Effect of Toothbrushing Frequency on Incidence and Increment of Dental Caries: A Systematic Review and Meta-Analysis

Toothbrushing is considered fundamental self-care behavior for maintenance of oral health, and brushing twice a day has become a social norm, but the evidence base for this frequency is weak. This systematic review and meta-analysis aims to assess the effect of toothbrushing frequency on the incidence and increment of carious lesions. Medline, Embase, Cinahl, and Cochrane databases were searched. Screening and quality assessment were performed by 2 independent reviewers. Three different meta-analyses were conducted: 2 based on the caries outcome reported in the studies (incidence and increment) with subgroup analyses of categories of toothbrushing frequency; another included all studies irrespective of the caries outcome reported with the type of dentition as subgroups. Meta-regression was conducted to assess the influence of sample size, follow-up period, diagnosis level for carious lesions, and methodological quality of the articles on the effect estimate. Searches retrieved 5,494 titles: after removing duplicates, 4,305 remained. Of these, 74 were reviewed in full, but only 33 were eligible for inclusion. Self-reported infrequent brushers demonstrated higher incidence (odds ratio [OR], 1.50; 95% confidence interval [CI], 1.34 to 1.69) and increment (standardized mean difference [SMD], 0.28; 95% CI: 0.13 to 0.44) of carious lesions than frequent brushers. The odds of having carious lesions differed little when subgroup analysis was conducted to compare the incidence between 2 times/d vs <2 times/d (OR: 1.45; 95% CI: 1.21 to 1.74) and 1 time/d vs <1 time/d brushers (OR: 1.56; 95% CI: 1.37 to 1.78). When meta-analysis was conducted with the type of dentition as subgroups, the effect of infrequent brushing on incidence and increment of carious lesions was higher in deciduous (OR: 1.75; 95% CI: 1.49 to 2.06) than permanent dentition (OR: 1.39; 95% CI: 1.29 to 1.49). Findings from meta-regression indicated that none of the included variables influenced the effect estimate.