Fluoride toothpastes of different concentrations for preventing dental caries

Tanya Walsh; Helen V Worthington; Anne‐Marie Glenny; Valeria CC Marinho; Ana Jeroncic

doi:10.1002/14651858.CD007868.pub3

. 2019 Mar 4;2019(3):CD007868. doi: 10.1002/14651858.CD007868.pub3

Fluoride toothpastes of different concentrations for preventing dental caries

Tanya Walsh ^1,^✉, Helen V Worthington ², Anne‐Marie Glenny ¹, Valeria CC Marinho ³, Ana Jeroncic ⁴

Editor: Cochrane Oral Health Group

PMCID: PMC6398117 PMID: 30829399

Abstract

Background

Caries (dental decay) is a disease of the hard tissues of the teeth caused by an imbalance, over time, in the interactions between cariogenic bacteria in dental plaque and fermentable carbohydrates (mainly sugars). Regular toothbrushing with fluoride toothpaste is the principal non‐professional intervention to prevent caries, but the caries‐preventive effect varies according to different concentrations of fluoride in toothpaste, with higher concentrations associated with increased caries control. Toothpastes with higher fluoride concentration increases the risk of fluorosis (enamel defects) in developing teeth. This is an update of the Cochrane Review first published in 2010.

Objectives

To determine and compare the effects of toothpastes of different fluoride concentrations (parts per million (ppm)) in preventing dental caries in children, adolescents, and adults.

Search methods

Cochrane Oral Health's Information Specialist searched the following databases: Cochrane Oral Health's Trials Register (to 15 August 2018); the Cochrane Central Register of Controlled Trials (CENTRAL; 2018, Issue 7) in the Cochrane Library (searched 15 August 2018); MEDLINE Ovid (1946 to 15 August 2018); and Embase Ovid (1980 to 15 August 2018). The US National Institutes of Health Ongoing Trials Register (ClinicalTrials.gov) and the World Health Organization International Clinical Trials Registry Platform were searched for ongoing trials (15 August 2018). No restrictions were placed on the language or date of publication when searching the electronic databases.

Selection criteria

Randomised controlled trials that compared toothbrushing with fluoride toothpaste with toothbrushing with a non‐fluoride toothpaste or toothpaste of a different fluoride concentration, with a follow‐up period of at least 1 year. The primary outcome was caries increment measured by the change from baseline in the decayed, (missing), and filled surfaces or teeth index in all permanent or primary teeth (D(M)FS/T or d(m)fs/t).

Data collection and analysis

Two members of the review team, independently and in duplicate, undertook the selection of studies, data extraction, and risk of bias assessment. We graded the certainty of the evidence through discussion and consensus. The primary effect measure was the mean difference (MD) or standardised mean difference (SMD) caries increment. Where it was appropriate to pool data, we used random‐effects pairwise or network meta‐analysis.

Main results

We included 96 studies published between 1955 and 2014 in this updated review. Seven studies with 11,356 randomised participants (7047 evaluated) reported the effects of fluoride toothpaste up to 1500 ppm on the primary dentition; one study with 2500 randomised participants (2008 evaluated) reported the effects of 1450 ppm fluoride toothpaste on the primary and permanent dentition; 85 studies with 48,804 randomised participants (40,066 evaluated) reported the effects of toothpaste up to 2400 ppm on the immature permanent dentition; and three studies with 2675 randomised participants (2162 evaluated) reported the effects of up to 1100 ppm fluoride toothpaste on the mature permanent dentition. Follow‐up in most studies was 36 months.

In the primary dentition of young children, 1500 ppm fluoride toothpaste reduces caries increment when compared with non‐fluoride toothpaste (MD ‐1.86 dfs, 95% confidence interval (CI) ‐2.51 to ‐1.21; 998 participants, one study, moderate‐certainty evidence); the caries‐preventive effects for the head‐to‐head comparison of 1055 ppm versus 550 ppm fluoride toothpaste are similar (MD ‐0.05, dmfs, 95% CI ‐0.38 to 0.28; 1958 participants, two studies, moderate‐certainty evidence), but toothbrushing with 1450 ppm fluoride toothpaste slightly reduces decayed, missing, filled teeth (dmft) increment when compared with 440 ppm fluoride toothpaste (MD ‐0.34, dmft, 95%CI ‐0.59 to ‐0.09; 2362 participants, one study, moderate‐certainty evidence). The certainty of the remaining evidence for this comparison was judged to be low.

We included 81 studies in the network meta‐analysis of D(M)FS increment in the permanent dentition of children and adolescents. The network included 21 different comparisons of seven fluoride concentrations. The certainty of the evidence was judged to be low with the following exceptions: there was high‐ and moderate‐certainty evidence that 1000 to 1250 ppm or 1450 to 1500 ppm fluoride toothpaste reduces caries increments when compared with non‐fluoride toothpaste (SMD ‐0.28, 95% CI ‐0.32 to ‐0.25, 55 studies; and SMD ‐0.36, 95% CI ‐0.43 to ‐0.29, four studies); there was moderate‐certainty evidence that 1450 to 1500 ppm fluoride toothpaste slightly reduces caries increments when compared to 1000 to 1250 ppm (SMD ‐0.08, 95% CI ‐0.14 to ‐0.01, 10 studies); and moderate‐certainty evidence that the caries increments are similar for 1700 to 2200 ppm and 2400 to 2800 ppm fluoride toothpaste when compared to 1450 to 1500 ppm (SMD 0.04, 95% CI ‐0.07 to 0.15, indirect evidence only; SMD ‐0.05, 95% CI ‐0.14 to 0.05, two studies).

In the adult permanent dentition, 1000 or 1100 ppm fluoride toothpaste reduces DMFS increment when compared with non‐fluoride toothpaste in adults of all ages (MD ‐0.53, 95% CI ‐1.02 to ‐0.04; 2162 participants, three studies, moderate‐certainty evidence). The evidence for DMFT was low certainty.

Only a minority of studies assessed adverse effects of toothpaste. When reported, effects such as soft tissue damage and tooth staining were minimal.

Authors' conclusions

This Cochrane Review supports the benefits of using fluoride toothpaste in preventing caries when compared to non‐fluoride toothpaste. Evidence for the effects of different fluoride concentrations is more limited, but a dose‐response effect was observed for D(M)FS in children and adolescents. For many comparisons of different concentrations the caries‐preventive effects and our confidence in these effect estimates are uncertain and could be challenged by further research. The choice of fluoride toothpaste concentration for young children should be balanced against the risk of fluorosis.

Plain language summary

Fluoride toothpastes of different strengths for preventing tooth decay

Review question  This review has been produced to assess the effects of toothpastes of different fluoride strengths on preventing tooth decay in children, adolescents and adults.

Background  Tooth decay (caries) is a widespread disease, affecting billions of people worldwide. Fluoride has long been used to prevent decay, through a variety of different methods including toothpaste, water, milk, mouthrinses, tooth gels and varnish. Regular toothbrushing is recommended to prevent decay and other oral diseases, and toothbrushing for 2 minutes twice daily with a fluoride toothpaste is generally recommended. The typical strength of regular or family toothpaste is around 1000 to 1500 parts per million (ppm) fluoride, but many other strengths are available worldwide. There is no minimum fluoride concentration, but the maximum permissible fluoride concentration for a toothpaste varies according to age and country. Higher concentrations are rarely available over the counter, and are classed as a prescription‐only medicine. Stronger fluoride toothpaste may offer greater protection against decay but also increases the risk of fluorosis (enamel defects) in developing teeth. This is an update of the Cochrane Review first published in 2010.

Study characteristics  Authors from Cochrane Oral Health carried out this review and the evidence is current up to 15 August 2018. It includes 96 studies published between 1955 and 2014: seven studies with 11,356 randomised participants reported the effects of fluoride toothpaste up to 1500 ppm on the primary teeth; one study with 2500 randomised participants reported the effects of 1450 ppm toothpaste on the primary and permanent dentition; 85 studies with 48,804 randomised participants reported the effects of toothpaste up to 2400 ppm on the permanent teeth of children up to 18 years of age; and three studies with 2675 randomised participants reported the effects of up to 1100 ppm toothpaste on the permanent teeth of adults. Most studies assessed decay after participants had been using the toothpastes for 36 months.

Main results  We present below findings for which there is moderate‐ or high‐certainty evidence.

In primary teeth of young children, brushing teeth with a toothpaste containing 1500 ppm fluoride reduced the amount of new decay when compared with non‐fluoride toothpaste; the amount of new decay was similar with 1055 ppm compared with 550 ppm fluoride toothpaste; and there was a slight reduction in the amount of new decay with 1450 ppm toothpaste compared with 440 ppm fluoride toothpaste.

Eighty‐one studies assessed the effects of different strengths of fluoride toothpaste compared against each other (seven different strengths in 21 combinations) in permanent teeth of children and adolescents. We found that there was less new decay when toothbrushing with toothpaste containing 1000 to 1250 ppm or 1450 to 1500 ppm fluoride compared with non‐fluoride toothpaste, and that toothbrushing with 1450 to 1500 ppm fluoride toothpaste reduced the amount of new decay more than 1000 to 1250 ppm toothpaste. We found that there was a similar amount of new decay when children and adolescents used a toothpaste of 1700 to 2200 ppm or 2400 to 2800 ppm fluoride compared to 1450 to 1500 ppm toothpaste. The evidence for the effects of other strengths of toothpaste was less certain.

In permanent teeth of adults of all ages, 1000 or 1100 ppm toothpaste reduced decay compared with non‐fluoride toothpaste.

Most studies did not measure harmful effects of toothpaste use, but when reported, effects such as soft tissue damage and tooth staining were minimal.

Certainty of the evidence  There is high‐certainty evidence that toothpaste containing 1000 to 1250 ppm fluoride is more effective than non‐fluoride toothpaste. There is moderate‐certainty evidence for the other findings reported in 'Main results' above. For other toothpaste strengths compared against each other or against non‐fluoride toothpaste, there are too few studies with too few participants to have any clarity about the effects.

Authors conclusions  There are benefits of using fluoride toothpaste at certain strengths to prevent tooth decay when compared with non‐fluoride toothpaste. The stronger the fluoride concentration, the more decay is prevented. For many of the comparisons of different strengths of toothpaste, the findings are uncertain and could be challenged by further research. The choice of fluoride toothpaste for young children should be balanced against the risk of fluorosis.

Summary of findings

Background

Description of the condition

Tooth mineral is lost and gained in a continuous process of de‐ and re‐mineralisation. Caries (dental decay) is a disease of the hard tissues of the teeth caused by an imbalance in this process over time, where there is demineralisation of tooth structure by organic acids formed from the interactions between cariogenic bacteria in dental plaque and fermentable carbohydrates (mainly sugars). The dental caries process is influenced by the susceptibility of the tooth surface, the bacterial profile, the quantity and quality of saliva and the presence of fluoride, which promotes remineralisation and inhibits demineralisation of the tooth structure. Aside from the pain arising from the dental carious lesions themselves, there is also the emotional distress of the disease and the potential consequences of medical intervention. Affected teeth cannot always be saved and may have to be extracted. This has particular consequences for young children, for whom general anaesthesia may be required. There is an associated social impact of this disease in terms of absence from school for the children and absence from work for their carers. There are also important financial implications for this disease with a substantial proportion of healthcare budgets being spent every year on treating caries. On a population basis, traditional treatment of oral disease (dental caries and periodontal diseases) is the fourth most expensive chronic disease to treat according to the World Health Organization (WHO) (Petersen 2009). Further, in low‐income countries, if treatment were available, the costs of treating dental caries in children alone would exceed the total healthcare budget for children (WHO 2019).

Caries in permanent teeth was the most prevalent condition among all those evaluated in the Global Burden of Disease 2016 study, affecting 2.4 billion people; the estimated prevalence of caries in deciduous teeth was 486 million children worldwide (GBD 2016). Whilst in some areas of middle‐income and high‐income countries, there has been evidence of a reduction in the prevalence and severity of dental caries in recent decades, social inequalities in dental health exist, with many individuals and communities having a clinically significant burden of preventable dental disease. Levels of dental decay vary considerably between and within countries, but children in lower socio‐economic status (SES) groups have higher caries levels than those in upper SES groups, and in high‐income countries, the association between socio‐economic position and caries might be stronger (Chen 1995; Reisine 2001; Schwendicke 2015).

Description of the intervention

The link between fluoride and oral health dates back to the 1930s, and today fluoride remains one of dentistry's key strategies to prevent dental caries. There are many methods of fluoride delivery e.g. toothpaste, fluoridated water, milk, mouthrinses, gels, varnish, etc. Evidence regarding the effectiveness of topical fluoride interventions has been synthesized in a series of Cochrane Reviews (Marinho 2003; Marinho 2004; Marinho 2013; Marinho 2015; Marinho 2016; Marinho a 2004; Marinho b 2003).

Toothbrushing is the process used to mechanically remove and control the dental biofilm to help prevent caries and oral diseases. Recommendations on toothbrushing and the use of fluoride toothpaste including choice of concentration, the frequency of toothbrushing and amount of toothpaste to be, can vary. Toothbrushing is usually carried out using a manual or powered toothbrush and a fluoride toothpaste, for 2 minutes twice daily. Toothbrushing with fluoride toothpaste is by far the most common form of caries control, and fluoride toothpaste use is commonly linked to the decline in caries prevalence in many countries. There is an argument that the effects of fluoride toothpaste are underestimated in 'short‐term' clinical trials of 2 to 3 years duration. It is reasonable to assume that a greater cumulative effect is conferred over time as fluoride toothpastes are used throughout life. In addition, some argue that the use of fluoride toothpaste in areas with community water fluoridation offers more protection than either alone. However, concern has been expressed that dental fluorosis, enamel defects caused by young children chronically ingesting excessive amounts of fluoride during the period of tooth formation (up to the age of 6 years), is increasing in both fluoridated and non‐fluoridated communities, and the early use of fluoride toothpaste by young children may be an important risk factor (Ellwood 1995; Horowitz 1992;Stookey 1994).

In higher‐income countries since the 1980s, nearly all commercially available toothpaste formulations contain fluoride. The formulation and fluoride concentration of toothpaste is diverse, with a variety of fluoride compounds used singly and in combination including sodium fluoride, sodium monofluorophosphate, amine fluoride and stannous fluoride, and, according to each manufacturer's specifications, these must be compatible with other basic ingredients, especially abrasive systems (which account for almost half of the entire toothpaste formulation). Fluoride toothpaste must be differentiated from fluoride prophylactic pastes, since their fluoride concentrations, methods and frequencies of application differ, as well as amounts of abrasives in their formulation (abrasives account for almost the entire content of a prophylactic paste). In addition, although some kinds of toothpastes are available in the translucent form of a gel, they differ from fluoride gels, which have higher fluoride levels, no abrasives and are applied much less frequently, usually by a professional. The usual concentration of fluoride in toothpaste ranges from 1000 to 1500 parts per million (ppm); toothpaste with higher and lower than conventional fluoride levels are available in many countries. There is no restriction on the minimum fluoride concentration, but the maximum permissible fluoride concentration for a toothpaste varies according to location and age. Higher concentrations are rarely available over the counter, classed as a prescription‐only medicine. In the UK, 2800 ppm sodium fluoride toothpaste can be prescribed to high caries‐risk patients aged 10 years and over, those with caries present, orthodontic appliances, or who have a highly cariogenic diet or medication; 5000 ppm toothpaste can be prescribed to high caries‐risk patients aged 16 years and over, with present or potential for root caries, dry mouth, orthodontic appliances, overdentures, or those with highly cariogenic diet or medication (Public Health England 2017).

Toothpaste containing higher concentrations of fluoride may confer greater protection against caries but increase the risk of fluorosis (enamel defects caused by chronic ingestion of excessive amounts of fluoride during the period of tooth formation). Chronic ingestion of fluoride from toothpaste in young children is common (Bentley 1999; Rojas‐Sanchez 1999) and despite the large variation in the amount swallowed, the younger children are, the more likely they are to swallow larger amounts, which often represent a substantial part of the total daily fluoride intake and can be enough to cause fluorosis (Levy 1994; Lewis 1996). Although the precise amount of fluoride (F) ingested beyond which fluorosis may occur is not known, a threshold of 0.05 mg F/kg to 0.07 mg F/kg body weight has been suggested (Burt 1992). A child‐sized toothbrush covered with a full strip of toothpaste holds approximately 0.75 g to 1.0 g of toothpaste, and each gram of fluoride toothpaste, contains approximately 1.0 mg of fluoride; children aged less than 6 years may swallow an estimated 0.3 g of toothpaste per brushing (0.3 mg of fluoride) and can inadvertently swallow as much as 0.8 g (Levy 1994). As a result, it is generally recommended that children of 6 years of age and under should be supervised when brushing their teeth with fluoride toothpaste and only a 'pea‐sized' amount of toothpaste, is used. The frequency of toothpaste use along with the method of rinsing after toothbrushing are other factors influencing the effectiveness and safety of fluoride toothpaste. Brushing twice a day or more, or rinsing less thoroughly or not rinsing at all would confer greater caries reductions than brushing once a day or less, or rinsing with larger volumes of water after toothbrushing (Ashley 1999; Chesters 1992; Chestnutt 1998; O'Mullane 1997). Although acute toxicity is rare, young children are particularly at risk of ingesting toxic doses of fluoride from a standard toothpaste tube of 125 g, containing 1100 ppm F (1.1 mg F/g paste). As the probable toxic dose is around 5 mg F/kg body weight (Whitford 1992), the accidental swallowing of one‐ or two‐thirds (45 g to 90 g) of a toothpaste tube is potentially life‐threatening for a 1‐year‐old (10 kg) or for a 5 to 6‐year‐old (20 kg) respectively (Ellwood 1998). For this reason, it is recommended that a fluoride toothpaste should be kept out of the reach of young children.

How the intervention might work

The most important anti‐caries effect of fluoride results from its local action on the tooth/plaque interface, through the promotion of remineralisation of early caries lesions and reduction in tooth enamel solubility (Featherstone 1988). The presence of fluoride at the time of the acid attack markedly reduces enamel demineralisation (mineral loss), and fluoride enhances mineral gain and provides a more resistant enamel structure (Ten Cate 1999). This occurs with all forms and concentrations of fluoride although to a variable extent. With high‐concentration topical fluoride vehicles such as varnishes and gels, calcium fluoride is precipitated on the enamel surface and in the plaque. This calcium fluoride acts as a fluoride reservoir, which is released when the oral pH falls. The amount of fluoride deposited in the subsurface lesion is greater after topical application with high‐concentration fluoride vehicles (Horowitz 1996; Ogaard 1994; Ogaard 2001). Regular use of fluoride toothpaste or mouthrinse (topical fluoride vehicles of relatively low concentration) results in sustained elevated fluoride concentrations in oral fluids during the demineralisation‐remineralisation cycle, as small amounts are maintained constantly in the mouth (Clarkson 1996).

Why it is important to do this review

Cochrane Oral Health undertook an extensive prioritisation exercise in 2014 to identify a core portfolio of titles that were the most clinically important ones to maintain on the Cochrane Library (Worthington 2015). Initially published in 2010 (Walsh 2010), the paediatric dentistry expert panel identified this Cochrane Review as a priority title (Cochrane Oral Health priority review portfolio). For this update, the scope of the review has been broadened to include adults.

