Histology of skin lesions establishes that the vesicular rash associated with COVID-19 is not 'varicella-like'.

Editor

Several articles recently reported a ‘varicella‐like’ rash in patients with COVID‐19. , We observed similar cases at our institution. However, although we agree that the clinical picture is original, we reject that ‘varicella‐like’ denomination since clinical presentation, as well as some histologic features that we wish to report here for the first time, make it clearly different from varicella.

Three patients with a vesicular rash associated with COVID‐19 (RT‐PCR test on a nasopharyngeal swab specimen positive for SARS‐CoV‐2 ARN) were seen at our institution in April, 2020. A biopsy of a vesicle was performed in each. Multiple levels with H&E stain were done; the slides were reviewed independently by two pathologists; only concordant data were validated. A test for SARS‐CoV‐2 was performed on a vesicle in two patients, and a direct immunofluorescence test on perilesional skin in one.

The main features of the cases are reported in the Table 1. Clinical lesions invariably consisted in small, monomorphic vesicles of 2–3 mm diameter, often excoriated at their top; trunk, especially back, was constantly involved (Fig 1a). Itching was absent or light. Evolution was sometimes irregular, until resolution that occurred without scaring after 10–22 days.

Table 1

Clinical and histologic data in three patients with COVID‐19‐associated vesicular rash

	Case 1	Case 2	Case 3
Age (years)	55	55	50
Sex	Female	Male	Female
General symptoms
Fever	Yes	Yes	No
Asthenia	No	Yes	No
Respiratory symptoms	No	Cough, dyspnoea, mild desaturation	Light cough
Anosmia	Yes	No	No
Time to resolution	2 days	18 days	2 days
Skin symptoms
Day of occurrence	Day 6	Day 6	Day 21
Itch	Light, irregular	No	Light, irregular
Involved sites	Trunk	Trunk	Trunk, upper limbs, face
Time to resolution	11 days	22 days	10 days
Any drug intake in the 5 days preceding rash	No	No	No
Histologic data
Date of biopsy	Day 10 of rash	Day 6 of rash	Day 7 of rash
Intraepidermal vesicle	Yes	Yes	Yes
Acantholysis	Yes	Yes	Yes
Dyskeratosis	Yes	Yes	Yes
Suprabasal clefts	Yes	Yes	No
Balloonization of cells	No	No	No
Nuclear viral inclusions	Yes	No	No
Multinucleated cells	Yes	No	No
Mild dermal inflammatory infiltrates	Yes	Yes	Yes
Dermal eosinophils	Yes	Yes	No

Figure 1

(a) COVID‐19‐ associated vesicular rash involving the back in a patient. (b–d) Histologic features of COVID‐19‐associated vesicular rash, haematoxylin‐eosin staining, (b) Acantholysis, intraepidermal vesicle, suprabasal clefts. (c) Prominent dyskeratosis with ‘pomegranate‐like’ aspect. (d) Suspected nuclear viral inclusions (black arrows), multinucleated cells (white arrows).

Histology showed a similar pattern in the three cases, with a prominent non‐ballooning acantholysis leading to the constitution of an intraepidermal unilocular vesicle, with in two patients a clear suprabasal location (Fig 1b). Eosinophilic dyskeratosis was also constant, with on occasion a striking ‘pomegranate‐like’ aspect (Fig 1c). Features more suggestive of a viral infection were present once (Fig 1d). No vasculitis was seen. The direct immunofluorescence performed in one patient and the two SARS‐CoV‐2 PCR tests performed on vesicles were negative.

The rash that we observed was similar to that reported by Marzano , and constituted a picture that we agree to be evocative of COVID‐19. But, in addition, the histologic pattern of prominent acantholysis and dyskeratosis with constitution of an unilocular intraepidermal vesicle in a suprabasal location, reported here for the first time, contributed to delineate a unique entity. Indeed, this pattern is very different from what is seen in varicella, in which major nuclear atypia, large multinucleated cells, acantholysis secondary to ballooning degeneration, involvement of the epidermis basal layer and vasculitis are regularly seen. Other acantholytic disorder (autoimmune or familial pemphigus, Grover's transient acantholytic dermatosis) do share some histologic features with our cases, but with a distinct clinical context. Actually, we are not aware of a viral rash in which such picture is present.

Considering pathology in COVID‐19, data are scarce and rely mainly on autopsies. Reports focused on the description of inflammatory and post‐inflammatory changes (such as fibrosis) in the lungs. Aspects suggestive of a more direct viral effect (multinucleated cells, viral inclusions) have been rarely reported. ,

In conclusion, this picture, clinical as well as histologic, has few in common with varicella, as with other known viral rashes. It appears to us that the denomination ‘varicella‐like rash’ is poorly indicative, and that such a misnomer should no longer be used. To name this entity, we suggest ‘COVID‐19‐associated acantholytic rash’, a denomination underlining a feature that we found striking.

Concerning pathogenesis, angiotensin‐converting enzyme 2, the receptor for SARS‐CoV‐2, has been identified on epidermis basal cell layer keratinocytes ; that might suggest a direct pathogenic effect of the virus here, leading to acantholysis and dyskeratosis. Finally, despite the SARS‐CoV‐2 research on vesicles in our patients was negative, the aspects evocative of a direct viral effect that were seen suggested the presence of viral replication; this should probably encourage isolation precautions until such lesions have resolved.