Ijaz/Croucher/MarinhoCaries Res 2010;44:579–592580tistry though [Marinho, 2003], and there is still room forimprovement in the quantity, in terms of subject areascovered, and quality of both the conduct and the report-ing of systematic reviews in the field [Glenny et al., 2003;Bader and Ismail, 2004; Richards, 2004; Major et al.,2006].The various topical fluoride interventions have oversix decades of experimental research supporting theirvalue as anticaries measures, and these were the focus ofsome of the first meta-analyses in dentistry [Stamm et al.,1984; Clark et al., 1985]. Existing systematic reviews oftopical fluoride interventions may reflect the availableprimary research in the field, and their reporting qualitymay vary. Nevertheless, the breadth of topical fluorideinterventions, comparisons, populations, and outcomescovered by published systematic reviews and meta-anal-yses, and the methodological quality of these reviews,have not been reported previously. Highlighting whichdirection future reviews in this area should take based onpast practice to avoid duplication of effort while advanc-ing knowledge reliably is important, and this should bedone in the light of ongoing methodological advance-ments so as to enhance their validity and applicability.In 1996, an international group developed a reportingguideline, the QUOROM Statement (QUality Of Report-ing Of Meta-analyses), to improve the reporting of sys-tematic reviews and meta-analyses in healthcare research[Moher et al., 1999]. The recent publication of ‘Preferredreporting items for systematic reviews and meta-analy-ses’, the PRISMA statement, is an improvement on thatguideline [Liberati et al., 2009; Moher et al., 2009]. In ad-dition to the QUOROM items, PRISMA now asks the re-view authors to make a protocol for the review accessible;to report at least one complete electronic search; to assessrisk of bias in and across included studies (along with theselective reporting of outcomes); to report limitations ofthe review, future research implications and sources offunding. The PRISMA statement has made it clear thatthe quality and content of systematic reviews is constant-ly being improved, and has also identified The CochraneCollaboration as the leader in the field since all these newitems are already an essential part of Cochrane reviews.Thus, new reviews incorporating these advancementsshould have enhanced quality and generate more reliableestimates of treatment effects.The aim of this paper is to describe the most importantfeatures of systematic reviews and meta-analyses on top-icalfluoridesforcariespreventionandtreatmentintermsof content and reporting quality by collating those pub-lished up to and including 2009. A comprehensive searchand an appraisal in the light of current reporting guide-lines (PRISMA) has enabled us to identify important is-sues that need consideration by researchers planning asystematic review in this area of dentistry.Materials and MethodsAn electronic search without date or language restrictionswas undertaken in December 2009 of: MEDLINE (via OVID in-cluding old MEDLINE), The Cochrane Database of SystematicReviews (CDSR) and The Centre for Reviews and Dissemination(CRD) databases [including DARE (Database of Abstracts of Re-views of Effects), NHSEED (National Health Service EconomicEvaluation Database), and HTA (Health Technology Assess-ment)]. The search terms ‘caries AND fluoride’ were used in allthe searches performed, but a more detailed search strategywas used in MEDLINE (online supplementary material, www.karger.com/doi/10.1159/000322132). This was supplemented bysearching the reference lists of the included reviews.Only completed published intervention systematic reviewsand meta-analyses, including any types of study design, assessingthe anticaries effects of any topical fluoride therapy (TFT), in-cluding toothpaste, mouthrinse, gel, varnish, paint-on solution,whether used alone, or in combination with another TFT, or aspart of a larger group of anticaries interventions in any populationgroup were considered relevant for inclusion. Any review that in-corporated at least two of the key characteristics of a systematicreview outlined in section 1.2.2 of The Cochrane Handbook of Sys-tematic Reviews of Interventions [Higgins and Green, 2008] wasconsidered.For each of the identified reviews included in this study,the following data were then extracted on a standard form: yearof publication; whether a clear objective and eligibility criteria(in terms of participants, interventions, comparisons and out-comes–PICOS)werestatedinthereview;typesofintervention(s),comparison(s) and population group(s) addressed; main out-comes addressed; number and types of studies included; methodsused for study identification, study selection, quality assessment(risk of bias) and data extraction of included studies; effect mea-sures used; whether a quantitative synthesis was done; statisticalmethods used in the synthesis (fixed effect or random effects); as-sessment, exploration and factors explored for heterogeneity; as-sessment and exploration of publication or reporting bias; wheth-er cost/economic evaluation was reported and in what format.These data were then summarized so as to identify the most im-portant issues in the conduct and reporting of topical fluoridereviews. Search strategy development, study selection and datacollection were done by one author (S.I.) in consultation with theother (V.C.C.M.).ResultsThe searches resulted in 150 citations from MED-LINE, 17 from CDSR, 39 from DARE, 17 from NHSEED,and 7 from HTA databases. These results (230 titles and
