Adverse Events
Dermatologic AEs are among the most common with nivolumab treatment. In a pooled retrospective analysis of melanoma patients, Weber et al9 found that 34% of 576 patients experience cutaneous any-grade AEs associated with nivolumab treatment, most commonly pruritus. It has been well documented that anti–PD-1 therapy AEs of the skin as well as other organ systems have a delayed onset of at least 1 month.9 The average time of onset for bullous eruptions associated with anti–PD-1 therapy has been reported to be approximately 12 weeks, with a range of 7 to 16.1 weeks.11 Our patient had a bullous eruption with an onset of 12 weeks following initiation of treatment.
Although lichenoid reactions appear to be relatively common AEs of anti–PD-1 therapy,2,5,6 only a small number of cases of bullous pemphigoid eruptions have been reported.7 It has been hypothesized that blockade of the PD-1/PD-L1 pathway increases production of hemidesmosomal protein BP180 autoantibody, which is involved in the pathogenesis of LPP.7 Bullous eruptions have not been reported in the use of anticytotoxic T-lymphocyte–associated protein 4 agents, which could indicate that such eruptions are specific to the anti–PD-1 class of drugs.7
Diagnosis
Our patient represents a rare drug reaction involving both lichenoid and bullous components. Our differential diagnosis included drug-induced bullous lichen planus (BLP) and drug-induced LPP. Differentiation of these diagnoses can be difficult. In fact, in 2017 Fujii et al12 found reason to reprise the hypothesis that BLP is a transitional step toward LPP. The histologic evaluation of LPP differs depending on the type of lesion biopsied and can be indistinguishable from BLP as well as bullous pemphigoid. Therefore, clinical history and immunofluorescence should be used to make a diagnosis. Lichen planus pemphigoides typically will have linear IgG deposition along the basement membrane zone on both DIF and IIF, findings that will be negative in patients with BLP.13 Direct immunofluorescence findings in BLP include shaggy deposits of fibrin along the basement membrane zone. In this patient, DIF was negative, which may have been caused by variability among lesions in LPP, but IIF was positive. Given the clinicopathologic correlation, the diagnosis of LPP was made. Further studies, such as immunoblot and enzyme-linked immunosorbent assay, also can be used to aid diagnosis.
A similar presentation has been documented in a patient with metastatic melanoma.14 The diagnosis in this patient was LPP induced by pembrolizumab, which is another agent within the anti–PD-1 class. The Naranjo probability scale scored our patient’s eruption as a possible adverse drug reaction.15 Thus, other etiologies, such as a paraneoplastic process, cannot be completely ruled out. However, our patient has not had recurrence after 1 year, and the timing of the eruption appeared to be related to drug therapy, making alternative etiologies less likely.
Management
Cessation of nivolumab therapy and a short course of oral corticosteroid therapy led to marked improvement of symptoms. Given the emergent treatment of our patient, the resolution of her symptoms cannot be solely attributed to the cessation of nivolumab or to treatment with prednisone. Oral rather than topical corticosteroids were chosen because of the severity of the eruption. Topical corticosteroids and oral antihistamines can provide relief in less severe cases of bullous reactions to anti–PD-1 therapy.7,11 This regimen also has proven to be effective in lichenoid dermatitis induced by anti–PD-1.2
Conclusion
We hope this case report will contribute to the growing body of evidence regarding recognition and management of unique reactions to cancer immunotherapies.