Poster Presentation 2014 Cutaneous Biology Meeting

Assessing Atopic Dermatitis in Pigmented Skin Patients (A Grey Area) (#35)

Cathy Zhao 1 2 , Asri Wijayanti 1 , Melanie J Doria 1 2 , Adam G Harris 1 2 , Swaranjali V Jain 1 , Kristine Legaspi 1 2 , Matthew G Law 1 3 , Dedee F Murrell 1 2
  1. Medicine, University of New South Wales, Sydney, NSW - New South Wales, Australia
  2. Dermatology, St George Hospital, Sydney, NSW - New South Wales, Australia
  3. Kirby Institute, Sydney, NSW - New South Wales, Australia

Background: Skin pigmentation can mask the perception of the severity of atopic dermatitis (AD). Outcome measures for AD patients with pigmented skin have not been validated.

Objective: To compare the reliability and validity of AD outcome measures in patients with various levels of skin pigmentation, and to determine the effect of erythema perception on reliability.

Methods: The inter-rater reliability, intra-rater reliability and construct validity of the four most commonly used AD outcome measures were evaluated in patients of various levels of skin pigmentation. These measures were the Three Items Severity index (TIS), Six Areas, Six Sites Atopic Dermatitis (SASSAD), EASI (Eczema Area and Severity Index) and SCORing Atopic Dermatitis (SCORAD). Each score’s non-erythema components’ variances were compared with its erythema components’ variances. Full body photos of 18 patients (four with highly pigmented skin, seven with mildly pigmented skin, and seven with non-pigmented skin) were assessed by five trained clinicians using the four outcome measures.

Results: For inter-rater reliability, the intra-class correlations (ICC) were poor for the highly pigmented group: TIS -0.21, SASSAD -0.071, EASI -0.054 and SCORAD -0.089; fair for the mildly-pigmented group: TIS 0.524, SASSAD 0.341, EASI 0.464 and SCORAD 0.588; fair to substantial for the non-pigmented group: TIS 0.403, SASSAD 0.667, EASI 0.640 and SCORAD 0.586. There were significantly different variances between the erythema and non-erythema components of highly pigmented groups’ SASSAD, EASI and SCORAD. This suggests the erythema component contributed to the poor ICCs. For intra-rater reliability, no differences were found between different levels of skin pigmentation.

Limitations: The patients were virtually scored, instead of in a routine patient clinic setting.

Conclusion: All outcome measures were highly unreliable in patients with very dark skin. A new outcome measure that uses shades of grey instead of erythema in pigmented skin patients with AD is implicated.