Validation of diagnostic criteria for atopic dermatitis and proposal of novel diagnostic criteria for adult and elderly Chinese populations: a multicentre, prospective, clinical setting-based study

A previous validation study showed a very low sensitivity and higher specificity associated with Hanifin and Rajka criteria (H&R) and the UK Working Party criteria (UKWP) in diagnosing AD vs. the Chinese criteria of atopic dermatitis (AD) for children (CCAD). However, their diagnostic efficacy in adult and elderly Chinese populations remains unknown.To validate the diagnostic efficacy of three sets of AD criteria in adult and elderly Chinese populations in a hospital setting.A total of 1034 patients (aged 19-95 years) from five university hospital dermatological clinics were recruited. Medical history, dermatological examination, AD diagnosis and evaluation of AD severity were done by dermatologists. Each patient was investigated by two dermatologist panels, one to establish a clinical diagnosis, and the other to identify and record the major or minor signs of H&R criteria, UKWP criteria and CCAD. Taking clinical diagnosis as the reference, the diagnostic efficacy of three sets of diagnostic criteria was evaluated. The χ2 test or rank sum test were used for between-groups comparisons.CCAD had a higher sensitivity (84.0%), especially among mild and moderate cases of AD (72.7% and 90.3%, respectively), than the H&R (58.0%; P < 0.001) and UKWP criteria (56.0%; P < 0.001) in diagnosing AD. The specificity of CCAD (92.7%) was slightly lower than the H&R (97.3%; P < 0.001) or UKWP criteria (97.4%; P < 0.001). The CCAD had the highest Youden index (0.77), accuracy rate (0.90) and Kappa value (0.76) of the three sets of diagnostic criteria.Consistent with results in a population of Chinese children, although the H&R and UKWP criteria had a high specificity for diagnosing AD, their low sensitivity limited their use in adult and elderly Chinese patients. Based on the high sensitivity and favourable diagnostic efficacy, the CCAD is proposed for AD diagnosis in adult and elderly Chinese populations, especially for cases of mild and moderate AD.