Fluorides for preventing early tooth decay (demineralised lesions) during fixed brace treatment

Summary of findings for the main comparison. Dentist/nurse‐applied fluoride: fluoride varnish compared to non‐fluoride (placebo) varnish for preventing early tooth decay (demineralised lesions) during fixed brace treatment

Fluoride varnish compared to non‐fluoride (placebo) varnish for preventing early tooth decay (demineralised lesions) during fixed brace treatment
Patient or population: orthodontic patients (any age) Setting: orthodontic clinics in Sweden Intervention: fluoride varnish Comparison: non‐fluoride (placebo) varnish
Outcomes	*Anticipated absolute effects^ (95% CI)**			Relative effect (95% CI)	Number of participants (studies)	Certainty of the evidence (GRADE)	Comments
	With non‐fluoride (placebo) varnish	With fluoride varnish	Difference
Number of participants with new DLs (new DLs) Assessed with: pre‐treatment and day of debond clinical intraoral photographs Follow‐up: 1.7 years	28.80%	15% (4 to 55.60)	13.80% fewer (24.80 fewer to 26.80 more)	RR 0.52 (0.14 to 1.93)	405 (2 RCTs)	⊕⊕⊝⊝ LOW^a,b	The evidence that professional application of fluoride varnish (7700 or 10,000 ppm F) every 6 weeks to the teeth of patients wearing fixed orthodontic braces reduces the number of new DLs is equivocal
Number of participants with more severe DLs (severity of DLs) (score 3 or 4 versus score 1 or 2) Follow‐up: 1.7 years	26%	12% (5.70 to 24.70)	14.10% fewer (20.30 fewer to 1.30 fewer)	RR 0.46 (0.22 to 0.95)	148 (1 RCT)	⊕⊕⊝⊝ LOW^c,d	The evidence suggests that when a dentist or nurse applies 7700 ppm F (ammonium fluoride) varnish every 6 weeks to a patient wearing a fixed orthodontic brace there may a reduction in the number of orthodontic patients with more severe DLs (score 3 or 4 versus score 1 or 2)
Number of participants with adverse effects (adverse effects) Follow‐up: 1.7 years	No evidence that the intervention had adverse effects				148 (1 RCT)	⊕⊕⊝⊝ LOW^d,e	Only 1 participant with an adverse event and not clear if this was directly related to the intervention
^*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; DLs: demineralised lesions; F: fluoride; ppm: parts per million; RCT: randomised controlled trial; RR: risk ratio.
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
^aDowngraded 1 level for indirectness (evidence from only 2 studies) (Stecksén‐Blicks 2007; Sonesson 2019). ^bDowngraded 1 level for imprecision (insufficient number of participants with new DLs (guideline 300 to 400 events)). ^cDowngraded 1 level for imprecision (insufficient number of participants with the more severe DLs (guideline 300 to 400 events)). ^dDowngraded 1 level for indirectness (evidence from only 1 study) (Sonesson 2019). ^eDowngraded 1 level for imprecision (insufficient number with adverse events and not clear if directly related to the intervention).

Summary of findings 2. Dentist/nurse‐applied fluoride: 12,300 ppm F APF foam compared to 0 ppm F placebo foam for preventing early tooth decay (demineralised lesions) during fixed brace treatment

12,300 ppm F APF foam compared to 0 ppm F placebo foam for preventing early tooth decay (demineralised lesions) during fixed brace treatment
Patient or population: orthodontic patients (any age) Setting: orthodontic department at dental hospital in China Intervention: 12,300 ppm F APF foam Comparison: 0 ppm F placebo foam
Outcomes	*Anticipated absolute effects^ (95% CI)**			Relative effect (95% CI)	Certainty of the evidence (GRADE)	Number of participants (studies)	Comments
	With 0 ppm F placebo foam	With 12,300 ppm F APF foam	Difference
Number of participants with new DLs (new DLs) Assessed with: clinical assessment Follow‐up: mean 18 months	48.90%	12.70% (5.40 to 27.90)	36.20% fewer (43.60 fewer to 21 fewer)	RR 0.26 (0.11 to 0.57)	⊕⊕⊝⊝ LOW^a,b	95 (1 RCT)	The evidence suggests that when foam, containing 12,300 ppm F, is applied by a dentist or a nurse, every 2 months, to the teeth of patients wearing fixed orthodontic appliances there might be a reduction in the number of patients who have at least 1 new DL
Number of participants with more severe DLs (severity of DLs)	None of the trials reported this outcome
Number of participants with adverse effects (adverse effects)	None of the trials 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). APF: acidulated phosphate fluoride; CI: confidence interval; DLs: demineralised lesions; F: fluoride; ppm: parts per million; RCT: randomised controlled trial; RR: risk ratio.
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
^aAlthough effect size is large this was only for 1 study at unclear risk of bias and therefore downgraded 1 level (Jiang 2013). ^bDowngraded 1 level for imprecision (insufficient number of participants with new DLs (guideline 300 to 400 events)).

Summary of findings 3. Patient‐applied/used fluoride: 5000 ppm F toothpaste compared to 1450 ppm F toothpaste for preventing early tooth decay (demineralised lesions) during fixed brace treatment

5000 ppm F toothpaste compared to 1450 ppm F toothpaste for preventing early tooth decay (demineralised lesions) during fixed brace treatment
Patient or population: orthodontic patients (any age) Setting: home use Intervention: 5000 ppm F toothpaste Comparison: 1450 ppm F toothpaste
Outcomes	*Anticipated absolute effects^ (95% CI)**			Relative effect (95% CI)	Number of participants (studies)	Certainty of the evidence (GRADE)	Comments
	With 1450 ppm F toothpaste	With 5000 ppm F toothpaste	Difference
Number of participants with new DLs (new DLs) Assessed with: pre‐treatment and post‐treatment clinical intraoral photographs Follow‐up: 1.8 years	26.60%	18.10% (12.20 to 26.60)	8.50% fewer (14.30 fewer to 0 fewer)	RR 0.68 (0.46 to 1.00)	380 (1 RCT)	⊕⊕⊝⊝ LOW^a,b	The evidence suggests that in patients wearing an orthodontic fixed brace use of a daily 5000 ppm F toothpaste compared with a daily 1450 ppm F toothpaste throughout treatment may lead to a reduction in the number of orthodontic patients with new DLs
Number of participants with more severe DLs (severity of DLs)	None of the trials reported this outcome
Number of participants with adverse effects (adverse effects)	None of the trials 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; DLs: demineralised lesions; F: fluoride; ppm: parts per million; RCT: randomised controlled trial; RR: risk ratio.
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
^aDowngraded 1 level due to single study at unclear risk of bias (Sonesson 2014). ^bDowngraded 1 level for imprecision (insufficient number of participants with new DLs (guideline 300 to 400 events)).

Summary of findings 4. Patient‐applied/used fluoride: 250 ppm F mouthrinse (100 ppm F amine F/150 ppm NaF) compared to 0 ppm F placebo mouthrinse for preventing early tooth decay (demineralised lesions) during fixed brace treatment

250 ppm F mouthrinse (100 ppm amine F/150 ppm NaF) compared to 0 ppm F placebo mouthrinse for preventing early tooth decay (demineralised lesions) during fixed brace treatment
Patient or population: orthodontic patients (any age) Setting: home use Intervention: 250 ppm F mouthrinse (100 ppm amine F/150 ppm NaF) Comparison: 0 ppm F placebo mouthrinse
Outcomes	*Anticipated absolute effects^ (95% CI)**			Relative effect (95% CI)	Number of participants (studies)	Certainty of the evidence (GRADE)	Comments
	With 0 ppm F placebo mouthrinse	With 250 ppm F mouthrinse (100 ppm amine F/150 ppm NaF)	Difference
Number of participants with new DLs (new DLs) Assessed with: QLF Follow‐up: 24.5 months	46.70%	30.30% (17.30 to 54.60)	16.30% fewer (29.40 fewer to 7.90 more)	RR 0.65 (0.37 to 1.17)	81 (1 RCT)	⊕⊝⊝⊝ VERY LOW^a,b	The evidence is very uncertain about the effect of a daily 250 ppm F mouthrinse (100 ppm amine F/150 ppm NaF) compared with a daily 0 ppm F placebo mouthrinse on the number of patients wearing a fixed orthodontic brace with new DLs
Number of participants with more severe DLs (severity of DLs)	None of the trials reported this outcome
Number of participants with adverse effects (adverse effects)	None of the trials 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; DLs: demineralised lesions; F: fluoride; NaF: sodium fluoride; ppm: parts per million; QLF: quantitative light‐induced fluorescence; RCT: randomised controlled trial; RR: risk ratio.
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
^aDowngraded 2 levels: single study with a relatively small number of participants (81), at high risk of bias due to high attrition (33%) (van der Kaaij 2015). ^bDowngraded 1 level for imprecision (insufficient number of participants with new DLs (guidance 300 to 400 events)).

See: Summary of findings for the main comparison Dentist/nurse‐applied fluoride: fluoride varnish compared to non‐fluoride (placebo) varnish for preventing early tooth decay (demineralised lesions) during fixed brace treatment; Summary of findings 2 Dentist/nurse‐applied fluoride: 12,300 ppm F APF foam compared to 0 ppm F placebo foam for preventing early tooth decay (demineralised lesions) during fixed brace treatment; Summary of findings 3 Patient‐applied/used fluoride: 5000 ppm F toothpaste compared to 1450 ppm F toothpaste for preventing early tooth decay (demineralised lesions) during fixed brace treatment; Summary of findings 4 Patient‐applied/used fluoride: 250 ppm F mouthrinse (100 ppm F amine F/150 ppm NaF) compared to 0 ppm F placebo mouthrinse for preventing early tooth decay (demineralised lesions) during fixed brace treatment; Summary of findings 5 Fluoride‐releasing materials: resin‐modified glass ionomer cement compared to light‐cured composite resin for bonding orthodontic brackets for preventing early tooth decay (demineralised lesions) during fixed brace treatment

Summary of findings 5. Fluoride‐releasing materials: resin‐modified glass ionomer cement compared to light‐cured composite resin for bonding orthodontic brackets for preventing early tooth decay (demineralised lesions) during fixed brace treatment

Resin‐modified glass ionomer cement compared to light‐cured composite resin for bonding orthodontic brackets for preventing early tooth decay (demineralised lesions) during fixed brace treatment
Patient or population: orthodontic patients (any age) Setting: 2 dental teaching hospitals and 4 specialist orthodontic practices in UK and Republic of Ireland Intervention: resin‐modified glass ionomer cement Comparison: light‐cured composite resin
Outcomes	*Anticipated absolute effects^ (95% CI)**			Relative effect (95% CI)	Number of participants (studies)	Certainty of the evidence (GRADE)	Comments
	With light‐cured composite resin	With resin‐modified glass ionomer cement	Difference
Number of participants with new DLs (new DLs) Assessed with: before and day of debond clinical intraoral photographs Follow‐up: 17.6 months	22.40%	26.20% (15.40 to 44.50)	3.80% more (6.90 fewer to 22.10 more)	RR 1.17 (0.69 to 1.99)	173 (1 RCT)	⊕⊕⊝⊝ LOW^a,b	The evidence suggests that resin‐modified glass ionomer cement for bonding orthodontic brackets compared with light‐cured composite resin results in little to no difference in the number of orthodontic patients with new DLs
Number of participants with more severe DLs of aesthetic concern (severity of DLs) Assessed with: pre‐treatment and day of debond clinical intraoral photographs Follow‐up: 17.6 months	8.00%	9.40% (3.60 to 24.80)	1.40% more (4.40 fewer to 16.90 more)	RR 1.18 (0.45 to 3.12)	173 (1 RCT)	⊕⊕⊝⊝ LOW^a,b	The evidence suggests that using resin‐modified glass ionomer cement for bonding orthodontic brackets compared with light‐cured composite resin results in little to no difference in the number of orthodontic patients with more severe DLs of aesthetic concern
Number of participants with adverse effects (adverse effects) Follow‐up: 17.6 months	No evidence that either intervention had adverse effects				173 (1 RCT)	⊕⊕⊝⊝ LOW^a,b	‐
^*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; DLs: demineralised lesions; RCT: randomised controlled trial; RR: risk ratio.
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
^aDowngraded 1 level: single study at unclear risk of bias (Benson 2019). ^bDowngraded 1 level for imprecision (insufficient number of participants with new DLs or more severe DLs (guidance 300 to 400 events)).

