P0479 - Sedative Use in the Setting of Critical Bronchiolitis

Background
Mechanical ventilation (MV) is often used for bronchiolitis in the Pediatric Intensive Care Unit (PICU), and typically requires analgesic and sedative medications to treat pain and agitation and to prevent dislodgement of life-sustaining equipment. Sedative side effects include bradycardia, hypotension, decreased respiratory drive, and withdrawal, but it is unknown if cumulative dosages of these medications within the first 36hrs of MV are associated with unfavorable outcomes.

Methods
With IRB approval, charts of PICU patients <2yrs with bronchiolitis who received MV between 5/2014 and 8/2018 were reviewed for demographics, comorbidities and administration (dose/kg) of medications of interest (chloral hydrate, dexmedetomidine, fentanyl, ketamine, lorazepam, midazolam, morphine, and propofol). Patients with a tracheostomy were excluded. As previously described, morphine and lorazepam doses were converted to fentanyl and midazolam, respectively. Duration of MV (primary outcome) and PICU length of stay (LOS; secondary outcome) were adjusted for illness severity by dividing the actual outcome by the value predicted using our previously described Critical Bronchiolitis Score, so that a “standardized” duration of MV less than 1.0 indicates a duration shorter than predicted. For fentanyl and midazolam, subjects were divided into four quartiles based on cumulative doses, and outcomes were compared between groups using Kruskal-Wallis ANOVA. For all other drugs, subjects were divided based on presence/absence of exposure, and outcomes compared using Wilcoxon rank-sum. Data shown as median (IQR) and p < 0.05 defined statistical significance.

Results
Among 74 patients, median age was 3 (1-7) months. Most (53%) were female and 38% had an identified comorbidity, most commonly prematurity. Median duration of MV was 7.8 (5.7-12.1) days and median ICU LOS was 11.0 (8.5-15.4) days.
Narcotics were used in 99% subjects in the first 36hrs of MV, but cumulative fentanyl dosage quartile was not associated with standardized duration of MV or ICU LOS. Similarly, benzodiazepine use (97%) in the first 36hrs of MV was not associated with either outcome. There were also no significant associations between standardized duration of MV and use of chloral hydrate (n=4), ketamine (n=4), or propofol (n=5) in the first 36hrs. Similar results were seen when medication usage in the first 72hrs of MV was analyzed.

Dexmedetomidine usage in the first 36hrs (n=3) was associated with shorter standardized duration of MV (0.297 [0.284-0.391] vs. 1.014 [0.693-1.663],p=0.006) and ICU LOS (0.599 [0.514-0.767] vs. 1.372 [0.970-2.347],p=0.014). Similar results were seen with dexmedetomidine use in the first 72hrs (n=7; 4 of whom were in the highest benzodiazepine usage quartile).

Conclusion
Cumulative dosages of narcotics and benzodiazepines early during MV were not associated with clinical outcomes, but dexmedetomidine use was associated with shorter duration of MV and shorter ICU LOS, even after adjusting for illness severity. Further study of sedative dosing impacting MV duration is needed.