International Bullous Diseases Group: Consensus on Diagnostic Criteria for Epidermolysis Bullosa Acquisita

C. Prost-Squarcioni; F. Caux; E. Schmidt; M.F. Jonkman; S. Vassileva; S.C. Kim; P. Iranzo; M. Daneshpazhooh; J. Terra; J. Bauer; J. Fairley; R. Hall; M. Hertl; J.S. Lehman; B. Marinovic; A. Patsatsi; D. Zillikens; V. Werth; D.T. Woodley; D.F. Murrell


The British Journal of Dermatology. 2018;179(1):30-41. 

In This Article


Altogether, 38 experts in EBA from 15 countries took part in this consensus, although not all were able to attend all three international meetings held in San Francisco, Rotterdam and Copenhagen. The actual results of the voting questions (Table S7; see Supporting Information) and the reported anonymized results in the final version of the approved wording or procedures were from 22 blistering disease experts from 14 different countries (Figure S1; see Supporting Information). Although not all members of the IBDG were present at the final voting meeting, all co–authors were in agreement with the EBA diagnostic criteria detailed herein.

Definition and Clinical Variants of Epidermolysis Bullosa Acquisita

EBA is defined as a subepithelial AIBD in which patients have tissue–bound autoAbs targeted against Col7 within AFs of the basement membrane zone (BMZ) of the DEJ or chorioepithelial junction (CEJ) in stratified squamous epithelia.

Several forms of EBA exist, with two major types, termed: (i) the classical/mechanobullous form; and (ii) the nonclassical/nonmechanobullous forms, first described by Gammon et al. in 1982.[18] The latter includes forms that may clinically resemble BP, MMP, Brunsting–Perry pemphigoid and linear IgA bullous dermatosis.[6,18–35]Table 1 and Figure 1 summarize the descriptions and definitions of these four phenotypes. The frequency of these subtypes varies in different countries, with the classical type being the most common in European reports and the BP–like form more common in Asia (Table S2).[36–39] It should be recognized that overlapping clinical presentations may also occur (proposal 09, cases 1–5 in Appendix S2). Overall, because EBA can appear clinically, histologically and immunologically similar to BP,[18,20–24] when the initial diagnostic consideration of the dermatologist is BP, EBA should also be considered. Furthermore, if certain clinical clues are somewhat atypical for BP (e.g. lesions that heal with scarring and the formation of milia, head and neck involvement, mucosal involvement, disease onset before 70 years of age),[40] the diagnostic possibility that the patient has EBA, rather than BP, rises considerably. Finally, it is important to classify cases of MM–predominant EBA (MM–EBA) because the severity of mucosal involvement dictates more aggressive, multidisciplinary management.

Figure 1.

Clinical subtypes of epidermolysis bullosa acquisita (EBA). (a–g) Classical/mechanobullous form of EBA: trauma–induced lesions are usually located on extensor skin surfaces, i.e. (a, b) back of the hands, (c) feet, (d) heel, (e) elbows and (f) knees. (b, d) Bullous/vesicular lesions or erosions are surrounded by noninflamed or scarred skin. Lesions heal with (e, f) scarring and (a, b, f) milia formation. (c) Nail dystrophy and (g) scarring alopecia are possible. (h–j) Bullous pemphigoid (BP)–like form of EBA: tense bullae on (h) erythematous or (i, j) urticarial skin suggestive of BP, in atypical locations for BP at extensor areas of the limbs. (k–p) Mucous membrane (MM)–EBA: all the MM lined by squamous epithelium may be involved in (k) buccal MM (particularly the tongue), (l) anus and (m) oesophagus, leading to strictures and (n) gastrostomy, (o) conjunctival scarring and (p) tracheal stenosis. (q) IgA–EBA with bullous eruption in a 'string of pearls'. (r) Brunsting–Perry–type EBA with recurrent blistering dermatosis confined to the head for 10 years. Images (c), (m) and (o) are reproduced from Figures 61·1(b), 61·1(c) and 61·1(d), and image (h) is reproduced from Figure 40·2(a) in Prost–Squarcioni C and Caux F. Management of epidermolysis bullosa acquisita. In: Blistering Disease: Clinical Features, Pathogenesis, Treatment (Murrell D, ed.). New York: Springer, 2015, with the permission of Springer. Image (g) is reproduced from Supplementary Figure 1(a) in Zumelzu C, Le Roux–Villet C, Loiseau P et al. Black patients of African descent and HLA–DRB1*15:03 frequency overrepresented in epidermolysis bullosa acquisita. J Invest Dermatol 2011; 131:2386–93, with the permission of the Journal of Investigative Dermatology. Image (p) courtesy of Professor Michel Brauner.