WHO state that poor oral hygiene and inadequate exposure to fluoride have negative effects on oral health (WHO 2018). Effective use of fluoride toothpaste in children and adolescents is a long‐held recommended strategy to prevent caries (Lancet 2009), with WHO guidance that twice‐daily toothbrushing with fluoride‐containing toothpaste (1000 to 1500 ppm) should be encouraged (O'Mullane 2016). Despite these recommendations, cost may prohibit the widespread use of toothbrushing with fluoride toothpaste in many low‐ and middle‐income countries.

Many systematic reviews have evaluated head‐to‐head (direct) comparisons of different fluoride. This Cochrane Review is the first to systematically evaluate a dose‐response effect through an assessment of the overall body of evidence, including both direct and indirect evidence (evidence drawn from trials comparing each of the interventions with a common comparator) from randomised controlled trials (RCTs). In addition, the certainty of this evidence is rated for the first time using the standard GRADE approach (GRADE 2004). A companion Cochrane Review has evaluated the effects of topical fluoride, including toothpaste of different concentrations, on dental fluorosis (Wong 2010). These Cochrane Reviews should be considered together to be fully informed of the potential caries‐preventive benefits of fluoride toothpastes of different concentrations and the potential risks of fluorosis.

The primary aim of this review update, therefore, is to provide a clear and robust summary of the current research evidence on the effects of toothpaste with higher fluoride concentrations, compared to lower or no concentration of fluoride to prevent dental caries in children, adolescents and adults.

Objectives

To determine and compare the effects of toothpastes of different fluoride concentrations (parts per million (ppm)) in preventing dental caries in children, adolescents, and adults.

Methods

Criteria for considering studies for this review

Types of studies

Parallel‐group randomised controlled trials (RCTs) (individually or cluster‐randomised), which compared toothbrushing with fluoride toothpaste with toothbrushing with a non‐fluoride toothpaste or toothpaste of a different fluoride concentration, with a follow‐up period of at least 1 year.

We excluded studies where random allocation was not used or indicated. Due to the high possibility of contamination of one part of the mouth from another, it would not be appropriate to use a split‐mouth design to evaluate the effects of this intervention, and so we excluded any studies with this design.

Types of participants

Children, adolescents or adults, irrespective of the initial level of dental caries, background exposure to fluoride, receipt of dental treatment, nationality, setting where the intervention was received or age at recruitment to the trial.

We excluded studies where the participants were selected on the basis of special (general or oral) health conditions.

Types of interventions

We included studies comparing toothbrushing with a fluoride toothpaste with toothbrushing with another fluoride toothpaste of a different concentration or with a non‐fluoride toothpaste or no toothpaste. On the basis of fluoride concentrations of toothpastes in regular use, we proposed the following categories:

0 (parts per million (ppm)) fluoride (F) (non‐fluoride or placebo toothpaste)
250 ppm F
440 to 550 ppm F
1000 to 1250 ppm F
1450 to 1500 ppm F
1700 to 2200 ppm F
2400 to 2800 ppm F.

There were no restrictions placed on the fluoride agents which could be used singly or in combination:

sodium fluoride (NaF)
sodium monofluorophosphate (SMFP)
stannous fluoride (SnF₂)
acidulated phosphate fluoride (APF)
amine fluoride (AmF).

Toothpastes could be formulated with any compatible abrasive system. These included dicalcium phosphate, sodium metaphosphate, calcium carbonate, silica, zirconium silicate, or calcium pyrophosphate.

There was no restriction on fluoride concentration (ppm), amount or duration of application, frequency of use, toothbrushing technique (including supervised toothbrushing), or post‐toothbrushing procedure.

We excluded studies where the intervention group or both the intervention and control groups received any additional active agent or caries preventive measure (e.g. chlorhexidine agent, fluoride varnish application, fluoride mouthrinse) as part of the study in addition to the fluoride or placebo toothpaste.

Studies where the intervention group alone received any additional potentially active agent in the toothpaste such as xylitol, triclosan, N‐lauroyl sarcosinate, and casein phosphopeptide‐amorphous calcium phosphate (CPP–ACP) were excluded; studies where both the intervention group and comparator group received any additional potentially active agent in the toothpaste were included. This protocol change was instigated to reflect the change in toothpaste formulation since the initial review. Many types of toothpaste now combine the caries preventive benefit of fluoride with other therapeutic agents to control plaque (antibacterial agents), tartar, and gum disease.

Studies where both the intervention and control groups included participants receiving additional measures as part of their routine oral care such as supervised brushing, fissure sealants, were included, as were studies that were undertaken in areas with fluoridation of the community water supply.

Types of outcome measures

Primary outcomes

Caries increment as measured by:

change from baseline in the decayed, (missing), and filled surface or teeth index (D(M)FS/T), in all permanent teeth erupted at the start and erupting over the course of the study (dental caries is defined here as being clinically and radiologically recorded at the dentine level of diagnosis);
change from baseline in the decayed, (missing/extraction indicated), and filled surface or teeth index (d(e/m)fs/t) in all primary teeth;
change in the proportion of participants developing new caries.

Where possible we extracted data on D(M)FS/T, d(e/m)fs/t, and the proportion of participants developing new caries. We acknowledged that the primary outcome of caries could be measured and reported in a number of different ways. To account for this, the choice of outcome followed an adapted version of the hierarchy presented in the earlier review (Walsh 2010):

DFS/T data would be chosen over DMFS/T data, and these would be chosen over DS/T or FS/T
- data for 'all surface types combined' would be chosen over data for 'specific types' only
- data for 'all erupted and erupting teeth combined' (E + U) would be chosen over data for 'erupted' (E) only, and these over data for 'erupting' (U) only
data for dentinal/cavitated caries lesions (D₃ level) would be chosen over data for enamel/non‐cavitated lesions (D₁ level)
net caries increment data would be chosen over crude (observed) increment data
data from 'clinical and radiological examinations combined' would be chosen over data from 'clinical' (cl) only, and these over 'radiological' (xr) only. We defined clinical examinations as an examination using visual or tactile methods or both (VT) using a conventional lighting source. We excluded studies reporting caries using a non‐clinical examination alone (e.g. electronic caries monitor, quantitative light‐induced fluorescence)
follow‐up nearest to 3 years (often the one at the end of the study period) would be chosen over all follow‐up.

Clinical and radiographic examinations provided the definition of different stages or grades of caries lesions. These have been grouped into two basic grades for each method of examination.

Clinical:

non‐cavitated incipient enamel lesions clinically visible as white spots or discoloured fissures (NCA);
lesions showing loss of enamel continuity that can be recorded clinically (undermined enamel, softened floor, walls) or showing frank cavitation (CA).

Radiographic:

any radiolucency in enamel/enamel‐dentine junction (ER);
radiolucency into dentine (DR).

Secondary outcomes

Adverse effects such as irritation, dental staining/discolouration, etc.

Search methods for identification of studies

Electronic searches

Cochrane Oral Health's Information Specialist conducted systematic searches in the following databases for randomised controlled trials and controlled clinical trials without language or publication status restrictions:

Cochrane Oral Health's Trials Register (searched 15 August 2018) (Appendix 1);
Cochrane Central Register of Controlled Trials (CENTRAL; 2018, Issue 7) in the Cochrane Library (searched 15 August 2018) (Appendix 2);
MEDLINE Ovid (1946 to 15 August 2018) (Appendix 3);
Embase Ovid (1980 to 15 August 2018) (Appendix 4).

Subject strategies were modelled on the search strategy designed for MEDLINE Ovid.

Searching other resources

The following trial registries were searched for ongoing studies:

US National Institutes of Health Ongoing Trials Register ClinicalTrials.gov (clinicaltrials.gov; searched 15 August 2018) (Appendix 5);
World Health Organization International Clinical Trials Registry Platform (apps.who.int/trialsearch; searched 15 August 2018) (Appendix 6).

Previously published systematic reviews of fluoride toothpastes were also screened to identify any reports that met the inclusion criteria (Ammari 2003; Bartizek 2001; Clarkson 1993; Steiner 2004; Twetman 2003).

We did not perform a separate search for adverse effects of interventions used, we considered adverse effects described in included studies only.

Data collection and analysis

Selection of studies

The downloaded set of records from each database were imported into the bibliographic software package EndNote. Duplicate records were identified and removed. Two review authors independently scanned all records for relevance on the basis of title and abstract (where available). We discarded irrelevant records and obtained the full text of the remaining records for further evaluation. Relevancy was assessed according to the characteristics of the participants, nature of the intervention, and comparison.

Following the initial screening, for studies appearing to meet the inclusion criteria, or for which there was insufficient information in the title and abstract to make a clear decision, we obtained the full report. Two review authors assessed these reports independently and in duplicate to establish whether the studies met the inclusion criteria. Disagreements were resolved by discussion. Studies written in a language not known by the review team were translated by members of Cochrane Oral Health and included or excluded as appropriate. Those studies awaiting translation are presented in the Studies awaiting classification section of the review.

We recorded studies rejected at this or subsequent stages in the Characteristics of excluded studies section with the reason for exclusion. Studies could be excluded for more than one reason.

Data extraction and management

All review authors undertook data extraction independently and in duplicate for all studies meeting the inclusion criteria using a piloted data extraction form. We recorded the following data for each included study in the Characteristics of included studies tables.

Study information: location, duration of data collection (follow‐up months), date of baseline collection^a, number of centres, other sources of fluoride exposure^b, duration of intervention, individual or cluster‐randomisation.
Participant information: age at baseline, number randomised, baseline caries^c, mean decayed missing and filled surfaces/teeth (standard deviation (SD)/standard error (SE)) for primary or permanent dentition or both (dmfs/t, DMFS/T).
Intervention: concentration and formulation of fluoride, abrasive system, frequency of brushing, supervised brushing, duration of intervention.
Assessment: teeth included, criteria for clinical diagnosis, calculation of change/increment, diagnostic threshold, net or crude caries increment.
Outcome information: final caries and number evaluated, proportion of children developing new caries, mean DMFS/dmfs (SD/SE) increment, mean DMFT/dmft (SD/SE) increment, level of compliance.
Reliability of primary outcome measurement: number of examiners and calibration details, method of clinical assessment.
Adverse effects: e.g. soft tissue damage, dental stain, irritation.

^aWhen data on the study start were not provided, we calculated a 'probable date' by subtracting the duration of the study (in years) plus 1 extra year, from the publication date of the study.  ^bBackground exposure to other fluoride sources encompassed data on the use (outside the trial) of topical fluorides/fluoride rinses or even fluoride toothpastes (in studies where the intervention was tested under supervision at school and no supply of any toothpaste had been provided for home use), and the consumption of fluoridated water/salt/tablets. Background use of other fluorides (rinses, gels, tablets, etc.) should be clearly reported as used by the majority in a study to be considered as such.  ^cFrom the study sample analysed (final sample) and in connection with the caries increment index chosen.

Assessment of risk of bias in included studies

We assessed all studies included in the review for risk of bias independently and in duplicate as part of the data extraction process, with reference to the Cochrane Handbook for Systematic Reviews of Interventions (Higgins 2011). We assessed included studies on the following:

sequence generation (selection bias): high, low, unclear;
allocation concealment (selection bias): high, low, unclear;
blinding of participants and personnel (performance bias) and outcome assessment (detection bias): high, low, unclear;
incomplete outcome data (attrition bias): high, low, unclear;
selective reporting (reporting bias): high, low, unclear;
other bias (baseline imbalance, contamination or co‐intervention): high, low, unclear.

A judgement of 'high' indicated a high risk of bias, 'low' indicated low risk of bias, and 'unclear' indicated either a lack of information or uncertainty over the potential for bias.

We categorised the overall risk of bias of each study (Risk of bias in included studies). Studies were categorised as being at low, high, or unclear risk of bias according to the following criteria:

low risk of bias (plausible bias unlikely to seriously alter the results) if all domains were at low risk of bias;
high risk of bias (plausible bias that seriously weakens confidence in the results) if one or more domains were at high risk of bias; or
unclear risk of bias (plausible bias that raises some doubt about the results) if one or more domains were at unclear risk of bias.

A risk of bias table was completed for each included study (Characteristics of included studies). Results are also presented graphically by study (Figure 1) and by domain over all studies (Figure 2).

Risk of bias summary: review authors' judgements about each risk of bias item for each included study.

Risk of bias graph: review authors' judgements about each risk of bias item presented as percentages across all included studies.

Measures of treatment effect

For continuous outcomes, we pooled data with the mean difference (MD), or standardised mean difference (SMD) if different measures were used to assess the same outcome. In the absence of an agreed consensus of minimally important clinical effect for caries increment, we chose an SMD value of 0.30 to indicate clinical importance, representing a small to moderate effect size. In terms of interpretation, mean caries increments are closely related to their standard deviations (approximately equal), and meta‐analyses using SMDs will yield materially similar results to those using prevented fractions (Salanti 2009).

For dichotomous outcomes, we pooled data with the risk ratios (RR).

Unit of analysis issues

Studies with multiple treatment arms

In studies with more than one relevant intervention group and a common comparator group, such as those comparing fluoride NaF and SMFP toothpastes of the same concentration against a different fluoride concentration, we combined summary statistics (the number of participants analysed, number of participants developing caries, mean caries increments, and standard deviations) from all relevant intervention groups (and from any relevant comparator groups, if this was the case) to obtain a measure of treatment effect.

Cluster‐randomised trials

When cluster‐randomised trials did not report results adjusted for clustering present in the data, we performed an approximately correct analysis by estimating the design effect for such trials (Higgins 2011) by using an intra‐class correlation coefficient (ICC) value of 0.05 (a value commonly used in caries prevention trials) to reduce the numbers in intervention and control groups to their 'effective sample size.'

Dealing with missing data

For the main outcome data, missing standard deviations for caries increments not revealed through contact with the original researchers were imputed through linear regression of log (standard deviations) on log (mean caries) increments as per Walsh 2010. This is a suitable approach for caries prevention trials as caries increments are closely related (similar) to their standard deviations, following approximately a Poisson distribution (van Rijkom 1998).

Assessment of heterogeneity

Clinical and methodological heterogeneity within treatment comparisons

We assessed the presence of clinical heterogeneity by examining the characteristics of the studies, the similarity between the types of participants (e.g. age, community water fluoridation), and the interventions (e.g. additional potential active agents added to the toothpastes, supervised toothbrushing). Meta‐analysis was restricted to studies of similar comparisons that reported the same outcomes.

Transitivity across treatment comparisons

Where possible, we planned to undertake a network meta‐analysis (NMA) to compare the caries increments of the different fluoride concentrations.

We assessed the assumption of transitivity by comparing the distribution of the potential effect modifiers, community water fluoridation and supervised brushing, across the different elements of the network.

If the assumption of transitivity could be considered to be violated, for example, in terms of substantially imbalanced distributions of effect modifiers, then we would not conduct an overall network meta‐analysis. Instead, we planned to revert to performing a series of independent meta‐analyses if we observed heterogeneity across treatment comparisons.

Assessment of reporting biases

If at least 10 studies were included in a meta‐analysis, we planned to assess publication bias according to the recommendations on testing for funnel plot asymmetry, as described in Section 10.4 of the Cochrane Handbook for Systematic Reviews of Interventions (Higgins 2011). If asymmetry was identified, we would examine possible causes. We were not able to assess publication bias in this way because, although we had a sufficient number of studies in our meta‐analyses for the main comparisons, they were analysed as pairwise comparisons of fluoride concentrations that each contained fewer than 10 studies.

Data synthesis

Traditional approaches to meta‐analysis have focused on direct (head‐to‐head) pairwise comparisons within studies. However, when many different interventions exist the number of possible pairwise comparisons can be large, making data analysis and interpretation difficult. In this review, we proposed seven different categories of fluoride concentration, ranging from non‐fluoride toothpaste (0 ppm F) through to 2800 ppm F, and resulting in 21 possible comparisons of fluoride concentration. Reducing the number of categories would limit the number of comparisons, but could potentially obscure any subtle concentration‐related differences in effect. Where sufficient data were available, where possible, we planned to undertake an NMA to compare the caries increments of the different fluoride concentrations. This method combines direct evidence from the head‐to‐head studies with indirect evidence from the included studies (Higgins 2011).

Methods for direct treatment comparisons

Where the number of available studies for each comparison was sparse, or where there was a disconnected network of comparisons, we planned to conduct standard pairwise meta‐analyses using random‐effects models using Review Manager (RevMan) software (Review Manager 2014).

Methods for indirect and mixed comparisons

We planned to conduct an NMA using a multivariate approach in Stata (Stata 2017), where included studies formed a connected network. A graphical representation of the different fluoride concentration interventions was undertaken to illustrate the nature of the network. Estimates of treatment effects (MD or SMD as appropriate) and random‐effects NMA were undertaken using the 'Network' module in the Stata software package (White 2017) and self‐programmed Stata routines available from www.mtm.uoi.gr.

We analysed studies evaluating the caries preventive effects on the primary and permanent dentition separately throughout, due to differences in their anatomical and subsequent disease sequelae manifestations.

Subgroup analysis and investigation of heterogeneity

Assumptions when estimating heterogeneity

In standard meta‐analyses we estimated different heterogeneity variances for each pairwise comparison. In NMA we assumed a common estimate for the heterogeneity variance across the different comparisons.

Measures and tests for heterogeneity

We assessed statistically the presence of heterogeneity within each comparison using a Chi² test, where a P value < 0.1 was taken to indicate statistically significant heterogeneity. We quantified heterogeneity using the I² statistic that measures the percentage of variability that cannot be attributed to random error. An I² statistic of:

0% to 40% might not be important;
30% to 60% may represent moderate heterogeneity;
50% to 90% may represent substantial heterogeneity; and
75% to 100% considerable heterogeneity.

This is according to Section 9.5.2 of the Cochrane Handbook for Systematic Reviews of Interventions (Higgins 2011).

We assessed statistical heterogeneity in the entire network based on the magnitude of the heterogeneity variance parameter (τ²) estimated from the NMA models. We intended to use the I² measure for NMA (i² option in mvmeta).

Assessment of statistical inconsistency

We evaluated the statistical agreement between the various sources of evidence in each network of interventions (consistency). Since different approaches may lead to different conclusions about the magnitude of inconsistency, we used both local and global approaches.

We used the loop‐specific method to examine the presence of inconsistency locally (Bucher 1997; Veroniki 2013). This method evaluates the consistency assumption in each closed loop of the network separately as the difference between direct and indirect estimates for a specific comparison in the loop (inconsistency factor). We assessed whether the inconsistency factor was incompatible with a zero null value using a 95% confidence interval (CI) and a loop specific z‐test.

To assess statistical inconsistency in the entire network we used the 'design‐by‐treatment' interaction model (Higgins 2012). The presence of inconsistency from any source in the entire network will be inferred from a Chi² test. The network command was used in Stata for the design‐by‐treatment model (White 2017).