Topical Fluorides Systematic Reviews Caries Res 2010;44:579–592 581abstracts) were read for identifying relevant systematicreviews and meta-analyses of topical fluoride interven-tions for caries. Reasons for exclusions at this stage were:duplicate citations (n = 64), traditional narrative reviews(n = 46), reviews of non-TFT interventions (n = 42); tech-nology or economic assessments that were not systematicreviews (n = 23).The remaining 55 papers were then read in full text.Of these, five were protocols of ongoing systematic re-views, six were response publications to already pub-lished meta-analyses, three were method papers, andthree were commentaries on a TFT systematic review.One did not report the review methodology in sufficientdetail while one focused purely on assessment of epide-miological data and did not assess effectiveness of theTFT intervention. These 19 papers were excluded.Thirty-six papers were included. Two other reviewswere identified from the reference lists of the includedreviews. Thus a total of 38 systematic reviews addressingeffectiveness of some form of TFT for caries were identi-fied. Tables 1 and 2 summarise the data extracted fromthese reports (general content and methodological fea-tures, and features specifically related to data synthesis,respectively). The oldest included review was a meta-analysis from 1984 and the latest from 2009. The greatestnumber of reviews was published between 2000 and 2004(fig. 1).Reporting of Objectives and Eligibility CriteriaAlthough all publications stated objectives at least interms of ‘intervention for condition’, most of these didnot provide complete information on eligibility criteria inPICOS format (population, intervention, comparison,outcome, study design), except in Cochrane reviews.Characteristics of Studies Included – PICOSPopulations. The most covered group was childrenand/or adolescents from the general population (n = 21,55%). Three reviews addressed fixed orthodontic appli-ance wearing patients, two addressed adults, and twocovered high caries risk individuals only. Eight did notspecify the population group addressed.Interventions. The topical fluoride intervention mostcommonly addressed was toothpaste (n = 10, 26%). Var-nish was the main intervention of interest in nine re-views, mouthrinses in four, and gels in three. Nine re-views addressed any TFT, while two addressed combina-tions of TFT. One review specifically addressed silverdiamine fluoride – a type of paint-on solution.Comparisons. The comparison most commonly ad-dressed was intervention (TFT) versus placebo or notreatment (n = 16, 42%). This was followed by compari-sons of one concentration of fluoride toothpaste versusanother (n = 5), and comparisons of one specific fluorideagent/compound in toothpaste versus another (n = 3).Two reviews addressed comparisons of one type of TFTagainst another, while 13 (34%) did not specify the com-parisons addressed.Outcomes. Caries increment was the most commonoutcome measure (n = 31, 81%), followed by caries inci-dence (n = 4), caries arrest (n = 3) and caries progression(n = 2). Adverse effects and patient-reported outcomeswere not measured as a primary/main outcome in anyreview (although adverse events were mentioned as sec-ondary outcomes in 11 reviews).Study Design. The most common type of study consid-ered was controlled clinical trial (n = 34, 90%). Four re-views considered various non-RCT designs (prospectivecontrolled studies, uncontrolled longitudinal studies,and even reviews and guidelines) for inclusion and tworeviews considered economic evaluations.Search Methods UsedSix reviews (16%) did not report any search methods(sources, dates, languages, search terms, etc.) for studies,although five of these were published before the first re-porting guideline came out in 1999 with clear recom-mendations on this aspect [Moher et al., 1999]. Only 12(31%) of the 38 reviews reported a complete search strat-egy for at least one database – mostly Cochrane reviews,while 17 reported search terms, key words or conceptsused for searching.Seventeen reviews (45%) reported searching multipledatabases (more than two) electronically while six report-ed searching MEDLINE and The Cochrane Library da-tabases, apparently treating The Cochrane Library as onedatabase. Nineteen (50%) reported complementing elec-tronic search with searching one or more other sources(reference/hand/grey literature/authors).Study Selection and Data ExtractionThree reviews (8%) did not report on the study selec-tion process. Study selection was reported to be per-formed in duplicate in 27 (71%) reviews. Of these, 20 re-views reported duplicate study selection in the entiresample while seven reported the same in a random onethird of the sample. For eight reviews it could not bejudged if it was done in duplicate.
Ijaz/Croucher/MarinhoCaries Res 2010;44:579–592586Authors +year ofpublicationStudiesincludedPrimaryeffectmeasuresMethod ofquantitativesynthesisAssessment/exploration of heterogeneity Assess./explo-ration of pub-lication biasEconomicevaluationRosenblattet al. 2 PF, NNT NR NR NR NRAzarpazhoohet al. 7 NR NR NR NR NRTwetman22 NR NR NR NR NRHeijnsbroeket al. 6 NR NR NR NR NRGriffinet al. 20 RR, MD RE MA 2, I2/sensitivity analysis for study design NR NRAmmariet al. 7 PF NR NR NR NRHiiriet al. 4 RR NR NR NR NRBensonet al. 15 WMD,Peto’s ORNR NR NR NRChadwicket al. 6 PF NR NR NR NRAxelssonet al. 24 NR NR NR NR NRBaderet al. 6 (FV) % cariesreductionNR NR NR NRPeterssonet al. 24 PF NR(likely done)NR NR NRTwetmanet al. 25 PF NR(likely done)NR/correlation analysis on intervention and populationfeaturesNR NRDerkset al. 15 PF NR NR NR NRMarinhoet al. [2004b]17 (15for MA)PF RE MA 2 + I2 tests/RE MR sensitivity subgroup analyses on pre-defined and not pre-defined aspects of study quality, andparticipants, intervention, and outcome-related featuresfunnel plot/formal testsNRMarinhoet al. [2004a]12 (9for MA)PF RE MA 2+ I2tests/RE MR sensitivity subgroup analyses on pre-defined and not pre-defined aspects of study quality, andparticipants, intervention, and outcome-related featuresfunnel plot/formal testsNRSteineret al. 4 PF RE + FE MA Cochran test/NR NR NRTwetmanet al. 54 PF NR NR NR NRAmmariet al. 7 (5 forMA)WMD FE MA NR/sensitivity analysis for intervention feature funnel plot/NRNRKällestålet al. 17 (2 FV,3 FMR)NR NR NR NR identified +quality-assessedeconomicevaluationsMarinho[2003a]144 (133for MA)PF RE MA 2+ I2tests/RE MR, sensitivity subgroup analyses on pre-defined and not pre-defined aspects of study quality, andparticipants, intervention, and outcome-related featuresfunnel plot/formal testsNRMarinho[2003c]36 (34for MA)PF RE MA 2+ I2tests/RE MR, sensitivity analyses on pre-defined andnot pre-defined aspects of study quality, and participants,intervention, and outcome-related featuresfunnel plot/formal testsNRMarinho[2003b]74 (70for MA)PF RE MA 2+ I2tests/RE MR, sensitivity analyses on pre-defined andnot pre-defined aspects of study quality, and participants,intervention, and outcome-related featuresfunnel plot/formal testsNRTable 2. Summary of data synthesis features in systematic reviews and meta-analyses of TFT
