Where can I find this paper?
What is this paper about (what is the research question)?
In the setting of simulated paediatric intubation, is use of a videolaryngoscope associated with higher “success” rates than traditional direct laryngoscopy?
Summary of the Paper
Design: Prospective case-control study (subjects acting as own controls)
Objective: To compare the effect of videolaryngoscopy performed by trained Emergency Physicians on outcomes related to intubation.
Control: Direct laryngoscopy performed using videolaryngoscope with video display off (equivalent to laryngoscopy with Miller 0, Miller 1 or Mac 3 blade for neonate, infant and adult respectively)
Exposure: Videolaryngoscopy performed using Storz videolaryngoscope, with participants directed to watch the screen rather than looking into the oropharynx.
Intervention: Attempted intubation of three mannequins (neonate, infant and adult) in randomised order with or without presence of “pharyngeal swelling ” (determined by coin toss).
Primary outcome: “Successful” intubation, defined as successful placement of an endotracheal tube in the trachea of the simulation mannequin before removal of the laryngoscopy blade. The outcome was determined by unblinded post-attempt review of 3-point video footage by study investigators.
Population: Fellows and faculty in Paediatric Emergency Medicine volunteering to participate at a single tertiary care paediatric hospital (Philadelphia, USA).
- Inclusion: Physicians with “adequate” training and experience with intubation
- Exclusion: Fellows in their first year who had not completed requisite airway management training in the operating room
Results: 26 subjects performed 156 intubations Data was available for 148 intubations.
- First attempt success in newborn: DL 20/24 (83%), VL 22/24 (92%). Difference 9% (95% CI: -9 to 28%)
- First attempt success in infant: DL 21/24 (88%), VL 17/24 (71%). Difference -17% (95% CI: -38% to 5%)
- First attempt success in adult: DL 10/26 (39%), VL 21/26 (81%). Difference 43% (95% CI: 18 to 67%, p=0.002)
First attempt success was significantly worse with VL than DL when pharyngeal swelling was turned on.
Paediatric emergency medicine providers had improved first-attempt success at intubation using videolaryngoscopy with a curved blade in an adult simulator; success rates with straight blades in neonatal and infant simulators were not significantly different between direct laryngoscopy and videolaryngoscopy.
On the study design
This was a small scale prospective study generating paired data. The subjects acted as their own “controls”, comparing normal DL technique with VL. There is considerable potential for selection bias in this study; subjects volunteered to participate. We can all imagine those of our colleagues most likely to volunteer to participate in this sort of study (and those who wouldn’t), and how this might skew or influence outcomes.
The scenario order was randomised, which increases the stringency of the methodology, but the lack of blinding of the investigators assessing the primary outcome is unfortunate as this might introduce an observation bias.
The introduction of the “pharyngeal swelling” component is an interesting one; the rationale for randomly including this complicating factor is unclear. While this might better represent “real life” clinical situations it does introduce a potential confounding factor for which the authors do not completely account. This might have been better placed in a separate study.
What were the results and what does this mean?
We can see from table 1 that there was a difference in time since last adult intubation between the “fellows” and the “attending physicians” which seems more significant than the for the other age groups, although there is no p-value given for this.
The investigators looked at comparative proportions of success between groups. The confidence intervals were wide but crossed zero in both the neonatal and infant mannequins, implying no evidence of significant difference in success rates. A bigger study would be expected to tighten these confidence intervals.
For the adult mannequin, the 95% confidence interval was similarly wide but did not cross zero. This suggests that the probability of observing a demonstrable difference in success rate between the groups if in fact there is no difference (the null hypothesis is true) is 0.2% (or 1 in 500). However the width of the confidence interval suggests that the “true” magnitude of the difference between the two observed groups could be between 18% and 67%.
What can we take from this paper into clinical practice?
Not much yet! The paper suggests that the use of a videolaryngoscope in the simulation setting does not significantly reduce intubation success in neonate and infant mannequins, and may improve performance in adult mannequins when used by skilled PED clinicians.
However, the controlled environment of the simulation room and the anatomy of the mannequin may well impact the realism of the situation, and the study is a proxy for clinical intubation success at best. More research is needed as the true value of videolaryngoscopy is still unclear.
More questions to ask
- How does the use of videolaryngoscopy translate into clinical ED practice, particularly in a pressured environment?
- Are these results reflected in physicians with less intubation experience?
- Would training/experience with the videolaryngoscope improve performance further?
Follow us on twitter: @PEMLit