Antineutrophil cytoplasmic antibodies and HLA class II alleles in minocycline-induced lupus-like syndrome

We report 14 patients with minocycline-induced lupus-like syndrome (four men, 10 women; mean age 27·8 years) who developed a lupus-like illness after chronic use of minocycline for acne (1–10 years, median 3·8). Clinical features resolved completely on drug withdrawal (mean follow-up 11 months) and reappeared in two patients who were rechallenged. Sera from all 14 patients contained antineutrophil cytoplasmic antibodies (ANCA) giving a perinuclear pattern on indirect immunofluorescence on ethanol-fixed human neutrophils (p-ANCA), whereas 14 control asymptomatic individuals taking minocycline for acne were ANCA-negative. Eleven of the 14 patients had elevated antimyeloperoxidase antibodies and 10 had antielastase antibodies on enzyme-linked immunosorbent assay, which diminished on extended follow-up, as did other serological abnormalities. Major histocompatibility complex class II typing demonstrated that all of the 13 patients tested were either HLA-DR4 (nine of 13) or HLA-DR2 (four of 13) positive, and all had an HLA-DQB1 allele encoding for tyrosine at position 30 of the first domain. Our findings suggest a model whereby the presence of p-ANCA may be a marker for the development of lupus-like symptoms in genetically susceptible individuals taking minocycline for acne.

Minocycline is a frequently prescribed antibiotic in the U.K. for the treatment of acne. There have been several recent reports of patients who have developed a lupus-like illness after the chronic use of minocycline, including two deaths from autoimmune hepatitis and one from pancytopenia. 1–9 The most common clinical features associated with minocycline-induced lupus-like syndrome (MILS) are arthritis, skin rash, hepatitis and fever, and some patients have required treatment with corticosteroids. It would appear that relatively few users of minocycline develop side-effects, 10 although a recent nested case–control study reported an 8·5-fold risk of developing a lupus-like syndrome with current use of minocycline that increased to 16-fold with long-term use for the treatment of acne. 11 Given the wide usage of minocycline in a relatively young and otherwise healthy population, it would be helpful to have markers for those individuals at increased risk of toxicity. Furthermore, the identification of a commonly used drug associated with a lupus-like syndrome may provide a useful experimental model for other idiopathic autoimmune conditions.

Drug-associated lupus-like syndromes are commonly associated with antibodies to nucleosomal constituents and especially components of histone. Several studies have reported the occurrence of antineutrophil cytoplasmic antibodies (ANCA) with drug-induced lupus. 12–14 For instance, Nassberger et al. 13 found antibodies to myeloperoxidase (MPO) and/or elastase in six of six patients with hydralazine-induced lupus syndrome. More recently, perinuclear immunofluorescent pattern ANCA (p-ANCA) antibodies have been reported in seven patients with MILS. 15 Serological abnormalities may precede symptoms, and usually resolve after withdrawal of the offending agent. 16

Host susceptibility factors have been reported, notably slow acetylator status, 17 for the two most commonly studied agents, procainamide and hydralazine. Major histocompatibility complex (MHC) loci have been implicated, with an association between hydralazine-induced lupus and HLA-DR4. 18 We report 14 further patients with a lupus-like syndrome associated with the contemporary use of minocycline, all of whom had p-ANCA on autoantibody screen. We have analysed the time course and specificity of autoantibody reactivity and also investigated possible immunogenetic risk factors residing at MHC class II loci that may have a bearing on pathogenesis.