Antecedentes

available in

Descripción de la afección

Durante el tratamiento ortodóntico con aparatos fijos, se colocan brackets en los dientes para sostener los alambres que proporcionan las fuerzas para alinear los dientes. Uno de los efectos adversos de los aparatos de ortodoncia fijos es que se acumula placa dental alrededor de los aparatos, lo cual da lugar a la acumulación de los tipos de bacterias que causan enfermedades dentales (Naranjo 2006). La acumulación de placa dental alrededor de los brackets ortodónticos se asocia con un mayor riesgo de lesiones desmineralizadas (LD, también conocidas como lesiones de manchas blancas), que pueden ser visibles en un plazo de seis meses (Tufekci 2011). La desmineralización es una etapa precoz, pero reversible, en el desarrollo de la caries dental. Las bacterias cariogénicas presentes en la placa dental transforman el azúcar de la dieta en ácidos orgánicos, que comienzan a dañar el esmalte dental. La eliminación efectiva de la placa evitará que ocurran LD; sin embargo, la presencia de aparatos de ortodoncia en la boca y el dolor dental asociado pueden hacer que la limpieza adecuada de los dientes y los aparatos ortodónticos sea más difícil. Las LD que se desarrollan en las superficies de los dientes cercanas a los labios y la boca durante el tratamiento ortodóntico pueden convertirse en un problema significativo durante el ciclo del tratamiento, que puede durar 18 meses o más, y dar lugar a una apariencia desagradable de los dientes después del alineamiento (Maxfield 2012). En los casos graves la caries puede dar lugar a un orificio en el diente que requerirá obturación (restauración), lo cual puede ser doloroso y costoso.

Enaia y colegas (Enaia 2011) utilizaron fotografías clínicas de los dientes tomadas antes y después del tratamiento con aparatos de ortodoncia fijos para mostrar que mientras que el 32% de las personas de su estudio tenían LD antes de que se les colocara el aparato, esta cifra aumentó al 74% después de que se les quitaron los aparatos. La mayoría de las LD eran menores, aunque una minoría significativa de los participantes (10%) presentó orificios en los dientes, lo que puede haber requerido la obturación. Aunque las LD tienden a desvanecerse con el tiempo a medida que sanan, un estudio ha demostrado que, incluso cinco años después del tratamiento, una proporción significativa de personas que habían usado aparatos ortodónticos fijos mostraron alguna evidencia de LD causadas por los aparatos ortodónticos en comparación con un grupo de control de personas que no habían utilizado un aparato ortodóntico fijo (Ogaard 1989).

Descripción de la intervención

Los ortodoncistas desean prevenir el desarrollo de las LD para que sus pacientes puedan tener la mejor apariencia posible después del tratamiento ortodóntico ‐ dientes alineados sin marcas. El fluoruro es importante en la prevención de la caries dental (ten Cate 2013). Marinho y colegas (Marinho 2016) encontraron una reducción definitiva en la caries dental en niños y adolescentes que realizaban enjuagues regulares supervisados con un enjuague bucal de fluoruro. También se ha demostrado que el fluoruro puede reducir el número de LD que se desarrollan durante el tratamiento con aparatos ortodónticos. Cuando los participantes con ortodoncia usaron un enjuague bucal, Geiger et al (Geiger 1992) encontraron una reducción del 30% en el número de participantes con LD y una reducción del 25% en la incidencia de dientes afectados por LD. Muchos ortodoncistas recomiendan el uso de un enjuague bucal diario de fluoruro durante el tratamiento con aparatos ortodónticos para prevenir las LD (Kerbusch 2012).

Se utilizan varios métodos (además de la pasta dental fluorada) para administrar fluoruro a los dientes de los pacientes durante el tratamiento ortodóntico. Estas incluyen las siguientes:

fluoruros tópicos (por ejemplo, enjuague bucal, gel, barniz o espuma) que son aplicados por el dentista o por personal de enfermería cuando atienden al paciente que utiliza el aparato ortodóntico para ajustarlo;
fluoruros tópicos (por ejemplo, pasta de dientes, enjuague bucal, gel o espuma) que el paciente que utiliza el aparato ortodóntico puede aplicar por sí mismo en su domicilio;
materiales que liberan fluoruro (por ejemplo, pegamentos utilizados para unir los brackets a los dientes y elásticos ortodónticos impregnados con fluoruro);
administración de suplementos con fluoruro en la dieta (por ejemplo, leche fluorada).

De qué manera podría funcionar la intervención

El fluoruro presente en la boca reduce el desarrollo de la caries a través de tres mecanismos: inhibición de la desmineralización del esmalte dental, mejoría de la remineralización del esmalte dental que produce una capa remineralizada resistente al ataque ácido e inhibición de las enzimas bacterianas que producen el ácido (Lynch 2006; ten Cate 2013).

La mayoría de los niños sometidos a un tratamiento ortodóntico estarán expuestos a algunas concentraciones bajas de fluoruro en el suministro de agua, concentraciones más altas en la pasta de dientes fluorada, o ambas. Es probable que el uso de otros fluoruros tópicos o fuentes de fluoruro diseñadas para suministrar fluoruro adicional al área en riesgo, o ambos, cerca de los brackets ortodónticos, reduzca aún más el riesgo de desarrollo de LD. Los fluoruros tópicos incluyen pastas dentales fluoradas, enjuagues bucales, geles, barnices y fuentes dietéticas (por ejemplo, leche fluorada). Las fuentes ortodónticas específicas de fluoruro incluyen adhesivos para brackets y bandas elásticas ortodónticas (ligaduras elastoméricas), que liberan lentamente el fluoruro en la boca. Todas estas fuentes de fluoruro liberan fluoruro en la saliva que se distribuye por toda la boca.

Por qué es importante realizar esta revisión

Varias revisiones sistemáticas han establecido la conclusión de que el uso de fluoruro tópico en diversas formas reduce la prevalencia y la incidencia de caries dental en niños y adolescentes (Marinho 2003a; Marinho 2003b; Marinho 2004; Marinho 2016). Por lo tanto, los ortodoncistas han recomendado de manera sistemática el uso de enjuagues bucales tópicos de fluoruro a los pacientes que usan aparatos fijos o aplican fluoruro en forma de barniz, gel o espuma para reducir el riesgo de que los pacientes que utilizan ortodoncia desarrollen caries dental. Otros utilizan materiales que liberan fluoruro para fijar los brackets o las bandas a los dientes. Existe una falta de evidencia clara con respecto a la concentración óptima de fluoruro tópico, la frecuencia óptima de uso y los efectos de los fluoruros tópicos y los materiales que liberan fluoruro durante el periodo completo del tratamiento ortodóntico.

Esta revisión Cochrane se publicó por primera vez en 2004 (Benson 2004). Se actualizó en 2013 con un protocolo enmendado para incluir solo los estudios aleatorizados diseñados de forma apropiada, que utilizaban grupos paralelos y que medían resultados relevantes a lo largo del periodo completo del tratamiento ortodóntico (Benson 2013). Ésta es la segunda actualización.

Objetivos

available in

El objetivo primario de esta revisión fue evaluar si el fluoruro tópico reduce la proporción de pacientes que utilizan ortodoncia con nuevas lesiones desmineralizadas (LD) después del uso de aparatos fijos.

Los objetivos secundarios fueron examinar la efectividad de los diferentes modos de administración de fluoruro para reducir las proporciones de pacientes que utilizan ortodoncia con nuevas LD, así como la gravedad de las lesiones, en términos del número, el tamaño y el color. Se debían incluir los resultados evaluados por los participantes, tales como las LD y los datos de la calidad de vida relacionados con la salud bucal, al igual que los informes de los efectos adversos.

Métodos

available in

Criterios de inclusión de estudios para esta revisión

Tipos de estudios

Se incluyeron ensayos controlados aleatorizados (ECA) en los que se administró fluoruro por cualquier método para prevenir el desarrollo de nuevas lesiones desmineralizadas (LD) del esmalte durante el tratamiento con aparatos de ortodoncia fijos. Debido a que los fluoruros tópicos son distribuidos por toda la boca por medio de la saliva, el uso de un diseño de estudio de boca dividida o en la misma persona para evaluar estas intervenciones es inapropiado debido a los efectos de arrastre potenciales (Pandis 2013), por lo tanto, se excluyó cualquier estudio que usara este diseño. Se examinarían los estudios con brazos múltiples y solo se incluirían en el análisis por pares las comparaciones entre dos intervenciones con fluoruro o intervenciones con fluoruro frente a una intervención sin fluoruro.

Tipos de participantes

Se incluyó a los participantes de cualquier edad que recibieron tratamiento ortodóntico con aparatos fijos en situaciones en las que se evaluaron las LD en los dientes que se encontraban en la boca al final del tratamiento ortodóntico (al momento del despegue, inmediatamente después de que se retira el aparato fijo activo). Se excluyeron los estudios que evaluaban la desmineralización de los dientes extraídos (ex vivo) o que se realizaron durante períodos cortos de tiempo, es decir, un periodo menor al tiempo completo que el aparato ortodóntico fijo estuvo en la boca.

Tipos de intervenciones

Fluoruro tópico en forma de pasta de dientes, enjuague bucal, gel, barniz, espuma o fuentes dietéticas en cualquier dosis, frecuencia, duración o método de administración, y con cualquiera de los siguientes agentes/ingredientes activos: NaF (fluoruro de sodio), SMFP (monofluorofosfato de sodio), SnF (fluoruro de estaño), APF (fluoruro de fosfato acidulado) y amina F (fluoruro de amina) aplicados por un profesional de cirugía odontológica o en el consultorio, o utilizados por el paciente en su domicilio.
Materiales que contienen fluoruro que se libera potencialmente durante el tratamiento, incluidos los materiales de resina compuestos para adhesión que liberan fluoruro, compómeros, cementos de ionómero de vidrio y ionómeros de vidrio modificados con resina para adherencia o colocación de bandas, dispositivos o microesferas de fluoruro de liberación lenta y ligaduras elastoméricas que liberan fluoruro.
Grupo de control compuesto por individuos que no fueron sometidos a la intervención con fluoruro, sino que fueron tratados con un placebo, como una pasta dental y enjuague bucal sin fluoruro, o a los que no se les administró ninguna intervención. También se incluyeron los estudios que incluían un control sometido a una intervención con fluoruro alternativa.

Tipos de medida de resultado

Resultados primarios

El resultado primario fue la diferencia en la proporción de participantes en el grupo experimental y de comparación con nuevas LD el día en que se retiró el aparato fijo, lo cual podía ser evaluado directamente a partir del participante (clínicamente) o preferiblemente a partir de fotografías del inicio y el final o imágenes fluorescentes de los dientes inmediatamente después de retirar el aparato fijo activo. Cuando no se registró el número de LD al inicio del tratamiento, el resultado fue la presencia o ausencia de LD al final del tratamiento ortodóntico, evaluadas de nuevo directamente a partir del participante o indirectamente a partir de fotografías o imágenes fluorescentes de los dientes. Es importante que la evaluación se haya realizado el día en que se retiró el aparato (o lo más cerca posible del mismo), ya que las marcas blancas tenderán a desaparecer (remineralizarse) después de retirar el aparato ortodóntico. La velocidad con la que se remineraliza la LD puede variar de un individuo a otro y es otro factor de confusión potencial.

