Bullous Tinea: Single-Center Retrospective Histopathologic Review of 25 Skin Biopsies

Our institutional review board reviewed and approved the study, classifying it as exempt (IRB protocol 263061). The study was a retrospective review of patients diagnosed with bullous dermatophyte infection by skin biopsy at the University of Arkansas for Medical Sciences (Little Rock) from 1998 through 2021. Case material was derived from our academic hospital clinic, serving adult patients, and from community dermatology clinics, serving patients of all ages. To find cases, diagnostic text searches were performed using the institution’s laboratory information system (currently Epic Beaker). Search terms included bullous tinea, bullous dermatophytosis, blister and tinea, bullous and tinea, and blister and dermatophytosis. From retrieved search results, any reports of tinea versicolor, suggestive of tinea versicolor, and tinea capitis were excluded. Remaining reports were reviewed to confirm the diagnosis of bullous tinea/bullous dermatophytosis. All patients with a positive biopsy result, as defined by the presence of fungal elements on hematoxylin-eosin (H&E)–stained sections or on special stained sections (periodic acid–Schiff [PAS] or Gomori/Grocott methenamine silver [GMS]) and evidence of bullae by histology and/or clinical description, were included. Reports without mention of blister (clinical or histopathologic) were excluded. H&E- and PAS- or GMS-stained slides from a final cohort of 25 cases were retrieved, de-identified, and examined by the study authors. Clinical information, including patient demographics, clinical description, and clinical preoperative diagnosis, was gathered from the pathology report. Histopathologic features were recorded and tabulated. Data were analyzed using basic descriptive statistical tests (Excel, Microsoft, Redmond, Washington) to determine the frequency of various histopathologic and clinical characteristics.

Twenty-five cases of biopsy-proven bullous dermatophytosis were retrieved; the summary of clinical parameters is provided in Table 1. The average patient age at the time of diagnosis of bullous dermatophytosis was 50 years (SD, 15.15 years; range, 20–83 years); 40% (n = 10) of patients were male. The foot was the most common anatomic site (12 cases; 48%), followed by the forearm (6 cases; 24%), arm (2 cases; 8%), thigh (2 cases; 8%), toe (2 cases; 8%), and hand (1 case; 4%) (Table 1). In only 26% (n = 6) of cases was tinea explicitly mentioned in the preoperative clinical differential diagnosis. Other frequent clinical considerations included spongiotic dermatoses (dyshidrosis, contact dermatitis, eczema), nonfungal infections (herpetic, viral, mycobacterial), autoimmune-mediated blistering diseases (bullous pemphigoid, epidermolysis bullosa acquisita, pemphigoid gestationis), psoriasis/palmoplantar pustulosis, Sweet syndrome, palisaded granulomatous dermatoses (granuloma annulare or palisaded and neutrophilic granulomatous dermatitis), and fixed drug eruption. In total, 14 cases (56%) described the presence of vesicles, blisters, or bullae or listed at least one vesiculobullous disorder in the clinical differential diagnosis, and 2 additional cases (8%) listed pustular disorders. One case included clinical consideration of an entity that can sometimes blister or vesiculate clinically (dyshidrosis). The clinical differential diagnoses in 6 cases included entities not classically associated with blister formation. No clinical differential diagnosis was provided for 3 cases, although 1 of these cases provided a clinical history of “clear, fluid-filled blisters.”

A summary of the visualized epidermal alterations is provided in Table 2. Histopathologic features that were present in more than half of cases included spongiosis (100%; n = 25), intraepidermal blister (92%; n = 23), intraepidermal neutrophils (92%; n = 23), and parakeratosis (68%; n = 17). Some degree of acanthosis was usually present, seen as either well-developed (60%; n = 15) or mild/focal (20%; n = 5). In contrast, histopathologic features that were less commonly identified included apparent acantholysis (36% present; n = 9 of 25), sandwich sign (32% present; n = 8) (Figure 1, A), subepidermal blister (28% present; n = 7), and Langerhans cell (LC) microabscesses (20% present; n = 5). Well-developed vacuolar interface changes were seen in only 1 case, whereas this reaction pattern was focally appreciated in 4 additional cases (16%). With regard to the anatomic location of blister formation, intraepidermal blister only was encountered in 18 cases, subepidermal blister only was seen in 2 cases, and 5 cases showed both intraepidermal and subepidermal blister formation (Figure 1, B).

Visualized dermal changes are tabulated in Table 3. Dermal edema was commonly seen (64%; n = 16), as were extravasated erythrocytes (68%; n = 17). Dermal inflammation was examined and characterized in 20 of the 25 cases (in 5 cases, the biopsy was too superficial to characterize inflammatory components). Lymphocytes were the predominant inflammatory cell type in 90% (n = 18) of cases (Table 3). Neutrophils were present in 85% (n = 17) of cases but were the predominant cell type in only 5% (n = 1) of cases. The depth of infiltrate was largely limited to the superficial dermis in 30% (n = 6) of cases, extended to the mid-dermis in 45% (n = 9) of cases, and extended to the deep dermis in 25% (n = 5) of cases (Figure 1, C).

