Introduction

Acute hemorrhagic edema of infancy (AHEI) is a cutaneous leukocytoclastic vasculitis usually affecting children under 3 years old [1,2,3]. The clinical presentation is often alarming with a sudden appearance of large targetoid purpuric lesions, typically on the face and extremities, associated with painful edema and fever in a previously healthy child [3, 4]. The diagnosis is usually based on clinical manifestations and can be supported by a skin biopsy when appropriate [5]. AHEI is a self-limited disease with spontaneous resolution within a few weeks. Possible triggers include infectious agents, especially viruses [3, 6].

AHEI has been documented in small retrospective case reports [3, 7,8,9,10,11,12,13,14,15,16] and in one prospective case series including 18 patients [17]. This uncommon disorder is often underrecognized and is frequently confused with IgA vasculitis (IgAV) [3, 5, 8, 9, 18], purpura fulminans [7, 19, 20], or urticaria multiforme [2, 3, 5, 17, 21, 22]. In 1996, Krause et al. suggested non-validated clinical criteria for the diagnosis of AHEI as follows: (1) age < 2 years, (2) purpuric or ecchymotic target-like skin lesions with edema on the face, ears, and extremities, (3) lack of visceral involvement, and (4) spontaneous recovery within a few days or weeks [23]. Appropriate diagnosis is important to distinguish this benign disease from more serious diseases to avoid superfluous investigations and treatments, iatrogenic harm, and unnecessary long-term follow-up.

The main objective of this study was to develop a diagnostic consensus for AHEI based on the opinions of expert dermatologists. The secondary objective was to compare the characteristics of AHEI classified as probable, doubtful, and unclear.

Materials and methods

Study design and setting

This study was a French multicenter observational study that retrospectively reviewed all cases of AHEI diagnosed between 1996 and 2021 in 22 centers.

Participants

Children under 3 years old with a diagnosis of AHEI were included. The diagnosis was made by hospital physicians (pediatric dermatologists, pediatric rheumatologists, or pediatricians).

Data collection

A questionnaire was sent to members of the Research Group of the French Society of Pediatric Dermatology, members of the French Society for Pediatric Rheumatology and Internal medicine, and pediatricians from five pediatric departments. The providers collected the data of all patients from their hospital centers diagnosed with AHEI during the study period, including demographic data, clinical characteristics, laboratory results, treatments, and follow-up data. Clinical photographs were collected. We excluded patients with missing data for age, clinical features, investigations, and clinical photographs. This non-interventional research was conducted in accordance with the French Data Protection Agency.

Outcomes

The primary outcome measure was the level of agreement among the panel of three experienced experts (pediatric dermatologists) from hospital tertiary centers (authors AM, CB, and MP) on the diagnosis of AHEI in previously diagnosed patients. All cases were reviewed independently. Data and photographs collected from the questionnaires were used to classify each patient into probable, doubtful, or unclear AHEI. The patient was classified into one of these three categories when at least two of the three experts agreed on the same classification. In cases of disagreement or incomplete data, the cases were classified as unclear AHEI. The final classification determined by the consensus of experts was considered the gold standard. Patients with an unclear diagnosis were excluded from further analysis. Because of no gold standard for diagnosing the disease (which is the rationale for this study), experts were asked to perform diagnosis based on history, photographs when available, clinical data, and complementary exams, with no pre-established diagnostic criteria. The clinical criteria suggested by Krause et al. [23] were known by the experts but were not used to classify AHEI. Indeed, these criteria are not sensitive and would have excluded potential patients with AHEI. The secondary outcome was to compare demographic characteristics and clinical manifestations between the children with probable and doubtful AHEI to highlight clinical signs that were either suggestive of or against the diagnosis of AHEI.

Statistical analyses

Descriptive statistics are expressed with mean (range) or median (Q1–Q3) for quantitative data and number (%) for categorical data. Chi-square or Fischer test was used to analyze categorical outcomes. A Wilcoxon test was used to analyze continuous outcomes. The Fleiss kappa index was calculated to evaluate the degree of agreement among the experts. All tests were two-tailed, and p values less than 0.05 were considered significant. No methods were used for missing data. Statistical analyses involved use of R 3.5.1 (R Foundation for Statistical Computing, Vienna, Austria).

Results

Participants and experts

The study screened 85 children, of which 16 were excluded (one duplicate patient and 15 with missing data). Figure 1 shows a flowchart of the process of patient inclusion. Overall, the data from 69 children were evaluated by the experts.

