Cutaneous Manifestations of COVID-19: A Systematic Review


 Background: The COVID-19 pandemic has affected 4.5 million people and killed over 300,000 patients. Although this virus primarily causes respiratory symptoms, an increasing number of cutaneous manifestations associated with this disease have been reported.Objective: The aim of this review was to collate and categorize the dermatologic findings reported in COVID-19 patients and identify specific lesions that may facilitate diagnosis and prognostication.Methods: A systematic review of the PubMed database was conducted on May 14th, 2020 using the search terms “Covid-19 skin,” “Covid-19 rash,” “Covid-19 exanthem,” and “Covid-19 chilblains.” Peer-reviewed publications containing original COVID-19 patient cases and a discussion of the associated cutaneous findings were included in the analysis.Results: The literature search identified 115 records, of which 34 publications describing 996 dermatologic patients were included. Case reports (n=15), case series (n=10) and observational studies (n=7) were the most common publication types. Pseudo-chilblains (PC) was the most frequent lesion identified (40.4% of cases), appearing in young adults (mean age, MA, 23.2 years) after the onset of extracutaneous COVID-19 symptoms (55/100 patients). Erythematous maculopapular rashes (EMR) affected 21.3% of patients, most frequently impacting middle-aged adults (MA 53.2 years) and occurring at the same time as non-cutaneous symptoms (110/187 patients). Vesicular rashes (VR) affected 13.0% of patients, appearing in middle-aged adults (MA 48.3 years) after the onset of other symptoms (52/84 patients). Urticarial rashes (UR) affected 10.9% of patients, appearing in adults (MA 38.3 years) and occurring at the same time as non-cutaneous symptoms (46/78 patients). Vascular rashes resembling livedo/purpura/necrosis (LPN) were uncommon (4% of cases), appearing in elderly patients (MA 77.5 years) and occurring at the same time as non-cutaneous COVID-19 symptoms (18/29 patients). Erythema multiforme-like eruptions (EME), although infrequent (3.7% of cases), affected mostly children (MA 12.2 years).Conclusions: VR may suggest an initial diagnosis of COVID-19, PC may be most appropriate for epidemiological uses, and LPN may be a useful prognostic marker for severe disease. As a potential correlate to disease severity, prognosis, or infectibility, it is critical that all health care professionals be well-versed in these increasingly common cutaneous manifestations of COVID-19.


Introduction
In December 2019, a new infectious pathogen named severe acute respiratory syndrome coronavirus 2 (SARS-CoV-2) was identi ed in Wuhan, China. 1 Transmitted through respiratory droplets, SARS-CoV-2, commonly referred to as coronavirus disease 2019 (COVID- 19), has rapidly spread across the globe to infect 188 countries. The combination of its high infectivity and lengthy asymptomatic latency period has culminated in a global pandemic affecting 4.5 million people and resulting in over 300,000 deaths to date. 2 Although known to primarily cause interstitial pneumonia and respiratory failure, recent reports from around the world have indicated that this novel coronavirus may be associated with speci c cutaneous manifestations. These dermatologic symptoms may be useful in identifying otherwise asymptomatic COVID-19 carriers, which may help slow the transmission of this highly infectious and dangerous virus. As such, a systematic review of peer-reviewed scienti c literature was conducted to collect clinically relevant information on the cutaneous signs and symptoms of patients with COVID-19.

Materials And Methods
To identify all peer-reviewed articles reporting on the cutaneous manifestations associated with COVID-19, a systematic review of the PubMed database was conducted on May 14 th , 2020 using the search terms "Covid-19 skin," "Covid-19 rash," "Covid-19 exanthem," and "Covid-19 chilblains." Records identi ed through the electronic database were initially screened by title and abstract content. The fulltext articles that remained were assessed for eligibility and inclusion in the systematic review. Peer-reviewed publications that contained original COVID-19 patient cases and a discussion of the associated dermatologic ndings were included in the analysis. Review or opinion articles that did not report on original patient cases, observational studies or case reports that did not discuss cutaneous symptoms, and articles that were published prior to the rst COVID-19 case in December 2019 were excluded.
Eligible articles were assessed for study type, location, setting, and Level of Evidence for clinical research. Demographic characteristics such as patient sample sizes, ages, and genders was collected. COVID-19 status and diagnostic modality were also noted. Descriptive characteristics of the cutaneous manifestations were recorded, including: rash type, location, duration, associated symptoms, relation to drug intake, and correlation to onset of other COVID-19 symptoms.

