Results of the search

This is an update of our previous review (Azarpazhooh 2011). For this 2016 update, we reviewed 33 potential citations. The study flow diagram is presented in Figure 1. We identified one new trial on ClinicalTrials.gov (www.clinicaltrials.gov/) completed and published (NCT01044030 as Vernacchio 2014) and a new trial that was not yet open for participant recruitment (NCT02592382). We saved this record for a future update (Characteristics of ongoing studies). Previously, our preliminary searches identified 84 potential citations. We read the titles and abstracts of these studies. We selected four studies for full text appraisal and these met the eligibility criteria for the first version of this review (Hautalahti 2007; Tapiainen 2002; Uhari 1996; Uhari 1998). This review included five clinical trials that involved 3405 children. Details of the participants enrolled in these studies are described in Characteristics of included studies.

Included studies

Vernacchio 2014 performed a pragmatic practice‐based RCT. Ninety‐seven practices of three paediatric practice–based networks (the Slone Center Office‐based Research Network at Boston University, the Pediatric Physicians’ Organization at Children’s (Boston), and the North Carolina Child Health Research Network at the University of North Carolina) referred 778 children, aged six to 71 months, to the study for eligibility review. These children were all otitis‐prone, that is, they had a history of at least three clinically diagnosed episodes of AOM in the previous 12 months with at least one in the previous six months and may have had middle ear effusions at the time of enrolment. Of these children, 326 were randomised to receive either a placebo syrup (n = 166, daily dose of 2.25 g sorbitol three times a day) or a viscous xylitol syrup (n = 160, daily dose of 5 g three times a day) for three months. All of the daily doses were administered by parents or other routine caregivers, not in a controlled daycare setting as in the previous trials. Compliance was monitored by asking parents, at each interview, how much of the recommended syrup the child had taken since the last interview. In total, about 28% of participants in the xylitol group and 24% of participants in the placebo group discontinued treatment before the end of the study period or the occurrence of the primary outcome. The most common reasons were gastrointestinal side effects (18 in the xylitol group and 11 in the placebo group), the participant refusing to take the solution (eight in each group), and the parent becoming too busy or finding the treatment too difficult to administer (eight xylitol, five placebo). At the end of the study, a blinded investigator reviewed medical records from each participant's primary care physician as well as any other healthcare provider whom the parent identified as having treated the participant during the study period. The clinical diagnoses of AOM was registered as reported in the medical records by a wide range of clinicians and were not otherwise verified. If medical records were not available, the diagnosis of AOM was registered as reported by the parents. By the end of the trial, two participants in the xylitol group and three in the placebo group declined further participation. Four participants in the xylitol group and three in the placebo group were lost to follow‐up without any outcome information. The details of the participants enrolled in this study are described in Characteristics of included studies.

Hautalahti 2007 randomised 663 healthy children enrolled in daycare in the city of Oulu, Finland to four groups to receive either a control product (n = 331, daily dose of 0.5 g in the control chewing gum group or in a syrup three times a day, pulled together as one control group) or xylitol (n = 332, daily dose of 9.6 g in chewing gum or in a syrup three times a day, pulled together as one test group) for three months. The total daily amount of received xylitol in this study was equal to that used in earlier trials but the size of a piece of chewing gum and the amount of syrup needed per dose were bigger and dosing was less frequent (three versus five daily doses). In working days, two daily doses were given at the daycare centre and one dose at home. At weekends and holidays all three doses were given at home. Compliance and the consumption of additional xylitol products were monitored by asking parents to record the doses actually given and any other xylitol products on a daily report sheet and also by counting and measuring the unused returned products at the end of the trial. The occurrence of the first AOM diagnosed during any period of respiratory symptoms during the follow‐up was the main outcome measure based on a finding of middle ear effusion in tympanometry (B, C or positive pressure curve) and confirmed with pneumatic otoscopy. Otorrhoea from a tympanostomy tube was counted as AOM. There was dropout of 96 participants (58 in the xylitol group and 38 in the control group). The baseline characteristics (demographic features, known AOM risk factors, or AOM history, earlier use of xylitol products) were similar in all four groups. By the end of trial, the dropouts range was statistically higher in the xylitol group (17%) versus the control group (11%). The reasons for dropouts and the related numbers per each group are as follows: child refused to take the product (control: 15 versus xylitol: 22); abdominal discomfort (control: seven versus xylitol: nine); chewing gum piece too big (control: one versus xylitol: five); duration of the trial too long (control: one versus xylitol: two); forgot to give the preparation when on holiday (control: three versus xylitol: three); illness (control: two versus xylitol: three); rash (control: three versus xylitol: three); parents tired of the trial (control: two versus xylitol: one); difficult to avoid additional xylitol products (control: zero versus xylitol: one); other (control: one versus xylitol: four) or unknown reasons (control: three versus xylitol: five). The details of the patients enrolled in this study are described in the Characteristics of included studies table.

