Does chlorhexidine mouthwash, with an anti‐discoloration system, reduce tooth surface discoloration without losing its efficacy? A systematic review and meta‐analysis

Abstract Objectives To investigate whether chlorhexidine mouthwash (CHX‐MW), with an anti‐discoloration system(ADS), is effective in preventing extrinsic tooth surface discoloration. Additionally, this paper seeks to evaluate whether CHX combined with an ADS maintains its efficacy with respect to reducing plaque and gingivitis scores. Material and methods MEDLINE‐PubMed and Cochrane‐Central were searched up to October 2018 to identify eligible studies. Papers evaluating the effect of CHX‐MW+ADS compared to CHX without an ADS were included. A descriptive analysis and when feasible a meta‐analysis was performed. Results Screening resulted in 13 eligible publications, presenting 16 comparisons. Six of these evaluated the MW in a non‐brushing model and ten as an adjunct to toothbrushing. A descriptive analysis demonstrated that the majority showed no differences in bleeding, gingivitis and plaque scores. This was confirmed by the meta‐analysis. In non‐brushing experiments, the difference‐of‐means (DiffM) for plaque scores was 0.10 (P = 0.45, 95%CI: [−0.15; 0.34]) and for the gingival index 0.04 (P = 0.15,95%CI: [−0.02; 0.11]). The DiffM in brushing studies for plaque scores was 0.01 (P = 0.29, 95%CI: [−0.01; 0.02]) and for the gingival index 0.00 (P = 0.87,95%CI: [−0.05; 0.06]). With respect to staining scores, the meta‐analysis revealed that in non‐brushing studies, the standardized mean difference was 3.19 (P = 0.0005,95%CI: [−3.98; −1.41]) while in brushing studies, the DiffM was 0.12 (P = 0.95,95%CI: [−3.32; 3.55]). Conclusion There is moderate quality evidence from non‐brushing studies that the addition of an ADS to CHX‐MW reduces tooth surface discoloration and does not appear to affect its properties with respect to gingival inflammation and plaque scores. In brushing studies, there is also moderate quality evidence that ADS does not affect the anti‐plaque and anti‐gingivitis efficacy of CHX. The majority of comparisons and the meta‐analysis including these indicate no significant effect of ADS on tooth staining in situations where the mouthwash is used in addition to toothbrushing.