Study characteristics and patient demographics
As depicted in Figure 1, 115 records were initially identi ed in the literature search. After screening for eligibility and inclusion criteria, 34 peer-reviewed publications were ultimately included in the systematic review.  Study and demographic characteristics are summarized in Table 1. The majority of publications were case reports (n = 15), followed by case series (n = 10), observational studies (n = 7), and survey-based studies (n = 2). All studies included were rated as Level 4 or 5 Evidence for clinical research. 37 Most studies originated from Europe (Italy, n = 10; Spain, n = 10; France, n = 7; Belgium, n = 2), followed by the United States (n = 3), the Middle East (Qatar, n = 1; Iran, n = 1), and Southeast Asia (Thailand, n = 1). Patients were identi ed from dermatology clinics (n = 19) and hospitals (n = 15). A total of 996 patients with dermatologic symptoms were included in the analysis, of which 54.3% were female and the mean age was 37.3 years (range 1.0 to 98.0).   9,18,20,[33][34][35][36] These non-laboratory con rmed patients all presented during the COVID-19 outbreak and, for those with a negative COVID-19 PCR result, antibody testing was not available to establish the possibility of prior infection. Case fatality rates ranged from 0.0% to 13.6%. 8,13,30 For studies that examined a larger cohort of COVID-19+ patients (both with cutaneous and extracutaneous symptoms), the prevalence of skin symptoms ranged from 4.9-20.4%. 3,12 3.2 Characteristics of cutaneous symptoms Table 2 summarizes the characteristics of the cutaneous symptoms described in the COVID-19 literature. These dermatologic manifestations were then grouped together into categories using common descriptive terminology and photographic evidence. These categories included: erythema multiforme-like eruptions (EME), erythematous maculopapular rashes (EMR), erythematous rashes not otherwise speci ed (ER-NOS), gurate erythema (FE), vascular rashes within the spectrum of livedo/purpura/necrosis (LPN), pityriasis rosea-like eruptions (PRE), pseudo-chilblains (PC), symmetrical drug-related intertriginous and exural exanthema (SDRIFE), urticarial rashes (UR), vesicular rashes (VR), and other rashes (OR).  (3) After (16) Statistically, PC was associated with greater patient age and EME was associated with more ventrally distributed lesions Fernandez-Nieto et al. 30 VR (24) Head (4) Chest (21) Trunk (14) Arms (8) Legs (10) Palms/soles (2) 10 days Pruritis (20)     Abbreviations: erythematous maculopapular rash (EMR), erythema multiforme-like eruption (EME), erythematous rash not otherwise specified (ER-NOS), figurate erythema (FE), livedo/purpura/necrosis (LPN), pityriasis rosea-like eruption (PRE), pseudo-chilblains (PC), symmetrical drugrelated intertriginous and flexural exanthema (SDRIFE), urticarial rash (UR), vesicular rash (VR), other rash (OR). Table 3, PC was the most common category identi ed (40.4% of all cases), followed by EMR (21.3%), VR (13.0%), UR (10.9%), OR (4.3%, including 1 eruptive cherry angioma, 1 reactivation of HSV-1, and 41 unspeci ed cases of eczematous, angiomatous or annular lesions), LPN (4%), EME (3.7%), and ER-NOS (2.1%). Cutaneous manifestations resembling SDRIFE, PRE and FE were the least common (0.3% combined).   Although not all studies reported on the mean duration of these cutaneous manifestations, most studies described signs and symptoms that resolved within 2 to15 days. One study reported persistent PC that required 2 to 4 weeks for complete resolution. 9 Of the dermatologic ndings that were symptomatic, the majority were pruritic (n = 295), followed by painful (n = 68) or burning (n = 24). Most patients had no new drug intake in the 2 weeks preceding rash onset. Some studies went into more descriptive detail to characterize the skin ndings. Fernandez-Nieto et al. observed two different types of VR: a diffuse polymorphic rash at various stages of evolution (observed in 75% of patients) and a localized monomorphic rash at the same stage of evolution (observed in 25% of patients). 30 The single case of a VR reported by Sachdeva et al. was described as resembling Grover's Disease. 31 Piccolo et al. observed two different lesion patterns for PC: erythematous-edematous types and blistering types. A few articles also described skin biopsy ndings, which almost uniformly revealed a perivascular mononuclear/lymphocytic in ltrate with occasional small vessel thrombosis. 8,15,16,19,22,26,32,36 One study in particular revealed a complement-mediated thrombogenic vasculopathy with deposition of C4d and C5b-9 on histopathology. 5 The association and time relationship between each dermatologic manifestation and the non-cutaneous symptoms of COVID-19 (like fever, cough, shortness of breath, malaise and myalgia) is depicted in Table 3. The appearance of most cutaneous ndings coincided with the onset of other COVID-19 symptoms (46.1% of all cases). Other rashes appeared shortly after (44.3%) or before (9.6%) the onset of non-cutaneous COVID-19 manifestations. EMR were the most likely to present concurrently with other symptoms (58.8%) compared to before (5.9%) or after (35.3%). UR also frequently coincided with the onset of non-cutaneous symptoms (59.0%) compared to before (6.4%) or after (34.6%). The same trend was observed for LPN, with the majority of cases appearing alongside other COVID-19 symptoms (62.1%) compared to before (6.9%) or after (31.0%). Of the cases reported, ER-NOS presented most frequently after other symptoms of COVID-19 (71.4%) rather than before (0%) or at the same time (28.6%). PC also more commonly appeared after (55.0%) compared to before (21.0%) or at the same time (24.0%) as extracutaneous symptoms. VR were also more likely to present after other symptoms (61.9%) compared to before (9.5%) or at the same time (28.6%).

