OBJECTIVE. The goal was to establish the final supportive therapy determinants of hospital length of stay for bronchiolitis.
METHODS. A retrospective case study of a randomly selected 25% of subjects <1 year of age who were hospitalized with bronchiolitis between April 1, 2003, and June 15, 2005 (n = 129), was performed. Records of 102 admissions to the general wards were reviewed (77 respiratory syncytial virus positive). Length of stay, pulse oxygen saturation profile, oxygen supplementation, feeding support, and nasal suction were determined. Infants admitted to the PICU (27 admissions) were excluded.
RESULTS. The majority of patients presented with feeding difficulties (82% at admission). Oxygen supplementation was not indicated initially for the majority of infants (22% with mean pulse oxygen saturation of 94%). However, oxygen treatment was required by 70% of infants by 6 hours, whereas the mean pulse oxygen saturation decreased by an average of only 2%. Feeding problems were resolved for 98% of infants by 96 hours, followed by oxygen supplementation resolving with an average lag of 66 hours. The mean pulse oxygen saturation at discharge was 95%. There was no significant correlation between pulse oxygen saturation at arrival at the emergency department and subsequent oxygen requirements or length of stay.
CONCLUSIONS. Oxygen supplementation is the prime determinant of the length of hospitalization for infants with bronchiolitis. Infants remaining in the hospital for oxygen supplementation once feeding difficulties had resolved did not experience deterioration to the extent of needing PICU support.
- oxygen inhalation therapy
- length of stay
- clinical practice variations
Bronchiolitis is the most common lower respiratory tract infection in infants <1 year of age, with respiratory syncytial virus (RSV) being the causative organism in ∼75% of cases.1 For many infants, the disease is self-limiting and does not require medical care. However, 2.9% of all infants each year require hospitalization for treatment of bronchiolitis,2 at an estimated annual cost of more than $500 million in the United States alone.3 RSV-related hospitalization creates significant distress for infants and children, caregivers, and families.4
The options for treatment of bronchiolitis are limited.5,6 Evidence does not support the use of β-adrenergic receptor agonists, epinephrine, corticosteroids, ribavirin, antibiotics, or physiotherapy, and usually hospitalization is to provide supportive measures only. Such measures include assisted (nasogastric) feeding, suction of excess nasal secretions, and supplemental oxygen therapy for treatment of hypoxia.5,6 Usually patients remain hospitalized until supportive therapies are no longer required.
Hospital admission rates and hospital length of stay (LOS) values for bronchiolitis have increased in the past 25 years.7,8 During the same period, mortality rates among hospitalized infants with bronchiolitis and emergency department (ED) visit rates for bronchiolitis have remained relatively constant, at 1% and 26 visits per 1000 US population, respectively,7,9–11 which suggests changes in admission criteria rather than increased virulence of the associated viruses. Reasons postulated for the increase in admission rates include greater day care attendance at younger ages, survival of greater numbers of premature infants, and increased sensitivity of pulse oximetry for detection of hypoxia (compared with clinical observation).12
Pulse oximetry, although first commercially available in 1976, was not in widespread use until the late 1980s. Pulse oximetry monitors are now used ubiquitously for hospitalized patients with respiratory disease, to provide early and ongoing assessments of possible hypoxia. The clinical expectation for pulse oximetry is that it will facilitate early identification of infants with impending significant hypoxia, to enable proactive early intervention.13 Before the widespread use of pulse oximetry, clinicians were limited to the use of clinical cyanosis (significant hypoxia) and intermittent blood gas determinations to guide oxygenation decision-making, with obvious clinical risks to patients.14
Clinicians are now influenced significantly in their decision-making by pulse oxygen saturation (Spo2) values. In one study, pediatricians working in EDs would change their admission rate for bronchiolitis from 43% to 83% with only a 2% change in Spo2 values (from 94% to 92%).15
There are no accepted Spo2 criteria for admission or discharge of infants with bronchiolitis, and practices vary widely around the world (Spo2 values of 89%–96%).11,16–19 To facilitate earlier discharge of patients recovering from bronchiolitis, some clinicians have provided supplemental oxygen therapy at home until Spo2 values are considered “acceptable.”20
It was suggested previously that there is a strong relationship between admission Spo2 values and hospital LOS.16,21 However, the only study we could identify that attempted to assess the role of Spo2 monitoring in determining LOS for infants with bronchiolitis included relatively few infants and was substantially limited by the lack of agreed policy for starting or stopping supplemental oxygen therapy.22 As a consequence, supplemental oxygen therapy could be stopped variably (as ordered by the physician of the day) at Spo2 values of 89% to 97%, which made any association between supplemental oxygen therapy and LOS difficult to ascertain.
