Petersson 1991.

Methods	Trial design: parallel (4 arms) Location: public dental clinic in Hyltebruk, Sweden Number of centres: 1 Recruitment period: not stated
Participants	Inclusion criteria: healthy 12‐ to 13‐year‐olds Exclusion criteria: not stated Baseline caries: not clearly stated (presented graphically) but there was a statistically significant difference between Gp A and Gp B Age at baseline: all children were 12 to 13 years old Gender: not stated Any other details of important prognostic factors: the study site was chosen due to a relatively high caries frequency and stable population; the water supply had a fluoride concentration of about 0.1 ppm Number randomised: 322 (Gp A: 78; Gp B: 83; Gp C: 78; Gp D: 83) Number evaluated: 284 (Gp A: 67; Gp B: 74; Gp C: 68; Gp D: 75)
Interventions	Comparison: the 4 arms were as follows: 1) xylitol toothpaste with sorbitol and normal level of fluoride 2) xylitol toothpaste with sorbitol and low level of fluoride 3) toothpaste with sorbitol and normal level of fluoride 4) toothpaste with sorbitol and low level of fluoride Gp A (n = 78): twice daily brushing with toothpaste containing 0.8% sodium monofluorophosphate, 3% xylitol, 6% sorbitol Gp B (n = 83): twice daily brushing with toothpaste containing 0.03% sodium fluoride, 3% xylitol, 6% sorbitol Gp C (n = 78): twice daily brushing with toothpaste containing 0.8% sodium monofluorophosphate, 9% sorbitol Gp D (n = 83): twice daily brushing with toothpaste containing 0.03% sodium fluoride, 9% sorbitol Duration of treatment: 3 years Sorbitol is considered to be inert (i.e. neither causes nor prevents caries) so we did not consider its inclusion in the syrups to be a problem We would have compared Gp A versus Gp C and Gp B versus Gp D in the analyses, if there had been any usable data
Outcomes	Caries: number of DFS. Assessed at 1, 2, and 3 years Microbial counts (not an outcome of interest in this review)
Notes	Sample size calculation: 65 to 75 participants per group to allow 95% power to detect a true difference between group means of approximately 4 DFS (theoretical calculations made via a pilot study) Adverse effects: not reported Funding source: Kema Nobel Consumer Goods Division (Stockholm, Sweden) provided toothpastes (also states they provided "economical support" but this may just be the toothpastes) Declarations/conflicts of interest: not stated
Risk of bias
Bias	Authors' judgement	Support for judgement
Random sequence generation (selection bias)	Unclear risk	Quote: "All children taking part were randomly distributed into four experimental groups" Comment: insufficient information on the method of sequence generation
Allocation concealment (selection bias)	Unclear risk	Quote: "All children taking part were randomly distributed into four experimental groups" Comment: allocation concealment not mentioned
Blinding of participants and personnel (performance bias) All outcomes	Low risk	Quote: "They were produced and delivered in 100‐gram tubes, marked with the name 'Toothpaste' in different colours...corresponding to groups 1‐4 respectively" and "The study was carried out double‐blind for subjects as well as for examiners" Comment: participants and personnel would not know which group a participant was assigned to
Blinding of outcome assessment (detection bias) All outcomes	Low risk	Quote: "The study was carried out double‐blind for subjects as well as for examiners" Comment: outcome assessment appears to have been adequately blinded
Incomplete outcome data (attrition bias) All outcomes	Low risk	12% of randomised participants were not included in the final analysis (Gp A: 14%; Gp B: 11%; Gp C: 13%; Gp D: 10%). This amount of attrition may be considered as low for a 3‐year study and the reasons were mainly due to them moving away from the area Comment: we do not believe that any of the above could pose a risk of bias significant enough to have led to a distortion of the true intervention effect
Selective reporting (reporting bias)	High risk	The caries data have not been reported adequately (on a small graph) or in a way that would allow them to be included in a meta‐analysis (i.e. there is no measure of variance (SD, SE or 95% CIs) reported for the mean DFS scores). Adverse events should be considered an important outcome in xylitol trials, but were not considered in this study
Other bias	Low risk	Quote: "The children were clinically examined...by one of the authors" and "A pilot study...was performed to train the examiner and to standardize the clinical and radiographic registrations" Comment: we consider that the risk of differential diagnostic activity was low