In-vitro differences between flexible and semi-rigid catheters
Currently, there exist no high-quality real-live in-vivo data on the application of different catheters during LISA. In 2017, Rigo and colleagues already published in vitro data asserting that intratracheal catheter placement is both faster and easier when using a rigid device and that the LISAcath® was found to be the preferred semi-rigid catheter by neonatologists during experimentation on airway dummies [
18]. In our study, using a more sophisticated airway model to simulate anatomic conditions in extremely premature infants (< 28 weeks of gestation) we found similar results in students, nurses and doctors as found by Rigo. Laryngoscopy and soft flexible catheter handling with Magill forceps is thought to be especially challenging if there is no prior nasotracheal intubation experience [
6]. However, with a higher level of medical expertise and neonatal intensive care training, the time difference of laryngoscopy became negligible and we did not find any difference in laryngoscopy duration in doctors with profound intubation experience. In a pan European survey (> 300 neonatologists, and 37 countries) the majority of neonatologists (56%) use standard flexible infant feeding tubes during LISA procedures, 34% use vascular catheters and 15% suction catheters. The size ranged from 2.5 to 5 French and two thirds use Magill forceps to insert the catheter [
21]. Despite widespread use of flexible catheters inserted with forceps, when tested in vitro against semi-rigid catheters, the overwhelming majority of participants favored the latter, which could be introduced faster and without the use of additional devices [
5,
17,
18]. The vast majority of nurses in our study were female. However, since there were similar numbers of female and male doctors and students and we did not find a sex difference, we consider this issue irrelevant.
Clinical implications of laryngoscopy time and catheter properties
Currently, no in vivo data exist on tracheal catheterization speed using different catheters. In our study of 31 premature infants, laryngoscopy failed or was interrupted 12 (27.9%) times (in 11 cases the first attempt was interrupted), but there was no significant difference in the number of failed attempts between the groups. In literature, the average rate of failed laryngoscopy on the first attempt was reported to be 10–48% using a sample with similar demographics of premature infants [
7,
13,
22]. Laryngoscopy and intrapharyngeal instrument handling are major causes of discomfort and pain during surfactant application [
23]. However, most randomized controlled trials do not use a standardized premedication before LISA / MIST procedures [
6,
24]. Especially when opting to avoid the use of pharmacological and potentially respiration compromising sedatives it is crucial to use a fast and straightforward method to minimize laryngoscopy time [
18,
24]. Early and correct tube placement equals less pain and discomfort.
Additionally, laryngoscopy causes several deleterious physiologic responses including hypoxemia, bradycardia and intracranial hypertension. Hypoxemia itself is linked to prolonged laryngoscopy and subsequent airway obstruction [
25]. Bradycardia during intubation is mostly caused by parasympathetic stimulation of catheter and laryngoscope and baroreceptor response to increased blood pressure [
25]. This is a reason why some neonatology units premedicate infants with atropine in order to avoid dangerous drops in heart rate during LISA/MIST [
26,
27]. As in previous studies, in order to reduce the risk of post-procedural respiratory and cardiovascular depression, our protocol does not make use of medical sedatives, analgetics or narcotics and therefore, efficiency is imperative to minimizing patient discomfort and additional clinical risk factors [
17,
18,
26,
28]. In our patient collective, the majority of laryngoscopies could be performed in under 60 s with a low rate of failure which signifies the staff’s high level of LISA and intubation proficiency. We did not find a significant difference in speed using one catheter over the other when intubation was performed by experienced neonatologists. Further, no significant difference has been found when analyses are adjusted for birthweight which we consider the most important anthropomorphic factor impacting laryngoscopy compared to gestational age, sex or body length. Figure
2 does however emphasize that longer durations of laryngoscopy were associated with significant higher drops in peripheral oxygen saturation. Our data shows no difference in occurrence of low bradycardias (< 70 bpm) in either group, but heart rates of less than 100 bpm were exclusively found in group 2. In these cases, the semi-rigid catheter may have caused more intense vagal stimulation during laryngeal insertion than the softer feeding tube even when inserted with Magill forceps. Randomized controlled trials using LISA protocols similar to ours without the preliminary use of narcotic, sedative or vasoactive drugs, observed similar results as our study [
7,
13,
15]. These studies report that all infants remained stable, without significant bradycardias and desaturations when LISA was performed with either flexible feeding tubes or semi-rigid vascular catheters [
7,
13,
15]. However, these studies reported only physiological parameters during surfactant application (laryngoscopy and drug instillation) and did not compare procedure properties when using flexible and semi-rigid catheters. Notably, no infant experienced soft tissue lacerations or airway injury following laryngoscopy. Two patients from each group experienced IVH on the first day of life. A known possible complication, besides prematurity itself, increased intracranial pressure during and after laryngoscopy poses a risk for intraventricular hemorrhage [
29]. In our case, MRI and sonography results deemed hemorrhaging not to be associated with treatment and likely originating prior to birth.
The vast majority of institutions who perform LISA procedures use catheters that lack medical device licensing and CE labeling for this particular application [
21]. Besides forceps, newly developed catheter mounts have been tried to facilitate tracheal insertion of flexible tubes, but clinical in vivo data is still missing and devices are solely tested during research studies and not readily available [
4]. LISAcath® is currently one of the few catheters, ISO licensed to be used in preterm infants for surfactant administration. Especially in institutions with less experienced staff, its ease of use, and overall positive feedback by participants and official licensing makes it a valid choice.
Airway simulator and premature infant mannequins do not reflect NICU real-live resuscitation circumstances. Even though we used a highly sophisticated airway model in the first part of the study, laryngoscopy time differences might be different in real-live. Further, the single center design and a small sample size are major limitations of this study. Smaller differences in laryngoscopy times may be undetected due to the small sample size. There is no published or established definition of fast versus slow laryngoscopy. Our definition of procedure speed was calculated by dividing the difference of our fastest and slowest recorded laryngoscopy completion times into thirds. However, choosing a different group calculation, would yield different results. In addition, since standardized operating procedures may vary between institutions, our data predominantly reflects our department. Despite this, the fact that our department has adhered to the same protocol for preterm resuscitation since 2007, doctors and nurses have been well trained and therefore laryngoscopy speed alterations are likely to reflect true time differences. Further, this study does not account for personal preferences of the doctors performing the procedure which may impact their performance. However, because all study members use both catheter devices in day to day clinical work and randomization was done immediately before neonatal resuscitation, this effect is minimized.