| INTRODUC TI ON
Gingivitis and periodontitis are perhaps the diseases most common among humans. 1 It has been established that teeth consistently surrounded by inflamed gingiva have a significantly higher risk of being lost than teeth surrounded by no or only slight inflammation.
Persistent gingivitis represents a risk factor for periodontal attachment loss and tooth loss. These may have a negative impact upon speech, nutrition, quality of life and self-esteem, and have systemic inflammatory consequences. 2,3 Gingivitis occurs due to the accumulation of an undisturbed layer of microbial plaque around the oral cavity and tooth surfaces. 4 Dental plaque deposit, the primary aetiologic factor for gingival inflammation, can be prevented by attaining and maintaining high standards of daily plaque removal. A manual or power toothbrush is recommended as a primary means of reducing plaque. 2 In addition, daily use of interdental cleaning devices ensures less interdental bleeding. 5 Using these techniques is generally sufficient to obtain satisfactory oral health. In this way, periodontitis is preventable and leads to reduced rates of tooth loss and improved quality of life. 2,6 Nonetheless, effective patient self-care is not an easy task for everyone. Many people fail to achieve optimal levels of oral care when just brushing their teeth with a dentifrice. If such mechanical cleaning is insufficient, chemical plaque control with adjunctive antimicrobial agents can be considered. 2,7 The anti-microbial agent most frequently advised is chlorhexidine mouthwash (CHX-MW), which can be used as an adjunct to daily oral hygiene for the prevention or treatment of gingival inflammation. Furthermore, CHX-MW can be prescribed after scaling and root planning or tooth extraction. 8 In periodontal surgery, CHX can be prescribed as an temporary alternative to mechanical plaque control. [9][10][11] A large body of literature exists that demonstrates the effectiveness of CHX-MW. Systematic reviews show that in particular, the parameters of plaque reduction and gingivitis significantly improved for those using a CHX-MW compared to those using a placebo. 12,13 Although CHX-MW is currently the most effective anti-microbial agent for reducing plaque and gingivitis, it does have several side effects. An increased calculus formation and decreased taste sensation (hypogeusia) are often reported. Hypogeusia induced by CHX concerns specifically salt and bitter. Salt perception will reach the lowest value on the second day of treatment while the bitter perception on the seventh day, in general, does not change till mouthrinses were interrupted. 14 Other less frequent complaints are a burning sensation, hypersensitivity, mucosal lesions and anaesthetized sensation. 12 However, its major side effect is extrinsic tooth staining, which may have a negative effect on patient compliance with rinsing. 9,12,14 For more than a decade, several commercial CHX-mouthwashes with an anti-discoloration system (ADS) have been available in different countries. Several studies have been performed; however, the results published regarding its effectiveness have been inconclusive. 15,16 It has been observed that an ADS can be effective in reducing stain, but it may potentially also reduce the clinical efficacy of CHX products. 16 This has been summarized in the past in the following simplified manner: "if it does not stain, it does not work." 17 The purpose of this systematic review (SR) is to synthesize the available scientific literature to investigate whether adding an ADS to CHX-MW is effective in preventing extrinsic tooth surface discoloration, as well as evaluating whether CHX combined with an ADS maintains its efficacy with respect to reduction of plaque and gingivitis.

| MATERIAL AND ME THODS
The preparation and presentation of this SR is in accordance with the Cochrane Handbook for Systematic Reviews of Interventions 18 and the guidelines of Transparent Reporting of Systematic Reviews and Meta-Analyses (PRISMA). 19 A protocol 20 was developed a priori following the initial discussion between the members of the research team. The focused questions of the review were as follows: • The first focused question: What is the effect of rinsing with a CHX-MW containing an ADS, as opposed to rinsing with a standard CHX-MW, on tooth surface discoloration?
• The second focused question: What is the effect of rinsing with a CHX-MW containing an ADS, as opposed to rinsing with a standard CHX-MW, on plaque and gingivitis scores?  the International Association for Dental Research (http://www.iadr. org). CHX product companies involved in the field of ADSs were contacted in an effort to trace unpublished or ongoing studies. Table 1 provides details regarding the search terms used. There were no restrictions regarding language or publication year.

| Screening and selection
Initially, the titles and abstracts (when available) of all studies identified through the searches were scanned by two reviewers independently (BVS and DES), who then selected studies that potentially met the inclusion criteria. After this phase, full-text versions were obtained for the studies that appeared to meet the inclusion criteria or for which the title and abstract provided insufficient information to make a clear decision. These studies were then categorized as "definitely eligible," "definitely not eligible" or "questionable." Disagreements concerning eligibility were resolved by consensus or, if disagreement persisted, by arbitration through a third reviewer (GAW). The papers that fulfilled all inclusion criteria were processed for data extraction. No language restriction was imposed.
The inclusion criteria were as follows:

| Methodological quality assessment
Two reviewers (BVS and DES) independently scored the individual methodological qualities of the studies included in this meta-analysis using the checklist presented in Appendix S1. Quality criteria were designated with a positive sign (+) if an informative description was present, and if the study design met the methodological criteria, a negative sign (-) if an informative description was present but the study design did not meet the criteria and a question mark (?) if information was missing or insufficient. A study was classified as having a "low risk of bias" when positive scores (+) were assigned to the criteria of random allocation, defined inclusion/exclusion criteria, blinding to product and examiner, balanced experimental groups, identical treatment between groups (except for the intervention) and reporting of follow-up. Studies that had six of these seven criteria were considered to have a potential "moderate risk of bias". If two or more of these seven criteria were absent, the study was considered to have a "high risk of bias". 21

