OK here is the article
Video Laryngoscopy is Associated With Increased First Pass Success and Decreased Rate of Esophageal Intubations During Urgent Endotracheal Intubation in a Medical Intensive Care Unit When Compared to Direct Laryngoscopy
These are the data now published that Dr Paul Mayo presented during his epic debates with Dr Scott Weingart. If you have not checked them out they are awesome!
I must admit at the time of the debates, Dr Mayo was very convincing in his arguement for sedative only intubation using a comprehensive team approach and VL. But reading their data and results now fully published, I am less than impressed. Sure they get the tube done but it carries with it a lot of complications . It may well be their data reflect a more detailed analysis of real world complications in an ICU but still some weird things I cant understand like the lack of paralytics for as far as I can tell philosophical reasons and the exclusion of anaesthesia doctors from teaching and supervising emergency intubations. ONCE again I sent out a call for comments from FOAMEd . Here is what I got …
A) Esophageal intubation rate of 19% in the DL group seems awfully high. In the big NEAR database (PMID 20434289), the esophageal intubation rate is <3% overall (245 immed recognized + 19 delayed recognized in 8937 intubations; only 69% were RSI and most of the rest were without any meds, presumably crash intubation)
-Also from NEAR: RSI 82% successful on 1st attempt, 95% on first 3, and 96% overall
compare to sedation only, which was 76% on 1st, 90% in 3, and 91% overall success.
-This NEAR study collected data from 1997-2002, and doesn’t report rates of VL, but presumably it’s very low.
-Sedation-only intubation is the worst of all the worlds. You get:
1) suboptimal conditions
2) with airway reflexes intact (i.e. they will move or vomit when you jab them in the throat)
3) the level of sedation required will often obliterate respirations (Jim DuCanto described this well: it only takes a little sedative to overpower the higher brain functions, but it takes a LOT to overpower the brainstem)
If they’re not breathing on their own anyway, then it’s a far better choice to both optimize conditions and remove airway reflexes. Of course if Mayo could show that SO intubation was just as good then that would beat my argument by physiology, which brings me to:
The 6-12% (10% overall) rate of hypoxia during intubation in Mayo’s study seems high. I tried finding some data to compare and unfortunately NEAR doesn’t report pulse ox (understandable).
In the GNYHA RSI vs SO debate, Mayo showed a 13% hypoxia rate (http://emcrit.org/podcasts/paralytics-for-icu-intubations/ at 13:38) which is a bit higher.
To be fair to Mayo, the best I could find in the ED had similar rates of hypoxia (9.4% in PMID 16373805), and I think the lessons from the recent well-done Peds airway studies suggest that we underestimate complications in emergency intubations, and Mayo’s methods were pretty rigorous as well and are unlikely to “undercount” complications.
Walls and others have demonstrated that RSI is a superior approach in the ED, and that was before VL, NODESAT, improved airway education, etc. It’s just important that we know how to handle paralyzed patients we can’t intubate (i.e. BMV, SGA, cric).
Seth Trueger, MD
Health Policy Fellow
A few issues come to mind on reading this:
Benefit of VL: when looking at the benefits of VL we are often actually looking at the combined impact of 2 separate things – firstly having a videolaryngoscope and secondly the shape of the blade to which it is attached. As such VLs are not generic devices & the interpretation of the benefits of VL must be applied specifically to the device being used in the study.
I think it’s clear on face value just from using VLs that employ a Mac blade such as the C-MAC (where the confounding element of blade shape is removed), that VL via a Mac blade improves intubation success (at least in clinicians experienced in the technique for using a Mac blade). Given that so many of us have had situations where DL has failed but we’ve been able to intubate with a C-MAC (adding only VL but with no “trade off” due to changed shape of the blade) this simply must be true – the DL element remains constant with the C-MAC and so would always have resulted in the same degree of success even with a non-VL Mac blade, but the VL component allows us to place the ETT in patients with who we would otherwise have had difficulty with DL alone. In other words the C-MAC & other Mac blade VLs are both DLs & VLs at the same time. From a Vortex perspective they offer a benefit in that one doesn’t have to choose between DL & VL and can therefore minimise the number of attempts by attempting DL & VL simultaneously. We don’t need research to tell us that VL improves intubation success in this situation, though it enables us to quantify by how much it might improve it.
