Assessment of the Cutaneous Direct Immunofluorescence Assay (2024)

Key Points

Question What is the frequency of immunoreactant positivity in direct immunofluorescence specimens associated with skin biopsies?

Findings In this quality improvement study spanning 8 years, including 2050 direct immunofluorescence biopsies, immunoglobulin G, immunoglobulin A, and C3 antibodies identified all primary immunobullous disease cases, and immunoglobulin A, C3, and fibrin antibodies identified all vasculitis cases. A total of 247 samples were submitted for clinical diagnoses not optimally supported by direct immunofluorescence.

Meaning The findings of this study suggest that reflexive use of a 6-antibody direct immunofluorescence panel is likely not necessary to diagnose diseases associated with positive findings; pathologists may tailor antibodies used, and clinicians their decision to use direct immunofluorescence, to the suspected clinical diagnosis, which may result in more cost-conscious care.

Importance Dermatologists submit direct immunofluorescence (DIF) biopsies on a daily basis, using an assay detecting immunoreactant deposition with a panel that has traditionally comprised immunoglobulin (Ig) G, IgA, IgM, C3, and fibrin, with or without albumin antibodies.

Objectives To evaluate and compare the frequency of immunoreactants in DIF biopsies submitted over an 8-year period and assess use by dermatologists based on clinical impression.

Design, Setting, and Participants A quality improvement study was conducted in a community outreach reference laboratory associated with a large academic medical center. Results of 2050 consecutive DIF skin biopsies submitted to the laboratory between April 1, 2012, and June 12, 2020, were analyzed by final pathologic diagnosis and antibody subtype positivity, in comparison with clinical impression. Biopsies in which the submitting physician had not performed the biopsy were excluded.

Main Outcomes and Measures Histopathologic findings and the results of DIF biopsies using the standard 6-antibody panel were evaluated in correlation with the submitted clinical diagnosis to assess immunoreactivity of the assay.

Results Of 2050 DIF biopsies submitted, 367 (17.9%) were positive; IgG, IgA, and C3 alone identified all primary immunobullous disease cases (pemphigoid, pemphigus, linear IgA, and dermatitis herpetiformis), and IgA, C3, and fibrin antibodies alone identified all vasculitis cases. A panel of IgG, IgA, IgM, and fibrin identified all cases of lupus erythematosus. DIF results were positive in less than half of cases of hematoxylin and eosin biopsy-confirmed lupus erythematosus (23 of 47 [49%]). A total of 247 biopsies were submitted for clinical diagnoses not optimally supported on DIF: lichen planus, porphyria, and connective tissue disease.

Conclusions and Relevance The findings of this study suggest that there is a knowledge gap among dermatologists relating to the opportunity for high-value, cost-conscious use of DIF. The practice of reflexive antibody testing using a 6-antibody panel for all DIF biopsies is likely unnecessary. DIF protocols tailored to the clinical diagnosis may enhance cost-effectiveness without loss of test sensitivity or specificity.

Immunofluorescence (IF) identifies antigen-bound antibodies in tissue or body fluids. First applied in the 1940s to detect tissue microorganisms,¹ assays were subsequently developed to diagnose autoimmune diseases, initially in the kidney and subsequently in skin.² There are 2 types of IF assays used in dermatology: direct IF (DIF) and indirect IF. In the more routinely used DIF assay, fluorochrome-labeled antibodies are used to detect the presence of immunoreactant deposition in the patient’s skin. Historically, a panel of antibodies to immunoglobulin (Ig) G, IgA, IgM, complement C3, fibrin, and albumin has been used. This panel allowed for characterization of cutaneous vesicobullous, autoimmune, and vasculopathic diseases, improving our understanding of the pathogenesis of these disorders.³

DIF is routinely used to diagnose primary immunobullous diseases (pemphigoid group, epidermolysis bullosa acquisita, pemphigus group, linear IgA bullous disease, and dermatitis herpetiformis). Determination of pemphigoid and pemphigus subtypes can be further delineated by a combination of findings from the analysis of tissue hematoxylin and eosin (H&E) staining and clinical presentation, as well as the use of indirect IF and/or enzyme-linked immunoassay if necessary.⁴ DIF can be useful in the workup of patients with suspected vasculitis, primarily to evaluate for IgA vasculitis in cases of suspected IgA vasculitis with systemic involvement (Henoch-Schönlein purpura). DIF has also been used in the diagnosis of lichen planus, porphyria, and connective tissue disease (CTD). In evaluation of patients with suspected CTD, the immunoreactant reported most frequently in patients with lupus erythematosus (LE) is continuous granular IgM at the basem*nt membrane zone, often also accompanied by continuous granular C3 deposition.^5-10 In addition, granular deposition of IgM and C3, with or without IgG, IgA, and fibrin, can be seen in sun-exposed skin samples from approximately 20% of healthy individuals as a nonspecific finding^11-13 and in other CTDs, including dermatomyositis.^5,14