Resultados secundarios

Diferencias en la gravedad de las nuevas LD, evaluadas de acuerdo al número, al tamaño y al color entre el grupo experimental y de control al final del tratamiento ortodóntico.
Cualquier evaluación cuantitativa de la pérdida de minerales del esmalte, como las técnicas de luz fluorescente o la microrradiografía, utilizadas con modelos de caries in situ (Benson 2010) al final del tratamiento.
Cualquier resultado evaluado por los participantes, como la percepción de la LD y los datos de la calidad de vida relacionados con la salud bucal.
Efectos adversos

Métodos de búsqueda para la identificación de los estudios

Búsquedas electrónicas

Cochrane Oral Health's Information Specialist conducted systematic searches in the following databases for randomised controlled trials and controlled clinical trials. There were no language, publication year or publication status restrictions:

Cochrane Oral Health's Trials Register (searched 1 February 2019) (Appendix 1);
Cochrane Central Register of Controlled Trials (CENTRAL; 2019, Issue 1) in the Cochrane Library (searched 1 February 2019) (Appendix 2);
MEDLINE Ovid (1946 to 1 February 2019) (Appendix 3);
Embase Ovid (1980 to 1 February 2019) (Appendix 4).

Subject strategies were modelled on the search strategy designed for MEDLINE Ovid. Where appropriate, they were combined with subject strategy adaptations of the highly sensitive search strategy designed by Cochrane for identifying randomised controlled trials and controlled clinical trials as described in the Cochrane Handbook for Systematic Reviews of Interventions Chapter 6 (Lefebvre 2011).

Búsqueda de otros recursos

The following trial registries were searched for ongoing studies (see Appendix 5 for details of the search terms used):

US National Institutes of Health Ongoing Trials Register ClinicalTrials.gov (clinicaltrials.gov; searched 1 February 2019);
World Health Organization International Clinical Trials Registry Platform (apps.who.int/trialsearch; searched 1 February 2019).

We searched the reference lists of included studies and relevant systematic reviews for further studies.

We checked that none of the included studies in this review were retracted due to error or fraud.

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

Obtención y análisis de los datos

Selección de los estudios

The search was designed to be sensitive and include controlled clinical trials, these were filtered out early in the selection process if they were not randomised.

Two review authors independently examined the title, keywords and abstract of reports identified through electronic searching for evidence of three criteria.

A randomised clinical trial of participants undergoing orthodontic treatment with fixed appliances.
A trial comparing the use of a fluoride‐containing product versus a non‐fluoride control or an alternative fluoride product.
A trial that assessed the prevalence or incidence of DLs either at the start and at the end of orthodontic treatment or just at the end of treatment, where the end was defined as the day of removal of the fixed appliance or as soon as possible thereafter.

For studies that appeared to meet the inclusion criteria, or for which data in the title and abstract were insufficient to allow a clear decision, the full report was obtained. We resolved disagreements by discussion.

No language restrictions were applied. Translations of foreign language articles were produced by contacts within Cochrane Oral Health.

Extracción y manejo de los datos

Two review authors extracted data independently, in duplicate, using specially designed data extraction forms. The data extraction forms were piloted on several papers and were modified as required before use. Any disagreement was discussed, and a third review author was consulted when necessary. We contacted all study authors for clarification of missing information. Data from studies in which the reporting was incomplete were not included in the analysis until the corresponding author of the study had supplied adequate clarification. If agreement could not be reached, data were excluded from the review. All studies that met the inclusion criteria underwent an assessment of the risk of bias. We extracted data from the published report; however if the report was unclear or lacking in important information then the corresponding author of the article was contacted by e‐mail. We recorded studies rejected at this or subsequent stages, along with reasons for exclusion, in the 'Characteristics of excluded studies' tables.

For each trial, we recorded the following data.

Year of publication and country of origin.
Study design.
Unit of randomisation.
Details of participants, including demographic characteristics and criteria for inclusion.
Details of types of interventions (method of delivery of fluoride, dose, duration of use).
Details of outcomes reported (number, size and severity of DLs), including method of assessment and mean duration of the study.

Evaluación del riesgo de sesgo de los estudios incluidos

This assessment was conducted by using the recommended approach for assessing risk of bias in studies included in Cochrane Reviews (Higgins 2011). We used the two‐part tool to address the six specific domains (namely, sequence generation, allocation concealment, blinding, incomplete outcome data, selective outcome reporting and other bias). Each domain includes one or more specific entries in a 'Risk of bias' table. Within each entry, the first part of the tool involves describing what was reported to have happened in the study. The second part of the tool involves assigning a judgement relating to the risk of bias for that entry: either low risk, unclear risk or high risk.

The domains of sequence generation, allocation concealment, incomplete outcome data and selective outcome reporting are addressed in the tool by a single entry for each study. For blinding, two entries were used because assessments need to be made separately for (1) participants and operators/orthodontists and (2) outcome assessors. When the operator/orthodontist assessed the outcome of the trial, this was noted. The final domain ('other sources of bias') was assessed as a single entry for studies as a whole.

Two review authors undertook the risk of bias assessment independently and in duplicate as part of the data extraction process. We resolved disagreements by discussion.

After taking into account additional information provided by the authors of the trials, review authors grouped studies into the following categories.

Risk of bias	Interpretation	Within a study	Across studies
Low risk of bias	Plausible bias unlikely to seriously alter the results	Low risk of bias for all key domains	Most information comes from studies at low risk of bias
Unclear risk of bias	Plausible bias that raises some doubt about the results	Unclear risk of bias for one or more key domains	Most information comes from studies at low or unclear risk of bias
High risk of bias	Plausible bias that seriously weakens confidence in the results	High risk of bias for one or more key domains	The proportion of information from studies at high risk of bias is sufficient to affect the interpretation of results

We completed a 'Risk of bias' table for each included study. We also presented the results graphically (Figure 1).

Medidas del efecto del tratamiento

For dichotomous outcomes, we expressed the estimate of effect of an intervention as risk ratios (RRs) together with 95% confidence intervals (CIs). For continuous outcomes, we estimated mean differences (MDs) and 95% CIs.

Cuestiones relativas a la unidad de análisis

In parallel‐group trials in which participants are randomly assigned to intervention or to control and a single outcome measure per participant is reported, the analysis is straightforward. When individuals are randomly assigned to treatment, each individual has a number of teeth exposed to the intervention or to the control. When the outcome is reported per number of teeth, the data should be adjusted for clustering within the mouth of each individual to avoid unit of analysis errors. If it was unclear from the reports of included trials whether clustering had been considered, authors were contacted to clarify how this dependence had been accounted for in the analysis.

Manejo de los datos faltantes

When data were not available in the printed report, or when data were unclear, we contacted the corresponding author of the study to obtain the missing data. The analysis generally includes only available data (ignoring missing data); we would, however, have used methods of estimating missing standard deviations as provided in Section 7.7.3 of the Cochrane Handbook for Systematic Reviews of Interventions (Higgins 2011), if appropriate. Otherwise, we did not undertake any imputations or use statistical methods to allow for missing data.

Evaluación de la heterogeneidad

Pooling of data and meta‐analysis were carried out only if sufficient similarities were noted between studies in types of participants, interventions and outcomes, including the time of the outcome measurement. If any trials were pooled, the significance of discrepancies in the estimates of treatment effects from the different trials was to be assessed by using Cochran's test for heterogeneity, by which heterogeneity was considered significant if P < 0.1 (Higgins 2011).

The I² statistic, which describes the percentage total variation across studies that is due to heterogeneity rather than to chance, was used to quantify heterogeneity, with I²greater than 50% considered to show substantial heterogeneity (Higgins 2011: Section 9.5.2).

Evaluación de los sesgos de notificación

Only a proportion of research projects conducted are ultimately published in an indexed journal and become easily identifiable for inclusion in systematic reviews. Reporting biases arise when reporting of research findings is influenced by the nature and direction of the findings of the research. We investigated and attempted to minimise in this review potential reporting biases, including publication bias, time lag bias, multiple (duplicate) publication bias and language bias.

If more than ten studies were included for one outcome, we would have constructed a funnel plot. Any asymmetry in the funnel plot indicating possible publication bias would have been investigated by statistical analysis using the methods introduced by Egger 1997 (continuous outcome) and Rücker 2008 (dichotomous outcome) (such analysis would have been done in STATA 11.0). However, insufficient trials were included in this review to enable the review authors to investigate publication bias.

Síntesis de los datos

A meta‐analysis was to be conducted only if studies of similar comparisons reported the same outcome measures. Risk ratios would have been combined for dichotomous data and mean differences for continuous data, using random‐effects models, provided more than three studies were included in the meta‐analysis.

Análisis de subgrupos e investigación de la heterogeneidad

We planned to investigate clinical heterogeneity by examining the different sources of fluoride. Provided sufficient studies were identified for each intervention and outcome, we planned a priori to conduct subgroup analyses for different sources of fluoride (mouthrinse, gel, varnish dentifrice, bracket adhesive, elastomeric ligature).

Análisis de sensibilidad

It was planned to undertake sensitivity analyses to examine the effects of quality assessment items on the assessment of overall estimates of effect. In addition, the effect on findings of the review of including unpublished literature was to be examined. However, insufficient trials were included in the review for a sensitivity analysis to be undertaken.

Summary of findings

We developed 'Summary of findings' tables for the comparisons that were considered most important for decision makers and the outcomes number of participants with new DLs on the day the fixed appliance was removed, number of participants with more severe DLs, and number of participants with adverse effects using GRADEpro GDT software (GRADEpro GDT 2015). We assessed the level of certainty in the findings with reference to the risk of bias assessments, the directness of the evidence, the inconsistency of the results, the precision of the estimates, and the risk of publication bias. The level of certainty for each of the comparisons was categorised as high, moderate, low, or very low.

Results

Description of studies

Results of the search

Searches were originally done in July 2003, then repeated in May 2012, January 2013, December 2016, January 2018 and February 2019.

The first version of this review included 14 trials, involving 613 participants. The review protocol was amended for the 2013 update to include only randomised studies (excluding quasi‐randomised or controlled clinical trials), using parallel groups and measuring relevant outcomes over the full length of orthodontic treatment. This resulted in the exclusion of all included studies from the first version, for the following reasons: five were quasi‐randomised (Banks 2000; Dyer 1982; Hirschfield 1978; Millett 2000; Sonis 1989), five were within‐person or split‐mouth designs (Chung 1998; Czochrowska 1998; Gillgrass 2001; Marcusson 1997; Twetman 1997) and three had ex vivo outcomes on extracted teeth (Gorton 2003; Ogaard 1986; Pascotto 2004). Ogaard 2001 was excluded because investigators compared fluoride versus fluoride plus an antiseptic solution.

The 2013 update included three studies (Luther 2005; Ogaard 2006; Stecksén‐Blicks 2007), involving 458 randomised participants (391 analysed). For the 2019 update a further seven studies were identified for inclusion (Benson 2019; He 2010; Jiang 2013; Jost‐Brinkman 2017; Sonesson 2014; Sonesson 2019; van der Kaaij 2015). Three ongoing studies were also identified (DRKS00012533; DRKS00012540; IRCT2016122531558N1).

For details of the studies examined and reasons for inclusion or exclusion, please see Characteristics of included studies and Characteristics of excluded studies tables. The search process and results are presented as a flow chart in Figure 2.

Figure 2

Study flow diagram.

Included studies

Characteristics of the trial participants and settings

Seven of the included studies were conducted in Europe, involving Sweden (Ogaard 2006; Sonesson 2014; Sonesson 2019; Stecksén‐Blicks 2007), the UK (Luther 2005), the UK and Republic of Ireland (Benson 2019), and the Netherlands (van der Kaaij 2015). One study was undertaken in Germany and Israel (Jost‐Brinkman 2017) and two in China (He 2010; Jiang 2013). Participant age ranged from 10 years at the start of treatment (Jost‐Brinkman 2017) to 60 years (Jost‐Brinkman 2017). All participants in the included trials were recruited at the start of their orthodontic treatment with fixed appliances and were followed until their fixed appliances were removed.

Characteristics of the interventions

We have grouped the included trials into three broad comparisons.

Dentist or nurse‐applied fluoride in the form of varnish (He 2010; Sonesson 2019; Stecksén‐Blicks 2007), foam (Jiang 2013), and gel (Jost‐Brinkman 2017), both home use once a week and professional application every three months.
Patient‐applied/used fluoride in the form of toothpaste/mouthrinse combinations (Ogaard 2006; Sonesson 2014; van der Kaaij 2015).
Fluoride‐releasing materials in the form of intraoral fluoride‐releasing glass beads device (Luther 2005) and resin‐modified glass ionomer cement for bonding orthodontic brackets (Benson 2019).

