No evidence for an association between alcohol consumption and Multiple Sclerosis risk: a UK Biobank study

Abstract

Multiple Sclerosis (MS) has been linked to a variety of environmental risk factors, including smoking, Epstein-Barr Virus infection, and childhood obesity. There is some evidence to support a relationship between alcohol consumption and MS risk, but this finding has been inconsistent across cohorts. A protective link between alcohol consumption and MS risk is seen in Swedish and Danish cohorts, however evidence from other cohorts and mendelian randomisation studies have failed to support this relationship. We assessed the relationship between alcohol consumption (never vs. ever drinking) and MS in 409,228 individuals (2100 with MS) from UK Biobank (UKB). We used multivariable logistic regression models adjusted for age and sex. To determine whether there was evidence of statistical interaction between alcohol consumption and HLA-DRB1*15:01 genotype, we calculated interaction on the additive and multiplicative scales. We analysed data from 2100 individuals with MS (72.3% female, median age at recruitment 56) and 407,128 controls (53.9% female, median age at recruitment 58). We found no evidence for an association between alcohol consumption and MS risk (OR = 1.12, 95% CI 0.61–2.08, p = 0.314). As expected, the HLA-DRB1*15:01 allele was strongly associated with MS risk (OR = 2.72, 95% CI 2.72–2.72, p < 2 × 10−16). We found no evidence of statistical interaction between non-drinking and MS risk on either the multiplicative scale (p = 0.8) or on the additive scale (Attributable Proportion = 0.03, 95% CI − 0.43–0.29, P = 0.45). Empirical power calculations indicated reasonable statistical power (85%) to detect a protective effect of alcohol consumption of Relative Risk ≤ 0.7. We were thus unable to replicate findings from other cohorts within UK Biobank. The inconsistent association seen between studies may reflect limited statistical power to detect a weak effect, differences in population characteristics, or the lack of a true causal association.