There is variability in specific DIF protocols between laboratories, with some having already discontinued the use of the anti-albumin antibody, yet many diagnostic laboratories still use the 6-panel antibody protocol on all DIF biopsies. To better understand the clinical utility and diagnostic value of performing DIF using the full antibody panel, we compared the frequency of positivity for each immunoreactant associated with submitted diagnosis in DIF biopsies submitted to the Massachusetts General Physician Organization (MGPO) Dermatopathology Associates Laboratory over 8 years.

Results of all specimens (N = 2220) submitted for DIF between April 1, 2012, and June 12, 2020, at MGPO Dermatopathology Associates, Newton, Massachusetts, were reviewed. Cases in which the referring physician had not performed the biopsy (170 cases [7.7%]) were excluded because these samples were not from the clinician providing direct care to the patient, bringing the total number of specimens to 2050. Cases were submitted by both dermatologists (1596 [77.9%]) and nondermatologists (454 [22.1%]). The study was reviewed by the Mass General Brigham Institutional Review Board and determined to be exempt research and did not require informed consent. This study followed the Standards for Quality Improvement Reporting Excellence (SQUIRE) reporting guideline.

All specimens were received in Michel’s Transport Medium and processed for DIF according to standard protocol, using fluorescein isothiocyanate–conjugated anti-IgG, anti-IgA, anti-IgM, anti-C3, anti-fibrinogen, and anti-albumin (all polyclonal ready to use; Ventana Medical Systems). An H&E section was also prepared from each DIF biopsy. All biopsies were evaluated by board-certified dermatopathologists with training and experience in interpreting DIF biopsies (J.D.R.R. and T.D.H.). Equivocal results were reviewed by a second dermatopathologist. Cases with both H&E biopsy and DIF biopsy were interpreted by the same dermatopathologist.

Data from each specimen were compiled into a spreadsheet comparing clinical impression with the final pathologic diagnosis, as well as the pattern and specific antibodies positive in each case. Clinical impression categories were grouped as follows: (1) pemphigoid group (bullous pemphigoid, pemphigoid gestationis, and cicatricial pemphigoid), (2) dermatitis herpetiformis, (3) pemphigus group (pemphigus vulgaris, pemphigus foliaceus, IgA pemphigus, paraneoplastic pemphigus, and pemphigus vegetans), (4) linear IgA disease, (5) immunobullous not otherwise specified (NOS), (6) CTD, (7) lichen planus, (8) porphyria, and (9) vasculitis. We used Excel version 2108 (Microsoft) for spreadsheet calculation. Sensitivity and specificity for the proposed antibody panels were performed according to standard binary diagnostic calculation, with 95% CIs calculated as exact binomial (Clopper-Pearson) CIs.

A total of 2050 samples from biopsies performed in 2041 individual patients were submitted for DIF during the study. Of those, the results of 367 (17.9%) were positive. Biopsy results were categorized by the following clinical impressions: pemphigoid (714 [34.8%]), immunobullous/blistering disorder NOS (552 [26.9%]), dermatitis herpetiformis (239 [11.7%]), vasculitis (210 [10.2%]), CTD (209 [10.2%]), pemphigus (73 [3.6%]), lichen planus (24 [1.2%]), linear IgA disease (15 [0.7%]), or porphyria (14 [0.7%]) (Table 1). Of all biopsies, DIF was positive for the suspected clinical entity in few cases (Table 1). Occasionally, a sample submitted for DIF for a specific immunobullous disease showed positivity for a different immunobullous disorder than was suspected clinically (46 [2.2%]) (Table 1).

The pattern of immunoreactant deposition in DIF-positive pemphigoid biopsies (n = 200) is described in Table 2. Of these, 120 (60.0%) cases positive for pemphigoid had a concurrent H&E biopsy, which showed H&E evidence of pemphigoid, except 1 case that showed interface dermatitis, 1 case showing a nonspecific erosion, and 1 case in which the H&E biopsy was of perilesional skin. Results on 11 (1.5%) of the 714 samples (1.5%) submitted with the clinical impression of pemphigoid were positive for a clinically unsuspected diagnosis on DIF (Table 1).