One study (Jost‐Brinkman 2017) investigated both a dentist/nurse‐applied and a patient‐applied fluoride intervention in the same participants. Adherence to home use is more difficult to assess than adherence to dentist or nurse‐applied therefore it was categorised in the latter. There were no studies investigating dietary fluoride supplementation.

Five studies were placebo‐controlled, comparing a fluoride‐containing product with the same product not containing fluoride, where participant, clinician/operator and assessor were all masked as to group allocation (Jiang 2013; Jost‐Brinkman 2017; Sonesson 2019; Stecksén‐Blicks 2007; van der Kaaij 2015). In four studies just the assessor was masked for group allocation, due to differences in the nature of the interventions. One study was a comparison of fluoride varnish versus placebo (He 2010). One study was a comparison of two products containing fluoride (Sonesson 2014), one was a comparison of two methods of delivering fluoride (Luther 2005) and one study compared one fluoride and one non‐fluoride containing bonding material (Benson 2019). One study compared two fluoride products, but the masking was unclear (Ogaard 2006).

Characteristics of the outcomes

Seven studies reported our primary outcome of number of participants in each group with new demineralised lesions (DLs) (Benson 2019; Jost‐Brinkman 2017; Luther 2005; Sonesson 2014; Sonesson 2019; Stecksén‐Blicks 2007; van der Kaaij 2015). One study reported the number of teeth in each group with new DLs (Jiang 2013), one study reported a mean index score (Ogaard 2006), and in one study the outcome was unclear (He 2010).

All reported some outcomes on the severity of DLs in terms of differences in the number, size, colour, or quantitative measurement of mineral loss. None of the included studies reported outcomes of participant perception of their DLs or oral health‐related quality of life, and only one study reported adverse effects (Sonesson 2019).

Excluded studies

The details and reasons for exclusion are outlined in the Characteristics of excluded studies table. The main reasons for exclusion were that participants were not followed to the end of their orthodontic treatment and an inappropriate research methodology was used (within‐person or split‐mouth design).

Risk of bias in included studies

Overall risk of bias assessments for all included studies are shown in Figure 1. Two studies were assessed at low risk of bias for all domains (Sonesson 2019; Stecksén‐Blicks 2007). Two studies were assessed as at unclear risk of bias in one or two domains (Benson 2019; Sonesson 2014) and four in more than two domains (He 2010; Jiang 2013; Jost‐Brinkman 2017; Ogaard 2006). Two studies were assessed as at high risk of bias in one domain (Luther 2005; van der Kaaij 2015) due to a high proportion of participants withdrawing or dropping out of the studies. The authors of the studies have been contacted for further information, but at the time of publication we are awaiting clarification from the authors of two studies (Jost‐Brinkman 2017; Ogaard 2006) and one is currently uncontactable (He 2010).

Allocation

We assessed six of the included studies at low risk of bias, because the method of sequence generation and clear allocation concealment were considered adequate from either the description in the reported or following further clarification from the study authors (Benson 2019; Luther 2005; Sonesson 2014; Sonesson 2019; Stecksen‐Blicks 2007; van der Kaaij 2015).

Three studies (He 2010; Jiang 2013; Ogaard 2006) reported the method of sequence generation (randomisation table), but did not mention how they achieved equal numbers of participants in each group (since clarified for Jiang 2013) or method of allocation concealment. One study described the method of allocation concealment, but did not report the method of sequence generation (Jost‐Brinkman 2017). They were all assessed as at unclear risk of selection bias.

Blinding

In five studies the participant, clinician and assessor were all considered masked to group allocation, because the authors reported that the substances provided to the active and control groups were identical in presentation, taste, appearance and consistency (Jiang 2013; Jost‐Brinkman 2017; Sonesson 2019; Stecksén‐Blicks 2007; van der Kaaij 2015). Due to the nature of the materials used in three studies it was not possible to mask the clinician as to group allocation and the participant could have guessed (Benson 2019; Luther 2005; Sonesson 2014). It is not clear how knowledge of group allocation might affect the behaviour of the clinician and participant, so these were judged to be an unclear risk of performance bias. In one study the masking was not clear (Ogaard 2006) and the author has been contacted. In one study, the participants received different preventative regimens, at different times, therefore neither participants or operators could be masked (He 2010).

Nine studies reported that the outcome assessors were masked as to group allocation. In four of these studies more than one expert judge was involved in assessing before treatment and day of debond or close to debond photographs for new DLs (Benson 2019; Sonesson 2014; Sonesson 2019; Stecksén‐Blicks 2007). In two studies only one assessor was involved, but using relatively objective measures of demineralisation (image analysis (Luther 2005), and quantitative light‐induced fluorescence (QLF) (van der Kaaij 2015)). These were judged to be at low risk of detection bias.

We are still awaiting clarification from Jost‐Brinkman 2017 about who and how many undertook the clinical assessments, whether they were calibrated and whether any assessment of reproducibility was undertaken.

Jiang 2013 reported that assessments were undertaken clinically by one masked assessor. Although the assessor was described as having participated in the 'initial calibration trial' and intraexaminer agreement was assessed before the start of the trial, it is not clear if the repeatability was good for the full length of the trial and whether the assessor agreed with other calibrated assessors. This study was therefore assessed to be an unclear risk of detection bias.

Ogaard 2006 indicates that clinical assessments and colour photographs were obtained prior to bonding and at debond, but does not state that the photographs were used in the assessment. There are no details about who carried out the assessments or whether they were calibrated. No repeatability or reproducibility assessments are reported and this study was therefore judged to be an unclear risk of detection bias.

In He 2010 the number of assessors was unclear. It was also judged to be an unclear risk of detection bias.

Incomplete outcome data

The proportion of post‐randomisation exclusions was reported as zero in one study (He 2010), and low in four studies (Benson 2019; Jiang 2013; Sonesson 2019; Stecksén‐Blicks 2007). The numbers and reasons for exclusion were similar for each group and the risk of attrition bias and the effect of this on the interpretation of the findings is outlined in the 'Risk of bias' tables for each study. The report by Benson 2019 had a flow diagram and fully reported the reasons for withdrawal or dropout, which were about equal in the two groups. Two studies were assessed as an unclear risk of attrition bias (Ogaard 2006; Sonesson 2014). The reasons for this are given in the 'Risk of bias' tables for each study. The report by Ogaard 2006 had neither a flow diagram, nor a detailed explanation of the reasons for withdrawal or dropout. Sonesson 2014 had a flow diagram and reported the reasons for withdrawal or dropout, however it reports that a number of participants "did not comply with the study protocol" and were excluded from the analysis, but the definition of non‐compliance is unclear. Due to inconsistencies on the number of exclusions from analyses and lack of clarity on non‐compliance Jost‐Brinkman 2017 was also assessed as at unclear risk of attrition bias. Two studies were assessed as at high risk of attrition bias, because a significant proportion of the participants who were randomised were not included in the analysis (47% (Luther 2005), 33% (van der Kaaij 2015)).

Selective reporting

Ogaard 2006 reported the change in the mean Gorelick Index scores, as well as the numbers of teeth with new DLs and was assessed as at unclear risk of reporting bias. The report by Luther 2005 had some information missing and the denominators were not stated, so this study was assessed at unclear risk of reporting bias. In He 2010 the definition of incidence is unclear (difference between start and finish) and it does not state how they took into account the clustering of teeth within the mouth, therefore the study was also assessed at unclear risk of reporting bias.

Other potential sources of bias

Four studies reported data for the number of participants with new DLs at or near the date of debond (Benson 2019; Sonesson 2014; Sonesson 2019; Stecksén‐Blicks 2007). All these studies assessed the number of new DLs (incidence) by comparing pre‐treatment and at or near debond photographs and were considered to be at low risk of other potential sources of bias. Other potential sources of bias are discussed for these studies, but they were assessed at low potential risk of bias, as was the study by Ogaard 2006. Luther 2005 was assessed at unclear risk of other bias, because of possible differences between the groups in terms of compliance, duration of orthodontic treatment and exposure to topical fluorides. van der Kaaij 2015 was also assessed as at unclear risk of bias, because it is unclear if all the QLF images were collected and analysed from the day of debond. The authors state that the "WSL assessments were made at an average of 52 d (days) after debonding (with a range of 0‐156 d)." Those undertaken at five months following debond might have undergone quite extensive remineralisation unrelated to mouthrinse use. Jiang 2013 was also assessed as unclear risk of bias in this domain, because it was not clear from the report how many operators were involved in the study and how the investigators controlled for other sources of fluoride. We are still awaiting clarification about other sources of bias from Jost‐Brinkman 2017 and were not able to contact the authors of He 2010.

Effects of interventions

The 10 studies included in this review evaluated different modes of fluoride application, which we have categorised into three broad methods.

Professionally‐applied (dentist or nurse‐applied) fluoride in the form of varnish (He 2010; Sonesson 2019; Stecksén‐Blicks 2007), foam (Jiang 2013) and gel (Jost‐Brinkman 2017), both home use once a week and professional application every three months.
Patient‐applied/used fluoride in the form of toothpaste/mouthrinse combinations (Ogaard 2006; Sonesson 2014; van der Kaaij 2015).
Fluoride‐releasing materials in the form of intraoral fluoride‐releasing glass bead device (Luther 2005) and resin‐modified glass ionomer cement for bonding orthodontic brackets (Benson 2019).

Although the study by Jost‐Brinkman and colleagues (Jost‐Brinkman 2017) investigated both a dentist/nurse‐applied and a patient‐applied fluoride intervention in the same participants, we have placed this in the dentist or nurse‐applied category for the reasons explained previously. Summaries of the methods, participants, interventions and outcomes for each study are provided in the Characteristics of included studies tables.

Dentist or nurse‐applied fluoride

Primary outcome

Stecksén‐Blicks 2007 reported a prevalence of DLs before treatment of 4.3% in participants who received the professionally‐applied fluoride varnish (experimental group) and 4.0% in those receiving the professionally‐applied non–fluoride placebo varnish (control group). At debond the proportions of participants with DLs was 11.7% in the experimental group and 29.7% in the control group, which the authors state as an incidence of 7.4% and 25.7% respectively. The authors report a risk reduction or preventive fraction (1‐RR) of 0.69 for new DLs. The authors calculated that nearly 6 (5.5) individuals would need to receive the fluoride varnish every orthodontic visit (approximately every six weeks) to prevent one patient from having a new DL (number needed to treat or NNT). This result should be treated with caution, because another trial by Sonesson and colleagues (Sonesson 2019) using a slightly lower concentration of fluoride varnish (7000 parts per million (ppm) fluoride) had different findings. They reported a prevalence of DLs before treatment of 8.0% (6 out of 75) in participants who received the professionally‐applied fluoride varnish (experimental group) and 9.6% (7 out of 73) in those receiving the professionally‐applied non‐fluoride placebo varnish (control group). At debond the proportions of participants with DLs (index scores of 2 and above) was 41.3% (31 out of 75) in the experimental group and 43.8% (32 out of 73) in the control group. Combining the data from these two studies the calculated risk ratio (RR) for the development of new lesions with professionally‐applied fluoride varnish was 0.52 (95% confidence interval (CI) 0.14 to 1.93; 405 participants; Analysis 1.1), which is a non‐significant reduction.

Jost‐Brinkman 2017, who applied fluoride gel every three months (as well as home‐used gel), found no difference in the development of DLs between those who were allocated to the fluoride gel and those to the placebo gel (RR 0.99, 95% CI 0.76 to 1.27; 312 participants; Analysis 2.1).

Jiang 2013 found that applying a fluoride‐containing foam professionally every two months reduces the incidence of new DLs after fixed orthodontic treatment (RR 0.26, 95% CI 0.11 to 0.57; 95 participants analysed; Analysis 3.1).

As previously stated we were not able to contact the authors of He 2010.

