Things aren’t the way they used to be. When I was training, opportunities abounded for opportunities to intubate infants. Then we did away with intubating vigourous infants born through meconium and now won’t be electively intubating them at all. Then you can add in the move towards use of non-invasive respiratory support instead of intubating and giving surfactant and voila…you have the perfect barrier for training residents and others how to intubate. On top of all of this the competition for learning has increased as the skill that was once the domain of the physician has now spread (quite rightly) to respiratory therapists, nurses in some places and with the growth of residency programs (ours is now 2.5X larger than when I trained) the scarce chances are divided among many.
Enter the Video Laryngoscope
To be clear this isn’t a post to promote a product but rather an examination of the effectiveness of a tool. I am putting this out there recognizing the possibility that someone out there might have heard of or have been contemplating purchasing one of these items. Those that are quite proficient at intubation (likely trained in the “good old days”) would likely question the need for such a device but I believe the device isn’t really aimed at that group except to use perhaps as a teaching tool. It really is targeted (at least I think) for those who don’t perform the skill often.
Does use of the video laryngoscope improve success rates at intubation?
This question has had an attempt now at being answered by Parmekar S et al in their paper Mind the gap: can videolaryngoscopy bridge the competency gap in neonatal endotracheal intubation among pediatric trainees? a randomized controlled study. The study involved taking 100 pediatric residents and randomizing them into two groups. The first would use the videolaryngoscope (VL group) and then intubate using the standard technique of direct laryngoscopy (DL group). The second group started with DL and then changed to VL. Both groups took part in a training session on intubation and then participated in three simulation scenarios from NRP. The findings demonstrated a couple interesting things. The first as shown in the graph was that the group that started with the laryngoscope had a near 90% success rate compared to about 60% for the traditional approach. When the groups swapped though they were both equal in effectiveness. This suggests that by visualizing the airway with the VL students were able to identify structures better after doing so such that success was improved simply by having used the device.
The other finding worth mentioning was that when the times to intubation were looked at, there was no difference between the two groups at all. If the intubation success is no different, why might the times be the same? Having used the video laryngoscope myself it does take some getting used to. Rather than looking directly at the airway you find yourself looking off to the side and adjusting the approach that is in front of you to place the ETT. No doubt this does take some getting used to.
What I would have liked to see is a repeat assessment a week later after using a few more trials with the VL as I suspect once you are used to it the speed of intubation would improve as well. I suppose though we will have to wait a little while until someone does such work but as a means of improving success in intubation I believe this tool has something to add.
We are the victims of our own success. Over the last decade, the approach to respiratory support of the newborn with respiratory distress has tiled heavily towards non-invasive support with CPAP. In our own units when we look at our year over year rates of ventilation hours they are decreasing and those for CPAP dramatically increasing. Make no mistake about it, this is a good thing. Seeming to overlap this trend is a large increase in demand by learners as we see the numbers of residents, subspecialty trainees, nurse practitioners on the rise. The combined effect is a reduction is the experience trainees can possibly hope to obtain when these rarer and rarer opportunities arise. The result of all of this is that at least by my eyes (although we haven’t documented it) the number of attempts for intubations seems to be much higher than it once was. It is not uncommon to see 3-4 attempts or sometimes more whereas in days gone by 1-2 attempts was the norm. We do our best to deal with these shortages using simulation as an example but nothing quite compares to dealing with the real thing even if it comes close.
The Less Practice You Get The More Adverse Events You Can Expect
This is just the way it is. Perfect practice makes perfect and it has been well documented that intubations can lead to many complications such as desaturation, bradycardia, bleeding, airway edema from multiple attempts and a host of other issues. Hatch and colleagues first described their experience with 162 intubations in which they found adverse events in 107 (39%) with 35% being classified as non-severe and severe events in 8.8%. Not surprisingly one of the factors associated with adverse events was the need for multiple intubation attempts. Based on this initial experience they determined that as a unit they could do better and soon after undertook a series of PDSA Quality Improvement cycles to see if they could reduce these events and that they did. What follows are the lessons learned from their QI project and it is my hope that some or all of these ideas may help others elsewhere who are experiencing similar frustrating rates.
Steps To A Better Intubation
The findings of their QI study were published last month in Pediatrics in their paper Interventions to Improve Patient Safety During Intubation in the Neonatal Intensive Care Unit. The strategies they used were threefold.