Investigation of heterogeneity and inconsistency

If we had observed important heterogeneity or inconsistency, or both, we planned to explore possible sources through subgroup analysis or meta‐regression if sufficient data were available. Two potential sources of heterogeneity were specified a priori: supervised toothbrushing and community water fluoridation.

Sensitivity analysis

We proposed two forms of sensitivity analysis: removing studies where both the intervention group and comparator group received any additional potentially active agent(s) in the toothpaste; and removing studies with the shortest observed follow‐up period (12 months).

Presentation of main results

Direct comparisons

We produced a 'Summary of findings' table for each comparison using GRADE methods (GRADE 2004) and GRADEpro GDT software (GRADEpro GDT 2015) for the outcomes caries increment (d(m)fs/t, D(M)FS/T), proportion of participants developing new caries, and adverse effects of toothpaste. We produced a separate table for each main comparison (young children (primary dentition), children and adolescents (immature permanent dentition), and adults (permanent dentition)). We assessed the certainty of the body of evidence for each direct comparison and outcome by considering the overall risk of bias of the included studies, the directness of the evidence, the inconsistency of the results, the precision of the estimates. We did not downgrade for reasons of reporting bias as too few studies contributed to the majority of the treatment estimates to draw meaningful conclusions. We categorised the certainty of each body of evidence as high, moderate, low, or very low.

NMA comparisons from the connected network

We used the Confidence in Network Meta‐Analysis tool (CINeMA) (CINeMA 2017) to evaluate and present our confidence in the findings from the NMA . The tool uses a methodological framework which considers six domains: within‐study bias, across‐studies bias, indirectness, imprecision, heterogeneity, and incoherence (Salanti 2014). The amount of information contributed by each study to the results of the NMA, as indicated by the contribution matrix, informs the confidence in the findings assessment.

Results

Description of studies

Results of the search

After applying the Cochrane randomised controlled trial (RCT) filter and removing duplicates, the search of electronic databases retrieved 3623 records.

The search for ongoing trials yielded no additional reports, as did the search of non‐electronic resources.

Following screening, we considered 250 records to be potentially eligible, and obtained them for further detailed assessment. This resulted in 96 included studies (168 records), 61 excluded studies (81 records), no ongoing studies and one study awaiting classification (the study has been completed but not yet written‐up). See Figure 3.

Included studies

See Characteristics of included studies table for details of included studies.

This version of the review includes 96 studies published between 1955 and 2014, of which 13 were new in this update: three studies published since the initial review (Cardoso 2014; Rao 2009; Vilhena 2010), five studies where additional information was obtained confirming eligibility (Biesbrock 2003a; Biesbrock 2003b; CL‐213 1983; CL‐216 1982; CL‐220 1986), one study previously awaiting classification (Takeuchi 1968), and four studies as a result of broadening the scope of the review (Jensen 1988; Lu 1980; Muhler 1957; Petersson 1991). The review includes placebo‐controlled trials and trials comparing one active intervention to at least one other active intervention, in the form of two‐, three‐, four‐ and five‐arm trials. One trial was cluster‐randomised (Sønju Clasen 1995) though reported as an individually‐randomised trial. The shortest follow‐up period was 12 months, the longest reported was 7 years. Multiple studies were conducted in the USA (37 studies), UK (22 studies), Sweden (six studies), Switzerland (six studies), Brazil (three studies), France (five studies), Canada (two studies), Germany (two studies), Italy (two studies), Australia (two studies), Guatemala (two studies). Single studies were conducted in Iceland, Denmark, China, Puerto Rico, India, Japan, and Lithuania.

Some publications have reported on multiple trials within the same publication due to studies conducted and reported separately: for different age groups (Marthaler 1965; Marthaler 1965a; Marthaler 1970; Marthaler 1970a; Zacherl 1970; Zacherl 1970a); from different locations (Forsman 1974; Forsman 1974a; Held 1968; Held 1968a; Held 1968b); or for different age groups in different locations (Torell 1965; Torell 1965a; Torell 1965b).

Three placebo‐controlled studies including 2675 randomised participants (2162 evaluated) assessed the effects of fluoride toothpaste on the mature permanent dentition in adults, whose age ranged from 18 to 93 years at the start of the studies. Reported mean decayed, missing, and filled permanent surfaces (DMFS) at baseline ranged from 27.3 to 53.35 carious surfaces (D₃ level (dentinal/cavitated)).

Eight studies including 13,856 randomised participants (9055 evaluated) assessed the effects of fluoride toothpaste on the primary dentition, with participants ranging from 1 to 4 years of age at the start of the studies. Reported mean decayed, missing, and filled primary surfaces (dmfs) at baseline ranged from 2.2 to 5.0 carious surfaces (three studies reporting baseline d₃ level), or from 0.95 to 1.85 non‐cavitated caries lesions (two studies reporting at white spot lesion level). Two studies used a zero fluoride concentration toothpaste as the comparator (placebo‐controlled) (Cahen 1982; Fan 2008). Supervised toothbrushing was used in four studies (Cardoso 2014; Lima 2008; Sønju Clasen 1995; Winter 1989).

Eighty‐six studies including 51,304 randomised participants (42,074 evaluated) assessed the effects of fluoride toothpaste on the immature permanent dentition in children and adolescents whose age ranged from 5 to 18 years of age at the start of the studies (one study also reported on the effects on the primary dentition). Reported mean DMFS at baseline ranged from 0.97 to 23.53 carious surfaces (D₃ level). 66 studies used a zero fluoride concentration toothpaste as the comparator (placebo‐controlled). Supervised toothbrushing was used in 24 studies (Ashley 1977; Biesbrock 2003a; Biesbrock 2003b; Blinkhorn 1983; Chesters 2002; Conti 1988; Di Maggio 1980; Fogels 1988; Glass 1978; Glass 1983; Held 1968; Held 1968a; Held 1968b; Hodge 1980; Howat 1978; Marks 1994; Mitropolous 1984; Peterson 1979; Rao 2009; Rule 1984; Segal 1967; Stookey 2004; Takeuchi 1968; Thomas 1966).

The most commonly reported primary outcome measure was caries increment at the tooth surface level, which was reported in all studies. Caries increment at the tooth level was reported in 60 studies (three studies in the primary dentition, 56 in the immature permanent dentition and one in the mature permanent dentition); the proportion of participants developing new caries was reported in 13 studies (three studies in the primary dentition, and 10 studies in the immature permanent dentition).

Adverse effects of the intervention were unreported in the majority of studies, but when reported included oral (soft tissue) damage and tooth staining. No trials reported on fluorosis.

Excluded studies

Reasons for exclusion of a trial from the review are given in the Characteristics of excluded studies table. The 61 studies were excluded for the following reasons: non‐random or systematic allocation; randomisation not stated or indicated; inappropriate randomisation (randomising two clusters, one to each of the groups compared); other fluoride‐based interventions in addition to fluoride toothpaste; or participants were institutionalised children or adolescents with specific health problems. A study could be excluded for more than one reason.

Risk of bias in included studies

Allocation

Method of randomisation

We assessed 33 (34%) studies at low risk of bias for this domain, where a clear statement of the randomisation method was reported. The remaining 63 studies did not report sufficient information to enable us to make a judgement and we assessed them as at unclear risk of bias. Statements such as 'were randomised' or 'were stratified' appeared most commonly.

Allocation concealment

We assessed eight (8%) studies as at low risk of bias for this domain, where a clear statement of the methods of concealing the allocation method was reported. The remaining 88 studies reported insufficient information on how the randomisation sequence was concealed from individuals involved in the enrolment and assignment of participants to enable us to make a judgement and we assessed them as at unclear risk of bias.

The overall risk of selection bias (randomisation and allocation concealment) was low in eight studies and unclear in 88 studies.

Blinding

Eighty‐six studies (90%) reported that they had made sufficient efforts to ensure that the fluoridated toothpastes were indistinguishable from each other and from the non‐fluoride toothpastes, and either clearly reported that the outcome assessor(s) was not aware of the participants' assignment or reported the trial as being 'double‐blind.' We assessed those studies as at low risk of bias for this domain.

For one study (Held 1968b) blinding of participants was not undertaken (placebo and fluoride toothpastes were packaged differently) and participants would have been aware of their assignment, thus introducing the potential for performance bias. We assessed this study as at high risk of bias. There was insufficient information from the remaining nine studies to form a judgement.

Incomplete outcome data

We applied an arbitrary threshold for a judgement of attrition given each study's length of follow‐up e.g. over 50% in 3 years, over 40% in 2 years, over 30% in 1 year resulting in a judgement of high risk of bias. We assessed 17 studies (18%) as at low risk of bias for this domain, with acceptable levels of attrition according to length of follow‐up and where the distribution was fairly evenly distributed across groups. 71 studies (74%) did not report the number or reasons for losses or exclusions by group or both and were judged to be at unclear risk of bias. High levels of attrition were reported in eight studies (8%), where overall attrition for length of follow‐up was not calculable as numbers randomised or evaluated or both were not provided for all groups (Beiswanger 1989; CL‐213 1983; Takeuchi 1968); attrition was unduly high for length of follow‐up, with or without differential losses between groups (Held 1968b; Lu 1987; Marthaler 1965a; Muhler 1957; Sønju Clasen 1995).

Selective reporting

We assessed 83 studies (86%) as at low risk of bias for this domain, as they reported all pre‐specified caries outcomes in full, according to the different units measured, methods of examination, diagnostic thresholds for caries, and approaches for reversals as stated in the methods section of each study report. Insufficient information was provided in 11 studies (11%) to form a judgement and we assessed those studies as at unclear risk of reporting bias.

We assessed the remaining two studies (2%) as at high risk of reporting bias due to no (Takeuchi 1968) or incomplete (Petersson 1991) reporting of the main outcome as stated in the methods section.

Other potential sources of bias

Other potential sources of bias considered included baseline imbalance for important factors such as caries and age, and contamination or co‐intervention principally from additional sources of fluoride.

Baseline characteristics

Baseline characteristics were reported and comparable between groups in 81 studies (84%). Baseline similarity was achieved most often through stratification by important prognostic variables. Five studies (5%) were assessed at high risk of bias where baseline imbalance in caries levels was reported in the different fluoride groups (Held 1968; Held 1968a; Held 1968b; Lima 2008; Petersson 1991). An unclear judgment was given for 10 (10%) studies when baseline characteristics were not reported or were not differentially reported.

Free of contamination or co‐intervention

One study (Biesbrock 2001) was judged to be at high risk of bias from contamination when a concurrent fluoride rinse programme was introduced to study participants. 67 trials (70%) were judged free from the possibility of any inadvertent application of the intervention being evaluated to people in the control group (contamination) or any additional treatment being given to one of the groups differentially (co‐intervention) or both, and hence were judged to be at low risk of bias. In 28 trials (29%) there was insufficient information to enable a judgement to be made.

Overall risk of bias

We assessed only one study (Mitropolous 1984) at low risk of bias for all domains, and therefore at low risk of bias overall.

We assessed 14 studies (Beiswanger 1989; Biesbrock 2001; CL‐213 1983; Held 1968; Held 1968a; Held 1968b; Lima 2008; Lu 1987; Marthaler 1965a; Muhler 1957; Muhler 1962; Petersson 1991; Sønju Clasen 1995; Takeuchi 1968) at high risk of bias for at least one domain, and therefore at high risk of bias overall. The most frequent high risk of bias judgements were in the incomplete outcome data domain (Beiswanger 1989; CL‐213 1983; Held 1968b; Lu 1987; Marthaler 1965a; Muhler 1957; Sønju Clasen 1995; Takeuchi 1968), followed by other potential sources of bias from baseline imbalance (Held 1968; Held 1968a; Held 1968b; Lima 2008; Petersson 1991), selective reporting (Petersson 1991; Takeuchi 1968), and potential contamination or co‐intervention (Biesbrock 2001).

We assessed 81 studies as being at unclear overall risk of bias. These studies had at least one domain judged to be at unclear risk of bias, but no domains judged to be at high risk of bias.

The results of the risk of bias assessments are presented graphically for each included study in Figure 1 and as percentages across all included studies in Figure 2.

Effects of interventions

See: Table 1; Table 2; Table 3; Table 4

Summary of findings for the main comparison. Fluoride toothpaste with higher versus lower concentration of fluoride for preventing caries in young children (primary dentition).

The effects of fluoride toothpaste with higher versus lower concentration of fluoride for preventing caries in young children (primary dentition)
Population: young children between 1 and 6 years of age at the start of the study  Setting: home, nursery, school  Intervention: toothbrushing with higher fluoride concentration toothpaste  Comparison: toothbrushing with lower fluoride concentration toothpaste Outcomes: primary dentition: d(m)fs, d(m)ft, proportion of children developing new caries, adverse effects of toothpaste
Comparison	Outcome	*Anticipated absolute effects^ (95% CI)**		Relative effect  (95% CI)	Number of participants  (studies)	Certainty of the evidence  (GRADE)	Comments
Comparison	Outcome	Risk with lower fluoride concentration toothpaste	Risk with higher fluoride concentration toothpaste	Relative effect  (95% CI)	Number of participants  (studies)	Certainty of the evidence  (GRADE)	Comments
Caries increment (surface index d₃mfs)
1500 ppm F compared with 0 ppm F	Caries increment (surface index d₃fs, adjusted for baseline value)  Follow‐up: 24 months	The mean caries increment with 0 ppm F was 4.73 d₃fs^a	The mean caries increment in the higher fluoride group was on average MD 1.86 lower (95% CI 2.51 lower to 1.21 lower)	‐	998  (1 RCT)	⊕⊕⊕⊝  moderate^b	Mean caries increment in the higher fluoride group was on average SMD 0.40 lower (95% CI 0.53 lower to 0.27 lower)
1450 ppm F compared with 250 ppm F	Caries increment (surface index d₃mfs)  Follow‐up: 22 months	The mean caries increment with 250 ppm F was 2.90 dmfs^c	The mean caries increment in the higher fluoridated group was on average MD 1.20 lower (95% CI 2.92 lower to 0.52 higher)	‐	172  (1 RCT)	⊕⊕⊝⊝  low^d	Effective sample size = 96 SMD 0.28 lower (95% CI 0.68 lower to 0.12 higher)
1055 to 1100 ppm F compared with 500 to 550 ppm F	Caries increment (surface index d₃mfs)  Follow‐up: range 20 to 36 months	The mean caries increment with 550 ppm F ranged from 2.05 to 2.52^e	The mean caries increment in the higher fluoridated group was on average MD 0.05 dmfs lower (95% CI 0.38 lower to 0.28 higher)	‐	1958  (2 RCTs)	⊕⊕⊕⊝  moderate^f	SMD 0.02 lower (95% CI 0.12 lower to 0.07 higher) 2 additional studies (12‐month follow‐up) reporting caries increment at the active non‐cavitated caries lesions (ANC) level with MD 0.31 lower (95% CI 0.93 lower to 0.32 higher) and SMD 0.20 lower (95% CI 0.67 lower to 0.27 higher)
Caries increment (tooth index d₃mft)
1450 ppm F compared with 250 ppm F	Caries increment (tooth index d₃mft)  Follow‐up: 22 months	The mean caries increment with 250 ppm F was 1.20 dmft^g	The mean caries increment in the higher fluoridated group was on average MD 0.40 lower (95% CI 1.14 lower to 0.34 higher)	‐	172  (1 RCT)	⊕⊕⊝⊝  low^d	Effective sample size = 96 SMD 0.22 lower (95% CI 0.62 lower to 0.18 higher)
1055 to 1100 ppm F compared with 500 to 550 ppm F	Caries increment (tooth index d₃mft)  Follow‐up: 36 months	The mean caries increment with 550 ppm F was 1.72^h	The mean caries increment in the higher fluoridated group was on average MD 0.27 dmft lower (95% CI 0.60 lower to 0.06 higher)	‐	905  (1 RCT)	⊕⊕⊝⊝  lowⁱ	SMD 0.11 lower (95% CI 0.24 lower to 0.02 higher)
1450 F compared with 440 ppm F	Caries increment (tooth index d₃mft)  Follow‐up: 60 months	The mean caries increment with 450 ppm F was 2.49^j	The mean caries increment in the higher fluoridated group was on average MD 0.34 dmft lower (95% CI 0.59 lower to 0.09 lower)	‐	2362 (1 RCT)	⊕⊕⊕⊝  moderate^k	SMD 0.11 lower (95% CI 0.19 lower to 0.03 lower)
Proportion of children developing new caries
1450 ppm F compared with 250 ppm F	Proportion of children developing new caries  Follow‐up: 22 months	455 per 1000^l	418 per 1000  (245 to 714)	RR 0.92  (0.54 to 1.57)	172  (1 RCT)	⊕⊕⊝⊝  low^d	Effective sample size = 69
1055 to 1100 ppm F compared with 500 to 550 ppm F	Proportion of children developing new caries  Follow‐up: 36 months	484 per 1000^m	416 per 1000  (358 to 479)	RR 0.86 (0.74 to 0.99)	905  (1 RCT)	⊕⊕⊝⊝  lowⁱ
1450 F compared with 440 ppm F	Proportion developing new caries Follow‐up: 60 months	577 per 1000ⁿ	502 per 1000  (467 to 542)	RR 0.87  (0.81 to 0.94)	2362  (1 RCT)	⊕⊕⊕⊝  moderate^k
Adverse effects of toothpaste
1500 ppm F compared with 0 ppm F	Adverse effects of toothpaste Follow‐up: 24 months			Not estimable	998  (1 RCT)	⊕⊕⊝⊝  low^o	0 events reported
1450 ppm F compared with 250 ppm F	Adverse effects of toothpaste	No studies reported this outcome
1055 to 1100 ppm F compared with 500 to 550 ppm F	Adverse effects of toothpaste  Follow‐up: 12 months			Not estimable	195 (1 RCT)	⊕⊕⊝⊝  low^p	"There were no reports on adverse effects, but some children complained about the taste of the dentifrice"
1450 F compared with 440 ppm F	Adverse effects of toothpaste	No studies reported this outcome
^*The risk in the intervention group (and its 95% CI) is based on the assumed risk in the comparison group and the relative effect of the intervention (and its 95% CI).  CI: confidence interval; d(m)fs/t: decayed, missing, filled primary surfaces/teeth; d₃: dentinal/cavitated caries lesions level; F: fluoride; ppm: parts per million; RR: risk ratio; MD: difference in means; SMD: standardised mean difference.
GRADE Working Group grades of evidence  High certainty: we are very confident that the true effect lies close to that of the estimate of the effect.  Moderate certainty: we are moderately confident in the effect estimate: the true effect is likely to be close to the estimate of the effect, but there is a possibility that it is substantially different.  Low certainty: our confidence in the effect estimate is limited: the true effect may be substantially different from the estimate of the effect.  Very low certainty: we have very little confidence in the effect estimate: the true effect is likely to be substantially different from the estimate of effect.

