Interstitial lung diseases (ILDs) in infants and children represent a heterogeneous group of rare and complex respiratory disorders that affect the parenchymal part of the lung, and are mostly chronic [1].

Disease complexity is reflected by the various underlying mechanisms initiated by the primary causes of injury. As a result, a major concern is the difficulty to define and to identify ILDs in children and especially in infants. A rigorous analytical process is required to explore correlation between pathological determinants and patient’s phenotypes, which in turn is essential to progress in disease classification and understanding.

ILDs are considered as rare diseases. In a growing number of countries worldwide, rare diseases have been recognized as a crucial health issue, and strategies are developed to implement appropriate public policies for patient care and research. In France, rare diseases have also been identified as a priority of public health needs, and among the various actions launched; a National Reference Centre for Rare Lung Diseases (RespiRare®, http://www.respirare.fr) was created in 2006 to care for all children in the country suspected or known to have one of them, including ILDs. The main objective of RespiRare® is to establish disease cohorts based on aggregation of patients followed over time, through a national organization with affiliated centres that was set up to cover the entire French pediatric population with respiratory diseases. For ILDs, a platform was built to provide appropriate resources to pediatricians to record patient information and to ensure prospective clinical data collection. After data standardization and monitoring for quality control, data is stored into a national ILD database through a web interface. The RespiRare® platform, which combines clinical, biological, functional and genetic information, is a unique national structure for the emergence of prospective registries and databases allowing the development of epidemiological, clinical and translational research studies.

Information provided herein describes the ILD database of the RespiRare® platform. Analyses of the data currently included are presented, as well as the potential for international collaborative studies.

In the present article, we describe the first national internet-based patient database for pediatric ILD. Analysis of the collected data provides information on the current diagnosis distribution of these heterogeneous disorders.

The need for a centralized database was clearly identified, with the main concern being how to build its infrastructure. Indeed, the system would have to serve clinicians treating the patients, researchers for conducting national and international programs, epidemiologists to gather demographic data, patients and their families seeking accurate information on their disease, as well as the pharmaceutical industry expressing increased interest for rare disease databases. At the present time, the system includes personal/demographic data and clinical information.

As for other patient registries and central databases, the major challenge is data standardization and quality. Data quality is the cornerstone of the registry system, and we are extremely focused on quality control and data accuracy. As indicated above, the RespiRare® system is headed by the reference centre in close connection with the affiliated centres in French university hospitals. After several months of development, involvement of affiliated centres is increasing. To ensure data quality, an intensive monitoring has been organized with the coordinating team and a database committee, that review all cases proposed for inclusion in the database. To facilitate switch from usual paper-based documentation and on-line implementation, several solutions, depending on local possibilities have been set up. Of major interest for improving data accuracy is the progressive use by physicians of the system as a daily aid for patient care through a web accessible medical record backed on the database.

Since its creation, RespiRare® has made significant efforts to collect valuable information on ILD and other rare lung diseases. Certainly, a major achievement has been the ability to gather and connect the entire medical teams involved in the network. Among the tasks in progress, current efforts include the standardization of diagnoses and algorithm for explorations in strong connection with Orphanet, as well as the definition of minimal datasets that need to be collected for evaluation of disease progression, and the link with biospecimen repositories for the development of clinical and research studies.

Obviously, our database is not a register and presents multiple selection bias. The fist one is based on the fact that physicians are responsible for the allocation of patients and can omit to propose some of them for inclusion. The second bias is that the participating physicians are pediatric pulmonologist, who are not working in intensive care units (ICU). Thus, extremely severe neonatal or pediatric cases leading to death, where RespiRare® physicians were not consulted are not included in the database. In order to improve this bias, we are currently reinforcing links between pulmonologists and local ICU physicians. Very mild ILD diseases can also represent a bias given that they are not always referred to a reference centre. Finally, children above 15 years old can, in France, be referred to an adult pulmonologic department. Our pediatric network can also miss those patients, despite our focused on reinforcing links with adult teams. As a result, our database is not exhaustive, and cannot at this time provide an incidence or prevalence. However, we are working hard to promote the database in France and Europe.

We introduce here the largest national database for children ILD. Information available in the database includes diagnosis breakdown, phenotypic data on age at diagnosis, date of diagnosis, length of follow up, clinical presentation, BAL, functional tests, genetic tests and lung biopsy. The number of patient entries is growing continuously [2]. Our present data confirms similarities with the 2 other ILD databases in Europe, which are currently used in Germany and United Kingdom. The first one reviews all diffuse parenchymal lung diseases (DPLD) in children in Germany and describes 38 DPLD children [3]. The calculated incidence was 1.32 cases per 1 million of children per year, which is close to our estimations. We observe a comparable sex ratio around 0.9 male to female and a large part of onset of disease before 2 years old (one third before 1 year old in the German cohort and 42% in this study).

