Comment
Chronic immunosuppression contributes to morbidity and mortality in patients with BP; therefore, accurate diagnosis of BP is of utmost importance.14 A meta-analysis described ELISA as a test with high sensitivity and specificity (87% and 98%–100%, respectively) for diagnosis of BP.3 Nevertheless, there are opportunities for misdiagnosis using ELISA, as demonstrated in our case. To determine if the reported sensitivity and specificity of ELISA is accurate and reliable for clinical use, individual studies from the meta-analysis were reviewed.4,5,7-10,13,15 Issues identified in our review included dissimilar diagnostic procedures and patient populations among individual studies, several reports of positive ELISA in patients without BP, and a lack of explanation for these false-positive results.
There are notable differences in diagnostic procedures and patient populations among reports that establish the sensitivity and specificity of ELISA for BP diagnosis.3-13 Studies have detected IgG that targets the NC16A domain of the BP180 kD antigen, the C-terminal of the BP180 kD antigen, or the entire ectodomain of the BP180 kD antigen. Study patient populations varied in disease activity, stage, and treatment. Control patients included healthy patients as well as those with many dermatoses, including pemphigus vulgaris, systemic scleroderma, systemic lupus erythematosus, rheumatoid arthritis, lichen planus, and discoid lupus erythematosus.3-13 Due to these differences between individual studies, we believe the results that determine the overall sensitivity and specificity of ELISA for BP diagnosis must be interpreted with caution. For ELISA statistics to be clinically applicable to a specific patient, he/she should be similar to the patients studied. Therefore, we believe each study must be evaluated individually for applicability, given the differences that exist between them.
Furthermore, there have been several reports of false-positive ELISA results in patients with other dermatologic disorders, specifically in elderly patients with pruritus who do not fulfill clinical criteria for diagnosis with BP.16-18 In a population of elderly patients with pruritus for which no specific dermatological or systemic cause was identified, Hofmann et al18 found that 12% (3/25) of patients showed IgG reactivity to BP180 despite having negative DIF results. In another study of elderly patients with pruritic dermatoses, Feliciani et al17 found that 33% (5/15) of patients had IgG reactivity against BP230 or BP180, though they did not fulfill BP criteria based on clinical presentation and showed negative DIF and IIF results. These findings suggest that IgG reactivity against BP autoantibodies as determined by ELISA is not uncommon in pruritic diseases of the elderly.
Explanations for false-positive ELISA results were rare. A few authors suggested that false-positives could be attributed to an excessively low cutoff value,7-9 which was consistent with reports that the titer of autoantibodies to BP180 correlates with disease severity, suggesting that the higher titer of antibodies correlates with more severe disease and likely more accurate diagnosis.10,19,20 It is important to consider that patients who have low titers of BP180 autoantibodies with inconsistent clinical characteristics and DIF results may not truly have BP. Furthermore, to determine the clinical value of ELISA in identifying patients in the initial phase of BP, sera of BP patients should be compared with sera of elderly patients with pruritic skin disorders because they comprise the patient population that often requires diagnosis.18
Given the issues identified in our review of the literature, the published sensitivity and specificity of ELISA for BP diagnosis are likely overstated. In conclusion, ELISA should not be relied on as a single criterion adequate for diagnosis of BP.12,21 Rather, the diagnosis of BP can be obtained with a positive predictive value of 95% when a patient meets 3 of 4 clinical criteria (ie, absence of atrophic scars, absence of head and neck involvement, absence of mucosal involvement, and older than 70 years) and demonstrates linear deposits of predominantly IgG and/or C3 along the basement membrane zone of a perilesional biopsy on DIF.15 The gold standard for diagnosis of BP remains clinical presentation along with DIF, which can be supported by histology, IIF, and ELISA.22