Secondary outcomes

Stecksén‐Blicks 2007 stated that they calculated a "progression score" by "subtracting the debond score from that registered at baseline." The progression scores were mean 0.8 (standard deviation (SD) 2.0) for participants who received the fluoride varnish and mean 2.6 (SD 2.8) for those who received the placebo varnish. We interpret this as the control group having a greater number of teeth with lesions, as well as a greater incidence. In terms of severity Figure 2 in the study shows that the great majority of new lesions were minor (Gorelick Index score 2: slight white spot formation, thin rim) and the aesthetic impact of the new DLs was not assessed.

Sonesson 2019 reported a reduction in the number of participants with the more severe DLs (index score of 3 or 4) at debond in the experimental group who received the fluoride varnish (12%, 9 out of 75), compared with the non‐fluoride control (26%, 19 out of 73). This is a RR of 0.46 (95% CI 0.22 to 0.95) and a number needed to treat of 7 (but with a wide confidence interval 3.80 to 71.10).

Jiang 2013 reported that the "mean net increment" in the scores was 0.7 (SD 2.8) for those who received the fluoride foam and 4.4 (SD 5.4) for those who received the non‐fluoride foam. This is a large increase, which we interpret as those not receiving the fluoride foam having a greater number of teeth affected by DLs. The authors cite a NNT, but we believe this is calculated on the number of teeth with new DLs, not participants, which is not appropriate. Again, the severity scores (Figure 2) were mostly minor (Gorelick Index 2) and the aesthetic impact is uncertain.

There were no data reported from any of the studies about participant perception of their DLs or any measure of oral health‐related quality of life. Only Sonesson 2019 reported adverse effects and that one participant who received the fluoride varnish withdrew from their study, due to a feeling of slight nausea during the trial.

Patient‐applied/used various fluoride toothpaste/mouthrinse combinations

Primary outcome

Sonesson 2014 found that the proportions of participants with DLs before orthodontic treatment were 17% in the high concentration fluoride (5000 ppm) toothpaste group and 19% in the standard concentration fluoride (1450 ppm) toothpaste group. These proportions increased to 35% and 45% respectively, on the day of debond. The calculated RR of developing a new DL is 0.68 (95% CI 0.46 to 1.00; 380 participants; Analysis 4.1). Although the upper confidence interval contacts the line of no difference, we believe that this provides evidence that the use of high fluoride toothpaste by patients wearing fixed orthodontic appliances does provide some protection against new DLs, with a risk reduction or preventive fraction (1‐RR) of 0.32 and a NNT of 12. As with the trial by Stecksén‐Blicks 2007, however, the results of this trial should be interpreted with caution until further clinical trials confirm this finding.

van der Kaaij 2015 states that none of their participants had DLs before the start of treatment. They report that 11 of 36 participants in the fluoride mouthrinse group developed at least one new DL, compared with 21 of 45 participants in the non‐fluoride mouthrinse group. However, the calculated RR is 0.65 (95% CI 0.37 to 1.17; 81 participants; Analysis 5.1) and the confidence interval crosses the line of no difference. The lack of statistical significance could be due to the small number of participants and the large proportion of withdrawals and dropouts.

Ogaard 2006 reported the outcome of new DLs at the tooth level with no indication of correction for clustering of teeth within the mouth. The author has been contacted to provide these data at the participant level.

Secondary outcomes

Sonesson 2014 reported that the mean numbers of DLs before the start of treatment were 0.3 (SD 1.0) in the high concentration fluoride (5000 ppm) toothpaste group and 1.0 (SD 1.8) in the standard concentration fluoride (1450 ppm) toothpaste group. These increased to 0.4 (SD 1.0) and 1.2 (SD 1.8) respectively, at debond. They also stated that the "vast majority of all new WSL were thin rims (score 2) in both groups," with only 1.2% in the high fluoride toothpaste group and 2.3% in the standard fluoride toothpaste group having teeth affected with the more severe Gorelick Index scores of 3 or 4. The aesthetic impact of the DLs is, therefore, unclear. Like most studies they report that lateral incisors were the most frequently affected teeth in both groups, followed by the canines and premolars.

van der Kaaij 2015 reported that the number of new DLs ranged from 1 to 5 in participants who received the fluoride mouthrinse (experimental group) and 1 to 15 in those receiving the non‐fluoride mouthrinse (control group). The quantitative light‐induced fluorescence (QLF) measurements revealed a mean fluorescence loss (delta F; DL – sound enamel) of 11.6% (SD 5.0) in the experimental group and 10.3% (SD 3.0) in the control group. The lesion depths were 0.9 mm² (SD 0.6) and 1.3 mm² (SD 1.6 mm) respectively. Neither were reported as statistically significant and there is no indication in the report how these QLF measurements can be interpreted in terms of aesthetic impact.

Ogaard 2006 found a greater mean change in the 'white spot lesion index' from baseline in a group of participants using a neutral sodium fluoride toothpaste (1400 ppm, pH 6.7) twice daily and a sodium fluoride mouthrinse (250 ppm fluoride (F), pH 6.3) at night compared with a group using amine fluoride/stannous fluoride toothpaste (Meridol 140 ppm F, pH 4.5) twice daily and an amine fluoride/stannous fluoride mouthrinse (250 ppm F, pH 4.0) after toothbrushing at bedtime (97 participants; Analysis 6.1). This suggest that the sodium fluoride combination was less effective than the amine fluoride/stannous fluoride. They also found a slightly larger increase in both the visible plaque index and the gingival bleeding index over the duration of treatment in the group exposed to sodium fluoride (97 participants; Analysis 6.2; Analysis 6.3). These differences, however, should be interpreted cautiously until the results can be independently replicated.

There were no data reported from any of the studies about participant perception of their DLs, any measure of oral health‐related quality of life, and/or reports of adverse effects.

Fluoride‐releasing materials

Primary outcome

Benson 2019 did not assess the number of DLs at baseline, rather the assessors were asked to look concurrently at the before treatment and day of debond images and determine if they saw new DLs. On this basis 23 out of 88 participants who received the fluoride‐containing bonding material (resin‐modified glass ionomer cement) were assessed as having new DLs versus 19 out of 85 who received the material containing no fluoride (light‐cured composite resin). The calculated RR was 1.17 (95% CI 0.69 to 1.99; 173 participants; Analysis 7.1); therefore, there was no difference statistically or clinically in the proportion of participants with new DLs between the two groups.

Luther 2005 reported that, of those participants completing the trial, 10 of the 18 receiving the fluoride‐containing glass beads developed DLs, compared with 7 out 19 receiving the fluoride mouthrinse. The calculated RR was 1.51 (95% CI 0.73 to 3.10; 37 participants; Analysis 8.1), which is not statistically significant.

Secondary outcomes

Unlike any of the other studies, Benson 2019 undertook an assessment of the aesthetic impact of the new DLs. They found that in only 15 of the 42 participants with new DLs were the lesions judged to be of aesthetic concern by a majority of the expert and lay assessors. Therefore, the incidence of aesthetically displeasing new DLs was 9% compared with the overall incidence of 24% and there was no difference between the groups (fluoride material 8 out of 23; non‐fluoride material 7 out of 19). Like other studies they found that upper lateral incisors were affected most, followed by upper central incisors and upper canines. This study also reported the proportion of bond failures, which was not an outcome in this review.

Luther 2005 reported that, for those participants who completed their trial, there was no statistically significant difference in the numbers of teeth affected by DLs between the two groups (fluoride‐releasing beads 17 out of 108 teeth; fluoride mouthrinse 19 out of 114).

There were no data reported from any of the studies about adverse effects, participant perception of their DLs and/or any measures of oral health‐related quality of life.

Discusión

available in

Resumen de los resultados principales

Hay evidencia insuficiente de una diferencia a partir de dos ensayos, considerados en riesgo bajo de sesgo, sobre la efectividad del barniz de fluoruro aplicado por un profesional cada seis semanas para reducir las nuevas lesiones desmineralizadas (LD) durante el tratamiento con aparatos de ortodoncia fijos (Resumen de resultados, Tabla 1). La evidencia de un ensayo considerado en riesgo incierto de sesgo, proporcionó un nivel bajo de certeza de que la espuma de fluoruro (12 300 partes por millón [ppm] de fluoruro), aplicada profesionalmente cada dos meses, reduce la incidencia de nuevas LD después del tratamiento con aparatos de ortodoncia fijos (Resumen de resultados, Tabla 2). De manera similar, un ensayo, evaluado como en riesgo incierto de sesgo, proporcionó un nivel bajo de certeza de que el uso de una pasta dental fluorada de alta concentración por parte de los pacientes durante su tratamiento ortodóntico fijo reduce la incidencia de nuevas LD, en comparación con una pasta dental fluorada de concentración convencional (Resumen de resultados, Tabla 3).

La evidencia es insuficiente para demostrar si el uso de un gel de fluoruro de amina aplicado profesionalmente reduce las nuevas LD (Tabla Adicional 1), o si el uso de un enjuague bucal de fluoruro de sodio/ fluoruro de amina por parte del paciente (Resumen de resultados, Tabla 4) o de una combinación de pasta dental/enjuague bucal de fluoruro de amina y fluoruro de estaño (Tabla Adicional 2) es más o menos efectivo que el uso de un enjuague bucal sin fluoruro o una combinación de pasta dental/enjuague bucal de fluoruro de sodio, respectivamente. Finalmente, no hay evidencia de que un material que contenga fluoruro para adherir los brackets (Resumen de resultados, Tabla 5) o un dispositivo intraoral de microesferas de vidrio que libere fluoruro (Tabla Adicional 3) reduzcan la incidencia de nuevas LD.

Compleción y aplicabilidad general de las pruebas

Al igual que con la última versión, la actualización de esta revisión ha incluido solo ensayos de grupos paralelos, en los que el participante individual es la unidad de aleatorización. Lo anterior se decidió debido a la posibilidad de contaminación cruzada entre los dientes experimentales y los de control en la misma boca, ya sea entre el arco superior e inferior o entre los lados de la boca, lo que podría dar lugar a una subestimación de la efectividad de cualquier producto de fluoruro.

Las intervenciones que dependen del paciente para la administración, como el enjuague bucal con fluoruro y la pasta de dientes, solo funcionarán si se utilizan con regularidad. Dependen en gran medida del cumplimiento del paciente para tener éxito; sin embargo, la evidencia sugiere que el cumplimiento con el enjuague bucal es deficiente entre los pacientes que utilizan ortodoncia. Un estudio (Geiger 1992) encontró que solo el 42% de los participantes utilizaron un enjuague bucal de fluoruro de sodio al menos cada dos días. Los resultados también mostraron que los que cumplían menos con los regímenes de enjuague con fluoruro tendieron a presentar más LD. Es importante considerar la aceptabilidad de las intervenciones tanto para los adolescentes como para los adultos con el fin de aumentar el cumplimiento de las prácticas de higiene dental recomendadas.

Las intervenciones que se aplican profesionalmente y administran fluoruro de forma "pasiva", como el barniz de fluoruro, los cementos que liberan fluoruro en los brackets y los elásticos que liberan fluoruro, evitan la necesidad de cumplimiento del paciente. Además, estos materiales administran fluoruro cerca del bracket, donde más se necesita. Muchos materiales fluorados liberan grandes cantidades de fluoruro inicialmente, aunque el nivel disminuye rápidamente y puede no ser suficiente para prevenir la caries durante el ciclo completo del tratamiento ortodóntico. Es probable que sea necesario volver a aplicar el barniz de fluoruro y reemplazar frecuentemente los elásticos que liberan fluoruro. En el ensayo de grupos paralelos de una intervención con barniz de fluoruro incluida en esta revisión, el barniz se volvió a aplicar cada seis semanas en cada cita de control ortodóntico. No se encontraron ensayos de grupos paralelos de cementos o elásticos que liberan fluoruro que cumplieran con los criterios de inclusión para esta revisión.

