Three cases of ICEP were identified from the national, internet-linked database (created in 2008) for pediatric interstitial lung disease (ILD) of the National Reference Center for Rare Lung Diseases (Respirare®, http://www.respirare.fr) 2 . Data were available for 217 cases of ILD, with “eosinophilic lung diseases” being one of the less-frequent etiologies (5/217) 2 .

As acute diseases are not included in the registry, cases of IAEP were retrieved by mailing to all members of the National Reference Center and to all members of the “Groupe Francophone de Réanimation et d’Urgences Pédiatriques”. This mailing process also permitted the identification of two supplementary cases of ICEP not yet registered and four other cases of eosinophilic lung diseases.

A detailed questionnaire regarding the initial presentation, investigations, elimination of differential diagnoses, treatment, outcome and follow-up was sent to referring physicians.

Inclusion criteria for ICEP and IAEP 3 were based on adult series (Table 1).

Footnote: * We excluded patients with evidence of: 1) exposure to drugs inducing pulmonary eosinophilia, according to the Pneumotox website (http://www.pneumotox.com); 2) ABPA; 3) positive stool parasite test or positive serologic test for Toxocara canis, Trichinella spiralis, Fasciola hepatica, Strongyloides stercoralis, Brugia malayi or Wuchereria bancrofti; 4) vasculitis or malignancy at the time of diagnosis.

Six cases of EP were excluded because of the following: hypereosinophilic asthma (two cases); Eosinophilic Granulomatosis with Polyangiitis diagnosis (three cases); positive serologic test for Wuchereria bancrofti (one case).

Due to the small number of patients in each group and the non normal distribution of values, we have presented median values with interquartile ranges.

Described by Carrington et al. 5 , ICEP is a rare disease accounting for < 3% of interstitial diseases, but it is the major cause of eosinophilic lung disease in countries with a low prevalence of parasitic infections. It predominates in females (sex ratio 0.5) with a mean age at diagnosis of 45 years, and less than 6% of patients are less than 20 years old 3 . More than 50% of cases have a prior history of atopy and asthma 6 . Patients are usually nonsmokers. ICEP has a progressive onset with development of respiratory (mild dyspnea, cough) and systemic (fatigue, malaise, fever, anorexia, night sweats and weight loss) symptoms over several weeks. The imaging features are rather characteristic with, in almost all cases, bilateral peripheral alveolar infiltrates with ill-defined margins on the chest radiograph and coexisting peripheral ground-glass on CT scan 4 . The blood eosinophil count is elevated in 90% of cases. BAL eosinophilia is now a major diagnostic procedure in ICEP: it is always present, commonly at around 40% in adults. ICEP responds dramatically, clinically, biologically and radiologically, to oral corticosteroids in adults and a short regimen of 6 weeks is now used. Relapses are frequent (> 50%) and require a longer regimen with low-dose corticosteroid therapy 7 .

ICEP is extremely rare in children. To date, only ten pediatric observations have been reported in the literature 8 9 10 11 12 13 14 , apart from those published by our center 15 16 17 , with no follow-up studies or series (Table 6). Diagnosis in children has relied on adult criteria 3 , as used in this study. Lung biopsy, no longer required for diagnosis, shows eosinophilic infiltration of the lung structures.

Footnote: F: female, M: male, nd: not determined, Eo: eosinophils, BALF: broncho-alveolar lavage fluid, AO: alveolar opacity, IO: interstitial opacities, Ad: adenopathy, BBO: basal band of opacities, GGO: ground-glass opacities.

Our study has collected comprehensive data for all the pediatric cases diagnosed and allows some comparisons to be made with the typical adult form 7 , even though the number of subjects is limited. Only two patients out of five had alveolar opacities and three had an interstitial aspect with ground-glass opacities and micronodular syndrome and for two of them a persistence of initial elements, with the late appearance of thin-walled cysts for both. We can thus distinguish two main subsets of patients.

The first subset (cases 1, 2 and 4) had a dramatic clinical, radiological and biological improvement after the first line of treatment, with no relapse to date. Cases 1 and 2 had a similar form to adults, and recovered fully after initial oral corticosteroid treatment. Case 4 had a 2.2% abnormal T-lymphocyte (TL) clone and a respiratory-only presentation with “patchy-only-interstitial” HRCT scan. The distinction between lymphoid-variant hypereosinophilic syndrome (L-HES) 18 and ICEP was questionable for this case. Indeed lung involvement in HES consists mainly of patchy ground-glass opacities as presented by our patient 19 . However, we considered that case 4 had ICEP, considering the respiratory-only setting (single-organ disease) without any other organ involvement at diagnosis and during follow-up (clinical follow-up and cardiac ultrasound performed at 1 month, 6 months and then yearly) 18 and the low level of the abnormal TL clone in comparison with levels reported in L-HES 20 – this clone having been described at low levels in the general population 21 . Moreover, Cottin et al. recently reported T-cell abnormalities in eosinophilic lung disease, showing that the pathophysiologic boundary with HES may be very thin (1/9 adult patients with ICEP had T-cell receptor gamma rearrangements and 1/57 adults with eosinophilic lung disease had an aberrant T-cell population immunophenotype) 22 .

