Article, Dermatology

Herpes associated erythema multiforme: A retrospective study

a b s t r a c t

Background: Erythema multiforme (EM), an acute dermatologic condition frequently encountered in the Emer- gency Department, classically presents with a targetoid rash. We reviewed all recent EM cases seen at the LAC- USC county hospital in order to ascertain the proportion of Herpes associated EM (HAEM) cases and to inform the diagnostic workup of these patients.

Methods: ICD-9 and ICD-10 codes were used to extract a list of EM cases at our institution from 2013 to 2019. Two non-blinded abstractors screened records to confirm an EM diagnosis and entered patient data utilizing a stan- dardized data abstraction form. Cohen’s kappa statistic was used to measure inter-rater reliability on various var- iables. Kappa (?) values ranged from 0.803 to 1.0.

Results: 70 pediatric and 56 adult EM patients were included in the study. A likely etiology was ascribed to 63% of pediatric and adult EM cases. Pediatric EM was most commonly attributed to upper respiratory infection (URI) (n = 23; 33%), Mycoplasma pneumoniae infection (n = 5; 7%), and medications (n = 4; 6%). Adult EM was most commonly attributed to HSV infection (n = 11; 20%), medications (n = 5; 9%), URIs (n = 4; 7%), and other infections (n = 4; 7%).

Conclusion: HSV-1/2 Serologic testing should be considered in most EM patients to potentially prevent repeated ED visits. In EM cases not clearly attributable to herpes or drug exposure, physicians can consider further workup: Mycoplasma serology, nasal PCR, and a respiratory viral panel in pediatric patients. Identification of an etiologic cause may suggest a different treatment approach and prevent mislabeling of medication allergies in patient charts.


Erythema multiforme (EM) is an acute dermatologic condition clas- sically presenting with a targetoid rash affecting the extremities or trunk. herpes simplex virus (HSV), Mycoplasma pneumoniae, and drug reactions are the three most frequently implicated causes [1]. In many studies, herpes is the most commonly reported etiology, estimated to cause 15-63% of EM cases in the general population and 18% of cases in children with non-recurrent disease [1-2]. Heterogeneity in patient age, disease recurrence, and mucosal involvement may contribute to the wide range in reported percentages of herpes associated EM (HAEM). We conducted a retrospective review of all EM cases at the LAC-USC County Hospital from 2013 to 2019 with the aim of

Abbreviations: EM, erythema multiforme; HAEM, herpes associated erythema multiforme; MIRM, Mycoplasma induced rash and mucositis; TEN, toxic epidermal necrolysis; RIME, reactive infectious mucoCutaneous eruption.

* Corresponding author at: University of Southern California Keck School of Medicine, 1441 Eastlake Avenue, Los Angeles, CA 90033, United States of America.

E-mail address: [email protected] (M. Hao).

determining the exact frequency of different causes of EM in our ethni- cally diverse patient population.


We used ICD-9 and ICD-10 codes to extract patient records dated 2013 to 2019 from the LAC-USC County Hospital database. Patient re- cords met study inclusion criteria if they were billed with ICD-9 code

695.10 or ICD-10 code L51.9 for EM. During the data abstraction process, patient records were manually checked and excluded if the patient note stated that another dermatologic condition was a more likely diagnosis than EM (i.e., urticaria multiforme). One-hundred twenty-six records were included in the final analysis.

A standardized data abstraction form (Supplementary Materials)

was created using REDCap Software. Variables included patient demo- graphics such as age at disease onset, current patient age, and sex. A sec- ond category assessed the clinical characteristics including disease duration, presence of typical or atypical target lesions, preceding illness, recent new drug ingestion, and mucous membrane involvement. We also collected data on clinical management, i.e., hospitalization status, diagnostic studies, and therapeutic outcomes. Finally, we abstracted 0735-6757/(C) 2020 Published by Elsevier Inc.

Pediatric cases“>2761.e2 M. Hao et al. / American Journal of Emergency Medicine 38 (2020) 2761.e12761.e3

data regarding noted associations in the provider’s patient note: signs or symptoms of a concomitant infection (HSV, Mycoplasma pneumoniae, or an upper respiratory infection) or a documented eruption occurring after a specific event (administration of a new medication or vaccine). Two non-blinded abstractors extracted data from patient charts.

Thirty eight out of 126 patient records were randomly selected, and rat- ings from these records were used to determine inter-rater reliability. Cohen’s kappa statistic was calculated for specific variables: mucous membrane involvement, recent new drug ingestion, and noted associa- tions. The lead abstractor audited all entries for completeness. Excel was used to analyze the data, and results were reported as proportions of pa- tients who met specific criteria.


We identified 126 EM cases out of 293 extracted patient records billed with an ICD-9 or ICD-10 code for EM. 70 patients were pediatric (under 18 years old), and 56 patients were adults.