Topical Fluorides Systematic Reviews Caries Res 2010;44:579–592 587For eleven reviews (29%) the data extraction methodwas not reported and for ten we could not judge whetherdata extraction was done in duplicate. Duplicate data ex-traction was clearly reported in 16 (42%) reviews; ninereviews reporting duplicate independent data extractionfor the entire sample and seven reporting the same in arandom one third of the total sample of studies. One re-view reported single abstraction of data. Only Cochranereviews and those from The Swedish Council of Technol-ogy Assessment in Health Care reported using pilotedforms for data extraction.Methodological Quality AssessmentThirteen reviews (34%) did not report assessing thequality of included studies. However, an equal number ofreviews reported independent duplicate assessment ofstudy quality in the entire sample. Another seven reviews(the series of Cochrane reviews 2002–2004) reported do-Authors +year ofpublicationStudiesincludedPrimaryeffectmeasuresMethod ofquantitativesynthesisAssessment/exploration of heterogeneity Assess./explo-ration of pub-lication biasEconomicevaluationMarinhoet al. [2002b]25 (23for MA)PF, SMD RE MA 2 + I2 tests/RE MR, sensitivity subgroup analyses on pre-defined and not pre-defined aspects of study quality, andparticipants, intervention, and outcome-related featuresfunnel plot/formal testsNRMarinhoet al. [2002a]9 (7 forMA)PF, SMD RE MA 2+ I2tests/RE MR, sensitivity analyses on pre-defined andnot pre-defined aspects of study quality, and participants,intervention, and outcome-related featuresfunnel plot/formal testsNRChaves andVieira-da-Silva22 % difference FE MA Q test/NR NR NRRozier7 (FV) PF, NNT NR NR NR NRBaderet al. 27 NR NR NR NR NRBartizeket al. 6 SMD MA (type NR) 2/NR NR NRStrohmengerand Brambilla3 WMD RE MA Q test/NR NR NRVan Rijkomet al. 17 PF MA (type NR) NR/MR analysis on population and intervention features funnel plot/NRreported costinfo + NNTVolpeet al. 4 WMD MA (type NR) NR/subgroup analysis for study design features NR NRHelfensteinand Steiner8 PF RE MA 2/correlation analysis on aspects of study quality, andparticipants, intervention, and outcome-related featuresNR NRJohnson10 (9 forMA)WMD MA (type NR) 2/subgroup analysis for study design features NR NRStookeyet al. 13 WMD MA (type NR) 2/subgroup analysis for study design features NR NRBeiswangerand Stookey20 % difference NR(likely done)NR NR NRClarket al. NR WMD RE MA NR NR NRStammet al. NR WMD MA (type NR) NR NR CEA.NR = Not reported; OR = odds ratio; MD = mean difference; PF = prevented fraction; RR = risk ratio; SMD = standardized mean difference;WMD = weighted mean difference; FE = fixed effect; MA = meta-analysis; MR = meta-regression; RE = random effect; CEA = cost-effective analysis;USD = US dollar.Table 2 (continued)
Ijaz/Croucher/MarinhoCaries Res 2010;44:579–592588main-based assessment of risk of bias in duplicate in arandom one third of included studies. Of the rest, oneused single assessment and four did not give enough de-tail to allow judgement of duplicate assessment.A total of 16 (42%) reviews (including the Cochranereviews and Swedish Council of Technology Assessmentin Health Care series of reviews) reported quality as-sessment in the domains of allocation concealment andblinding. Four reviews reported using scores for qualityassessment – all published after 2000. An assessment ofrisk of selective outcome reporting bias within studieswas not specified by any of the included reviews, reflect-ing its very recent addition to the risk of bias aspect.Quantitative SynthesisSummary effect measures could not be identified for6 reviews. Prevented fraction was the most common ef-fect measure used (n = 18, 47%) followed by weightedmean difference (n = 7). Other effect measures used in-cluded risk ratio, standardized mean difference, andnumbers needed to treat.Of the 20 (53%) reporting a meta-analysis, 11 reporteda random effects meta-analysis and two reported usingfixed effect analysis. One reported using both fixed andrandom effects models while six did not specify the typeof meta-analysis performed.Assessment of heterogeneity was reported by 17 ofthose reporting a meta-analysis (85%); 2and I2statisticswere the most commonly used tests (n = 12) to assess het-erogeneity. Fourteen reviews (70%) reported additionalanalyses for exploration of heterogeneity between stud-ies. Cochrane reviews pre-specified the aspects to be as-sessed in these additional meta-regression, subgroup orsensitivity analyses, and reported any post-hoc decisionfor further analyses undertaken. These reviews exploredheterogeneity through regression analysis on aspects ofstudy quality, baseline caries level, additional fluorideexposure, F–concentration, frequency of TFT use, andmode and form of TFT use. Another eight reviews usedsensitivity or subgroup analyses for the aspects of studyquality or study design, with five of these published be-fore 2000.Nine reviews (24%) explored publication bias by ob-serving funnel plot asymmetry, and seven of these (allCochrane