Bronchiolitis: More Evidence, Fewer Interventions—Shifting Paradigms With Evidence‐based Diagnostics

Young, nonverbal children presenting to the emergency department (ED) with nonspecific complaints pose a significant diagnostic challenge to the emergency physician (EP). Given the paucity of pediatric research and clinical decision rules (CDR)1 for many common conditions that present to the ED, this can lead to overinvestigation, which can have negative consequences for the patient and families.2 Traditionally, rigorous evidence requires derivation of a rule, testing and validating the rule in small contexts, and externally validating the CDR in multiple settings.3 This has been done successfully in various contexts, including the traumatic head injury rules in children and management of low-risk ankle fractures.45 In the era of Choosing Wisely, and the recognition that “more testing” does not equate to better care, it is important to have CDRs that can both minimize unnecessary testing and give a clinician confidence in ordering specific tests.2, 6 The American Academy of Pediatrics published evidence-based guidelines for the management of bronchiolitis in 20067 and updated them in 2015.8 One study examining the effect of the 2006 guidelines on the investigation and treatment of children with bronchiolitis noted a significant reduction in investigations (labs 35% vs. 29%, respiratory syncytial virus testing 61% vs. 41%, chest X-ray [CXR] 61% vs. 52%) and treatment (bronchodilator 65% vs. 58%, steroids 25% vs. 16%).9 Antibiotic administration rates were unchanged (34% vs. 33%).9 The study showed that although releasing national guidelines had an impact in decreasing unnecessary tests and interventions, there is room to improve.10 The guidelines do not provide an ED CDR or an algorithmic pathway that could aid EPs in their approach to a child with suspected bronchiolitis. In Volume 23 Issue 10 Academic Emergency Medicine (AEM), Chao et al.11 present a systematic review and meta-analysis of the clinical predictors of airspace disease in pediatric patients presenting to the ED with bronchiolitis. The authors’ literature search and article selection were clearly outlined. In addition, they used the quality assessment of studies diagnostic accuracy tool to assess level of evidence from the five articles included. The authors conducted a pooled analysis of the 15 studies and 1,139 patients under the age of 2 years and found a weighted prevalence of 17% for a positive CXR with airspace disease. The only predictor was an oxygen saturation of <95%, with a likelihood ratio12 of 2.3. Fever, tachypnea, crackles, retractions, or breath sound asymmetry did not significantly change the post-test probability of a positive CXR. Low oxygen saturations did not necessitate the need for a CXR to rule out an infiltrate. However, in a novel approach, the authors created a mathematical diagnostic calculator based on the Pauker and Kassirer13 decision threshold model. Their model takes into account all important determinants like operating characteristics of a CXR in identifying superimposed bacterial infection, the risk of exposure to ionizing radiation, the risk of overtreatment with antibiotics, and anticipated benefit of treatment. The authors were clear that benefit from treatment of bacterial pneumonia with antibiotics has never been measured against placebo; therefore, using the best possible evidence they quantified and assigned reasonable conservative estimates of risk/benefit for every category. Based on their calculations, a CXR is only justifiable in patients with bronchiolitis and respiratory failure or severe bronchiolitis requiring admission. In patients with mild bronchiolitis who are eligible for discharge from the ED, the treatment benefit is estimated only at 5%, indicating that even when oxygen saturation is less than 95%, the harm will outweigh the benefit of finding airspace disease or treating it. Based on those findings, the authors suggest children with mild bronchiolitis should not be X-rayed, which is consistent with the national guidelines in North America.8, 14 Given that the findings focus primarily on oxygen saturation, it is worth discussing previous studies examining oxygen saturations in bronchiolitis diagnosis and management. In a double-blinded randomized controlled trial, Schuh and colleagues15 examined the outcome of children with mild to moderate bronchiolitis with initial oxygen saturation of ≥88%, with one group having their oxygen saturation artificially elevated by 3%. The group that had elevated oxygen saturation had almost half the admission rate when compared to the “true” saturation (25% vs. 41%), despite no physiologic difference in oxygen delivery. In another study, Principi et al.16 attached a blinded oxygen saturation probe to infants diagnosed with bronchiolitis who were discharged home. They found that 64% of children had desaturation events at home, with some having saturations as low as 70% for 1 minute or longer. The majority of desaturations took place during feeding or sleeping, and children with or without desaturations had comparable rates of return visits. This study helped build the evidence base that oxygen desaturations alone do not necessarily predict severe outcomes. With the increasing use of percutaneous oxygen saturation measurements for bronchiolitis, admission rates for children less than 6 months increased by 239% from 1980 to 1996 without any change in mortality.17, 18 Children with spot saturations > 89% while awake can be safely discharged. Oxygen saturation should only be used in conjunction with infant's risk factors, respiratory distress, and ability to feed safely to make a decision regarding further workup and disposition. This systematic review reinforces existing guidelines and supports it with pooled data. Unfortunately, there were no clear predictors for positive CXR from the study, and thus no significant clinical decision tools were put forward from the aggregated data. It is interesting that there were only five diagnostic studies with just over 1000 patients with high risk of bias. For good reasons, researchers were focused on therapeutics in randomized controlled trials for bronchiolitis that showed little benefit. We are now starting to understand that bronchiolitis is a heterogeneous illness that could be subtyped into groups.19 Future studies need to be more selective of their “bronchiolitis patients” when designing clinical risk scoring systems.20 After decades of research we still need a reliable, easy-to-apply, and validated clinical scoring system that takes into account risk factors, history, and respiratory assessment for infants presenting with wheezing to the ED. This coupled with diagnostic research and incorporating cutting-edge rapid microbiologic testing will allow us to diagnostically cluster infants with wheezing more accurately. A well-developed CDR would need to be designed and rigorously evaluated in a prospective randomized fashion in a multicenter trial against “usual care.” Until a breakthrough in therapeutics occurs, we should now focus on decreasing morbidity from overinvestigating and overtreating children with mild to moderate bronchiolitis. This review marks the first of AEM's Evidence-Based Diagnostic series focusing on pediatric issues. There are a myriad of other topics in pediatrics that should be reviewed, given their clinical frequency and practice variations. For example, investigating and reviewing diagnostics in children with a limp, renal stones or appendicitis are topics that should be reviewed and discussed. We hope that this review stimulated interest in this topic, and AEM looks forward to future submissions relevant directly to the EP who sees children.

Récupéré en direct depuis OpenAlex et désinversé. Les résumés ne sont pas conservés dans cette base de données : les index inversés représentent 8,6 Go des 9,3 Go de texte de la base, et le serveur dispose de 13 Go libres.