  • Pediatric cases

The average pediatric patient age +- standard deviation was 5.3 +-

5.0 years, and 42.9% of pediatric patients were female. A likely etiology was identified in 63% of cases. The most commonly implicated causes of EM were upper respiratory infection (URI) (n = 23; 33%), Myco- plasma pneumoniae infection (n = 5; 7%), and medications (n = 4; 6%) (Table 1). Only 2 HAEM cases (n = 2) were reported, although no- tably only 7.1% of pediatric patients were tested for HSV (culture, PCR, or serology). 1 out of 2 patients with suspected HAEM had serologic testing, which confirmed positive HSV-1 IgM and IgG antibodies.

Eight patients (11%) were diagnosed with EM due to a combination

of factors, i.e. medications and infection. The cause of EM was undeter- mined in 26 children (37%). Only one patient had recurrent disease to our knowledge, although this may be limited based on follow up.

Seven (10%) pediatric patients had mucosal (oral, ocular or genital) involvement with oral involvement being the most common. Interest-

  • Adult cases

The average adult patient age +- standard deviation was 38.0 years

+-15.5, and 62.5% of adult patients were female. An etiologic association was also reported in 63% of adult EM cases. The most common causes were HSV infection (n = 11; 20%) and medications (n = 5; 9%); URIs and other infections each accounted for 4 more cases (7% each) (Table 2). Notably, only 30.4% of adult EM patients were worked up for HSV. IgG antibodies were ordered without IgM antibodies 60% of the time. diagnostic yield was b50% for all HSV studies, except for HSV1 IgG titers (n = 9 out of 11; 82%).

A combined etiology was ascribed to 8 patients (14%). In 21 adults (38%), the cause of EM was undetermined. Four patients (7%) had recur- rent disease. Twenty-three percent (n = 13) of our adult patients exhib- ited mucosal involvement, of which 77% (n = 10) included the oral mucosa. Of mucosa involved cases, 46% were diagnosed as HAEM. Twenty-seven patients received systemic steroids, and eighteen re- ceived antiviral therapy of which 66% and 61% of patients exhibited a partial or complete response, respectively. As EM often spontaneously resolves, this response rate may be falsely elevated.

  • Inter-rater reliability

Cohen’s kappa statistic was calculated using paired ratings from 38 patient records. Kappa (?) values were found to be 0.803 for recent new drug ingestion, 0.907 for presence of mucous membrane involve- ment, and 1.0 for each of the seven EM etiologic associations included in this study (HSV infection, Mycoplasma pneumoniae infection, Otitis media, medication, vaccination, menses, other infection, and idio- pathic). Based on these values, inter-rater reliability can be interpreted to be substantial [3].

Table 2

Breakdown of adult erythema multiforme cases by etiology.

ingly, mucosal involvement was associated with M. pneumoniae infec-

tion 57% (n = 4) of the time; this is likely due to the correct diagnosis

Etiology Cases



being a distinct diagnostic entity separate from HAEM termed Myco- plasma induced rash and mucositis (MIRM).

Table 1

Breakdown of pediatric erythema multiforme cases by etiology.

herpes simplex virus 11 HSV-1 (5)

HSV-1/2 (1)

NOSa [5]

Herpes simplex virus + medication 2 Augmentin, HSV-1 (1)

Plaquenil, HSV-1 (1)

Medication 5 Amitryptiline [1] Ampicillin, Tamiflu [1]

Etiology Cases


Upper respiratory infection 23


Upper respiratory infection 4

Penicillin [1]

Clindamycin [1]

Plaquenil [1]

Upper respiratory infection + medication 3 Amoxicillin [2]

Ibuprofen [1]

Upper respiratory infection

+ medication

2 Amoxicillin [1]

Dextromethorphan [1]

Mycoplasma Pneumoniae 5 IgM+ [1]

IgG+ [1]

NOSa [3]

Mycoplasma Pneumoniae + Medication 1 Lamictal [1]

Medication 4 Amoxicillin [3]

Augmentin [1]

Herpes simplex virus 2 HSV-1 (2)

Other infection 2 Gastroenteritis


Other infection 4 Unspecified Preceding Illness [1] Tick Bite [1]

HCV [1]

Pulmonary Aspergillosis, Pneumocystis Pneumonia, HIV/AIDS [1]

Other infection + medication 1 HBV/HCV & ?Med [1]

Mycoplasma Pneumoniae 2 IgG+ [1]

NOSa [1]

Upper respiratory infection + otitis media

+ medication

Viral Prodrome [1]

2 Amoxicillin [1] Cephalosporin [1]

Mycoplasma Pneumoniae

+ medication Herpes simplex virus

+ Mycoplasma Pneumoniae

+ medication

2 Bactrim [1]

Lamotrigine, Risperidone [1]

1 ?Med, HSV-1 (1)

Otitis media + medication 2 ?Med [1] Augmentin [1]

Upper respiratory infection

+ other infection

1 HIV [1]

Undetermined 26 No Workup (24)

Workup [2]

a NOS: Not otherwise specified; diagnosis based on clinical grounds alone, without confirmatory laboratory testing or despite negative laboratory testing.

