Association between vitiligo and smoking: A nationwide population-based study in Korea

No study has examined the associations between vitiligo and smoking. The purpose of this study was to investigate the incidence of vitiligo according to smoking status. We used clinical data from individuals aged over 20 years who received a health examination in the National Insurance Program between 2009 and 2012 (n = 23,503,807). We excluded individuals with pre-existing vitiligo who had ever been diagnosed with vitiligo before the index year (n = 35,710) or who were diagnosed with vitiligo within a year of the index year (n = 46,476). Newly diagnosed vitiligo was identified using claims data from baseline to date of diagnosis or December 31, 2016 (n = 22,811). The development of vitiligo was compared according to self-reported smoking status by a health examination survey. The hazard ratio of vitiligo in current smokers was 0.69 (95% confidence interval; 0.65–0.72) with a reference of never-smokers after adjustment for age, sex, regular exercise, drinking status, body mass index, diabetes mellitus, hypertension, dyslipidemia, history of stroke, and history of ischemic heart diseases. The decreased risk of vitiligo in current smokers persisted after subgroup analysis of sex and age groups. The results suggested there are suppressive effects of smoking on the development of vitiligo. Further studies are needed to evaluate the mechanism of smoking on the development of vitiligo.


Subgroups according to smoking status.
Smoking status was obtained from a self-reported questionnaire during the health examination. Study individuals (n = 22,991,641) were divided into three groups according to smoking status; never-smoker (n = 14,345,458), ex-smoker (n = 3,042,684), and current smoker (n = 5,603,499). (Fig. 1) The amount of smoking was sub-grouped into <10 cigarettes per day, 10-19 cigarettes per day, and ≥20 cigarettes per day, and smoking duration was divided into <10 years, 10-29 years, and ≥30 years.
comorbidities. To adjust for comorbidities, the presence of diabetes mellitus, hypertension, and dyslipidemia were defined using ICD-10 codes: diabetes mellitus (E11- 14), hypertension (I10-13 and I15), and dyslipidemia (E78) with medication in KNHIS database. History of stoke or ischemic heart disease was obtained by self-reported questionnaire during the National Health examination. Statistical analysis. We considered comorbidities including diabetes mellitus, hypertension, dyslipidemia, ischemic heart disease, and stroke as possible confounders to adjust in our analyses. Cox's proportional hazard regression models with an age timescale were used to identify the associations between smoking status and newly diagnosed vitiligo. Hazard ratio (HR) and 95% confidence intervals (CI) for smoking status were compared to the reference (never-smoker). We performed the subgroup analyses separately for men and women and age groups (20-39 years, 40-65 years, ≥65 years). Proportional hazard assumptions were checked using log-log cumulative survival graphs and the time-dependent variable Cox model after adjustment for baseline covariates including age, sex, regular exercise, drinking status, BMI, diabetes mellitus, hypertension, dyslipidemia, history of stroke, and history of myocardial infarction, according to smoking status. All statistical analyses were performed using SAS software (ver. 9.4; SAS Institute, Cary, NC, USA).

Results
characteristics of the study population. Among the 22,991,641 individuals, 14,345,458 were never-smokers, 3,042,684 were ex-smokers, and 5,603,499 were current smokers. The current smokers (43.34 ± 12.87 years) were younger than never-smokers (48.86 ± 14.8 years) and ex-smokers (50.11 ± 13.4 years). The percentage of men was significantly higher in ex-smokers and current smokers than in never-smokers. Drinking habits (2-3 times a month) were higher in ex-smokers and current smokers. Comorbidities including diabetes mellitus, hypertension, and dyslipidemia were highest in ex-smokers. History of previous stroke and ischemic heart disease was also higher in ex-smokers. (Table 1). newly diagnosed vitiligo according to smoking status. There were 16,515 newly diagnosed cases of vitiligo in never-smokers, 3,003 cases in ex-smokers, and 3,293 cases in current smokers. The incidence of newly diagnosed vitiligo was 2.63 per 10,000 person-years in never-smokers, 2.23 per 10,000 person-years in ex-smokers, and 1.35 per 10,000 person-years in current smokers. The cumulative incidence of newly diagnosed vitiligo after adjustment for covariates is shown in Fig. 2 according to smoking status. Current smokers had a significantly lower risk of vitiligo (HR 0.51, 95% CI, 0.50-0.53) compared with never-smokers. Decreased risk of newly diagnosed vitiligo in current smokers persisted after setting an age timescale in Model 1 (HR 0.56, 95% CI, 0.54-0.59) and adjustment for covariates in Model 2 (HR 0.68, 95% CI, 0.64-0.71) and Model 3 (HR 0.69, 95% CI, 0.65-0.72). Sensitivity analysis also showed decreased vitiligo risk in current smokers regardless of age or sex ( Table 2). www.nature.com/scientificreports www.nature.com/scientificreports/ Subgroup analysis of amount and duration of smoking. Current smokers showed a decrease in HR for newly diagnosed vitiligo in proportion to the amount of smoking compared with never-smokers; HR 0.70, 95% CI, 0.64-0.76) in current smokers who smoked less than 10 cigarettes a day, HR 0.51, 95% CI, 0.49-0.52 in current smoker who smoked 10-29 cigarettes a day, and HR 0.47, 95% CI, 0.44-0.49 in current smoker who smoked more than 30 cigarettes a day. (Table 3) Current smokers showed decreased HR for vitiligo regardless of smoking duration compared with never-smokers.

