We have carefully read the excellent review by Hernandez et al.,1 with regard to skin lesions that occur during treatment with antagonist of tumor necrosis factor (anti-TNF), and we would like to make some additional comments with respect to cutaneous lupus erythematosus (LE) induced by such drugs.

As the authors report, the development of autoantibodies is a frequent event in patients receiving anti-TNF drugs,1 with an estimated prevalence of ANA positivity ranging from 25% to 80% and anti-DNA ranging from 5% to 15%.2 However, as they state, the appearance of LE is quite rare.1 Postmarketing studies estimate the incidence of induced LE at 0.19%–0.22% for infliximab, 0.18% for etanercept and 0.10% for adalimumab.2 The slightly higher frequency of LE induced with infliximab and etanercept may simply reflect more years of exposure of patients compared with adalimumab. In connection with more recently introduced anti-TNF agents, certolizumab and golimumab, a case of induced LE has been described with the first3 and one subacute cutaneous LE exacerbation has been related to the second.4 Considering the high prevalence of autoantibodies and the large number of patients treated, one would expect a higher frequency of induced LE. One probable explanation for this discrepancy is that the type of autoimmune response induced by anti-TNF agents is mainly restricted to nonpathogenic IgM or IgA isotypes, and although the main reactivity is anti-DNA, it is rare to develop other LE related antibodies, such as anti-ENA or hypocomplementemia. In addition, the titles of anti-DNA IgM tend to fluctuate over time and disappear quickly after removal of the drug.5

Identified risk factors for the development of LE during anti-TNF treatment are advanced age and the presence of increased baseline anti-DNA, but not of ANA.6 Another factor that could influence this is the underlying disease. Although the production of autoantibodies is similar among the different diseases treated with these agents, most cases have been described in Ra patients, as evidenced by a review of Costa et al.,7 who found that of 33 published cases of induced LE due to anti-TNF drugs, 76% of patients had RA. The frequency with which these cases appear in the literature contrasts to those described in RA clinical trials with long-term follow up, so it should be noted that these cases are generally based on retrospective observations that often lack serological data prior to starting anti-TNF therapy and there may be some overlap of RA and LE before treatment.5

LE cases induced by anti-TNF comply with 4 or more ACR classification criteria in 40%, 3 criteria in 21%, and 2 or less in 39%.2 Up to 67% of cases have cutaneous manifestations,8 corresponding generally to maculopapular, pruritic erythematous rash affecting photosensitive areas, as mentioned by the authors,1 however, the spectrum is much broader. Both LE-specific lesions (cutaneous acute, subacute and discoid), and other nonspecific findings including urticarial lesions, scarring, alopecia and purpura may occur.8 Within difficult to classify cutaneous LE lesions, there has also been published cases of LE tumidus and lupus perniosis (LP) induced by anti-TNF. LE tumidus is characterized by the appearance of papules on exposed areas, erythematous plaques or nodules without other associated epidermal changes; one of the cases found in the literature occurred with infliximab and adalimumab9 in another,10 both in RA patients. Our group conducted a review of 5 cases of LP associated with anti-TNF,11 a rare form of cutaneous LE characterized by papules or plaques with erythematous violaceous acral distribution that simulate ischemic injury. Four of these cases occurred in patients with RA and one in ankylosing spondylitis.

In summary, although induced LE is a rare adverse event seen during anti-TNF treatment, it is important to have in mind because of its varied clinical expression, especially on the skin, and to identify those cases that actually are due to this entity, given the trend that may lead to overdiagnosis.