Efficiency in patch testing: the number needed to test to get one relevant result as a new approach in the evaluation of baseline series

Dear Editor, Patch testing is a common procedure, in which the baseline series is frequently used as a first screening. The number and the specific allergens of these series varies from country to country and keep changing over the years. Although the European Society of Contact Dermatitis (ESCD) made recommendations about patch testing,1 the decisions about the European baseline series were traditionally delegated to the European Environmental and Contact Dermatitis Research Group (EECDRG).2 A different, more reactive, approach has been proposed recently by the ESCD task force,3 using data provided by the European Surveillance System on Contact Allergies (ESSCA) to generate an initial discussion. The main criteria for the addition of a new allergen is that the prevalence of allergy is 0·5–1% or higher,4 and positive results should be relevant. As a new indicator that considers both the frequency of positive tests and their relevance in a simple measure, we propose using the number of tests needed to get one relevant result to quantify efficiency (number of tests done/number of relevant results). REVAC (Red Española de Vigilancia de Alergia de Contacto) is a surveillance system across Spain working with the ESSCA, with eight participants. It has been approved by the Hospital del Mar Ethics Committee (2008/3159/I). We reviewed all consecutive patch tests done between 2004 and 2014, including 9767 patients. The clinical relevance of the results was analysed in each case by the participating dermatologist and classified as present or past relevance by describing the association between exposure and contact dermatitis symptoms (e.g. allergic contact dermatitis).4,5 The study includes the European baseline series, the Spanish baseline series, and some other allergens that participants test in all patients and which could become part of the baseline series. A few allergens show high testing efficiency (between 16 and 52 tests needed to detect one positive currently relevant result). Many of the allergens are inefficient, with up to 2437 tests needed to detect one positive currently relevant result (Table 1). The results are slightly better for ever‐relevant results. Number of patch tests needed to detect relevant positive results Excipients used: pet, Vaseline; aq, water. Number of patch tests needed to detect relevant positive results Excipients used: pet, Vaseline; aq, water. The advantages of this study is that it is highly representative of patch testing in specialized care, and that all participants use the same method and are experts in the topic, decreasing measurement error. It is likely that results in less specialized settings could be less efficient, as the prevalence of contact allergy is lower. The main limitations to consider are the subjective definition of relevant results and the common difficulty to identify with certainty the source of exposure to the contact allergen. However, we have chosen positive and relevant results as the main outcome for our measure, instead of positive results, because only positive and relevant results offer the opportunity to modify management and improve our patients’ health. Similarly, this index gives a low value to negative results, because although they can support the diagnosis of irritative contact dermatitis, they do not modify patient management. These findings suggest that some tests of the baseline series are efficient and some are highly inefficient. The number needed to test to get one currently relevant result could be very useful for the decision to include tests in the baseline series, as it considers not only frequency but also relevance of positive tests. If we accept the original criteria for the addition of a new allergen to the baseline series, that ‘prevalence of allergy should be 0·5–1% or higher and positive results should be relevant’,4 that would mean having in the baseline series only allergens with a number needed to test lower than 100–200. Our results open the discussion about the possibility of using a shortened series for screening, in general settings, and a specialized series in contact dermatitis clinics. This should be the topic of further economic and decision analysis. we thank Wolfgang Uter for his support for data collection. Funding sources: no external funding. Conflicts of interest: none declared.