The other subset (cases 3 and 5) had a persistent interstitial form with the appearance of thin-walled cysts for both after more than five years of stability of the initial lesions for case 5 and two years of follow-up for case 3. They, moreover, had very limited general symptoms and less peripheral blood hypereosinophilia as well as lower BAL eosinophilia compared to the adults. One patient had, at diagnosis, a biopsy showing an initially typical histological aspect of ICEP. The other patient had a T-cell clonality, as recently described in ICEP 22 . Their clinical and radiological follow-up might, nevertheless, suggest revisiting the final diagnosis. Indeed, mild BALF eosinophilia has been described in adult idiopathic pulmonary fibrosis (IPF), in up to 50% of patients with desquamative interstitial pneumonia (DIP), in cryptogenic organizing pneumonias and nonspecific interstitial pneumonitis 23 24 25 26 27 . Peripheral blood eosinophilia is rarely found in UIP and DIP (around 10% of cases) 27 . ICEP can evolve, moreover, spontaneously into lung fibrosis with a DIP-like reaction 28 . All of these data indicate that there is still a place for lung biopsy in the diagnosis of ICEP when the initial setting and evolution are not usual. The modality of these biopsies also needs to be clarified, as case 3, with an unusual evolution, was initially described as having a standard histopahologic finding of ICEP on transbronchial biopsies. In cases not responding to initial oral corticosteroid treatment, diagnostic biopsies should be performed after a period of weaning off treatment, to contribute to the differential diagnosis between ICEP and idiopathic interstitial pneumonia. However, the optimal length and the feasibility of this weaning period need to be established.

The frequency of unusual forms in this first pediatric national series suggests the specification of a more strict pediatric definition, including the definition of the requirement for biopsy. The treatment regimens of these patients should be modified and the use of corticosteroid-sparing agents in the context of an eosinophil-triggered disease evaluated (omalizumab, mepolizumab) 29 . Low-dose cyclosporine, used by other teams in lymphoid-variant HES and idiopathic HES 30 , should be further evaluated in the context of eosinophilic pneumonias 31 , and especially in persistent forms of ICEP.

Regarding hypereosinophilic asthma, the boundary 32 can be very difficult to specify in young children, and two patients who were excluded from our cohort were finally classified as asthmatic. Hypereosinophilic asthma can be observed in pediatric patients with > 25% eosinophils in BALF. Asthma attacks are induced by viral infections in 80% of cases 33 and alveolar opacities and atelectasis are frequent. Limits can be difficult to establish initially as more than 50% of ICEP patients have previously been diagnosed as asthmatic 7 . Only the clinical evolution of the patient can help the evaluation of the clinician.

In order to consider these pediatric specificities, the definition of ICEP could be clarified in children as:

i) diffuse pulmonary alveolar consolidation with air bronchogram and/or ground-glass opacities, especially with peripheral predominance;

ii) BALF eosinophilia > 20% or peripheral blood eosinophilia > 1×10⁹ cells/L;

iii) respiratory symptoms present for more than 4 weeks;

iv) absence of other known causes of eosinophilic lung disease;

v) consistent open lung biopsy for cases without initial dramatic clinical AND radiological improvement on first-line treatment.

Two items of adult classification concerning imaging and the exclusion of differential diagnoses were not changed. Clinical and biological aspects were adapted in order to be fulfilled by most of the pediatric cases. As discussed previously, the requirement for open lung biopsy has been specified in cases of resistance to first-line treatment.

A new classification algorithm for eosinophilic pneumonia in children is also proposed (Figure 3), including two main differential diagnoses to ICEP: hypereosinophilic asthma and idiopathic interstitial pneumonia. The overall impression is that idiopathic chronic eosinophilic pneumonia is part of the clinical spectrum of a wider group of lung diseases, with some potential overlap between them, making a clear distinction sometimes difficult. Molecular analysis should likely help in the clarification of this aspect and better circumvent underlying biological processes.

We describe three cases of IAEP in adolescents. Specific descriptions of pediatric IAEP are rare 34 35 36 , but such patients have also been included in series mixed with adults 37 . To date, no specific pediatric series have been described but pediatric cases are often adolescents. Acute exposure to irritants was found in our series (smoke, cannabis, paint stripper and tear gas), as in adults. The severity of respiratory distress requiring ECMO 24 hours after hospitalization and 48 hours after the beginning of symptoms has never been described in adults. This does not correspond to a delay in medical care but rather to the rapid development of an aggressive pathology. Other clinical, biological and radiological aspects of these patients seem similar to those in adults. The length of steroid treatment was highly variable, as in adults, and should be reduced to two weeks 38 . Treatment resulted, as in adults, in a dramatic clinical improvement, with rapid weaning of ventilator support. None of the patients relapsed. CT scan control was normal and initial spirometric follow-up showed only minor alterations.