Article: Cunningham S, et al. “Oxygen saturation targets in infants with bronchiolitis (BIDS): a double-blind, randomised, equivalence trial.” Lancet. 2015 Sep 12;386(9998):1041-8. doi: 10.1016/S0140-6736(15)00163-4.
Link to Article: click here
Why this article?: The advent of pulse oximetry has been implicated as one of the main reasons why we have seen such a dramatic increase in hospitalization for infants with bronchiolitis over the last 20-30 years (without any other signs of increasing morbidity/mortality associated with the infection). And, oxygen saturation has been cited as one of the main determinants for LOS for children admitted with bronchiolitis. With differing recommendations between the US and UK on target oxygen saturation for these children (90% and 94%, respectively), the authors of this study aimed to assess whether a target of starting oxygen supplementation at SpO2 <90% was equivalent to a higher threshold of <94%.
Methods: The study was carried out by enrolling infants hospitalized at several different children’s hospitals in the UK over a span of 2 bronchiolitis seasons. They included infants ages 6 weeks to 12 months of age (corrected for prematurity) who were clinically diagnosed with bronchiolitis. Infants were excluded from the study if they had any of the following issues:
- born at < 37 weeks gestation and received oxygen within 4 weeks of admission
- cyanotic or hemodynamically significant heart disease
- CF or interstitial lung disease
- admitted to ICU
Children were randomized to a standard pulse oximeter or a modified pulse oximeter which displayed a value of 94% if the measured SpO2 was actually 90%. Infants were to be started on supplemental oxygen if the displayed oxygen saturation was < 94%.
The primary outcome for equivalence was time to resolution of cough.
Other outcomes measured included:
- time to feed adequately (> 75% usual)
- time to parental perception of “back to normal”
- SpO2 measured at 28 days
- timing of readiness for discharge and actual discharge
- adverse events (PICU transfers, readmits, etc)
Results: There was a total of 615 patients included in the study: 308 in the standard group, 307 in the modified group. The modified group did include significantly more boys than girls and also had more preterm infants than the standard group. Otherwise, there were no differences between groups.
- Median time to cough resolution was 15 days in both groups.
- SpO2 values at day 28 were not different between the groups.
- Infants in the modified group returned to adequate feeding 2.7 hrs sooner and were considered “back to normal” by parents one day sooner.*
- Transfers to ICU were greater in the modified group (13 vs. 8)*
- Readmissions to the hospital were higher in the standard group (26 vs. 12)*
- Supplemental oxygen was given to 73% of infants in the standard group and to 56% of infants in the modified group. Oxygen was used for significantly longer periods of time in the standard group (27.6 hrs vs. 5.7 hrs).
- Infants in the modified group were ready for discharge sooner (30.2 hrs vs. 44.2 hrs ) and were discharged sooner (40.9 hrs vs. 50.9 hrs).
*Not statistically significant
Conclusions: The authors conclude that using a target SpO2 of 90% or higher is as safe and effective as one 94% or higher and they also report some findings which indicate that outcomes may actually be better. But…the study probably deserves a little closer scrutiny.
Discussion: I think the first thing that needs to be addressed is the primary outcome: Time to resolution of cough. I was (and still am wondering) exactly why this was chosen as the primary outcome as I have a hard time seeing how modifying our threshold to start oxygen is going to impact the disease course (i.e. length of cough) for children with bronchiolitis and would not have expected a difference in the first place. The authors briefly address this stating that cough is “associated with airway inflammation and might be influenced by hypoxia.” Two articles are cited at that point, which upon my reading, both appear to outline the mechanisms of airway inflammation for children with viral LRTIs, with neither one commenting on the role of hypoxia as a factor. Using this measure as a primary outcome to determine equivalency seems very odd and is concerning when it is used to argue that the SpO2 targets are not only equivalent, but equally safe.
There are also a few other things I would’ve liked to have seen measured to make the study a little more robust. One thing that I would have loved to see would have been measures of clinical severity to ensure the groups were equal in disease severity, which would have an impact on many of their secondary outcomes. They could have also looked at other interventions to ensure there wasn’t any other possible confounding variables between the groups (such as frequency and method of nasal suctioning). While harder to measure, looking at patient comfort level could have also been considered to ensure infants weren’t being too bothered by a nasal cannula when their pulse ox was 92% while also ensuring they weren’t overly distressed by mild hypoxia of 90% which would have been treated with oxygen in the standard group.
When looking at other outcomes measured in this study, I think we should give some pause as far as the claim for safety goes with the higher rate of PICU transfers noted in the modified group (though this wasn’t statistically significant and if we’re keeping track of insignificant results, there were also 2 deaths in the standard group vs. 0 deaths in the modified group). Neurocognitive effects of hypoxia is usually brought up as a safety concern as well, but I tend to side with the authors on this that most of the evidence we have to date is encouraging that children with mild and brief hypoxia during short-lived illnesses is unlikely to cause any long-term consequences.
Despite having some concerns about the choice of primary outcome for this study, overall, my takeaway is that there appears to be no clear safety concerns and more potential benefits to having a lower threshold for starting oxygen; including having kids on oxygen less, discharging patients home sooner, parents feeling like their children get better faster, having children resume usual feeds sooner and having fewer readmissions.
Lastly, as I side note…I was impressed at the low rates of both antibiotic and bronchodilator use on arrival that was reported for this cohort of patients (about 8% and 5%, respectively)! Would like to know how they did that!