The Prevalence of Vitiligo: A Meta-Analysis

Objective To conduct a meta-analysis assessing the prevalence of vitiligo. Methods Literatures that reported prevalence rates of vitiligo were identified using EMBASE, PubMed, the Cochrane Library, China National Knowledge Infrastructure (CNKI), Wanfang database and Weipu database for the period from inception to May 2016. We performed stratified analyses on possible sources of bias, including areas difference, years of publication, gender and age. Publication bias was assessed with Egger’s test method. Results A total of 103 studies were eligible for inclusion. The pooled prevalence of vitiligo from 82 population- or community-based studies was 0.2% (95%CI: 0.1%–0.2%) and from 22 hospital-based studies was 1.8% (95%CI: 1.4%–2.1%). A relatively high prevalence of vitiligo was found in Africa area and in female patients. For population- or community-based studies, the prevalence has maintained at a low level in recent 20 years and it has increased with age gradually. For hospital-based studies, the prevalence has showed a decreased trend from 60s till now or from young to old. No significant publication bias existed in hospital-based studies (t = 0.47, P = 0.643), while a significant publication bias existed in population- or community-based studies (t = 2.31, P = 0.026). Conclusion A relatively high prevalence of vitiligo was found in Africa area and in female patients. The prevalence has maintained at a low level in recent years. It showed an inverse trend with age increment in population- or community-based studies and hospital-based studies.


Introduction
Vitiligo refers to an acquired, idiopathic, and common de-pigmentation disorder of the skin [1]. The clinically characteristic symptoms of the vitiligo are pale or milk-white macules or patches due to the selective destruction of melanocytes. They occur on the skin in different parts of the body and sometimes also on the mucous membranes. The exact pathogenesis of vitiligo is still to be elucidated. Multiple mechanisms, including metabolic abnormalities, oxidative stress, generation of inflammatory mediators, cell detachment and autoimmune responses, might contribute to the pathogenesis. In particular, the autoimmune mechanism is now clearly established. Vitiligo may appear at any age and affect both sexes. It tends to occur or recur in spring and/or summer [2,3].
Some previous reports on vitiligo epidemiology were based on population surveys, while others were performed in patients of dermatology clinics. However, the prevalence of vitiligo varies in different geographic regions and different sample size, and the data have limitations and localizations. Besides, the disorder afflicts various ethnic populations with varying prevalence estimates ranging from 0.1% to 2.0% based on the general populations in previous studies [4,5]. But recently, some papers suggested that previous epidemiological data were exaggerated. To date, no meta-analysis on the prevalence of vitiligo has been conducted. Accordingly, it seems that an international and pooled estimate based on the various populations is necessary.
The main objective of this meta-analysis is to summarize all available data to give a description of a worldwide picture on the prevalence of vitiligo. The information was collected from both population-or community studies and hospital-based studies. Various epidemiological characteristics of vitiligo were studied in order to understand this disease more clearly.

Search Strategy
We conducted a systematic search of scientific databases, including EMBASE, PubMed, the Cochrane Library, China National Knowledge Infrastructure (CNKI), Wanfang database and Weipu database to find relevant papers published from inception to May 2016. The search medical subject heading (MeSH) terms and keywords were "vitiligo" OR "leucoderma" AND "prevalence" OR "epidemiology". In addition, a manual search was supplemented by verifying a secondary review of the reference lists of key publications to confirm additional relevant citations.

Inclusion and Exclusion Criteria
The criteria of included studies were as follows: (1) had sufficient information to estimate the pooled prevalence of vitiligo; (2) population-based, community-based or hospital-based; (3) published in either English or Chinese language.

Data Extraction
The whole potentially relevant information from the included studies was independently reviewed by two investigators (Yuhui Zhang, Meihui Shi) using a standardized form which was designed in advance. When there was a disagreement about whether selecting articles should be resolved for analysis or not, a third investigator (Yunfei Cai) made the final decision. The following information was extracted from each suitable study: first author's name, years of publication, country, survey age, gender of the participants, survey year, total sample size, numbers of vitiligo and prevalence rate.