We wished to assess the relative impact of supportive therapies in determining hospital LOS for infants admitted to the hospital with acute viral bronchiolitis. The aim was to identify the number of infants for whom oxygen supplementation was the final determinant of LOS, with a view to considering how best such infants could receive their care in the future and where to aim future research.
Study Design and Subjects
This was a retrospective, cohort, observational study of infants (<1 year of age) admitted from the ED to a tertiary care university hospital with a diagnosis of bronchiolitis. A total of 495 infants were admitted during the study period (between April 1, 2003, and June 15, 2005). A random 25% sample (determined with SPSS 14 for Windows; SPSS, Chicago, IL) was extracted, providing an assessment of 129 admissions. Hospital charts, together with information from a hospital patient information database, were used to extract data for predefined data collection sheets. Admissions to the PICU were excluded.
One trained abstractor extracted key data elements from each set of hospital charts, including data on demographic features, LOS, PICU admission, and virologic features. The ED triage time and hospital discharge order time were used to calculate the duration of hospitalization to the nearest hour. Clinical details at presentation to the ED were noted, including Spo2, social issues (leading to admission), and the presence or absence of feeding difficulties, suction requirement, and apnea.
One hundred two admissions were to the acute medical wards. Review of the hospital charts was performed to assess the supportive therapies that had been required during each 6-hour block of time, starting with the time of arrival at the ED. Six-hour blocks were used because we considered that this would be a sufficient length of time to smooth out short-term variability in the outcomes assessed. For each 6-hour period, a note was made regarding whether each infant required (1) nasal suction, (2) nasogastric tube feeding, or (3) oxygen supplementation; in addition, it was noted whether any apnea occurred and whether intravenous fluids were required. We noted Spo2 values and any unresolved feeding problems and/or social issues at discharge.
Children admitted during the study period were provided with supportive care only. The use of medicines for treatment of bronchiolitis was discouraged as ineffective.5,6 Infants were provided with nasal suction for excess nasal secretions or apnea, nasogastric tube feeding for feeding volumes <75% of normal or excessive work of breathing, and supplemental oxygen (through nasal cannulae) for Spo2 of ≤93% in air. Supplemental oxygen therapy was stopped once values for Spo2 in air were >93%. Spo2 monitoring was continuous (Nellcor model N-550 oximeter; Nellcor, Pleasanton, CA), with an 8-hour observation period after removal of supplemental oxygen treatment before discharge (if there was no other reason for the patient to remain in the hospital). During oxygen supplementation, Spo2 values were assessed in air at 2-hour intervals, for consideration of weaning/discontinuation.
Given the objective nature of these data, we did not blind the abstractor to the study objective. All data were transcribed onto standardized study forms, and the process was monitored by Dr Cunningham.
Data were analyzed by using Microsoft Excel X for Mac Service Release 1, with additional analysis with SPSS 14 for Windows (SPSS, Chicago, IL). Summary statistics were calculated, together with the Spearman rank correlations between LOS and duration of nasogastric tube feeding, oxygen supplementation, and nasal suctioning and social problems.