reviews) used formal tests for its assessment.One review reported carrying out supplemental analysisincluding studies published later, one assumed publica-tion bias to be nonexistent, and two just mentioned pub-lication bias in the text.Only three (8%) reviews addressed economic data. Ofthese, one identified economic evaluations on fluoridemouthrinses and varnishes and quality-assessed/gradedthem, another, assessing fluoride gels, reported numbersneeded to treat as a measure of resource intensiveness,and one reported a cost-effectiveness analysis of mouth-rinse studies.DiscussionThis study is the first to provide a comprehensive as-sessment of the content and reporting quality of system-atic reviews and meta-analyses on TFT for dental caries.We included 38 reviews published from 1984 to 2009.This is a relatively large number of systematic reviews/meta-analyses identified in one particular area of den-tistry. This may be partly explained by the fact that twoof the first meta-analyses in dentistry appeared exactlyon this topic as early as in the mid-1980s [Stamm et al.,1984; Clark et al., 1985], and also by the large amount ofexperimental evidence published over five decades. Inaddition, we were deliberately inclusive in classifying apublication as a systematic review in order to get a broad-er perspective on the development of evidence synthesisin this particular area. The highest number of publica-tions between 2000 and 2004 were associated with thepublication of a series of Cochrane reviews and system-atic reviews from the Swedish Council of Technology As-sessment in Health Care assessing the effectiveness offluoride toothpastes, mouthrinses, gels and varnishes.0Publishedreviews(n)5101520251984–1989 1990–1994 1995–1999 2000–2004 2005–2009Year of publicationFig. 1. Frequency of systematic reviews and meta-analyses on top-ical fluoride by year of publication.
Topical Fluorides Systematic Reviews Caries Res 2010;44:579–592 589This study identified some key issues in the publishedTFT systematic reviews. We found that in terms of con-tent (PICOS), children and adolescents, fluoride tooth-pastes, no treatment/placebo comparisons, and the out-come of caries increment have been addressed repeatedly.Some important population groups (e.g. adults), TFT in-terventions (e.g. paint-on solutions), comparisons (e.g. ofvarious application features), and outcomes (e.g. adverseeffects, economic) have not been addressed sufficientlythough, or have not been covered at all. Some Cochranereviews that could be identified as protocols at the timeof the search, and have now been published, are likely tonarrow some of the existing gaps, such as those related tothe assessment of direct comparisons of toothpaste ap-plication features (fluoride concentration) [Walsh et al.,2010] and of adverse effects (fluorosis) [Wong et al., 2010].We had expected that the dissemination of methodsfor conducting systematic reviews from the mid 1990s[CRD, 1996; Clarke and Oxman, 2002], and the publica-tion of reporting guidelines for systematic reviews since1999 [Moher et al., 1999] would result in an overall im-proved reporting quality in the TFT systematic reviewsthat followed. Higher quality reviews that were publishedpost QUOROM have been reported in other healthcareareas [Delaney et al., 2005; Al Faleh and Al-Omran,2009]. This appears to be the case for TFT systematic re-views as well, where the reviews published after QUO-ROM were of better overall reporting quality than thosepublished before it. There remains room for improve-ment though. Even very recently published reviews didnot adopt the clear structured format (PICOS) advised bythe guidelines. Reporting of search methods was alsofound to be largely inadequate, particularly the completereporting of a search strategy which was rare outside ofCochrane reviews. In addition, searching multiple sourc-es for studies was reported by only half of the reviewseven if pre-QUOROM publications are disregarded.Published reviews sometimes reported searching theCochrane Library without specifying the databasessearched within it. The Cochrane Library includes spe-cific databases for searching various types of studies suchas Cochrane and non-Cochrane systematic reviews, con-trolled trials, methods studies, economic evaluations,and health technology assessments. It should be specifiedwhich of these were used to locate studies. Nevertheless,methods used for ‘selection of studies’ were found to bemore in line with the current reporting guidelines – donein duplicate – and reported in more detail than they werefor data extraction, although both aspects are advised tobe equally explicitly reported.Another important finding was that the quality/risk ofbias assessment in most reviews did not meet current re-porting guidelines, except in Cochrane reviews. For ex-ample, the use of summary scores for risk of bias assess-ment, even though guidelines are explicitly discouragingthis approach now [Higgins and Altman, 2008; Liberatiet al., 2009]. The assessment of risk of bias in includedstudies according to the latest PRISMA recommen-dations is a major development that should be widelyadopted.The use of quantitative synthesis, meta-analysis, wasreported in half of the reviews with reporting of themethods of synthesis variable across them. Although het-erogeneity was also formally assessed in nearly half of thereviews, its presence was usually not adequately exploredby means of subgroup, sensitivity, or meta-regressionanalyses. The exceptions were, again, Cochrane reviews,which predefined and detailed the analyses and the in-vestigations of potential reasons for heterogeneity be-tween studies to be undertaken, and reported any post-hoc decisions on analyses or deviations from protocolwith reasons for doing so. In line with a previous studyassessing systematic reviews in health care [Moher et al.,2007], assessment of publication bias was also a relativelyrare feature in