PP) = 90% (where PE is the percentage of subjects having an AE under erythritol, and PP the percentage of subjects having an AE under placebo) and the higher boundary of the confidence interval of (PE‐PP) ≤ 39%. On the basis of the above hypotheses, the number of children required to be randomised in a given sub‐study was 56 Analyses were performed using SAS 9.1.3 Service Pack 4. Statistical analyses accounting for the cross‐over design of the study were of risk type 1 error (α), set at 0.10 for the assessment of the threshold level dose and at 0.05 for all other statistical tests Primary analysis: qualitative variables were analysed using one‐tailed Mainland Gart's test, which takes into account that the product groups were not independent and that products were not given in the same order. This analysis was performed after each sub‐study in order to go to next step (risk type 1 error set at 0.1). To ensure that erythritol was not inferior to placebo when no statistically significant difference was found, a confidence interval at 90% was computed applying the May and Johnson method Secondary analysis: one‐tailed Mainland Gart's test was also used for binary variable. For nominal or ordinal variable bilateral Kappa test was used. Qualitative variables were described as frequencies and percentages of the number of individuals examined. Quantitative variables were analysed using repeated measures ANOVA with product group, sequence and product group by sequence interaction as effects. Models were reduced in a stepwise manner until only significant (p ≤ 0.05) terms or product group remained. Quantitative variables, including those related to demographic data, were summarised in tables displaying sample sizes, means, SDs, medians, minimum and maximum values |
| Further information | |
| Data from this unpublished study report were subsequently published in Jacqz‐Aigrain et al. (2015) | |
| Study ID | |
| RefID (DistillerSR) | 759 |
| Reference (authors, year, title, other info) | Storey, D., Lee, A., Bornet F., & Brouns, F. (2007). Gastrointestinal tolerance of Erythritol and Xylitol ingested in a liquid. European Journal of Clinical Nutrition, 61, 349–354. https://doi.org/10.1038/sj.ejcn.1602532 |
| Source (published/unpublished) | Published |
| Study design | |
| Study type | HCT – randomised, placebo‐controlled, cross‐over study |
| Type of blinding | Double‐blinded |
| Duration of the study and length of follow‐up | 1 day per test material |
| Subjects | |
| Number of participants in the study | 70 participants. Subjects were randomly recruited from the student population of the University of Salford |
| Number of exposed/non‐exposed subjects or number of cases/controls (if applicable) | 65 exposed |
| Sex (male/female) | 34 males and 36 females |
| Age (mean or range as reported) | 18–24 years old |
| Geography (country) | UK |
| Ethnicity | Not reported |
| Confounders and other variables as reported | Not reported |
| Special health condition of participants | Healthy |
| Inclusion and exclusion criteria in the study | Potential recruits were included according to the procedures as described by Lee and Storey a |
| Other information |
Five male subjects failed to complete the study: 3 dropped out due to illnesses unrelated to the study and 2 due to adverse GI effects following consumption of xylitol‐containing test products For one subject, reporting was incomplete for all GI symptoms except for data on stools, therefore data were based on 64 subjects |
| Intervention/exposure | |
| Test material | Erythritol |
| Description of the intervention and estimated dietary exposure |
20 g, 35 g and 50 g erythritol in test drinks consumed in subjects' homes on test days whilst maintaining their normal diets Test materials were provided as 400 mL orange‐flavoured non‐carbonated drinks (2 × 200 mL glass bottles; Cerestar R&D, Vilvoorde, Belgium). For all test drinks, sweetness intensity was corrected to 11.25 sucrose equivalent value using aspartame. Consumption of test drinks was separated by 1‐week washout periods. Dietary restrictions included no prior consumption of polyol‐containing products. Subjects consumed the drinks either mid‐morning following consumption of a normal breakfast or mid‐afternoon following consumption of a normal lunch. Drinks were consumed within 15 min and subjects requested not to consume food or drinks (except up to 300 mL water) for the following 2 h. Subjects were debriefed within 24 h to determine adherence to dietary restrictions and consumption and to assess GI responses |
| Co‐exposure description (if applicable) | Not applicable |
| Endpoint measured, measurement time points and methods | Subjects were given printed sheets on which to report the prevalence and magnitude of flatulence, borborygmi, bloating, colic, nausea, bowel movements and the passage of faeces of an abnormally watery consistency for the 24 h period following consumption. A hedonic scale was used: 0 = normal function, 1 = slightly more symptom than usual, 2 = noticeably more symptom than usual and 3 = considerably more symptom than usual. Symptom scores were derived by summing each subject's GI responses. Subjects recorded the number of bowel movements to pass faeces of normal, hard or watery consistency, where watery faeces were defined as those of an abnormally watery consistency (e.g. loss of firm shape owing to high‐water content) |
| Were sub‐groups analyses predefined? (yes/no, including justification) | Not applicable |
| Results | |
| Findings reported by the study author/s |
Consumption of 20 and 35 g erythritol in a liquid by healthy volunteers was well tolerated and without symptoms. At 50 g erythritol, there were significant increases only in the number of subjects reporting borborygmi (p < 0.05) and mild nausea (p < 0.01) Xylitol produced significantly more watery faeces than erythritol when comparing all intake levels: 50 g xylitol versus 35 g erythritol (p < 0.001), 50 g xylitol versus 20 g erythritol (p < 0.001) and 35 g xylitol versus 20 g erythritol (p < 0.05) |
| Statistical analysis | |
| Statistical methods (including power analyses, multiple comparison, potential sources of bias, adjustment for confounders, test for interactions) | Symptoms were classified as categorical, and considered to be non‐parametric. GI responses were compared by 2 × 2 contingency table analysis (chi‐squared) according to the methods of McNemar. b The binomial test was used when the expected frequency in each cell of the contingency table was less than 5. The chi‐squared test was used to test for differences in the occurrence of multiple symptoms following consumption. The frequency of bowel movements to pass normal, watery and hard faeces were analysed by one‐way ANOVA, followed by Dunnet's post hoc test to detect differences in case of an overall significant treatment effect |
| Further information | |
| Test drinks containing sucrose (45 g) and xylitol (20, 35 and 50 g) were also tested | |
Lee, A., & Storey, D. M. (1999). Comparative gastrointestinal tolerance of sucrose, lactitol, or d‐tagatose in chocolate. Regulation of Toxicology Pharmacology, 29, S78–S82.
McNemar, Q. (1947). Note on the sampling error of the difference between correlated proportions or percentages. Psychometrika, 12, 153–157.