Four cases of bullous tinea were diagnosed on H&E, without the use of special stains; hyphae were detected in the stratum corneum in these cases. The remaining cases required special staining. PAS stain was performed in 21 cases. No cases had GMS performed. Fungal hyphae were found in the stratum corneum in 100% of these cases (n = 21), in the blister cavity in 4 cases (19%), and within the hair follicle in 1 case (5%) (Figure 2, A and B). Excluding PAS, additional staining and/or testing was ordered in 24% (n = 6 of 25) of cases. This included a separate biopsy for direct immunofluorescence in 5 cases and 1 case in which additional staining for herpes simplex virus, varicella zoster virus, mycobacteria (Fite stain), and bacteria (Gram stain) was performed. Of the 5 direct immunofluorescence tests performed, 80% (n = 4) were negative and 20% (n = 1) showed nonspecific perivascular fibrin and C3 deposition.

For the cases in which PAS stain was performed (n = 21), the H&E was carefully reexamined for visible fungal hyphae (Figure 3). In 10 cases (48%), hyphae were visualized in the stratum corneum only with PAS and not seen on H&E. In the 4 cases with hyphae seen in the blister cavity, these were apparent only on PAS-stained sections. The single case that included hair follicle involvement, defined as hyphae seen in/around the hair shaft at the base of the hair follicle with deep perifollicular inflammation, had hyphae visualized on both H&E and PAS.

Bullous dermatophytosis is a rare blistering disorder resulting from fungal infection. To date, limited literature exists regarding the spectrum of clinical and histopathologic changes of this treatable condition. Within this case series, we report a summary of histopathologic and clinical features present in 25 biopsy-proven cases of bullous tinea.

Clinically, bullous dermatophytosis often presents as pruritic, sometimes painful, vesicular or bullous eruption with possible underlying erythema.⁷  In general, bullous tinea is thought to occur primarily in adults and most frequently presents as bullous tinea pedis. Diagnosis of bullous tinea may be made on clinical examination alone, but within this patient cohort, biopsy was necessary for diagnosis. Within our patient cohort, the average age was 50 years, and the most common site was the foot, as expected. However, approximately half of our cases arose at other anatomic sites, highlighting the need to keep bullous tinea within the clinical and histopathologic differential diagnosis even when anatomic sites besides the feet are being evaluated. Admittedly, it is unlikely that 50% of bullous tinea infections occur off the foot, as our study design relied on biopsy-proven bullous tinea. As bullous tinea may be more readily recognized clinically when presenting on the foot, a subset of cases never require skin biopsy for diagnosis. Further studies incorporating clinical/in-office diagnosis in addition to biopsy diagnosis are required to more accurately assess the incidence of bullous tinea occurring at nonpedal sites.

With regard to the preoperative clinical impression, bullous tinea was rarely suspected and mimicked spongiotic disorders that can vesiculate, as well as other infections, immune-mediated blistering disorders, palmoplantar pustulosis, fixed drug eruption, Sweet syndrome, and palisaded granulomatous dermatoses. As this study was not designed to thoroughly review clinical charts, we were not able to determine exactly what percentage of patients presented with clinically evident blisters. However, with respect to the provided clinical information for each case, 14 cases described the presence of vesicles, blisters, or bullae or listed at least 1 vesiculobullous disease in the clinical differential diagnosis; we can presume that these patients all had clinical evidence of blistering/vesiculation. Two additional cases described the presence of pustules clinically, and 1 case listed dyshidrosis, an entity that often blisters/vesiculates and may have had clinical evidence of blistering, as a clinical impression. These data suggest that at least two-thirds of our cohort had some clinical evidence of blistering or pustulation, whereas the others may have been limited to microscopic blistering.

Classically taught histopathologic clues to the diagnosis of nonbullous dermatophyte infection include neutrophils in the epidermis and/or stratum corneum, compact stratum corneum, parakeratosis, and a spongiotic reaction pattern.^8,9  The compact orthokeratosis stratified between parakeratosis, otherwise known as sandwich sign, has also been described as a clue to dermatophytosis.^10,11  However, it has been acknowledged that no single histopathologic feature is completely specific or pathognomonic for dermatophyte infection.⁶  Detection of hyphae is strongly suggestive of the diagnosis; however, even this feature must be taken in context with the clinical presentation, as it may be incidentally identified in other dermatologic conditions. Like nonbullous tinea, our cases of bullous tinea consistently demonstrated spongiosis (100%; n = 25) and intraepidermal neutrophils (92%; n = 23). Nearly all cases showed intraepidermal blistering (92%; n = 23), although a subset of cases showed subepidermal blister formation (28%; n = 7), either alone or in association with the intraepidermal blister. Both intraepidermal and subepidermal vesiculation are reflective of the prominent edema universally present in the epidermis and dermis of these biopsies.