The pattern of immunoreactant deposition in 18 DIF-positive pemphigus biopsies is reported in Table 2. Twelve of 18 (66.7%) cases with results positive for pemphigus had a concurrent H&E biopsy, which showed H&E evidence of pemphigus. Four (5.5%) of the 73 samples submitted with the clinical impression of pemphigus showed clinically unsuspected DIF findings (Table 1).

Granular deposition of IgA accentuated at the dermal papillae was present in all 33 patients with dermatitis herpetiformis. Twenty-two (66.7%) of these positive cases had a concurrent H&E biopsy that showed dermatitis herpetiformis, except for 9 cases (5 showing hypersensitivity reaction, 1 bacterial folliculitis, and 3 nonspecific excoriation changes). Three (1.3%) of the 239 samples submitted with the clinical diagnosis of dermatitis herpetiformis showed clinically unsuspected DIF findings (Table 1).

All 5 cases diagnosed as linear IgA disease on DIF showed linear deposition of IgA at the basem*nt membrane zone, except 1 case, which showed weak linear deposition of IgM at the basem*nt membrane zone in addition to linear IgA. Three of the 5 samples were submitted with concurrent H&E biopsies, which showed findings typical of linear IgA disease. One of the 15 cases with the clinical impression of linear IgA showed DIF and H&E findings consistent with pemphigoid (Table 1).

The pattern of immunoreactant deposition in 69 samples positive for vasculitis on DIF is reported in Table 3. Fifty-one (73.9%) of these samples scored as positive for vasculitis had a concurrent H&E biopsy. The results of all but 5 biopsies showed H&E changes of leukocytoclastic vasculitis; 4 showed nonspecific changes on H&E (likely due to a very early lesion biopsied) and 1 demonstrated urticarial vasculitis on H&E.

The pattern of immunoreactant deposition in 30 cases positive for CTD is reported in Table 4. The diagnosis of CTD (including LE) was supported by a concurrent H&E biopsy in 23 (76.7%) of these cases. DIF results were positive in less than half of cases of H&E biopsy-confirmed LE (23 of 47 [48.9%]). The results of an additional 24 biopsies from patients clinically suspected to have CTD were negative on DIF, yet the concurrent H&E biopsy showed findings consistent with CTD (including LE). Two (1.0%) of 209 samples submitted with the clinical impression of CTD were positive for a clinically unsuspected diagnosis on DIF (Table 1).

All 8 DIF samples positive for lichen planus showed cytoid bodies positive for IgM with or without other immunoreactants. Five of 8 samples had a concurrent H&E biopsy and all showed lichen planus.

Samples from 14 biopsies were submitted for evaluation of porphyria. Of those, 4 (28.6%) samples were positive on DIF; 3 showed deposition of IgG in dermal vessel walls (with or without other immunoreactants in the vessel walls and/or at the basem*nt membrane zone) and 1 sample demonstrated only IgM deposition in dermal vessel walls. Two of the 4 samples were submitted with concurrent H&E biopsies, which showed findings typical of porphyria.

Five samples showed an immunoreactant deposition pattern and histopathologic findings in the H&E biopsy that were difficult to categorize and were signed out descriptively. Of the 552 samples submitted with the clinical impression of immunobullous/blistering disorder NOS, 25 (4.5%) were positive, with pemphigoid the most common diagnosis (13 [2.4%]), followed by vasculitis (5 [0.9%]), pemphigus (4 [0.7%]), immunobullous NOS (1 [0.2%]), dermatitis herpetiformis (1 [0.2%]), and CTD (1 [0.2%]).

Within the 1678 samples negative for specific immunoreactant deposition, 131 (7.8%) showed sparse deposition of various immunoreactants, mainly granular deposition of C3 or IgM or fibrin at the basem*nt membrane zone in biopsies from sun-exposed skin of the extremities (107 [81.7%]) or intravascular or perivascular fibrin deposition in biopsies from the lower extremities (24 [18.3%]).

Only IgG, IgA, and C3 were 100% sensitive (95% CI, 98.6%-100.0%) and 100% specific (95% CI, 99.7%-100.0%) in detecting primary immunobullous disorders, and only IgA, C3, and fibrin were 100% sensitive (95% CI, 94.8%-100.0%) and 100% specific (95% CI, 97.5%-100.0%) in detecting vasculitis. None of the DIF biopsies showed specific staining with the albumin antibody. Considering the 107 biopsies from sun-exposed skin with nonspecific deposition of immunoreactants (described in the previous paragraph), in the present study the sensitivity for DIF biopsy in the diagnosis of LE including CTD was 48.9% (95% CI, 34.1%-63.9%) and the specificity was 46.8% (95% CI, 39.7%-53.9%).

