A Living Systematic Review of Nebulized Hypertonic Saline for Acute Bronchiolitis in Infants

Grewal and Klassen¹ in JAMA Pediatrics note the frustrations in interpreting evidence about bronchiolitis. Evidence is spread across a prior meta-analysis and other trials. The Grewal and Klassen editorial¹ encourages living systematic reviews that are updated as new trials emerge. We use the topic of nebulized hypertonic saline for bronchiolitis to propose the method of a living systematic review.

In this meta-analysis, we started by including the same trials in the Cochrane review by Zhang et al.² We then searched for newer trials in the Cochrane Central Register of Controlled Trials and articles in Web of Science that cited the Zhang et al study.² Our methods are detailed online at the living review (http://openmetaanalysis.github.io/Hypertonic-Saline-for-Bronchiolitis/).

We identified 11 new trials (4 only available at http://clinicaltrials.gov).¹ The meta-analysis showed that hypertonic saline significantly reduced the length of stay (LOS) among hospitalized infants. Heterogeneity was largely owing to variation in LOS in the control groups of trials. Benefit was confined to studies with a long LOS; however, even within this group study, results of recent trials were negative. Among infants given multiple doses, symptoms were improved and hospitalization was reduced. Forest plots, meta-regressions, and risk of bias assessment are available online. Quality of evidence as assessed by the GRADEprofiler was low owing to imprecision and other factors detailed in the GRADEprofiler online.

Prior research was comprehensively summarized by the Cochrane review. The addition of the newer trials attenuated the results of all outcomes. However, all outcomes were statistically significant owing to reduction in hospitalization in the subgroup analysis of infants who received multiple doses of treatment. The reduction in LOS was confined to older trials with a longer LOS. We rated the quality of evidence lower than the Cochrane review. This is likely owing to downgrading evidence according to criteria listed in our summary of findings table and methods (http://openmetaanalysis.github.io/Hypertonic-Saline-for-Bronchiolitis/). The small number of patients studied limit both generalizability and the certainty of conclusions.

This topic validates the Grewal and Klassen¹ recommendation for living systematic reviews. Further, a practice guideline³ and meta-analysis⁴ on the same topic were published this year and neither included the newer trials. Living reviews provide opportunities (Box) to advance knowledge creation, discourage unnecessary trials, and prevent under-recognition of prior research by trialists⁵ and authors.⁶ In addition, living reviews may help medical education via experiential learning (Box).

Box Section Ref ID

Repositories of living systematic reviews may be hosted at GitHub (http://openmetaanalysis.github.io/), as we did, or at the Systematic Review Data Repository (http://srdr.ahrq.gov/). GitHub provides web hosting and sharing repositories with version control. In addition, it offers wikis, issue tracking, communication, and social networking features that facilitate rapid collaborative knowledge work. Beyond sharing code, GitHub now hosts projects ranging from simple documents to complex governmental legislation.

Methods for living systematic reviews will evolve both technically and procedurally. For example, should open repositories include exhaustive interpretations of data? We propose that repositories should focus on the data and then link to interpretations of data in journals or tertiary publications, such as online textbooks and wikis. How can living reviews be designed to complement rather than compete with exhaustive reviews published by Cochrane and others? One solution is that when the published review is current, the living review will indicate there are no recent trials and link to the exhaustive review.

Available evidence, although low quality, suggests that hypertonic saline for bronchiolitis decreases the LOS for hospitalized children and may reduce symptoms and the rate of hospitalization. We encourage colleagues to help maintain this review, create other reviews, and advance the methods of living systematic reviews. Analyses and figures will be updated at http://openmetaanalysis.github.io/Hypertonic-Saline-for-Bronchiolitis/ as new trials are published.

Corresponding Author: Robert G. Badgett, MD, Departments of Internal Medicine, Preventive Medicine, and Public Health, Kansas University School of Medicine, 1010 N Kansas St, Wichita, KS 67214-3199 (rbadgett@kumc.edu).

Published Online: June 22, 2015. doi:10.1001/jamapediatrics.2015.0681.

Author Contributions: Dr Badgett had full access to all of the data in the study and takes responsibility for the integrity of the data and the accuracy of the data analysis.

Study concept and design: Badgett, Stirnaman.

Acquisition, analysis, or interpretation of data: Badgett, Vindhyal, Gibson, Halaby.

Drafting of the manuscript: Badgett, Vindhyal, Stirnaman.

Critical revision of the manuscript for important intellectual content: Badgett, Stirnaman, Gibson, Halaby.

Statistical analysis: Badgett, Vindhyal, Gibson.

Administrative, technical, or material support: Badgett, Vindhyal, Stirnaman.

Study supervision: Badgett.

Conflict of Interest Disclosures: None reported.

Additional Contributions: We acknowledge the patient assistance provided directly and indirectly by Jeroen Ooms, PhD, Department of Statistics, University of California Los Angeles, creator of https://www.opencpu.org/, and Guido Schwarzer, PhD, Institute for Medical Biometry and Statistics, Medical Center; University of Freiburg, Germany, author of http://cran.r-project.org/web/packages/meta/. Neither individual received financial compensation.