PEARLS FROM THE POSTERS: NEW ANDNOTEWORTHY RESEARCH FINDS

P7805Erythroderma: Searching for diagnostic clues in a series of 164 patients

Denis Miyashiro, MD, Hospital das Cl�ınicas, S~ao Paulo, Brazil; Alice Nunes, MD,Hospital das Cl�ınicas, S~ao Paulo, Brazil; Jose Antonio Sanches, PhD, Hospital dasCl�ınicas, S~ao Paulo, Brazil

Background: Erythroderma is a severe syndrome with challenging etiologicdiagnosis. Despite histologic, immunophenotypic, and molecular analysis improve-ment, many cases remain as idiopathic ones.

Objectives: Analyze clinical, laboratory, and histologic findings of 164 erythrodermicpatients to find clues to the etiologic diagnosis and propose an adequate form ofinvestigation to facilitate the management of these patients. Survival data wereanalyzed according to the etiologic diagnosis.

Methods: In a retrospective study performed at the Hospital das Cl�ınicas, Universityof S~ao Paulo Medical School (2007-2012) including 164 cases of acquirederythroderma, clinical, laboratory, and histologic data were collected according toa protocol of investigation. For qualitative variables, we the used chi-square or Fisherexact tests, and for quantitative variables we used KruskaleWallis 1-way analysis ofvariance.

Results: The median age was 55.7 years, with a male:female ratio of 1.98:1. Theatopic dermatitis group developed erythroderma in an earlier age (median, 22.5years; P\.0001). Acute onset was associated with drug reactions (P ¼ .017) andpsoriasis (P ¼ .041). Psoriasis was the most frequent etiology (20.1%), followed byeczema (18.9%), drug eruption (17.1%), S�ezary syndrome (9.8%), atopic dermatitis(8.5%), mycosis fungoides (6%), and other diagnoses (5.5%), represented bypemphigus foliaceus, adult T-cell lymphoma, lichen planus, pityriasis rubra pilaris,paraneoplastic erythroderma, and HIV-associated erythroderma. In 14% of thepatients, it was not possible to elucidate the cause of erythroderma. We did notobserve a significant statistical correlation between the groups for both clinical andlaboratory findings, except for high levels of immunoglobulin E in the atopicdermatitis group. Pathologic examination was consistent with the final diagnosis in73.9% of cases. Monoclonal T cell proliferation in the skin was observed essentiallyin the mycosis fungoides and S�ezary syndrome groups. At the last assessment, 65.2%of patients were alive with evidence of disease, 26.8% were alive without disease,and 7.9% had died with disease.

Conclusions: Erythroderma is a challenging syndrome with difficult diagnosisapproach and reserved prognosis. Young age and high immunoglobulin E levelsare associated with atopic dermatitis; acute onset is observed in drug eruptions andpsoriasis. Histopathologic and molecular biology tests are important tools in theinvestigation of erythroderma.

MAY 201

cial support: None identified.

Commer

P7622A population-based study on the association between bullous pemphigoidand neurologic disease

Katherine Brick, MD, Mayo Clinic, Rochester, MN, United States; CarilynWieland, MD, Mayo Clinic, Rochester, MN, United States; Chad Weaver, MD,Mayo Clinic, Rochester, MN, United States; Christine Lohse, MS, Mayo Clinic,Rochester, MN, United States; Lawrence Gibson, MD, Mayo Clinic, Rochester,MN, United States; Mark Pittelkow, MD, Mayo Clinic, Rochester, MN, UnitedStates; Michael Camilleri, MD, Mayo Clinic, Rochester, Minnesota, United States

The association between bullous pemphigoid (BP) and multiple neurologic diseaseshas been well-documented, but population-based studies from the United States arelacking. BP antigen 1 has similar isoforms in neuronal and cutaneous tissues, whichis one possible link to explain this association. However, the etiology of thisrelationship is not well understood. A population-based study was performed toexamine this phenomenon. The Rochester Epidemiology Project was used toidentify 87 patients who were residents of Olmsted County, Minnesota at their firstlifetime diagnosis of BP between January 1, 1950 and December 31, 2009. Anadditional 3 age- and sex-matched subjects who were also residents of OlmstedCounty were used as controls for each BP case, and all charts were reviewed for thepresence of neurologic disorders (ie, dementia, Alzheimer disease, Parkinsondisease or other movement disorders, epilepsy, stroke, or multiple sclerosis) beforeor after the diagnosis of BP. Any previous diagnosis of neurologic disease wasassociated with a 6.85-fold increased odds of developing BP (odds ratio, 6.85; P\.001). A history of dementia increased the odds of developing BP 6.75-fold (oddsratio, 6.75; P ¼ .002), but in the 4 patients with Parkinson disease (n ¼ 3) or amovement disorder (n¼ 1), the increased odds were not statistically significant. Forthe cohort arm of this study, those with a history of BP were significantly more likelyto develop Parkinson disease (hazard ratio, 8.56; P ¼ .014) as well as any type ofneurologic disease (hazard ratio, 2.02; P¼.012) during follow-up. A limitation of ourstudy is the diagnosis of dementia could not be further subdivided because thediagnosis was based on retrospective chart review. Our study confirms theassociation of neurologic disorders and bullous pemphigoid, and specificallyprovides population-based evidence for the association with dementia andParkinson disease. This supports need for additional investigation into the specificmechanisms that may underlie these relationships.

cial support: None identified.