There has been confusion in the literature as to which forms constitute an 'inflammatory' form of EBA.[13,20,22,23,36–39] The IBDG members agreed that the BP–like form is inflammatory but that both MM–EBA and IgA–EBA may also be inflammatory. However, Brunsting–Perry–like EBA is usually a noninflammatory form of EBA.

Laboratory Testing for Epidermolysis Bullosa Acquisita

The IBDG reached agreement on the following proposal: routine histopathology, direct IF (DIF) microscopy and indirect IF (IIF) – which are widely available laboratory tests – allow a diagnosis of subepithelial AIBD but are not able to distinguish EBA from another subepidermal AIBD. Routine histopathology on a biopsy obtained from lesional skin or MM of a patient with EBA shows (i) subepidermal or subepithelial cleavage; (ii) great variability in the amount or type of inflammatory infiltrate; and (iii) milia cysts and fibrosis in older lesions. Routine DIF microscopy of perilesional skin or MM shows (i) linear immune deposits along the BMZ of the DEJ or CEJ, and (ii) no labelling of dermal blood vessels. The profile of immune deposits includes IgG and C3 but occasionally IgA or IgM (Table S3, case report 5; see Supporting Information).[41–43] Routine IIF microscopy on monkey, rat or rabbit oesophagus or human skin can detect anti–BMZ autoAbs, but it is often at a low titre.

Currently, a diagnosis of EBA should be confirmed by at least one of the following tests, which are only performed in academic centres and are not widely available to the average dermatologist worldwide:[13,15,35] (i) among tests requiring skin or MM biopsies – electron microscopy (EM) and direct IEM,[44] serration pattern analysis by DIF[41,45,46] and fluorescent overlay antigen mapping (FOAM);[47–49] and (ii) among serological tests for the detection of circulating autoAbs to Col7 – enzyme–linked immunosorbent assay (ELISA), IIF on BIOCHIP™ with noncollagenous 1 (NC1) Col7–transfected human cells,[50] immunoblotting (IB), IIF on skin deficient in Col7[51,52] and/or indirect IEM.[44] Obviously, all these serological tests require that the patient with EBA has autoAbs to Col7 circulating in their blood and it must be kept in mind that anti–Col7 autoAbs are also present in bullous systemic lupus erythematosus (BSLE).[53]

Alternative laboratory tests when none of these tests is available include DIF and IIF on salt–split skin (SSS),[54] but they will not absolutely confirm the diagnosis of EBA.[55,56]

Standard transmission electron microscopy and immunoelectron microscopy. In patients with EBA, standard transmission EM shows an electron–dense band immediately below the lamina densa (LD) in the AF zone. Another EM finding suggestive of EBA is that the cleavage occurs below the LD, which remains attached to the roof of the blister. A paucity of AFs also supports a diagnosis of EBA.[44]

Direct IEM on perilesional skin shows in vivo bound immune deposits that are very thick and located in the AF zone, and a cleavage under immune deposits, if EBA is present (Figure 2).[44] The IBDG agreed on the limitations of direct IEM (proposal 29). Sixteen members of the consensus group attending the EADV round of voting had EM available at their site, but only about half of the group had experience in using EM and only seven were currently using IEM.

Figure 2.

Immunoelectron microscopy (IEM) in epidermolysis bullosa acquisita. (a) Direct IEM using the pre–embedding immunoperoxidase technique. Thick immune deposits are observed in the anchoring fibril zone below the lamina densa (LD) and split (asterisk) under them. (b, c) Indirect IEM using (b) pre–embedding immunoperoxidase and (c) the immunogold technique. Immune deposits (arrows) decorate the ends of anchoring fibrils. Images (a), (b) and (c) are reproduced from Figures 19·10(b), 19·12(a) and 19·12(b) in Prost–Squarcioni C. Electron microscopy and immunoelectron microscopy. In: Blistering Disease: Clinical Features, Pathogenesis, Treatment (Murrell D, ed.). New York: Springer, 2015, with the permission of Springer.

