Cutaneous Manifestations in Confirmed COVID-19 Patients: A Systematic Review

Simple Summary Patients diagnosed with COVID-19 and concomitant skin rashes have been frequently reported. We summarized the cases described to date, including only patients with positive RT-PCR testing from nasopharyngeal swabs. Six hundred and fifty-five patients were found who presented different types of skin rashes, from maculopapular, vascular, vesicular, urticarial, to atypical forms and ocular involvement. Chilblains which are lesions resembling frostbite have been more frequent in the younger population and seemed to predict a milder disease course. Vascular-purpuric lesions appeared in older patients and were linked to a more severe evolution. In the case of vesicular rashes, the possibility of herpesvirus co-infections was raised. Moreover, cutaneous hydroxychloroquine drug reactions have been described. For patients with conjunctivitis, eye discharge might be contagious. These skin manifestations may help identify asymptomatic COVID-19 carriers in some cases or predict a more severe evolution in others. Abstract There have been increasing reports of skin manifestations in COVID-19 patients. We conducted a systematic review and included manuscripts describing patients with positive RT-PCR coronavirus testing from nasopharyngeal swabs who also developed cutaneous manifestations. A total of 655 patients were selected, with different types of skin rashes: Erythematous maculopapular (n = 250), vascular (n = 146), vesicular (n = 99), urticarial (n = 98), erythema multiforme/generalized pustular figurate erythema/Stevens-Johnson syndrome (n = 22), ocular/periocular (n = 14), polymorphic pattern (n = 9), generalized pruritus (n = 8), Kawasaki disease (n = 5), atypical erythema nodosum (n = 3), and atypical Sweet syndrome (n = 1). Chilblain-like lesions were more frequent in the younger population and were linked to a milder disease course, while fixed livedo racemosa and retiform purpura appeared in older patients and seemed to predict a more severe prognosis. For vesicular rashes, PCR determined the presence of herpesviruses in the vesicle fluid, which raised the possibility of herpesvirus co-infections. The erythema-multiforme-like pattern, generalized pustular figurate erythema and Stevens-Johnson syndrome were most frequently linked to hydroxychloroquine intake. A positive PCR determination of SARS-COV-2 from conjunctival swabs suggest that eye discharge can also be contagious. These cutaneous manifestations may aid in identifying otherwise asymptomatic COVID-19 carriers in some cases or predict a more severe evolution in others.

Simple Summary: Patients diagnosed with COVID-19 and concomitant skin rashes have been frequently reported. We summarized the cases described to date, including only patients with positive RT-PCR testing from nasopharyngeal swabs. Six hundred and fifty-five patients were found who presented different types of skin rashes, from maculopapular, vascular, vesicular, urticarial, to atypical forms and ocular involvement. Chilblains which are lesions resembling frostbite have been more frequent in the younger population and seemed to predict a milder disease course. Vascular-purpuric lesions appeared in older patients and were linked to a more severe evolution. In the case of vesicular rashes, the possibility of herpesvirus co-infections was raised. Moreover, cutaneous hydroxychloroquine drug reactions have been described. For patients with conjunctivitis, eye discharge might be contagious. These skin manifestations may help identify asymptomatic COVID-19 carriers in some cases or predict a more severe evolution in others.