| Data extraction
All studies that met the inclusion criteria were selected for data extraction and a "risk of bias" assessment. Independent data extraction was performed by two reviewers (BVS and DES) using a specially designed standardized data extraction form. Data recorded from the studies included here were based directly on the focus of the research questions, including details of the population, intervention, comparison outcome and study characteristics. Disagreement between the reviewers was resolved through discussion until a consensus was reached. Any persisting disagreements were resolved by discussion with a third reviewer (GAW). If any missing data or information were identified, an attempt was made to contact the authors of the publication to request additional information.

| Assessment of clinical and methodological heterogeneity
The factors used to assess the clinical heterogeneity of the outcomes of the various studies were as follows: characteristics of participants, groups, variation of the CHX concentration in the MW, evaluation period, side effects and industry funding. Factors to assess the methodological heterogeneity were diversity in study design. When clinical or methodological heterogeneity was considered to be too high across studies, sources of heterogeneity were investigated with subgroup and/or sensitivity analyses. When the individual studies were sufficiently similar with respect to included patients, treatments and outcomes, pooling of results was considered and statistical heterogeneity assessed.

| Assessment of statistical heterogeneity
Poor overlap of confidence intervals generally indicates the presence of statistical heterogeneity. Heterogeneity was statistically tested by the chi-square and I 2 tests. Tau-squared was used to estimate the between-studies variation. A chi-square test resulting in a P < 0.1 is considered an indication of significant statistical heterogeneity. As an approximate guide to assessing the possible magnitude of inconsistency across studies, an I 2 statistic of 0-40% was interpreted to indicate unimportant levels of heterogeneity. An I 2 statistic of 30%-60% may represent moderate heterogeneity an I 2 statistic of 50%-90% may represent substantial heterogeneity while a statistic of greater than 75% was interpreted to indicate considerable heterogeneity. This form of heterogeneity was assessed with subgroup and or sensitivity analysis to assess the effect modification. 22

| Descriptive methods
As a summary of data, a descriptive data presentation was used for all studies. It was decided "a priori" to categorize the studies into either monotherapy studies (non-brushing studies) or studies that also included self-performed daily oral hygiene (brushing studies).
Discoloration scores, plaque, bleeding and gingivitis were taken into account.

| Quantitative methods
If quantitative methods were feasible, a meta-analysis was performed to explore the effectiveness of CHX-MW+ADS vs CHX-MW alone within various parameters. Analysis was carried out using Review Manager version 5.3 according to the PRISMA guidelines. 19 In studies consisting of multiple treatment arms, and in which data from one particular group were compared to the data of more than one other group, the number of subjects (n) in the group was divided by the number of comparisons. In cases where it was not possible to perform a meta-analysis, only a descriptive analysis is reported. A metaanalysis was performed if more than one study could be included.
When the pooled outcome of several studies was measured using the same unit, then it was expressed as a difference-of-means (DiffM) with its associated 95% confidence interval. When the primary outcome was measured using different units across studies, then the standardized mean difference (SMD) was used to combine the outcomes in the meta-analyses. 23 The DiffM between test and control was calculated using both the "random and fixed effects" model where appropriate. When there is heterogeneity that cannot readily be explained, one analytical approach was incorporated into a random-effects model.
Random-effects models are well suited for meta-analysis with heterogeneous effects. A fixed-effect model was presented if there were fewer than four comparisons, because the estimate of between-study variance is poor for analyses with low numbers of studies. 18 The testing for publication bias per outcome was used as proposed by Egger et al. 24 If the meta-analysis involved sufficient trials to make visual inspection of the plot meaningful (a minimum of 10 trials), funnel plots were used as a tool to assess publication bias. The presence of asymmetry in the inverted funnel would suggest a systematic difference between large and small trials in their estimates of treatment effects-a difference that may occur, for example, because of publication bias. 18,19

| Grading the "body of evidence"
The Grading of Recommendations Assessment, Development and Evaluation (GRADE) was used to rank the evidence. 25 Any disagreement between the two reviewers was resolved through additional discussion.

