Bronchiolitis is a common self-limiting viral illness generally affecting children under 12 months of age. The illness is marked by acute inflammation of the mucous membranes of the nasal cavities (a coryzal illness) with a subsequent viral infection of the lower airway, associated with poor feeding, cough, increased work of breathing and hypoxaemia (low blood oxygen levels).
Normoxaemia and hypoxaemia
Normoxaemia is the range of oxygen levels within the blood of healthy individuals. An oxygen saturation of 94% or more is seen in 97.5% of the population up to 1500 m altitude, and 94% is commonly accepted as defining the lower limit of normoxaemia.1
Hypoxaemia (an oxygen saturation of < 94%) is common, particularly in respiratory disease, and results from poor ventilation or perfusion or both.
Clinical approach to hypoxaemia in respiratory disease
Cyanosis has been a core clinical sign of hypoxaemia for over three centuries. With ready supply of supplemental oxygen from the 1950s, cyanosis was corrected during disease. The clinical availability of arterial pulse oximetry in the 1980s, becoming ubiquitous in the 1990s, has enabled clinicians to have a more finely tuned understanding of arterial oxygenation. This precision, however, has vexed clinicians about how to interpret safely small changes in oxygen saturation. Clinical cyanosis is distinguishable at approximately 85% oxygen saturation, so what is the clinical impact of increments in oxygen saturation between 85% and 94%? Since the early 1990s, clinicians have grown accustomed to interpreting changes in oxygen saturation without an evidence base on which to do so; in acute bronchiolitis the rate of admission to hospital would double when clinicians were provided with scenarios depicting only a 2% difference in oxygen saturation (from 94% to 92%).2
Clinical response to supplemental oxygen in those with hypoxaemia
Supplemental oxygen is provided for both acute and chronic respiratory disease to treat hypoxaemia. There are different recommendations for oxygen saturation targets from which to supplement oxygen in developed and developing health-care settings. The evidence for recommendations is limited or absent.
In developed health-care settings, in adults with acute respiratory disease, it is recommended that a target oxygen saturation of 94–98% be maintained;3 however, supplemental oxygen has no effect on acute respiratory symptoms,4,5 and there is no evidence that it has any effect on duration of illness. Among adults with chronic respiratory disease (chronic obstructive pulmonary disease), domiciliary oxygen supplementation confers a survival benefit on those who are severely hypoxaemic but not on those with mild or moderate hypoxia.6 At the other end of the age spectrum, it has been found that preterm infants managed to a target oxygen saturation of 84–89% (median oxygen saturation attained 89%) were significantly more likely to die than those managed to a target oxygen saturation of 90–95% (median oxygen saturation attained 92%).7 The small differences in median oxygen saturation observed in this study highlight the potential for significant health implications from minor variance of oxygen saturation in vulnerable populations.
In developing health-care settings, it is recommended that a target oxygen saturation of ≥ 90% be maintained in those with acute respiratory disease.8 Provision of supplemental oxygen to meet this recommendation is associated with reduced mortality in children with acute respiratory infection.9 The World Health Organization has recommended that the oxygen saturation target in acute respiratory infection be derived from expert interpretation of oxygen dissociation dynamics combined with a pragmatic wish to optimise resource allocation for maximal reduction of mortality in resource-poor settings.10
Before the present study, there were no randomised studies assessing the role of oxygen supplementation in acute respiratory infection in children.
Hypoxaemia in acute viral bronchiolitis
One in five infants develops symptomatic bronchiolitis in the first year of life, and approximately 3% of all infants require hospital admission to receive supportive care,11 predominantly for help with feeding and to receive supplemental oxygen for hypoxaemia. Oxygen saturation below 92% leads to oxygen supplementation being provided in 70% of infants admitted to hospital with acute viral bronchiolitis.12 In 57% of infants admitted to hospital, hypoxaemia will prolong length of stay after all issues have resolved (i.e. feeding). Length of stay for all infants with bronchiolitis is a mean of 72 hours, but in those requiring supplemental oxygen the mean is 96 hours.12
Controversies in approach to hypoxaemia in bronchiolitis
Hypoxaemia is common in bronchiolitis; however, the majority of infants do not have clinical cyanosis. The controversies lie in how to address oxygen saturation values just below the limit of normoxaemia. Could there be clinical benefit from elevating oxygen saturation in this region to normoxaemia (as is suggested by adult recommendations for acute respiratory disease)? What harm may come from accepting oxygen saturation in this region during acute respiratory disease (because small changes in oxygen saturation targets cause significant harms in preterm infants)?