Abstract: There have been increasing reports of skin manifestations in COVID-19 patients. We conducted a systematic review and included manuscripts describing patients with positive RT-PCR coronavirus testing from nasopharyngeal swabs who also developed cutaneous manifestations. A total of 655 patients were selected, with different types of skin rashes: Erythematous maculopapular (n = 250), vascular (n = 146), vesicular (n = 99), urticarial (n = 98), erythema multiforme/generalized pustular figurate erythema/Stevens-Johnson syndrome (n = 22), ocular/periocular (n = 14), polymorphic pattern (n = 9), generalized pruritus (n = 8), Kawasaki disease (n = 5), atypical erythema nodosum (n = 3), and atypical Sweet syndrome (n = 1). Chilblain-like lesions were more frequent in the younger population and were linked to a milder disease course, while fixed livedo racemosa and retiform purpura appeared in older patients and seemed to predict a more severe prognosis. For vesicular rashes, PCR determined the presence of herpesviruses in the vesicle fluid, which raised the possibility of herpesvirus co-infections. The erythema-multiforme-like pattern, generalized pustular figurate erythema and Stevens-Johnson syndrome were most frequently linked to hydroxychloroquine intake. A positive PCR determination of SARS-COV-2 from conjunctival swabs suggest that eye

Introduction
Since December 2019, COVID-19 [1] has spread throughout the world at a staggering pace, gradually becoming a pandemic. As of September 19th, the confirmed number of cases has reached 30.827.639 globally and has claimed 958.514 lives [2]. The quantitative reverse transcriptase polymerase chain reaction (RT-PCR) is used to identify the viral nucleic acid in respiratory specimens or blood samples [3]. Common clinical features of COVID-19 include fever, cough, myalgia, fatigue, headache, and diarrhoea [4,5]. Although known to primarily affect the lungs and the respiratory function, recent reports from around the world have brought to our attention the possibility of cutaneous involvement [6]. These dermatologic symptoms may aid in identifying otherwise asymptomatic COVID-19 carriers or predict a more severe evolution in other cases. Therefore, we conducted a systematic review in order to collect clinically relevant information on the dermatologic effects of COVID-19.

Materials and Methods
A systematic review was elaborated following the Preferred Reporting Items for Systematic Reviews and Meta-Analyses (PRISMA) guidelines. A search of PubMed and Science.gov databases was performed for the period 2019-2020 using the terms: Coronavirus, COVID in combination with each of the following: Dermatology, skin, rash. Only articles in English were selected. The last search was run on 19 September 2020. Only manuscripts reporting on patients with positive RT-PCR-SARS-COV-2 testing from nasopharyngeal swabs who also developed cutaneous manifestations were included. Patients with underlying skin conditions were excluded. Eligible articles were assessed according to the Oxford Centre for Evidence-Based Medicine 2011 guidelines [7]. Review articles, meta-analyses, observational studies, case reports, survey snapshot studies, letters to the editor, and comments to the letters were all included. Other potentially relevant articles were identified by manually checking the references of the included literature.
Skin manifestations in patients with positive RT-PCR-SARS-COV-2 testing from nasopharyngeal swabs were assessed. The relation to COVID symptomatology and medication were analyzed. Histopathologic parameters, as well as RT-PCR testing from skin lesions and conjunctival swabs were summarized.
An independent extraction of articles was performed by two investigators according to the inclusion criteria. Disagreement was resolved by discussion between the two review authors. Since the study designs, participants, treatment measures, and reported outcomes varied markedly, we focused on describing the dermatologic findings, their relation to COVID-19 symptomatology, medication, histopathologic parameters, and other relevant investigations.
Limitations of this review lie in the form of confirmation bias in reporting, since every clinician interpreted these cases as either virus-related or medication-related. Moreover, epiphenomena cannot be excluded, as patients with skin rashes and flu-like symptoms might be increasingly likely to seek medical care, considering the actual pandemic situation. In order to limit bias in reporting COVID-19 related cutaneous manifestations, we only included patients with positive RT-PCR testing from nasopharyngeal swabs. Additionally, we objectively presented the skin manifestations as described by the initial authors and encompassed them into a table where mentions regarding the relation to systemic symptoms and medication intake were made. Moreover, histopathologic parameters, where available, were described for each type of skin rash. The interpretation of the initial authors was excluded. Skin manifestations were only categorized by morphology and not labelled as either virus-induced or medication-induced, thus leaving the final interpretation to the reader.

Results
A total of 1629 records were initially identified in the literature search, of which 117 were duplicates. After screening for eligibility and inclusion criteria, 113 publications were ultimately included ( Figure 1). The study and clinical characteristics are summarized in Table 1. The majority of publications were letters to the editor (n = 48), followed by case reports (n = 28), case series (n = 11), observational prospective studies (n = 4), and comments to letters (n = 2). All studies included were rated as level 4 or 5 evidence for clinical research as detailed in the Oxford Centre for Evidence-Based Medicine 2011 guidelines [7]. A total of 655 patients with dermatologic symptoms and positive RT-PCR-SARS-COV-2 testing from nasopharyngeal swabs were included.
