New insights in COVID-19-associated chilblains: A comparative study with chilblain lupus erythematosus.

To the Editor: An unexpected outbreak of chilblains has been reported in association with COVID-19. Severe acute respiratory syndrome coronavirus 2 (SARS-CoV-2) infection has been shown in a few documented cases of chilblains. Chilblains may also be observed in acquired lupus and rarely as a manifestation of a familial disorder related to interferonopathies. To enhance understanding of these epidemic chilblains (EC) cases and their relevance to SARS-CoV-2 infection, we studied clinical, hematoimmunologic, histopathologic, immunohistochemical, and virologic characteristics of 7 EC cases and compared them with 11 previous cases of chilblain lupus erythematosus (CLE).

Patients with EC were included between February and April 2020 and were suspected of COVID-19 because they presented with COVID-19 symptoms or were in close contact with patients with presumed/confirmed COVID-19. Exclusion criteria were patients with previous chilblains episode, cold exposure preceding chilblains occurrence, and history of known autoimmune disorder. For each patient, we collected demographic data and clinical and laboratory test results, including exhaustive hematoimmunologic screening, cutaneous histology (including immunostaining for CD123, a plasmocytoid dendritic cell marker, and MxA, a type I interferon [IFN-I]–induced protein), and virologic studies.

The clinicobiological findings of EC and CLE cases are summarized in Table I . Hands, ears, or nose localization were more frequently observed in the CLE group (82% vs 0%). Antinuclear antibodies were detected only in the CLE group (91% vs 0%). Age at onset of chilblains, sex, pre-existing Raynaud phenomenon, and other immunologic abnormalities did not differ between groups. Antineutrophil cytoplasmic antibodies (ANCAs) and lupus-type circulating anticoagulant were found in 2 and 1 patients with EC, respectively, without any clinical manifestation of ANCA vasculitis or thrombosis. No patient with EC had cryoprotein, cold agglutinin, or anticardiolipin antibodies.

Table I

Clinical and biological findings in EC and CLE

Variable	EC (N = 7)	CLE (N = 11)	P value
Female, n (%)	4 (57)	7 (64)	>.99
Age, y, mean (SD)	42 (10)	49 (15)	.27
Previous Raynaud phenomenon, n (%)	4 (57)	4 (36)	.63
Previous other cutaneous symptoms, n (%)	3∗ (43)	8 (73)	.33
Localized to feet, n (%)	7 (100)	2 (18)	<.01
COVID-19 symptoms, n (%)	5 (71)	NA	—
Potential SARS-CoV-2 contact, n (%)	4 (57)	NA	—
Positive antinuclear antibodies, n (%)	0 (0)	10 (91)	<.01
Presence of other immunologic abnormalities, n (%)	3† (43)	9 (82)	.14

CLE, Chilblain lupus erythematosus; EC, epidemic chilblains; NA, not applicable; SARS-CoV-2, severe acute respiratory syndrome coronavirus 2; SD, standard deviation.

Two patients had acrocyanosis, and 1 patient had photosensitivity.

Two patients had antineutrophil cytoplasmic antibodies, and 1 patient had lupus-type circulating anticoagulant.

Our 7 EC cases were histologically similar to CLE. High expression of CD123 and MxA were observed in both groups (Table II ).