Tapiainen 2002 randomised 1277 healthy children enrolled in daycare in the city of Oulu, Finland after screening with tympanometry to receive either the control mixture (n = 212), xylitol mixture (n = 212), control chewing gum (n = 280), xylitol chewing gum (n = 286) or xylitol lozenges (n = 287) during an acute respiratory infection. The parents began administering the products to their children at the onset of acute respiratory infection (ARI) symptoms, which was defined as the appearance of one or more of the following symptoms: clear or purulent discharge from the nose, congestive nose, cough, conjunctivitis, sore throat or earache. Compliance was monitored by asking the parents to list the doses actually given on the daily symptom sheets and by counting the unused pieces of chewing gum and lozenges returned at the last appointment and measuring the volume of the unused mixture. The follow‐up lasted until resolution of the symptoms or up to three weeks. AOM was diagnosed based on a finding of middle ear effusion in tympanometry (B, C or positive pressure curve) and confirmed with pneumatic otoscopy. A total of 1253 of the 1277 randomised children were eligible for the analysis of the primary outcome. The children who dropped out (n = 24) were excluded from the statistical analysis but those who prematurely stopped using the products but still visited the clinic (n = 35) were included. The reasons for dropouts and the related numbers per each group are as follows: parents got tired (only two in the xylitol syrup and four in the xylitol chewing gum); child disliked the product (two only in the xylitol lozenge); other antibiotics for ARI (only one in the xylitol syrup and one in the control chewing gum); left the area (only one in the control syrup, four in the xylitol chewing gum and two in the xylitol lozenge); and unknown (syrup: zero in the control versus two in the xylitol group; chewing gum: two in control versus one in xylitol; lozenge: two in xylitol). Moreover, the administration was discontinued due to abdominal discomfort (syrup: one in control versus three in xylitol; chewing gum: one in control versus one in xylitol; lozenge: five in xylitol); child disliked the product (one in xylitol chewing gum and 10 in xylitol lozenge); or unknown (syrup: two in control versus two in xylitol; chewing gum: three in control versus three in xylitol; five in xylitol lozenge). The details of the participants enrolled in this study are described in Characteristics of included studies.

Uhari 1996 randomised 306 children in the city of Oulu, Finland daycare (157 in the xylitol group and 149 children in the sucrose group). Each child was instructed to chew two pieces of gum five times a day after meals or snacks until there was no taste (or until five minutes), making a total dose of 8.4 g xylitol per day for two months. Compliance was assessed by asking the parents to report the actual number of pieces of chewing gum used at the end of each month of the trial and also by reporting on the use of additional xylitol products and possible medications. Nasopharyngeal samples were taken before the sugar challenge, at two weeks and at the end of the study. Pneumococci were identified by (alpha) haemolysis on sheep blood agar plates, colony morphology and optochin sensitivity. Parents completed symptom sheets, time and reasons for being absent from daycare, and the use of additional xylitol products and possible medications. At a physician appointment, clinical diagnosis and the drugs prescribed were registered. AOM was diagnosed based on symptoms and signs of ARI and simultaneous signs of middle ear effusion: a cloudy tympanic membrane or impaired tympanic membrane motility in pneumatic otoscopy. The baseline characteristics were similar in the two randomised groups. By the end of trial, there was a dropout of 30 participants; 10 in the xylitol group and 20 in the control group, representing a dropout rate of 6% and 13%, respectively. The reasons for dropouts and the related numbers per each group are as follows: child got tired of eating chewing gum (sucrose: eight versus xylitol one), dental caries (sucrose three versus xylitol four), moved from area (sucrose five versus xylitol one), parents got tired (sucrose two versus xylitol two), insufficient follow‐up data (sucrose: two versus xylitol zero) and loose stools (sucrose zero versus xylitol two). There were no differences in the duration of cough or other symptoms of infections between the groups as reported by parents. There was also no difference in the pneumococcal carriage rates during the study and between the groups as well as the numbers of upper respiratory tract infections without AOM, acute bronchitis, sinusitis and conjunctivitis. The details of the participants enrolled in this study are described in Characteristics of included studies.