Biology 2020, 9,449 3 of 31 labelled as either virus-induced or medication-induced, thus leaving the final interpretation to the reader.

Results
A total of 1629 records were initially identified in the literature search, of which 117 were duplicates. After screening for eligibility and inclusion criteria, 113 publications were ultimately included ( Figure 1). The study and clinical characteristics are summarized in Table 1. The majority of publications were letters to the editor (n = 48), followed by case reports (n = 28), case series (n = 11), observational prospective studies (n = 4), and comments to letters (n = 2). All studies included were rated as level 4 or 5 evidence for clinical research as detailed in the Oxford Centre for Evidence-Based Medicine 2011 guidelines [7]. A total of 655 patients with dermatologic symptoms and positive RT-PCR-SARS-COV-2 testing from nasopharyngeal swabs were included. Various cutaneous manifestations have been described. The most frequent were erythematous maculopapular (n = 250), followed by vascular lesions (n = 146), vesicular (n = 99), urticarial (n = 98), erythema multiforme/generalized pustular figurate erythema/Stevens-Johnson syndrome (n = 22), ocular/periocular (n = 14), polymorphic pattern (n = 9), generalized pruritus (n = 8), Kawasaki disease (n = 5), atypical erythema nodosum (n = 3), and atypical Sweet syndrome (n = 1).

Erythematous Maculopapular
Erythematous skin rashes (n = 250) have been the most frequent cutaneous manifestations in patients with RT-PCR confirmed COVID-19, with a majority of maculopapular patterns (n = 189) and macular erythema (n = 44), followed by papulo-squamous (n = 14) and one of each: Pityriasis rosea-like, Grover disease-like, SDRIFE-like . The majority presented the rash onset after the appearance of COVID-19 symptoms (28 out of 33 cases that specified the timeframe).
Rosell-Diaz et al. conducted a retrospective case series on 12 adult patients with a mean age of 66 years (47-79 years). They had pneumonia and were on hydroxychloroquine, lopinavir/ritonavir treatment. After an average of 20.4 days (10-28) all patients developed papular exanthema, seven of which further developed erythema multiforme-like lesions and three of them presented fever and facial edema. Cutaneous biopsies were performed in two of these patients and were compatible with a drug reaction (Table 2). Underlying viral infections may increase the risk of adverse drug reactions, as it has already been described for the ampicillin rash in infectious mononucleosis or the increased risk of drug reactions in AIDS patients [8]. Antiviral immune responses may facilitate drug allergy and excessive production of proinflammatory cytokines, which has been observed in COVID-19 [101]. The authors strongly recommend that all COVID-19 patients with exanthema and eosinophilia be investigated for drug sensitization [8].   four maculopapular rash, two chickenpox like, one of which presented with Grover disease-like pattern, two livedoid exanthemas Papular phase: Exocytosis with minimal vacuolar changes near the dermal-epidermal junction. In one of these patients, there were nests of intraepidermal Langerhans cells associated with signs of vasculitis and extravasation of red blood cells.
Chickenpox eruption: Characteristic clefts in the lower epidermis, as well as dyskeratotic keratinocytes in the granular layer and also near the basement membrane.
Livedoid exanthematous eruption: Nest of Langerhans cells in the epidermis. In the deep dermis and occasionally in the superficial dermis, there were microthrombi admixed with nuclear and eosinophilic debris. There was a significant degree of interstitial and perivascular neutrophilia with prominent leukocytoclasia. IHC showed striking and extensive deposition of C5b-9 within the microvasculature.
Superficial vascular ectasia and an occlusive arterial thrombus within the deeper reticular dermis in the absence of inflammation. Extensive vascular deposits of C5b-9, C3d, and C4d were observed throughout the dermis, with marked deposition in an occluded artery. A biopsy of normal-appearing deltoid skin also showed conspicuous microvascular deposits of C5b-9.
Modest perivascular lymphocytic infiltrate in the superficial dermis along with deeper seated small thrombi within rare venules of the deep dermis, in the absence of a clear vasculitis. Significant vascular deposits of C5b-9 and C4d.