Es interesante agregar que desde que se llevó a cabo la revisión inicial, se han desarrollado materiales que producen una liberación lenta y sostenida de fluoruro (Luther 2005). Este ensayo fue pequeño y tuvo un riesgo alto de sesgo, y la evidencia fue insuficiente para revelar si estos dispositivos son más o menos efectivos que un enjuague bucal para reducir el desarrollo de LD. Es posible que con un mayor refinamiento, esta técnica pueda ser potencialmente efectiva. Los dispositivos intraorales que liberan fluoruro deben evaluarse mediante ensayos controlados aleatorizados de grupos paralelos con un enmascaramiento apropiado de los participantes, los médicos/operadores y los asesores.

Al examinar la efectividad de un producto de fluoruro en la prevención de la caries dental, se deben considerar dos aspectos: primero, si el producto de fluoruro reduce el número de LD que aparecen durante el tratamiento, y segundo, si reduce la gravedad de las LD en términos del tamaño o área de la superficie dental afectada, la cantidad de mineral perdido o la profundidad de la caries. Banks et al (Banks 2000) desarrollaron el Enamel Decalcification Index, que es un índice ordinal que incluye una evaluación del área cubierta. La evaluación del tamaño de la lesión es una medida de resultado útil, aunque ninguno de los estudios incluidos en esta revisión informó este resultado.

Lo ideal es que la apariencia del diente se registre antes y después del tratamiento ortodóntico, de modo que se mida el cambio en la apariencia del diente (incidencia), no solo su apariencia al final (prevalencia). Existen muchas causas diferentes de las marcas de color en los dientes, muchas de las cuales ocurren durante su desarrollo. Es importante que estas lesiones del desarrollo, así como las caries que se han producido antes de la colocación del aparato ortodóntico, se excluyan del análisis, por lo cual se deben tomar fotografías clínicas o imágenes fluorescentes antes del tratamiento. La medición tanto de la incidencia como de la gravedad dependerá del método utilizado para registrar las LD. Se pueden utilizar dos métodos principales: la inspección visual y las imágenes clínicas. Ambos métodos están asociados con problemas. Un problema con la inspección visual es que el examinador o los examinadores requerirán calibración al inicio y recalibración regular durante el período experimental para asegurar la consistencia de la medición. La duración del experimento, incluyendo el reclutamiento y la recolección de datos, será larga debido a que, como se discutirá más adelante, el producto debe ser evaluado a lo largo de todo el tratamiento ortodóntico. El mismo puede tener una duración de entre 18 y 30 meses, a veces más. Otro problema con el registro visual clínico implica enmascarar al evaluador con respecto a la intervención asignada. Para reducir el sesgo, el examinador no debe saber si el participante ha recibido un producto de fluoruro lo cual complicará la forma en que se realiza el experimento.

Las imágenes tienen la ventaja de proporcionar un registro permanente de la apariencia del diente. Las evaluaciones pueden ser llevadas a cabo por varias personas de forma independiente o en grupos, con lo que se logra un consenso. Las imágenes pueden ser colocadas en orden aleatorio y los jueces enmascarados para la asignación de los grupos. Además, debido a que la evaluación puede realizarse en un período corto de tiempo, se reduce el problema de la deriva del examinador, por la cual un evaluador puede cambiar sutilmente su evaluación con el tiempo. El reto del uso de fotografías clínicas consiste en lograr consistencia en la iluminación y la reducción de los reflejos que pueden enmascarar o imitar LD. Sin embargo, cuando se aplica una técnica fotográfica cuidadosa, las ventajas de las fotografías superan sus desventajas potenciales. Existen varios métodos ópticos y fluorescentes para medir las lesiones en los dientes (Angmar‐Mansson 1996). Estos métodos requieren un equipo especializado, que aumentaría de manera considerable el coste de un estudio clínico, aunque proporcionan una medición objetiva de la cantidad de caries en términos de la pérdida de minerales o de la profundidad de la lesión, o ambos.

Calidad de la evidencia

Se consideró que dos estudios incluidos de esta revisión estaban en riesgo bajo de sesgo en todos los dominios; sin embargo, fue necesario que algunos de los autores aclararan algunas cuestiones antes de poder confirmar estos juicios. Tanto el diseño como el informe de los ensayos sobre los fluoruros para la prevención de las LD han mejorado desde la primera versión de esta revisión y se espera que esta posibilidad pueda continuar. Sin duda, el creciente uso de las guías CONSORT para el informe de los ensayos controlados aleatorizados ha sido un factor.

Se consideró que varios de los estudios presentaron algún riesgo de sesgo como resultado de que la proporción de participantes aleatorizados a la intervención no se incluyó en el análisis después del retiro o el abandono. Es inevitable que algunos pacientes se alejen durante un ciclo de tratamiento ortodóntico, aunque se debe hacer todo lo posible para recopilar registros y datos en los momentos adecuados, incluso si el tratamiento se interrumpe antes de tiempo.

Otro problema en la investigación ortodóntica es que a menudo no es posible enmascarar al médico/operador a la intervención asignada. En estas situaciones, el participante también puede determinar a qué grupo ha sido asignado y, por lo tanto, el triple cegamiento es difícil de lograr, si no imposible. Sin embargo, no está claro en qué medida este hecho afectará el comportamiento del médico y del participante. Se sabe que los pacientes responden de forma favorable al hecho de simplemente formar parte de un ensayo clínico, independientemente de si han recibido la intervención activa o el control (McCarney 2007). Este es el llamado efecto Hawthorne, que debe su nombre a los trabajos de una empresa eléctrica de Chicago, cuyos trabajadores fueron sometidos a una serie de experimentos durante las décadas de 1920 y 1930. Los investigadores notaron que sin importar cómo cambiaban las condiciones de trabajo de los trabajadores, su productividad mejoraba, y establecieron la conclusión de que este hecho se debía a que sabían que eran parte de un experimento. Se cree que sería duro juzgar todos los ensayos que no son triple ciego como en riesgo potencial alto de sesgo cuando no se conocen los efectos y quizás esta área debería investigarse en el futuro.

Sesgos potenciales en el proceso de revisión

Se realizó una búsqueda sensible en varias fuentes electrónicas, complementada con búsquedas en listas de referencias. No se impusieron restricciones de idioma ni de estado de publicación. Los autores de la revisión han intentado, en la medida de lo posible, identificar todos los estudios posibles que pudieran cumplir con los criterios de inclusión para esta revisión. Se estableció contacto con los autores de los estudios y muchos de ellos han respondido; sin embargo, algunos no han podido proporcionar la información solicitada, ya que sus registros habían sido destruidos o se habían perdido.

Cuando un producto, como un material de adhesión, puede aplicarse a un solo diente, es tentador utilizar un diseño experimental en el que el material que se está evaluando se utiliza en dos cuadrantes de la boca y el material de control se utiliza en los otros dos cuadrantes. Este diseño se denomina «de boca dividida». La principal ventaja del diseño de boca dividida sobre un diseño de estudio convencional de grupos paralelos, en el que los dos materiales se evalúan en dos grupos separados de individuos, es que el material experimental se examina en la misma boca, bajo las mismas condiciones que el material de control. En teoría, cualquier diferencia en el resultado entre los dos materiales se debe solo a sus propiedades, no a otros factores, como las diferencias en la higiene oral y la dieta entre los participantes (con un diseño paralelo) o incluso las diferencias en la higiene oral y la dieta a lo largo del tiempo dentro de los mismos participantes (con un diseño cruzado).

Desafortunadamente, cuando se examina la capacidad de los productos de fluoruro para reducir la caries, es altamente improbable que el fluoruro liberado se limite solo a los cuadrantes/dientes en los que se ha colocado el material experimental, y es inevitable cierta contaminación de los dientes "no tratados". Esta contaminación reducirá la diferencia en los resultados entre los dientes tratados y los no tratados. La versión anterior de esta revisión incluyó estudios de boca dividida, que no lograron mostrar ninguna diferencia entre los dientes tratados y los no tratados, lo cual puede deberse a la contaminación cruzada entre el lado de control y el lado experimental y puede reflejar lo que se cree sobre la contaminación. Por esta razón, se decidió excluir los estudios de boca dividida de la actualización anterior de la revisión.

Acuerdos y desacuerdos con otros estudios o revisiones

En la bibliografía se han informado otras tres revisiones sistemáticas que reúnen evidencia sobre los medios más efectivos para prevenir la caries/desmineralización durante el tratamiento con aparatos ortodónticos fijos. Derks et al (Derks 2004) examinaron todas las medidas preventivas de la desmineralización, no solo los productos fluorados. Estos autores de la revisión también tuvieron que excluir muchos estudios publicados debido a un diseño inadecuado de la investigación o a un informe deficiente, y no pudieron proporcionar recomendaciones sólidas y basadas en la evidencia en cuanto a la prevención de las LD durante el tratamiento con aparatos de ortodoncia fijos.

Una segunda revisión sistemática (Chadwick 2005) investigó la efectividad de los fluoruros tópicos utilizados solos para prevenir la desmineralización durante el tratamiento ortodóntico. Estos revisores incluyeron siete estudios en su revisión; sin embargo, estos estudios fueron excluidos de la revisión debido a que los resultados no eran apropiados (DMFT/DMFS [dientes/superficies cariadas, perdidas y obturadas]), o los participantes no fueron examinados inmediatamente después de la extracción de los aparatos fijos. Aunque sugieren que de acuerdo con su medida de resultado (fracción preventiva), alguna evidencia muestra que el agregado de un preparado de fluoruro tópico ayuda a prevenir la desmineralización durante el tratamiento con aparatos de ortodoncia fijos, esta conclusión debe considerarse con cautela, debido a que estos autores de la revisión no pudieron calcular los intervalos de confianza. Los autores de la revisión apoyan su solicitud de que los investigadores diseñen e informen sus estudios utilizando resultados estándar, de modo que en el futuro se puedan agrupar los datos y se puedan proporcionar recomendaciones generales sobre las medidas preventivas.

Rogers et al (Rogers 2010) incluyeron 10 estudios en su revisión sistemática que investigaba la efectividad de los adhesivos que contienen fluoruro utilizados en los aparatos de ortodoncia para prevenir la desmineralización. Cinco de estos estudios se excluyeron de la revisión debido a que no eran aleatorizados, y otros tres estudios se excluyeron debido a que los datos del informe fueron insuficientes y los autores de los estudios, cuando se estableció contacto con ellos, no pudieron proporcionar los datos solicitados. Las conclusiones de Rogers son consistentes con las de esta revisión con respecto al diseño de los ensayos y la calidad de los informes y los análisis estadísticos.

Wang 2013 incluyó 20 estudios (19 artículos) en su revisión, que está escrita en chino y solo incluye datos de estudios chinos. Incluyeron estudios con diseños cuasialeatorizados y no está claro si incluyeron estudios con un diseño de boca dividida (en la misma persona) y los que no realizaron el seguimiento de los pacientes durante el ciclo completo del tratamiento ortodóntico.

También se examinaron las siguientes revisiones: Höchli 2017; Lapenaite 2016; Lopatiene 2016; Nascimento 2016; Rahimi 2017; Sardana 2019a; Sardana 2019b; Tasios 2019. Estas revisiones investigaron intervenciones distintas del fluoruro para prevenir las LD, como el tratamiento de las LD con fosfopéptido de caseína‐fosfato de calcio amórfico (CCP‐ACP, por sus siglas en inglés), después de la extracción del aparato o incluyeron estudios cuasialeatorizados y no aleatorizados, estudios con un seguimiento corto o resultados que no son particularmente pertinentes para los pacientes y los médicos. No se identificaron estudios nuevos a partir de las referencias de estos artículos para incluir en esta revisión.

Figure 1

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

Figure 2

Study flow diagram.

Analysis 1.1

Comparison 1 Professionally‐applied F varnish versus non‐F (placebo) varnish, Outcome 1 Number of participants with new DLs.

Analysis 1.2

Comparison 1 Professionally‐applied F varnish versus non‐F (placebo) varnish, Outcome 2 Number of participants with more severe DLs (scores 3 or 4 versus scores 1 or 2).

Analysis 2.1

Comparison 2 12,500 ppm F (NaF/olaflur/dectaflur) gel versus 0 ppm F placebo gel, Outcome 1 Number of participants with new DLs.

