Objective:
A previous meta-analysis demonstrated a moderate relationship between dental caries in primary and permanent dentition and secondhand smoke exposure (prenatal or postnatal) in children and adolescents. The present study aimed to compare the effect size of secondhand smoke on dental caries and its heterogeneity by potential sources of the difference with subgroup meta-analysis.

Methods:
Articles identified in our previous review and collected through updated PubMed search were used. A generic inverse variance method (random-effects model) was conducted to combine the data from eligible studies and calculate the summary adjusted odds ratio (OR) as pooled effect size estimates. Subgroup meta-analysis compared differences of the pooled effect size by study design (cross-sectional or longitudinal), country (Japan, the U.S. and Europe), adjustments for socio-economic status (with or without), and status of secondhand smoke exposure (current and former). The heterogeneity across studies was assessed using the I2 statistics and Cochrane’s Q.

Results:
Nineteen articles were included in the meta-analysis. The pooled effect size estimate for studies in Japan (OR = 1.61, p < 0.001) was similar to that in the U.S. and slightly higher than that in Europe. Studies with adjustments for socio-economic status showed a lower effect size (OR = 1.51, p < 0.001) compared to those without (OR = 1.66, p < 0.001). The estimated OR was 1.46 (p < 0.001) and 1.13 (p = 0.11) for current and former exposure, respectively. Heterogeneity analysis found a moderate heterogeneity in most subgroups, a high heterogeneity (Q = 13.6, p = 0.004; I2 = 78%) in longitudinal studies and a mild heterogeneity (Q = 9.1, p = 0.17; I2 = 34%) in studies of Europe.

Conclusions:
The effect of secondhand smoke on dental caries and its heterogeneity were influenced by geographical and methodological factors. Unexpected factors for explaining heterogeneity still remain.