References Age Sex Rash Type Histopathology
Rosell-Díaz et al. [8] papular exanthema; seven patients developed target-like areas; three developed fever and facial edema; One of them showed a superficial perivascular inflammation with eosinophils and the other showed a lichenoid pattern with eosinophils.
Rubio-Muniz et al. [24] six maculopapular rash, two chilblain, four targetoid lesions, two palpable purpura, two urticaria Maculopapular: In early-onset cases, histopathology showed moderate epidermal spongiosis and perivascular lymphocytic infiltrate with eosinophils in the dermis, whereas the analysis of the delayed lesions showed perivascular lymphocytic infiltrate and histiocytes among collagen fibers without mucin deposits.

Sachdeva et al. reported three cases of COVID-19 with different cutaneous manifestations:
A 72 year old female with a vesicular eruption, a 77 year old female with a morbilliform eruption on the trunk and legs, with purpuric areas on her legs and a 71 year old female with a maculopapular rash, which, interestingly, resembled Grover disease. In all three cases, the rash appeared after the onset of symptoms and two of them after a few days of treatment with HCQ, lopinavir/ritonavir, ceftriaxone [9].
Another peculiar skin rash was described by Estebanez et al.: A 28 year old female presenting with pruritic, confluent erythematous-yellowish papules on both heels, which appeared 10 days after the last dose of acetaminophen. No other skin lesions were present [5].
Tamai et al. described the onset of a maculopapular rash after 11-22 days from the initial symptoms of COVID-19, in three patients. One of them had six days of hydroxychloroquine and favipiravir treatment when the rash appeared. A drug induced eruption was excluded since erythema was relieved without discontinuing the medication [10].
A rash reminiscent of symmetrical drug-related intertriginous and flexural exanthema (SDRIFE) was reported by Mahé et al. It involved a 64 year old female who took acetaminophen 4 days previously. However, the rash disappeared after 5 days in spite of acetaminophen continuation, which excludes its involvement [11]. Sanchez et al. had an elderly patient who developed a digitate papulosquamous eruption clinically reminiscent of pityriasis rosea, 8 days after the onset of symptoms and 2 days after cefpodoxime discontinuation. He was admitted into the hospital after RT-PCR-SARS-COV-2 from nasopharyngeal swabs was positive. RT-PCR performed on a fresh skin biopsy specimen was negative for SARS-CoV-2 ( Table 3). The PCR blood test for Epstein-Barr virus (EBV) was positive, with a viral load of 4.6 log10 copies/mL reflecting EBV replication. Serologic markers indicated reactivation and ruled out acute mononucleosis. The spontaneous resolution of the rash occurred within a week [12].
In a study by Freeman et al., chilblain-like lesions appeared in patients with relatively mild COVID-19 disease courses: Five out of thirty one hospitalized, two deaths [22]. They suggest that the underlying mechanism might be protective, making pernio (chilblain) a marker of a robust, effective host anti-viral response, limiting COVID-19 complications [36]. COVID-19 has been widely suspected as the etiological agent for these lesions, especially since they have been appearing in warm weather conditions. Clinical and histopathologic similarities to chilblain lupus erythematosus have been suggested. A type 1 interferon (IFN-I) mediated immune response is triggered in COVID-19 patients which plays an important part in the antiviral host defense, similar to the one in lupus erythematosus. It has been hypothesized that young patients exhibit an early IFN-I response, therefore muting early viral replication, but still inducing microangiopathic changes that cause a chilblain lupus erythematosus-like eruption. Older patients, however, may have an inadequate or delayed IFN-I response leading to an exacerbated hypercytokinemia with subsequent increased morbidity and mortality [68]. Galván et al. also linked the livedoid/necrotic lesions to older patients and severe disease (10% mortality) [15]. Fixed livedo racemosa, retiform purpura, and true acral ischemia appeared in critically ill patients [15,22,68]. Livedo racemosa and retiform purpura are hallmark manifestations of cutaneous thrombosis, appearing due to partial and complete occlusion of cutaneous blood vessels, respectively [51]. The appearance of livedo reticularis can be explained by the inflammatory effect of SARS-CoV-2 on endothelial cells or vessel-associated smooth muscle cells, both expressing angiotensin converting enzyme 2-receptor on their surface, which is the target of SARS-CoV-2-spike protein [97].