Analysis 3.1

Comparison 3 12,300 ppm F APF foam versus 0 ppm F placebo foam, Outcome 1 Number of participants with new DLs.

Analysis 4.1

Comparison 4 5000 ppm F toothpaste versus 1450 ppm F toothpaste, Outcome 1 Number of participants with new DLs.

Analysis 5.1

Comparison 5 250 ppm F mouthrinse (100 ppm AmF/150 ppm NaF) versus 0 ppm F placebo mouthrinse, Outcome 1 Number of participants with new DLs.

Analysis 6.1

Comparison 6 1400 ppm/250 ppm F (AmF/SnF) versus 1400 ppm/250 ppm F (NaF) toothpaste/mouthrinse combinations, Outcome 1 White spot index.

Analysis 6.2

Comparison 6 1400 ppm/250 ppm F (AmF/SnF) versus 1400 ppm/250 ppm F (NaF) toothpaste/mouthrinse combinations, Outcome 2 Visible plaque index.

Analysis 6.3

Comparison 6 1400 ppm/250 ppm F (AmF/SnF) versus 1400 ppm/250 ppm F (NaF) toothpaste/mouthrinse combinations, Outcome 3 Gingival bleeding index.

Analysis 7.1

Comparison 7 Resin‐modified glass ionomer cement (RM‐GIC) versus light‐cured composite resin (LCC), Outcome 1 Number of participants with new DLs.

Analysis 7.2

Comparison 7 Resin‐modified glass ionomer cement (RM‐GIC) versus light‐cured composite resin (LCC), Outcome 2 Number of participants with more severe DLs of aesthetic concern.

Analysis 8.1

Comparison 8 Intraoral F‐releasing glass bead device versus F mouthrinse only, Outcome 1 Number of participants with new DLs.

Fluoride varnish compared to non‐fluoride (placebo) varnish for preventing early tooth decay (demineralised lesions) during fixed brace treatment
Patient or population: orthodontic patients (any age) Setting: orthodontic clinics in Sweden Intervention: fluoride varnish Comparison: non‐fluoride (placebo) varnish
Outcomes	*Anticipated absolute effects^ (95% CI)**			Relative effect (95% CI)	Number of participants (studies)	Certainty of the evidence (GRADE)	Comments
	With non‐fluoride (placebo) varnish	With fluoride varnish	Difference
Number of participants with new DLs (new DLs) Assessed with: pre‐treatment and day of debond clinical intraoral photographs Follow‐up: 1.7 years	28.80%	15% (4 to 55.60)	13.80% fewer (24.80 fewer to 26.80 more)	RR 0.52 (0.14 to 1.93)	405 (2 RCTs)	⊕⊕⊝⊝ LOW^a,b	The evidence that professional application of fluoride varnish (7700 or 10,000 ppm F) every 6 weeks to the teeth of patients wearing fixed orthodontic braces reduces the number of new DLs is equivocal
Number of participants with more severe DLs (severity of DLs) (score 3 or 4 versus score 1 or 2) Follow‐up: 1.7 years	26%	12% (5.70 to 24.70)	14.10% fewer (20.30 fewer to 1.30 fewer)	RR 0.46 (0.22 to 0.95)	148 (1 RCT)	⊕⊕⊝⊝ LOW^c,d	The evidence suggests that when a dentist or nurse applies 7700 ppm F (ammonium fluoride) varnish every 6 weeks to a patient wearing a fixed orthodontic brace there may a reduction in the number of orthodontic patients with more severe DLs (score 3 or 4 versus score 1 or 2)
Number of participants with adverse effects (adverse effects) Follow‐up: 1.7 years	No evidence that the intervention had adverse effects				148 (1 RCT)	⊕⊕⊝⊝ LOW^d,e	Only 1 participant with an adverse event and not clear if this was directly related to the intervention
^*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; DLs: demineralised lesions; F: fluoride; ppm: parts per million; RCT: randomised controlled trial; RR: risk ratio.
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
^aDowngraded 1 level for indirectness (evidence from only 2 studies) (Stecksén‐Blicks 2007; Sonesson 2019). ^bDowngraded 1 level for imprecision (insufficient number of participants with new DLs (guideline 300 to 400 events)). ^cDowngraded 1 level for imprecision (insufficient number of participants with the more severe DLs (guideline 300 to 400 events)). ^dDowngraded 1 level for indirectness (evidence from only 1 study) (Sonesson 2019). ^eDowngraded 1 level for imprecision (insufficient number with adverse events and not clear if directly related to the intervention).

Summary of findings 2. Dentist/nurse‐applied fluoride: 12,300 ppm F APF foam compared to 0 ppm F placebo foam for preventing early tooth decay (demineralised lesions) during fixed brace treatment

12,300 ppm F APF foam compared to 0 ppm F placebo foam for preventing early tooth decay (demineralised lesions) during fixed brace treatment
Patient or population: orthodontic patients (any age) Setting: orthodontic department at dental hospital in China Intervention: 12,300 ppm F APF foam Comparison: 0 ppm F placebo foam
Outcomes	*Anticipated absolute effects^ (95% CI)**			Relative effect (95% CI)	Certainty of the evidence (GRADE)	Number of participants (studies)	Comments
	With 0 ppm F placebo foam	With 12,300 ppm F APF foam	Difference
Number of participants with new DLs (new DLs) Assessed with: clinical assessment Follow‐up: mean 18 months	48.90%	12.70% (5.40 to 27.90)	36.20% fewer (43.60 fewer to 21 fewer)	RR 0.26 (0.11 to 0.57)	⊕⊕⊝⊝ LOW^a,b	95 (1 RCT)	The evidence suggests that when foam, containing 12,300 ppm F, is applied by a dentist or a nurse, every 2 months, to the teeth of patients wearing fixed orthodontic appliances there might be a reduction in the number of patients who have at least 1 new DL
Number of participants with more severe DLs (severity of DLs)	None of the trials reported this outcome
Number of participants with adverse effects (adverse effects)	None of the trials 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). APF: acidulated phosphate fluoride; CI: confidence interval; DLs: demineralised lesions; F: fluoride; ppm: parts per million; RCT: randomised controlled trial; RR: risk ratio.
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
^aAlthough effect size is large this was only for 1 study at unclear risk of bias and therefore downgraded 1 level (Jiang 2013). ^bDowngraded 1 level for imprecision (insufficient number of participants with new DLs (guideline 300 to 400 events)).

Summary of findings 2. Dentist/nurse‐applied fluoride: 12,300 ppm F APF foam compared to 0 ppm F placebo foam for preventing early tooth decay (demineralised lesions) during fixed brace treatment

Summary of findings 3. Patient‐applied/used fluoride: 5000 ppm F toothpaste compared to 1450 ppm F toothpaste for preventing early tooth decay (demineralised lesions) during fixed brace treatment

5000 ppm F toothpaste compared to 1450 ppm F toothpaste for preventing early tooth decay (demineralised lesions) during fixed brace treatment
Patient or population: orthodontic patients (any age) Setting: home use Intervention: 5000 ppm F toothpaste Comparison: 1450 ppm F toothpaste
Outcomes	*Anticipated absolute effects^ (95% CI)**			Relative effect (95% CI)	Number of participants (studies)	Certainty of the evidence (GRADE)	Comments
	With 1450 ppm F toothpaste	With 5000 ppm F toothpaste	Difference
Number of participants with new DLs (new DLs) Assessed with: pre‐treatment and post‐treatment clinical intraoral photographs Follow‐up: 1.8 years	26.60%	18.10% (12.20 to 26.60)	8.50% fewer (14.30 fewer to 0 fewer)	RR 0.68 (0.46 to 1.00)	380 (1 RCT)	⊕⊕⊝⊝ LOW^a,b	The evidence suggests that in patients wearing an orthodontic fixed brace use of a daily 5000 ppm F toothpaste compared with a daily 1450 ppm F toothpaste throughout treatment may lead to a reduction in the number of orthodontic patients with new DLs
Number of participants with more severe DLs (severity of DLs)	None of the trials reported this outcome
Number of participants with adverse effects (adverse effects)	None of the trials 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; DLs: demineralised lesions; F: fluoride; ppm: parts per million; RCT: randomised controlled trial; RR: risk ratio.
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
^aDowngraded 1 level due to single study at unclear risk of bias (Sonesson 2014). ^bDowngraded 1 level for imprecision (insufficient number of participants with new DLs (guideline 300 to 400 events)).

Summary of findings 3. Patient‐applied/used fluoride: 5000 ppm F toothpaste compared to 1450 ppm F toothpaste for preventing early tooth decay (demineralised lesions) during fixed brace treatment

250 ppm F mouthrinse (100 ppm amine F/150 ppm NaF) compared to 0 ppm F placebo mouthrinse for preventing early tooth decay (demineralised lesions) during fixed brace treatment
Patient or population: orthodontic patients (any age) Setting: home use Intervention: 250 ppm F mouthrinse (100 ppm amine F/150 ppm NaF) Comparison: 0 ppm F placebo mouthrinse
Outcomes	*Anticipated absolute effects^ (95% CI)**			Relative effect (95% CI)	Number of participants (studies)	Certainty of the evidence (GRADE)	Comments
	With 0 ppm F placebo mouthrinse	With 250 ppm F mouthrinse (100 ppm amine F/150 ppm NaF)	Difference
Number of participants with new DLs (new DLs) Assessed with: QLF Follow‐up: 24.5 months	46.70%	30.30% (17.30 to 54.60)	16.30% fewer (29.40 fewer to 7.90 more)	RR 0.65 (0.37 to 1.17)	81 (1 RCT)	⊕⊝⊝⊝ VERY LOW^a,b	The evidence is very uncertain about the effect of a daily 250 ppm F mouthrinse (100 ppm amine F/150 ppm NaF) compared with a daily 0 ppm F placebo mouthrinse on the number of patients wearing a fixed orthodontic brace with new DLs
Number of participants with more severe DLs (severity of DLs)	None of the trials reported this outcome
Number of participants with adverse effects (adverse effects)	None of the trials 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; DLs: demineralised lesions; F: fluoride; NaF: sodium fluoride; ppm: parts per million; QLF: quantitative light‐induced fluorescence; RCT: randomised controlled trial; RR: risk ratio.
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
^aDowngraded 2 levels: single study with a relatively small number of participants (81), at high risk of bias due to high attrition (33%) (van der Kaaij 2015). ^bDowngraded 1 level for imprecision (insufficient number of participants with new DLs (guidance 300 to 400 events)).

Resin‐modified glass ionomer cement compared to light‐cured composite resin for bonding orthodontic brackets for preventing early tooth decay (demineralised lesions) during fixed brace treatment
Patient or population: orthodontic patients (any age) Setting: 2 dental teaching hospitals and 4 specialist orthodontic practices in UK and Republic of Ireland Intervention: resin‐modified glass ionomer cement Comparison: light‐cured composite resin
Outcomes	*Anticipated absolute effects^ (95% CI)**			Relative effect (95% CI)	Number of participants (studies)	Certainty of the evidence (GRADE)	Comments
	With light‐cured composite resin	With resin‐modified glass ionomer cement	Difference
Number of participants with new DLs (new DLs) Assessed with: before and day of debond clinical intraoral photographs Follow‐up: 17.6 months	22.40%	26.20% (15.40 to 44.50)	3.80% more (6.90 fewer to 22.10 more)	RR 1.17 (0.69 to 1.99)	173 (1 RCT)	⊕⊕⊝⊝ LOW^a,b	The evidence suggests that resin‐modified glass ionomer cement for bonding orthodontic brackets compared with light‐cured composite resin results in little to no difference in the number of orthodontic patients with new DLs
Number of participants with more severe DLs of aesthetic concern (severity of DLs) Assessed with: pre‐treatment and day of debond clinical intraoral photographs Follow‐up: 17.6 months	8.00%	9.40% (3.60 to 24.80)	1.40% more (4.40 fewer to 16.90 more)	RR 1.18 (0.45 to 3.12)	173 (1 RCT)	⊕⊕⊝⊝ LOW^a,b	The evidence suggests that using resin‐modified glass ionomer cement for bonding orthodontic brackets compared with light‐cured composite resin results in little to no difference in the number of orthodontic patients with more severe DLs of aesthetic concern
Number of participants with adverse effects (adverse effects) Follow‐up: 17.6 months	No evidence that either intervention had adverse effects				173 (1 RCT)	⊕⊕⊝⊝ LOW^a,b	‐
^*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; DLs: demineralised lesions; RCT: randomised controlled trial; RR: risk ratio.
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
^aDowngraded 1 level: single study at unclear risk of bias (Benson 2019). ^bDowngraded 1 level for imprecision (insufficient number of participants with new DLs or more severe DLs (guidance 300 to 400 events)).