Uhari 1998 randomised 857 healthy children enrolled in the city of Oulu, Finland daycare to one of five treatment groups. Participants received control syrup (n = 165), xylitol syrup (n = 159), control chewing gum (n = 178), xylitol gum (n = 179) or xylitol lozenge (n = 176) for three months. Two pieces of chewing gum or lozenges were chewed for at least five minutes five times a day after meals (three doses given by the personnel at the childcare centres during the day and the rest given by the parents at home). Compliance was monitored by asking the parents to list the doses actually given on the daily symptom sheets and also by counting and measuring the unused returned products at the end of the trial. AOM was diagnosed based on a finding of middle ear effusion in tympanometry (B‐ or C‐curve), verified otoscopically with signs of inflammation in the tympanic membrane and the presence of symptoms of ARI (rhinitis, cough, conjunctivitis, sore throat, earache). The baseline characteristics (demographic features, known AOM risk factors or AOM history) and the mean number of forgotten dosages were similar in all five groups. By the end of trial, the dropouts range from 4.5% to 18.9% and were statistically higher in the xylitol syrups (18.9%) and xylitol lozenges (14.8%) as compared to their control groups of respectively control syrups (10.3%) and control chewing gum (4.5%). The reasons for dropouts and the related numbers per each group are as follows: unwilling to continue taking the product (syrup: nine in control versus 14 in xylitol; chewing gum: seven in control versus eight in xylitol; lozenge: 14 in xylitol), left the area (syrup: zero in control versus two in xylitol; chewing gum: zero in control versus two in xylitol; lozenge: one in xylitol), abdominal discomfort (syrup: five in control versus eight in xylitol; chewing gum: zero in control versus one in xylitol; lozenge: seven in xylitol), and unknown reason (syrup: three in control versus six in xylitol; chewing gum: one in control versus one in xylitol; lozenge: four in xylitol). The details of the patients enrolled in this study are described in the Characteristics of included studies table. It should be noted that these trials were all conducted in similar populations in Finland. The methodology appears to be similar with the small change in the control group after the first study (change from sucrose to low‐dose xylitol in the control group), the population in the 2002 study (from healthy children to children with ARI) and the dosing in the 2007 study (five times per day to three times per day).

Excluded studies

We identified a recent trial that evaluated the effectiveness of xylitol paediatric topical oral syrup to reduce the incidence of dental caries among very young children (Milgrom 2009). In the published paper, it was mentioned that the effect of xylitol in reducing AOM will be published in a subsequent study. However, after contacting the primary investigator, it was understood that due to a lack of reliability of data regarding AOM, the plan to publish AOM results has been suspended.

Allocation

With the exception of Vernacchio 2014 for which the sequence generation is unclear, the other four included trials had adequate sequence generation (low risk). For allocation concealment, three trials were considered to have low risk of bias (Hautalahti 2007; Tapiainen 2002; Uhari 1996) and the other two trials an unclear risk (Uhari 1998; Vernacchio 2014).

Blinding

Three out of five included studies had adequate blinding (Hautalahti 2007; Uhari 1996; Vernacchio 2014). The blinding procedure was not mentioned in one study (Uhari 1998) but it was judged to be 'probably done' as per the authors' earlier publication. Tapiainen 2002 was blinded as far as the mixture and chewing gum groups were concerned but open between the xylitol lozenge and control chewing gum groups.