Discussion
The repercussions of the SARS-CoV-2 pandemic is substantial, impacting millions of patients medically, nancially, and socially. Unfortunately, this highly-virulent pathogen is extremely di culty to contain given its prolonged asymptomatic latency period and respiratory droplet transmission pattern. Cutaneous manifestations of COVID-19 may help assist in the identi cation of carriers, enabling health care o cials to implement appropriate measures to stem the spread and, if appropriate, provide earlier COVID-19speci c care to these individuals.
Reports of the dermatologic signs and symptoms associated with COVID-19 is mounting in the literature, but these studies have yet to be comprehensively evaluated. This manuscript presents the most current and extensive systematic review of the COVID-19 scienti c literature to identify and present the cutaneous manifestations of SARS-CoV-2. Thirty-four studies on 996 dermatologic patients from 8 different countries and 4 different continents were included in this report.
The most commonly reported dermatologic ndings included erythema multiforme-like eruptions (EME), erythematous maculopapular rashes (EMR), vascular rashes within the spectrum of livedo/purpura/necrosis (LPN), pseudo-chilblains (PC), urticarial rashes (UR), vesicular rashes (VR), and otherwise unspeci ed erythematous rashes (ER-NOS). Most of these rashes may be attributable to a COVID-19-speci c viral exanthem; however, some dermatologic manifestations may be by-products of the thrombogenic and immune deregulatory effects of SARS-CoV-2, as evidenced by the cutaneous reaction patterns and histopathological ndings. 38 Pseudo-chilblains were the most common type of rash presented in the literature, primarily affecting young adult patients and presenting after the onset of non-cutaneous COVID-19 symptoms. According to Galvan Casas et al., PC were signi cantly associated with younger aged patients and with a milder disease course. Patients with PC typically required less treatments than those with LPN, EMR, UR and VR skin ndings. 13 Of note, these PC lesions were largely unrelated to cold exposure (secondary to social isolation and stay-at-home orders) or a prior history of Raynaud's disease. 9,20,32,33,34,36 As such, these cutaneous ndings may be categorized as COVID-19-induced pseudo-chilblains rather than primary pernio disease. The majority of PC cases were associated with milder disease, resolving disease, or negative laboratory testing. In one study, PC was found in many close patient contacts. 4 Since this rash affected a large number of patients, it may be useful for detecting individuals more likely to unknowingly transmit the disease. As such, pseudochilblains may be most appropriate for epidemiological uses rather than diagnostic applications.
Erythematous maculopapular rashes were the second most common skin manifestation, affecting mostly middle-aged patients and presenting at the same time as other symptoms. Vesicular rashes and urticarial rashes were the next most frequent dermatologic ndings. Both rash types affected adults, but VR typically presented after the onset of extracutaneous COVID-19 manifestations while UR occurred simultaneously. UR and EMR eruptions are often drug-induced, which lessens the ability of these cutaneous reactions to serve as a COVID-19 diagnostic marker. Vesicular lesions, however, tend to be more speci c to viral exanthemas. As such, VR lesions may be a more useful COVID-19 diagnostic tool. Unfortunately, the VR pattern varies amongst COVID-19 studies, ranging from diffuse polymorphic to localized monomorphic distributions. Further studies evaluating the VR form most strongly associated with COVID-19 is warranted prior to determining the diagnostic utility of this particular cutaneous lesion.