| Search and selection results
Searching

| Assessment of clinical heterogeneity
Heterogeneity was observed in the 13 clinical trials with respect to participants, and mouthwash (MW) brands used in the brushing/rinsing regimen among the studies.

| Side effects
The papers used in this meta-analysis did not report any serious ad-

| Industry funding
Three studies do not mention any details regarding funding or con-

| Assessment of methodological heterogeneity
All studies were RCTs, of which seven used a crossover design (I 15

| Methodological quality assessment
The potential risk of bias was estimated based on the methodological quality aspects of the selected studies, as presented in the online Appendix S1. Based on a summary of the proposed bias-assessment criteria, the potential risk of bias was estimated to be moderate for Studies I 15 and III 37 and low for the other studies. Sub-analysis was performed only for studies with a low risk of bias.

| Study outcome results
The online Appendix S2, sub-sections a-d, presents the results of the data extraction that was performed on the selected studies in various clinical indices. When available, the baseline, end scores and changes between baseline and end scores are presented.

| Description of findings
In detail, Table 3  showed that an ADS negatively influenced plaque score, and one positively influenced it.

| Meta-analysis
It was possible to perform a meta-analysis for the comparisons between products assessing stain scores for non-brushing and brushing studies. For non-brushing studies, a significant difference was found in the SMD for end scores as well as for the incremen-  Table 4a,b summarizes the detailed data of the outcomes of the meta-analysis. Online Appendices S3-S8 present the corresponding forest plots. A test for publication bias could not be performed because fewer than 10 studies were included in the meta-analysis, which would result in insufficient statistical power. 18,24 Consequently, publication bias cannot be ruled out. Sub-analysis of studies that possessed a low risk of bias did not reveal any significant discrepancies with the original analysis. The heterogeneity is exposed and stays unclarified (Appendix S9a-c).

| Sensitivity analysis
In the meta-analysis of those studies that evaluated the intervention under non-brushing circumstances, considerable heterogeneity was observed. For instance, the meta-analyses for stain scores showed an I 2 of 87%, 97% and 97% for baseline, end and incremental scores, respectively. A sensitivity analysis was performed to explore the source of heterogeneity which showed that without the outlying study X 28 , lower heterogeneity was present between the outcomes of the studies both at baseline (I 2 = 0%) and for incremental scores (I 2 = 69). For the end score, the I 2 remained high (94%) for which no obvious explanation was found. The studies included in the meta-analysis do differ by study design, being either crossover or parallel. In the meta-analysis on plaque scores, the sensitivity analysis was performed by study design. Meta-analysis that only included those with a crossover design showed a decrease in I 2 for end scores of plaque from 99% to 89%, and for incremental scores from 94% to 70%. When only the parallel designs are taken into account, no evident explanation was found as well. The I 2 still remained high. For gingival scores, if the study VII 16 with the smallest sample (N = 8) size is excluded, the I 2 for end scores and incremental scores decreased from 94% and 95%, respectively, to 0% for both. None of these sensitivity analyses did affect the overall result and conclusions. Table 5 presents a summary of the various criteria with which the quality of the evidence was rated and with which the strength and TA B L E 3 A descriptive summary of statistical significance levels of the use of chlorhexidine mouthwashes with or without an antidiscoloration system, with or without alcohol and without brushing or as adjuvant to toothbrushing on the parameters of interest direction of recommendations were appraised according to Guyatt et al. 2008. 41 The addition of an ADS to CHX-MW is favourable with respect to reducing tooth surface discoloration, and it does not appear to affect the inhibition of plaque and gingivitis scores. Given the strength of the recommendation, there is a weak-to-moderate certainty that the addition of an ADS does not negatively influence the effect of CHX-MW on plaque scores and gingival inflammation.