The UK database (BPOLD) focuses on 9 diseases including ILD, pulmonary alveolar proteinosis, pleural and pulmonary lymphangiectasia, and idiopathic pulmonary haemosiderosis. All together, it includes 29 children [4, 5]. The described diagnosis breakdown of ILD is similar to the German study and confirms that the main diagnoses of ILD are surfactant disorders, haemosiderosis, granulomatous diseases (sarcoidosis) and alveolar proteinosis [1, 3].

Regarding surfactant disorders, we confirm that SFTPC mutation I73T (c.218 T > C) is the more prevalent mutation, but numerous other mutations have been described [6, 7]. Recently, SFTPC mutations have been described in an adult cohort of familial idiopathic fibrosis [8]. This observation together with the extremely heterogeneous phenotypes of SP-C deficiency, ranging from early fatal respiratory failure to children and adults chronic respiratory disease should encourage to enlarge indications of SFTPC sequencing in children and adults with chronicle respiratory symptoms and alveolo-interstitial pattern on CT-scan [7, 9–11]. Recessive mutations in the ABCA3 gene were first attributed to fatal respiratory failure in term neonates but are increasingly being recognized as a cause of ILD in older children and young adults [12]. Over 100 ABCA3 mutations have been identified in neonates with respiratory failure and in older children with ILD [13–20]. The 5 children we report here are still alive and the oldest is now 16 years old. Our 7 diagnosed children for SP-B deficiency died within the first days of life in a refractory hypoxemia condition. Mutations in the TITF-1 gene have been associated with “brain-lung-thyroid syndrome” [21–27]. We report 2 children with this extremely rare disease, of which only one is still alive at 5 years old. So far, few mutations have been reported, mostly in exon 3 [28, 29].

The present study reports the largest pediatric haemosiderosis cohort with 23 cases. Haemosiderosis is a diagnosis of exclusion of diffuse alveolar haemorrhage syndromes (DAH) based on patient presentation with acute, subacute, or recurrent DAH, on the results of BAL showing more than 10% siderophages and lung biopsy showing evidence of ‘bland’ pulmonary haemorrhage (i.e., without capillaritis or vascularitis) [1, 30]. When diagnosing idiopathic pulmonary haemosiderosis in a child, particular attention has to be paid to diseases induced by environmental factors such as pesticide and cow’s milk (Heiner’s syndrome) [31]. One child in our study presented this particular acute situation. After the initial potential extreme anaemia, outcome is good in children with steroid treatment. Some children in our study needed additional hydroxychloroquin treatment as it is recommended for adults’ haemosiderosis. No capillaritis is accounted for in the database. Vascularitis related ILD are extremely rare in children, but are probably under-diagnosed. This database is a starting point and enables us to outline missing data. This dynamic process is ongoing and efforts are currently made to sensitize physicians to recognize and to include those patients.

Sarcoidosis represents also a large part of ILD in children. We report here 21 pediatric sarcoidosis, which are mostly teenagers. Clinical findings are various, depending on the different organ involvement [1, 32–36]. We also confirm that lung function abnormalities are frequently observed in children with restrictive pulmonary pattern and abnormal diffusing capacity [37]. All children with sarcoidosis were treated with steroids. This probably highlights that only the most severe sarcoidosis were included by physicians in the database.

Of importance, the present study highlights that more than a quarter of ILDs remain with no diagnosis, and the number is a third for patients included before 2 years old. This is in agreement with other previous reports. Undiagnosed cases are heterogeneous in terms of initial presentation, follow-up and management. An important challenge is certainly to reduce the number of unclassified ILDs, especially in youngest children, maybe with a precise phenotypic description and a better frequency and observation of lung biopsy. In this perspective, a more systematic approach to diagnostic is required, ensuring that the entire appropriate examinations currently available have been preformed at presentation and during the follow-up. Although huge efforts have been made in genomics, the phenotypic description of ILD cohort is still insufficient [38]. To progress, a well-implemented database with biologic, functional and histological markers represents an essential step for a large phenomic study on pediatric ILD.

To conclude, we introduce here the French network for rare lung diseases and the first national Internet based database for pediatric ILD. The database is now prospectively growing and more accurate data is being implemented in order to get better phenoms of these rare lung diseases. Moreover, a biologic bank related to the web-based database is currently being created. This database is a great potential for improving patient care and disease understanding. A European network of physicians, scientists and families’ foundations is currently emerging and we are working, together with the US ChILD foundation on a simplified European database register as a first step to European large studies [39–41].