For each admission, the times to resolution of the supportive therapies were identified and the difference between these times (“lag”) was calculated. Summary data are presented for patients for whom the final support to be resolved was oxygen supplementation or nasogastric tube feeding.
One hundred twenty-nine admissions were assessed. The mean age of admitted infants was 23 weeks (interquartile range [IQR]: 10–33 weeks), with a mean LOS of 79.5 hours (IQR: 22–116 hours). Twenty-seven admissions were directly to the PICU from the ED and therefore were excluded from the study population.
One hundred two admissions (81.6% of all admissions; mean age: 24 weeks; IQR: 11–37 weeks) were to the acute receiving pediatric ward. Demographic features of the study population are provided in Table 1 and were comparable to those in previous studies.9,21,23 None of the infants in the study population experienced deterioration to the extent of PICU admission. Overall, feeding problems were experienced by 82% of admitted infants (n = 84) and 70% required supplemental oxygen (Spo2 of ≤93%; n = 71), with only 40% requiring nasal suction (n = 41). Social factors were not identified as a reason for admission in any case.
The mean LOS was 72 hours, with a range of 6 to 371 hours (IQR: 20–104 hours). There was a strong correlation between the duration of oxygen supplementation and LOS (r = 0.891; P < .001) (Fig 1). Correlations of LOS with the duration of nasogastric tube feeding (r = 0.441; P < .001) and the need for nasal suction (r = 0.567; P < .001) were less strong. There was no significant correlation between admission Spo2 levels and LOS (r = −0.101; P = .398).
Feeding problems resolved by 96 hours (4 days) in 98% of cases (Fig 2). The mean duration of feeding problems was 27 hours (IQR: 6–42 hours). Intravenous fluids were not required for any infant. The mean Spo2 value in the ED was 94% (IQR: 92%–97%). As a consequence, only 22% of infants presented with an immediate need for supplemental oxygen. The proportion requiring supplemental oxygen increased to 70% by 6 hours, with the mean Spo2 value decreasing by an average of only 2% (SD: 4%). The correlation between Spo2 at 6 hours and LOS was not significant (r = 0.110; P = .464). It must be presumed that the increase in oxygen requirements by 6 hours was not attributable to acute deterioration in such a large proportion of cases. Continuous Spo2 monitoring after admission might have identified lower values for such infants. Oxygen supplementation resolved (Spo2 of >93%) by 180 hours (7.5 days) in 98% of cases. The mean length of oxygen support was 56 hours (IQR: 0–96 hours). The mean Spo2 value at discharge was 95% (IQR: 95%–97%).
The need for oxygen was the final determinant of LOS in 58 cases (57%). The average time (lag), after all other problems had resolved, until supplemental oxygen was no longer required was 66 hours (2.75 days) (Fig 2). Feeding problems were the final determinant in only 27 cases (26%). Feeding and oxygen problems resolved simultaneously in 13 cases (13%). The duration of problems was not documented in 3 cases (3%), and feeding problems were not resolved at discharge (unrelated to the bronchiolitis) in 1 case (1%). In the group for which oxygen supplementation was the prime determinant, the mean LOS (94 hours; IQR: 52–129 hours) was significantly greater (P < .001) than that in the feeding problem group (30 hours; IQR: 16–35 hours), although the mean duration of oxygen supplementation was 56 hours (IQR: 0–96 hours).
This study identified the relationship of supportive measures for the hospital-based treatment of acute viral bronchiolitis to LOS and ascertained that oxygen supplementation was the principle determinant of LOS for 57% of infants admitted to acute receiving wards. For such infants, the additional LOS after resolution of feeding difficulties and nasal suction was, on average, 66 hours (2.75 days). Once oral feeding had been reestablished, infants in this study did not experience regressive deterioration to the extent of requiring PICU care.