the reviews despite its known importancein affecting reviews’ findings. However, a note of cautionwould be always required in assessing and interpretingheterogeneity and publication bias; there will often be in-sufficient data for these investigations to be done reliablyusing statistical methods, and they would be of question-able value when based on very few studies [Higgins et al.,2002; Sterne et al., 2008]. In addition, explorations of het-erogeneity that are devised after heterogeneity is identi-fied can at best lead to the generation of hypotheses[Deeks, 2008].We found that economic evaluations in TFT system-atic reviews have been uncommon so far. Current meth-odological developments from the Cochrane and theCampbell Collaboration (www.campbellcollaboration.org) support the introduction of economic evaluations insystematic reviews, and these have the potential to en-hance the applicability of new TFT review’s findings,since such evaluations are becoming increasingly impor-tant for decision makers to help identify more cost-effec-tive treatments.Our findings were consistent with those from previousstudies of this nature in that the Cochrane reviews per-formed better than others in reporting quality [Moja etal., 2005; Jørgensen et al., 2006, 2008; Moher et al., 2007;Mrkobrada et al., 2008; Lundh et al., 2009]. Our broad
Ijaz/Croucher/MarinhoCaries Res 2010;44:579–592590inclusion criteria for systematic reviews may be consid-ered a reason for the low review quality observed gener-ally. However, in a recent study of similar nature, restrict-ing inclusion to reviews clearly labelled as systematicreviews did not result in improved overall quality, nordid it explain the differences in quality between Coch-rane and non-Cochrane systematic reviews [Lundh etal., 2009].The suboptimal reporting outside of Cochrane re-views may indicate a low level of awareness regarding ex-isting reporting guidelines for systematic reviews amongauthors and dental journals, which is understandableconsidering that these involve relatively new methodolo-gy. However, although information on the endorsementby journals of reporting guidelines for systematic reviews(QUOROM/PRISMA) is not available (www.prisma-statement.org), the list of around 350 scientific journalsofficially endorsing the reporting guideline for clinicaltrials – CONSORT (CONsolidated Standards Of Report-ing Trials) included only nine dental journals in early 2010(www.consort-statement.org). Recent studies suggest thatendorsement of reporting guidelines by journals can leadto improved overall quality [Moher et al., 2001, 2007; Plintet al., 2006; Mrkobrada et al., 2008; Alvarez et al., 2009].With this regard, it has also been suggested that directingcontinuing education efforts towards journal editors mayimprove the quality of published research faster than oth-er interventions [Sørensen and Rothman, 2010]. Thus, theendorsement of PRISMA by dental journals may be thesimplest way to ensure up-to-the-mark reporting of sys-tematic reviews and meta-analyses in dentistry.As regards potential limitations of our study, it shouldbe noted that we restricted our search to MEDLINE andcertain dedicated databases of systematic reviews. Wemay have missed out on publications not covered bythese and available from other sources. In addition, selec-tion of included studies, and data extraction were doneby one author with a second author consulted at eachstage and consensus achieved. These were not undertak-en independently, which may have allowed some degreeof bias. Finally, it should be pointed out that we essen-tially assessed included reviews for reporting quality withregard to adherence to reporting guidelines. We did notuse a review quality assessment tool such as AMSTAR[Shea et al., 2007, 2009] or OQAQ [Oxman and Guyatt,1991] for their appraisal, although most aspects assessedwould also be covered by these tools.In summary, we have identified some content and re-porting quality issues for consideration in future topicalfluoride reviews.In terms of reporting quality, complete reporting ofobjectives and eligibility criteria (in PICOS format), andof the search strategy, assessment of risk of bias, explora-tion of heterogeneity, and explicit reporting of the reviewprocess – what was and was not done in the review – areimportant issues in current publications that should beaddressed more thoroughly in future reviews. We foundthe methods and reporting in the Cochrane topical fluo-ride reviews to be superior to others in general, and in linewith PRISMA recommendations. Therefore, future sys-tematic reviews on topical fluorides will benefit from fol-lowing the PRISMA statement.In terms of content, these are assessments of interven-tions and of direct comparisons of application featureswithin interventions, population groups, and outcomesnot covered to date or sufficiently. In addition, the relativeeconomic value of topical fluoride therapies should alsobe assessed more widely. The CDSR in The Cochrane Li-brary provides a registry of systematic reviews, makingcomplete and ongoing Cochrane reviews an essential ref-erence resource when undertaking a new systematic re-view, as they provide an indication of content areas al-ready covered so that effort is not duplicated. An initia-tive to develop an international registry for prospectiveregistration of protocols for all systematic reviews is nowunderway [Booth et al., 2010] and will be invaluable inassisting those planning new reviews and updating exist-ing ones.AcknowledgementsThis work was performed independent of any funding. S.I. isa PhD student supervised by V.C.C.M. and R.E.C. at Queen Mary,University of London.Disclosure StatementV.C.C.M. is an editor of the Cochrane Oral Health Group andthe lead author in a series of Cochrane Topical fluoride reviews.S.I. and R.E.C. are review authors with the same group. None ofthe authors have any financial conflict of interests to declare.