Vascular lesions belonging to the spectrum of livedo/purpura/necrosis, while relatively uncommon, were primarily seen in the elderly at the onset of non-dermatologic COVID-19 symptoms. According to Galvan Casas et al., LPN lesions were signi cantly correlated with more advanced age and/or severe symptomatology. Unsurprisingly, these vascular rashes were then associated with higher rates of hospital admission and mechanical ventilation compared to EMR, UR, VR and PC. 13 Given their correlation to disease severity, these LPN ndings might represent a COVID-19-speci c complication in which the virus-induced pro-thrombotic state provokes vascular occlusion and ischemia. 38 Consequently, these rashes may be a useful prognostic marker that can help guide medical management.
Erythema multiforme-like eruptions were infrequent and typically occurred in young patients, particularly children. OR (including one HSV-1 reactivation and one eruptive cherry angioma), EME, PRE and SDRIFE were also rarely observed. As such, these cutaneous manifestations may not be directly correlated to the COVID-19 virus. However, the true incidence of EME and PRE may be under-reported, as some studies grouped many unspeci ed rashes of annular or angiomatous appearance into the OR category. 36 The frequency of PRE in particular warrants further investigation, as this rash has previously been linked to other viruses. 39 As with any cutaneous eruption, one must consider that it may have been triggered by a medication. Drugs used to treat COVID-19, such as antimalarials, azithromycin, remdesivir, antiretrovirals, steroids, and biologics, are known to cause acute urticaria, vasculitis and other pruritic lesions. The rashes included in this systematic review that coincided with recent medication intake were determined to be nondrug induced, as most authors reported that, for each questionable patient, the administered drug had either previously been taken without complication or was not known to cause the speci c rashes observed. Of note, some COVID-19 patients with previous skin conditions like rosacea, acne, eczema and atopic dermatitis experienced a are during the course of their disease. 41 Finally, the case fatality rate reported by the studies included in this analysis ranged from 0.0% to 13.6%. This span encompasses the most recent and accurate estimation of overall COVID-19 mortality rate of 6.8%. 2 As a rapidly evolving situation, COVID-19 peer-reviewed reports are rarely in the form of high-powered, strictly controlled clinical trials. As such, one of the limitations of this study was the inability to collect complete and standardized data sets to allow for in-depth comparisons between rash subgroups. Additionally, not all studies included in this analysis reported on patients with a laboratory-con rmed diagnosis of COVID-19; however, this may also be attributable to the unprecedented turmoil induced by this novel coronavirus, as testing is often limited or inaccessible.

Conclusion
A study of 552 hospitals in mainland China initially indicated a 0.2% prevalence of COVID-19 skin symptoms. 40 However, the prevalence suggested by this systematic analysis implies that the true rate is much greater, affecting up to 20.4% of COVID-19 patients. As a potential correlate to disease severity, prognosis, or infectibility, it is critical that all health care professionals be well-versed in these increasingly common cutaneous manifestations of COVID-19. Hand surgeons and podiatrist must be even more aware of these lesions, as they most often appear on the hands and feet. Additional standardized studies of the COVID-19 rashes are warranted to further establish the diagnostic validity and utility of these visible ndings.

Declarations
Funding sources: The authors have no funding sources to disclose.
Con icts of interest/Competing interests: The authors have no con icts of interest or other competing interests to disclose.
Ethics approval: This article does not contain studies on human or animal subjects that require approval by an ethics board.
Consent to participate/Consent for publication: This article does not discuss any individuals or material that would require consent for participation or publication.

Figure 1
Flow Diagram of the Search Strategy