Use of DIF is important in the diagnosis of certain skin diseases. Although some variation exists in how laboratories perform DIF (because a number of laboratories no longer use albumin), the practice of using a 5- to 6-antibody panel remains widespread and, based on the data presented herein, appears unnecessary. Results of this quality improvement study suggest that the antibodies used in DIF can be optimized to the relevant clinical question, and the data also highlight the need to educate clinicians in a more optimal cost-conscious use of DIF, particularly regarding its use in the workup of patients with suspected CTD, lichen planus, porphyria, and vasculitis.

The continued use of the albumin antibody is mainly due to historical precedent, because the presence of albumin and/or fibrin in tissue is largely associated with the degree of vascular permeability present, as might be seen in the setting of a decrease in collagen matrix concentration due to edema associated with increased leakage of albumin and fibrin into skin.^15,16 With the exception of fibrin within vessel walls in vasculitis, the presence of either albumin or fibrin deposition in the skin in the absence of other immunoreactants is nonspecific and does not substantially contribute diagnostically to the DIF assay. In this study and in our cumulative experiences as diagnosticians, no specific pattern of albumin deposition was seen in any DIF biopsies. Although many laboratories no longer use albumin, some still do. The College of American Pathologists laboratory checklist does not identify a requirement that albumin be used in IF microscopy.

In the present series, 100% of samples positive for a diagnosis of primary immunobullous disease (pemphigoid, pemphigus, linear IgA, and dermatitis herpetiformis) were identified using antibodies to IgG, IgA, and C3. The reliable presence of these immunoreactants in immunobullous disease is supported by previous studies characterizing these disorders.^3,4,17 In instances in which the result of DIF biopsy is negative and yet a primary immunobullous disorder is still suspected (eg, if H&E findings are suggestive and/or the clinical presentation is suspicious for an immunobullous disorder), antigen-specific serologic testing is more sensitive and less expensive than redoing or expanding the DIF biopsy and also allows for identification of specific target antigens.¹⁸

In this study, 100% of samples positive for vasculitis were identified using antibodies to IgA, C3, and fibrin. DIF biopsy in the setting of suspected vasculitis is typically performed on a lesion that is less than 24 hours old to specifically evaluate for IgA vasculitis in suspected cases of systemic IgA-mediated syndrome: Henoch-Schönlein purpura.¹⁹ Diagnosis of Henoch-Schönlein purpura requires evidence of systemic involvement in addition to demonstration of IgA-mediated cutaneous small vessel vasculitis. If the clinical question is only whether a purpuric eruption is vasculitis, an H&E biopsy (preferably of a lesion 24-48 hours old) is more cost-effective than DIF biopsy and is sufficient to assess for vasculitis.²⁰

In this series, 247 samples were submitted for clinical diagnoses not optimally supported by findings on DIF: lichen planus, porphyria, and CTD. The lack of necessity for a DIF biopsy is particularly true for lichen planus for which an H&E biopsy and clinical findings are sufficient for diagnosis, and porphyria in which H&E biopsy and laboratory testing of body fluids are pivotal in establishing the diagnosis.²¹ For cases in which lichen planus pemphigoides is suspected, a biopsy for H&E and a DIF panel in which IgM and fibrin are added to IgG, C3, and IgA would be warranted.

The value of DIF as an aid in the diagnosis of LE/CTD is questionable. The sensitivity and specificity of DIF samples taken from sun-exposed skin do not justify use of the test, owing to the known false-positive rate and low sensitivity of the test (further supported in the present study). A positive lupus band on biopsy of nonlesional non–sun-exposed skin is highly specific for systemic LE^22,23 but usually adds little useful information in addition to clinical history, routine histologic characteristics, and serologic evaluation. In particular, a positive lupus band test has been shown to be associated with the presence of anti-dsDNA on serologic evaluation and has a similar sensitivity (approximately 60%) and specificity (approximately 97%) for systemic LE,^24-26 suggesting that DIF biopsy in addition to serologic evaluation is redundant.

