Prevalence of contact allergy to metals in the European general population with a focus on nickel and piercings: The EDEN Fragrance Study

Background Studies on sensitization to metals in the general population are scarce. Objectives To determine the prevalence of sensitization to metals in the general population, and factors associated with nickel sensitization. Methods In 5 European countries (The Netherlands, Germany, Italy, Portugal and Sweden), a random sample (N = 3119) from the general population (aged 18‐74 years) was patch tested and interviewed by use of a questionnaire on exposure to metals, piercing, and jewellery. Results Overall, the age‐standardized prevalences of sensitization to nickel, cobalt and chromium were, respectively, 14.5%, 2.1%, and 0.8%. The highest prevalence of nickel sensitization was seen in Portugal (18.5%) and the lowest (8.3%) in Sweden. The prevalence of cobalt sensitization varied between 3.8% (The Netherlands) and 0.9% (Italy), and the prevalence of chromium sensitization varied between 1.3% (Portugal) and 0.2% (Sweden). Significant associations were observed between nickel allergy and female sex (odds ratio [OR] 5.19; 95% confidence interval [95%CI]: 3.99‐6.74), past piercing use (OR 3.86; 95%CI: 2.85‐5.24), and currently having ≥3 piercings (OR 5.58; 95%CI: 4.02‐7.76). Conclusions The prevalence of sensitization to metals in the European general population was high, mostly because of nickel. The lowest prevalence of contact allergy to nickel and chromium observed in Sweden supports the effectiveness of long‐standing regulation.


| INTRODUCTION
Nickel, chromium and cobalt are patch tested in the European baseline series. Results on contact allergy to metals in consecutive dermatitis patients in Europe have been reported frequently, 1,2 whereas studies on sensitization to metals in the general population are scarce. Thyssen et al reported a median nickel allergy prevalence of 8.6% (range 0.7%- 27.8%) based on data mostly from western Europe and North America.
The authors concluded that nickel was an important cause of contact allergy in the general population, and that contact allergy was

| Data collection and analysis
As reported previously, the interview was conducted face-to-face with a trained interviewer, and consisted of three parts. 4,5 The interview comprised, among other things, demographic and personal characteristics, and a description of a previous diagnosis of atopic dermatitis made by a physician or dermatologist. A lifetime prevalence of hand eczema was registered as follows: the interviewer asked whether the participant had ever experienced an itchy skin rash that lasted for >3 days, and subsequently the location of this rash on the body. Hand eczema was defined as being present if the location was the hand, and a diagnosis of contact dermatitis, atopic dermatitis or other dermatitis was registered. In addition, a detailed history of exposure to metals, piercing and jewellery was taken.
All statistical analyses were performed with SPSS 23 (IBM, Armonk, New York). Descriptive data are presented in tables as numbers with percentages and 95% confidence intervals (95%CIs). The main information is also presented in strata of sex and country of origin. Prevalences are presented as both crude estimates and age-standardized estimates with accompanying 95%CIs. Age standardization was performed according to the direct method. 7 The European standard population was taken as the reference for standardization. Measures of association between nickel allergy and main variables are expressed as odds ratio (ORs) with 95%CIs. OR estimates were obtained by univariate logistic regression analysis in a first step. Then, a multivariate analysis including all variables that were significant in the univariate model was performed in order to control for potential confounders. ORs were considered to be statistically significant when 1 was not included in the 95%CI.

| Socio-demographic characteristics
The socio-demographic characteristics of the 3119 patch tested subjects are shown in Table 1, together with the characteristics of the prevalence sample; these data have already been partly presented in Diepgen et al. 8 It can be seen that the subsample patch tested was quite comparable to the prevalence sample of the general population.
The percentage of females was slightly higher (54.9% females vs 45.1% males) in both samples. Age distribution did not differ substantially between the 2 samples, showing a slightly lower percentage of patch tested subjects in the oldest age group (71-84 years). In the subsample patch tested, 51.7% of the subjects had ever had a piercing and 10% had currently ≥3 piercings; these numbers were also comparable to those in the prevalence sample. The prevalence of ever having a piercing was significantly higher in females (81.5%) than in males (14.9%; P < .01). The prevalence of currently having ≥3 piercings was also higher in females than in males (17.0% females vs 1.5% males). Concerning results stratified by sex, across all study centres, the prevalence of nickel contact allergy was much higher in females (22.0%) than in males (5.3%); the prevalence of cobalt contact allergy was also higher in females (3.0%) than in males (1.1%), whereas, for chromium, the prevalence in females (0.6%) was lower than in males (1.0%). The highest prevalence of nickel allergy in Portugal was found in females (29.5%) but not in males (4.3%). In males, the highest prevalence of nickel allergy was found in The Netherlands (8.9%), whereas the lowest prevalence was found in Sweden (2.9%). Regarding chromium, the subdivision by sex showed comparable prevalences in males and females in Germany, The Netherlands, and Sweden. In Italy, the prevalence of chromium allergy was higher in males (0.9%) than in females (0.0%); this was also observed in Portugal (males, 1.7%; females, 1.0%).