Fig. 1
figure 1

Flow diagram showing the selection of cases of acute hemorrhagic edema of infancy reported by pediatricians, pediatric dermatologists, and pediatric rheumatologists. GRSFDP: Research Group of the French Society of Pediatric Dermatology, SOFREMIP: French Society for Pediatric Rheumatology and Internal medicine

Full size image

The experts classified the patients into probable AHEI in 40 cases (58%), doubtful AHEI in 22 cases (32%), and unclear AHEI in seven cases (10%). Patients diagnosed by dermatologists were classified as probable AHEI, doubtful AHEI, and unclear AHEI in 69%, 28%, and 3% of cases, respectively, and those diagnosed by pediatricians in 47%, 38%, and 15% of patients. Photographs of the lesions were available in 38 cases (55%), mostly performed by dermatologists (27/38, 71%). Among the 69 patients included, 14 (20%) had skin biopsies performed both by pediatricians (7/14) and dermatologists (7/14).

Primary outcome results

The degree of agreement between the experts for probable AHEI was moderate with a category-wise kappa of 0.45. There was a low degree of agreement on the classifications of doubtful and unclear AHEI cases (category-wise kappa of 0.36 and 0.15, respectively) (Table 1 online). One expert classified 25 cases as unclear compared to 6 and 9 for the two other experts and partially explained the moderate and low consensus obtained.

Table 1 A comparison between clinical characteristics of patients classified as probable (n = 40) and doubtful (n = 22) acute hemorrhagic edema of infancy
Full size table

Secondary outcome results

Demographic characteristics and clinical manifestations in children with probable and doubtful AHEI are shown in Table 1. Triggering factors were evidenced in 43 patients (28 patients with probable AHEI [70%] and 15 patients with doubtful AHEI [68%]) and were acute infections and vaccination. The mean duration between the onset of infection and diagnosis of AHEI (41/43 patients) was 6.8 days (range 0–20, missing data for three patients). Between the vaccination and disease onset (2/43 patients), data was available for one patient and was four days. The overall condition was good in 83% (n = 33/40) and 91% (n = 20/22) of children with probable and doubtful AHEI, respectively. Clinical signs significantly associated with probable AHEI were purpura localized on the face, ears, arms/forearms, or thighs/legs and edema localized on the hands with the absence of pruritus. The morphology of the purpura and the localization of the purpura on the hands or feet were not distinguishing features of probable versus doubtful AHEI (Figs. 2 and 3).

Fig. 2
figure 2

Distribution of skin lesions in the 40 patients: frequency (%) of purpura (left) and edema (right)

Full size image
Fig. 3
figure 3

a Edema of the hands and feet associated with rosette-shaped purpura on the face and the limbs in a 14-month-old boy diagnosed with acute hemorrhagic edema of infancy. b Purpura of the ear in an infant with acute hemorrhagic edema of infancy (Coll Pr Boralevi). c Coexistence of lesions of different stages: purpuric and post-ecchymotic lesions (Coll Pr Boralevi)

Full size image

A sensitivity analysis, which was not initially planned, was conducted to analyze clinical characteristics of patients with clinical photographs provided (Table 2 online). Results were similar to the main analysis, except for triggering factors, which were significantly more frequent with probable AHEI, and pruritus, which was more frequent but not significantly with doubtful AHEI.

In children with probable AHEI, the diagnosis was delayed by a mean of 2.7 days (0–11 days). AHEI was the original diagnosis in 60% of patients (n = 24/40). Other initial diagnoses were purpura fulminans in 10 patients treated by antibiotics and discontinued after negative culture results, IgAV in two patients, drug-induced anaphylaxis in two patients treated with antihistamines and corticosteroids, a viral infection in one patient treated with antibiotics and corticosteroids, and bullous impetigo in one patient due to a bullous purpura.