| Evidence profile
Given that only in studies with a non-brushing design, it significantly reduces tooth surface discoloration, the direction of the recommendation for situations where toothbrushing is not involved is moderately in favour of the use of CHX-MW+ADS.

| D ISCUSS I ON
There is a strong body of evidence in support of CHX-MW 12,13 ; however, one of the most prominent side effects is tooth staining. The

| Interpreting of staining analysis
In the non-brushing studies using a non-brushing model, all six comparisons in the descriptive analysis (Table 3), except Study II, 36 significantly favoured CHX-MW+ADS for stain scores. This was confirmed by the meta-analysis, based on five comparisons. It should be taken into account that the trials comparing CHX-MW+ADS to CHX-MW for tooth staining included subjects diagnosed with both gingivitis and periodontitis. The meta-analysis for stain scores was performed using the standardized mean difference as a summary statistic because studies measure the outcome in a variety of ways. 18 A total of ten comparisons used the mouthwash as an adjunct to toothbrushing. In the descriptive analysis (Table 3)   Overall recommendation If CHX-mouthwashes are prescribed for gingivitis and plaque reduction, and there is interest to reduce staining, there is moderate evidence for non-brushing situations to consider a product containing an anti-discoloration system even though the authors were contacted for additional information.
Nevertheless, based on the combined data from the descriptive analysis of non-brushing and brushing designs and meta-analysis of the non-brushing designs, the addition of an ADS to CHX-MW appears favourable with respect to preventing tooth surface discoloration. As a result, the first question can be answered: CHX-MW+ADS has the potential to reduce the side effect of tooth staining.

| Possible mechanisms
The staining side effect associated with CHX rinsing is attributed to

| Anti-microbial activity
The intra oral structures. It seems plausible that these components may interfere with CHX. 36 This would accord with findings indicating that a reduction in the tendency to stain may also lead to a loss of plaque inhibition. It is also possible that in vivo, there is a continuous competition between anti-plaque and anti-staining processes.
Another uncertainty is the difference in outcomes of "in vitro" compared to "in vivo" research regarding it clinical relevance. In an early "in vitro" study, it was shown that no significant difference in staining existed between the ADS rinses and the positive control rinse. 17 In addition, in a polyspecies biofilm model, the effect of CHX-MW+ADS was evaluated, showing that all solutions containing CHX significantly reduced the number of microorganisms in biofilms. The CHX without an ADS proved most effective in reducing the total number of bacterial colonies. It was therefore proposed that regular CHX mouth rinses are best confined to short-term therapeutic use and the addition of ADS solutions would be indicated for a long-lasting prophylactic application. 42 This conflicts with the results of the current review. A similar phenomenon was observed when a sodium lauryl sulphate (SLS) dentifrice was used in combination with a CHX-MW. In vitro, SLS and CHX may act as antagonists. 43 Based on a recent SR of clinical trials, the combined use of an SLS-containing dentifrice and CHX-MW is not contraindicated. 44 Therefore, it may be concluded that CHX does not act similarly in vitro compared to "in vivo." MW. 46 Part of a pre-experimental period for this specific model is a professional prophylaxis and optimal self-performed plaque control to establish a healthy gingiva. 46 All non-brushing experiments provide such a prophylaxis, but only Study VII 16 also concluded the pre-experimental preparatory phase of optimal oral hygiene practices. In addition, in the past, it has been proposed that the period without mechanical plaque control should extent over at least 14 days. 46 Table 2). Study designs may influence the heterogeneity. 18 The present review excluded surgical procedures as part of a study protocol of interest. The search revealed that some papers are published on the topic CHX with or without ADS as adjunct used by periodontal flap surgery. 10,47,48 Non-surgical periodontal therapy differs from resective or regenerative procedures by its origin and indication. As the non-brushing studies in this review mostly refer to experimental gingivitis conditions, and not post-surgery use of the mouthwash, it seems of interest to evaluate the staining properties with the specific study model of periodontal surgery in a future systematic review.