The effects of fluoride toothpaste with higher versus lower concentration of fluoride for preventing caries in children and adolescents ‐ NMA estimates
Population: children and adolescents between 5 and 15 years of age at the start of the study Setting: home, school Intervention: toothbrushing with higher fluoride concentration toothpaste Comparison: toothbrushing with lower fluoride concentration toothpaste Outcomes: immature permanent dentition: D(M)FS, D(M)FT (all follow‐ups: closest to 36 months)
Comparison	Relative effect SMD (NMA estimate)  (95% CI)	Studies^*	Certainty of evidence  (GRADE)	Relative effect SMD (NMA estimate)  (95% CI)	Studies^*	Certainty of evidence  (GRADE)
Comparison	Caries increment (surface index D₃(M)FS)			Caries increment (tooth index D₃(M)FT)
250 versus 0 ppm F	‐0.15 (‐0.25 to ‐0.05)	3	low^{a, b, c}	‐0.11 (‐0.27 to 0.05)	1	low^{a, b, c}
440 to 550 versus 0 ppm F	‐0.12 (‐0.31 to 0.07)	2	low^{a, c, d}	‐0.18 (‐0.41 to 0.04)	2	low^{a, c}
1000 to 1250 versus 0 ppm F	‐0.28 (‐0.32 to ‐0.25)	55	high^{a, d}	‐0.26 (‐0.31 to ‐0.21)	41	high^a
1450 to 1500 versus 0 ppm F	‐0.36 (‐0.43 to ‐0.29)	4	moderate^{a, d}	‐0.39 (‐0.49 to ‐0.28)	4	moderate^a
1700 to 2200 versus 0 ppm F	‐0.32 (‐0.41 to ‐0.22)	‐	low^{a, d}	‐0.33 (‐0.56 to ‐0.11)	‐	low^{d, e}
2400 to 2800 versus 0 ppm F	‐0.41 (‐0.49 to ‐0.33)	3	low^{a, b, c}	‐0.39 (‐0.52 to ‐0.25)	2	low^{a, c, d}

440 to 550 versus 250 ppm F	0.03 (‐0.17 to 0.23)	1	low^{a, b, c}	‐0.08 (‐0.32 to 0.17)	1	low^{a, c, d}
1000 to 1250 versus 250 ppm F	‐0.14 (‐0.24 to ‐0.04)	7	low^{a, b, c}	‐0.15 (‐0.31 to 0)	3	low^{a, c, d}
1450 to 1500 versus 250 ppm F	‐0.21 (‐0.33 to ‐0.10)	‐	low^{a, d}	‐0.28 (‐0.47 to ‐0.09)	‐	low^{a, d}
1700 to 2200 versus 250 ppm F	‐0.17 (‐0.30 to ‐0.04)	‐	low^{a, d}	‐0.23 (‐0.50 to 0.05)	‐	low^{d, e}
2400 to 2800 versus 250 ppm F	‐0.26 (‐0.38 to ‐0.14)	‐	low^{a, d}	‐0.28 (‐0.48 to ‐0.08)	‐	low^{a, d}

1000 to 1250 versus 440 to 550 ppm F	‐0.16 (‐0.35 to 0.03)	1	low^{a, d}	‐0.08 (‐0.30 to 0.15)	1	low^{a, c, d}
1450 to 1500 versus 440 to 550 ppm F	‐0.24 (‐0.44 to ‐0.04)	‐	low^{a, d}	‐0.20 (‐0.45 to 0.04)	‐	low^{a, d}
1700 to 2200 versus 440 to 550 ppm F	‐0.20 (‐0.41 to 0.01)	‐	low^{a, d}	‐0.15 (‐0.46 to 0.16)	‐	low^{a, d}
2400 to 2800 versus 440 to 550 ppm F	‐0.29 (‐0.49 to ‐0.08)	‐	low^{a, d}	‐0.20 (‐0.46 to 0.05)	‐	low^{a, d}

1450 to 1500 versus 1000 to 1250 ppm F	‐0.08 (‐0.14 to ‐0.01)	10	moderate^a	‐0.13 (‐0.23 to ‐0.02)	4	low^{a, b}
1700 to 2200 versus 1000 to 1250 ppm F	‐0.03 (‐0.12 to 0.06)	5	low^{c, e}	‐0.07 (‐0.30 to 0.15)	1	very low^{c, d, e}
2400 to 2800 versus 1000 to 1250 ppm F	‐0.12 (‐0.20 to ‐0.05)	6	low^{a, b, c}	‐0.12 (‐0.25 to 0.01)	3	very low^{b, c, e}

1700 to 2200 versus 1450 to 1500 ppm F	0.04 (‐0.07 to 0.15)	‐	moderate^a	0.05 (‐0.19 to 0.30)	‐	low^{a, b, d}
2400 to 2800 versus 1450 to 1500 ppm F	‐0.05 (‐0.14 to 0.05)	2	moderate^a	0 (‐0.16 to 0.17)	‐	low^{a, f}

2400 to 2800 versus 1700 to 2200 ppm F	‐0.09 (‐0.20 to 0.02)	1	low^{a, e}	‐0.05 (‐0.27 to 0.17)	‐	very low^{b, e, g}
^Indirect evidence only is indicated by ‐  CI:* confidence interval; D(M)FS: decayed, missing, filled surfaces (permanent dentition); D(M)FT: decayed, missing, filled teeth (permanent dentition); D₃: dentinal/cavitated caries lesions level; F: fluoride; NMA: network meta‐analysis; ppm: parts per million; SMD: standardised mean difference.
GRADE Working Group grades of evidence  High certainty: we are very confident that the true effect lies close to that of the estimate of the effect.  Moderate certainty: we are moderately confident in the effect estimate: the true effect is likely to be close to the estimate of the effect, but there is a possibility that it is substantially different.  Low certainty: our confidence in the effect estimate is limited: the true effect may be substantially different from the estimate of the effect.  Very low certainty: we have very little confidence in the effect estimate: The true effect is likely to be substantially different from the estimate of effect.

The effects of fluoride toothpaste with higher versus lower concentration of fluoride for preventing caries in adults
Population: adults aged between 18 and 93 years of age at the start of the study  Setting: home  Intervention: toothbrushing with higher fluoride toothpaste concentration  Comparison: toothbrushing with lower fluoride toothpaste concentration Outcomes: mature permanent dentition: D(M)FS, D(M)FT, proportion of adults developing new caries, adverse effects of toothpaste
Comparison	Outcomes	*Anticipated absolute effects^ (95% CI)**		Relative effect  (95% CI)	Number of participants  (studies)	Certainty of the evidence  (GRADE)	Comments
		Risk with lower fluoride concentration toothpaste	Risk with higher concentration toothpaste
1000 or 1100 ppm F compared with 0 ppm F	Caries increment (surface index D₃MFS)  Follow‐up: 12 to 24 months	The mean caries increment with 0 ppm F ranged from 0.69 to 4.99 D₃MFS^a	The mean caries increment in the higher fluoride group was on average MD 0.53 lower (95% CI 1.02 lower to 0.04 lower)	‐	2162  (3 RCTs)	⊕⊕⊕⊝  moderate^b	Mean caries increment in the higher fluoride group was on average SMD 0.17 lower (95% CI 0.29 lower to 0.06 lower)
1000 or 1100 ppm F compared with 0 ppm F	Caries increment (tooth index D₃MFT) Follow‐up: 24 months	The mean value with 0 ppm F was 1.52 D₃MFT^c	The mean caries increment in the higher fluoride group was on average MD 0.46 lower (95% CI 0.93 lower to 0.01 higher)	‐	247 (1 RCT)	⊕⊕⊝⊝ low^d	Mean caries increment in the higher fluoride group was on average SMD 0.24 lower (95% CI 0.49 lower to 0.01 higher)
1000 or 1100 ppm F compared with 0 ppm F	Proportion of adults developing new caries	No studies reported this outcome
1000 or 1100 ppm F compared with 0 ppm F	Adverse effects of toothpaste	No studies reported this outcome
^*The risk in the intervention group (and its 95% CI) is based on the assumed risk in the comparison group and the relative effect of the intervention (and its 95% CI).  CI: confidence interval; D(M)FS: decayed, missing, filled surfaces (permanent dentition); D(M)FT: decayed, missing, filled teeth (permanent dentition); D₃: dentinal/cavitated caries lesions level; F: fluoride; MD: difference in means; ppm: parts per million; RCT: randomised controlled trial; SMD: standardised mean difference.
GRADE Working Group grades of evidence  High certainty: we are very confident that the true effect lies close to that of the estimate of the effect.  Moderate certainty: we are moderately confident in the effect estimate: the true effect is likely to be close to the estimate of the effect, but there is a possibility that it is substantially different.  Low certainty: our confidence in the effect estimate is limited: the true effect may be substantially different from the estimate of the effect.  Very low certainty: we have very little confidence in the effect estimate: The true effect is likely to be substantially different from the estimate of effect.

Fluoride concentration	Supervised toothbrushing		Community water fluoridation > 0.5 ppm F
Fluoride concentration	D(M)FS	D(M)FT	D(M)FS	D(M)FS
0 ppm F (placebo toothpaste)	15/61 (25%)	13/46 (28%)	6/41 (15%) (unreported 20 studies)	6/31 (unreported 15 studies)
250 ppm F	1/4 (25%)	1/2 (50%)	1/4 (25%)	1/2 (50%)
440 to 550 ppm F	No studies	No studies	No studies	No studies
1000 to 1250 ppm F	4/20 (20%)	3/7 (43%)	3/15 (20%) (unreported 5 studies)	2/7 (29%)
1450 to 1500 ppm F	2/8 (25%)	2/2 (100%)	1/4 (25%) (unreported 4 studies)	1/2 (50%)
1700 to 2200 ppm F	0/5 (0%)	0/1 (0%)	1/5 (20%)	0/1 (0%)
2400 to 2800 ppm F	1/6 (17%)	0/3 (0%)	0/3 (0%) (unreported 3 studies)	0/3 (0%)

Fluoride concentration comparison (ppm F)	NMA SMD (95% CI) NMA SMD (95% PI)	Direct SMD (95% CI) Indirect SMD (95% CI) Direct contribution
250 versus 0	‐0.15 (‐0.25 to ‐0.05) ‐0.15 (‐0.37 to 0.07)	‐0.09 (‐0.25 to 0.07)  ‐0.19 (‐0.33 to ‐0.06) 42.9%
440 to 550 versus 0	‐0.12 (‐0.31 to 0.07) ‐0.12 (‐0.40 to 0.15)	‐0.06 (‐0.29 to 0.17)  ‐0.27 (‐0.64 to 0.10) 71.9%
1000 to 1250 versus 0	‐0.28 (‐0.32 to ‐0.25) ‐0.28 (‐0.48 to ‐0.09)	‐0.28 (‐0.32 to ‐0.24)  ‐0.33 (‐0.45 to ‐0.21) 91.5%
1450 to 1500 versus 0	‐0.36 (‐0.43 to ‐0.29) ‐0.36 (‐0.57 to ‐0.15)	‐0.34 (‐0.46 to ‐0.22)  ‐0.37 (‐0.46 to ‐0.29) 33.8%
1700 to 2200 versus 0^*	‐0.32 (‐0.41 to ‐0.22) ‐0.32 (‐0.53 to ‐0.10)	‐ ‐0.32 (‐0.53 to ‐0.10) ‐
2400 to 2800 versus 0	‐0.41 (‐0.49 to ‐0.33) ‐0.41 (‐0.62 to ‐0.20)	‐0.523 (‐0.689 to ‐0.36)  ‐0.372 (‐0.47 to ‐0.28) 24.8%

440 to 550 versus 250	0.03 (‐0.17 to 0.23) 0.03 (‐0.25 to 0.31)	‐0.01 (‐0.25 to 0.23)  0.11 (‐0.26 to 0.49)  70.2%
1000 to 1250 versus 250	‐0.14 (‐0.24 to ‐0.04) ‐0.14 (‐0.36 to 0.08)	‐0.10 (‐0.21 to 0.01)  ‐0.36 (‐0.63 to ‐0.09)  86.1%
1450 to 1500 versus 250^*	‐0.21 (‐0.33 to ‐0.10) ‐0.21 (‐0.44 to 0.02)	‐ ‐0.21 (‐0.33 to ‐0.10) ‐
1700 to 2200 versus 250^*	‐0.17 (‐0.30 to ‐0.04) ‐0.17 (‐0.41 to 0.07)	‐ ‐0.17 (‐0.30 to ‐0.04) ‐
2400 to 2800 versus 250^*	‐0.26 (‐0.38 to ‐0.14) ‐0.26 (‐0.49 to ‐0.03)	‐ ‐0.26 (‐0.38 to ‐0.14) ‐

1000 to 1250 versus 440 to 550	‐0.16 (‐0.35 to 0.03) ‐0.16 (‐0.44 to 0.11)	‐0.11 (‐0.35 to 0.14)  ‐0.26 (‐0.58 to 0.06)  63.5%
1450 to 1500 versus 440 to 550^*	‐0.24 (‐0.44 to ‐0.04) ‐0.24 (‐0.52 to 0.04)	‐ ‐0.24 (‐0.44 to ‐0.04) ‐
1700 to 2200 versus 440 to 550^*	‐0.20 (‐0.41 to 0.01) ‐0.20 (‐0.48 to 0.09)	‐ ‐0.20 (‐0.41 to 0.01) ‐
2400 to 2800 versus 440 to 550^*	‐0.29 (‐0.49 to ‐0.08) ‐0.29 (‐0.57 to 0.00)	‐ ‐0.29 (‐0.49 to ‐0.08) ‐

1450 to 1500 versus 1000 to 1250	‐0.08 (‐0.14 to ‐0.01) ‐0.08 (‐0.28 to 0.13)	‐0.10 (‐0.17 to ‐0.03)  0.02 (‐0.12 to 0.15)  77.8%
1700 to 2200 versus 1000 to 1250	‐0.03 (‐0.12 to 0.06) ‐0.03 (‐0.25 to 0.18)	‐0.02 (‐0.12 to 0.07)  ‐0.19 (‐0.58 to 0.20)  94.4%
2400 to 2800 versus 1000 to 1250	‐0.12 (‐0.20 to ‐0.05) ‐0.12 (‐0.33 to 0.09)	‐0.09 (‐0.18 to 0.00)  ‐0.20 (‐0.35 to ‐0.06)  71.1%

1700 to 2200 versus 1450 to 1500^*	0.04 (‐0.07 to 0.15) 0.04 (‐0.18 to 0.27)	‐ 0.04 (‐0.07 to 0.15) ‐
2400 to 2800 versus 1450 to 1500	‐0.05 (‐0.14 to 0.05) ‐0.05 (‐0.26 to 0.17)	‐0.11 (‐0.28 to 0.07)  ‐0.03 (‐0.14 to 0.09)  28.5%

2400 to 2800 versus 1700 to 2200	‐0.09 (‐0.20 to 0.02) ‐0.09 (‐0.32 to 0.13)	‐0.01 (‐0.22 to 0.20)  ‐0.13 (‐0.26 to 0.01)  29.6%

Comparison	1000 to 1250			440 to 550
Study	Mean	SD	n	Mean	SD	n	SMD (95% CI)
Stookey 2004	6.27	4.58	174	6.24	4.49	168	0.01 (‐0.21 to 0.22)
Comparison	1450 to 1500			440 to 550
Study	Mean	SD	n	Mean	SD	n	SMD (95% CI)
Biesbrock 2003b	0.21	4.02	180	0.26	4.03	169	‐0.05 (‐0.90 to 0.80)
Comparison	1450 to 1500			1000 to 1250
Study	Mean	SD	n	Mean	SD	n	SMD (95% CI)
Marks 1994	4.27	3.91	1116	4.33	3.91	1120	‐0.02 (‐0.10 to 0.07)
Comparison	1700 to 2200			1000 to 1250
Study	Mean	SD	n	Mean	SD	n	SMD (95% CI)
Marks 1994	3.78	3.91	2126	4.33	3.91	1120	‐0.14 (‐0.21 to ‐0.07)
Comparison	2400 to 2800			1000 to 1250
Study	Mean	SD	n	Mean	SD	n	SMD (95% CI)
Biesbrock 2003a	1.25	3.59	153	1.47	3.50	168	‐0.06 (‐0.28 to 0.16)
Marks 1994	3.46	3.91	1112	4.33	3.91	1120	‐0.22 (‐0.31 to ‐0.14)
Comparison	1700 to 2200			1450 to 1500
Study	Mean	SD	n	Mean	SD	n	SMD (95% CI)
Marks 1994	3.78	3.91	2126	4.27	3.91	1116	‐0.13 (‐0.20 to ‐0.05)
Comparison	2400 to 2800			1450 to 1500
Study	Mean	SD	n	Mean	SD	n	SMD (95% CI)
Marks 1994	3.46	3.91	1112	4.27	3.91	1116	‐0.21 (‐0.29 to ‐0.12)
Comparison	2400 to 2800			1700 to 2200
Study	Mean	SD	n	Mean	SD	n	SMD (95% CI)
Marks 1994	3.46	3.91	1112	3.78	3.91	2126	‐0.08 (‐0.15 to ‐0.01)

Date	Event	Description
28 May 2019	Amended	Minor typo corrected.
21 February 2019	New citation required and conclusions have changed	New co‐author; broadened scope of review to include adults; review update including 13 new studies bringing the total to 96 included studies; methods updated; and 'Summary of findings' tables included
15 August 2018	New search has been performed	Searches updated to 15 August 2018
20 January 2010	Amended	Minor edits (contact details and acknowledgements)

Outcome or subgroup title	No. of studies	No. of participants	Statistical method	Effect size
1 DMFS	3	2162	Mean Difference (IV, Random, 95% CI)	‐0.53 [‐1.02, ‐0.04]
2 DMFT	1		Mean Difference (IV, Random, 95% CI)	Subtotals only

Outcome or subgroup title	No. of studies	No. of participants	Statistical method	Effect size
1 dmfs/ANC	4		Mean Difference (IV, Random, 95% CI)	Subtotals only
1.1 dmfs	2	1958	Mean Difference (IV, Random, 95% CI)	‐0.05 [‐0.38, 0.28]
1.2 ANC	2	285	Mean Difference (IV, Random, 95% CI)	‐0.31 [‐0.93, 0.32]
2 dmft	1		Mean Difference (IV, Random, 95% CI)	Subtotals only
3 Proportion developing new caries	1		Risk Ratio (M‐H, Random, 95% CI)	Subtotals only

Outcome or subgroup title	No. of studies	No. of participants	Statistical method	Effect size
1 dmft	1		Mean Difference (IV, Random, 95% CI)	Subtotals only
2 Proportion developing new caries	1		Risk Ratio (M‐H, Random, 95% CI)	Subtotals only

Outcome or subgroup title	No. of studies	No. of participants	Statistical method	Effect size
1 D(M)FS closest to 3 years	3	1460	Std. Mean Difference (IV, Random, 95% CI)	‐0.09 [‐0.19, 0.02]
2 D(M)FT	1		Std. Mean Difference (IV, Random, 95% CI)	Subtotals only
3 Proportion developing new caries	2	684	Risk Ratio (M‐H, Random, 95% CI)	1.07 [0.91, 1.27]