Incomplete outcome data

In Vernacchio 2014, 62/160 allocated to xylitol (38.8%) declined participation, were lost to follow‐up or discontinued intervention versus 58/166 allocated to placebo (34.9%). Intention‐to‐treat (ITT) analysis was performed. The other four trials clarified the number of and the reasons for dropouts, but did not perform an ITT analysis.

Selective reporting

No selective reporting was identified in the five included trials.

Other potential sources of bias

In four Finnish trials the test material was donated by industry (Hautalahti 2007; Tapiainen 2002; Uhari 1996; Uhari 1998). All five trials had governmental funding. No conflict of interest was declared in any of the included trials; however, it should be noted that the study authors in Finnish trials have a US patent for the use of xylitol in treating respiratory infections.

Primary outcome

Number of children with at least one AOM episode during the follow‐up period

The included trials reported on the primary outcome of number of children with at least one AOM episode during the follow‐up period when xylitol was administered to healthy children (Hautalahti 2007; Uhari 1996; Uhari 1998), to healthy children with a respiratory infection (Tapiainen 2002), and to otitis‐prone healthy children (Vernacchio 2014). In particular, for the latter, the population was not homogenous, the total dose was increased by 50% (15 g daily in the form of viscous xylitol solution) and the outcome was measured as what had been documented in the patients' medical charts. Due to the heterogeneity among these three groups, the results are presented separately for each outcome.

Xylitol in any form versus control in any form

Figure 3 summarises available data on our primary outcome. We combined the result of three trials among 1826 healthy children to compare the effect of xylitol in any form (syrup, chewing gum or lozenges) versus control of any kind (syrup or gum) (Hautalahti 2007; Uhari 1996; Uhari 1998). There was a statistically significant 25% reduction in the risk of occurrence of AOM among healthy children at daycare centres in the xylitol group compared to the control group (RR 0.75, 95% CI) 0.65 to 0.88 and RD ‐0.07, 95% CI ‐0.12 to ‐0.03; quality of evidence: moderate ‐ downgraded for inconsistency). However, the effect of xylitol in reducing AOM among healthy children during a respiratory infection (Tapiainen 2002: 99 out of 765 children in the xylitol group versus 56 out of 488 children in the control group, RR 1.13, 95% CI 0.83 to 1.53; quality of evidence: moderate ‐ downgraded for imprecision), or among otitis‐prone healthy children (Vernacchio 2014: 53 out of 160 children in the xylitol group versus 61 out of 166 children in the control group, RR 0.90, 95% CI 0.67 to 1.21 and RD ‐0.07, 95% CI ‐0.14 to 0.07; quality of evidence: low ‐ downgraded for imprecision and high attrition) was inconclusive.

Figure 3

---

---

Forest plot of comparison: 1 Xylitol in any form (syrup, gum or lozenge) versus control in any form (gum, syrup), outcome: 1.1 Final diagnosis of at least one episode of AOM.

Secondary outcomes

Subgroup analysis

Based on the methodology of the identified studies, we defined the subgroup analyses of a) younger children unable to chew gum; b) older children who were able to chew and c) comparison between different types of xylitol vehicles. The subgroup analyses were tested among the population of healthy children, healthy children during respiratory infection, as well as otitis‐prone healthy children. It should be noted that although Hautalahti 2007 included the subgroups of chewing gum and syrup (for both xylitol group and control group), the data were presented collectively for control and xylitol group with no breakdown of the AOM in the subgroups. We contacted the trial authors several times but no further information was obtained and therefore the results of the subgroups of Hautalahti 2007 were not included in the subgroup analysis.

