Prospective evaluation of hydroxychloroquine in pediatric interstitial lung diseases: Study protocol for an investigator-initiated, randomized controlled, parallel-group clinical trial

Interstitial lung disease in children (chILD) is an umbrella term for more than 200 different entities that affect the parenchyma of the lungs, leading to a chronic lung disease. Previous studies have reported an incidence of 0.1–16 per 100,000 children per year and a prevalence of 1.3–3.6 per million children [1‐3]. Often the children are dependent on oxygen (O₂) for a long period of time or require mechanical ventilation [4]. The overall mortality in childhood is around 15% [4]. No successfully proven treatments are available and there only exist few pharmacological treatment options based on expert agreement [5] and anecdotal experience collected from individual treatments. Symptomatic treatment consists of O₂ supplementation if necessary, adequate nutrition adapted to the high energy demand generated by the disease due to the increased breathing effort required, as well as immunization against respiratory pathogens to prevent exacerbations through respiratory infections [6]. Common drug treatments include glucocorticosteroids, hydroxychloroquine (HCQ), and steroid-sparing agents, such as azathioprine, cyclophosphamide, cyclosporine or methotrexate, are also tried [6]. A few cases with possible responses to macrolide antibiotics have been reported.

Due to the rarity of the diseases, only international studies can collect sufficient numbers of patients to evaluate treatment in controlled settings. We used the European-wide project on chILD, the chILD-EU project [7], to prepare and implement a randomized controlled, Phase 2a study to evaluate treatment with HCQ in chILD. With this study we aim to: (1) evaluate the efficacy of HCQ against placebo in chILD, (2) evaluate the safety of the mid-term use of HCQ, (3) be able to make a decision on the risks and benefits of the use of HCQ and (4) help in standardizing the pharmacological treatment of chILD.

HCQ is approved for the treatment of malaria, rheumatoid arthritis, juvenile idiopathic arthritis, lupus erythematosus and dermatological conditions caused or aggravated by sunlight [8]. Due to the assumed various immunomodulatory mechanisms of HCQ it has been used frequently for more than 40 years off-label for different autoimmune diseases (e.g., Sjögren’s syndrome and inflammatory osteoarthritis) in a wide range of clinical conditions and, in particular, in chronic or acute and often very severe pediatric diffuse parenchymal lung diseases.

Eighty-three cases of chILD were reported between 1984 and 2013 to have been treated with chloroquine (CQ) or HCQ alone or in combination with other drugs [9]. From these data it was not possible to predict which patients would benefit from the treatment. Clear positive clinical improvements, usually occurring within 4 weeks, have been reported in 15 of 16 published cases when HCQ was given alone, and in 37 of 53 published cases, when given in combination with glucocorticosteroids. Obviously, a publication bias towards successful cases has to be considered.

Applications even for many years appeared safe with very few reported side effects during off-label usage [9]. Again, a significant underreporting of side effects or unsuccessful trials has to be expected. The oral doses ranged from 3.5–10 mg/kg bodyweight per day [9]. In this respect it is notable that there exist two pharmaceutical forms of HCQ, HCQ base and the HCQ sulfate (correction factor 200 mg sulfate = 155 mg base) [9]. The majority of case reports and small series did not indicate which of the two pharmaceutical forms the dose is referring to. Hydroxychloroquine treatment may have several side effects including rare retinal changes and, most frequently, abdominal discomfort and pain. The retinal changes are dose-dependent and can be irreversible [10]. In one series of patients who received HCQ, abdominal pain in the first week of treatment was reported in a minority of cases [11].

The use of HCQ in a trial in chILD has several important challenges. Firstly, it is used in a broad age and weight range, from birth to adulthood, with not much pharmacological data available [9]. Secondly, disease severity varies from acute respiratory distress and failure to thrive to chronic, mild partial respiratory insufficiency. Randomization of acutely ill patients to treatments which may potentially be harmful may be very difficult to achieve and consent by the caregivers, even if the treatments might be helpful, may not be given, thus reducing the number of subjects that can be included. Nevertheless, patients in a chronic disease state have a different response pattern than those who are acutely sick. All this must be addressed by an appropriate study design. Thirdly, no randomized controlled studies have been performed in chILD. Therefore, outcome variables have not been used in studies and their responsiveness to treatments is unknown. Lastly, the extreme rarity and heterogeneity of the diseases targeted, as well as their wide geographical scattering, induces significant burden and huge logistic challenges for such a clinical trial. The goal of this Phase 2a study is to include all chILD patients who are planned to be, or are actively being, treated with HCQ. We intend to gather as much information as possible to improve the knowledge in this group of patients. Therefore, the character of this study is explorative and not confirmatory. It was designed to closely accommodate the current clinical care situation.

