Topical diacerein for epidermolysis bullosa: a randomized controlled pilot study

Abstract

Blistering in epidermolysis bullosa simplex type Dowling-Meara (EBS-DM) is associated with an inflammatory phenotype, which can be disrupted by diacerein in vitro. In this pilot study we hypothesized, that a topical formulation of diacerein 1% reduces blistering. Five patients initially applied diacerein underneath both armpits. Then, each participant received 1% diacerein-cream for one armpit, and placebo for the other (randomized withdrawal). The number of blisters was reduced significantly (left: -78%; right: -66% of baseline) within two weeks and remained significantly below the initial level even during withdrawal in four patients. These findings point to a relevant effect of diacerein and provide important information for a confirmative study.

Figure 1

Study design comprising an open phase and a controlled-randomized, double-blinded withdrawal. (A) Five patients were included in the pilot study, all of them harboring a mutation in K14 protein position 125. Initial blister numbers are given. *Note that, contradictory to an aforegoing patient inquiry, patient DIDM005 did not have any blisters at the first measurement. (B) Representative pictures of the right armpit of DIDM002 during t1 and t8. (C) The pilot study was a two-phase study with a 6 weeks open phase (P1), followed by a six-weeks double-blinded, randomized and controlled phase (P2). During P1, patients daily applied 1% diacerein cream underneath both armpits. During P2, one armpit of each patient was assigned to placebo in a double-blinded and randomized manner. For further data analysis, the study was divided into eight equidistant time spans (t), each comprising six counts. Patients counted blisters every second day, starting from day 1.

Figure 2

Results of blister counts show a significant reduction during P1. Patients received diacerein (grey graphs) to apply underneath both armpits during the open phase P1. During P2, one armpit was assigned to placebo (black graph) in a double-blinded, randomized manner. (A) During P1, there a significant reduction of blister numbers (given in%) from time span t1 to time span t2 and from t2 to t3 (t1:100%, t2:mean left: 22%/CI6-38; mean right: 34%/CI20-48). On both sides, t2 to t8 show a significant reduction of blisters compared to t1 (upper black line, * indicating a p-value ≤ 0.05). Blister numbers remained low until the end of the study period. The mean of blister numbers is given with 95% confidence intervals, whereby t1 is normalized to 100%. (B) During P2, there is a significant difference between placebo and diacerein at t7 (* indicates a p-value ≤ 0.05).