Introduction

Viral bronchiolitis is a very common cause of outpatient visits in children younger than 2 years and accounts for 16% of hospitalizations in this age group [1]. Although several viruses may cause bronchiolitis, respiratory syncytial virus (RSV) is responsible for about 80% of infections, especially in infants [2]. Whereas the infection has a benign and self-limiting course in the majority of affected children, it can be a serious and potentially life-threatening disease in some patients with complex underlying diseases such as congenital heart disease or bronchopulmonary dysplasia [3].

Unfortunately, a causative treatment of acute viral bronchiolitis does not exist due to its special pathophysiology. Infection of the terminal bronchiolar epithelial cells causes edema, excessive formation of mucus, and eventually necrosis of bronchiolar epithelia. Hence, therapy is restricted to symptomatic measures until the respiratory epithelia have regenerated. Such supportive care may include a variety of medications, including chest physiotherapy and breathing aids, and different caregivers have been using different approaches in the past. In 2014, a landmark paper by the American Association of Pediatrics was published which has been regarded a universal clinical practice guideline since [4]. Surprising to many, the AAP guideline does not support the routine use of commonly used therapies such as chest physiotherapy, bronchodilators (e.g. albuterol, or salbutamol), epinephrine inhalation, corticosteroids, or antibiotics. Recommendations were only made for administration of nebulized hypertonic saline and nasogastral or intravenous fluids, both in hospitalized children only.

The AAP guideline was developed by a group of very renowned experts based on published study results. However, it is increasingly noted that the results of many clinical trials are not reported in a timely manner or not reported at all. Such selective reporting of study results, known as publication bias or publication gap has been observed in a various fields of pediatrics [5–8]. Given the discrepancy between the few treatment options recommended by the AAP guideline and the widespread use of various other therapeutic approaches, we were wondering whether the actual decision making might be influenced by such a publication gap.

Methods

Results

Publication status of studies

We identified a total of 124 studies that were reported as completed in the ClinicalTrials.gov database. Thirty of these studies were not related to viral bronchiolitis, the majority of which studies on bronchiolitis obliterans following lung transplantation. Of the remaining 94 clinical trials, 71 investigated therapeutic interventions in acute viral bronchiolitis. Two of these studies, both unpublished, were completed less than one year before close of the database. Because the U.S. Food and Drug Administration (FDA) allows a time frame of one year between completion and publication of the study as specified in the FDA Drug Administration Amendments Act (FDAAA) [10], these two studies were excluded from the analysis. Of the remaining 69 studies, 50 were published and 19 were unpublished (S1 Table). Among them were two clinical trials that had been reported as conference papers only. One of these abstracts gave sufficient data that supported a therapeutic recommendation. We therefore regarded the study as published. The other abstract, however, only contained vague information on the study results, therefore we considered the study unpublished. All but one principal investigator of the unpublished studies could be contacted by email or social networks. Of these 18 authors, three replied and confirmed that the study results had not been published yet (Fig 1). Publication rates considerably varied between different countries of the sponsor (Table 1) and main topics of the investigations (Table 2).

Download:

• PPT
  PowerPoint slide
• PNG
  larger image
• TIFF
  original image

Fig 1. Flowsheet: Details of the study selection process.

https://doi.org/10.1371/journal.pone.0237801.g001

Download:

• PPT
  PowerPoint slide
• PNG
  larger image
• TIFF
  original image

Table 1. Countries of the sponsor.

https://doi.org/10.1371/journal.pone.0237801.t001

Download:

• PPT
  PowerPoint slide
• PNG
  larger image
• TIFF
  original image

Table 2. Main study topics.

https://doi.org/10.1371/journal.pone.0237801.t002

The numbers of published and unpublished studies for each year of study completion (2003–2018) is shown in Fig 2. While 80% of the registered clinical trials that had been completed up to 2014 have been subsequently published, it is noticeable that almost 40% of the studies completed after the release of the AAP guideline still remain unpublished.