- Standardized checklist before intubation – This used a “do-confirm” approach in which the individuals on the team “do” what they need to prepare and then confirm with the group that they are done. An example might be an RRT who states “I have three sizes of ETT ready with a stylet already inserted, surfactant is thawed and the ventilator is set with settings of … if needed etc”. Another critical part of the checklist includes ensuring that everyone knows in advance their roles and who is responsible for what.
- Premedication algorithm – Prior to this project the use of premedication was inconsistent, drug selection was highly varied and muscle relaxation was almost non-existent. The team identified from the literature that a standard approach to premedication had been associated with reductions in adverse events in other centres so adopted the same here using fentanyl with atropine if preterm and muscle relaxation optional.
- Computerized order set for intubation – interestingly the order set included prompts to nursing to make sure intervention 1 and 2 were done as well.
The results of there before and after comparison were numerous but I have summarized some of the more important findings in the table below.
||Period 1 (273 intubations)
||Period 2 (236 intubations)
|<10 intubations experience
The median number of attempts were no different but the level of experience in the second epoch was less. One would expect with less experienced intubators this would predict higher risk for adverse events. What was seen though was a statistically significant reduction in many important outcomes as listed in the table. I can only speculate what the results might have been if the experience of the intubators was similar in the first and second periods but I suspect the results would have been even more impressive. The results seem even more impressive in fact when you factor in that the checklist was used despite all of the education and order set 73% of the time and muscle relaxation was hardly used at all. I believe though what can be taken out of these results is that taking the time to plan each intubation and having a standard approach so that all staff practice in the same way reaps benefits. If you already do this in your unit then congratulations but if you don’t then perhaps this may be of use to you!
What About Intubation For INSURE?
We are in the process of looking in our own centre at the utility of providing premedication when planning to give surfactant via the INSURE technique. I couldn’t help but notice that this paper also looked at that very issue. Their findings in 17 patients all of whom were provided premedication were that only one could not be extubated right after surfactant. The one who was not extubated however was kept intubated for several hours without any reasoning provided in the records so it may well be that the infant could have been electively kept ventilated when they may have indeed been ready for extubation. The lesson here though is that we likely do not need to exclude such patients from premedication it will reduce the likelihood of complications without prolonging the time receiving positive pressure ventilation.
Whatever your thoughts may be at this time one of the first questions you should ask is what is our local rate of complications? If you don’t know then do an audit and find out. Whatever the result, shouldn’t we all strive to lower that number if we can?
In the spirit of full disclosure I have to admit I have never placed a laryngeal mask airway (LMA) in a newborn of any gestational age. I have played with them in simulated environments and on many occasion mentioned that they are a great alternative to an ETT especially in those situations where intubation may not be possible due to the skill of the provider or the difficulty of the airway in the setting of micrognathia for example.
In recent years though we have heard of examples of surfactant delivery via these same devices although typically these were only case reports. More recently a small randomized study of 26 infants by Attridge et al demonstrated in the group randomized to surfactant administration through an LMA that oxygen requirements were reduced after dosing. This small pilot provides sufficient evidence to show that it is possible to provide surfactant and that at least some gets into the airway of the newborn. This proof of concept though while interesting, did not answer the question of whether such delivery of surfactant would be the same or better than through an ETT. As readers of my blog posts know, my usual stance on things is that the less invasive the better and as I look through the literature, I am drawn to concepts such as this to see if they can be added to our toolbox of non or less invasive strategies in the newborn.
A Minimally Invasive Technique For The Masses?
This past month, a small study by Pinheiro et al sought to answer this question by using 61 newborns between 29 0/7 – 36 6/7 weeks and greater than 1000g and randomizing them to either surfactant via the INSURE technique or LMA. I cannot stress enough so will get it out of the way at the start that this strategy is not for those <1000g as the LMA is not designed to fit them properly and the results (to be shown) should not be generalized to this population. Furthermore then study included only those infants who needed surfactant between 4 – 48 hours of age, were on CPAP of at least 5 cm H2O and were receiving FiO2 between 30 – 60%. All infants given surfactant via the insure technique were premedicated with atropine and morphine while those having an LMA received atropine only. The primary outcome of the study was need for subsequent intubation or naloxone within 1 hour of surfactant administration. The study was stopped early after an interim analysis (done as the fellow involved was finishing their fellowship – on a side note I find this an odd reason to stop) demonstrated better outcomes in the group randomized to the LMA.
