LAUSR

Background Despite our awareness of the significant effect of smoking on multiple sclerosis (MS), there was a serious lack of information about the effect of different types of smoking habit on MS as well as the dose-response relationship between smoking and MS, but this gap was addressed by this meta-analysis. Methods Major electronic databases, including PubMed, Web of Science and Scopus were searched until July 2015. Epidemiological studies addressing the association between smoking and MS were enrolled. The heterogeneity across studies was investigated by Q-test and I2 statistic. The probability of publication bias was explored using Begg's and Egger's tests. The results were expressed as odds ratio (OR) with 95% confidence intervals (CI) using a random effects model. Results We identified a total of 10 687 references and included 34 studies involving 623 852 participants. Compared with nonsmokers, the OR estimate of MS was 1.46 (1.33, 1.59) among ever smokers, 1.57 (1.34, 1.80) among current smokers, 1.36 (1.27, 1.46) among ex-smokers and 1.12 (0.87, 1.36) among passive smokers. In addition, analyzing the effect of cigarette pack-year on MS indicated that the OR estimate of MS was 1.34 (1.06, 1.61) for 1-5 cigarette pack-years, 1.56 (1.15, 1.97) for 6-10 cigarette pack-years, 1.74 (1.16, 2.33) for 11-15 cigarette pack-years and 1.46 (0.49, 2.43) for more than 15 cigarette pack-years. Conclusions Smoking habits are significantly associated with MS, although the association is not very strong. However, there is a dose-response relationship between the smoking habits and MS.