| Clinical and methodological heterogeneity
Out of the 13 included studies, two (V 31 and XI 33 ) had industry involvement. It is well-established that publication bias may be associated with the source of funding for a study. The main origin of this bias is failure to publish negative or null findings. The consequence is that it may lead to overestimates of treatment effects in meta-analyses. 49 Industry involvement did however not provide an explanation as a potential source of observed heterogeneity. Moreover, grey literature did not reveal any unpublished studies.

| Statistical heterogeneity
I 2 is the ratio of true heterogeneity to the total variation in observed effect, which can be interpreted as a signal-of-noise ratio. It is not sensitive to the metric of the effect size nor to the number of included studies. 50 I 2 was found to be 0%-26% for brushing studies (see Table 4b). This was interpreted as potentially unimportant with respect to heterogeneity. 22 For non-brushing studies in the metaanalysis, considerable heterogeneity was mainly observed in the end scores and incremental difference scores (I 2 = 97%-99% There are several options to address (statistical) heterogeneity.
For the present review, it was chosen to explore heterogeneity by performing sensitivity analysis. This is a repeat of the primary analysis or meta-analysis, substituting alternative decisions or ranges of values for decisions that were arbitrary or unclear. It is an informal comparison made between different ways of estimating the same thing. Some sensitivity analyses can be prespecified in the study protocol, but often only identified during the review process where the individual peculiarities of the studies under investigation are identified. 18 The latter was the case during preparation of this systematic review. For the sensitivity analyses, different factors as source of heterogeneity were explored being outliers, study design and sample size. The overall result and conclusions were not affected by the sensitivity analyses although it had an effect on the statistical heterogeneity expressed by I 2 . Consequently, the results of this review can be regarded with a higher degree of certainty. However, when the testing for heterogeneity is significant, the reader should always exercise caution in using the effect size that emerges from the meta-analysis, because the estimate may not reflect the actual effect in any particular population being studied. 22

| CON CLUS ION
There is moderate quality evidence from non-brushing studies that the addition of an ADS to CHX-MW reduces tooth surface discoloration and does not appear to affect its properties with respect to gingival inflammation and plaque scores. In brushing studies, there is also moderate quality evidence that ADS does not affect the anti-plaque and anti-gingivitis efficacy of CHX. The majority of comparisons and the meta-analysis including these indicate no significant effect of ADS on tooth staining in situations where the mouthwash is used in addition to toothbrushing.

| Scientific rationale for the study
The most common side effect of chlorhexidine is extrinsic staining of oral surfaces. An anti-discoloration system presumably reduces staining while maintaining the chlorhexidine efficacy.

| Principal findings
A significant benefit was found in favour of chlorhexidine mouthwash with an anti-discoloration-system in 4-21 days non-brushing studies for stain scores. No differences in the clinical parameters of plaque, bleeding and gingival index scores were found for either brushing or non-brushing studies (ie experimental gingivitis conditions).

| Practical implications
When a chlorhexidine mouthwash is prescribed, there is moderate evidence for 4-21 days non-brushing situations, that a product containing an anti-discoloration system can be considered in order to reduce side effects. This may potentially improve patient compliance.

ACK N OWLED G EM ENTS
The authors acknowledge the support of Thalina Wamelink-van

CO N FLI C T O F I NTE R E S T
The authors declare that they have no conflict of interests.

AUTH O R CO NTR I B UTI O N
All authors gave final approval and agreed to be accountable for all aspects of work ensuring integrity and accuracy. BVS: contributed to design, search and selection, analysis and interpretation, and drafted the manuscript. GAW: contributed to conception and design, analysis and interpretation, and critically revised the manuscript. EB: contributed to analysis and interpretation, and critically revised the manuscript. FG: contributed to design, analysis and interpretation, and critically revised the manuscript. DES: contributed to conception and design, search and selection, analysis and interpretation, and critically revised the manuscript.