Download:

• PPT
  PowerPoint slide
• PNG
  larger image
• TIFF
  original image

Fig 2. Distribution of published (n = 50) and unpublished (n = 19) trials by year of completion.

Close of database was October 1st 2019.

https://doi.org/10.1371/journal.pone.0237801.g002

Patient numbers

All studies involved both genders. Published trials contained data from n = 9403 patients, whereas n = 4687 patients were enrolled in unpublished trials. Median size of published trials was 93 (IQR 48–175), range 12–1636 individuals, whereas median size of unpublished trials was 94 (IQR 60–146), range 33–2580 participants. Fig 3 shows that in some years, the number of patients enrolled in unpublished studies significantly exceeded those in published studies.

Download:

• PPT
  PowerPoint slide
• PNG
  larger image
• TIFF
  original image

Fig 3. Distribution of patient count stratified by publication status and year.

Close of database was October 1st 2019.

https://doi.org/10.1371/journal.pone.0237801.g003

Time to publication

Median time to publication was 20 months (IQR 13–31), range 2 to 58 months. No trend could be identified that either older or more recent studies were published faster, and only 8 of 50 trials were published within 12 months after completion as warranted by the FDAAA (Fig 4).

Download:

• PPT
  PowerPoint slide
• PNG
  larger image
• TIFF
  original image

Fig 4. Time to publication (time between completion of the trial and publication of results) in months by year of completion.

Close of database was October 1st 2019.

https://doi.org/10.1371/journal.pone.0237801.g004

Discussion

Few other guidelines have gained quasi universal global acceptance as the 2014 AAP guideline on bronchiolitis [4]. Therefore, our analysis of the registered clinical trials on the treatment of acute viral bronchiolitis had two key questions: (1) Could a publication gap have influenced the guideline’s recommendations and (2) did the guideline affect the publication status of subsequent clinical trials?

For the first part of our analysis, only clinical trials the results of which could have been available prior to completion of the AAP guideline were taken into account. Few therapeutic options have been recommended in the guideline: oxygen (given an oxygen saturation below 90%), adequate fluid intake, and inhalation with hypertonic saline. No unpublished studies on supplemental oxygen or fluid intake were identified. However, two studies on the use of hypertonic saline have been completed, but not published before the AAP guideline was released. The guideline gives a weak recommendation to administer nebulized hypertonic saline to infants and children hospitalized for bronchiolitis, but does not recommend its use in the emergency department (ED). One of the unpublished studies compared hospitalization rates of infants presenting to the ED with bronchiolitis during the year of use of nebulized hypertonic saline versus the two previous years when nebulized hypertonic saline was not used in a single center in France (NCT01460524). This study alone enrolled 2580 patients, which is about 10% more than the sum of all patients (2294) from 14 studies analyzed in the AAP guideline. So the results of this trial might well have affected the recommendations for the use of hypertonic saline in the ED. The other unpublished clinical trial (NCT01238848) studied the effects of hypertonic saline versus normal saline in combination with albuterol in 82 children already hospitalized for moderate bronchiolitis. It is unlikely that this study would have had an impact on the AAP recommendations.