Using only 3 antibodies (IgG, IgA, and C3) for diagnosis of immunobullous disease, 3 antibodies (IgA, C3, and fibrin) for vasculitis, and omitting the 247 DIF samples submitted for diagnoses with minimal or no associated diagnostic utility from DIF analysis would have resulted in a savings of $756 840 in the present data set based on Medicare-allowable payments of $148.84 per first antibody and $108.82 per subsequent antibodies. In 21 of the 2050 (1.0%) of submitted specimens, positive findings were noted consistent with an immunobullous disorder other than what was suspected clinically. All these unsuspected diagnoses would have been identified by the panel that would have been used based on the clinical impression, except for 1 case when the clinical impression was pemphigoid and the DIF biopsy showed hom*ogeneous granular IgM, with H&E findings consistent with CTD. Because the sensitivity and specificity of DIF in diagnosis and subclassification of CTD are low, the H&E findings would have been enough to establish the diagnosis of CTD.

In 552 samples (26.9%), the submitting diagnosis was nonspecific immunobullous disease NOS; 4.5% of the results were positive. All positive diagnoses would have been identified with the 3-antibody panel (IgG, IgA, and C3), except for a case of CTD for which, as discussed previously, an H&E biopsy of lesional skin with or without serologic testing would have sufficed. It would be reasonable in these situations for the laboratory or dermatopathologist to begin with the basic panel of IgG, IgA, and C3 and then expand the panel to include IgM and fibrin if there is a concurrent H&E biopsy showing findings consistent with CTD. In addition, if there is no concurrent biopsy for H&E evaluation and additional clinical history to narrow the clinical impression is not obtainable, the full panel of IgG, IgA, IgM, C3, and fibrin is likely warranted. Pathologists working in conjunction with the submitting clinician should make these judgments.

Based on extensive clinical experience regarding DIF immunoreactant patterns in cutaneous diseases in conjunction with the present study results, we propose the following recommendations for the use of DIF testing and evaluation:

1. Limit the panel to IgG, IgA, and C3 in suspected immunobullous disease and IgA, C3, and fibrin in suspected vasculitis.

2. Educate technical staff on use of DIF protocols tailored to the submitting diagnosis, including lists of specific clinical differential diagnoses for which a 3-antibody panel rather than the full panel is warranted.

3. As appropriate or in cases in which the clinical impression is broad or nonspecific (ie, immunobullous disease NOS), addition of IgM and fibrin to the basic IgG, IgA, and C3 panel may be necessary. Adding 1 or more antibodies based on clinical pathologic correlation should not result in a substantial delay in diagnosis, nor will this additional testing negatively impact care rendered to the patient.

4. These data highlight a knowledge gap and opportunity to adopt a more high-value, cost-conscious use of DIF for clinical use that local and national educational societies should consider in planning curricula. Regardless of the rationale for performing a skin biopsy, accurate communication between dermatologist and dermatopathologist is necessary. This may be especially true for a high-complexity and costly test, such as DIF.

Further details on the recommendations are available in Table 5.

This study has limitations. It was retrospective and follow-up clinical data were not available. Although the sample size was relatively large, more cases would have been helpful. The fact that the DIF biopsies were interpreted by several different board-certified dermatopathologists rather than just one is also a potential limitation.

This comprehensive evaluation of DIF biopsy results from a single institution found that the standard practice of reflexive antibody testing using a 6-antibody panel for all DIF biopsies is unnecessary. A DIF protocol tailored to the submitting diagnosis may enhance cost-effectiveness without loss of test sensitivity and specificity. In addition, not using DIF biopsy in cases of suspected lichen planus, porphyria, and CTD, as well as limiting its use in vasculitis to cases of suspected systemic IgA vasculitis, may reduce costs without compromising patient care.

Accepted for Publication: August 13, 2021.

Published Online: October 6, 2021. doi:10.1001/jamadermatol.2021.3892

Corresponding Author: Thomas Horn, MD, MBA, Department of Dermatology, Massachusetts General Hospital, 55 Fruit St, Boston, MA 02114 (thorn@mgh.harvard.edu).

Author Contributions: Drs Reimann and Horn had full access to all of the data in the study and take responsibility for the integrity of the data and the accuracy of the data analysis.

Concept and design: Reimann, Horn.

Acquisition, analysis, or interpretation of data: All authors.

Drafting of the manuscript: All authors.

Critical revision of the manuscript for important intellectual content: Reimann, Horn.

Statistical analysis: Reimann, Moynihan.

Administrative, technical, or material support: Moynihan, Horn.

Supervision: Reimann, Horn.

Conflict of Interest Disclosures: None reported.

Additional Contributions: Tammie Ferringer, MD (Geisinger Commonwealth School of Medicine), provided critical review of the manuscript, and Joseph Tung, MD (University of Pittsburgh School of Medicine), and Kevin Moore, MD (Harvard Medical School), assisted with database development. No compensation was received.