Serration pattern analysis. DIF microscopy of a perilesional skin biopsy can distinguish EBA from other subepithelial AIBD by showing a 'u–serrated' linear pattern of immunoglobulin deposits along the BMZ in EBA and BSLE, and an 'n–serrated' pattern of immunoglobulin deposits in BP, antilaminin 332 MMP and anti–p200/laminin γ1 pemphigoid (Figure 3).[41,45,46,55,56] No agreement was obtained on limitations of the serration pattern analysis (proposal 26). Indeed, it can be performed with routine DIF microscopy (new, unpublished data on the exact requirements can be read in Supplementary Data S2 of Appendix S2). However, to date, this test is not widely available and so far only the teams in Groningen and Lübeck have been able to master the technique (proposal 26bis).

Figure 3.

Serration pattern analysis by direct immunofluorescence. The u–serrated pattern is characterized by closed arches at the bottom appearing like 'growing grass', whereas the n–serrated pattern shows closed arches at the top. (a) Direct immunofluorescence (DIF) of epidermolysis bullosa acquisita skin showing linear IgG deposition along the epidermal BMZ in a u–serrated pattern (× 400). (b) DIF of bullous pemphigoid skin showing linear IgG deposition along the epidermal basement membrane zone in an n–serrated pattern (× 400). Images courtesy of Dr Gilles Diercks.

Fluorescent overlay antigen mapping analysis. In EBA, FOAM, using routine IF microscopy with either an image analysis system or laser scanning confocal microscopy,[47–49] shows in vivo bound immune deposits below the basal keratinocyte membrane, lamina lucida and LD components (Figure 4).

Figure 4.

Fluorescence overlay antigen mapping (FOAM) in skin biopsies. In epidermolysis bullosa acquisita (EBA), in vivo bound immune deposits are below the α6β4 integrin of the basal keratinocyte membrane and below the components of the lamina lucida and the lamina densa (laminin 332 and type IV collagen). (a) In a patient with EBA, in vivo bound IgG (green) is below type IV collagen (red). (b) In a patient with a bullous pemphigoid, in vivo bound IgG (green) is above laminin 332 (red). (c) In a patient with mucous membrane pemphigoid (MMP), in vivo bound IgG (green) is below or co–localized (yellow) with laminin 332 (red). (d) In a patient with MMP, in vivo bound IgG (green) is above or co–localized (yellow) with type IV collagen (red). Images courtesy of Dr Katarzyna Wozdniak.

Enzyme–linked immunosorbent assay. Commercially available ELISAs using recombinant NC1/noncollagenous 2 (NC2) Col7,[57–60] or NC1 Col7,[50] are widely available. Sensitivities vary depending on the selection criteria. It is very high (79–96·7%) on preselected positive sera by IIF on SSS, with floor labelling (Table S4; see Supporting Information). The sensitivity of ELISA to NC1/NC2 is lower (30–54%) in studies on unselected EBA sera.[39,61]

An ELISA using full–length Col7 is more sensitive than an ELISA using NC1/NC2 Col7 but is not commercially available (Table S5; see Supporting Information).[39,42,62,63]

ELISAs for Col7 are not highly specific as they may be positive in patients with Crohn disease or ulcerative colitis without cutaneous manifestations of EBA,[64] atypical AIBD[65] and patients with recessive dystrophic EB (RDEB) (Tables S5 and S6).[39,63,66,67] Of note, the presence of circulating autoAbs against BP180, laminin 332 and the p200/laminin γ1 chain (detected by IB or ELISA), which may occur as a result of the epitope spreading phenomenon, does not rule out a diagnosis of EBA (supplementary cases 2 and 4 in Appendix S2).[37,68–75]

Novel technique: indirect immunofluorescence using noncollagenous 1 type VII collagen–transfected cells. Patient serum autoAbs could label NC1 Col7–transfected cells on a special slide, a so–called BIOCHIP™ method (Figure 5). This test is now commercially available and could be used as a substitute for the ELISA outlined above. It is a sensitive, specific and rapid assay for testing preselected positive sera by IIF on SSS with floor labelling.[50] Like serological testing by ELISA, its sensitivity is lower in studies on unselected EBA sera.[39]

Figure 5.

