Background Erythroderma has protean underlying causes. There have been isol
ated case reports suggesting an association between erythroderma and the hu
man immunodeficiency virus (HIV).
Objective To describe and characterize further the prevalence, etiology, an
d metabolic sequelae of erythroderma in HIV positive and negative patients.
In a subset of patients, clinicopathologic correlation was performed.
Method One hundred and thirty-eight consecutive patients were prospectively
recruited over a one and a half year period at the skin clinic of King Edw
ard VIII Hospital. Demographis, clinical, biochemical, and histologic data
were recorded.
Results Seventy-five per cent of the patients were black, 22.5% Indian, and
2.5% white. University of The men to women ratio was 1.9 : 1. The mean age
was 34.7 years (range, 1 month to 85 years). Forty-three per cent of patie
nts were HIV positive, of whom 90% were black. The commonest causes of eryt
hroderma in the total sample were atopic dermatitis (23.9%), psoriasis (23.
9%), and drug reactions (22.5%). The commonest cause in the HIV positive gr
oup was drug reactions (40.6%), the commonest being ethambutol (30.8%). HIV
positive patients had a significantly lower (P < 0.05) white cell count (7
.6 vs. 10.5 x 10(9)/L), hemoglobin (11.1 vs. 12.6 g/dL), platelets (278.3 v
s. 378.0 x 10(9)/L), and albumin (25.4 vs. 28.7 g/L) and significantly high
er serum urates (0.6 vs. 0.4 mM/L) than HIV negative patients. HIV positive
patients did not have a significant increase in the number of episodes of
erythroderma. Clinicopathologic correlation was greatest with psoriasis in
the HIV negative group and with psoriasis and drug reactions in the HIV pos
itive group.
Conclusions A large proportion of erythrodermic patients in this study were
HIV positive. Inflammatory dermatoses were the commonest cause of erythrod
erma in all the patients studied. Drug reactions were the commonest cause i
n HIV positive patients. In the young black patient, erythroderma may be a
marker for HIV infection.