Thrombotic occlusive vasculopathy in a skin biopsy from a livedoid lesion of a patient with COVID-19.

dear editor, Some authors have reported the presence of cutaneous lesions related to the new coronavirus disease 2019 (COVID‐19), caused by SARS‐CoV‐2, in up to 20·4% of cases. However, these lesions are not well characterized either clinically or histopathologically.1 A recently highlighted finding is congested and oedematous blood vessels along with hyaline thrombi in the alveolar septum, and also in the heart, liver and kidney of three autopsied patients who died due to severe infection by SARS‐CoV‐2.2 Thus, anticoagulant treatment has been proposed to decrease mortality in certain cases of severe COVID‐19 disease.3 We report a case of livedoid purple lesions along with acrocyanosis in a patient with confirmed SARS‐CoV‐2 infection with positive nasopharyngeal swab, showing an underlying obstructive cutaneous vasculopathy and sweat gland necrosis, both previously undescribed in this disease.

A 61‐year‐old man, otherwise apparently healthy except for a history of diabetes, was accepted as an inpatient at the intensive care unit with a diagnosis of COVID‐associated severe bilateral pneumonia complicated with diabetic ketoacidosis. He simultaneously started showing cutaneous lesions. There was no previous history of immunosuppression, new drugs or endovascular manipulation. He was treated with low‐molecular‐weight heparin. Physical examination showed purple ischaemic digits involving the second and fourth distal phalanges of the right hand, and milder on the second, fourth and fifth digits of the left hand. The second, third and fifth digits of the left foot were also involved. In addition, we observed livedoid purplish retiform and roundish patches on the other fingertips, and on both the volar and dorsal areas of both feet and hands, some with angled edges (Figure 1a). No cutaneous or mucosal lesions in other areas were observed.

Figure 1

(a) Purplish retiform and roundish patches (livedoid lesions) on the dorsal areas of the left foot. The contralateral foot presented similar, milder lesions. Some of them presented angled edges. No other cutaneous or mucosal lesions in other areas were observed. (b) Panoramic picture from a skin biopsy showing upper‐epidermal necrosis and intravascular thrombi in a larger muscular dermo–hypodermal artery. (c) Detail of the thrombi in the larger muscular vessel. (d) Detail of fibrinoid necrosis and leucocytoclasia surrounding some of the vessels in the reticular dermis. Endotheliitis is present. (b–d) Haematoxylin and eosin stain, original magnification (b) × 2, (c) × 20, (d) × 40.

A biopsy showed a slightly necrotic upper epidermis. In the papillary dermis, dilated blood vessels were found, most of them filled with hyaline thrombi and a few with a mild neutrophilic component surrounding them. In some areas, larger arterial vessels located in the dermo–hypodermal interface showed focal fibrinoid necrosis surrounded by a scarce neutrophilic infiltrate (Figure 1b–d). Orcein staining demonstrated that the larger vessel was an artery. Sweat gland necrosis and degeneration were present, more evident in the secretory portion of the eccrine sweat coil, with preserved eccrine ducts. Microbiological cultures from skin biopsy and blood, and polymerase chain reaction for SARS‐CoV‐2 from the skin biopsy were negative.

Blood tests showed increased fibrinogen (up to 995 mg mL−1) and D‐dimer levels (60·8 mg L−1) and leucopenia, with later progressive normalization. Coagulation tests showed a heterozygous factor V Leiden mutation with normal antithrombin III, homocysteine, protein C and protein S levels. No prothrombin G20210A mutation, cryoglobulin, anticardiolipin antibodies, anti‐beta‐2 glycoprotein IgG and IgM antibodies, or lupus anticoagulant were demonstrated. Follow‐up after 17 days showed some improvement of the cutaneous lesions and he was extubated shortly after.

Regarding the livedoid and necrotic lesions observed in our case, a recent series described seven cases of COVID‐19 with presence of finger cyanosis and skin bulla, all of them showing coagulation alterations, although none of them were studied histopathologically.4 Five of those patients (71%) died, within a mean time of 12 days.4 Very recently Kolivras et al. reported the first histopathology of a chilblain‐like lesion showing a superficial infiltrate and deep lichenoid, perivascular and perieccrine infiltrates of lymphocytes without any thrombi.5

The retiform purpura and livedoid lesions in our patient could be explained as a cutaneous manifestation of an underlying systemic coagulopathy related to COVID‐19, as has been previously described in other coagulopathies.6 Our patient's heterozygous mutation of factor V Leyden is an additional known risk factor favouring thrombosis.7 Our case's sweat gland necrosis was similar to that present in non‐drug‐related coma‐associated bulla.7 However, in our case, thrombi and even fibrinoid necrosis were observed not only in the upper‐dermis vessels but also in all the other vessels in the biopsy, even in the dermo–hypodermal junction. This unexpected finding could be related to more extensive endothelial damage secondary to COVID‐19 infection, as this has recently been reported after observation of viral inclusion bodies in electron microscopy of the kidney, small bowel and lung of severely affected patients.8

To our knowledge, this is the first report of a patient with COVID‐19 where the presence of an occlusive vasculopathy at the cutaneous level has been demonstrated. In our case, we also observed a striking sweat gland necrosis, a finding previously reported as being associated with ketoacidosis coma but not with COVID‐19 infection. Additional studies are needed to characterize completely the tissue damage associated with the virus, including within the gamut of lesions observed in the skin.