Examining the causal association between 25-hydroxyvitamin D and caries in children and adults: a two-sample Mendelian randomization approach

• Published in: Wellcome open research 2021, Vol.5, p.281
• Main Authors: Dodhia, Serena A, West, Nicola X, Thomas, Steven J, Timpson, Nicholas J, Johansson, Ingegerd, Lif Holgerson, Pernilla, Dudding, Tom, Haworth, Simon
• Format: Article
• Language: English
• Subjects: 25-hydroxyvitamin D; Dental caries; Mendelian randomization; Vitamin D
• ISSN: 2398-502X; 2398-502X
• DOI: 10.12688/wellcomeopenres.16369.2

Prior observational studies have reported that higher levels of vitamin D are associated with decreased caries risk in children. However, these studies are prone to bias and confounding so do not provide causal inference. Genetic variants associated with a risk factor of interest can be used as proxies, in a Mendelian randomization (MR) analysis, to test for causal association with an outcome. The objective was to estimate the causal association between serum 25-hydroxyvitamin D (25(OH)D) (the commonly measured vitamin D metabolite in blood) and dental caries using a two-sample MR approach which estimates the causal effect of an exposure on an outcome. A total of 79 genetic variants reliably associated with 25(OH)D were identified from genome-wide association studies and used as a proxy measure of 25(OH)D. The association of this proxy measure with three outcome measures was tested; specifically: caries in primary teeth (n=17,035, aged 3-12 years), caries in permanent teeth in childhood and adolescence (n=13,386, aged 6-18 years), and caries severity in adulthood proxied by decayed, missing and filled tooth surfaces (DMFS) counts (n=26,792, aged 18-93 years). The estimated causal effect of a one standard deviation increase in natural log-transformed 25(OH)D could be summarized as an odds ratio of 1.06 (95%CI: 0.81, 1.31; P=0.66) for caries in primary teeth and 1.00 (95%CI: 0.76, 1.23; P=0.97) for caries in permanent teeth in childhood and adolescence. In adults, the estimated casual effect of a one standard deviation increase in natural log-transformed 25(OH)D was 0.31 fewer affected tooth surfaces (95%CI: from 1.81 fewer DMFS to 1.19 more DMFS; P=0.68) The MR-derived effect estimates for these three measures are small in magnitude with wide confidence intervals and do not provide evidence for a causal relationship between 25(OH)D and dental caries.