Methods	Trial design: 2‐armed, double‐blind, placebo‐controlled and stratified RCT  Location: UK  Number of centres: 3 comprehensive schools (Henbury, Brislington, Withywood) in Bristol, UK  Recruitment period: study commenced in 1970
Participants	Inclusion criteria: first‐year intakes at each school  Exclusion criteria: not reported  Baseline caries: 6.9 DFS (Gp A: 6.85 DFS (SD 4.38); Gp B: 6.97 DFS (SD 4.72)). Baseline characteristics (TAR, DFS, DMFS, DFT, DMFT, ECSI) "balanced"  Age at baseline (years): 11 to 12 years (Gp A: 11.72 (SD 0.35); Gp B: 11.70 (SD 0.34)). Baseline characteristic (age) "balanced" Sex: overall: 343 F:397 M (Gp A: 171 F:193 M; Gp B: 172 F:204 M). Evaluated participants reported only  Any other details of important prognostic factors: no background exposure to fluoride reported; natural fluoride level of community water supply ranged from 0.07 to 0.14 ppm F  Number randomised: 846 (Gp A: 418; Gp B: 428)  Number evaluated: 740 at 3 years (present at final assessment. Gp A: 364; Gp B: 376)  Attrition: 13% dropout after 3 years. Main reasons for attrition described: moved away, absent at final examination; no differential group losses
Interventions	Comparison: FT versus PL  Gp A (n = 418): SMFP 1000 ppm F; Al oxide trihydrate abrasive system; home use/unsupervised, daily frequency assumed  Gp B (n = 428): placebo; Al oxide trihydrate abrasive system; home use/unsupervised, daily frequency assumed
Outcomes	Primary: 3‐year net DFS increment ‐ (E + U) (CA) cl + (ER) xr; DMFS; DFT; DMFT; PF‐DMFS; MD‐BL‐DMFS; MD‐DMFS; O‐DMFS; ECSI (at 3 years follow‐up)  Secondary: none assessed  Assessments irrelevant to this review's scope: oral debris index; gingival index  Follow‐up duration: 3 years
Notes	Adverse effects: not reported  Funding source: "supported by Gibbs Dental Research"  Declarations/conflicts of interest: not reported  Data handling by review authors: none  Other information of note: clinical (VT) caries assessment by 2 examiners, diagnostic threshold = CA. Radiographic assessment (2 postBW) by 2 examiners, diagnostic threshold = ER. State of tooth eruption included = E/U. Reproducibility ratio was less than 0.22 for intra‐examiner reproducibility of clinical and radiographic caries diagnosis; "significant differences between examiners could not have affected caries increment figures since each examined same children annually"
*Risk of bias*
Bias	Authors' judgement	Support for judgement
Random sequence generation (selection bias)	Unclear risk	Quote: "Following baseline examinations, the children were grouped on the basis of age, sex, previous caries experience and the number of erupted second permanent molars; they were then randomly assigned to either the test or control group" Comment: not enough information provided
Allocation concealment (selection bias)	Unclear risk	No information provided
Blinding (performance bias and detection bias)   All outcomes	Low risk	Quotes: "Radiographs were examined ...... without reference to the clinical examination data..." "The test dentifrice contained MFP.... The toothpastes were packed in similar but distinguishable tubes. The investigators did not know which of the tubes contained the test paste nor which of the pastes any child was using" Comment: blind outcome assessment and use of placebo described
Incomplete outcome data (attrition bias)   All outcomes	Unclear risk	Overall dropout for length of follow‐up: 13% in 3 years. Dropout by group: 54/418 FT, 52/428 PL. Reasons for losses: did not like taste of paste (1 from control group), changed school or moved away (63), exclusion due to absence at last examination Comment: numbers lost were not unduly high for the length of follow‐up, with no differential loss between groups. It is unclear if reasons for missing outcome data are acceptable and balanced. Caries data used in analysis pertain to participants present at final examination
Selective reporting (reporting bias)	Low risk	Outcomes reported:  DFS increment ‐ (E + U) (CA) cl + (ER) xr, reported at 3 years follow‐up  DMFS  DFT  DMFT  PF‐DMFS  MD‐BL‐DMFS  MD‐DMFS  O‐DMFS  ECSI Comment: trial protocol not available. All pre‐specified outcomes (in Methods) were reported and were reported in the pre‐specified way
Baseline characteristics balanced?	Low risk	Prognostic factors reported:  DFS: 6.30 (4.04) FT, 6.43 (4.31) PL age (years): 11.73 (0.33) FT, 11.69 (0.32) PL TAR: 17.25 (4.35) FT, 17.35 (4.40) PL dental age: 21.75 (4.51) FT, 21.77 (4.47) PL DFT: 4.52 (2.56) FT, 4.42 (2.66) PL DMFT: 5.04 (2.68) FT, 4.96 (2.99) PL DMFS: 8.80 (6.55) FT, 9.10 (7.25) PL ECSI: 12.03 (8.34) FT, 12.41 (8.66) PL Comment: initial caries appears balanced
Free of contamination/co‐intervention?	Low risk	Quote: "The distribution of toothpastes and toothbrushes was the responsibility of two ladies called 'home visitors', whose duties were to visit each home every 5 weeks to supply enough of the appropriate toothpaste for the needs of the whole family.... and maintaining the interest and co‐operation of participants throughout the trial" Comment: there is sufficient indication overall of prevention of contamination/co‐intervention

Methods	Trial design: 4‐armed, double‐blind, placebo‐controlled and stratified RCT  Location: UK  Number of centres: 9 comprehensive high schools in London, UK  Recruitment period: study commenced 1973
Participants	Inclusion criteria: first‐year high school students attending 1 of 9 selected schools  Exclusion criteria: not reported  Baseline caries: 9.1 DFS (Gp A: 8.44 DFS (SD 5.58); Gp B: 9.79 DFS (SD 7.28)). Baseline characteristics (DFS, DMFS, DMFT) "balanced"  Age at baseline (years): mean 12 years (Gp A: 12.33; Gp B: 12.28). Baseline characteristic (age) "balanced" Sex: overall 279 F:210 M (Gp A: 138 F:108 M; Gp B: 141 F:102 M)  Any other details of important prognostic factors: no background exposure to fluoride. Water supply fluoridated to 0.3 ppm F  Number randomised: 1135 (numbers for relevant groups not reported)  Number evaluated: overall: 1002 at 2 years; relevant groups to this review: 489 at 2 years (present at final assessment. Gp A: 246; Gp B: 243)  Attrition: 12% dropout (for all study groups combined) after 2 years. Natural losses; any differential group losses not assessable
Interventions	Comparison: FT versus PL  Gp A (n = not reported): SMFP 1000 ppm F; IMP (main abrasive) abrasive system; school use/supervised, daily, 1 g applied for 1 min, post‐brushing water rinse done (non‐fluoride toothpaste provided to all for home use)  Gp B (n = not reported): placebo; IMP (main abrasive) abrasive system; school use/supervised, daily, 1 g applied for 1 min, post‐brushing water rinse done (non‐fluoride toothpaste provided to all for home use)
Outcomes	Primary: 2‐year net DFS increment ‐ (E + U) (NCA) cl + (ER) xr; DFS (U); PF‐DFS; MD‐BL‐DFS; MD‐DFS  Secondary: none assessed  Assessments irrelevant to this review's scope: compliance; inflamed gingival units/person; calculus sites/person  Follow‐up duration: 2 years
Notes	Adverse effects: not reported  Funding source: "support of the Warner‐Lambert Research Institute" (manufacturers of Listerine mouthwash and Dentyne chewing gum)  Declarations/conflicts of interest: not reported  Data handling by review authors: Gps C and D (fluoride mouthrinse and fluoride mouthrinse + fluoride toothpaste) irrelevant to this review  Other information of note: clinical (V) caries assessment by 1 examiner (FOTI used), diagnostic threshold = NCA. Radiographic assessment (postBW) by 1 examiner, diagnostic threshold = ER. State of tooth eruption included = E/U. Intra‐examiner reproducibility checks for incremental caries data (ICC for clinical 0.95, for radiographic 0.8); reversal rate between 12% and 7% of observed DFS increment in study groups
*Risk of bias*
Bias	Authors' judgement	Support for judgement
Random sequence generation (selection bias)	Low risk	Quote: "Using a table of random numbers, subjects were allocated within each school to one of four study groups"
Allocation concealment (selection bias)	Unclear risk	No information provided
Blinding (performance bias and detection bias)   All outcomes	Low risk	Quotes: "The control dentifrice was identical, except that it did not contain sodium MFP..." "The study was organised on a double‐blind basis..." "Records of earlier examinations were not available at the subsequent examination sessions" Comment: blind outcome assessment and use of placebo described
Incomplete outcome data (attrition bias)   All outcomes	Unclear risk	Overall dropout for length of follow‐up: 12% in 2 years (133/1135, all 4 groups combined). Dropout by group: not reported. Reasons for losses: mainly due to moving from the area Comment: numbers lost were not unduly high given length of follow‐up; it is unclear if there were any differential losses, and if reasons for missing outcome data are acceptable and balanced. Caries data used in the analysis pertain to participants present at baseline and final exams
Selective reporting (reporting bias)	Low risk	Outcomes reported:  DFS increment ‐ (E + U) (NCA) cl + (ER) xr, reported at 2 years follow‐up  PF‐DFS  MD‐BL‐DFS  MD‐DFS  DFS (U) Comment: trial protocol not available. All pre‐specified outcomes (in Methods) were reported and were reported in the pre‐specified way
Baseline characteristics balanced?	Low risk	Prognostic factors reported:  DFS: 8.44 (5.58) FT, 9.79 (7.28) PL DMFT: 5.35 (3.03) FT, 6.06 (3.66) PL DMFS: 9.89 (6.94) FT, 11.05 (7.98) PL age: 12.33 FT, 12.28 PL Comment: initial caries appears balanced between groups
Free of contamination/co‐intervention?	Low risk	Quote: "...all subjects received ample supplies of the non‐fluoride control toothpaste and toothbrushes. This ensured that the exposure of the subjects to fluoride dentifrice or rinse was restricted to the experimental regime" Comment: there is sufficient indication overall of prevention of contamination/co‐intervention

Methods	Trial design: 4‐armed, double‐blind, placebo‐controlled and stratified RCT  Location: UK  Number of centres: 7 secondary schools, Greater Manchester, UK  Recruitment period: study commenced in 1972
Participants	Inclusion criteria: not reported  Exclusion criteria: not reported  Baseline caries: 8.2 DMFS (Gp A: 7.83 DMFS (SD 5.17); Gp B: 8.48 DMFS (6.29)). Baseline characteristics (DMFS, DMFT, SAR) "balanced" (DFS baseline data not reported)  Age at baseline (years): 11 to 12 years (not reported by group)  Sex: 225 F:143 M (groups relevant to review: Gp A: 115 F:69 M; Gp B: 110 F:74 M) (evaluated participants only)  Any other details of important prognostic factors: no background exposure to fluoride was reported. Fluoride content of the water supplies was less than 0.10 mg/L  Number randomised: 826 (groups relevant to review: 410: Gp A: 205; Gp B: 205)  Number evaluated: 751 at 3 years (present at final assessment. Groups relevant to review: 368: Gp A: 184; Gp B: 184)  Attrition: 10% dropout after 3 years. Reasons for attrition described with respective total numbers for all 4 arms: 57 left school, 12 withdrawn by parents, 6 absent at final examination; no differential group losses: Gp A: 21/ 205; Gp B: 21/205
Interventions	Comparison: FT versus PL  Gp A (n = 205): SMFP 1000 ppm F; IMP (main abrasive) abrasive system; school use/supervised, daily, for 1 min, post‐brushing water rinse done. FT provided to all groups for use at home  Gp B (n = 205): placebo; IMP (main abrasive) abrasive system; school use/supervised, daily, for 1 min, post‐brushing water rinse done. FT provided to all groups for use at home
Outcomes	Primary: 3‐year net DFS increment ‐ (E+U) (CA) cl + (DR) xr; PF‐DFS; MD‐BL‐DFS; MD‐DFS; posterior MD‐DFS; DMFT; DMFT (U); anterior DMFT; posterior DMFT; DFS (U) (at 3 years follow‐up)  Secondary: none assessed  Assessments irrelevant to this review's scope: SAR  Follow‐up duration: 3 years
Notes	Adverse effects: not reported  Funding source: "supported from a grant from the Colgate‐Palmolive Company"  Declarations/conflicts of interest: none reported  Data handling by review authors: study contained 4 arms. Gp 1: fluoride paste + fluoride rinse; Gp 2: placebo paste + placebo rinse; Gp 3: fluoride paste + placebo rinse; Gp 4: placebo paste + fluoride rinse. Only Gps 2 (Gp B) and 3 (Gp A) are used in this review  Other information of note: clinical (V) caries assessment by 1 examiner, diagnostic threshold = CA. Radiographic assessment (1 postBW) by 1 examiner, diagnostic threshold = DR. State of tooth eruption included = E/U. Intra‐examiner reproducibility checks for incremental clinical and radiographic caries data in 10% sample (ICC score 0.9)
*Risk of bias*
Bias	Authors' judgement	Support for judgement
Random sequence generation (selection bias)	Unclear risk	Quote: "The children were allocated to four groups by stratified random sampling at two levels: school and dental age" Comment: not enough information provided
Allocation concealment (selection bias)	Unclear risk	No information provided
Blinding (performance bias and detection bias)   All outcomes	Low risk	Quotes: "The trial was organized on a double‐blind basis, neither the children nor the examiner being aware of who was receiving test or control products" and "...another group used the fluoride dentifrice..... and a fourth group....a placebo dentifrice" Comment: blind outcome assessment and use of placebo described
Incomplete outcome data (attrition bias)   All outcomes	Unclear risk	Overall dropout for length of follow‐up: 10% in 3 years. Dropout by group: 21/ 205 FT, 21/205 PL. Reasons for losses: left school (57), withdrawn by parents (12), absent at final examination (6) (not reported by group) Comment: numbers lost were not unduly high for the length of follow‐up with no differential loss between groups. It is unclear if reasons for missing outcome data are acceptable and balanced. Caries data used in analysis pertain to participants present at final examination
Selective reporting (reporting bias)	Low risk	Outcomes reported:  DFS increment (E+U) (CA) cl + (DR) xr, reported at 3 years follow‐up  PF‐DFS  MD‐BL‐DFS  MD‐DFS posterior MD‐DFS  DFS (U)  DMFT  anterior DMFT  posterior DMFT DMFT (U) Comment: trial protocol not available. All pre‐specified outcomes (in Methods) were reported and were reported in the pre‐specified way
Baseline characteristics balanced?	Low risk	Prognostic factors reported:  DMFT: 4.94 (2.86) FT, 5.26 (3.47) PL DMFS: 7.83 (5.17) FT, 8.48 (6.29) PL SAR: 93.41 (21.30) FT, 93.61 (20.43) PL Comment: initial caries appears balanced between groups
Free of contamination/co‐intervention?	Low risk	Quote: "..both dentifrice tubes and rinse bottles were colour coded so that the children received the correct products. Independent laboratory checks of the dispensed rinse and dentifrice were made at regular intervals to assess the reliabilty of the supervisors who dispensed agents. The coded dentifrice and rinse was dispensed in the school..." Comment: there is sufficient indication overall of prevention of contamination/co‐intervention

Methods	Trial design: 5‐armed, double‐blind, placebo/active‐controlled and stratified RCT  Location: USA  Number of centres: not reported, only that the sample derived from a "large school population"  Recruitment period: study commenced 1961
Participants	Inclusion criteria: not reported  Exclusion criteria: not reported  Baseline caries: 15.4 DFS (Gp A: 16.88 DFS (SD 12.81); Gp B: 14.03 DFS (SD 10.16); Gp C: 14.89 DFS (SD 10.19); Gp D: 15.70 DFS (SD 10.96)). Baseline characteristics (DFS, DFT, DMFS, DMFT) "balanced"  Age at baseline (years): range 7 to 16 years. Baseline characteristic (dental age) "balanced"  Sex: 650 F:628 M (Gp A: 168 F:167 M; Gp B: 164 F:177 M; Gp C: 150 F:129 M; Gp D: 168 F:155 M). Baseline characteristic (sex) "balanced"  Any other details of important prognostic factors: data unavailable for site fluoridation status  Number randomised: 2156 (numbers for relevant groups not reported)  Number evaluated: 1278 at 2 years (present at all assessments. relevant groups to this review: Gp A: 335; Gp B: 341; Gp C: 279; Gp D: 323)  Attrition: 25% dropout after 2 years. Reasons for attrition not reported; any differential group losses not assessable
Interventions	Comparison: FT (3 groups)^a versus PL  Gp A (n = evaluated 335): SnF₂ 1000 ppm F; Ca pyrophosphate abrasive system; home use/unsupervised, daily frequency assumed (GP 1)  Gp B (n = evaluated 341): SnF₂ 1000 ppm F; IMP abrasive system; home use/unsupervised, daily frequency assumed (GP 4) Gp C (n = evaluated 279): APF 1000 ppm F; IMP abrasive system; home use/unsupervised, daily frequency assumed (GP 3)  Gp D (n = evaluated 323): placebo; abrasive system not reported; home use/unsupervised, daily frequency assumed (GP 2)
Outcomes	Primary: 2‐year DFS increment ‐ cl + xr; DFT; DMFS; DMFT (at 2 years follow‐up)  Secondary: none assessed  Assessments irrelevant to this review's scope: none assessed  Follow‐up duration: 2 years
Notes	Adverse effects: not reported  Funding source: "supported in part by the Bristol‐Myers company, New York" manufacturers of toothpastes used by Gps B (experimental formula, Bristol‐Myers) and C (Ipana Durenamel, Bristol‐Myers)  Declarations/conflicts of interest: not reported  Data handling by review authors: ^aNaF‐secondary Ca pyrophosphate toothpaste Gp E not considered (abrasive system known to be incompatible with NaF). Adjusted mean values and SEMs used in analysis. Gps A, B and C combined versus Gp D  Other information of note: clinical (VT) caries assessment by 2 examiners; diagnostic threshold = CA. Radiographic assessment (10 BW) by 1 examiner; diagnostic threshold not reported. State of tooth eruption included not reported. Diagnostic errors not reported
*Risk of bias*
Bias	Authors' judgement	Support for judgement
Random sequence generation (selection bias)	Unclear risk	Quote: "At the initial exam, the record cards of these youngest, or master, siblings were stratified (ordered) simultaneously according to 12 characteristics....The ordered cards of the 'master' siblings were then divided into 5 dentifrice groups by superimposing the numbers 1 through 5 in random sequence. The same dentifrice was assigned automatically to the other, or "trailing", siblings in his household" Comment: not enough information provided
Allocation concealment (selection bias)	Unclear risk	No information provided
Blinding (performance bias and detection bias)   All outcomes	Low risk	Quotes: "Each (of 2) examiner assessed about half of the subjects in each group, and each subject had the same dentist‐examiner throughout the study. Separate records were used for each examination, and previous records were never available to the examiner. All observations were recorded in code for subsequent transfer to machine data processing. The radiographs were read and recorded independently by a third dentist. At no time was it possible for the examiners to identify a subject with a dentifrice group" and "An independent laboratory was assigned the responsibility of coding, packaging, and shipping all dentifrices in this study... NaF dentifrice was compared to.....and a fluoride free dentifrice" Comment: blind outcome assessment and use of placebo described
Incomplete outcome data (attrition bias)   All outcomes	Unclear risk	Overall dropout for length of follow‐up: 24.7% in 2 years (534/2156, all 5 groups combined). Dropout by group: not reported. Reasons for losses: not reported Comment: numbers lost were not unduly high given length of follow‐up; It is unclear if there were any differential losses, and if reasons for missing outcome data are acceptable and balanced. Caries data used in the analysis pertain to participants examined after 2 years
Selective reporting (reporting bias)	Low risk	Outcomes reported:  DFS increment ‐ cl + xr, reported at 2 years follow‐up  DMFS  DMFT  DFT Comment: trial protocol not available. All pre‐specified outcomes were reported and were reported in the pre‐specified way
Baseline characteristics balanced?	Low risk	Prognostic factors reported:  DFS: 16.88 (12.81) FT 1; 14.03 (10.16) FT 2; 14.89 (10.19) FT 3; 15.70 (10.96) PL DFT: 8.53 (5.49) FT 1; 7.61 (4.80) FT 2; 6.04 PL; 7.59 (5.01) FT 3; 8.07 (5.02) PL DMFT: 8.84 (5.86) FT 1; 7.87 (4.80) FT 2; 2.94 PL; 7.91 (5.34) FT 3; 8.35 (5.21) PL DMFS: 18.43 (13.91) FT 1; 15.33 (11.08) FT 2; 16.48 (12.86) FT 3; 17.09 (11.68) PL dental age: 21.12 (6.59) FT 1; 22.28 (6.47) FT 2; 20.49 (6.51) FT 3; 21.70 (6.29) PL Comment: initial caries appears balanced although adjustment for baseline imbalance was made in the analysis
Free of contamination/co‐intervention?	Low risk	Quotes: "As the study group was assembled, all siblings were noted to permit limitiation of one dentifrice code to a family" and "New shipments supplied every 8 to 10 weeks, and new toothbrushes supplied every 6 months" Comment: there is sufficient indication overall of prevention of contamination/co‐intervention