In addition to these typical histopathologic findings, several novel characteristics were observed in our population that may be unique to cases of bullous tinea. Although the importance of intraepidermal neutrophils is commonly stressed in diagnosing tinea, our study suggests that dermal neutrophils are seen as the nonpredominant cell type in the majority of cases of bullous tinea (85%; n = 17) and can serve as an additional clue to diagnosis. In addition, the presence of deep inflammation, as seen in 25% (n = 5) of our cases, does not exclude the possibility of a superficial infection. Although classically considered an important clue to the diagnosis of dermatophytosis,¹¹  the sandwich sign may be less relevant in the setting of bullous tinea infection, as we detected it only in 32% (n = 8) of our cases. Alternatively, the sandwich sign may be overall less common than is often touted; one study that reviewed 18 biopsy samples of tinea faciei and tinea corporis detected a sandwich sign in only 22% of cases.¹⁰ 

LC microabscesses in skin biopsies have been correlated to positive patch testing in allergic contact dermatitis.¹²  In our series, LC microabscesses were rare, present in only 20% (n = 5) of cases. Further investigation of the relationship between LC microabscesses and dermatophyte infection may provide a useful clinical insight. An inverse correlation could help steer a practitioner’s differential away from the diagnosis of tinea infection and toward other spongiotic dermatoses. Lastly, detection of apparent acantholysis was seen in 9 cases (36%). Defined as a rounding up and retraction of keratinocyte cytoplasm, acantholysis is detected when there are hereditary deficiencies or autoantibodies generated against desmosome proteins. Considering the edema associated with tinea infection, the likely true mechanism by which blistering occurs in bullous tinea is spongiosis and massive edema rather than actual desmosome malfunction (hence our term apparent acantholysis). However, the histopathologic detection of apparent acantholysis in a subset of our cases of bullous tinea represents a potential pitfall and could prompt unnecessary consideration for immune-mediated or inherited blistering disorders. Recognition that this feature may be observed in tinea infections is important to prevent misdiagnosis.

Identification of hyphae on H&E alone in biopsies of dermatophytosis ranges widely, from 33% to 78%, within the literature.^6,10,13  Therefore, many cases require use of special stains for identification.^6,9  GMS and PAS represent the most widely used special stains in pathology for the detection of fungal elements.⁸  A recent review article⁷  regarding the utility of PAS and GMS staining in dermatopathology did not identify studies directly comparing the sensitivity or specificity of GMS versus PAS in non-nail fungal infections, and so no conclusion regarding the superiority for either ancillary test could be drawn. Within our study, only 16% (n = 4) of our cases had sufficient hyphae to preclude the need for special staining during routine sign-out, although in our retrospective review of the biopsies, hyphae could actually be identified in 60% (n = 15) of cases on H&E, which is in line with previous studies. Importantly, this means that 40% (n = 10) of cases required PAS or GMS to reach the correct diagnosis. Hyphae were most commonly identified in the stratum corneum adjacent to the blister, but not within the blister cavity itself, suggesting this to be the best place to examine the H&E for organisms. Similarly, Wang et al¹³  retrospectively reviewed 110 cases of (nonbullous) tinea and found that fungal elements were most frequently observed in the stratum corneum, and rarely within intracorneal pustules.

Our study has some limitations. As this was a study designed to evaluate biopsy-proven cases of bullous tinea, there is the possibility that it overrepresents unusual clinical presentations of the disorder. More classic presentations may have been diagnosed clinically or by skin scrapings, thereby avoiding skin biopsy. Our overall sample size is small, and our search may have missed cases in which bulla formation was not explicitly mentioned in the pathology report. Additionally, tinea infections are common in children. Although children were not excluded from this study, no biopsies were retrieved from children. This likely reflects the predominantly adult population served by our academic medical center and the higher threshold dermatologists may have in obtaining a skin biopsy from children. Lastly, the lack of associated culture data prevents any speculation regarding fungal species that could be more closely associated with clinical or microscopic blistering; this represents an avenue for future study.

This is the first series documenting clinical and histopathologic parameters of biopsy-detected bullous tinea. Although rare, bullous tinea should be considered within the clinical differential of blistering disorders, particularly those occurring on acral sites. It may mimic allergic contact or dyshidrotic dermatitis, immune-mediated blistering diseases, fixed drug eruption, palmoplantar pustulosis, or Sweet syndrome. For pathologists, it is important to remember that bullous dermatophyte infection is rarely suspected clinically, and that H&E frequently lacks obvious fungal hyphae; for this reason, PAS or GMS stains should be routinely considered for biopsies showing intraepidermal and/or subepidermal blister formation and prominent neutrophil infiltration.