Indirect immunofluorescence microscopy with noncollagenous 1 type VII collagen (NC1 Col7)–expressing human cells by BIOCHIP™ technology. Patient serum autoantibodies could label molecularly engineered epidermal cells that express human NC1 Col7 on a special slide. (a) On a standard–sized slide, there are five incubation fields each with two different BIOCHIPs: one with human embryonic kidney (HEK)293 cells transfected with pTriEx–1, which serve as negative control, and one with human HEK293 cells transfected with NC1 Col7. (b) Autoantibodies in the serum of a patient with epidermolysis bullosa acquisita labelled NC1 Col7–expressing HEK293 cells (right) but not non–NC1 Col7–expressing cells (left). (c) No reactivity of either NC1 Col7–expressing or non–NC1–Col7–expressing cells is seen with normal human serum. Images courtesy of Dr Aucouturier, Paris.

Immunoblotting. IB is a serological technique performed on extracts of tissues or cells or recombinant Col7 proteins (Figure 6).[13,15,35] IB is effective for diagnosing EBA by detecting autoAbs in patient sera that label the Col7 α–chain. IB can substitute for the commercially available ELISA. The IBDG agreed on the limitations of IB (proposal 33).

Figure 6.

Immunoblotting in epidermolysis bullosa acquisita (EBA). By immunoblotting with dermal extract, EBA serum recognized a band at 290 kDa, which is the alpha chain of type VII collagen (Col7), different from laminin γ1/p200 protein. A second band at 145 kDa, which is the amino terminal noncollagenous 1 domain of the alpha chain of Col7, could be seen. NHS, normal human serum.

Indirect immunofluorescence on type VII collagen–deficient human skin and indirect immunoelectron microscopy. A definitive diagnosis of EBA can be demonstrated by IIF when presumptive EBA sera label the DEJ of normal human skin but do not label skin from generalized severe RDEB with no Col7 (Figure 7).[51,52] Obviously, this is not offered in commercial laboratories and is limited by access to skin specimens from patients with this exceedingly rare disease.

Figure 7.

Indirect immunofluorescence (IIF) microscopy on type VII collagen (Col7)–deficient skin vs. normal skin. (a) IIF microscopy with epidermolysis bullosa acquisita (EBA) serum on Col7–containing normal human skin showing positive IgG binding along the epidermal basement membrane zone (BMZ) (× 400). (b) IIF microscopy with EBA serum on Col7 knockout skin from a patient with severe generalized recessive dystrophic epidermolysis bullosa showing negative IgG binding along the epidermal BMZ (× 400). Images courtesy of Dr Hendri Pas, Groningen, the Netherlands.

A definitive diagnosis of EBA can also be demonstrated when presumptive EBA sera label AFs by indirect IEM (Figure 2).[44]

Direct and indirect immunofluorescence microscopy on salt–split skin. DIF and IIF on SSS are alternative laboratory tests that only give a probable diagnosis of EBA. DIF is performed after splitting the skin biopsy of the patient using NaCl 1 mol L−1 (Supplementary Data S3 in Appendix S2).[54,76] It is not possible with MM biopsies. IIF may be performed on either normal human SSS using a similar procedure,[19,77–79] or monkey SSS, which is commercially available. Immune deposits in patients with EBA remain on the dermal floor of the separation, whereas BP immune deposits remain with the epidermal roof (Figure 8). However, this dermal labelling is not specific to EBA. It is also seen in antilaminin 332 MMP and anti1–p200/laminin γ1 pemphigoid (Figure 4). Additional tests are necessary to exclude reactivity against these molecules and finally diagnose EBA. Overall, IIF on SSS is more sensitive than IIF on monkey or rat oesophagus or unsplit human skin for detecting anti–BMZ autoAbs.[77,79]

Figure 8.

Indirect immunofluorescence microscopy on salt–split skin. An artificial cleavage is induced by NaCl 1 mol L−1 in (a, b) monkey skin (counterstained with Evans blue; images Courtesy of Dr Aucouturier, Paris) or (c, d) normal human skin. (a, c) Labelling of the floor of the cleavage by the serum of a patient with EBA. (b, d) Labelling of the roof of the cleavage by the serum of a patient with bullous pemphigoid.