C-reactive protein (CRP) level was obtained for 29 of 40 patients and was elevated in 23 (58%; range 16 to 183 mg/l; median 54 mg/l). In 6 patients, CRP levels were above 100 mg/l, with negative bacteriological cultures. No data on leukocytes, sedimentation rates, and D-dimers were available. Creatinine level was normal in all 17 patients who were tested. Proteinuria was investigated in 29 patients and was found slightly positive in one case, at 0.5 g/24 h. The following bacteriological tests were performed in 22 patients among 40 (55%): blood culture in 12 (30%), mycoplasma serology in eight (20%), spinal puncture in six (15%), meningococcal PCR in four (10%), and urinary culture in five (13%). All bacterial tests were negative except for one urine sample positive for Escherichia coli. Viral tests were performed in 16 patients (40%). Nasal swabs of two patients were positive for rhinovirus, and the cerebrospinal fluid of one patient was positive for enterovirus. Skin biopsies were performed in 13 patients (33%) with probable AHEI and showed a leukocytoclastic vasculitis with presence of C3 on direct immunofluorescence in 6 out of 9 patients. Sixty percent of children (n = 24/40) were hospitalized for 1 to 11 days (median 3 days, missing data for 16 patients) in pediatric departments, including two patients in an intensive care unit. One patient with a urinary tract infection was treated with antibiotics. Four patients received H1-antihistamines and 1 patient oral corticosteroids. For 34/40 patients, no treatment was initiated for managing AHEI. The purpura and edema completely resolved without sequelae in all cases within 2 to 44 days (mean 17 days). Recurrences with new cutaneous manifestations occurred in two patients 1 and 2 weeks later, respectively. Eleven patients who were monitored at follow-up with urinalysis had normal urine findings.

The diagnosis of AHEI was doubtful in 22 patients. Photographs were available in 14 cases and were consistent with urticaria multiforme in 11 cases, isolated petechial purpura on the lower limbs in two cases, and IgA vasculitis in one case. One patient had a skin biopsy showing a leukocytoclastic vasculitis. Eight patients had no photographs, but for seven patients, the clinical history was consistent with urticaria multiforme and in 1 with IgAV.

The diagnosis of AHEI was unclear in seven patients: six of them due to incomplete data and one due to expert disagreement. The latter was a 12-month-old patient with ecchymotic purpura who was in good overall condition classified as probable AHEI by the first expert, doubtful by the second expert due to purpura and edema affecting only the feet, and unclear by the third expert due to incomplete data with no photographs.

Discussion

Key results

Our study found that AHEI is an uncommon disorder often misdiagnosed by both dermatologists and pediatricians. Nearly 40% of cases were not considered as probable AHEI by the experts. Urticaria multiforme was often mistaken for AHEI, and AHEI was often misdiagnosed as purpura fulminans leading to unnecessary hospitalizations and procedures.

Interpretation

In agreement with previous studies [3, 5, 17], our study indicates that AHEI typically presents with a triad of the sudden appearance of purpuric lesions, edema, and fever in an otherwise healthy boy with a median age of 11 months. The morphology of the purpura was typically targetoid or rosette-shaped and/or ecchymotic and predominantly affected the arms, the legs, the face, and the ears with relative sparing of the trunk (Fig. 3a, b). Edema predominantly affected the hands and feet and could be painful. An infectious trigger was found in most cases and was consistent with a recent study [24]. Pruritus was uncommon and was mostly observed for patients with doubtful AHEI. Although visceral involvement is uncommon [25], two patients with probable AHEI, for whom the diagnosis of IgA vasculitis was excluded, had abdominal pain. Lesions in different stages at the same age were also observed with coexistence of purpuric and post-ecchymotic lesions (Fig. 3c) [26]. However, the children in our study had greater impaired general condition (17%) than reported in a previous study (general appearance severely or mildly reduced in 8% of cases) [3]. This finding might be due to our large recruitment from pediatric emergency departments (14/40, 35%, in the probable AHEI group).

Most children classified as doubtful AHEI had clinical features consistent with urticaria multiforme. Urticaria multiforme is a cutaneous hypersensitivity reaction more common in young children and presents with large annular and polycyclic erythematous wheals with dusky ecchymotic centers, sometimes in association with acral edema. Lesions are commonly transient, diffuse, and pruriginous, and treatment with antihistamines is usually effective [2, 3, 5, 17, 21, 22]. IgAV also shares similar clinical features with AHEI, and some suggest that AHEI is a benign variant of IgAV occurring in young children [3, 5, 8, 9, 18]. Reports of the simultaneous appearance of these two diseases in a brother and a sister suggest that AHEI and IgAV may be variants of the same clinical entity [27]. However, IgAV and AHEI have important differences, and the distinction between these two diseases is essential to avoid unnecessary investigations, unjustified concern, and long-term renal monitoring. IgAV affects older children between three and 12 years old. The purpuric lesions of IgAV are predominantly localized to the lower extremities. IgAV can present with edema of the extremities and is associated with abdominal, articular, or renal involvement [28]. Purpura fulminans is another differential diagnosis reported by several publications [7, 19, 20]. In purpura fulminans, the purpura typically starts on the lower limbs, quickly spreads to become extensive, ecchymotic, and necrotic, and the patients typically have a very poor overall condition [29].

