Methods
Volunteer neonatal intensive care unit (NICU) providers including neonatologists, fellows and neonatal nurse practitioners were recruited from four tertiary perinatal medical centres (University of Michigan C.S. Mott Children’s Hospital, Ann Arbor, Michigan, USA; Advocate Children’s Hospital, Oak Lawn, Illinois, USA; St. John Providence Hospital, Detroit, Michigan, USA; St. Joseph Mercy Hospital, Ann Arbor, Michigan, USA) for this randomised crossover simulation study.
Using a prestudy survey, participants indicated their professional role, years of clinical neonatology experience, self-perceived confidence in laryngoscopy skill (5-point rating scale) and preferred laryngoscope blade size for intubating an extremely preterm newborn. Using four different laryngoscope blades, participants orally intubated a high-fidelity neonatal manikin that simulated a 25-week gestation newborn (Premature Anne, Laerdal Medical, Wappingers Falls, New York, USA). The study blades included one size-0 Miller and one size-00 Miller blade made by two different manufacturers (Welch-Allyn (WA), Skaneateles Fall, New York, USA and Rusch-Teleflex (RT), Morrisville, North Carolina, USA) (figure 1). The blades were selected to reflect those used in routine clinical practice at participating centres. The sequence that each blade was presented was randomly allocated using a computer-generated table (http://www.graphpad.com/quickcalcs/randomize1.cfm) and assignments were concealed in opaque sealed envelopes. Each intubation was performed using an uncuffed 2.5 mm internal diameter tracheal tube (Mallinckrodt, Mallinckrodt Pharmaceuticals, St. Louis, Missouri, USA) and an optional intubating stylet (Sheri-I-Slip, Hudson RCI/Teleflex, Morrisville, North Carolina, USA). Participants assembled their equipment and positioned the manikin on an adjustable radiant warmer as they would for a typical neonatal intubation. To ensure standardisation and avoid bias, participants were not allowed to practice intubation with the study blades; however, they were shown a photograph of the manikin’s airway and vocal cords for orientation purposes. One investigator handed the participant each blade in the assigned order and was available to provide cricoid pressure if requested. Using each of the four blades, participants were allowed as many attempts and as much time as required to successfully intubate the manikin. An attempt was defined as insertion of the blade into the manikin’s mouth, regardless of whether the tracheal tube was inserted.5 The duration of an intubation attempt was defined as the interval from insertion of the blade into the manikin’s mouth until it was fully removed. All attempts using a single blade are referred to as a blade trial. A digital video camera (Logitech HD Pro C920, Logitech, Newark, California, USA) was used to record each intubation attempt for subsequent time analysis by an investigator who was masked to the blade sequence. The laryngoscopy time for each attempt was recorded. If more than one attempt was required, the data for each attempt were summed to calculate the total laryngoscopy time for each blade trial. Successful intubation was indicated by chest movement during positive-pressure ventilation as assessed by the on-site investigator. After each successful intubation, the participant was offered a rest period and asked to reset the intubating environment before the next blade trial. On completing all four blade trials, participants completed a poststudy survey that asked which single blade was most preferred and which was least preferred. Using a 5-point rating scale, participants were asked how accurately the manikin represented a premature newborn’s airway.
Figure 1Study laryngoscope blades. (A) Size-0 Miller Welch-Allyn (top) and size-0 Miller Rusch-Teleflex (bottom). (B) Size-00 Miller Welch-Allyn (top) and size-00 Miller Rusch-Teleflex (bottom). (C) Size-0 Miller Welch-Allyn (right) and size-0 Miller Rusch-Teleflex (left). (D) Size-00 Miller Welch-Allyn (right) and size-00 Miller Rusch-Teleflex (left).
We hypothesised that intubation using the size-00 blade compared with the size-0 blade would decrease the total laryngoscopy time. Secondary outcome measures included the proportion of trials with first attempt success in <30 s and the total number of attempts required to successfully intubate the manikin. Planned subanalyses included laryngoscopy time and success rate stratified by trial number and by the operator’s prestudy blade preference. For the primary analysis (size-0 vs size-00), total laryngoscopy time for each participant using the two blades of the same size were averaged. In secondary analyses, results for each of the four blades were analysed individually. The baseline laryngoscopy time and distribution were estimated from the study by Bismilla et al describing neonatal intubations at a high-risk perinatal centre.4 We defined a clinically significant difference as the interval between this baseline (50 s) and the current NRP intubation time standard (30 s).10 We estimated a sample size of 48 participants would be required to find this 20 s difference in mean laryngoscopy time when evaluating each of the four blades separately (m1=50 s, m2=30 s, SD=28 s, four groups with six pairwise comparisons, alpha 0.05, power 0.80).4 13
Continuous variables are reported as medians and IQRs. Group comparisons for independent continuous variable were conducted using the Kruskal-Wallis test. Paired continuous variables were compared using the Wilcoxon matched-pairs signed rank test or Friedman's test with Dunn’s multiple comparisons post-test. Categorical variables are reported as counts with percentages and outcomes were compared with the Χ2 test or McNemar’s test for matched-pairs with continuity correction. All analyses were two-tailed and a p value <0.05 was considered to be statistically significant. Statistical analyses were conducted using Prism V.6.01 (GraphPad Software, La Jolla, California, USA).
The study was reviewed and assigned exempt status by the University of Michigan Institutional Review Board. Written consent was obtained from all study participants.