Outcome or subgroup title	No. of studies	No. of participants	Statistical method	Effect size
1 D(M)FS closest to 3 years	2	816	Std. Mean Difference (IV, Random, 95% CI)	‐0.07 [‐0.21, 0.06]
2 D(M)FT	2	816	Std. Mean Difference (IV, Random, 95% CI)	‐0.28 [‐0.72, 0.15]

Outcome or subgroup title	No. of studies	No. of participants	Statistical method	Effect size
1 D(M)FS	55	31819	Std. Mean Difference (IV, Random, 95% CI)	‐0.28 [‐0.33, ‐0.24]
2 D(M)FT	41	24837	Std. Mean Difference (IV, Random, 95% CI)	‐0.26 [‐0.31, ‐0.21]
3 Proportion developing new caries	7	1898	Risk Ratio (M‐H, Random, 95% CI)	0.90 [0.77, 1.06]

Outcome or subgroup title	No. of studies	No. of participants	Statistical method	Effect size
1 D(M)FS	4	4406	Std. Mean Difference (IV, Random, 95% CI)	‐0.34 [‐0.51, ‐0.18]
2 D(M)FT	4	4406	Std. Mean Difference (IV, Random, 95% CI)	‐0.37 [‐0.50, ‐0.24]
3 Proportion developing new caries	1		Risk Ratio (M‐H, Random, 95% CI)	Subtotals only

Outcome or subgroup title	No. of studies	No. of participants	Statistical method	Effect size
1 D(M)FS closest to 3 years	3	1989	Std. Mean Difference (IV, Random, 95% CI)	‐0.84 [‐1.29, ‐0.38]
2 D(M)FT	2	1209	Std. Mean Difference (IV, Random, 95% CI)	‐1.57 [‐3.96, 0.82]

Outcome or subgroup title	No. of studies	No. of participants	Statistical method	Effect size
1 D(M)FS	1	766	Std. Mean Difference (IV, Random, 95% CI)	‐0.01 [‐0.15, 0.13]
2 D(M)FT	1		Std. Mean Difference (IV, Random, 95% CI)	Subtotals only

Outcome or subgroup title	No. of studies	No. of participants	Statistical method	Effect size
1 D(M)FS	6	3047	Std. Mean Difference (IV, Random, 95% CI)	‐0.11 [‐0.20, ‐0.02]
2 D(M)FT	3	2087	Std. Mean Difference (IV, Random, 95% CI)	‐0.16 [‐0.27, ‐0.05]
3 Proportion developing new caries	2	676	Risk Ratio (M‐H, Random, 95% CI)	1.00 [0.63, 1.59]

Outcome or subgroup title	No. of studies	No. of participants	Statistical method	Effect size
1 D(M)FS	7	13094	Std. Mean Difference (IV, Random, 95% CI)	‐0.10 [‐0.14, ‐0.05]
2 D(M)FT	4		Std. Mean Difference (IV, Random, 95% CI)	Totals not selected
3 Proportion developing new caries	2	4328	Risk Ratio (M‐H, Random, 95% CI)	1.02 [0.93, 1.11]

Outcome or subgroup title	No. of studies	No. of participants	Statistical method	Effect size
1 D(M)FS	5	14196	Std. Mean Difference (IV, Random, 95% CI)	‐0.02 [‐0.05, 0.01]
2 D(M)FT	1		Std. Mean Difference (IV, Random, 95% CI)	Subtotals only

Outcome or subgroup title	No. of studies	No. of participants	Statistical method	Effect size
1 D(M)FS	6	10553	Std. Mean Difference (IV, Random, 95% CI)	‐0.08 [‐0.12, ‐0.04]
2 D(M)FT	3	7044	Std. Mean Difference (IV, Random, 95% CI)	‐0.06 [‐0.12, 0.00]

Methods	Trial design: 3‐armed, double‐blind, placebo‐controlled and stratified RCT  Location: USA  Number of centres: 14 schools in San Francisco Bay area  Recruitment period: study commenced in/before 1976
Participants	Inclusion criteria: not reported  Exclusion criteria: not reported  Baseline caries: 3.2 DFS (Gp A: 3.34 DFS; Gp B: 3.06 DFS; Gp C: 3.13 DFS; "balanced")  Age at baseline (years): range 5 to 12 years. Mean not reported  Sex: not reported  Any other details of important prognostic factors: background exposure to fluoride not reported. Natural fluoride level of community water supply ranged from 0.015 to 0.093 ppm F  Number randomised: 2210 (Gp A: 740; Gp B: 721; Gp C: 749)  Number evaluated: 1141 at 3 years (present at final assessment: Gp A: 383; Gp B: 378; Gp C: 380)  Attrition: 48% dropout after 3 years. Reasons for high dropout described: change of residence, absenteeism, non‐adherence to study protocol; no differential group losses
Interventions	Comparison: FT (2 groups) versus PL   Gp A (n = 740): SnF₂ 1000 ppm F; silica gel abrasive system; home use/unsupervised, daily frequency assumed  Gp B (n = 721): SnF₂ 1000 ppm F; Ca pyrophosphate abrasive system; home use/unsupervised, daily frequency assumed Gp C (n = 749): placebo; silica gel abrasive system; home use/unsupervised, daily frequency assumed
Outcomes	Primary: 3‐year net DFS increment ‐ (E + U) (CA) cl + (ER) xr; MD‐DFS; DFS rate; DFT; DFT rate; DMFS; DMFT (at 1, 2 and 3 years follow‐up)  Secondary: none assessed  Assessments irrelevant to this review's scope: none assessed  Follow‐up duration: 3 years
Notes	Adverse effects: not reported  Funding source: "grant from the Lever Brothers Company", Lever manufacture Aim (Gps A, and C ‐ without fluoride)  Declarations/conflicts of interest: not reported  Data handling by review authors: combined 2 SnF₂ 1000 groups with different abrasive systems  Other information of note: clinical (VT) caries assessment by 2 examiners, diagnostic threshold = CA. Radiographic assessment (postBW) by 2 examiners, diagnostic threshold = ER. State of tooth eruption included = E/U. Intra‐ and inter‐examiner reproducibility of clinical caries diagnosis (DFS) assessed annually by duplicate examination of 10% random sample (percentage of times diagnosis replicated in all 3 examinations ranged 42% to 97% and 77% to 92% for both examiners and for each respectively)
*Risk of bias*
Bias	Authors' judgement	Support for judgement
Random sequence generation (selection bias)	Unclear risk	Quote: "Children were randomly assigned to one of 3 treatment groups. A stratified sequential sampling technique was used within each school to balance the sample size with respect to sex and grade level for each dentifrice" Comment: not enough information provided
Allocation concealment (selection bias)	Unclear risk	No information provided
Blinding (performance bias and detection bias)   All outcomes	Low risk	Quotes: "A 3 year double‐blind study of a dentifrice containing 0.4% stannous fluoride and a placebo..." "The examiners at all times were unaware of the children's dentifrice assignment" Comment: blind outcome assessment and use of placebo described
Incomplete outcome data (attrition bias)   All outcomes	Unclear risk	Overall dropout for length of follow‐up: 48% in 3 years. Dropout by group: 357/740 FT1, 343/721 FT2, 369/749 PL. Reasons for losses: change of residence, absenteeism, exclusion due to non‐adherence to study requirements Comment: numbers lost were high for the length of follow‐up. No differential loss between groups. It is unclear if reasons for missing outcome data are acceptable and balanced. Caries data used in analysis pertain to participants present at final examination
Selective reporting (reporting bias)	Low risk	Outcomes reported:  DFS increment ‐ (E + U) (CA) cl + (ER) xr, reported at 1, 2 and 3 years follow‐ups  DMFT  DMFS  DFT  MD‐DFS  DFT rate  DFS rate Comment: trial protocol not available. All pre‐specified outcomes (in Methods) were reported and were reported in the pre‐specified way
Baseline characteristics balanced?	Low risk	Prognostic factors reported: DFS: 2.90 FT1, 3.28 FT2, 2.94 PL Comment: initial caries appears balanced
Free of contamination/co‐intervention?	Low risk	Quotes: "Provisions were made ensuring the randomization process to assure that only one dentifrice code would be available in each household.... A letter to parents was attached, giving brushing instructions and urging use of only the assigned dentifrice" Comment: there is sufficient indication overall of prevention of contamination/co‐intervention

PERMALINK

Fluoride toothpastes of different concentrations for preventing dental caries

Tanya Walsh

Helen V Worthington

Anne‐Marie Glenny

Valeria CC Marinho

Ana Jeroncic

Abstract

Background

Objectives

Search methods

Selection criteria

Data collection and analysis

Main results

Authors' conclusions

Plain language summary

Summary of findings

Background

Description of the condition

Description of the intervention

How the intervention might work

Why it is important to do this review

Objectives

Methods

Criteria for considering studies for this review

Types of studies

Types of participants

Types of interventions

Types of outcome measures

Primary outcomes

Secondary outcomes

Search methods for identification of studies

Electronic searches

Searching other resources

Data collection and analysis

Selection of studies

Data extraction and management

Assessment of risk of bias in included studies

1.

2.

Measures of treatment effect

Unit of analysis issues

Studies with multiple treatment arms

Cluster‐randomised trials

Dealing with missing data

Assessment of heterogeneity

Clinical and methodological heterogeneity within treatment comparisons

Transitivity across treatment comparisons

Assessment of reporting biases

Data synthesis

Methods for direct treatment comparisons

Methods for indirect and mixed comparisons

Subgroup analysis and investigation of heterogeneity

Assumptions when estimating heterogeneity

Measures and tests for heterogeneity

Assessment of statistical inconsistency

Investigation of heterogeneity and inconsistency

Sensitivity analysis

Presentation of main results

Direct comparisons

NMA comparisons from the connected network

Results

Description of studies

Results of the search

3.

Included studies

Excluded studies

Risk of bias in included studies

Allocation

Method of randomisation

Allocation concealment

Blinding

Incomplete outcome data

Selective reporting

Other potential sources of bias

Baseline characteristics

Free of contamination or co‐intervention

Overall risk of bias

Effects of interventions

Summary of findings for the main comparison. Fluoride toothpaste with higher versus lower concentration of fluoride for preventing caries in young children (primary dentition).

Methods	Trial design: 4‐armed, triple‐blind, active‐controlled and stratified RCT  Location: USA  Number of centres: not reported  Recruitment period: study commenced 1983
Participants	Inclusion criteria: not reported  Exclusion criteria: orthodontic appliances; unsuitable medical history  Baseline caries: not reported for all groups (Gp A: mean 1.93 DMFS (SEM 0.075); Gp B: not reported; Gp C: mean 1.87 DMFS (SEM 0.072); Gp D: not reported). Baseline characteristics similar according to DMFS (reported for SMFP 1100 ppm F group and NAF 1100 ppm F group)  Age at baseline (years): 6 to 16 years (Gp A: mean 9.07 years; Gp B: not reported; Gp C: 9.02 years; Gp D: not reported). Baseline characteristics similar according to age (reported for SMFP 1100 ppm F group and NAF 1100 ppm F group)  Sex: Gp A: 841 F:803 M; Gp B: not reported; Gp C: 831 F:815 M; Gp D: not reported  Any other details of important prognostic factors: background exposure to fluoride in drinking water, containing approximately 0.5 ppm F from natural sources  Number randomised: 3290 (1100 ppm Gps only. Not reported for 1700 ppm Gps). Gp A: 1644; Gp B: not reported; Gp C: 1646; Gp D: not reported  Number evaluated: 4458 at 3 years (present at final assessment. Gp A: 1122; Gp B: 1109; Gp C: 1151; Gp D: 1076)  Attrition: overall percentage dropout after 3 years not calculable as numbers randomised at baseline not presented for 1700 ppm groups. Percentage dropout for 1100 ppm F groups 1017/3290 = 31%
Interventions	Comparison: FT versus FT  Gp A (n = 1644): SMFP 1100 ppm F; hydrated silica abrasive system; home use (unsupervised) and school use (not clear whether supervised brushing), at least once a day  Gp B (n = not reported): SMFP 1700 ppm F; hydrated silica abrasive system; home use (unsupervised) and school use (not clear whether supervised brushing), at least once a day Gp C (n = 1646): NaF 1100 ppm F; hydrated silica abrasive system; home use (unsupervised) and school use (not clear whether supervised brushing), at least once a day  Gp D (n = not reported): NaF 1700 ppm F; hydrated silica abrasive system; home use (unsupervised) and school use (not clear whether supervised brushing), at least once a day
Outcomes	Primary: 3‐year net DMFS increment ‐ (CA) cl + DR; DMFS (at 2 and 3 years follow‐ups)  Secondary: none assessed  Assessments irrelevant to this review's scope: mean number of sealed occlusal surfaces after 3 years  Follow‐up duration: 3 years
Notes	Adverse effects: not reported  Funding source: not reported, however intervention (Gp C: Crest toothpaste) is manufactured by Procter & Gamble, therefore probable trial was funded by the manufacturer as a co‐author is also employed by them  Declarations/conflicts of interest: not reported  Data handling by review authors: Gps A + C, and Gps B + D combined in analyses  Other information of note: additional information provided in Bartizek 2001. Caries assessment determined by visual‐tactile examinations according to Radike criteria by single examiner and supplemented with radiographs. Diagnostic threshold not stated, CA assumed
*Risk of bias*
Bias	Authors' judgement	Support for judgement
Random sequence generation (selection bias)	Unclear risk	Quote: "Within each stratum, the subjects were randomly assigned to one of the two dentifrice groups" Comment: random sequence generation not stated
Allocation concealment (selection bias)	Unclear risk	No information provided
Blinding (performance bias and detection bias)   All outcomes	Low risk	Quotes: "Both dentifrices were formulated using the same …. And were essentially identical with regard to common excipients" and "At no time during the course of the study did the subjects or the examiner know to which dentifrice group any subject was assigned" Comment: investigators and participants blinded
Incomplete outcome data (attrition bias)   All outcomes	High risk	Comment: overall dropout for length of follow‐up: not calculable as numbers randomised not provided for all groups. Caries data used in analysis pertain to participants present at final examination
Selective reporting (reporting bias)	Low risk	Outcomes reported: DMFS increment ‐ (CA) cl (DR) xr, reported at 24 and 36 month follow‐up mean number of reversals after 3 years mean number of sealed occlusal surfaces after 3 years Comment: trial protocol not available. Fluoride concentration for 1 group not reported so unable to include in meta‐analysis
Baseline characteristics balanced?	Unclear risk	Prognostic factors reported: age, caries Comment: all appear balanced but only reported for 2 of the 4 FT groups
Free of contamination/co‐intervention?	Low risk	Quote: "Care was exercised to ensure that siblings were assigned to the same treatment group to avoid having different dentifrices in the same household" Comment: contamination not observed

Methods	Trial design: 4‐armed, double‐blind, active‐controlled and stratified RCT  Location: USA  Number of centres: not reported. Sample comprised of elementary schoolchildren resident in urban central Ohio  Recruitment period: study commenced in/before 2001
Participants	Inclusion criteria: not reported  Exclusion criteria: conditions preventing thorough oral examination (e.g. orthodontic or extensive prosthetic appliances)  Baseline caries: 5.3 DMFS (Gp A: 5.29 DMFS (SEM 0.184); Gp B: 5.49 DMFS (SEM 0.181); Gp C: 5.19 DMFS (SEM 0.178); Gp D: 5.39 DMFS (SEM 0.180)). Baseline characteristics (DMFS, DMFT) "balanced"  Age at baseline (years): range 6 to 15 years; mean 9.5 years (Gp A: 9.47; Gp B: 9.47; Gp C: 9.47; Gp D: 9.50). Baseline characteristics (age) "balanced"  Sex: overall: 2777 F:2662 M (Gp A: 702 F:659 M; Gp B: 685 F:675 M; Gp C: 702 F:657 M; Gp D: 688 F:671 M)  Any other details of important prognostic factors: background exposure to fluoride: community water supply < 0.3 ppm F  Number randomised: 5439 (Gp A: 1361; Gp B: 1360; Gp C: 1359; Gp D: 1359)  Number evaluated: 4431 at 1 year (present at 1 year assessment. Gp A: 1127; Gp B: 1129; Gp C: 1082; Gp D: 1093)  Attrition: 18.5% dropout (for all groups combined) after 1 year. Reasons for attrition not reported
Interventions	Comparison: FT versus FT (4 groups)  Gp A (n = 1361): NaF 1100 ppm F; silica abrasive system; home use/unsupervised, daily frequency assumed  Gp B (n = 1360): NaF 1700 ppm F; silica abrasive system; home use/unsupervised, daily frequency assumed Gp C (n = 1359): NaF 2200 ppm F; silica abrasive system; home use/unsupervised, daily frequency assumed Gp D (n = 1359): NaF 2800 ppm F; silica abrasive system; home use/unsupervised, daily frequency assumed
Outcomes	Primary: 1‐year net DMFS increment cl + xr, reported at 1 year follow‐up; DMFS increment by surface (at 1, 2, and 3 years follow‐ups); DMFT increment (at 1 year follow‐up)  Secondary: none assessed  Assessments irrelevant to this review's scope: none assessed  Follow‐up duration: 3 years
Notes	Adverse effects: not reported  Funding source: "supported by the Procter & Gamble Company"  Declarations/conflicts of interest: first author is employed by Procter & Gamble  Data handling by review authors: only results for 1 year follow‐up analysed in review. Results at years 2 and 3 confounded by a concurrent fluoride rinse programme, which involved half of the study population. Gps A versus B + C versus D in analysis  Other information of note: clinical (VT) caries assessment and radiographic assessment carried out by a single examiner established as "repeatably sensitive" based on prior trial experience. Additional information published in Bartizek 2001
*Risk of bias*
Bias	Authors' judgement	Support for judgement
Random sequence generation (selection bias)	Unclear risk	Quote: "randomly assigned to one of the four dentifrice groups"
Allocation concealment (selection bias)	Unclear risk	Insufficient information provided
Blinding (performance bias and detection bias)   All outcomes	Low risk	Quotes: "...double‐blind study..." "They [dentifrices] were supplied in plain white 2.7 oz tubes" and "Subject and examiner blindness to treatment were maintained throughout the study"
Incomplete outcome data (attrition bias)   All outcomes	Unclear risk	Comment: 38% dropout at 3 years. No reasons are given for those not examined but similar attrition rate in each of the 4 groups. Not stated but assumed that ITT analysis carried out for those present at exam. No imputation carried out
Selective reporting (reporting bias)	Unclear risk	Comment: unclear but DMFS data not presented by surface for years 2 and 3, unlike year 1
Baseline characteristics balanced?	Low risk	Quote: "..well balanced with respect to ....mean caries experience as measured by DMFS and DMFT at baseline"
Free of contamination/co‐intervention?	High risk	Quote: "Results at years 2 and 3 confounded by a concurrent fluoride rinse programme"  Comment: after 1 year schools participated to varying degrees in a fluoride rinse programme. Only results for 1 year follow‐up analysed in review