| Patch test reactivity to nickel
The grades of patch test results of all patients who were positive for nickel, marked by country, are shown in Table 3. Strong (++) and extreme (+++) patch test reactions were combined as 1 group (++/+++). Overall, more strong/extreme positive reactions (10.3%) were observed than weak positive reactions (4.2%) to nickel. The distribution between strong/extreme and weak was not similar across the different countries. In Sweden, almost all sensitized subjects had strong/extreme positive reactions (8.2%) rather than weak positive reactions (0.1%). Also in Italy and Portugal, many more strong/extreme positive reactions (14.3%) were seen than weak positive reactions (1.1%). In Germany and The Netherlands, the differences between strong/extreme and weak reactions were not so obvious, being, respectively, 7.9% and 6.0% for Germany, and 9.1% and 6.7% for The Netherlands. The highest prevalence of irritant reactions was observed in Sweden (3.7%), and the lowest prevalence of irritant reactions was observed in Portugal (0%). The highest prevalence of doubtful reactions was observed in Germany

| Nickel allergy stratified by sex and piercings
The prevalence of nickel allergy stratified by sex, age group and never having had a piercing, having ever had a piercing but not currently, currently having 1 to 2 piercings, and currently having ≥3 piercings, respectively, is shown in Table 4. Overall, subjects who ever had a piercing but did not currently have a piericing showed a higher percentage of nickel allergy than the group who never had a piercing, namely 20.8% and 6.4%, respectively. The frequency of nickel allergy increased with the number of current piercings. Subjects with ≥3 piercings currently showed the highest prevalence of nickel allergy (27.6%), followed by subjects with 1 to 2 piercings currently, who had a prevalence of nickel allergy of 21.4%. Subdivision into age groups showed that, in subjects with ≥3 piercings currently, high prevalences of nickel allergy were found in all age groups:

| Factors associated with nickel contact allergy
The results of a logistic regression analysis to assess different risk factors for contact allergy to nickel are shown in Table 5   to 3% of the general population were sensitized in previous studies. 3 In the current study, we found similar prevalences, namely, cobalt allergy in 2.1% and chromium allergy in 0.8%.

| Nickel allergy and regulation in different European countries
The prevalence of nickel sensitization showed wide variation among the different countries; high age-standardized prevalences were seen Although the prevalence of nickel allergy has decreased since implementation of the EU nickel restriction, nickel is still a common cause of contact allergy, both in the general population and in the clinical population. This can partly be attributed to the lack of restriction regarding the many short and frequent contacts of consumers with everyday products containing nickel. 16 Another reason for the ongoing high prevalence of nickel allergy may be the risk of nickel exposure from consumer products such as mobile phones, laptop computers, and tablet computers, as the release of nickel from these products may not comply with the regulation. 17 In Germany, the Federal Institute for Risk Assessment reported on nickel in toys and metal construction kits for children. Overall, 41 of 168 toys exceeded the legal limit value for nickel release of 0.5 μg/cm 2 of toy per week, and 29 of 32 metal construction kits exceeded the legal limit. 18 Factors other than nickel regulation, such as occupational exposure, may also contribute to the ongoing high prevalence in the general population.
Interestingly, the proportion of strong or extreme patch test reactions (++/+++) varied substantially between the countries ( This indicates that the prevalence of nickel allergy does not depend on the age group, but on having piercings. Although the overall prevalence of nickel allergy was somewhat lower in the oldest group aged 61 to 74 years, this was less obvious in those with ≥3 piercings (23.1%).
In the multivariate regression analysis, in the groups aged 31 to 45 years and 46 to 60 years there was an increased risk of nickel contact allergy, with the youngest age group as a reference. In the oldest group aged 61 to 74 years, there was a decreased risk of nickel allergy. The prevalence of nickel allergy decreases with increasing age, owing to different frequencies of ear piercing in different generations, and probably also because of a decrease in exposure to jewellery. It has also been reported that the immune response of the skin diminishes with ageing, owing to senescence of the immune system. 23 This can also explain the lower prevalence of nickel allergy in the oldest age group. A recent pilot study by Lusi et al reported a higher prevalence of nickel allergy in an overweight female population. 24 In the univariate analysis, we found the opposite; however, this effect was no longer significant in the multivariate model, and there were also no differences in effects between males and females.

| Chromium and cobalt allergy
In the current study, the prevalence of chromium allergy was lowest Leather products have been described as important causes of chromium contact allergy in the clinical population. 26 Leather shoes were the most frequent sources of relevant exposure in patients with chromate allergy, more so in females than in males. 27 Leather glove exposure occurred more often in males than in females. The use of chromium in leather tanning could be a contributory factor to the high prevalence of chromium allergy in Portugal, as the prevalence in females was also high (1.0%), and people may wear shoes without The prevalence of cobalt allergy in the general population was higher in females (3.0%) than in males (1.1%), which can possibly be explained by exposure to cobalt in jewellery. In a clinical population, pronounced concomitant reactivity between nickel and cobalt was observed, especially in females. 1 Other sources of cobalt exposure are other metal consumer objects, prosthetics, paints, and pigments.
Concerning cobalt, there is no legislation yet to limit the amount of cobalt in consumer products.

| Limitations
The response rates might constitute a study limitation. Selection bias at the first stage of recruitment cannot be ruled out, owing to the response rates, and might have been a reason for some of the international differences observed.

| CONCLUSION
The data presented show that the prevalence of metal contact allergy in the general population was high, mostly because of nickel. The ongoing high prevalence of nickel allergy shows the importance of complying with the regulation, also including new consumer products.