Droesch et al. described three cases of livedo racemosa and one case of retiform purpura. All four patients had markedly elevated D-dimer levels and complement including C5b-9 in skin biopsy samples [51]. Coagulopathy in the context of severe inflammation (elevated D-dimer, fibrinogen, or C-reactive protein levels) has been reported in patients with COVID-19 (Table 4) [102]. Elevated D-dimer is the most common laboratory abnormality and appears to be related to mortality [103]. The case reported by Bosch-Amate et al. highlights the concomitant presentation of cutaneous microthrombi presenting as retiform purpura and macrothrombi presenting as pulmonary thromboembolism in the setting of COVID-19 coagulopathy [44]. These morphologies are different from pernio-like lesions, as suggested by the histopathologic findings: Non-inflammatory to pauci-inflammatory thrombi without other findings linked to pernio, such as vacuolar interface changes, papillary dermal edema, and dermal lymphocytic infiltrate. This suggests that thrombotic disease in critically ill COVID-19 patients has a cutaneous correspondent manifesting in the forms of livedo racemosa, retiform purpura, or acro-ischemia. One study implicated activation of the alternative complement pathway in cutaneous thrombosis pathophysiology [22,72]. Table 4. Modifications of cutaneous markers in COVID-19 patients.
Droesch et al. [51] livedo racemosa on hands and forearms in three patients and retiform purpura on hands and forearms All four patients had D-dimer levels of more than 3 µg/mL (normal range, 0-0.229 µg/mL) Magro et al. [72] retiform purpura, buttocks elevated D-dimer of 1024 ng/mL (normal range 0-229) on presentation, which peaked at 2090 ng/mL on hospital day 19, and a persistently elevated INR of 1.6-1.9, but a normal PTT and platelet count.
dusky purpuric patches, palms, and soles markedly elevated D-dimer of 7030 ng/mL, but normal INR and PTT reticulated eruptions, consistent with livedo racemosa, chest, legs, and arms D-dimer was elevated at 1187 ng/mL, with a normal platelet count and PTT, but an elevated INR of 1.4.
Suarez-Valle et al. [91] chilblain-like eruption D-dimer was elevated in the three of them and fibrinogen in two, but no other coagulation abnormalities were detected Verheyden et al. [97] symmetric livedo reticularis elevated D-dimers Almost 20% of COVID-19 patients present with severe disease consisting of microangiopathic ARDS and extrapulmonary thrombotic complications associated with markedly elevated D-dimers, which indicate an excessive activation of the coagulation pathway [72,104]. Magro et al. used pulmonary and cutaneous biopsy and autopsy samples from five patients with severe COVID-19 and demonstrated that critically ill cases are associated with generalized thrombotic injury. They appeared to be complement mediated: Extensive deposits of C4d, C5b-9, and MASP2 in the lungs of the two autopsied cases and C4d, C5b-9 in the three cases with a retiform and purpuric rash [72]. Similarly, Rotman et al. described a pauci-inflammatory thrombogenic vasculopathy with deposits of C3d, C4d, C5b-9, and MASP2 within the microvasculature and positive SARS-CoV-2 envelope/spike glycoprotein and ACE2 receptors in biopsy specimens [85].
Zhang Y et al. conducted a retrospective study on seven critically ill COVID-19 patients. They all had acral ischemia in the form of finger or toe cyanosis, skin bullae, and dry gangrene. D-dimer, fibrinogen, and fibrinogen degradation products were significantly elevated in most patients. Four patients were diagnosed with disseminated intravascular coagulation (DIC) and finally five patients died. The median time from acro-ischemia to death was 12 days [100].