Topical Fluorides Systematic Reviews Caries Res 2010;44:579–592 591ReferencesAl Faleh K, Al-Omran M: Reporting and meth-odologic quality of Cochrane Neonatal re-view group systematic reviews. BMC Pediatr2009;9:38.Alvarez F, Meyer N, Gourraud PA, Paul C: Con-sort adoption and quality of reporting ofrandomized controlled trials: a systematicanalysis in two dermatology journals. Br JDermatol 2009;161:1159–1165.Ammari A, Bloch-Zupan A, Ashley PF: System-atic review of studies comparing the anti-caries efficacy of children’s toothpaste con-taining 600 ppm of fluoride or less with highfluoride toothpastes of 1,000 ppm or above.Caries Res 2003;37:85–92.Ammari JB, Baqain ZH, Ashley PF, Ammari JB,Baqain ZH, Ashley PF: Effects of programsfor prevention of early childhood caries. Asystematic review. Med Princ Pract 2007;16:437–442.Antman EM, Lau J, Kupelnick B, Mosteller F,Chalmers TC: A comparison of results ofmeta-analyses of randomized control trialsand recommendations of clinical experts:treatments for myocardial infarction. JAMA1992;268:240–248.Axelsson S, Söder B, Nordenram G, PeterssonLG, Dahlgren H, Norlund A, Källestål C,Mejàre I, Lingström P, Lagerlöf F, Holm AK,Twetman S: Effect of combined caries-pre-ventive methods: a systematic review of con-trolled clinical trials. Acta Odontol Scand2004;62:163–169.Azarpazhooh A, Main PA, Azarpazhooh A,Main PA: Fluoride varnish in the preventionof dental caries in children and adolescents:a systematic review. J Can Dent Assoc 2008;74:73–79.Bader J, Ismail A: Survey of systematic reviewsin dentistry. J Am Dent Assoc 2004;135:464–473.Bader JD, Rozier RG, Lohr KN, Frame PS: Physi-cians’ roles in preventing dental caries inpreschool children: a summary of the evi-dence for the U.S. Preventive Services TaskForce. Am J Prev Med 2004;26:315–325.Bader JD, Shugars DA, Bonito AJ: A systematicreview of selected caries prevention andmanagement methods. Community DentOral Epidemiol 2001;29:399–411.Bartizek RD, Gerlach RW, Faller RV, Jacobs SA,Bollmer BW, Biesbrock AR: Reduction indental caries with four concentrations of so-dium fluoride in a dentifrice: a meta-analysisevaluation. J Clin Dent 2001;12:57–62.Beiswanger BB, Stookey GK: The comparativeclinical cariostatic efficacy of sodium fluo-ride and sodium monofluorophosphate den-tifrices: a review of trials. J Dent Child 1989;56:337–347.Benson PE, Parkin N, Millett DT, Dyer F, Vine S,Shah A: Fluorides for the prevention of whitespots on teeth during fixed brace treat-ment. Cochrane Database Syst Rev 2004;3:CD003809.Booth A, Clarke M, Ghersi D, Moher D, Petti-crew M, Stewart L: An international registryof systematic-review protocols. Lancet 2010,E-pub ahead of print.CRD: CRD Report 4. Undertaking systematicreviews of research on effectiveness: CRDguidelines for those carrying out or commis-sioning reviews. York, Centre for Reviewsand Dissemination, University of York, 1996.Chadwick BL, Roy J, Knox J, Treasure ET: Theeffect of topical fluorides on decalcificationin patients with fixed orthodontic appli-ances: a systematic review. Am J OrthodDentofacial Orthop 2005;128:601–606.Chaves SC, Vieira-da-Silva LM: Anticaries effec-tiveness of fluoride toothpaste: a meta-anal-ysis (in Portuguese). Rev Saude Publica 2002;36:598–606.Clark DC, Hanley JA, Stamm JW, Weinstein PL:An empirically based system to estimate theeffectiveness of caries-preventive agents: acomparison of the effectiveness estimates ofAPF gels and solutions, and fluoride var-nishes. Caries Res 1985;19:83–95.Clarke M, Oxman A: Cochrane reviewers hand-book 4.1.5 (updated April 2002); in TheCochrane Library issue 2. Oxford, UpdateSoftware 2002.Deeks JJ: Chapter 9: Analysing data and under-taking meta-analyses; in Higgins JPT, GreenS (eds): Cochrane Handbook for SystematicReviews of Interventions version 501 (updat-ed September 2008). The Cochrane Collabo-ration, 2008.Delaney A, Bagshaw S, Ferland A, Manns B,Laupland K, Doig C: A systematic evaluationof the quality of meta-analyses in the criticalcare literature. Crit Care 2005;9:R575–R582.Derks A, Katsaros C, Frencken JE, van’t Hof MA,Kuijpers-Jagtman AM: Caries-inhibitingeffect of preventive measures during or-thodontic treatment with fixed appliances: asystematic review. Caries Res 2004;38:413–420.Glenny A, Esposito M, Coulthard P, Worthing-ton HV: The assessment of systematic re-views in dentistry. Eur J Oral Sci 2003;111:85–92.Griffin SO, Regnier E, Griffin PM, Huntley V:Effectiveness of fluoride in preventing cariesin adults. J Dent Res 