This study assessed a large cohort over 2 winters, to provide a representative sample of infants with acute viral bronchiolitis. Our hospital policy for the treatment of acute bronchiolitis limits unnecessary variability in care; in particular, infants are not exposed to ineffective treatments (inhaled or orally administered corticosteroids, inhaled bronchodilators, or nebulized epinephrine), and oxygen supplementation is initiated and discontinued when Spo2 values in air are ≤93% and >93%, respectively.5,23
The observational nature of this study might have introduced bias, but we aimed to reduce this with the use of predefined data collection sheets and criterion-based definitions. The use of 6-hour time windows enabled us to cope with missing data points in nursing observation charts, which the use of specific times after admission would not have done. Another problem with the observational nature of this study was that we were unable to determine the unsupplemented Spo2 values for these infants, because the values were assessed but not charted. Such information might very helpfully demonstrate the rate of improvement in Spo2 values over time during recovery from bronchiolitis, providing a profile of Spo2 values to demonstrate what values are associated with most of the LOS prolongation.
None of the infants experienced apnea on the ward. It was noted previously that apnea is an initial manifestation of bronchiolitis and that, if it does not develop within the first 2 to 3 days of the hospital stay, then it is unlikely to occur.24 Our study demonstrated that infants who required admission to the PICU did not do so during the recovery phase, once feeding had been reestablished. This may help those who wish to plan for earlier discharge for infants who are slow to recover to satisfactory Spo2 levels once feeding has been reestablished.20
In comparison with previous studies, we had a larger sample size, randomized from a large sample of 495 infants over 2 years, and had no ambiguity in the treatment of those infants.22 Our study population was comparable to those of some previous studies with respect to LOS and RSV status.9,21,23 However, most previous studies involved children <2 years of age, rather than infants (<1 year of age) as in our study.9,21 Peak incidence occurs in children <12 months of age, and 3 months is the mean age of infants hospitalized with RSV infection.2,25 The average LOS in our study was 3 days, which is shorter than that reported by Wilson et al17 but similar to findings of other previous studies.9–11 Schroeder et al22 suggested that only 26% of children <2 years of age had a prolonged stay because of a “perceived need for oxygen.” The average lag in their study was 1.6 days. However, the management of bronchiolitis in that study was inconsistent, with variable accepted Spo2 levels and aerosol treatments, similar to other studies,16 and it is not possible to extrapolate the results of that study to other infant cohorts.
It was suggested previously that the need for oxygen at admission predicts strongly the LOS.7 In our study, there was no correlation between admission Spo2 values and LOS, whether determined at admission or at 6 hours. This is similar to the findings by Brown et al.21
Our study assessed infants with bronchiolitis by using continuous Spo2 monitoring. Continuous oximetry may provide useful clinical data, but it has been postulated to increase LOS,26 with pulse oximetry suggested to be one of the main reasons for the increase in hospital admissions in the past 15 years.12 The use of intermittent Spo2 monitoring has been proposed for infants recovering from bronchiolitis; to date, however, there have been no studies to determine what effect such monitoring has on time to discharge or at what Spo2 level intermittent monitoring should start and with what frequency and for how long Spo2 should be checked.5,6
There is no consensus regarding what constitutes safe appropriate Spo2 limits for admission, inpatient care, and discharge for infants with bronchiolitis. Guidelines and practices vary significantly among hospitals, regions, and countries.16–19,25,27 Accepted Spo2 values range from 89% to 96%.5,6,9,19,28 Slight differences in accepted Spo2 levels influence admission rates hugely.15 Our data demonstrated that a 2% decrease in mean Spo2 levels 6 hours after admission increased threefold the proportion of infants receiving supplementary oxygen. With our data, we were unable to suggest by how much LOS could be reduced by decreasing the accepted Spo2 level by 3% (for example), because of insufficient data on Spo2 levels measured in air while the infants were receiving supplementary oxygen. The effects of lower Spo2 limits on feeding, time to clinical recovery, and subsequent health would need to be explored.
