Atypical and Typical Presentation of Erythema Nodosum: Clinical Differences in Treatment and Outcome

Erythema nodosum (EN) is the most common form of panniculitis and typically presents itself with painful subcutaneous nodules [1, 2]. EN nodules are typically located on the lower legs and can result in considerable disease burden. While EN can occur in patients of all ages, the highest incidence occurs in patients between 20 and 40 years [3, 4]. EN is often idiopathic. The etiology cannot be determined in 32.5–74% of cases [2, 5, 6]. Many triggers of EN have been suggested in the literature. The most common trigger factors include bacterial as well as viral infections, such as streptococcal and non-streptococcal upper airway infections and tuberculosis. Other conditions such as sarcoidosis, inflammatory bowel disease, Behcet’s disease as well as pregnancy and – seldom – Hodgkin’s disease are also associated with EN [1, 2, 6, 7]. Oral contraceptives, penicillin, and sulfonamides are among the most common pharmaceutical triggers. Importantly, EN can occur before the onset of systemic diseases. Symptoms that can appear before or during disease onset are fever, malaise, sore throat, coughing, and arthritis. Further findings include leukocytosis, elevated erythrocyte sedimentation rate and CRP [6, 8, 9].

The pathogenesis of EN is controversial. It is believed to be a hypersensitivity reaction to various antigens and subsequent involvement of neutrophil granulocytes, leading to the activation of reactive oxygen intermediates [1, 10]. In the septal venules, immune complex deposition can be found, and tumor necrosis factor (TNF)-alpha production is increased [10]. While female patients are more likely to be affected than males [5, 11, 12], in pediatric patients, female predominance does not seem to be as prominent [7, 13]. An effect of estrogen on EN has therefore been discussed [11].

Why EN predominantly affects the anterior lower legs remains unclear. It has been proposed that the absence of muscle pumps, combined with diminished arterial blood supply, may favor the development of EN within this region [14]. However, EN lesions may also appear in other atypical locations, including the upper legs, arms, trunk, and face, which was so far only reported in case reports. Furthermore, no data exist on whether EN in atypical sites differs from typical EN in terms of gender involvement, age of onset, etiology, or prognosis. In this study, we aimed to investigate clinical differences of EN cases with atypical locations compared to those occurring on the lower legs.

After obtaining ethical approval (BASEC No. ID 2018-00854), a search in the hospital database for patients with “erythema nodosum” in their medical history was performed. The database was only screened for patients above 18 years of age. Then, all patients with EN as a diagnosis or differential diagnosis who were treated at the University Hospital Zurich between January 1, 2008, to December 31, 2017, were screened (n = 202) (Fig. 1). Ninety-seven patients then had to be excluded because the EN diagnosis was discarded after further diagnostics such as biopsy, new symptoms in the course of the disease not matching EN, or reevaluation by a senior dermatologist (Fig. 1).

The included 105 patients all had a skin status in their medical history and therefore could be considered for further analysis. The primary objective was to investigate possible differences between typical and atypical EN. Typical EN was defined as affecting only the lower legs. Atypical EN was defined as having lesions on the upper legs, trunk, arms, or face, only or additionally to lesions on the lower legs. In the latter group, only the atypical lesions were observed. Biopsies were only taken if clinically indicated. Overall, 45 patients (43%) had a biopsy taken. The medical data of the patients was assessed for age, gender, dermatologic history, time until first medical consultation, time to diagnosis and until remission, recurrent course of disease, as well as therapy regimen. Recurrent EN was defined as a new episode of EN 1 month after full remission of a previous episode. For the therapy regimen, we analyzed what drugs or interventions were prescribed. The first therapy prescribed was counted as first-line therapy. If at a later point a new therapy was prescribed, this was counted as second-line therapy. A third adjustment or change of therapy in the course of disease was counted as third-line therapy.

For statistical analysis, SPSS^® version 29.0 was used. To analyze categorical data, the χ² test was used, while the Mann-Whitney U test was used for numerical data since normality distribution was not fulfilled by our data set. Threshold for significance, alpha, was set at 0.05.

Sixty-six percent of the study population (n = 70) had typical EN, while 33% (n = 35) had atypical EN. In the atypical EN group, 66% (n = 23) had painful nodules on their upper legs, 57% (n = 20) had nodules on their arms, and 11% (n = 4) had lesions on their trunk. No patients with facial EN were documented. Overall, 94% (n = 33) of patients with atypical lesions had additional nodules on their lower legs (Fig. 2). Accordingly, only 2 patients of the atypical EN cohort did not have lesions on their lower legs.

The study population was predominantly female. Recurring disease occurred in slightly more often in atypical EN cases; however, the difference was not statistically significant (Table 1). In the atypical EN group, the mean age at first manifestation was slightly higher, but the difference was not statistically significant. While mean time from first presentation until diagnosis was 1.14 ± 4.68 months in the typical EN group, the time until diagnosis in the atypical EN group was 0.46 ± 1.14 months, being significantly shorter (p = 0.034). Also, average healing time was slightly shorter in the atypical EN group; however, this difference was not significant (Table 1). A total of 37% (n = 26) of patients in the typical EN group and 54% (n = 19) in the atypical group had a biopsy taken.

