Acute bronchiolitis (BA) is a disease of the lower respiratory tract caused by viral infection, especially by respiratory syncytial virus (VRS), that predominantly affects infants and exhibits a seasonal pattern.1 It is the leading reason for hospital admission in children aged less than 2 years worldwide, with an annual rate of hospitalisation of approximately 25 per 1000 children aged less than 2 years in Spain.2 Although its clinical picture has been known for a long time, it was in 1940 that the term “bronchiolitis” was first introduced to refer to it.3 This term has become the standard name for the disease. In 1967, the National Library of Medicine of the United States included bronchiolitis as a MeSH descriptor (bronchiolitis, viral).

The term bronchiolitis is currently used worldwide in hundreds of thousands of diagnoses each year, and yet there are still problems about its definition.4 There are no universally accepted diagnostic criteria, so that the same clinical picture may be diagnosed as AB or given a different label: infant asthma, bronchopneumonia, episode of wheezing or bronchitis, possibly accompanied by any of several qualifiers (spasmodic, asthmatic, catarrhal, etc.). Some authors have even questioned its being a distinct entity.5 Various scientific societies, quality assessment agencies and individual experts have developed criteria for the clinical diagnosis of AB, which differ on significant aspects.4 The case definitions and inclusion criteria used in clinical trials of AB are also heterogeneous, and many studies have considered eligible any infants with “signs and symptoms consistent with bronchiolitis”, with the authors simply stating that they applied one or other “widely accepted” definition of AB.6,7

There are also inconsistencies in clinical practice. The criteria used to diagnose AB vary, and the diagnostic label assigned to a patient with manifestations of AB determines the approach to treatment.8 Recent guidelines have established that there is no effective pharmacological treatment for AB,9,10 but clinicians do not adhere to the treatment recommendations given in current guidelines.11 This may be due to the discrepancies that exist between different guidelines,12 but also to differences in opinion about what is or is not AB.13–15 Similarly, the heterogeneity in diagnostic criteria generates disagreement in regard to the assessment of the long-term prognosis.16

The evidence on the criteria used to diagnose AB in clinical practice is scarce. A study conducted in Portugal found substantial variability in the applied criteria both in general physicians and in paediatricians.17 In Spain, the AB guideline published in 2007 by the Ministry of Health did not include a case definition,18 and published Spanish studies usually cite one of the most common definitions in the international literature. It is not known whether Spanish paediatricians use consistent criteria to diagnose AB.

The aim of our study was to find out which criteria paediatricians in Spain use to diagnose AB, including both experts on the subject and the collective of paediatricians that manage children with AB, and to analyse possible contributors to variability.

We conducted 2 consecutive studies. In the first study, we assessed whether there was agreement between Spanish experts on the diagnostic criteria for AB using the Delphi method.19 The second one was a cross-sectional survey of clinical paediatricians carried out by means of an online questionnaire that explored their opinions regarding the diagnosis of AB. Appendix B (available in the online version of this article) provides detailed information on the methodology used.

Main findings. The criteria used to diagnose AB in Spain are heterogeneous. Among experts, there is consensus on a minimum of criteria that does not extend to aspects as relevant as the age limit for diagnosis. Clinical paediatricians do not frequently use standardised criteria, but most consider that the diagnosis should be restricted to first episodes and to the first 24 months of life. The importance given to different signs or symptoms varied depending on the type of professional activity. Hospital specialists, such as intensivists or neonatologists, underscored the relevance of dyspnoea, while coryza in AB was mainly taken into consideration by primary care paediatricians. We also found moderate differences based on the age of respondents. However, our data suggest that geographical factors do not contribute to the heterogeneity of the diagnostic criteria.

Interpretation. Over 40 years ago, McIntosh stated, in regard to bronchiolitis, that “Clinicians who care for small children all know what this means”.21 This certainty seems to now be faltering, assuming it ever existed. The criteria used for diagnosis differ between countries (for instance, between the United States and northern Europe) and between the physicians in charge of these patients.17 It is known that the treatment approach chosen in children with manifestations of AB is associated with the diagnostic “label” assigned to the case,8 and some authors have even argued that the alleged ineffectiveness of pharmacotherapy in AB is due to an inadequate definition of AB.5

In the Delphi process, the most significant disagreement between the experts concerned the age limit for diagnosis. Rather than increasing agreement, the second round further polarised the positions of the experts (split between 12 and 24 months). The same division has been observed in the United Kingdom.22 This is alarming, as age is the most influential factor in giving a patient with compatible signs and symptoms a diagnosis of AB.8 All widely used definitions limit the diagnosis of AB to infants, and several establish an upper limit of 24 months,9,20,23 which was the age limit that two-thirds of the Spanish paediatricians surveyed in our study agreed with. There are other definitions, however, that propose a limit of 12 months,21,24–27 and some highlight that most cases of AB occur in infants aged less than 6 months.14,28,29 Definitions proposed by Scandinavian authors tend to have a lower age limit.29–31 Among the scarce evidence that may contribute to resolve this issue, we ought to highlight a Spanish community-based study of incident cases that identified a specific wheezing pattern consisting of a single episode typically occurring before age 13 months, with the incidence peak at 7 months.32