Children’s interstitial lung diseases (chILD) cover a large, heterogenic group of rare pediatric pulmonary disorders. They are difficult to diagnose, many are not yet genetically defined, there are no evidence-based pharmacological treatments and the natural disease course cannot be reliably predicted, as such information is frequently lacking. Some years ago, we implemented a web-based international European registry and biobank – the European Management Platform for Children’s Interstitial Lung Diseases (chILD EU Register) [7] to provide structures for the collection and care of such “orphaned” patients. The platform also has the technical and legal requirements to perform investigator-initiated, randomized controlled treatment trials. This study is embedded in the chILD_EU Register and is the first prospective, international, interventional study of chILD. At the beginning of the project in 2011 we did not anticipate the broad diversity and depth of the burdens and pitfalls that we were facing. Here we will discuss some decisions that we had to make from the beginning to the successful implementation of this trial.

As the entities for which empiric treatment with HCQ was successful were not molecularly defined, we decided to be as inclusive as possible. Nevertheless, our inclusion criteria were clearly set to select so far: (1) subjects of all ages (premature infant to adults) with genetically defined surfactant-dysfunction syndromes or molecularly defined diseases caused by mutations in genes associated with diffuse parenchymal lung diseases and anecdotal or possible responsiveness to HCQ; (2) mature newborn (≥ 37 weeks of gestational age) to young adults with lung biopsies showing any of the histopathology pattern and clinical phenotypes known to be associated with these well-defined conditions and (3) children with idiopathic pulmonary hemosiderosis, in the absence of a lung biopsy, due to some evidence for HCQ responsiveness [12]. In an amendment, the following entities were removed: microvasculopathy, cryptogenic organizing pneumonia, diffuse alveolar damage/acute interstitial pneumonia, acute fibrinous and organizing pneumonia, giant-cell pneumonia, chILD clinically diagnosed (CDB-ILD), respiratory bronchiolitis-interstitial lung disease, sarcoidosis, hypersensitive pneumonitis, neuroendocrine-cell hyperplasia of infancy and pulmonary interstitial glycogenosis. Reasons to exclude the first six entities were that they were non-specific and non-predicting for the diagnosis of chronic diffuse parenchymal lung disease; the next three entities were excluded as there are established and specific treatment responses in adult patients and the last two have generally a favorable natural course and lack any previous data of HCQ responsiveness. These restrictions narrow the disease spectrum, thus the revised collection includes either genetically well-defined entities or complementary histopathologically defined chronic diffuse parenchymal lung diseases, previously clarified as idiopathic interstitial pneumoniae [13]. Despite some potential disadvantages associated, e.g., a great variability of disease severity as currently known, a broad range of different molecular entities, potentially new and different underlying pathomechanisms, a wide age-range from infancy into adulthood, we expected significant advantages to this “basket” approach. An important justification comes from the molecular definition of the diseases, at least allowing “n of 1” descriptions and later grouping of similar observations. As this Phase 2b study is exploratory, some entities with good responses may be identified and allow later focusing on these. Additionally, the approach selected here mirrors everyday care and clinical routine, i.e., treating the majority of this group of chILD patients empirically, enhancing their chances of participating in this study.

To avoid disadvantages for families and patients living in remote areas, we tried to open as many trial centers as possible. However, it was soon evident that the financial and, in particular, the administrative burden was too big and in Germany in the beginning of 2019, 12 study centers were open. While the Ethics Committee would allow expansion of a clinical trial site to another hospital (“flying doctor”) to prevent excluding patients for no good reason, the administrative office of the sponsor and the lead clinical trial unit did not agree to this approach.

As already described above, we designed separate START and STOP blocks, which was helpful to recruit subjects otherwise lost because subjects were already taking the medication. It will be interesting to analyze the results with statistical approaches, that take care of such design features. To our knowledge, similar approaches have not been used widely. Due to the heterogeneity of the diseases with a wide spectrum of clinical symptoms and severity it is not possible to determine one specific HCQ mode of duration that will fit the majority of subjects. Physicians anticipate the treatment of some patients who will be treated with HCQ only for some weeks as well as patients who will need HCQ treatment for several years. Due to the variety of clinical conditions and our intention to be as inclusive as possible in this exploratory trial, we allowed a variable time period in which patients were treated with HCQ between the START and STOP blocks. As this is an investigator-initiated trial, reimbursement of the sites for including subjects is limited and is very much dependent on the funding organizations (DZL €4000 and E-rare €6000 for START and STOP). Lastly, and despite more than 50% of those known to the study centers being recruited into the trial, it was soon evident that the number of cases observed in Germany was insufficient to complete the trial. Luckily, we could obtain an international grant, which would allow us to start to expand the study into six other European countries (Portugal, Spain, Austria, Poland, Italy and Turkey). We hope that we can successfully perform this study and that it will generate some important insights into the fascinating areas of chILD.