2007;86:410–415.Heijnsbroek M, Paraskevas S, Van der WeijdenGA: Fluoride interventions for root caries: areview. Oral Health Prev Dent 2007;5:145–152.Helfenstein U, Steiner M: Fluoride varnishes(Duraphat): a meta-analysis. CommunityDent Oral Epidemiol 1994;22:1–5.Higgins J, Altman D: Chapter 8: Assessing riskof bias in included studies; in Higgins JPT,Green S (eds): Cochrane Handbook for Sys-tematic Reviews of Interventions version 501(updated September 2008). The CochraneCollaboration, 2008.Higgins JPT, Green S (eds): Cochrane Handbookfor Systematic Reviews of Interventions ver-sion 501 (updated September 2008). TheCochrane Collaboration, 2008.Higgins J, Thompson S, Deeks J, Altman D: Sta-tistical heterogeneity in systematic reviewsof clinical trials: a critical appraisal of guide-lines and practice. J Health Serv Res Policy2002;7:51–61.Hiiri A, Ahovuo-Saloranta A, Nordblad A, Mä-kelä M: Pit and fissure sealants versus fluo-ride varnishes for preventing dental decay inchildren and adolescents. Cochrane Data-base Syst Rev 2006;4:CD003067.Johnson MF: Comparative efficacy of NaF andSMFP dentifrices in caries prevention: ameta-analytic overview. Caries Res 1993;27:328–336.Jørgensen A, Hilden J, Gøtzsche PG: Cochranereviews compared with industry supportedmeta-analysis and other meta-analysis of thesame drugs: systematic review. BMJ 2006;333:782–785.Jørgensen A, Maric K, Tendal B, Faurschou A,Gøtzsche P: Industry-supported meta-anal-ysescomparedwithmeta-analyseswithnon-profit or no support: differences in method-ological quality and conclusions. BMC MedRes Methodol 2008;8:60.Källestål C, Norlund A, Söder B, Nordenram G,Dahlgren H, Petersson LG, Lagerlöf F, Axels-son S, Lingström P, Mejàre I, Holm AK,Twetman S: Economic evaluation of dentalcaries prevention: a systematic review. ActaOdontol Scand 2003;61:341–346.Kay EJ, Locker D: Is dental health education ef-fective? A systematic review of current evi-dence. Community Dent Oral Epidemiol1996;24:231–235.Liberati A, Altman DG, Tetzlaff J, Mulrow C,Gøtzsche PC, Ioannidis JPA, Clarke M, De-vereaux PJ, Kleijnen J, Moher D: The PRIS-MA statement for reporting systematic re-views and meta-analyses of studies thatevaluate healthcare interventions: explana-tion and elaboration. BMJ 2009;339:b2700.Lundh A, Knijnenburg SL, Jørgensen AW, vanDalen EC, Kremer LC: Quality of systematicreviews in pediatric oncology: a systematicreview. Cancer Treat Rev 2009;35:645–652.Major MP, Major PW, Flores-Mir C: An evalua-tion of search and selection methods used indental systematic reviews published in En-glish. J Am Dent Assoc 2006;137:1252–1257.Marinho VC: Systematic reviews of controlledtrials in general and oral health care. Braz JOral Sci 2003;2:215–226.Marinho VC, Higgins JP, Logan S, Sheiham A:Fluoride varnishes for preventing dental car-ies in children and adolescents. CochraneDatabase Syst Rev 2002a;3:CD002279.
Ijaz/Croucher/MarinhoCaries Res 2010;44:579–592592Marinho VC, Higgins JP, Logan S, Sheiham A:Topical fluoride (toothpastes, mouthrinses,gels or varnishes) for preventing dental car-ies in children and adolescents. CochraneDatabase Syst Rev 2003a;4:CD002782.Marinho VC, Higgins JP, Logan S, Sheiham A:Fluoride mouthrinses for preventing dentalcaries in children and adolescents. CochraneDatabase Syst Rev 2003c;3:CD002284.Marinho VC, Higgins JP, Logan S, Sheiham A:Combinations of topical fluoride (tooth-pastes, mouthrinses, gels, varnishes) versussingle topical fluoride for preventing dentalcaries in children and adolescents. CochraneDatabase Syst Rev 2004a;1:CD002781.Marinho VC, Higgins JP, Logan S, Sheiham A:One topical fluoride (toothpastes, or mouth-rinses, or gels, or varnishes) versus anotherfor preventing dental caries in children andadolescents. Cochrane Database Syst Rev2004b;1:CD002780.Marinho VC, Higgins JP, Logan S, Sheiham A:Fluoride gels for preventing dental caries inchildren and adolescents. Cochrane Data-base Syst Rev 2002b;2:CD002280.Marinho VC, Higgins JP, Sheiham A, Logan S:Fluoride toothpastes for preventing dentalcaries in children and adolescents. CochraneDatabase Syst Rev 2003b;1:CD002278.Moher D, Cook DJ, Eastwood S, Olkin I, RennieD, Stroup DF: Improving the quality of re-ports of meta-analyses of randomised con-trolled trials: the QUOROM statement. Lan-cet 1999;354:1896–1900.Moher D, Jones A, Lepage L, for the CONSORTGroup: Use of the consort statement andquality of reports of randomized trials: acomparative before-and-after evaluation.JAMA 2001;285:1992–1995.Moher