FC-07   THE AGE AT ONSET OF ALOPECIA AREATA

AJG McDonagh, *SP Macdonald-Hull, AG Messenger, MP Birch, MJ Cork, R. Tazi-Ahnini. Department of Dermatology & Division of Genomic Medicine, Royal Hallamshire Hospital, Sheffield & *Pontefract General Infirmary, UK.

Childhood alopecia areata (AA) has an unfavourable prognosis but hitherto, distribution of the age at onset in AA has been little studied. In psoriasis, age at onset analysis led to a classification showing significant differences in HLA associations between early and late onset subgroups. To examine the possibility of such age at onset subgroups in AA we have analysed a series of 500 cases, also taking account of associated diseases and family history data. Age at onset of hair loss followed a bimodal distribution with onset peaks in the teenage years and around age 30. The bimodal pattern was apparent both in the overall dataset and when patchy AA and other subtypes were plotted separately. There was no significant difference between men and women but the median age at onset was lowest in alopecia universalis (20 yrs, P<0.001) and significantly higher in patchy (29 yr) and diffuse AA (51 yr) than other clinical subtypes (P<0.001). The frequency distribution of age at onset in atopic individuals was unimodal, in contrast to the non-atopic who showed a prominent secondary peak of late onset cases. In keeping with the distribution of age at onset, we divided the patients into two subgroups with the boundary at age 20 years. Nail involvement, atopy and a family history of AA were all commoner in group I (39% vs 21%, p<0.001, 40% vs 29%, p<0.02, 30% vs 23%, p=0.053 respectively). Patchy AA was commoner in group II (64% vs 49%, p<0.01). When age at onset data has been included in published series, authors have usually presented the data in aggregated form using wide age groups leaving little new information to be obtained by reanalysis. A single study (Sanz y Benitez, 1936) included a frequency distribution providing comparative data showing a distribution similar to our series. Our findings provide an empirical basis for subdivision of AA patients by age at -26- onset. This is an observation worthy of further investigation with potential importance for future genetic analysis.