Prevalence of metal hypersensitivity in patients with shoulder pathologies

Abstract

To date, medical history and dedicated questionnaires are the fastest and easiest way to assess risks of joint metal hypersensitivity. No published studies determined the overall prevalence of hypersensitivity to metals in patients with shoulder pathologies. The purpose of this study was therefore to estimate the prevalence of metal hypersensitivity reported by patients with shoulder pathologies, and to identify patients at risk of joint metal hypersensitivity based on a dedicated questionnaire. The authors prospectively asked all adult patients consulting for shoulder pathologies between September 2018 and February 2019 at 10 centers to fill in a form. The main outcome was "reported hypersensitivity to metals,"comprising belt buckles, coins, earrings, fancy jewelry, keys, leather, metallic buttons, piercings, spectacles, watch bracelets, or zips. A total of 3217 patients agreed to fill in the survey, aged 55 ± 16 (range, 18-101) with equal proportions of men (51%) and women (49%), and a majority of patients consulting for cuff pathology (55%). A total of 891 (28%) patients had professions considered at risk formetal hypersensitivity. The most frequently reported metal hypersensitivities were fancy jewelry (15%), earrings (13%), and watch bracelets (9%). A total of 629 (20%) patients, of which the vast majority were women, reported hypersensitivity to 1 or more metals. This survey of 3217 patients identified 20% who reported metal hypersensitivities, though only 2.2% had done patch tests. Matching profiles of those with positive patch tests to those with no patch tests revealed that 9.4% of the total cohort had similar sex and self-reported metal hypersensitivities. Factors associated with a positive patch test were female sex, self-reported cutaneous allergy, and self-reported metal hypersensitivity. The clinical applicability of these estimates remains uncertain as there is insufficient evidence that allergy to metal implants can be predicted by questionnaires or patch tests.