How Frequently Are Drugs Associated With the Development or Exacerbation of Subacute Cutaneous Lupus?

IN 1985, Reed et al¹ published the first case series of subacute cutaneous lupus erythematosus (SCLE) that was linked to the administration of a drug. They reported that the onset of the cutaneous eruption occurred during the administration of hydrochlorothiazide treatment and resolved with its discontinuation. In addition, 1 of their patients was treated with a second administration of hydrochlorothiazide that was followed by a reappearance of the eruption, which then cleared again with discontinuation of treatment with the offending drug. Their patients were anti-Ro/SSA positive, and in 1 of the 3 patients subsequently tested, the antibody disappeared.

In 1986, my colleagues and I reported our experience with SCLE.² Following the initial review of our manuscript, we were asked by one of the reviewers to assess the potential occurrence of drug-induced disease in our group of patients. At that time we did not find any case of drug-induced or drug-worsened disease. In a follow-up report that included 88 patients, we again noted no relationship between a drug and SCLE.³ However, in a study of SCLE that appeared in a recent issue of the ARCHIVES, coworkers and I found 3 patients with drug-induced SCLE among 76 patients with the disease.⁴ Two of these formed part of our report of 5 cases of terbinafine-induced or -exacerbated SCLE published in the ARCHIVES last fall.⁵ In the report by Srivastava et al⁶ in the present issue of the ARCHIVES, in 15 of 70 patients with Ro/SSA-positive cutaneous lupus erythematosus (LE), the potential cause was drug induction. This frequency is striking and suggests that in the population referred for tertiary medical care to Johns Hopkins Medical Institutions, there might be an overrepresentation of drug-induced disease compared with community-based practices.

There are several issues raised by Srivastava et al.⁶ First, is the frequency outside of a tertiary center higher or lower than that observed in their report? Second, how might a practitioner diagnose drug-induced SCLE? Is the presence of the Ro/SSA antibody necessary? What types of skin lesions are required? Are positive dechallenge and rechallenge needed? And how long an interval from initiation of a drug to the appearance of the eruption is needed? A third question is that of mechanism of induction of lesions—is the appearance of the skin lesions in these patients merely a reflection of the development of a photosensitive eruption in a patient with underlying SCLE, or is this occurrence due to induction of antibodies that are pathogenetically involved in the production of the lesions. Finally since many of the patients with drug-induced disease continue to have cutaneous LE following the "induction" by the drug, what is the exact role of the drug—an initiator of disease or the awakening of an underlying disorder?

The frequency of drug-induced disease found by Srivastava et al⁶ is higher than previously reported and may reflect the referral nature of patients to Johns Hopkins Medical Center, or it may be due to the inclusion of only Ro/SSA-positive patients in the report. The diagnosis of SCLE is not dependent on the identification of the Ro/SSA antibody; rather it is a clinical diagnosis that is confirmed histopathologically. Therefore, there may be many more patients with SCLE who were seen at this medical center than the 70 included in the report, and their inclusion would then result in a lowering of the percentage of patients with SCLE linked to drug administration, which would then be in line with the experience of Black et al.⁴ What I believe is needed is an examination of patients seen in a nonreferral setting, which might reflect the true frequency in a nonreferral setting.

I also question whether the choice by the authors to include any patient with SCLE diagnosed within 6 months of the administration of a "new" drug is a reasonable one. Most drug eruptions occur within a short period of the administration of the drug and resolve relatively rapidly on discontinuation of treatment with the drug. However, there are reports of minocycline-induced systemic LE that occur after long periods of administration of the drug, but these resolve rapidly on discontinuation of the minocycline therapy. In the report by Srivastava et al,⁶ the eruption began between 4 and 20 weeks after treatment with the offending drug was begun and resolved, generally with minimal intervention, between 6 and 12 weeks after discontinuation of the suspect drug treatment.

Most of the patients developed a photosensitive erythema rather than classic annular or papulosquamous lesions of SCLE. Many of the drugs reported to be involved in this report are known to be photosensitizing agents, and therefore one might wonder if the drug reactions reported are merely phototoxic drug reactions in a patients who happen to possess Ro/SSA antibodies. This is rendered less likely by the observation that all of the patients had histopathologic and immunoflouresence changes compatible with cutaneous LE. Another consideration for many of the patients with previously reported annular lesions of SCLE is whether the occurrence is actually an erythema multiformelike eruption in a Ro/SSA-positive patient. In this instance, the histologic findings might also result from an interface dermatitis. In my opinion, this issue is not a factor in this report.

Srivastava et al⁶ have added 2 new drug classes to the already expanding list of drugs associated with drug-induced SCLE: statins (lipid-lowering agents) and interferon alfa. They have also noted a unique diagnostic problem. The 2 patients in whom a statin was presumed to be the cause of the disease also had elevated creatine kinase levels and were weak. However, through careful evaluation, it became evident that these patients actually had 2 separate drug reactions—namely, drug-induced myopathy and cutaneous LE. Therefore, the patients in whom statins have been linked to dermatomyositislike disease should be reexamined for the possibility that they actually have SCLE. The authors of this report make a point to note that these 2 patients were Jo-1 negative. However, Jo-1 is not frequently present in dermatomyositis, and therefore its absence is not helpful in excluding a diagnosis of dermatomyositis. The basis of the diagnosis of drug-induced dermatomyositis should be the presence of clinically characteristic skin lesions, not a photosensitive eruption only.

The authors also identified 2 patients in whom the disease was due to a biologic agent—interferon alfa in one and interferon beta in another. Recent reports have documented the appearance of SCLE or subacute LE from etanercept and infliximab,⁷ and these anti–tumor necrosis factor α agents should be added to the table in the report by Srivastava et al⁶ of drugs that might induce LE-like disease.

Whether the appearance of the skin lesions in these patients is a reflection of the development of a photosensitive eruption in a patient with underlying SCLE or is due to the induction of Ro/SSA antibodies is not known. However, the observation by Srivastava et al⁶ that the antibody titers dropped in most and became undetectable in at least 3 patients suggests that the drugs may induce antibodies that might be involved in the pathogenesis of the disease. However, there are multiple reports in which the patient's clinical and serologic disease continues following drug treatment cessation. In these instances, the disease may be awakened by the drug and, once awakened, may continue. Further studies of prospectively identified patients might help us to elucidate the mechanism of this phenomenon.

What conclusions should be drawn from these and other observations of drug-induced SCLE? First, when making a diagnosis of SCLE, the practitioner must examine the drugs that the patient is taking to evaluate the possibility that a drug has caused or exacerbated the disease. Second, patients with "naturally-occurring" SCLE should be advised about the risk of drugs that might exacerbate their disease. However, it should be stressed that it is not known what percentage of patients with SCLE will experience an exacerbation of their disease with the appropriate use of antihypertensive agents including thiazides, angiotensin-converting enzyme inhibitors, and calcium channel blockers. Therefore, when these drugs are deemed the best choice for the patient's treatment, they should be prescribed. Finally, the exacerbation of SCLE skin lesions due to a drug, while distressing to the patient and physician, is not life threatening and usually resolves with drug treatment discontinuation and standard forms of therapy used for LE.

I would like to thank Carol L. Kulp-Shorten, MD, and J. Mark Jackson, MD, for reviewing this editorial and for their thoughtful critique.