Lastly, the IBDG updated consensus criteria for EBA diagnosis (2015) include combinations of the following tests: (1) a bullous disorder within the defined clinical spectrum; (2) histopathology revealing a subepidermal or subepithelial blister; (3) a positive DIF microscopy of perilesional skin or MM with linear IgG, C3, IgA and/or IgM deposits within the epithelial BMZ; (4) detection of circulating autoAbs against Col7 by IB, ELISA and/or IIF microscopy on Col7–expressing human cells; (5) labelling AFs by indirect IEM or negative IIF microscopy on Col7–deficient skin; (6) a 'u–serration' pattern by DIF microscopy; (7) direct IEM of perilesional skin demonstrating immune deposits within AFs zone ± the lower LD; (8) in vivo bound immune deposits below type IV collagen by FOAM; (9) alternatively to items (4)–(8), dermal labelling by DIF and/or IIF on SSS.

The IBDG did reach agreement that the ideal scenario is for a patient with putative EBA disease (criterium 1) to exhibit a subepidermal bulla by histology (criterium 2 optional), a positive DIF microscopy (criterium 3) and an ELISA (or another serological test) showing that the patient's serum autoAbs target Col7 [criterium (4) or (5)]. In this scenario, a highly probable diagnosis of EBA is made and no further tests need to be done for confirmation. Unlike this ideal scenario for the diagnosis of EBA, the problem that often arises is when a patient with EBA lacks circulating autoAbs and therefore IIF, SSS IIF and ELISA are negative.[39,61] Then the diagnosis of EBA could be considered definitive if criteria (1) and (3) and at least one of criteria (6)–(8) are present [criterium (2) is optional]. Lastly, if tests (6)–(8) cannot be done, a diagnosis of EBA is possible if items criteria (1), (3) and (9) are present; then diagnosis has to be confirmed by exclusion of autoimmunity against laminin 332 or the p200/laminin γ1 chain. The consensus conference was not able to determine one procedure that would be applicable worldwide. There was an animated discussion about what should be considered routine (Proposal 18). Which test(s) that the practitioner chooses will likely be determined by the clinical presentation of the patient (classical/mechanobullous type or not), the geography of the practitioner and which test is most logistically accessible. Figure 9 summarizes the different diagnostic investigative pathways.

Figure 9.

Flow chart for diagnosis of epidermolysis bullosa acquisita (EBA). Type VII collagen (Col7)–specific enzyme–linked immunosorbent assays (ELISAs) include noncollagenous (NC)1, NC1/NC2 or full–length Col7 ELISAs. BIOCHIP™ technology uses NC1 Col7–transfected human embryonic kidney (HEK)293 cells as a substrate for indirect immunofluorescence (IIF) microscopy (IIFT test). Col7–deficient skin is obtained from patients with generalized severe recessive dystrophic epidermolysis bullosa. DIF, direct immunofluorescence; DEJ, dermoepidermal junction; CEJ, chorioepithelial junction; AIBD, autoimmune blistering disease; autoAb, autoantibody; Col7, type VII collagen; WB, Western blotting; MM, mucous membranes; IEM, immunoelectron microscopy; FOAM, fluorescence overlay antigen mapping; Col4, type IV collagen; LL, lamina lucida; LD, lamina densa; BP, bullous pemphigoid; MMP, mucous membrane pemphigoid; SSS, salt–split skin (artificial cleavage obtained by incubation of normal monkey or human skin with NaCl 1 mol L−1). aNC1 ELISA, NC1/NC2 ELISA, NC1 Col7–transfected HEK293 cells and monkey SSS are commercially available; bnegativity on Col7–deficient skin is significant if IIF microscopy on normal human skin is positive; cinclude WB with dermal extract or recombinant C–terminus laminin γ1 or IIF on laminin 332–deficient skin.

In conclusion, this consensus of the criteria for the diagnosis of EBA provides a general framework for establishing a diagnosis of EBA and takes into account the clinical presentation and available laboratory testing. These criteria should help prevent clinicians from misdiagnosing other AIBDs and missing the diagnosis of EBA. They will also be useful for future studies designed to define the natural history and therapeutic outcomes of EBA.