The lag between the resolution of feeding problems and the need for oxygen supplementation is sizeable and poses the following question: what is the role of oxygen supplementation in the recovery phase of acute viral bronchiolitis once feeding problems have resolved? Although oxygen is one of the most commonly administered drugs in hospitals, there are no published studies on the effects of supplemental oxygen therapy on the recovery and subsequent health of children with acute respiratory infections. The health and social costs of such prolonged stays in hospital for infants with bronchiolitis justify such a study.
- Accepted August 20, 2007.
- Address correspondence to Steve Cunningham, MBChB, MRCPCH, PhD, Department of Respiratory and Sleep Medicine, Royal Hospital for Sick Children Edinburgh, Sciennes Rd, Edinburgh, EH9 1LF, United Kingdom. E-mail:
Financial Disclosure: Dr Cunningham received a fee from Abbott Laboratories for providing a nonpromotional postgraduate lecture on bronchiolitis.
- ↵Deshpande SA, Northern V. The clinical and health economic burden of respiratory syncytial virus disease among children under 2 years of age in a defined geographical area. Arch Dis Child.2003;88 (12):1065– 1069
- ↵Pelletier AJ, Mansbach JM, Camargo CA Jr. Direct medical costs of bronchiolitis hospitalizations in the United States. Pediatrics.2006;118 (6):2418– 2423
- ↵Leidy NK, Margolis MK, Marcin JP, et al. The impact of severe respiratory syncytial virus on the child, caregiver, and family during hospitalization and recovery. Pediatrics.2005;115 (6):1536– 1546
- ↵Scottish Intercollegiate Guidelines Network. Bronchiolitis in children: a national clinical guideline. Available at: www.sign.ac.uk/pdf/sign91.pdf. Accessed January 12, 2008
- ↵Lieberthal AS, Bauchner H, Hall CB, et al. Diagnosis and management of bronchiolitis. Pediatrics.2006;118 (4):1774– 1793
- ↵Langley JM, LeBlanc JC, Smith B, Wang E-EL. Increasing incidence of hospitalization for bronchiolitis among Canadian children, 1980–2000. J Infect Dis.2003;188 (11):1764– 1767
- ↵McMillian JA, Tristram DA, Weiner LB, Higgins AP, Sandstrom C, Brandon R. Prediction of the duration of hospitalization in patients with respiratory syncytial virus infection: use of clinical parameters. Pediatrics.1988;81 (1):22– 26
- ↵Mallory MD, Shay DK, Garrett J, Bordley WC. Bronchiolitis management preferences and the influence of pulse oximetry and respiratory rate on the decision to admit. Pediatrics.2003;111 (1). Available at: www.pediatrics.org/cgi/content/full/111/1/e45
- ↵Willson DF, Horn SD, Hendley JO, Smout R, Gassaway J. Effect of practice variation on resource utilization in infants hospitalized for viral lower respiratory illness. Pediatrics.2001;108 (4):851– 855
- Wang EEL, Law BJ, Boucher FD, et al. Pediatric Investigators Collaborative Network on Infections in Canada (PICNIC) study of admission and management variation in patients hospitalized with respiratory syncytial virus lower respiratory tract infection. J Pediatr.1996;129 (3):390– 395
- ↵Weiss J, Annamalai VR, Willson DF. Discharge criteria for bronchiolitis patients. Pediatrics.2003;111 (2):445
- ↵Bajaj L, Turner CG, Bothner J. A randomized trial of home oxygen therapy from the emergency department for acute bronchiolitis. Pediatrics.2006;117 (3):633– 640
- ↵Brown L, Reiley DG, Jeng A, Green SM. Bronchiolitis: can objective criteria predict eligibility for brief hospitalization? Can J Emerg Med.2003;5 (4):239– 244
- ↵Christakis DA, Cowan CA, Garrison MM, Molteni R, Marcuse E, Zerr DM. Variation in inpatient diagnostic testing and management of bronchiolitis. Pediatrics.2005;115 (4):878– 884
- Copyright © 2008 by the American Academy of Pediatrics