Table 1. Summary of findings for 12,500 ppm F (NaF/olaflur/dectaflur) fluoride gel compared to 0 ppm F placebo gel for preventing early tooth decay (demineralised lesions) during fixed brace treatment

Patient or population: orthodontic patients (any age) Setting: dental clinics in Germany and Israel and home use Intervention: 12,500 ppm F (NaF/olaflur/dectaflur) fluoride gel Comparison: 0 ppm F placebo gel
Outcomes	*Anticipated absolute effects^ (95% CI)**			Relative effect (95% CI)	Number of participants (studies)	Certainty of the evidence (GRADE)	Comments
	With 0 ppm F placebo gel	With 12,500 ppm F (NaF/olaflur/dectaflur) fluoride gel	Difference
Number of participants with new DLs (new DLs) Assessed with: clinical examination Follow‐up: mean 16.6 months	42.90%	42.50% (32.60 to 54.50)	0.40% fewer (10.30 fewer to 11.60 more)	RR 0.99 (0.76 to 1.27)	312 (1 RCT)	⊕⊝⊝⊝ VERY LOW^a,b,c	The evidence is very uncertain about the effect of an amine fluoride gel compared with a placebo gel on the number of patients wearing fixed orthodontic braces with new DLs
Number of participants with more severe DLs (severity of DLs)	None of the trials reported this outcome
Number of participants with adverse effects (adverse effects) Follow‐up: mean 16.6 months	0.60%	0.00% (0 to 0)	0.60% fewer (0.60 fewer to 0.60 fewer)	Not estimable	312 (1 RCT)	⊕⊝⊝⊝ VERY LOW^a,c,d	Authors write "Most AEs (96.4%) were unrelated to the study treatment. Only a single AE (hay fever) was considered related to the study treatment (placebo)." They go on to state "The use... for a maximum of 32.5 months was not associated with any unexpected ARs or AEs. Thus, there are no safety issues to be considered for long‐term use of elmex® gel"
^*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). AE: adverse effects; CI: confidence interval; DLs: demineralised lesions; F: fluoride; NaF: sodium fluoride; ppm: parts per million; RCT: randomised controlled trial; RR: risk ratio.
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
^aDowngraded 1 level due to single study at unclear risk of bias (Jost‐Brinkman 2017). ^bDowngraded 1 level for imprecision (insufficient number of participants with new DLs (guidance 300 to 400 events)). ^cDowngraded 1 level for publication bias (data collection for the trial was completed in 2011. The results were obtained from the study report that was published in 2016, but not yet submitted to a peer‐reviewed journal). ^dDowngraded 1 level for imprecision (insufficient number with adverse events).

Table 2. Summary of findings for 1400 ppm/250 ppm F (amine F/SnF) toothpaste/mouthrinse compared to 1400 ppm/250 ppm F (NaF) toothpaste/mouthrinse for preventing early tooth decay (demineralised lesions) during fixed brace treatment

Patient or population: orthodontic patients (any age) Setting: home use Intervention: 1400 ppm/250 ppm F (amineF/SnF) toothpaste/mouthrinse Comparison: 1400 ppm/250 ppm F (NaF) toothpaste/mouthrinse
Outcomes	*Anticipated absolute effects^ (95% CI)**		Relative effect (95% CI)	Number of participants (studies)	Certainty of the evidence (GRADE)	Comments
	Risk with 1400 ppm/250 ppm F (NaF) toothpaste/mouthrinse	Risk with 1400 ppm/250 ppm F (amineF/SnF) toothpaste/mouthrinse
Number of participants with new DLs (new DLs) Follow‐up: 1.5 years	Outcome not reported. White spot lesion index, visible plaque index and gingival bleeding index reported instead: MD 0.05 lower (0.10 lower to 0) white spot index, MD 0.09 lower (0.18 lower to 0) visible plaque index, MD 0.07 lower (0.15 lower to 0.01 higher) gingival bleeding index for 1400 ppm/250 ppm F (amine F/SnF) toothpaste/mouthrinse		‐	97 (1 RCT)	⊕⊝⊝⊝ VERY LOW^a,b,c	The evidence is very uncertain about the effect of daily 1400 ppm/250 ppm F (amine F/SnF) versus 1400 ppm/250 ppm F (NaF) toothpaste/mouthrinse combinations on the white spot index, visible plaque index and on the gingival bleeding index in patients wearing fixed orthodontic braces
Number of participants with more severe DLs (severity of DLs)	None of the trials reported this outcome
Number of participants with adverse effects (adverse effects)	None of the trials 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; DLs: demineralised lesions; F: fluoride; MD: mean difference; NaF: sodium fluoride; ppm: parts per million; RCT: randomised controlled trial; SnF: stannous fluoride.
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
^aDowngraded 1 level due to single study assessed at unclear risk of bias (Ogaard 2006). ^bDowngraded 1 level for indirectness (outcomes assessed were not the most useful for answering this question). ^cDowngraded 1 level for imprecision (insufficient number of participants with new DLs (guideline 300 to 400)).

Table 3. Summary of findings for intraoral fluoride‐releasing glass bead device compared to 250 ppm F mouthrinse for preventing early tooth decay (demineralised lesions) during fixed brace treatment

Patient or population: orthodontic patients (any age) Setting: teaching hospital in UK and home use Intervention: intraoral fluoride‐releasing glass bead device Comparison: 250 ppm F mouthrinse
Outcomes	*Anticipated absolute effects^ (95% CI)**			Relative effect (95% CI)	Number of participants (studies)	Certainty of the evidence (GRADE)	Comments
	With 250 ppm F mouthrinse	With intraoral fluoride‐releasing glass bead device	Difference
Number of participants with new DLs (new DLs) Assessed with: clinical intraoral photographs Follow‐up: 19 months	36.80%	55.60% (26.90 to 100)	18.80% more (9.90 fewer to 77.40 more)	RR 1.51 (0.73 to 3.10)	37 (1 RCT)	⊕⊝⊝⊝ VERY LOW^a,b	The evidence is very uncertain about the effect of intraoral fluoride‐releasing glass bead devices compared with daily 250 ppm F mouthrinses on the number of patients wearing a fixed orthodontic brace with new DLs
Number of participants with more severe DLs (severity of DLs)	None of the trials reported this outcome
Number of participants with adverse effects (adverse effects) Follow‐up: 19 months	‐	‐	‐	‐	37 (1 RCT)	⊕⊝⊝⊝ VERY LOW^a,b,c	Current design of fluoride glass bead inadequate. The authors write "an improved method of attachment is needed before it can be established whether FGBs (fluoride glass beads) are more effective than FR (fluoride mouthrinse)."^c
^*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; DLs: demineralised lesions; F: fluoride; ppm: parts per million; RCT: randomised controlled trial; RR: risk ratio.
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
^aDowngraded 2 levels due to single study at high risk of bias (high attrition (47%)) (Luther 2005). ^bDowngraded 1 level for imprecision (insufficient number of participants with new DLs (guidance 300 to 400 events)). ^cHigh attrition partly due to breaks of the glass beads. The authors write "Some patients in the FGB (fluoride glass bead) group experienced a large number of bead breakages, resulting in a number of patients requesting to leave the trial." Numbers not supplied.

Comparison 1. Professionally‐applied F varnish versus non‐F (placebo) varnish

Outcome or subgroup title	No. of studies	No. of participants	Statistical method	Effect size
1 Number of participants with new DLs Show forest plot	2	405	Risk Ratio (M‐H, Random, 95% CI)	0.52 [0.14, 1.93]

2 Number of participants with more severe DLs (scores 3 or 4 versus scores 1 or 2) Show forest plot	1	148	Risk Ratio (M‐H, Random, 95% CI)	0.46 [0.22, 0.95]

Comparison 1. Professionally‐applied F varnish versus non‐F (placebo) varnish

Comparison 2. 12,500 ppm F (NaF/olaflur/dectaflur) gel versus 0 ppm F placebo gel

Outcome or subgroup title	No. of studies	No. of participants	Statistical method	Effect size
1 Number of participants with new DLs Show forest plot	1	312	Risk Ratio (M‐H, Random, 95% CI)	0.99 [0.76, 1.27]

Comparison 2. 12,500 ppm F (NaF/olaflur/dectaflur) gel versus 0 ppm F placebo gel

Comparison 3. 12,300 ppm F APF foam versus 0 ppm F placebo foam

Outcome or subgroup title	No. of studies	No. of participants	Statistical method	Effect size
1 Number of participants with new DLs Show forest plot	1	95	Risk Ratio (M‐H, Random, 95% CI)	0.26 [0.11, 0.57]

Comparison 3. 12,300 ppm F APF foam versus 0 ppm F placebo foam

Comparison 4. 5000 ppm F toothpaste versus 1450 ppm F toothpaste

Outcome or subgroup title	No. of studies	No. of participants	Statistical method	Effect size
1 Number of participants with new DLs Show forest plot	1	380	Risk Ratio (M‐H, Random, 95% CI)	0.68 [0.46, 1.00]

Comparison 4. 5000 ppm F toothpaste versus 1450 ppm F toothpaste

Comparison 5. 250 ppm F mouthrinse (100 ppm AmF/150 ppm NaF) versus 0 ppm F placebo mouthrinse

Outcome or subgroup title	No. of studies	No. of participants	Statistical method	Effect size
1 Number of participants with new DLs Show forest plot	1	81	Risk Ratio (M‐H, Random, 95% CI)	0.65 [0.37, 1.17]

Comparison 5. 250 ppm F mouthrinse (100 ppm AmF/150 ppm NaF) versus 0 ppm F placebo mouthrinse

Comparison 6. 1400 ppm/250 ppm F (AmF/SnF) versus 1400 ppm/250 ppm F (NaF) toothpaste/mouthrinse combinations

Outcome or subgroup title	No. of studies	No. of participants	Statistical method	Effect size
1 White spot index Show forest plot	1	97	Mean Difference (IV, Random, 95% CI)	‐0.05 [‐0.10, 0.00]

2 Visible plaque index Show forest plot	1	97	Mean Difference (IV, Random, 95% CI)	‐0.09 [‐0.18, ‐0.00]

3 Gingival bleeding index Show forest plot	1	97	Mean Difference (IV, Random, 95% CI)	‐0.07 [‐0.15, 0.01]

Comparison 6. 1400 ppm/250 ppm F (AmF/SnF) versus 1400 ppm/250 ppm F (NaF) toothpaste/mouthrinse combinations

Comparison 7. Resin‐modified glass ionomer cement (RM‐GIC) versus light‐cured composite resin (LCC)

Outcome or subgroup title	No. of studies	No. of participants	Statistical method	Effect size
1 Number of participants with new DLs Show forest plot	1	173	Risk Ratio (M‐H, Random, 95% CI)	1.17 [0.69, 1.99]

2 Number of participants with more severe DLs of aesthetic concern Show forest plot	1	173	Risk Ratio (M‐H, Random, 95% CI)	1.18 [0.45, 3.12]

Comparison 7. Resin‐modified glass ionomer cement (RM‐GIC) versus light‐cured composite resin (LCC)

Comparison 8. Intraoral F‐releasing glass bead device versus F mouthrinse only

Outcome or subgroup title	No. of studies	No. of participants	Statistical method	Effect size
1 Number of participants with new DLs Show forest plot	1	37	Risk Ratio (M‐H, Random, 95% CI)	1.51 [0.73, 3.10]

Comparison 8. Intraoral F‐releasing glass bead device versus F mouthrinse only