Fernandez-Nieto et al. described two different morphological patterns: A diffuse pattern, found in 18 patients (75%), which consisted of small papules, vesicles, and pustules of varying sizes, at different stages simultaneously, affecting more than one corporal area, including the palms and soles in two cases; a localized pattern found in six patients (25%), consisting of monomorphic lesions, at the same stage of evolution, with no more than one central area affected. Vesicular rashes appeared before COVID-19 symptoms in two cases, concomitant in three cases, and after in 19 cases with a median latency time of 14 days (range 4-30 days). An important mention is that seven patients had received HCQ, lopinavir/ritonavir, and azithromycin before the rash appeared. Multiplex PCR for herpesvirus and real-time RT-PCR for SARS-CoV-2 from the vesicle content performed in four cases were negative [55].
Llamas et al. also performed a herpesvirus family microarray PCR from the vesicle fluid in three patients with vesicular rash. The results showed a combination of HSV-1, HSV-6, and Epstein Barr virus (EBV) in case #1, HSV-1 and HSV-7 in case #2, and Varicella Zoster virus (VZV) in case #3. SARS COV-2 PCR in the vesicle fluid could not simultaneously be performed, which cannot completely rule out its additional involvement [70]. Vesicular exanthems are caused by viruses that can replicate in epidermal cells: DNA viruses, such as poxviruses, herpes simplex, varicella zoster [105], and some RNA viruses, including coxsackieviruses [106]. These findings bring into attention the possibility of a co-infection with herpesviruses, which might be responsible for the vesicular type of rash. However, the etiopathogenic role of SARS-COV-2 cannot be completely ruled out. Furthermore, reactivations of herpes simplex one HSV-1 [16] and herpes zoster have been reported in the evolution of COVID-19 [52,53,87].

Urticarial Rash
Urticarial rashes have been relatively frequent too.
Unique urticaria presentations include a case of annular urticaria plaques involving the upper limbs, chest, neck, abdomen of a 39 year old male, which appeared concomitantly with a 39 • C fever and no history of drug intake [41]. Moreover, a case of urticaria and angioedema of the face and hands in a 46 year old female appeared 2 days before the onset of COVID-19 symptoms and no history of drug intake, but with a known history of hay fever and mild asthma [60].
Robustelli et al. described the case of a 70 year old woman presenting with a widespread eruption on an erythematous-oedematous base, with scattered pustules and scales, involving the face, trunk, upper limbs, and symmetric targetoid lesions over the buttocks, thighs, and legs. This eruption had a rapid onset and it appeared 3 days after treatment withdrawal: Lopinavir/ritonavir and hydroxychloroquine (HCQ) for 10 days for COVID-19 pneumonia. The patient had no personal or family history of psoriasis. A histopathological description was consistent with acute generalized exanthematous pustulosis (AGEP) [84].
AGEP is characterized by a sudden onset of widespread non-follicular sterile pustules arising within large areas of oedematous erythema and it has been secondary to drug intake in 90% of cases [107]. Rarely, atypical cases of AGEP with the development of target-like lesions have been described, especially in patients taking HCQ [108][109][110]. HCQ has been described as one of the main drugs involved in triggering AGEP [111,112]. AGEP has a rapid onset of up to 48 h of ingestion, often with an acute onset of fever and leukocytosis. Schwartz et al. brings into discussion another diagnostic entity, generalized pustular figurate erythema (GPFE), which has a longer onset, of 2 to 3 weeks after drug ingestion (range 4-27 days) and it is typically due to hydroxychloroquine. Previously described as atypical AGEP, GPFE may at first be evident as erythematous papules and plaques on the face, with facial edema and generalized urticaria, with development of nonfollicular pustules atop and finally erythematous and sometimes atypical targetoid erythema multiforme-like plaques on the trunk and extremities [107]. Taking into account the atypical clinical presentation and the longer onset of AGEP, the case reported by Robustelli et al. may fall into the GPFE category. Additionally, Abadias et al. described two cases of GPFE following 2 and 3 weeks, respectively, of HCQ treatment [37].
A retrospective case series by Rosell-Díaz et al. analyzed the appearance of a papular exanthema in 12 adult patients with RT-PCR diagnosed COVID-19 pneumonia. They all received hydroxychloroquine and lopinavir/ritonavir in combination with different other medications for 10 to 28 days before the rash onset. Seven of them developed target-like lesions and most of them had eosinophilia. A histopathological examination was performed in only two patients and was compatible with a drug reaction. The authors suggest that antiviral immune responses may induce drug sensitization via excessive production of proinflammatory cytokines, which has been observed in COVID-19. Moreover, exanthema and eosinophilia might be an indicator for an adverse drug reaction [8].
