Quantification of aluminium release from Finn chambers under different in vitro test conditions of relevance for patch testing

Contact allergy to aluminium (Al) might pose a risk of false‐positive readings of patch‐test results when testing with Finn chambers.


| INTRODUCTION
Contact allergy to aluminium (Al) among dermatitis patients in general has not often been reported, and as this substance is not included in the baseline series in most countries, the real frequency of this contact allergy is unknown. In a recent French study, 1 a surprisingly high frequency of contact allergy to Al (21.6%) was reported in consecutively patch-tested children. Vaccines and immunotherapy seem to be main causes of the development of contact allergy to Al 2-6 ; however, a recent study could not confirm the clear role of immunotherapy. 2 In a Swedish prospective cohort study comprising 4758 children, 0.83% of vaccinated children developed intensely itching subcutaneous nodules at the injection site for Al-adsorbed vaccines. 7 Generally, the higher the Al dose and the more frequently injections are given, the higher the risk for developing contact allergy to Al 2,7 ; individuals with atopic dermatitis seem to have an increased risk. 2 Once sensitized, other elicitation sources can be cosmetics, deodorants, Al metal, eardrops, toothpaste, and tattoos, 3 but the bioavailability of different Al sources is not well investigated. A recent Danish questionnaire study of 177 Al-allergic children and their parents compared with a reference group concluded that itching vaccination granulomas and Al allergy have a considerable negative impact on these children and their families, causing for instance, reduced adherence to vaccination programs and a lower score on overall life quality as compared to the reference group. 8 Contact allergy to Al is not easily diagnosed, as the elicitation threshold might be higher than the patch test substance used 9 and since there is a considerable individual over-time variation in patchtest results, 10 resulting in a high risk of false-negative results in Alallergic individuals. A recent study on 241 children with previous vaccine-induced itching nodules found that patch testing with 2% Al chloride hexahydrate in pet. gave more positive reactions as compared with an empty Al Finn chamber. 11 It has been reported, however, that false-positive reactions to various allergens applied in Finn chambers can occur in Al-allergic individuals. 12 The objectives of this study were to (a) quantify the release of Al from Al Finn chambers and Finn chambers Aqua (covered Al chambers) as compared to common patch-test skin doses of Al chloride hexahydrate and (b) to quantify the release of Al from Al Finn chambers containing different baseline patch-test substances.

| Al release testing
The Al release from Finn chambers (about 0.05 g weight) and Finn chambers Aqua was tested in vitro in artificial sweat (ASW, 5.0 g/L NaCl, 1.0 g/L urea, 1.0 g/L lactic acid, pH adjusted to 6.5 ± 0.05 with NaOH). 13 All chemicals for artificial sweat were of analytical grade and obtained from VWR, Sweden. The total surface area of the Finn chambers (both sides) was approximately 1.57 cm 2 (their outer diameter was 10 mm) and the total volume of the ASW was 1.57 mL, giving a surface area to volume ratio of 1 cm 2 /mL. side, so that only the paper-covered front side (0.785 cm 2 ) was exposed to the artificial sweat (1.57 mL), corresponding to a surface area to volume ratio of 0.5 cm 2 /mL. Finn chambers of different types and after different surface and cleaning conditions were compared (Table 1). In addition, Al Finn chambers (batch 3) with applied patchtest substances were tested (Tables 1-2). Thirty-two different patchtest substances ( Table 2) were applied on 2-3 Finn chambers for each substance, with 15 μL for aqueous solutions and 20 mg for substances in pet., which is in agreement with European guidelines for patch testing in 8 mm (inner diameter) Finn chambers. 14 For the application of pet. substances, the Finn chamber was placed on a balance (0.1 mg accuracy) and 20 mg of the pet. substance was added by means of a cleaned stainless steel spoon onto the Finn chamber, covering an area of 0.3-0.5 cm 2 . For application of the aqueous solution, a round (0.5 cm diameter) paper filter (Finn Chambers Filter Paper Discs, Lot# ALBAGIRM, SmartPractice) was placed on the Finn chamber and 15 μL of the aqueous patch-test substance was added onto that paper filter. The Finn chamber with the patch-test substance was then placed in an acid-cleaned (10 vol% nitric acid for at least 24 hours followed by four times rinsing with ultrapure water) centrifuge tube and the ASW was added. Due to the conical shape of the centrifuge tube, the entire Finn chamber was exposed to the ASW solution. Batch 1 and batch 3   Appendix S1, together with the raw data.

| Calculations and data presentation
The amount of Al release is presented in the unit μg/cm 2 and calculated by:  Table 1).

| Optical microscopy
Ten unexposed Finn chambers and 3 of batch 2 exposed to ASW were inspected visually and by optical microscopy, showing a clear difference with a slightly brownish appearance of the exposed Finn chambers, while all unexposed Finn chambers appeared shiny metallic.
This brownish discoloration after exposure to ASW was distributed over the whole exposed surface and there was no specific discoloration along the lacquer, which was used to seal the back side of these Finn chambers, indicative of no crevice corrosion. A Leica DM2770M light optical microscope equipped with a Leica DFC295 camera was used to take representative images at different magnifications up to 100 times (Appendix S1, Table S1).