Discussion
We report a negative association between tobacco smoking and newly diagnosed vitiligo using a nation-wide cohort database. No study has reported on the incidence of vitiligo associated with smoking status.
In contrast, protective effects of smoking have also been reported for several skin diseases. Case-control studies report that pemphigus vulgaris and foliaceous occur less frequently in current and ex-smokers [28][29][30][31][32][33] . Development of aphthous ulcers is associated with cessation of smoking [44][45][46][47][48][49][50] . The beneficial effects of smoking on oral ulcers in Behçet's disease has been shown in several reports [34][35][36] . Behçet's disease oral lesions appear after cessation of smoking; therefore, the beneficial effects of nicotine have been suggested [51][52][53][54] . However, some studies www.nature.com/scientificreports www.nature.com/scientificreports/ report that smoking is not a significant risk factor for Behçet's disease 55 , and smoking is a risk factor for vasculitis and neurological manifestations of the disease 56,57 . (Table 4) In addition to skin diseases, Parkinson's disease [58][59][60] is inversely associated with smoking without exact known mechanisms. Ulcerative colitis is well-known to be inversely associated with smoking 61 . Until now, the association between vitiligo and smoking has not been evaluated or reported.
The pro-oxidant and anti-oxidant effects of smoking are suggested in several studies. Tobacco contains a variety of reactive oxygen species (ROS). Therefore, smoking has deleterious effects on human health. Otherwise, tobacco also contains compounds that inhibit the activity of monoamine oxidase (MAO) and can reduce the level of ROS produced by MAO. There has been a report that smokers have lower MAO activity than never-smokers 62,63 . Another study reported that nicotine chelates ferrous ions that produce oxygen radicals and functions as an antioxidant 64 .
In this study, the risk of vitiligo decreased in current smokers in relation to smoking dose. This result was counter to our expectation that the oxidative stress incurred from smoking would cause vitiligo development. Vitiligo is an acquired depigmentation skin disease caused by destruction of melanocytes in affected skin. Oxidative stress and autoimmunity are important in pathogenic events in melanocyte loss 65 . Increased levels of ROS, mainly H 2 O 2 , are reported in the epidermis of vitiligo skin 66 . Elevated NADPH oxidase activity and increased production of ROS and reactive nitrogen species are reported in vitiligo patients 67 . Increased activity of MAO is reported in vitiligo epidermis that results in accumulation of H 2 O 2 68 . Low levels of antioxidants, such as catalase, glutathione peroxidase, and superoxide dismutase are reported in the epidermis and serum of vitiligo patients 69,70 . The inverse relationship between vitiligo and smoking might be explained by inhibitory effects on MAO due to tobacco smoking 62 . However, the exact mechanism needs to be evaluated by further studies.
Considering the harm caused by smoking and the lack of a distinct benefit for vitiligo development, we do not think smoking should be considered as a prevention or treatment option for vitiligo. We suggest investigating the mechanism of smoking in vitiligo development in vitro and in vivo in further studies. We present a new topic in vitiligo research in a field with little progress and limited development of treatment options.
There are several limitations in our study. First, the genetic susceptibility of individuals or family history of vitiligo was not evaluated. Second, causal links between smoking and vitiligo could not be identified in an epidemiologic study. Third, the status of smoking was obtained from self-reported questionnaires during the health examination and a response bias caused by respondents failing to provide correct information cannot be ruled www.nature.com/scientificreports www.nature.com/scientificreports/ out. And the absence of data on time since quitting smoking in ex-smokers is considered as a limitation in this study. The role of smoking in vitiligo development remains for additional clinical analysis or in vitro studies.
Several strengths in this study are that first, we examined the relationship between vitiligo and smoking in the South Korean population. This study used a nation-wide population-based cohort. In Korea, a general health examination, either biannually or annually according to occupation, is mandatory for local household owners, Figure 2. An adjusted cumulative incidence of vitiligo for baseline covariates including age, sex, regular exercise, drinking status, BMI, diabetes mellitus, hypertension, dyslipidemia, history of stroke, and history of myocardial infarction, according to smoking status.  www.nature.com/scientificreports www.nature.com/scientificreports/ office employees, and family members over the age of 40 years. The total number of individuals who underwent National Health Examinations in Korea from 2009 until 2012 was 23,503,802 that was more than half of the total population of Korea. Considering that we excluded subjects less than 20 years old, the results of this study could generalize the majority of Koreans who receive a medical examination. We identified the possible suppressive effects of smoking on the development of vitiligo. Current smokers showed a decreased risk of vitiligo development compared to non-smokers regardless of age and sex, and the amount of smoking was negatively correlated with vitiligo development. The suppressive mechanism of smoking on the cutaneous autoimmune disease, vitiligo, is unknown. Further in vitro studies are needed to investigate the exact mechanism between smoking and vitiligo. However, this study helps to broaden the understanding of vitiligo pathogenesis.  Psoriasis [16][17][18][19] Aphthous ulcer [44][45][46][47][48][49][50] Hidradenitis suppurativa [21][22][23][24][25][26][27] Oral ulcer in Behçet's disease 34,35,[51][52][53][54] Systemic lupus erythematosus [37][38][39] Wound healing 40 Skin aging 41-43 Table 4. Effects of smoking on skin disease.