Erythema-multiforme (EM) pattern was reported in four cases by Rubio-Muniz et al., [24] one by Chaabane et al. [20], and one case of fever and rash at presentation, by Navaeifar et al., in a one year old male [75]. Jimenez-Cauhe et al. reported four cases of erythema-multiforme after 10-16 days of hydroxychloroquine, lopinavir/ritonavir, and azithromycin treatment [63]. Sakaida et al. described a case of EM pattern in a 52 year old female after 3 days of cefcapene and loxoprofen treatment for a dental procedure, 7 days after which she developed COVID-19 symptoms. In this case, drug eruption appeared in the latency period and it could be explained by a drug hypersensitivity expressed in some COVID-19 patients [86].
Davoodi et al. reported on a rare case of Stevens-Johnson syndrome in a 42 year old female with COVID-19 pneumonia which appeared after 2 days of treatment with HCQ and acetaminophen [46]. The rash resolved after HCQ discontinuation and lopinavir/ritonavir switch, thus demonstrating, at least in this case, that the cutaneous drug reaction was due to HCQ, and not lopinavir/ritonavir.
Ping et al. evaluated 28 patients with positive RT-PCR testing from nasopharyngeal swabs. Of them, 11 presented ocular involvement in the form of conjunctival hyperemia, chemosis, epiphora, and increased secretions and two patients had positive findings for SARS-CoV-2 in their conjunctival swabs, as well [80]. Using the guideline on diagnosis and treatment of the novel coronavirus pneumonia issued by the National Health Commission of the People's Republic of China, three cases were judged as moderate, two as severe, and six as critical [113], thus suggesting that ocular abnormalities occur in patients with more severe systemic manifestations.
Additionally, Kalner et al. reported two cases of recurrent dusky red, nonpruritic, nonblanching periorbital dyschromia in a 43 year old female and a 50 year old male, with moderate systemic symptoms. The dermatitis resolved with the resolution of systemic symptoms of COVID-19 [66]. Similarly, Olisova et al. reported the case of a 12 year old patient with periocular macular erythema with purpuric areas and strawberry tongue. The rash spontaneously resolved within 3 days [76].
Ocular involvement was reported in three other cases, one involving a 4 year old boy with polymorphic pattern: Bilateral nonpurulent conjunctivitis, strawberry tongue, erythematous lacy rash on the palms [98], and two cases with Kawasaki disease-like presentation [64,83].
Polymorphic patterns are unusual, especially when they appear in the same individual. A hypothesis to explain this polymorphism may be that in some cases there are alternative causes, which are different virus strains or different host reactions [15]. Aghazadeh et al. described the case of a 9 year old girl with vesicular oral eruption and acral erythematous papules and plaques [38]. Young et al. reported a 68 year old male with polymorphic pattern consisting of a morbilliform rash on his trunk, acral purpura and an ulcerated, purpuric plaque with livedoid borders on his buttocks [99]. The cases reported by Freeman et al. were a combination of either morbilliform and urticarial rash or morbilliform and pernio-like lesions [22]. Wolfe had a 4 year old patient who presented with bilateral nonpurulent conjunctivitis, strawberry tongue, and erythematous lacy rash on the palms [96]. Chaabane et al. described a pruriginous rash on the upper chest concomitant with unilateral livedo reticularis in a 35 year old woman [20]. Klimach et al. had a 13 year old patient with a maculopapular rash on his legs and chilblain-like lesions on the soles of his feet [67] and Olisova et al. a 12 year old patient with periocular macular erythema with purpuric areas and strawberry tongue [76]. Except for two cases [67,76], who had a few days history of acetaminophen intake, none of the others had any relation to a drug intake.
A noteworthy mention is the presence of cutaneous hyperesthesia [114]. In a letter to the editor, Krajewski et al. described the cases of two COVID-19 patients who developed abnormal hypersensitivity after the onset of fever and general symptoms, but which subsided 10 days after treatment inception [114].