The AAP guideline explicitly advises against various other therapeutic measures: bronchodilators (albuterol or salbutamol), antibiotics, epinephrine, steroids, and chest physiotherapy. In our analysis, we did not find unpublished studies on the use of bronchodilators or antibiotics. One unpublished clinical trial with 82 participants (NCT00435994) investigated the production of VEGF from nasal washing after inhalation with epinephrine in infants with bronchiolitis and healthy controls, respectively. Thus, published results of this study would not have affected the guideline. We identified one unpublished trial with 94 patients on the use of steroids in acute bronchiolitis (NCT02571517). In this study, children that had been hospitalized for moderate to severe bronchiolitis received intravenous methylprednisolone and/or oral prednisolone for seven days. Several clinical endpoints (disease severity scores, chest X-rays, admission to the PICU, need for mechanical ventilation) were defined as primary and secondary outcomes on day 7. Interestingly, the AAP guideline focuses on hospital admission rates regarding the use of steroids, but does not give evidence for a lack of efficiency in children already hospitalized for bronchiolitis. Thus, the results of this unpublished study might have had an impact on the official recommendations. Several of the recommendations in the guideline are supported by low-quality evidence, e.g. the recommendation not to administer supplemental oxygen if SpO₂ >90% (Expert Opinion; Evidence quality D). Many infants with bronchiolitis upon discharge from the hospital have lower basal SpO₂ than healthy control infants and an elevated oxygen desaturation index during sleep [11]. The pathophysiological and clinical consequences of mild nocturnal hypoxemia and especially of the intermittent type are unknown, therefore, further research is needed.

There are also therapies for bronchiolitis on which clinical trials are listed in the register, but that are not addressed in the guideline. Montelukast is mentioned in Appendix 1 as a MedLine search term, but its use is not discussed in the text. The Clinical trials database lists a total of six studies on montelukast in bronchiolitis, two of which only investigated the pharmacokinetics. Of the remaining four clinical trials, two have been published. A small study with 53 participants found no benefit of montelukast in patients with acute disease [12]. A second multicenter study with 1125 participants focused on post-RSV bronchiolitis during an observation periods of 24 weeks [13]. Again, beneficial effects of montelukast were not observed. On the other hand, we identified two unpublished randomized controlled trials on the use of montelukast in acute bronchiolitis, both with significant numbers of participants (NCT01370187: n = 146; NCT00863317: n = 141). It can be speculated that the AAP would have advised against montelukast in the guideline if the results of the two unpublished studies had come to the same conclusion as the published ones.

Finally, we analyzed the database to find out whether the AAP guideline might have affected the publication status of clinical trials in children with acute bronchiolitis. It was impressive that the rate of unpublished studies doubled after the release of the guideline. Among the unpublished studies completed after 2014, four were on the use of nebulized hypertonic saline (NCT02538458, NCT02233985, NCT03143231, NCT03614273). It is likely that given the guideline’s at least weak recommendation to use it in hospitalized patients, there is a publication bias regarding positive results. In other words, the mere confirmation of the official recommendations may not have been worth the effort of publication. The same may hold true for two other unpublished studies on the combined use of hypertonic saline, dexamethasone, and epinephrine (NCT01834820) and chest physiotherapy, respectively (NCT02853838). On the other hand, there are possible therapies that might be addressed in a revision of the guideline in the near future, such as the use of high flow nasal cannula (HFNC). Since there are several clinical studies that found positive effects of HFNC in acute bronchiolitis [14–16], maybe this will motivate the principal investigators to publish the results of three recent studies that are yet unpublished (NCT02791711, NCT01498094, NCT02856165).

This study has several limitations. We did not investigate other clinical trials databases (e.g., EU-CTR or the German Clinical Trials Register), because ClinicalTrials.gov is generally considered the largest and most important clinical trial registry. Only registered clinical trials could be analyzed because the existence of non-registered trials was not transparent to us. In order to avoid a clinical trial being erroneously classified as unpublished, we conducted a semantic literature search in PubMed and GoogleScholar and we contacted investigators and sponsors. We assumed that the data submitted to ClinicalTrials.gov were accurate and complete as mandated by the FDAAA [10]. We did not compare whether the pre-specified statistical analysis plan and the pre-specified research questions of the study were consistent with the published reports, because this information was not in the public domain for all studies.

In conclusion, publication bias or publication gap is an ubiquitous issue in clinical science. Increasingly, study results are not reported in a timely manner, or not published at all. The therapy of acute viral bronchiolitis is a major task for pediatricians during the winter season, and it is supported by a highly appreciated guideline that is followed worldwide. But even this guideline might have been affected by a publication gap and in turn might have discouraged investigators to publish their results after its release.

Supporting information