On the other hand, both experts and clinical paediatricians agreed that AB is a term that should only be applied to a first episode. McConnochie20 was the first to include this criterion in his definition of AB in 1983. Several clinical guidelines currently include it in their definitions.25–27,29 Others simply warn that recurrent episodes call for the exploration of other diagnoses, such as asthma or “viral-induced wheeze”.10 Many clinical trials include only infants with a first episode to forestall potential questioning of their inclusion criteria.5

It is also worth commenting on the agreement found in Spain that AB may be diagnosed any time of the year, despite the fact that children admitted during the winter epidemic season have distinct characteristics: they are more likely to be infected by RSV, have more severe presentations, are more likely to have a history of tobacco exposure during intrauterine life, are less likely to have a family history of asthma and lower plasma eosinophil counts.33 Some definitions do include the season of presentation among the criteria that are important for diagnosis.6,25,34

In the online survey, we identified 3 factors associated with the clinical manifestations of AB that we labelled dyspnoea, coryza and auscultation. Of all 3, dyspnoea is the one that explains the largest proportion of the variance in diagnosis. These factors had unequal weights in clinicians of different subspecialties. We believe that these differences are partly due to overall differences in the severity of the AB cases managed in each health care setting.

Maybe the most surpraising finding in our study is the agreement on considering AB a distinct disease, which stands in contrast to the scarce agreement on what its characteristic features are. In recent years, the conceptualization of respiratory tract diseases has been changing, as their pathophysiology is broken down into components and researchers seek to identify phenotypes and endotypes that may respond differently to treatment.35 Acute bronchiolitis has also been subject to this process. Several severe AB phenotypes have already been identified, and research in this field continues to grow.36 Thus, some of the current evidence suggests that AB due to rhinovirus has specific clinical and epidemiological characteristics37 as well as a different long-term prognosis in relation to atopic asthma.38,39 A viral aetiology is emphasised in descriptions of AB or specifically included in the definitions of disease proposed by some guidelines,6,9,25–27,34 and many clinical trials only enrol children with infection by RSV.7 Guidelines do not recommend routine testing for identification of the causative agent. It is possible that detection of RSV would make clinicians lean towards a diagnosis of AB and treatment in adherence to current guidelines, but the benefits of this approach are not actually clear.40 In our panel of Spanish experts there was consensus that viral identification was useful in the diagnosis of AB, but clinical paediatricians did not agree with this opinion.

Our results can be compared with those of a recent study conducted in Portugal.17 In the latter, only 32% of paediatricians applied the “first episode only” criterion, and 76% set the age limit at 24 months. Portuguese paediatricians considered those 2 aspects less important for diagnosis compared to general physicians, which were also surveyed in the study.

Limitations. The Delphi method can be applied in different ways. We followed international recommendations19 for the selection of experts and in establishing criteria to stop the process and to define the presence of a consensus. As is the case of any form of expert opinion, the results of the consensus process are not based on rigorous evidence, which is a limitation shared by all the diagnostic criteria for AB proposed to date. The online survey methodology entails an inherent participation bias, which is reflected by the fact that nearly 94% of respondents reported frequently managing children with AB. This bias, however, is of little relevance, as the target population of the study was precisely paediatricians that manage children with AB in Spain. We expected participation to be lower among paediatricians who were less interested in this disease, which is why we offered the incentive of the draw for a congress registration, although there is no way of knowing what effect this measure had, if any. The method used for recruitment was flawed, given the gaps that may exist in the databases of the collaborating scientific societies and that some paediatricians may not check their electronic mail often. Although it would be difficult to assess how representative the sample was, we think it is important to consider that there were participants from all autonomous communities and from all subspecialties involved in the management of children with AB. On the other hand, the answers given in the survey may not provide an accurate reflection of real-world clinical practice, and investigation of actual practices was not within the scope of our study.

The criteria applied by Spanish paediatricians to diagnose AB are heterogeneous, differ between experts and clinicians, and vary according to factors such as paediatric subspecialty. There is no question that this can surely hinder adherence to the recommendations of clinical practice guidelines. Initiatives have been launched, such as the one by the European Respiratory Society (https://taskforces.ersnet.org/item/standardizing-definitions-and-outcome-measures-in-acute-bronchiolitis), to attempt to standardise the diagnosis of AB, but we ought to wonder whether the resulting standard would not be just as arbitrary and flawed as previous definitions. In our opinion, it would be more fruitful to direct our efforts to deepen our understanding of the heterogeneity of AB and other respiratory diseases and to apply diagnostic labels fitting clearly differentiated and easily identifiable endotypes.

This study was funded by the Fundación Ernesto Sánchez Villares (Project 02/2017).

The authors have no conflicts of interest to declare.