Methods	Trial design: 2/3‐armed, placebo/active‐controlled and stratified RCT (placebo‐controlled until 9 months when placebo participants re‐allocated to active groups)  Location: Guatemala  Number of centres: not reported. Sample comprised of elementary schoolchildren resident in an urban area of Guatemala  Recruitment period: not reported
Participants	Inclusion criteria: not reported  Exclusion criteria: conditions preventing thorough oral examination (e.g. orthodontic or extensive prosthetic appliances)  Baseline caries: 7.71 DMFS (Gp A: 8.02 (SD 5.61); Gp B: 9.14 (SD 6.30); Gp C: 7.95 (SD 6.10); Gp D: 7.47 (SD5.67)). Evaluated participants at 21 months reported. Baseline DMFS appears balanced  Age at baseline (years): range 9 to 12 years; mean 10.4 years (Gp A: 10.4 (SEM 0.07); Gp B: 10.4 (0.07); Gp C:10.4 (SEM 0.08))  Sex: 214 F:430 M (Gp A: 72 F:144 M; Gp B: 70 F:142 M; Gp C: 72 F:144 M)  Any other details of important prognostic factors: background exposure to fluoride in community water < 0.3 ppm F  Number randomised: 644 (Gp A: 216; Gp B: 212; Gp C: 216)  Number evaluated: 494 at 21 months (Gp A: 83; Gp B: 90; Gp C: 168; Gp D: 153)  Attrition: 23.3% dropout after 21 months, reasons for dropout reported. Highest dropout observed in highest fluoride concentration group
Interventions	Comparison: FT versus FT Gp A (n = not reported): placebo ‐ NaF 1100 ppm F; hydrated silica abrasive system; home use (unsupervised) with supervised brushing at school twice daily  Gp B (n = not reported): placebo ‐ NaF 2800 ppm F; hydrated silica abrasive system; home use (unsupervised) with supervised brushing at school twice daily Gp C (n = 216): NaF 1100 ppm F; hydrated silica abrasive system; home use (unsupervised) with supervised brushing at school twice daily  Gp D (n = 212): NaF 2800 ppm F; hydrated silica abrasive system; home use (unsupervised) with supervised brushing at school twice daily
Outcomes	Primary: 21‐month net DMFS increment ‐ (CA) cl + (DR) xr; DMFS‐O; DMFS‐BL; DMFS‐MD (at 9 and 21 months follow‐up)  Secondary: none assessed  Assessments irrelevant to this review's scope: none assessed  Follow‐up duration: 21 months
Notes	Adverse effects: not reported  Funding source: "Support for this study was provided by the Procter & Gamble Company"  Declarations/conflicts of interest: 4 of 5 authors (AR Biesbrock, RW Gerlach, SA Jacobs, and RD Bartizek) are employed by Procter & Gamble. L Archila is an academic at University of Texas Health Science Center Dental School, Texas, USA  Data handling by review authors: after 9 months the participants in the placebo group were randomised to the fluoride groups. The data used for analysis are for the 2 active intervention groups receiving the same fluoride concentration for the duration of the study (Gps C and D). Results reported for 3 examiners. Integrated results (VT + radiograph) reported for Examiner A used for analysis  Other information of note: clinical (VT) and radiographic (DR) caries assessment by 3 trained and calibrated examiners according to a modification of the Radike criteria, diagnostic threshold = CA. Weighted Kappa scores of 0.77, 0.94, 0.96 for VT on 50 children
*Risk of bias*
Bias	Authors' judgement	Support for judgement
Random sequence generation (selection bias)	Unclear risk	Quote: "and randomly assigned…" Comment: random sequence generation not stated
Allocation concealment (selection bias)	Unclear risk	No information provided
Blinding (performance bias and detection bias)   All outcomes	Low risk	Quotes: "Subject and examiner blindness to treatment were maintained throughout the study" and "Dentifrice products were labeled with the subject's name and a unique identification number" Comment: participants and examiners blinded
Incomplete outcome data (attrition bias)   All outcomes	Low risk	Comment: overall dropout rate of 23.3% with slight difference in rate (20% for initial placebo group and 23% and 29% for fluoride groups). Caries data used in analysis pertain to participants present at final examination
Selective reporting (reporting bias)	Low risk	Outcomes reported: DMFS increment ‐ (CA) cl + (DR) xr, reported at 9 and 21 months follow‐ups Comment: trial protocol not available. All outcomes listed in Methods section were reported (DMFS)
Baseline characteristics balanced?	Low risk	Prognostic factors reported: age, sex, caries Comment: baseline DMFS appears balanced
Free of contamination/co‐intervention?	Unclear risk	Quote: "Siblings residing in the same household were automatically assigned to minimise risk from cross‐usage of assigned dentifrice" Comment: contamination not observed

Methods	Trial design: initially 3‐armed, placebo/active‐controlled and stratified RCT (placebo‐controlled until 9 months when placebo participants re‐allocated to active intervention groups)  Location: Guatemala  Number of centres: 2 elementary schools, in urban Guatemala  Recruitment period: not reported
Participants	Inclusion criteria: not reported  Exclusion criteria: conditions preventing thorough oral examination (e.g. orthodontic or extensive prosthetic appliances)  Baseline caries: 9.8 DMFS (Gp A: 9.64 (SD 6.73); Gp B: 10.27 DMFS (SD 6.39); Gp C: 9.56 DMFS (SD 6.56); Gp D: 9.96 (SD 5.30)). Evaluated participants at 21 months reported. Baseline DMFS appears balanced  Age at baseline (years): 9 to 12 years; mean 10.3 years (Gp A: 10.3; Gp B: 10.2; Gp C: 10.3; Gp D 10.3)  Sex: 289 F:245 M (Gp A: 51 F:46 M; Gp B: 51 F:37 M; Gp C: 89 F:80 M; Gp D: 98 F:82 M)  Any other details of important prognostic factors: background exposure to fluoride in community water < 0.3 ppm F  Number randomised: 657 (Gp A: 219; Gp B: 218; Gp C: 220)  Number evaluated: 534 at 21 months (present at final assessment. Gp A: 266 (original allocation: 169; Gp C re‐allocation: 97); Gp B: 268 (original allocation: 180; Gp C re‐allocation: 88); Gp C: participants re‐randomised after 9 months and allocated between Gps A and B)  Attrition: 19.8% dropout after 21 months, reasons for dropout not reported. No differential dropout
Interventions	Comparison: FT versus FT   Gp A (n = not reported): placebo ‐ NaF 500 ppm F; hydrated silica abrasive system; home use (unsupervised) with supervised brushing at school twice daily Gp B (n = not reported): placebo ‐ NaF 1450 ppm F; hydrated silica abrasive system; home use (unsupervised) with supervised brushing at school twice daily Gp C (n = 219): NaF 500 ppm F; hydrated silica abrasive system; home use (unsupervised) with supervised brushing at school twice daily Gp D (n = 218): placebo ‐ NaF 1450 ppm F; hydrated silica abrasive system; home use (unsupervised) with supervised brushing at school twice daily
Outcomes	Primary: 21‐month net DMFS increment ‐ (CA) cl + DR (xr); DMFS‐O; DMFS‐BL; DMFS‐MD (at 9 and 21 months follow‐up)  Secondary: none assessed  Assessments irrelevant to this review's scope: none assessed Follow‐up duration: 21 months
Notes	Adverse effects: not reported  Funding source: "Support for this study was provided by the Procter & Gamble Company"  Declarations/conflicts of interest: 4 authors are employed by Procter & Gamble  Data handling by review authors: after 9 months the participants in the placebo group were randomised to the fluoride groups. The data used for analysis are for the 2 active intervention groups receiving the same fluoride concentration for the duration of the study. Results reported for 3 examiners. Integrated results (VT + radiograph) reported for Examiner A used for analysis  Other information of note: clinical (VT) and radiographic (DR) caries assessment by a trained and calibrated examiner according to a modification of the Radike criteria, diagnostic threshold = CA. Weighted Kappa scores of 0.77, VT on 20 children. Sensitivity of 98%, specificity of 91% relative to expert consensus panel for radiographic interpretation
*Risk of bias*
Bias	Authors' judgement	Support for judgement
Random sequence generation (selection bias)	Unclear risk	Quote: ".. and randomly assigned…" Comment: random sequence generation not stated
Allocation concealment (selection bias)	Unclear risk	No information provided
Blinding (performance bias and detection bias)   All outcomes	Low risk	Quotes: "Subject and examiner blindness to treatment were maintained throughout the study" and "Each tube was labeled with the subject's name and a unique identification number" Comment: participants and examiners blinded
Incomplete outcome data (attrition bias)   All outcomes	Low risk	Comment: overall dropout rate of 18.8% with slight difference in rate (15.9% for initial placebo group and 22.8% and 17.4% for fluoride groups. Caries data used in analysis pertain to participants present at final examination
Selective reporting (reporting bias)	Low risk	Outcomes reported: DMFS increment ‐ (CA) cl + (DR) xr, reported at 9 and 21 months follow‐up Comment: trial protocol not available. All outcomes listed in Methods section were reported (DMFS)
Baseline characteristics balanced?	Low risk	Prognostic factors reported: age, sex, caries Comment: baseline DMFS appears balanced
Free of contamination/co‐intervention?	Unclear risk	Quote: "Siblings residing in the same household were assigned to the same treatment group to eliminate the potential for cross‐usage of test dentifrices" Comment: contamination not observed

Methods	Trial design: 3‐armed, active‐controlled and stratified RCT  Location: Brazil  Number of centres: 5 public primary schools in Bauru (each in a different geographical region of the city), Brazil  Recruitment period: study commenced in 2010
Participants	Inclusion criteria: (from NCT01049503) ≥ 2 and ≤4 years old; absence of very large carious lesions/dentine sensitivity during the study  Exclusion criteria: participation in other trials in prior 3 months; orthodontic appliance use  Baseline caries: 0.947 ANC; (Gp A: 1.14; Gp B 0.84)  Age at baseline (years): range 2 to 4 years; mean 3.4 (SD 0.6) (distribution of mean age between groups not reported)  Sex: 167 F:148 M (distribution of sex between groups not reported)  Any other details of important prognostic factors: background exposure to fluoride in community water supply (0.6 to 0.8 ppm F)  Number randomised: 315 (Gp A: 104; Gp B: 211)  Number evaluated: 195 at 1 year (present at final assessment. Gp A: 71; Gp B: 124)  Attrition: 38% dropout after 1 year, reasons for dropout reported. No differential dropout
Interventions	Comparison: FT versus FT  Gp A (n = 104): NaF 1100 ppm F (pH 7.0); abrasive system not reported; home use (unsupervised) with supervised brushing at school at least twice daily Gp B (n = 211): NaF 550 ppm F (pH 4.5 (n = 104) and 7.0 (n = 107)); abrasive system not reported; home use (unsupervised) with supervised brushing at school at least twice daily
Outcomes	Primary: 12‐month net white spot lesion increment ‐ (ANC) cl (at 1 year)  Secondary: adverse effects  Assessments irrelevant to this review's scope: compliance; change in fluorescence; lesion area (mm)  Follow‐up duration: 1 year
Notes	Adverse effects: "There were no reports on adverse effects, but some children complained about the taste of the dentifrice"  Funding source: "This study was funded by FAPESP (2008/58402‐9, 2010/11916‐8, and 2010/01944‐4)." FAPESP is São Paulo's state‐funded research foundation. "...experimental formulations, manufactured by Dental Prev Ind. Com. Ltda (Lorena, Brazil)"  Declarations/conflicts of interest: "University of São Paulo has a patent request in Brazil (INPI) for 'Low‐fluoride acidic liquid dentifrice and its use'"  Data handling by review authors: data entered separately for caries‐active and caries‐inactive at baseline. Both 550 ppm F arms combined in this review under Gp B  Other information of note: clinical (VT) caries assessment by 2 examiners, diagnostic threshold = NCA. Kappa scores of 0.88 and 0.78 for intra‐rater reliability and 0.78 for inter‐rater reliability at baseline on 20 children and at 0.84 and 0.75 at the end of the study 0.75 inter‐rater reliability
*Risk of bias*
Bias	Authors' judgement	Support for judgement
Random sequence generation (selection bias)	Low risk	Quotes: "was randomly reassigned into three subgroups…" and "The software Excel generated random numbers ranging from 0 to 0.99" Comment: random sequence generation stated
Allocation concealment (selection bias)	Unclear risk	No information provided
Blinding (performance bias and detection bias)   All outcomes	Low risk	Quote: "All formulations were NaF‐based and identical, except for the dentifrices' color. The colors' code was kept by one person not involved in the examination of the children" Comment: probably done. Participants and examiners blinded
Incomplete outcome data (attrition bias)   All outcomes	Low risk	Quote: "38% loss, mainly due to the fact that parents changed their children's school" Comment: overall dropout rate of 38.1% with no differential dropout. Reasons for dropout reported. Caries data used in analysis pertain to participants present at final examination
Selective reporting (reporting bias)	Low risk	Outcomes reported: progression of caries in caries inactive group (lesions/child) net increment for children in caries active group ‐ (NCA)cl, reported at 12‐month follow‐up Comment: all outcomes listed in Methods section were reported
Baseline characteristics balanced?	Unclear risk	Baseline caries unbalanced in the caries active group mean 1.55 active white spot lesions in the 550 ppm group and mean 2.61 in the 1100 ppm group Comment: some imbalance in caries in the active caries group. No other prognostic factors reported
Free of contamination/co‐intervention?	Low risk	Quotes: "Kits containing dentifrices and toothbrushes were supplied for the whole family every 3 months to guarantee the use of experimental LD by the children, facilitating compliance with study protocol" and ".. and were also asked not to use other oral hygiene products" Comment: contamination not observed

Methods	Trial design: 2‐armed, double‐blind, active‐controlled and stratified RCT  Location: Lithuania  Number of centres: 28 secondary schools in Vilnius, Lithuania  Recruitment period: study commenced in 1999
Participants	Inclusion criteria: ≥ 1 erupted second permanent molar; CVA D₃MFS score of 2 to 24  Exclusion criteria: low‐caries subjects; parent/guardian unwilling to disclose medical/dental history; intra‐oral x‐ray for caries diagnosis in prior 6 months; > 2 occlusal surfaces in second permanent molars restored/clinically cavitated/sealed; conduct of oral examination liable to cause unacceptable stress to participant; medical/dental conditions with potential to affect caries development (including antibiotic therapy); heart condition; cancer treatment receipt; fixed orthodontic appliances present at baseline prohibiting assessment of all erupted teeth  Baseline caries: baseline characteristic (baseline caries) "well balanced." Baseline values D₃MF not reported but ''not statistically different". Baseline D₁MFS for assessed participants at 24 months: 32.48 D₁MFS (Gp A: mean 32.95 (SEM 0.40); Gp B: mean 32.01 (SEM 0.37)) Age at baseline (years): range 11 to 14 years, mean 13 years (Gp A: 13.0; Gp B: 13.1)  Sex: 1330 F:1057 M (Gp A: 665 F:528 M; Gp B: 665 F:529 M). Baseline characteristic (sex) "well balanced"  Any other details of important prognostic factors: background exposure to fluoride not reported  Number randomised: 2387 (Gp A: 1193; Gp B: 1194)  Number evaluated: 2011 at 2 years (present for final assessment. Gp A: 994; Gp B: 1017)  Attrition: 15.8% dropout (for all study groups combined) after 2 years. Gp A: 16.7%; Gp B: 14.8%. Reasons for attrition absent from assessment (Gp A: 24, Gp B: 17); withdrawn (Gp A: 175, Gp B: 160)
Interventions	Comparison: FT versus FT  Gp A (n = 1193): SMFP 1000 ppm F; silica abrasive system; home use twice daily/unsupervised; daily brushing at school  Gp B (n = 1194): SMFP 2500 ppm F; silica abrasive system; home use twice daily/unsupervised; daily brushing at school
Outcomes	Primary: 2‐year net DMFS increment cl(DSTM) FOTI at D₃ all radiographic lesions; D₁MFS increment (DSTM only); D₃MFS increment (DSTM only); D₁MFS events (DSTM); D₃MFS events (DSTM). Reported at 1 and 2 year follow‐ups  Secondary: none assessed  Assessments irrelevant to this review's scope: none  Follow‐up duration: 2 years
Notes	Adverse effects: not reported  Funding source: "funded by Unilever Dental Research"  Declarations/conflicts of interest: 5 (including lead author) of 14 authors employed by Unilever: RK Chesters, E Huntington, JR Matheson; JA Nicholson, D Savage  Data handling by review authors: n/a  Other information of note: examinations carried out by a single examiner. Intra‐examiner reliability: repeat DSTM and FOTI examinations held throughout the baseline, 12 and 24 month examinations on 5% to 10% of subjects. For radiography, baseline and 12 and 24 month radiographs re‐assessed for 5% to 10% of subjects. Reproducibility "excellent", Kappa values > 0.8
*Risk of bias*
Bias	Authors' judgement	Support for judgement
Random sequence generation (selection bias)	Low risk	Quotes: " . .randomized to one of two silica‐based dentifrices" and "...stratified into 12 strata ...allocated to a product group according to a pre‐prepared list of randomized blocks?"
Allocation concealment (selection bias)	Unclear risk	Comment: insufficient information
Blinding (performance bias and detection bias)   All outcomes	Low risk	Quotes: "... double‐blind study" and "Neither the subjects, clinical examiners, nor those distributing the test products were aware of the product identities at any time during the trial. The investigators were supplied with sealed code‐break envelopes that could be opened in an emergency. This was not required and the integrity of the product code was confirmed with regular GPC monitoring and independent audit" and "The products were identical except for the fluoride level and different coloured packaging for each product code?"
Incomplete outcome data (attrition bias)   All outcomes	Low risk	2387 randomised (994/1193 included in final main analysis in low fluoride group; 1017/1194 in high fluoride group) Comment: not unreasonable dropout rate; similar in both groups. Reasons unlikely to be due to intervention. Numbers absent and withdrawn are given for each group. Well balanced between groups. No further information about dropouts given
Selective reporting (reporting bias)	Unclear risk	Comment: results reported traditional increment and DSTM increment at different levels of diagnosis. DMFT and proportion developing new caries missing. DSTM, FOTI and radiographic assessments
Baseline characteristics balanced?	Low risk	Comment: balance of sex and baseline DMFS
Free of contamination/co‐intervention?	Unclear risk	Comment: unlikely as used different colours for toothpaste tubes/cartons, but possibility of contamination during school brushing sessions