Since the early 1990s, the number of admissions to hospital with bronchiolitis have increased, while mortality from the condition has remained the same.13 Death occurs in approximately 0.9% of hospital patients with bronchiolitis, mostly in those with pre-existing conditions.14 Some have questioned whether or not the increase in admissions may be (in part) because of the increased use of oxygen saturation monitoring.15
Two recent evidence-based guidelines have considered oxygen saturation in bronchiolitis. The Scottish Intercollegiate Guidelines Network (SIGN) guideline No. 9116 considered that, with no evidence available, current practice should prevail and infants should be considered ready for discharge once normoxaemia (oxygen saturation ≥ 94%) has been restored. The American Academy of Pediatrics (AAP) Guideline on Bronchiolitis,17 published in the same year, also accepted that there was no evidence but considered (by expert opinion of the group) that infants could be discharged home once they are clinically stable with an oxygen saturation ≥ 90%. The recommendation of an oxygen saturation at discharge below normal ranges has been controversial.18 Since publication of these two evidence-based guidelines, there continues to be variation in practice and recommendations, possibly highlighting clinician uneasiness at the interpretation of oxygen saturation in sick infants. There are a variety of recommended oxygen saturation targets present within national guideline portals in the USA,19 the UK,20 Spain21 and Australia.22 In addition, guidelines have recommended a lower target for starting supplemental oxygen than for stopping, typified by current guideline synthesis provided by the Agency for Healthcare Research and Quality at the US Department for Health and Human Services.23 The clinical logic for the different targets to start and stop supplemental oxygen is not clear, particularly the lower oxygen saturation target for commencing supplemental oxygen at a time when infants are typically clinically less stable.
Potential health-care impact of clinical response to hypoxaemia in bronchiolitis
We performed a pilot study in 62 infants admitted to hospital with bronchiolitis who were provided with supplemental oxygen for low oxygen saturation.24 Nursing observations noted oxygen saturation in air and amount of supplemental oxygen (if required) every 2 hours during the hospital stay. The median time for infants to improve from 90% to 94% oxygen saturation in air was 33 hours (stability for 4 hours at each level was required). In some infants, oxygen saturation was stable at ≥ 90% before feeding had been re-established. Taking feeding into account identified that a move from discharge at ≥ 94% saturation to ≥ 90% could facilitate discharge 22 hours earlier (30%) than currently is the case. If projected to the 70% of infants with bronchiolitis requiring oxygen during their admission, this would be equivalent to 18,434 bed-days per year gained for the UK NHS (an equivalent US value would be 95,608 bed-days per year). This represents substantial cost savings to the UK NHS, and, hence, assuming no significant cost increases through implementation of this changed regime and equal or improved health benefits, this would represent a highly cost-effective intervention.
There is a significant pressure to improve hospital logistics and costs associated with bronchiolitis. In the USA, an estimated 149,000 infants were admitted in 2002, for an average of 3.3 days, at a total cost of US$543M. Paediatric hospitals are logistically challenged each winter, particularly during the 6-week peak of respiratory syncytial virus (RSV) infection, as infants who are clinically stable remain in hospital for oxygen supplementation. In response, some centres now provide short-term supplemental oxygen at home, using primary care (USA)25 or secondary care home nursing teams (Australia).26 The burden on primary care services and families of home oxygen is significant.27 Studies have not provided a health-economic perspective on this service development.28
Study aims
Bronchiolitis has no effective treatment.16,17 Care is supportive to those who need it. In Scotland, approximately 2000 children under 12 months of age are admitted each year with bronchiolitis.16 In the UK, in 2007, an estimated 28,728 infants under 12 months of age were admitted to hospital with bronchiolitis, similar to the number of all children (aged 0–14 years) admitted each year with acute asthma.
If the AAP recommendation of an oxygen saturation target of ≥ 90% were widely adopted and used as a target for both starting and stopping supplemental oxygen, this could reduce time in hospital for infants. To be acceptable to clinicians, however, the practice would need to be demonstrated to be safe, without clinical detriment and without undue additional burden on primary care or families caring for children at home during an earlier part of their illness. To answer this clinical question we performed a double-blind, randomised controlled, equivalence trial in children with acute viral bronchiolitis to determine whether or not a target oxygen saturation of ≥ 90% would be equivalent to ≥ 94% for resolution of illness and also to compare clinical, safety and parental outcomes.