D, Liberati A, Tetzlaff J, Altman DG, forthe PG: Preferred reporting items for sys-tematic reviews and meta-analyses: ThePRISMA statement. BMJ 2009;339:b2535.Moher D, Tetzlaff J, Tricco AC, Sampson M, Alt-man DG: Epidemiology and reporting char-acteristics of systematic reviews. PLoS Med2007;4:e78.Moja LP, Telaro E, D’Amico R, Moschetti I, CoeL, Liberati A, on behalf of the MetaqualityStudy G: Assessment of methodologicalquality of primary studies by systematic re-views: results of the metaquality cross sec-tional study. BMJ 2005;330:1053.Mrkobrada M, Thiessen-Philbrook H, HaynesRB, Iansavichus AV, Rehman F, Garg AX:Need for quality improvement in renal sys-tematic reviews. Clin J Am Soc Nephrol2008;3:1102–1114.Oxman A, Guyatt G: Validation of an index ofthe quality of review articles. J Clin Epide-miol 1991;44:1271–1278.Oxman A, Guyatt G: The science of reviewingresearch. Ann N Y Acad Sci 1993;703:125–133.Petersson LG, Twetman S, Dahlgren H, NorlundA, Holm AK, Nordenram G, Lagerlöf F,Söder B, Källestål C, Mejàre I, Axelsson S,Lingström P: Professional fluoride varnishtreatment for caries control: a systematic re-view of clinical trials. Acta Odontol Scand2004;62:170–176.Plint AC, Moher D, Morrison A, Schulz K, Alt-manDG,Hill C,GabouryI:Doestheconsortchecklist improve the quality of reports ofrandomised controlled trials? A systematicreview. Med J Aust 2006;185:263–267.Richards D: 10 years after. Evid Based Dent 2004;5:87.Rosenblatt A, Stamford TC, Niederman R: Silverdiamine fluoride: a caries ‘silver-fluoridebullet’. J Dent Res 2009;88:116–125.Rozier RG: Effectiveness of methods used bydental professionals for the primary preven-tion of dental caries. J Dent Educ 2001;65:1063–1072.Shea B, Grimshaw J, Wells G, Boers M, Anders-son N, Hamel C, Porter A, Tugwell P, MoherD, Bouter L: Development of AMSTAR: ameasurement tool to assess the methodolog-ical quality of systematic reviews. BMC MedRes Methodol 2007;7:10.Shea BJ, Hamel C, Wells GA, Bouter LM, Krist-jansson E, Grimshaw J, Henry DA, Boers M:AMSTAR is a reliable and valid measure-ment tool to assess the methodological qual-ity of systematic reviews. J Clin Epidemiol2009;62:1013–1020.Sørensen HT, Rothman KJ: The prognosis for re-search. BMJ 2010;340:c703.Stamm JW, Bohannan HM, Graves RC, DisneyJA: The efficiency of caries prevention withweekly fluoride mouthrinses. J Dent Educ1984;48:617–626.Steiner M, Helfenstein U, Menghini G: Effect of1000 ppm relative to 250 ppm fluoride tooth-paste: a meta-analysis. Am J Dent 2004;17:85–88.Sterne JAC, Egger M, Moher D: Chapter 10: Ad-dressing reporting biases; in Higgins JPT,Green S (eds): Cochrane Handbook for Sys-tematic Reviews of Interventions, version501 (updated September 2008). The Coch-rane Collaboration, 2008.Stookey GK, DePaola PF, Featherstone JD, Fejer-skov O, Möller IJ, Rotberg S, Stephen KW,Wefel JS: A critical review of the relativeanticaries efficacy of sodium fluoride andsodium monofluorophosphate dentifrices.Caries Res 1993;27:337–360.Strohmenger L, Brambilla E: The use of fluoridevarnishes in the prevention of dental caries:a short review. Oral Dis 2001;7:71–80.Twetman S: Prevention of early childhood caries(ECC): review of literature published 1998–2007. Eur Arch Paediatr Dent 2008;9:12–18.Twetman S, Axelsson S, Dahlgren H, Holm AK,Källestål C, Lagerlöf F, Lingström P, MejàreI, Nordenram G, Norlund A, Petersson LG,Söder B: Caries-preventive effect of fluoridetoothpaste: a systematic review. Acta Odon-tol Scand 2003;61:347–355.Twetman S, Petersson L, Axelsson S, DahlgrenH, Holm AK, Källestål C, Lagerlöf F, Ling-ström P, Mejàre I, Nordenram G, Norlund A,Söder B: Caries-preventive effect of sodiumfluoride mouthrinses: a systematic reviewof controlled clinical trials. Acta OdontolScand 2004;62:223–230.Van Rijkom HM, Truin GJ, van’t Hof MA: Ameta-analysis of clinical studies on the car-ies-inhibiting effect of fluoride gel treat-ment. Caries Res 1998;32:83–92.Volpe A, Petrone M, Davies R, Proskin H: Clini-cal anticaries efficacy of NaF and SMFP den-tifrices: overview and resolution of the scien-tific controversy. J Clin Dent 1995;6(specNo):1–28.Walsh T, Worthington H, Glenny A, Appelbe P,Marinho V, Shi X: Fluoride toothpastes ofdifferent concentrations for preventing den-tal caries in children and adolescents. Coch-rane Database Syst Rev 2010;1:CD007868.Wong MCM, Glenny AM, Tsang BWK, Lo ECM,Worthington HV, Marinho VCC: Topicalfluoride as a cause of dental fluorosis in chil-dren. Cochrane Database Syst Rev 2010;1:CD007693.