Methods	Trial design: 3‐armed, single‐blind, active‐controlled RCT  Location: UK  Number of centres: "examinations in 808 schools, in nine districts, throughout the North West of England." Districts include: Blackburn; Bolton; Burnley; Oldham; Salford; Skelmersdale; South Sefton; Tameside; Wigan  Recruitment period: study commenced in 1993 (5 health districts), and 1994 (4 health districts)
Participants	Inclusion criteria: not reported  Exclusion criteria: not reported  Baseline caries: 0 dmfs. Baseline characteristics not quantitatively reported  Age at baseline (years): 1 year. Baseline characteristics not reported  Sex: baseline characteristics not reported  Any other details of important prognostic factors: background exposure to fluoride not reported. Community water supply < 0.1 ppm F  Number randomised: 7422 (Gp A: 2488; Gp B: 2472; Gp C: 2462)  Number evaluated: 3467 (Gp A: 1186; Gp B: 1176; Gp C: 1369)  Attrition: 32% dropout after 5 years. Reasons for attrition: refused to participate (9%), change of residence (19%); product related and dental recommended withdrawals in high fluoride group only (0.07%)
Interventions	Comparison: FT versus FT^a  Gp A (n = 2488): NaF 1450 ppm; abrasive system not reported; home use/unsupervised, daily frequency assumed Gp B (n = 2472): NaF 440 ppm; abrasive system not reported; home use/unsupervised, daily frequency assumed Gp C (n = 1369): no intervention control
Outcomes	Primary: 5‐year dmft increment ‐ cl; prevalence of caries experience (dmft > 0). Reported at 5 years follow‐up  Secondary: none assessed  Assessments irrelevant to this review's scope: cost per tooth saved; cost per child saved from caries experience/extraction experience  Follow‐up duration: 5 years
Notes	Adverse effects: not reported  Funding source: not reported  Declarations/conflicts of interest: institutional affiliations reported only  Data handling by review authors: ^aComparator group (n = 2462) receiving no intervention also reported but not considered in this review  Other information of note: clinical (VT) caries assessments by trained, standardised, calibrated examiners. Clinical data only. Reliability values not reported
*Risk of bias*
Bias	Authors' judgement	Support for judgement
Random sequence generation (selection bias)	Low risk	Quotes: "..randomised controlled parallel group clinical trial" and "..centrally allocated to either one of the two test groups or a control group using random number tables"
Allocation concealment (selection bias)	Low risk	Quote: "..centrally allocated to either one of the two test groups or a control group" Comment: centralised allocation
Blinding (performance bias and detection bias)   All outcomes	Unclear risk	Quote: "Dental examinations were conducted under blind conditions but as "off the shelf" toothpaste was delivered to the participants, subjects and their families were aware of which toothpaste they were using" Comment: clinical assessors blinded, but participants and their families were not. Participants very young children so knowledge of intervention unlikely to influence outcome
Incomplete outcome data (attrition bias)   All outcomes	Low risk	1677/2472 available for examination in low fluoride group; 1696/2488 available in the high fluoride group. Total dropout rate of 32% Comment: dropout rate mainly due to refusal to participate, change of residence; product related and dental recommended withdrawals in high fluoride group only but this number is very small. Reasons for dropouts primarily unlikely to be due to intervention
Selective reporting (reporting bias)	Low risk	Comment: routine caries diagnosis. No radiographs taken; clinical examination only. Caries indices reported: mt, dmft, caries free
Baseline characteristics balanced?	Unclear risk	Comment: no baseline data presented. Study undertaken in deprived areas of North West of England with comparable caries prevalence in 5 year olds
Free of contamination/co‐intervention?	Unclear risk	Comment: contamination possible. Toothpaste supplied for use by children participating in the trial only and not to other family members

Methods	Trial design: 3‐armed, double‐blind, active/placebo‐controlled and stratified RCT  Location: China  Number of centres: not reported. Chengdu area, China  Recruitment period: study commenced in/before 2005
Participants	Inclusion criteria: not reported  Exclusion criteria: children with orthodontic appliances, participating in any other clinical study during the 3 months prior to baseline examination, or a condition impeding participation at baseline were excluded  Baseline caries: 3.6 dfs (Gp A: 3.54 (SD 5.34); Gp B: 3.53 (SD 5.62); Gp C: 3.60 (SD 6.07)). Baseline characteristic (dfs) "well balanced" (evaluated participants at 2 years reported only)  Age at baseline (years): range 4.0 to 4.5 years; mean 4 years (Gp A: 4.23 years (SD 0.134); Gp B: 4.28 years (SD 0.141); Gp C: 4.19 years (SD 0.129)). Baseline characteristic (age) "well balanced" (evaluated participants at 2 years reported only)  Sex: 431 F:567 M (Gp A: 136 F:193 M; Gp B: 141 F:200 M; Gp C: 154 F:174 M). Baseline characteristic (sex) "well balanced" (evaluated participants at 2 years reported only)  Any other details of important prognostic factors: background exposure to fluoride: community water supply fluoridated 0.3 ppm F  Number randomised: 1200 (group distribution not reported)  Number evaluated: 998 at 2 years (present for all assessments) (Gp A: 329; Gp B: 341; Gp C:328)  Attrition: dropouts not obtainable. Reasons for attrition not fully reported: "Subjects who did not complete the study dropped out for reasons unrelated to the use of the treatments." Differential losses not assessable
Interventions	Comparison: FT (2 groups) versus PL  Gp A (n = assessed 329): SMFP 1500 ppm F; abrasive system Ca‐carbonate; home use/unsupervised, twice daily frequency assumed  Gp B (n = assessed 341): SMFP 1500 ppm F; abrasive system not reported silica; home use/unsupervised, twice daily frequency assumed Gp C (n = assessed 328): placebo; abrasive system not reported Ca‐carbonate; home use/unsupervised, twice daily frequency assumed
Outcomes	Primary: 2‐year dfs increment ‐ cl (at 2 years)  Secondary: adverse effects  Assessments irrelevant to this review's scope: compliance  Follow‐up duration: 2 years
Notes	Adverse effects: "Throughout the study, there were no adverse effects on the oral hard or soft tissues, which were observed by the dental examiner, or reported by the subjects when questioned in this regard"  Funding source: "This clinical study was funded by the Colgate‐Palmolive Company"  Declarations/conflicts of interest: half of the authors were employed by the product manufacturer (3 of 6 authors work at Colgate‐Palmolive Technology Center, New Jersey, USA: Yun Po Zhang; Anthony R Volpe; William DeVizio)  Data handling by review authors: Gp A + B versus C for analysis. Combined 2 SMFP 1500 groups with different abrasive systems  Other information of note: analysis of covariance adjusted for baseline dfs. Clinical (VT) assessment by 1 examiner. A subsample of 40 children were re‐assessed Kappa > 0.9
*Risk of bias*
Bias	Authors' judgement	Support for judgement
Random sequence generation (selection bias)	Unclear risk	Quote: "Subjects were randomly assigned..."
Allocation concealment (selection bias)	Unclear risk	Comment: insufficient information
Blinding (performance bias and detection bias)   All outcomes	Low risk	Quotes: "...employed a double‐blind..." and "Dentifrices were packaged in white tubes or overwrapped with white tape so as to mask their identity"
Incomplete outcome data (attrition bias)   All outcomes	Unclear risk	Quotes: "A total of 1200 qualifying children...entered the study" and "Subjects who did not complete the study dropped out for reasons unrelated to the use of the treatments"  Comment: number randomised/excluded/withdrawn not reported
Selective reporting (reporting bias)	Low risk	Mean dfs increment. Clinical (VT) assessment only
Baseline characteristics balanced?	Unclear risk	Comment: analysis adjusted for baseline dfs. Baseline data reported for participants completing the trial only. However, groups analysed are similar with respect to sex and mean dfs score at baseline
Free of contamination/co‐intervention?	Unclear risk	Comment: insufficient information. Possibility of contamination during brushing sessions

Methods	Trial design: 4‐arm, double‐blind, active/placebo‐controlled RCT  Location: Sweden  Number of centres: not reported. Schools in Växjö, Sweden  Recruitment period: study commenced in/before 1970
Participants	Inclusion criteria: not reported  Exclusion criteria: not reported  Baseline caries: 5.02 DMFS (Gp A: 4.53; Gp B: 4.87; Gp C: 5.57; Gp D: 5.16 DMFS). Baseline characteristic (DMFS) "balanced"  Age at baseline (years): range 10 to 11 years (not reported by group)  Sex: 255 F:304 M (Gp A: 64 F:73 M; Gp B: 63 F:77 M; Gp C: 62 F:75 M; Gp D: 66 F:79 M)  Any other details of important prognostic factors: background weekly supervised exposure to fluoride mouthrinse (NaF 1000 ppm), continued fortnightly throughout study. Community water supply naturally fluoridated (< 0.2 ppm F)  Number randomised: 681 (group numbers not reported)  Number evaluated: 559 at 2 years (present at final assessment. Gp A: 137; Gp B: 140; Gp C: 137; Gp D: 145)  Attrition: 18% dropout after 2 years. Reasons for attrition described with respective total numbers: change of residence/school, orthodontic treatment, did not wish to continue; no differential group losses reported (but not assessable)
Interventions	Comparison: FT (3 groups) versus PL  Gp A (n = 137 evaluated): NaF 250 ppm; silicon dioxide abrasive system; home use/unsupervised, daily frequency assumed  Gp B (n = 140 evaluated): SMFP 250 ppm; silicon dioxide system; home use/unsupervised, daily frequency assumed Gp C (n = 137 evaluated): SMFP 1000 ppm ; silicon dioxide system; home use/unsupervised, daily frequency assumed  Gp D (n = 145 evaluated): placebo; silicon dioxide system; home use/unsupervised, daily frequency assumed
Outcomes	Primary: 2‐year DMFS increment ‐ (NCA) cl; (BLMD‐DFS) cl; (MD‐DFS) xr; proportion of children with new smooth surface caries (at 2 years)  Secondary: none assessed  Assessments irrelevant to this review's scope: none assessed  Follow‐up duration: 2 years
Notes	Adverse effects: "In Växjö, where type K was used, some pupils complained after a short time that they had stains on their teeth" Complainants examined, and blue‐green colouring agent mostly removed. Authors report participants (and parents) understanding and cooperative throughout study irrespective of issue  Funding source: materials provided by the manufacturer, Barnängen Company, Stockholm. Financial support provided by Barnängen Company, Stockholm and the Swedish patent revenue research fund (Patentmedelsfonden för odontologisk profylaxforskning)  Declarations/conflicts of interest: not reported  Data handling by review authors: Gps A + B (250 ppm F NaF and SMFP) combined for meta‐analysis. Same trial report as Forsman 1974a, intervention differed solely according to abrasive system  Other information of note: clinical (VT) caries assessment by 1 examiner, diagnostic threshold = NCA. Radiographic assessment (postBW) by 1 examiner, diagnostic threshold = ER. State of tooth eruption included not reported. Diagnostic errors not reported
*Risk of bias*
Bias	Authors' judgement	Support for judgement
Random sequence generation (selection bias)	Unclear risk	Quote: "From lists for girls resp. boys in all classes each fourth child on the Vaxjo lists and each third child on the Ljungby lists was randomly selected for the respective groups" Comment: not enough information provided
Allocation concealment (selection bias)	Unclear risk	No information provided
Blinding (performance bias and detection bias)   All outcomes	Low risk	Quotes: "The toothpaste was delivered in tubes with the word 'Toothpaste' printed in different colours. During the period of investigation, only the manufacturer knew the code" and "...study was designed as a double‐blind experiment" Comment: blind outcome assessment and use of placebo described
Incomplete outcome data (attrition bias)   All outcomes	Unclear risk	Overall dropout for length of follow‐up: 17.9% (122/681) in 2 years. Dropout by group: not reported. Reasons for losses: orthodontic treatment (6), moved away (39), did not wish to continue (77) (not reported by group) Comment: numbers lost are not unduly high for length of follow‐up. It is unclear if there were any differential losses, and if reasons for missing outcome data are acceptable and balanced. Caries data used in the analysis pertain to participants continuing the study up to year 2 (children completing tests)
Selective reporting (reporting bias)	Low risk	Outcomes reported:  DMFS increment ‐ (NCA) cl, reported at 2 years follow‐up  (BLMD‐DFS) cl  (MD‐DFS) xr  proportion of children with new smooth surface caries Comment: trial protocol not available. All pre‐specified outcomes were reported and were reported in the pre‐specified way
Baseline characteristics balanced?	Low risk	Prognostic factors reported:  DMFS: 5.57 FT 1, 4.87 FT 2, 4.53 FT 3, 5.16 PL dental age: 18.89 FT 1, 19.08 FT 2, 18.66 FT 3, 19.03 PL Comment: initial caries appears balanced
Free of contamination/co‐intervention?	Low risk	Quote: "The dentifrice was distributed every second month in amounts calculated to meet the needs of the whole family, to ensure as far as possible that the participants did not have access to other toothpastes" Comment: no reported indication of contamination/co‐intervention

Methods	Trial design: ^a3‐armed, double‐blind, placebo‐controlled and stratified RCT  Location: USA  Number of centres: 3 rural elementary schools in North‐Central Indiana  Recruitment period: study commenced in/before 1994
Participants	Inclusion criteria: children lacking dental care and caries‐susceptible; good general health; no serious medical condition/transmissible diseases  Exclusion criteria: absence of dmfs/t/DMFS/T; undergoing orthodontic treatment  Baseline caries: 4.2 DMFS (Gp A: 4.38 DMFS (SD 5.35)/2.81 DMFT (SD 2.81); Gp B: 3.95 DMFS (SD 4.53)/2.73 DMFT (SD 2.58)). Baseline characteristics (DMFS, DMFT) "balanced"  Age at baseline (years): range 10 to 11 years, mean 10.7 years (Gp A: 10.7 years; Gp B: 10.6 years). Baseline characteristic (age) "balanced"  Sex: Gp A: 42 M:45 F; Gp B: 42 M:45 F. Baseline characteristic (sex) "balanced" Any other details of important prognostic factors: no background exposure to fluoride reported. Community water supplies < 0.4 ppm F; request to dentists in area not to apply topical fluoride to study population children  Number randomised: 174 (Gp A: 87; Gp B: 87)  Number evaluated: 156 at 1 year (present at final assessment. Gp A: 77; Gp B: 79)  Attrition: 10% dropout after 1 year (study duration = 1 year). Main reasons for attrition: changes in residence, few exclusions for initiation of orthodontic treatment; no differential group losses
Interventions	Comparison: FT (+ Al rinse) versus PL (+ Al rinse)^b  Gp A (n = 87): NaF 1100 ppm F; abrasive system: silica; home use/unsupervised, daily frequency assumed  Gp B (n = 87): placebo; abrasive system: silica; home use/unsupervised, daily frequency assumed
Outcomes	Primary: 1‐year DMFS increment ‐ (CA) cl + (ER) xr; DMFT; proportion of children with new DMFS; proportion of children remaining caries free (at 6 months, 1 year)  Secondary: adverse effects (oral soft tissues lesions) (at 6 months, 1 year)  Assessments irrelevant to this review's scope: compliance  Follow‐up duration: 1 year
Notes	Adverse effects: "No adverse oral effects attributable to any treatment regimens were observed during the study. Due to the low incidence of soft tissue aberrations in this population, the soft tissue findings are not presented"  Funding source: not reported  Declarations/conflicts of interest: institutional affiliations reported only  Data handling by review authors: results of 1 examiner chosen (findings consistent throughout). ^aThird trial arm (fluoride toothpaste and no mouthrinse) excluded from this review due to no eligible comparator arm. ^bRinsing with 500 ppm Al solutions performed daily at school in both relevant groups compared. Clinical (VT) caries assessment by 2 examiners, diagnostic threshold = CA; state of tooth eruption included = E/U. Radiographic assessment (postBW) by 2 examiners (independently), diagnostic threshold = ER. Reversals were small in both groups and equally common
*Risk of bias*
Bias	Authors' judgement	Support for judgement
Random sequence generation (selection bias)	Unclear risk	Quote: "Subjects with evidence of caries activity were stratified according to age, sex... then randomly assigned to one of the balanced groups" Comment: not enough information provided
Allocation concealment (selection bias)	Unclear risk	No information provided
Blinding (performance bias and detection bias)   All outcomes	Low risk	Quotes: "A double blind comparison of three parallel groups of children... who used a test or placebo dentifrice for a twelve month period" and "Radiographs were scored independently by each examiner at a later date.." Comment: blind outcome assessment and use of placebo described
Incomplete outcome data (attrition bias)   All outcomes	Low risk	Overall dropout for length of follow‐up: 10% in 1 year. Dropout by group: 10/87 FT, 8/87 PL. Reasons for losses: changes in residence, exclusion based on orthodontic treatment Comment: numbers lost were not unduly high given the length of follow‐up with no differential losses between groups. Reasons for the missing outcome data are acceptable. Caries data used in analysis pertain to participants present at final examination
Selective reporting (reporting bias)	Low risk	Outcomes reported:  DMFS increment ‐ (CA) cl + (ER) xr, reported at 6 months and 1 year follow‐ups  DMFT  proportion of children remaining caries free  proportion of children with new DMFS  oral soft tissues lesions Comment: trial protocol not available. All pre‐specified outcomes (in Methods) were reported and were reported in the pre‐specified way
Baseline characteristics balanced?	Low risk	Prognostic factors reported:  DMFS: 5.06 (0.58) FT, 4.78 (0.50) PL DMFT: 3.31 (0.32) FT, 3.32 (0.27) PL Age: 10.7 FT, 10.6 PL Sex: 42 M, 45 F (FT); 42 M, 45 F (PL). Comment: initial caries appears balanced between groups
Free of contamination/co‐intervention?	Low risk	Quote: "Sufficient quantities of the respective products were provided for the participants and their families to use throughout the study. Participants with the same telephone number or address were assigned to the same group to avoid confusion with different test products in the same household" Comment: there is sufficient indication overall of prevention of contamination/co‐intervention