AHEI can have a frightening appearance; therefore, children were hospitalized in 60% of cases, and in two cases, the child was placed in an intensive care unit. These patients underwent invasive procedures such as spinal puncture, received irradiation from radiological exams, and were given ineffective treatments. These superfluous investigations and treatments increase the risk of iatrogenic harm. In our study, purpuric lesions localized on the face/ears, arms/forearms, and thighs/legs, edema of the hands, and the absence of pruritus were the distinguishing features between AHEI and other diagnoses, including urticaria multiforme and IgAV. A good overall condition usually discriminates between purpura fulminans and AHEI [7, 19, 20], but in cases of doubt, antibiotics and hospitalization are justified.

The diagnosis of AHEI can usually be made based on clinical manifestations; therefore, a skin biopsy is not essential [3, 5, 17]. However, misdiagnoses are common and the classification criteria suggested by Krause et al. are not sensitive, include an evolutive feature not present at the diagnosis, and are not validated. All skin biopsies performed in our series showed a leukocytoclastic vasculitis, some with presence of C3 on direct immunofluorescence [30]. Although leukocytoclastic vasculitis can be found on the skin biopsies of patients with IgAV [31] and purpura fulminans, skin biopsies can be helpful to differentiate between diagnoses [31,32,33]. In urticaria multiforme, a skin biopsy shows diffuse dermal edema, perivascular and interstitial lymphocytic infiltrate with variable number of eosinophils and/or neutrophils, and, rarely, a leukocytoclastic vasculitis [33]. IgA deposits on direct immunofluorescence are highly suggestive of IgAV [31]. A bacterial culture of the skin biopsy specimen can suggest purpura fulminans even after the initiation of antibiotics [32, 34].

In uncertain cases, an expert opinion may also be sought, in which case photographs of the cutaneous lesions become critical. In our study, patients were classified as probable AHEI in 69% of cases reported by dermatologists versus in 47% of cases reported by pediatricians. This discrepancy was likely due to the lack of available photographs in cases reported by pediatricians.

Limitations and strengths

Our study is limited by its retrospective design; we could not avoid recall bias and incomplete data, such as missing photographs. Moreover, the interexpert disparity observed was probably linked to the current lack of criteria for diagnosing condition, the lack of clinical photographs for some patients, and the missing data due to retrospective collection, although we excluded patients with missing data for important data. However, we described the largest series of AHEI published to date, with all cases of AHEI diagnosed by three independent experts. Recent literature included case series with a lower sample size (16 and 26 patients) [6, 17] and systematic reviews of the Acute Hemorrhagic Edema Bibliographic Database, but patients/reports included were not reviewed by experts [24, 25, 30]. Indeed, the main strength of our work is that we included only patients with a probable diagnosis of AHEI defined by three experts. We believe that cases reported in the literature could be misdiagnosed as AHEI, because in the present work, 32% of cases had doubtful AHEI. In addition, we compared patients with probable and doubtful AHEI to highlight clinical features that could help clinicians in the diagnosis of AHEI.

Conclusion

In conclusion, AHEI, which the diagnosis is made on clinical examination, is often misdiagnosed causing unnecessary hospitalizations, procedures, and treatments. Given its dramatic presentation, it is essential to accurately diagnose AHEI to avoid any undue stress to parents and medical staff. Clinical signs that can help distinguish AHEI from other diagnoses are purpuric lesions localized on the face/ears, arms/forearms, and thighs/legs with relative sparing of the trunk and edema localized on the hands in a young child (under 3 years old) with good overall condition and the absence of pruritus. In uncertain cases, a photograph of the lesions shared with an AHEI expert may be helpful. A urine test should be performed at diagnosis, and other investigations (e.g., skin biopsy, blood sample) are not required, unless in case of atypical features (e.g., impaired general condition). For the follow-up, urinary investigation should be proposed in case of proteinuria at the diagnosis. A clinical evaluation by a general practitioner pediatrician or dermatologist could be proposed to ensure resolution of symptoms.