Serum pemphigoid antibodies in 76 patients: ocular and other sites compared

Published: 27-07-2020| Version 1 | DOI: 10.17632/7pxbkx84r3.1
John Dart,
Jane Setterfield,
Richard Groves,
John Mee,
Gilles Diercks,
Hendri Pas,
Darwin Minassian


Our data was collected to evaluate the sensitivity, specificity, and differences compared to controls, of a panel of tests and their combinations, for serum pemphigoid autoantibodies in 76 mucous membrane pemphigoid (MMP) patients. This was done to compare those with or without a positive direct immunofluorescence (DIF+) result to test our hypothesis that a panel of serum pemphigoid autoantibody tests, and their combinations, might be used to confirm an immunopathological diagnosis of MMP in DIF negative patients with ocular involvement. The hypothesis was tested by evaluating the sensitivity and specificity of tests and their combinations for cases with that of 45 age, sex, race matched controls. The MMP patients had a clinical diagnosis of MMP with or without a positive DIF using defined criteria in a data collection tool. The number of controls was chosen a priori to give an 80% power to detect a difference in the proportions of BP180-NC16a autoantibodies. This was calculated using data from Wieland CN et al. [Wieland CN, Comfere NI, Gibson LE, et al. Anti-bullous pemphigoid 180 and 230 antibodies in a sample of unaffected subjects. Arch Dermatol 2010;146(1):21-5] on age, sex stratified controls having detectable levels of these in 14/337 (4.15%), and our pilot data in our MMP cases in which 8/32 (25%) had detectable levels. Seventy-six MMP patients were prospectively phenotyped using a data collection tool and matched with 45 healthy age sex matched controls. All patients and controls had serum collected and this was tested for the presence of pemphigoid serum autoantibodies using a panel of 5 tests. The sensitivity and specificity of autoantibody detection, Youden’s index, and significant differences, compared to controls, for individual tests and test combinations were evaluated.


Steps to reproduce

This file includes in Spreadsheet 1 “Patient and Control dataset” clinical and serology phenotyping data for 76 mucous membrane pemphigoid patients and 45 matched controls from a prospective cross-sectional study. The serology results used for inclusion and analysis in this dataset were derived from the results of repeating the tests in three laboratories for which the results are shown in in Spreadsheet 2 “Serology results”. The Service laboratory results were retested when we found that the control results could not be reproduced at St John’s laboratory. Following this all the tests were repeated at the Groningen laboratory which identified 51/304 tests (16.8%) that were different between the two labs. The tests used for the analysis (Spreadsheet 1) were chosen from these three datasets using the following plan. The Groningen results were employed for the analysis with 6 exceptions only using the following protocol: the discrepancies for the results of tests between the Groningen and Service laboratories were retested, masked as to the results from the other laboratories, by St Johns; 45/51 (88.2%) of the Groningen result discrepancies vs Service laboratory results were confirmed as correct by retesting at St John’s and were used for the analysis. For the 6/51 results where the Service laboratory findings concurred with the results of repeat testing at St John’s these were used for the analysis instead of the Groningen results. Laminin 332 reactivity was tested using the Groningen keratinocyte footprint assay (KFA)33 results. The Service laboratory results could not be confirmed for 45/304 (14.8%).