Symptoms were distributed similarly in both typical and atypical EN. All patients in the atypical EN group reported suffering from pain, while nearly all patients (97%) in the typical EN group reported pain. Accordingly, the difference was not statistically significant (p = 0.313, χ²). Local warmth was also equally common in both groups with 47% (n = 33) in the typical EN group and 49% (n = 17) in the atypical EN group. Patients in the atypical EN group were affected by fever slightly more often than patients in the typical EN group (14% [n = 5] versus 9% [n = 6]); however, the difference was not statistically significant (p = 0.367). Swelling of affected limbs was reported more frequently in the atypical EN group (37% [n = 13]), compared to the typical EN group (31% [n = 22]); however, the difference was also not statistically significant (p = 0.558).

Sarcoidosis was the only etiologic factor that differed significantly between the typical and atypical EN groups (Table 2). It was significantly more common in the atypical EN group, accounting for 23% (n = 8) of cases, compared to 9% (n = 6) in the typical EN group. Out of the 16 sarcoidosis patients, 6 patients had Lofgren syndrome. While Lofgren syndrome also was more frequent in the atypical group (8% vs. 4%), this was not statistically significant. Streptococcal infections were equally distributed in typical and atypical EN. Tuberculosis was only found in the typical EN group; however, the difference was not statistically significant. Other infectious diseases were also found more frequently in the typical EN group with 24% (n = 17), compared to 14% (n = 5) in the atypical group; however, the difference was again not statistically significant. Pregnancy was suggested as an etiologic factor in 4% (n = 3) of typical EN cases and in 3% (n = 1) of atypical cases. Crohn’s disease and colitis ulcerosa were also similarly distributed in both groups. Further, autoimmune diseases, including Sjogren’s syndrome and Graves’ disease, were rare in both groups. Cancer was described as a possible etiologic factor in one case in the typical EN group only. Two cases of atypical EN were associated with drugs. Cases were termed idiopathic in 31% (n = 22) of typical and 34% (n = 12) of atypical EN patients (Table 2; Fig. 3).

In the atypical EN group, the prescribed first-line therapy was as follows: in 80% (n = 28), nonsteroidal anti-inflammatory drugs (NSAIDs); in 60% (n = 21), topical steroids; in 29% (n = 10), compression therapy; and in 9% (n = 3), opioids. In comparison, in the typical EN group, 83% (n = 58) received NSAIDs; 67% (n = 47), topical steroids; and 60% (n = 42), compression as first-line therapy. Oral steroids were given to 26% (n = 9) of patients in the atypical EN group as first-line therapy, compared to 19% (n = 13) in the typical EN group. Additionally, in one case, colchicine was prescribed in the atypical EN group. In the typical EN group, potassium iodatum was prescribed in one case and hydroxychloroquine in another case.

Second-line therapy was more frequently used in the atypical EN group with 40% (n = 14), compared to 30% (n = 21) in the typical EN group. In the atypical EN cohort, second-line therapy included oral steroids in 20% (n = 7) of cases. Furthermore, colchicine was prescribed in one case and diaminodiphenyl sulfone (dapsone) in two cases in the atypical EN cohort. In the typical EN cohort, 10% (n = 7) of patients received oral steroids, 6% (n = 4) received potassium iodatum, 3% (n = 2) received methotrexate, and in one case, colchicine was prescribed. Biologics were prescribed in the atypical EN cohort only as 5 patients (14%) received adalimumab treatment. A total of 6% of both patient cohorts received third-line therapy.

This study is to our knowledge the first study, apart from case studies, focusing on EN with atypical location. We were surprised to find that time until diagnosis was significantly shorter in our atypical EN group. A possible explanation for this could be that general practitioners referred these cases faster to a specialist than typical EN cases. Regarding etiology, only subtle differences were found, except for sarcoidosis, which was significantly more frequent in the atypical EN group. No statistically significant difference was found regarding symptoms. When analyzing the therapies used, we found that only patients in the atypical EN group received biologics.

In our study, 19% (n = 20) of patients suffered from lesions on the arms, 4% (n = 4) on the torso, and 22% (n = 23) on the upper legs. While EN has also been described in the face [15], we did not encounter a patient with facial lesions. Due to patients suffering from lesions on multiple locations, overall, only 35 of 105 patients (33%) showed EN on atypical locations. In contrast, other authors reported that only 7% of patients had appearances of EN on atypical sites [16]. This difference may partly be caused by a selection bias as due to a general practitioner-centered healthcare model, common in Switzerland, primarily difficult and severe cases are sent to a university hospital.

In our study, even though more areas were affected in atypical EN, prognosis for typical and atypical EN cases seems to be similar. Accordingly, mean time until resolution did only differ slightly in cases that affected only the shins compared to cases that further affected atypical areas.

