Commentary

It has been proposed that CHX-V, with its antibacterial properties can reduce caries risk1, through its effect on mutans streptococci 2 However, two earlier reviews3, 4 of the effectiveness of CHX-V on caries in children and adolescents have proved inconclusive, as has a review by Twetman 5 for arresting root caries in elderly patients with xerostomia. There is an increased risk of root caries as people increasingly retain their teeth through their life course,6 and therefore a need to find ways of effectively preventing root surface caries within this growing cohort. This systematic review is therefore timely.

The review's question could perhaps have been slightly better focused, PICO style,7 to enable readers to understand exactly which population, intervention, comparison and outcome were being reviewed. As Cochrane recommends, a systematic review should identify all relevant studies from a number of different sources.8 A clear, pre-determined strategy was used to determine inclusion of randomised controlled trials, with independent assessment by two assessors. Slot searched a number of important databases, and included follow up of the reference lists from all selected studies, but did not include searches of grey literature, unpublished studies (e.g. clinical trials databases) or non-English-language databases (e.g. Lilacs, the Latin-American and Caribbean database.) Such exclusions would both reduce the overall papers reviewed, and increase the risk of publication bias. As noted by Egger9 authors are more likely to report positive results in an international journal, and more likely to report negative results in a local non English publication.

The authors give both quality parameters (internal and external validity, statistical methods) and a score to each of the reviewed papers. The individual rating of each domain is admirable as it allows readers to easily judge for themselves the quality of each paper under each specific feature of the study. Scoring systems such as the CEBM or Jadad scales, yielding a summary score, are however not recommended by Cochrane due to the inherent bias in calculating such a score.8

Within the reviewed papers there was a lack of comprehensive information as suggested by the CONSORT statement (www.consort-statement.org). There was no information given on the baseline and end point data for root caries activity, and no consistent measurement of DMF-RS. Characteristics of the patients (e.g. age) were either omitted or not comparable across papers, making it difficult to generalise the specific results. It may have been possible to retrieve this information through contact with the authors, but this does not seem to have occurred. Within the selected papers there was considerable heterogeneity, with variation in CHX-V concentrations, inclusion/exclusion criteria, the comparison group (control, fluoride varnish) and application regime (including cleaning regimes). There were also different methods used to measure both caries indices and activities. Such heterogeneity meant that only three papers were included in the meta-analysis. With the information provided it is difficult for the reader to determine whether the papers were similar enough to perform a meta-analysis. One way of overcoming this uncertainty would have been to include a forrest plot to provide a graphic illustration of the variation between studies. Within a Cochrane systematic review it is customary, within a meta-analysis, to use a form of weighted average, where some of the studies (depending on their size, level of bias etc) contribute more to the total than others. Slot's meta-analysis calculated weighted mean differences using Cochrane's Review Manager; it would however have been useful to have an explanation as to how the weighting was constructed.

The results were well presented with both descriptive and individual data. The results were presented as means +/- standard deviations. As Petrie10 suggests, the use of +/- should be avoided in data presentation, with the use of confidence intervals a preferred format for this display. The overall reduction of 0.65 in DMF-RS calculated through the meta-analysis had a reasonably wide confidence interval indicating the uncertainty of the analysis. As well as using systems to select and review individual papers, the authors also used the systematic scoring system GRADE 11 based on the trial type quality to assess the quality of this review as ‘weak’. The scoring system within this paper was not shown as part of this review.

In the conclusion it states that for special care patients in the absence of regular toothbrushing and oral hygiene instructions, CHX-V may provide a beneficial effect. While the evidence presented is suggestive of a greater effect in this group of patients, the quality of the evidence does not seem sufficient to make this a firm recommendation.

Practice points

  • There is currently only weak evidence of a positive effect of CHX-V on root caries and further high quality studies are necessary to confirm effectiveness.