| Statistical analysis
A Student's t-test of unpaired data with unequal variance (KaleidaGraph 4.0) was used to calculate whether differences among samples were statistically significant (P-value <.05).

| RESULTS
Al release testing from Finn chambers of batch 1 is shown in Figure 1A (left). The Al release into ASW was between 125 and 350 times larger as compared to the release into ultrapure water (P < .05). Although   Background concentrations of all patch-test substances, the paper filters, and the metal-free lacquer used to seal the back side of the Finn chambers Aqua and batch 2 were all found to be negligible (<3 μg/L) and close to blank concentrations (Table S3, Appendix S1), as compared to the sample concentrations with Finn chambers of batch 3 (591-21 536 μg/L) (Table S2, Appendix S1).

| DISCUSSION
Al belongs to the passive metals that are protected by a thin surface oxide that hinders corrosion and dissolution effectively in neutral aqueous solutions. 15 However, Al metal is susceptible to localized corrosion and sometimes other types of corrosion in salt solutions, solutions containing certain anions and organic acids, in contact with other metals, and strongly acidic or alkaline solutions. 16 Chlorides have particularly strong effects on localized corrosion of Al metal, 15,16 which can explain the high release from Finn chambers observed in  Table 2). The error bars show the standard deviation of two to five independent replicate samples (c.f. Tables 1 and 2). The dashed line marks the Al release from the empty Finn chambers without the presence of a patch-test substance. L, statistically significantly lower Al release as compared to the empty Al Finn chambers of the same batch; H, statistically significantly higher Al release as compared to the empty Al Finn chambers of the same batch this study in the presence of chloride-containing patch-test substances and in ASW as compared to pure water.
The release of metals from passive metals is further strongly influenced by the surface preparation or storage conditions of the metals prior to testing (for as-received surfaces). 17,18 Generally, longer storage time and a more humid, warmer, and acidic storage atmosphere will result in lower subsequent release of metals when tested without any further surface preparation (as-received). In this study, three different batches of Finn chambers with slightly different (unknown) age and storage conditions were tested and showed partially statistically significant (up to 30-fold) differences in Al release at similar test conditions for as-received (non-treated) Al Finn chambers.
This result is interesting, as it could possibly explain the difference observed in reactivity to empty Finn chambers observed in different studies. 6,9 Further studies are required to understand this baseline variation of Al release from empty Finn chambers. In addition to different transport and storage conditions, the Al release for different batches could also be influenced by factors caused and controlled during manufacturing, such as inclusions, differences in impurities, or in internal stresses (residual tensile or compressive stresses originating from external stresses during manufacturing such as bending).
Another difference is the sealing of the back side. The batch that served as reference for the Finn Aqua chambers (batch 2), and was therefore sealed on the back side, released the highest amounts of Al both in absolute concentration and dose per surface area (Appendix S1). It cannot be determined in this study whether this was caused by a difference in storage conditions, batch, or an effect of the sealing procedure. Sealing a passive metal susceptible to localized chlorideinduced corrosion poses a risk of causing a micrometer-thick crevice that might serve as initiation site for localized corrosion and therefore accelerate the process of Al release, 19 but no crevice corrosion was observed visually and by optical microscopy in this study. Instead, the corrosion type resembled pitting corrosion and was distributed over the entire exposed surface (Appendix S1, Figure S1).
Several of the investigated patch-test substances contain chlorine, as covalent bound chloride in organic molecules, as easily dissociated chloride in metal salts, or as hydrochloric acid (Appendix S1 nor any other Al patch-test substance is currently included in the Swedish national or in international baseline series, it might be difficult to recognize Al allergy, and hence there is a risk of false-positive reactions, and consequent misdiagnosis, to other haptens in Al-sensitized individuals. Al allergy is relatively common in some countries and age groups (about 1% of general population) and might therefore pose a serious risk of jeopardizing a correct diagnosis using patch testing with Finn chambers. Several Swedish studies are currently ongoing to investigate whether Al release from Finn chambers could influence diagnostic outcomes. This has also been discussed recently for isolated palladium allergy. 23 From a chemical point of view, release of Al from Al Finn chambers could be of concern for current patch-test diagnostic outcomes.

| CONCLUSION
The amount of Al released from empty Al Finn chambers corresponded to a skin dose of approximately 0.03%-0.5% Al chloride hexahydrate applied in plastic chambers. Finn chambers Aqua released significantly lower (16-to 4100-fold) amounts of Al.
Although most patch-test substances reduced the release of Al from the Al Finn chambers, some substances significantly increased the release of Al from the Finn chambers, most notable for Caine mix II 10% pet., M. pereirae resin 25% pet., and sodium tetrachloropalladate hydrate 3.0% pet. (corresponding to 0.5% Al chloride hexahydrate).
The release of Al from Finn chambers should be considered in further development of diagnostic patch testing. We strongly recommend patch testing of Al chloride hexahydrate 10% pet. in a plastic chamber as a control substance if Al Finn chambers are used for patch testing.

CONFLICTS OF INTERESTS
The authors declare they have no conflicts of interest.