Table II

Histologic and immunohistochemical comparison between EC and CLE

Variable	EC (N = 7)	CLE (N = 11)	P value
Epidermis, n (%)
Lymphocyte exocytosis	3 (43)	7 (64)	.63
Confluent necrosis	1 (14)	0 (0)	.39
Apoptotic keratinocytes	2 (29)	4 (36)	>.99
Vacuolized basement membrane zone	1 (14)	8 (73)	.049
Papillary dermis
Edema, n (%)	4 (57)	2 (18)	.14
Lymphocyte infiltrate intensity score,∗ median (range)	2 (1-3)	2 (1-3)	.34
Lymphocyte infiltrate localization, n (%)
Perivascular localization	7 (100)	11 (100)	>.99
Interstitial localization	3 (43)	8 (73)	.33
Other inflammatory cell infiltrate, n (%)	2 (29)	3 (27)	>.99
Lymphocytic vasculitis, n (%)	5 (71)	1 (9)	.01
Congestive vessels, n (%)	2 (29)	0 (0)	.13
Red blood cell extravasation, n (%)	4 (57)	1 (9)	.047
Reticular and deep dermis
Lymphocyte infiltrate intensity score,∗ median (range)	2 (1-3)	2 (0-3)	.77
Lymphocyte infiltrate localization, n (%)
Perivascular	7 (100)	10 (91)	>.99
Interstitial	0 (0)	0 (0)	>.99
Perieccrine	6 (86)	7/10 (70)†	.60
Perineural	4 (57)	7/9 (78)‡	.59
Other inflammatory cell infiltrate, n (%)	2 (29)	3 (27)	>.99
Lymphocytic vasculitis, n (%)	4 (57)	7 (64)	>.99
Leukocytoclastic vasculitis, n (%)	1 (14)	1 (9)	>.99
Congestive vessels, n (%)	3 (43)	1 (9)	.24
Neural section, median (range)	5 (2-9)	3 (0-4)	.008
Hypodermis§
Perivascular lymphocyte infiltrate, n (%)	2/2 (100)	0/2 (0)	.33
Immunohistochemical features
Case with MxA+ cells, n (%)	7 (100)	10/10 (100)‖	>.99
MxA expression, median (range)	180 (105-280)	270 (120-300)	.28
Case with CD123+ cells, n (%)	6 (86)	9/10 (90)‖	>.99
CD123 expression, median (range)	50 (0-60)	15 (0-100)	.32
Positive cutaneous DIF, n (%)	0/3 (0)¶	1/2 (50)#	.4

Bold values are statistically significant.

CLE, Chilblain lupus erythematosus; DIF, direct immunofluorescence; EC, epidemic chilblains; MxA, myxovirus resistance protein A; SD, standard deviation.

Intensity was scored as follow: 0, absence; 1, rare; 2, moderated; 3, intense.

One CLE biopsy sample did not show the eccrine gland.

Two of CLE biopsy samples did not show the nerve.

Hypodermis was observed in 2 biopsy samples in each groups.

One CLE did not have immunohistochemistry analysis.

Three DIF analyses were performed in the EC group.

Two DIF analyses were performed in the CLE group.

SARS-CoV-2 RNA detection performed at a median delay of 23 days after symptom onset (range, 10-36 d) showed negative results in nasopharyngeal, skin biopsy, and plasma samples. Repeated SARS-CoV-2 immunoglobulin (Ig) G/IgA test results were negative for all patients, except for 1 who showed an isolated IgA positivity (time between first symptoms and serologic tests range, 21-51 d).

Active human herpes virus types 6, 7, and 8 and Epstein-Barr virus infections were excluded by reliable tests (polymerase chain reaction).

These results confirmed that chilblains may be considered as a manifestation of high production of IFN-I as observed in interferonopathies. These patients may exhibit only IFN-I associated symptoms or minor forms of COVID-19 infection. High level of IFN-I was associated with moderate cases of COVID-19. Interferon-induced proteins such as IFITM (interferon-induced trans-membrane) 1, 2, and 3 inhibit early replication of several enveloped RNA viruses, such as Middle East respiratory syndrome coronaviruses. In addition, active viral replication may not be necessary to mount an efficient IFN response in SARS-CoV infection. IFN-I may also suppress antibody responses, which might explain the negative serology test results in most patients with EC.

SARS-Cov-2 infection may induce, in some predisposed patients, a high production of IFN-I responsible for a high innate immune protective response. This hypothesis provides additional arguments to propose early IFN treatment for infected high-risk patients.