Discussion
Erythematous maculopapular skin rashes were the most frequent cutaneous manifestation in COVID-19 patients. Cutaneous biopsies, although performed only in two of these patients, indicated a drug reaction [8]. However, the role of SARS-COV-2 cannot be completely excluded. It has been suggested that underlying viral infections may increase the risk of adverse drug reactions, as it was already established for the ampicillin rash in infectious mononucleosis or the increased risk of drug reactions in AIDS patients [8,101]. Patients with exanthema and eosinophilia might benefit from drug sensitization investigation.
Vascular lesions linked to COVID-19 were the second most frequently described in the literature. Chilblain-like lesions seem to be linked to a milder disease course and to affect the younger population [22,36,68]. Conversely, fixed livedo racemosa, retiform purpura, and true acral ischemia appeared in older, critically ill patients and seemed to predict a more severe prognosis [15,22,72]. This might be explained by an earlier IFN-I mediated immunologic response in young patients which mutes early viral replication. On the other hand, a delayed IFN-I response in older patients [68] along with coagulopathy [51,100,102,103] and alternative complement pathway activation [36,72,85,104] led to a generalized thrombotic state, clinically manifested as retiform purpura and pulmonary thromboembolism [44]. Vesicular rashes have been described as two different patterns, diffuse and localized. The majority appeared after the onset of COVID-19 symptoms and seven patients had received HCQ, lopinavir/ritonavir, and azithromycin before the rash appeared [55]. A herpesvirus family microarray PCR from the vesicle fluid was positive for different combinations of HSV-1, HSV-6, EBV, HSV-7, VZV in three patients. These findings raise the possibility of herpesvirus co-infection, which might be responsible for the vesicular type of rash [70], although the etiopathogenic role of SARS-COV-2 cannot be completely excluded. Multiplex PCR for herpesvirus and real-time RT-PCR for SARS-COV-2 from the vesicle content performed in four cases were negative [55]. Reactivations of HSV-1 [16] and herpes zoster [52,53,87] have also been reported in COVID-19 patients.
Erythema-multiforme-like (EM) pattern, generalized pustular figurate erythema (GPFE), and Stevens-Johnson syndrome (SJS) have been described in patients with COVID-19, the majority of whom had a history of drug intake [8,37,46,63,84,86]. A histopathological examination was performed in two of the cases and was compatible with a drug reaction [8]. Several articles in the literature have linked AGEP and GPFE to hydroxychloroquine (HCQ) [37,108,111,112]. Additionally, Davoodi et al. reported a SJS case after 2 days of HCQ and acetaminophen. The skin rash resolved, even though lopinavir/ritonavir was introduced instead of HCQ [46], which takes the focus off the lopinavir/ritonavir role, at least for this patient.
Ocular involvement in the form of conjunctivitis appears to be linked to a more severe form of COVID-19 disease. Ping et al. determined SARS-COV-2 in the conjunctival swabs of two patients, which suggests that eye discharge can also be contagious [80].

Conclusions
The high diversity of cutaneous manifestations linked to the novel coronavirus, coupled with the low availability of RT-PCR tests from nasopharyngeal swabs is still interfering with our ability to accurately classify each morphology as either COVID-19-related or drug-related. The histopathological examination, as well as RT-PCR testing from eye discharge and vesicular fluid bring us closer to an accurate diagnosis.
Alternative etiopathogenetic factors might be involved in the appearance of skin rashes in COVID-19 patients, such as herpesvirus co-infections or re-activations and drug reactions, especially hydroxychloroquine-related. The latter was proved to be more frequent in patients with underlying viral infections. However, vascular rashes, either IFN-I mediated or as a consequence of coagulopathy and alternative complement pathway activation, seem to be linked to the novel coronavirus. Furthermore, conjunctivitis appearing in COVID-19 patients might be highly contagious.
Given the actual pandemic restrictions, face-to-face consultations have been temporarily replaced by telemedicine. The possibility of photo and video sharing, as well as communication via text and voice messages have made the continuation of dermatologic examinations possible. Additionally, treatment adherence has been improved by telemedicine, as shown in a study conducted my Marasca et al. [115].