Younger children unable to chew gum

Figure 4 presents data from three trials that compared the effect of xylitol syrup versus control syrup in younger children unable to chew gum (Tapiainen 2002; Uhari 1998; Vernacchio 2014). Due to the heterogeneity in these studies, the results were not pooled. Among healthy children, Uhari 1998 showed that xylitol syrup resulted in a statistically significant 30% decrease in the occurrence of AOM among healthy younger children at daycare centres who were unable to chew gum (46 out of 159 children in the xylitol group versus 68 out of 165 children in the control group, RR 0.70, 95% CI 0.52 to 0.95; RD ‐0.12, 95% CI ‐0.23 to ‐0.02). However, the effect of xylitol syrup in reducing AOM among healthy children during a respiratory infection (Tapiainen 2002: 34 out of 207 children in xylitol group versus 32 out of 211 children in control group, RR 1.08, 95% CI 0.70 to 1.69), or among otitis‐prone healthy children (Vernacchio 2014: 53 out of 160 children in the xylitol group versus 61 out of 166 children in the control group, RR 0.90, 95% CI 0.77 to 1.21; RD ‐0.04, 95% CI ‐0.14 to 0.07) was inconclusive.

Figure 4

---

---

Forest plot of comparison: 3 Xylitol syrup versus control for younger children unable to chew, outcome: 3.1 Final diagnosis of at least one episode of AOM.

Older children who were able to chew

The results of Uhari 1996 and Uhari 1998 were combined to compare the effect of xylitol chewing gum or lozenges versus control chewing gum among the subgroup of older healthy children able to chew gum. Based on the pooled result of these two studies, compared to control chewing gum, xylitol resulted in a statistically significant 34% decrease in the occurrence of AOM if it was administered in chewing gum or lozenges (Figure 5 RR 0.66, 95% CI 0.50 to 0.87; RD ‐0.09, 95% CI ‐0.14 to ‐0.03), and in a statistically significant 41% decrease in the occurrence of AOM if it was administered in chewing gum only (Figure 6 RR 0.59, 95% CI 0.42 to 0.81; RD ‐0.10, 95% CI ‐0.16 to ‐0.04). The effect of xylitol lozenge in reducing AOM was inconclusive (only reported in Uhari 1998, RR 0.80, 95% CI 0.56 to 1.16; RD ‐0.05, 95% CI ‐0.14 to 0.04). Among the healthy children during a respiratory infection (only reported in Tapiainen 2002), there was no difference in the outcome whether xylitol was administered in chewing gum or lozenges (Figure 5 RR 1.34, 95% CI 0.86 to 2.10), in chewing gum only (Figure 6 RR 1.29, 95% CI 0.78 to 2.14), or in lozenges only (RR 1.40, 95% CI 0.85 to 2.29). Vernacchio 2014 did not report this subgroup.

Figure 5

---

---

Forest plot of comparison: 4 Xylitol chewing gum/lozenges versus control chewing gum for older children able to chew, outcome: 4.1 Final diagnosis of at least one episode of AOM.

Figure 6

---

---

Forest plot of comparison: 5 Xylitol chewing gum versus control chewing gum for older children able to chew, outcome: 5.1 Final diagnosis of at least one episode of AOM.

Comparison of different types of xylitol vehicles

We also compared the outcome of the final diagnosis of at least one episode of AOM between the active ingredients groups to assess whether the mode of delivery for the ingredient makes any difference. We performed this analysis on healthy children (Uhari 1998), and also on children with a respiratory infection (Tapiainen 2002). It should be noted that this comparison is limited to one study in each population group: Uhari 1996 did not test different modes of delivery for xylitol (chewing gum only); and although Hautalahti 2007 used two modes of delivery for xylitol (chewing gum and syrup), the data were collectively presented as one test group. The analyses show that xylitol chewing gum is superior to xylitol syrup in preventing AOM among healthy children (RR 0.59, 95% CI 0.39 to 0.89), but not for children with a respiratory infection (RR 0.68, 95% CI 0.43 to 1.07). There was no difference between xylitol lozenges and xylitol syrup in preventing AOM among healthy children (RR 0.77, 95% CI 0.53 to 1.11), or among children with a respiratory infection (RR 0.74, 95% CI 0.47 to 1.14). Similarly, no difference was noted between xylitol chewing gum and xylitol lozenges in preventing AOM among healthy children (RR 0.73, 95% CI 0.47 to 1.13), or among children with a respiratory infection (RR 0.92, 95% CI 0.59 to 1.46). Vernacchio 2014 did not report this subgroup.