Before we get into the results let’s address the possible shortcomings of the study as they might already be bouncing around your heads. This study could not be blinded and therefore there could be a significant bias to the results. The authors did have predetermined criteria for reintubation and although not presented, indicate that those participating stuck to these criteria so we may have to acknowledge they did the best they could here. Secondly the study did not reach their numbers for enrolment based on their power calculation. This may be ok though as they found a difference which is significant. If they had found no difference I don’t think I would be even writing this entry! Lastly this study used a dose of surfactant at 3 mL/kg. How well would this work with the formulation that we use BLES that requires 5 mL/kg?
What were the results?
What do these results tell us? The majority of failures occurred within an hour of delivery of surfactant in the ETT group? How does this make any sense? Gastric aspirates for those in the LMA group but not the INSURE group suggest some surfactant missed the lung in the former so one would think the intubation group should have received more surfactant overall however it would appear to be the premedication. The rate of needing surfactant afterwards is no different and in fact there is a trend to needing reintubation more often in the LMA group but the study was likely underpowered to detect this difference. Only two patients were given naloxone to reverse the respiratory depressive effects of morphine in those given the INSURE technique so I can’t help but speculate that if this practice was more frequent many of the reintubations might have been avoided. This group was quite aggressive in sticking to the concept of INSURE as they aimed to extubate following surfactant after 5 – 15 minutes. I am a strong advocate of providing RSI to those being electively intubated but if the goal is to extubate quickly then I believe one must be ready to administer naloxone soon after extubation if signs of respiratory depression are present and this did not happen effectively in this study. Some may argue those getting the INSURE technique should not be given any premedication at all but that is a debate that will go on for years I am sure but they may have a valid point given this data.
Importantly complications following either procedure were minimal and no different in either group.
Where do we go from here?
Despite some of the points above I think this study could prove to be important for several reasons. I think it demonstrates that in larger preterm infants it is possible to avoid any mechanical ventilation and still administer surfactant. Many studies using the minimally invasive surfactant treatment (MIST) approach have been done but these still require the skill of laryngoscopy which takes a fair bit of skill to master. The LMA on the other hand is quite easy to place and is a skill that can be taught widely. Secondly, we know that even a brief period of over distension from PPV can be harmful to the lung therefore a strategy which avoids intubation and direct pressure to the lung may offer some longer term benefit although again this was not the study to demonstrate that.
Lastly, I see this as a strategy to look at in more rural locations where access to highly skilled level III care may not be readily available. We routinely field calls from rural sites with preterm infants born with RDS and the health care provider either is unable to intubate or is reluctant to try in favour of using high flow oxygen via mask. Many do not have CPAP either to support such infants so by the time our Neonatal Transport team arrives the RDS is quite significant. Why not try surfactant through the LMA? If it is poorly seated over the airway and the dose goes into the stomach I don’t see them being in any worse shape than if they waited for the team to arrive. If some or all of the dose gets in though there could be real benefit.
Might this be right for your centre? As we think about outreach education and NRP I think this may well become a strong reason to spend a little more time on LMA training. We may be on to something!
As those of you who have been following this blog are aware, I am always on the lookout for strategies that can help minimize blood work without sacrificing care in the NICU. At particular risk our the very premature infants in our units who for example at 1 kg have about 80-90 mL of blood. It does not take very many 0.5 – 1 mL “small” draws to create anemia. In a recent study (free article in link) of infants less than 1500g entitled A mathematical modeling approach to quantify the role of phlebotomy losses and need for transfusions in neonatal anemia, the authors studied 26 infants over a one month period. The results were staggering in that these infants experienced 138 +/- 21 blood draws with an average of nearly four transfusions per patient. While the authors do not specify what type of testing was done they did find a shocking statistic that 59% of the blood collected by weight of sample was discarded. This certainly stresses the point that we should aim to minimize the volume of sample collected in each case to that which is only necessary for the equipment to run. Furthermore, strategies to minimize sample draws should be utilized where possible and if accuracy permits point of care technology may further reduce volumes required and provide immediate results at the bedside. Lastly where possible, utilizing non-invasive technology to avoid blood draws needs to be explored when possible and was the subject of another post on Masimo non-invasive HgB measurement (http://wp.me/p5NWfD-1t).