Compared to other studies, time until remission was longer in our study, with 3.61 months in the typical and 3.05 months in the atypical EN group. In other studies, the mean duration until remission was reported to only be 11.5 days and 18 days, respectively [7, 13]. This could be explained by the fact that hard-to-treat cases are more likely to be sent to a university hospital, due to the already mentioned general practitioner-centered healthcare model in Switzerland. This notion is further supported by the high number of reoccurring EN cases in our study (30 out of 105 cases, 40%). A comparable study with a similar sample size in France only registered eight recurring cases in a study population of 129 patients (6%) [11]. The mentioned study further analyzed a possible correlation between etiology and clinical outcome but did not find significant correlations [11]. Meanwhile, a smaller study in Turkey found recurring EN in 19 out of 50 cases (38%), with a statistically significant correlation of idiopathic EN to a recurrent course of disease [6]. On the contrary, a different study suggests that recurrent EN may be a warning sign of tuberculous lymphadenitis [17]. Which factors truly influence the prognosis of EN thus remains a topic of debate and demands for further research. In our study, we could not detect a correlation between recurrent disease and atypical or typical EN location.

Regarding treatment, in both typical and atypical EN, nearly all our patients received first-line analgesic therapy. This is not surprising as almost all patients reported pain (97% of atypical EN patients, 100% of typical EN patients). Noteworthily, however, in the atypical study group, 9% of patients received opioids as first-line analgesic therapy. This indicates that the level of pain may be higher in these patients. While other studies demonstrated that application of NSAIDs generally leads to improvement of symptoms within days, this does not seem to apply to our study population [6]. Even though NSAIDs were prescribed in nearly all cases as first-line therapy (83% in the typical EN and 80% in the atypical EN group), 30% of patients in the typical EN group and 40% of patients in the atypical EN group required further second-line therapy. This suggests that for a considerable number of patients admitted to specialized clinics, first-line therapy is not sufficient. In the atypical EN group, 14% received the antibody adalimumab as treatment, in contrast to none in the typical EN group. Adalimumab has previously been described as an effective treatment in EN [18, 19]. However, this is still an off-label use and combined with the high cost usually reserved for harder-to-treat patients. The observed use of opioids in the atypical EN group only, as well as the fact that the atypical EN group reported symptoms more frequently in all categories, compared to the typical EN group, could indicate a generally higher burden of disease. This could thus justify the more expensive treatment choice of adalimumab only used for atypical EN, as well as the more frequent use of systemic steroids (26% in the atypical EN cohort, 19% in the typical EN cohort). The more aggressive treatments chosen in the atypical EN cohort could also explain the average shorter observed time until remission in the atypical EN group compared to the typical EN group.

This is the first study focusing on atypical EN, and it is one of the largest studies on EN overall. However, given the small data size, it would require larger cohorts to detect differences between cohorts with adequate statistical significance. The main study limitation is the rather small cohort size. Even though it represents the most common panniculitis, there are no studies with more than 129 patients [11]. This may be explained by the fact that most general practitioners are familiar with this disease, and thus, patients do not reach medical research hospitals often enough, to obtain relevant sample sizes.

Our results suggest that only minor differences between EN on typical versus EN on atypical location exist. In atypical EN, time to diagnosis was significantly shorter and sarcoidosis as an etiologic factor significantly more frequent. Even though prognosis for typical and atypical EN seems to be similar, only patients in the atypical EN group were prescribes biologics. Whether these differences are of clinical relevance is debatable. Due to the limited sample size, it is possible that other, more relevant differences exist but could not be detected. Further research maybe therefore needed to focus on collaboration with general practitioners and dermatologist in private practices to overcome the problem of small sample sizes in EN studies.

This is the first study on erythema nodosum in atypical locations.

This study protocol was reviewed and approved by the Cantonal Ethics Committee of the canton of Zurich, approval number ID 2018-00854. Written consent of participants was not required according to the Ethics Committee of the canton of Zurich.

F.A. has served as an advisor and has held presentations for AbbVie, Almirall, Amgen (Celgene), Bristol-Myers Squibb, Eli Lilly, Galderma, Leo Pharma, Janssen-Cilag, MSD, Novartis, Roche, Sanofi, and UCB. J.-T.M. has served as advisor and/or received speaking fees and/or participated in clinical trials sponsored by AbbVie, Almirall, Amgen, BMS, Celgene, Eli Lilly, LEO Pharma, Janssen-Cilag, MSD, Novartis, Pfizer, Pierre Fabre, Roche, Sanofi, and UCB. L.V.M. has served as advisor and/or received speaking fees and/or participated in clinical trials sponsored by Almirall, Amgen, Eli Lilly, MSD, Novartis, Pierre Fabre, Roche, and Sanofi outside of the current work. N.M., T.N., F.F., and T.K. have no conflict of interest.

No funding was obtained for this project.

Nina Meienberger: data collection, statistical analysis, and writing of the manuscript. Julia-Tatjana Maul, Fabienne Fröhlich, Lara Valeska Maul, and Thomas Kündig: writing and editing of the manuscript. Thierry Nordmann: concept of the study, writing and editing of the manuscript. Florian Anzengruber: ethical approval, concept of study, and writing and editing of the manuscript.

Thierry Nordmann and Florian Anzengruber: last co-authorship.

All data generated or analyzed during this study are included in this article and its online supplementary material. Further inquiries can be directed to the corresponding author.