Single Dose of Furosemide to Improve Respiratory Distress in Moderate to Severe Bronchiolitis

Randomization and Blinding:

Subjects will be randomly assigned by a 1:1 ratio to either furosemide or placebo. The Biostatistics Unit will develop a randomization procedure using a permuted block design. The randomization will be stratified by mode of delivery of furosemide: IV or PO. Blinding will be maintained at all times throughout the study. The only unblinded study personnel will be the pharmacy department who will dispense either of placebo or furosemide to each patient. Blinding will be broken only in the case of allergic reaction in which case a pediatric attending physician, not involved with the study (designated prior to study initiation) will be notified to access the intervention of that patient. If such an event should occur, the primary investigator and IRB will be notified by e-mail/phone, and an adverse event will be documented.

Materials / Methods:

All patients aged 0-4 years old with primary diagnosis of bronchiolitis on assessment by the Emergency Department will be evaluated for inclusion to study. The goal patient enrollment is 100 patients, which is required to achieve statistical significance (see below). As this study focuses on moderate (admitted) to severe (requiring mechanical ventilation) patients with bronchiolitis, further documentation would be completed on a data collection sheet. Randomization will occur upon consent to study. In collaboration with the Biostatistics Unit, patients will be randomly assigned by a 1:1 ratio to either furosemide or placebo. There will be a randomization procedure using a permuted block design. The randomization will be stratified by mode of delivery of intervention: IV or PO. The corresponding intervention (placebo or furosemide) will then be dispensed by pharmacy for administration in a blinded manner by the patient's nurse and/or physician. Either a single dose furosemide (1mg/kg or 10mg - whichever is less) will be administered orally PO (or intravenously [IV] if IV is present), or a placebo of equal volume. Ideally, the intervention dose will be administered as early as possible (immediately after decision to admit was made and consent obtained). Subsequent retrieval of that data will be done by study personnel. Data recorded on the data collection sheet can include: O2 Saturation/Weight (pre-intervention and, 2 and 4 hours post-furosemide/control), Age, Gender, Prematurity (\<32 weeks gestation), history of mechanical ventilation, history of wheeze / asthma, history of furosemide use, interventions prior to furosemide (racemic epinephrine, albuterol, or non-invasive ventilation), RSV / Influenza status, degree of dehydration (mild / moderate / severe), onset of symptoms prior to triage, +/- apnea, timing from initiation of visit until furosemide administration and/or non-invasive/invasive ventilator support, and reason for admission to ICU if necessary.

Considerations for all analyses:

All analyses will be carried out under the intention to treat principle; patients will be analyzed according to the groups to which they were randomized. (However, it is likely that the intention to treat population will be the same as the per protocol population.)

The primary analyses will be carried out for all patients. Mode of delivery of furosemide (IV vs. PO) will not be included in the analysis, as the aim of the study is to examine the effects of furosemide, regardless of mode of delivery. Randomization will be stratified by mode of delivery, so that the distribution of mode of delivery will be the same for the two treatment groups. Secondary subgroup analyses will be performed analyzing each mode of delivery (IV, PO), separately. These secondary subgroup analyses will be the same as the primary analyses.

Analyses for Specific Aims:

1. The association between treatment (furosemide) and control (placebo) groups will be compared using the chi-square test for the categorical outcomes (25% decrease in respiratory rate, intubation, etc).

Percent change in respiratory rate will be compared using the two sample t-test. If the usual assumptions required for the t-test are not met, then either a suitable transformation, or an appropriate non-parametric test (the Mann-Whitney test), will be used.

Length of stay (LOS) will be compared using the Mann-Whitney test. In the event that not all participants are followed until discharge from the hospital (in-hospital mortality, not discharged by end of study period), survival methods will be used (LOS will be estimated using the product-limit method, and the treatment groups will be compared using the log-rank test).

Sample Size Justification:

Based on clinical experience, we believe that very few patients who are admitted with bronchiolitis will have a response (as described in the primary outcome: a decrease in respiratory rate of 25% or more, or a return to their age appropriate respiration rate), at 2 hours post administration of placebo. To be conservative, it is assumed that the response to placebo will be no higher than 5%. It is believed that a clinically meaningful response rate would be 25% of patients at 2 hours after administration of furosemide. The proposed sample size of 49 subjects per group will yield 80% power to detect such a difference using a chi-square test with a 0.05 significance level.

Clinical Outcomes Justification:

Given the transient nature of bronchiolitis, as symptoms typically peak treatment is sought. Along with furosemide's half life of approximately 6hrs, there could be a correlation drawn between those children intubated (respiratory failure) or not, as it relates to when furosemide was given. Twenty-four hours is a reasonable observation period to correlate to the intervention. Intubation beyond that time point may infer that similar treatments of furosemide were warranted given the disease.