Data Analysis
All statistical analyses were made using Stata software (version 12.0; Stata Corporation, College Station, Texas, USA) and the meta package was used to produce the pooled estimates, forest plots and publication bias assessment. Initially, the pooled prevalence estimates of vitiligo and 95% confidence intervals (CIs) were calculated assuming a fixed-effect model when significant heterogeneity was absent (P>0.1, I 2 <50%). If significant heterogeneity was present (P<0.1, I 2 >50%), a random-effect model was selected. To determine possible causes of heterogeneity, subgroup analyses were conducted by areas, years of publication, gender, and age. The areas covered Asia (India, China, Saudi Arabia, Sri Lanka, Turkey, Nepal, Iran, Korea, Kuwait, Thailand, Japan, Jordan), Africa (Tanzania, Egypt, Mali, Mozambique, Nigeria, Congo), America (USA, Brazil, Mexico, West Indies), Europe (Denmark, Sweden, Italy, Germany, Romania, France), Oceania (Australia) and Atlantic (Faroe Islands). For publication years, studies were grouped into eight periods, including 1 period before 80s and 7 periods after 80s with an interval of 5 years. For subgroup analysis according to age, it was grouped into four sections with an interval of 20 years. Publication bias was assessed by visually inspecting funnel plots and applying Egger's tests to evaluate sources of variability. For all tests, P value < 0.1 was considered to be statistically significant.

Literature search
A total of 1731 titles and/or abstracts of relevant studies were retrieved, and 1586 papers were removed due to irrelevance or review. The full-texts of remaining 145 papers were further reviewed, and 42 papers were excluded because of duplication and not providing sufficient information. Finally, 103 studies met the inclusion criteria and were included in this metaanalysis . The flow chart of study selection process was shown in Fig 1.

Study characteristics
Of the 103 studies, 82 were population-or community-based studies and 22 were hospitalbased studies. The countries were Faroe Islands, India, Denmark, USA, Australia, Sweden, Brazil, China, Italy, Germany, Tanzania, Saudi Arabia, Romania, Sri Lanka, Egypt, France, Turkey, Mali, Mozambique, Nepal, Iran, Korea, Mexico, Kuwait, West Indies, Thailand, Nigeria, Japan, Congo, Jordan and the areas covered Asia, Africa, America, Europe, Oceania and Atlantic. The years of publication ranged from 1964 to 2015. The sample size of included studies ranged from 102 to 50593516. The prevalence of vitiligo ranged from 0.004% to 9.98%. The characteristics of included studies were summarized in Table 1.
When stratified by publication years, the prevalence of vitiligo was 0.6% (0.4%, 0.9%) before the 80s. It decreased to 0.2%~0.3% in the 80s. The prevalence rebounded to 0.6% (-0.1%, 1.3%) in the first half of 90s. After that, the prevalence drastically decreased and maintained at a low level of 0.1%~0.2%.

First Author Publication Year Country Survey Age (years) Survey Year Sample (N) Vitiligo (n) Prevalence
Population or community-based studies The subgroup analyses of hospital-based studies ( Table 2) With regard to hospital-based studies, the prevalence of vitiligo was as high as 2.5% (1.6%, 3.4%) of Africa, compared with 1.6% (1.1%, 2.0%) of Asia and 1.5% (-0.1%, 3.1%) of America. Before the 80s, the prevalence of vitiligo was 2.6% (2.3%, 2.9%). The data of the 80s and the first half of 90s were not available. The prevalence of the latter half of 90s and the first 5 years of 21 st century were 1.9% (0.7%, 3.1%) and 2.0% (1.0%, 2.9%), respectively. In recent 10 years, the prevalence has decreased to 1.5%~1.6%. It has showed a decreased trend from 60s till now.

Publication Bias
There was significant publication bias in population-or community-based studies (t = 2.31, P = 0.026), while no significant publication bias existed in hospital-based studies (t = 0.47, P = 0.643). The funnel plot of publication bias was shown in