Commer

4

P7571Immunohistochemical detection of BRAF mutation in melanoma

Michelle Vernali, University of North Carolina School of Medicine, Chapel Hill,NC, United States; Dan Zedek, MD, Department of Dermatology, University ofNorth Carolina School of Medicine, Chapel Hill, NC, United States; David W.Ollila, MD, Department of Surgical Oncology, University of North Carolina Schoolof Medicine, Chapel Hill, NC, United States; Nancy E. Thomas, MD, PhD,Department of Dermatology, University of North Carolina School of Medicine,Chapel Hill, NC, United States

BRAF mutational status is a major factor in the clinical management of patients withmetastatic melanoma. Ninety percent of mutations within this gene are accountedfor by a missense substitution at amino acid position 600 (V600E), substitutingvaline for glutamic acid. This results in constitutive phosphorylation of downstreamtargets. Vemurafenib, a targeted BRAF kinase inhibitor, has demonstrated signifi-cantly increased overall and progression-free survival as compared to dacarbazine ina phase III clinical trial. BRAF mutation status, and therefore eligibility for BRAFinhibitors, is currently determined by sequencing methods. Recently, a monoclonalantibody (VE1) against the BRAF V600E mutant protein has been developed. In thisstudy, we assessed the validity of VE1 in the detection ofmutant BRAF V600E proteinas compared to the UNC CLIA-certified method of DNA pyrosequencing. Weanalyzed 96 primary and metastatic melanomas from 76 patients with knownmutational status. Primary melanomas were immunostained and results matched totheir associated metastases to determine if staining remained consistent betweenthe two lesions. The staining intensity of these specimens was assessed by adermatopathologist blinded to all clinical and genetic data and scored from 0 to 3 (0¼ no staining, 1¼weak background staining, 2¼moderately positive staining, and3 ¼ strongly positive staining). Scores of 0 and 1 were considered to be negativestaining while scores of 2 and 3 were considered to be positive staining. The VE1antibody demonstrated a sensitivity of 81% and a specificity of 100% as compared topyrosequencing. In addition, there was 100% concordance between primary andmetastatic lesions. Other BRAF mutations, including V600K, V600Q, and V600R didnot positively with the VE1 immunostain. The results of this study demonstrate theclinical utility of the VE1 immunostain in the management of melanoma patients.

cial support: None identified.

Commer

P8324Melanoma associated with TNFa inhibitors: A research on adverse drugreactions (RADAR) report

Beatrice Nardone, MD, PhD, Department of Dermatology, NorthwesternUniversity, Chicago, IL, United States; Dennis P. West, PhD, Department ofDermatology, Northwestern University, Chicago, IL, United States; Dennis W.Raisch, PhD, College of Pharmacy, University of New Mexico, Albuquerque, NM,United States; Gil Abramovici, MD, Department of Dermatology, NorthwesternUniversity, Chicago, IL, United States; Jeave Reserva, MD, Department ofDermatology, Northwestern University, Chicago, IL, United States; Josh A.Hammel, MD, Department of Dermatology, Northwestern University, Chicago,IL, United States

Introduction: Tumor necrosis factor-a inhibitors (TNFa Is), are widely used fortreatment of several dermatologic, rheumatologic, and gastrointestinal inflamma-tory disorders. Despite their therapeutic benefit, there is evidence linking theseagents with the occurrence of malignancies. For 4 out of 5 TNFa Is, the FDA labelstates that ‘‘melanoma has been reported in patients treated with these agents.’’

Methods: We searched the FDA Adverse Event Reporting System (FAERS) databasefor terms related to melanoma (M) combined with TNFa Is and we calculatedempirical Bayes geometric means (EBGM) for detection of safety signals. We alsosearched a large urban academic electronic medical record (EMR) database ([2.2million individual records), for which we calculated the relative risk (RR) ofmelanoma in subjects exposed to TNFa Is versus nonexposed subjects.

Results: Since the year of the drug approval date through December 2012, 1041reports of melanoma-associated with a TNFa I were identified in the FAERS database.A safety signal was detected for infliximab (I), n ¼ 434 (EBGM, 7.90; 95% CI, 7.13-8.6); golimumab (G), n¼ 10 (EBGM, 5.34; 95% CI 2.41-9.88); etanercept (E), n¼ 347(EBGM, 2.49; 95% CI, 2.24-2.76); and adalimumab (A), n¼ 237 (EBGM, 2.49; 95% CI,2.19-2.83). Certolizumab pegol (CP), n¼ 13 had no detectable safety signal. For theEMR cohort, 6,045were exposed to TNFa Is, and 35 cases of Mwere detected (I, n¼3; G, n¼ 1; E, n¼ 17; A, n¼ 14; CP, n¼ 0). Significant RRwas detected for A (RR, 1.8;95% CI, 1.06-3.0; P¼.02) and E (RR, 2.35; 95% CI, 1.46-3.77; P¼.0004). For TNFa Isas a class of drugs, a safety signal was detectable in FAERS database (EBGM, 3.30; 95%CI, 3.10-3.52) and RR was significant in EMR database (RR, 1.75; 95% CI, 1.25-2.43;P\.0009).

Conclusion: We identified a significant association for exposure to TNFa Is and M,for all drugs except CP in the FAERS dataset and for A and E in the academic dataset.Interestingly, the current FDA labeling for certolizumab pegol does not includemelanoma as an adverse effect. Our findings add to existing evidence linking theseagents with the occurrence of melanoma. Additional investigations are required tofurther explore this association and the risk of melanoma with TNFa Is therapy.

cial support: None identified.

Commer

J AM ACAD DERMATOL AB1