Certainly in sick neonates whether they be term or preterm the drawing of blood gases to monitor ventilation contributes to the anemia of prematurity which often culminates in a transfusion. Sicker infants with greater lability due to respiratory compromise are deserving of optimal ventilation and this is achieved by monitoring pCO2 levels in arterial or venous samples. There have been different strategies employed to replace the sampling of CO2 by blood gas analysis which have not been very successful but there is one that I believe has promise that I will discuss at the end.
Transcutaneous pCO2 measurement was introduced in the 1980s. While this technology does allow measurement of pCO2 the variation between true arterial pCO2 and tcPCO2 can be wide making the technology difficult to implement on a consistent basis. In particular the accuracy in infants <28 weeks has been quite poor leading to increased numbers of arterial and venous samples to “check” ow closely the results correlate. As was described in 2005 by Aliwalas LL et al the technology in this group who actually have the highest number of blood draws does not meet the required standard to replace arterial pCO2 measurements (http://www.ncbi.nlm.nih.gov/pubmed/15496874)
Another method is of directly sampling exhaled CO2 in ventilated patients. Traditionally such measurements were taken with proximal gas sampling and in neonates in particular the results were discouraging. Problems encountered with proximal end tidal sampling were related to the lack of cuffed endotracheal tubes in part as the measured gas would be diluted with air in the presence of any leak around the tube leading to underestimation of true CO2 levels. Furthermore, in the presence of significant pulmonary disease the clearance of CO2 may be impaired such that the arterial pCO2 – ETCO2 difference may be quite large. For a review see the free article by Malloy and Deakins Are carbon dioxide detectors useful in neonates? The agreement between arterial and proximal sampling measured in this way has been quite variable and as such the technology has not really caught on to any great degree for monitoring ventilated infants. That being said it can be quite useful at determining if the endotracheal tube is in the trachea or esophagus. The presence of the waveform even if not yielding an accurate level confirms proper placement although where the tube sits in the trachea still needs confirmation.
The final method for sampling CO2 is the one which I believe holds the most promise for actually reducing blood draws and by extension risk of anemia and pain in the neonate. Kugelman and colleagues in Haifa, Israel published the following paper (free article in the link) A novel method of distal end-tidal CO2 capnography in intubated infants- comparison with arterial CO2 and with proximal mainstream end-tidal CO2. This creative study utilized a double lumen endotracheal tube which had been designed for surfactant installation and distal pressure measurement to instead sample pCO2 near the carina. This strategy was postulated to eliminate the issue with dilution of gas from proximal sampling and provide a closer measurement of true pCO2 when compared to arterial CO2 and proximal sampling. They studied 27 infants with varying degrees of pulmonary condition severity although most had RDS. When comparing the three methods of pCO2 measurement the following was found.
This demonstrates that while proximal measurement was quite poorly correlated with true arterial pCO2 the distal measurement was much more accurate. In fact the mean differences between arterial pCO2 and distal measurement was -1.5 mm Hg while that of proximal measurement -10.2 mm Hg albeit with wide confidence intervals. As found in other studies of proximal end tidal CO2 measurement, worse pulmonary disease correlated with worse accuracy as shown in table 2.
As the pCO2 rises above 60 the accuracy is less but remains much better than proximal measurements. Interestingly the same group has published an additional trial using high frequency ventilation and confirmed the measurements remain accurate. (http://www.ncbi.nlm.nih.gov/pubmed/22328495)
So what does the future hold? in VLBW infants one concern may be the internal diameter of the smallest double lumen tubes and the effect of upsizing to a larger tube and risk of subglottic stenosis. After a personal communication with Dr. Kugelman I understand that this has not been an issue in their unit as they tend to use these double lumen tubes in most if not all of the their infants. The accuracy is sufficient enough from my point of view that units should be able to implement this strategy at least in larger infants at first (those who would need a 3.0 ETT and larger) to see the effect on blood sampling. I suspect that one blood gas a day to determine accuracy in a given patient would be sufficient most of the time if the numbers were found to correlate well.
I would welcome feedback from people who work in units where this strategy has been utilized. How effective is it? Did it reduce your blood gas draws or increase them due to unreliability? Have you seen a rise in subglottic stenosis? Please send your feedback to either this site or at my Facebook page at www.facebook.com/AllThingsNeonatal.