Discussion
To our knowledge, this is the first meta-analysis examining the prevalence of vitiligo. The results of this study showed that the pooled prevalence of 82 population-or community-based studies was 0.2% and of 22 hospital-based studies was 1.8%. The latter data derived from hospital-based surveys was obviously high.
Although vitiligo occurs worldwide, it is known that the reported prevalence of vitiligo is various. Prevalence distributions might differ in areas. In the included population-or community-based studies, the lowest prevalence was in Asia and Atlantic, the second-highest in Africa and in Europe, and the highest in Oceania. But only 1 study was included, the result of Oceania was not definite. In hospital-based studies, prevalence was the lowest in America and the highest in Africa. So we could draw a common conclusion from the two types of studies that Africa had a high prevalence of vitiligo. These results were in accordance with previous studies that vitiligo frequently occurred in darker-skinned individuals [109]. However, differences did exist  The Prevalence of Vitiligo: A Meta-Analysis in various reports. We suspect the differences that vitiligo is more prevalent in some geographic areas may result from the following factors. Firstly, different skin types and ethnic groups may play important roles in the discrepancy of the prevalence among different areas. Environmental conditions as well as genetic factors may solely or synergistically contribute to the various prevalence distribution in different geographic areas [7]. Secondly, the populations in many surveys were ethnically and culturally diverse such as the survey in USA. Several generations of immigrants or various population's lifestyles in this region might contribute to these differences. Thirdly, unbalanced number of included studies in geographic regions might compromise accurate and sufficient information for heterogeneity. Studies were extremely more conducted in Asia, while only 1 study was conducted in Oceania or Atlantic. Lastly, small amounts of participants in some included studies may contribute to imprecise estimates. In general, the prevalence of vitiligo showed a relatively decreased trend with increase in the times. Especially, it has remained at a low level in recent two decades in both population-or community-based studies and hospital-based studies. The association between vitiligo and its autoimmune diseases, such as autoimmune thyroid diseases, psoriasis, pernicious anemia, The Prevalence of Vitiligo: A Meta-Analysis Addison's disease et al has been frequently described in the literatures. As vitiligo may accompany with other diseases or disorders, we assume that the decreasing prevalence may be beneficial from development of diagnostic tools or improvement of screening programs or therapeutic methods of vitiligo-related diseases or disorders. The prevalence went up to 0.6% in the first half of 90s in population-or community-based studies. We found there was a literature written by Xue SQ about skin diseases of workers and technical personnel in Taigang company. The prevalence of vitiligo of this community population reached 1.27%, higher than other papers in this period. Chemical elements or decolorization may result in the increase of patients with vitiligo. The exact etiology and pathogenesis of vitiligo have not been completely unraveled. It involves a series of known and unknown environmental factors or immunological factors acting over time.
Our results also demonstrated that the pooled prevalence of vitiligo was slightly higher in females than in males, both for the population-or community-based studies (0.5% compared with 0.2%) and the hospital-based studies (1.3% compared with 1.1%). This result was different from previous literatures, which revealed that male and female patients were affected equally by vitiligo [4,58,110] or men were more affected than women [111]. Women usually incline to concern about pigmentation changes of their skin and the impact on their social life, and women may be more diligent in seeking treatment. This could be a possible reason for the greater number of female patients in this study [112].
Besides, this study revealed that the prevalence of vitiligo increased with age in populationor community-based studies, increasing gradually from 0.2% in the 0~19 years age-group to 0.7% in the ! 60 years age-group. Similar results have been found in some previous studies [8,58]. The increase possibly correlates to a cumulative effect, because vitiligo is a long-lasting disease and is life-long in most patients. However, in hospital-based studies, the prevalence in the 0~19 years age-group was higher than that in the ! 60 years age-group. Youngers may occupy more important position in the family and parents will take them to see a doctor as soon as The Prevalence of Vitiligo: A Meta-Analysis they find the children's conditions. In contrast, some elderly patients do not pay greater awareness of their appearance and will not see doctors unless necessary.
Despite we have conducted a comprehensive searching of the epidemiology of vitiligo, several limitations should be considered in this meta-analysis. The available publications/ studies were from 31 countries. The data of unavailable countries are required to reflect the wide variation. Some characteristics of the patients, such as clinical types, site or age of onset, risk factors, etc., were not included in the subgroup analyses. These might exert an important influence on the prevalence of vitiligo. Another possible limitation of this study was related to publication bias. The result from Egger's test showed an evidence of publication bias in population-or community-based studies. It may result from unbalanced number of studies and year of publications. For example, the number of included studies of 2001~2005 was 23, in contrast, the number of 1986~1990 was only 1. Finally, some included studies had noted methodological flaws, especially related to selection and recruitment of samples. Special subjects, such as soldiers, teachers, miners and other professional workers, participating in the investigations could not be representative of other samples. Control group with other diseases such as diabetes was also selected in some studies. As a result, the estimates of prevalence may have been influenced in unpredictable ways and need continuous perfectibility for verifying our conclusion.
In conclusion, we investigated a worldwide prevalence of vitiligo with population-or community-based and hospital-based data. A relatively high prevalence of vitiligo was found in Africa area and in female patients. The prevalence has maintained at a low level in recent years. It showed an inverse trend with age increment in the two types of studies. The current study provided a basic result for further studies. Future researches should be done to find key factors that contribute to the prevalence.