Undetermined 21 No Workup (20)

Workup [1]

a NOS: Not otherwise specified; diagnosis based on clinical grounds alone, without confirmatory laboratory testing or despite negative laboratory testing.

M. Hao et al. / American Journal of Emergency Medicine 38 (2020) 2761.e12761.e3 2761.e3


In any case of EM with mucosal involvement (but particularly in children), Mycoplasma status should be ascertained. MIRM is marked by severe mucositis involving at least two mucous membranes and var- iable skin involvement, features similar to those seen in EM major and drug-induced toxic epidermal necrolysis (TEN). Meyer et al. demon- strated that pediatric patients presenting with mucocutaneous erup- tions, URI signs, and a positive M. pneumoniae PCR are very likely to have MIRM [4]. Recognition of MIRM is critical to prevent the mislabeling of drug allergy in patients’ medical records and consider- ably lowers concern for life-threatening TEN. Other pathogens can cause a similar clinical presentation to MIRM, including Chlamydophila pneumoniae, human metapneumovirus, human parainfluenzavirus 2, rhinovirus, enterovirus, and influenza B virus. As a group including MIRM, these are termed RIME (reactive infectious mucocutaneous eruption) [5]. Clinicians should consider these other RIME-associated pathogens if M. pneumoniae testing is negative as well as a full respira- tory virus panel. RIME is a distinct entity from EM major, the latter of which is characterized by typical target lesions. In our cohort, Myco- plasma infection was infrequently reported. However, 37% or 47 EM cases in the current study had no known cause. Of these cases, only 6% or 3 cases had any workup done. Thus, it is probable that

M. pneumoniae and HSV along with other infections were under- reported as etiologies for patients’ rashes due to lack of testing.

This study found that URIs far outranked HSV as an etiologic cause of pediatric EM (33% vs. 3%), although in the ED most clinicians were not checking HSV status in pediatric patients, and 37% of our pediatric EM patients had no known cause. In contrast, in adults where HSV status was more readily ascertained, HAEM accounted for the majority of cases with known cause (20%). Although EM is not often an emergent diagnosis, in order to accurately ascribe causality for the eruption and potentially prevent further visits to the ED and over-utilization of Healthcare resources, it is prudent for ER physicians to consider ordering both IgG and IgM HSV1/ 2 antibodies in any patient with EM. In addi- tion, it may be helpful in future management of these patients for ED doctors and/or specialists to consider the following work-up in cases of EM not clearly attributable to herpes or drug exposure: Mycoplasma IgM, IgG, and nasal PCR, along with a respiratory viral panel in pediatric patients.

Funding sources


CRediT authorship contribution statement

Michelle Hao:Methodology, Formal analysis, Investigation, Writing

– original draft, Visualization.Peter Zang:Software, Methodology, For- mal analysis, Investigation, Writing – review & editing.Melanie Miller: Conceptualization, Methodology, Writing – review & editing.Lauren Cutler:Writing – review & editing.Scott Worswick:Conceptualization, Methodology, Writing – review & editing, Supervision, Project administration.

Declaration of competing interest

None declared.

Appendix A. Supplementary data

Supplementary data to this article can be found online at https://doi. org/10.1016/j.ajem.2020.05.084.


  1. Clark Huff J, Weston WL, Tonnesen MG. Erythema multiforme: a critical review of characteristics, diagnostic criteria, and causes. J Am Acad Dermatol. 1983;8(6):763-5.
  2. Zoghaib S, Kechichian E, Souaid K, Soutou B, Helou J, Tomb R. Triggers, clinical mani- festations, and management of pediatric erythema multiforme: a systematic review. J Am Acad Dermatol. 2019;81(3):813-22.
  3. McHugh ML. Interrater reliability: the kappa statistic. Biochem Med (Zagreb). 2012; 22(3):276-82.
  4. Meyer Sauteur PM, Theiler M, Buettcher M, Seiler M, Weibel L, Berger C. Frequency and clinical presentation of mucocutaneous disease due to Mycoplasma pneumoniae infection in children with community-acquired pneumonia. JAMA Dermatol. 2020; 156(2):144-50.
  5. Ramien ML, Bruckner AL. Mucocutaneous eruptions in acutely ill pediatric patients- think of mycoplasma pneumoniae (and other infections) first. JAMA Dermatol. 2020;156(2):124-5.