With other VLs the nett benefit is less clear without studies such as these, as whilst the use of VLs with non-standard blades clearly offers benefit in some patients (perhaps some in whom a Mac style VL might also be unsuccessful) it is unclear whether the different blade shape/technique offsets this in other patients making laryngoscopy more difficult. A choice therefore has to be made between intubating via DL or VL for the first attempt when VLs with non-standard blades are used.
This study illustrates the above competing issues, demonstrating that, in novices at least, laryngoscopy success tends to improve overall when using the Glidescope’s non-standard blade. Similarly it demonstrates that in cases where VL with a non-standard blade fails, DL with a Mac blade may still be used successfully.
I don’t think this necessarily translates to the fact that “DL is dead” because:
The novice element is signifcant as it is feasible that Glidescope VL is less dependent on laryngoscopy technique than DL with a Mac blade, allowing an increased success rate in novices which may not necessarily translate to those experienced in Mac blade DL technique.
As Reuben has pointed out, there are situations where VL may not be available or successful, so DL technique remains an important skill.
If we don’t train airway clinicians in DL we may end up just trading the benefits of VL for a loss of skill at laryngoscopy technique with no nett change in intubation success. To really reap the benefits of VL we need to superimpose them on top of our best laryngoscopy technique, not use them as a substitute for this.
A couple of cautions about VL in general:
In my experience VL often improves the view but not always passage of the ETT. This is particularly true where an experienced laryngoscopist has failed at DL implying that the issue is anatomical (very anterior larynx) rather than technique related (normal larynx but poor laryngoscopic technique preventing view being obtained) where once a view is obtained, the alignment of structures still allows easy passage of the ETT.
The trainees in this study were supervised by the attending who could intervene if necessary. Whilst they illustrate the potential value of VL in circumstances where by necessity novices are forced to be responsible for intubation, I’d be concerned about these results being used to justify cutting corners in training staff in laryngoscopy before releasing them to unsupervised practice. One of my concerns in general with VLs is that they might increase the threshold for junior staff to call for help so that a trainee who is relatively inexperienced or faced with a patient who might possibly be difficult prompts the trainee to think “maybe we should get help – oh no, don’t worry we’ve got the VL”. The more appropriate response is that if trainees recognise that their skills or the patients anatomy suggest that DL might not be adequate and that VL might be more necessary, that this becomes a trigger for seeking help as the case is not straightforward.
Overall these results need to be interpreted in terms of the study group being novices and specific to the use of the Glidescope VL.
No muscle relaxant: no reason is given for not using muscle relaxant for intubation of these patients. Presumably it was perceived that there was some added safety in not paralysing these patients in case intubation was difficult. If so I think this is spurious reasoning creating a false sense of security. The problem with potential interruption to alveolar oxygen delivery during intubation arises because the patient becomes apnoeic & loses muscle tone, regardless of the mechanism for this. Certainly induction agents/opioids/BZD’s given alone, particularly in the sort of unwell patients who would have required UEI in this study, can induce apnoea and/or loss of airway tone for long enough periods to cause an interruption to alveolar oxygen delivery sufficient to produce critical desaturation (and likely this period of interruption to AOD would not be reliably shorter than the time taken for recovery from sux or within which reversal or roc can be undertaken using suggamadex). In Vortex terms, it is induction, not necessarily paralysis, that takes you out of the “green zone” and into the funnel of the Vortex. Whilst not paralysing the patient offers no real benefit it certainly adds some potential dangers though:
Vomiting: without paralysis, if insufficiently sedated, there is the potential for laryngoscopy to precipitate vomiting and possible aspiration.
Increased difficulty obtaining a non-surgical airway: in the absence of paralysis FM ventilation, insertion of a SGA and ETI all become more difficult. Thus not using relaxant provides no increased margin of safety for spontaneous recovery of airway patency/ventilation over that derived from using short acting or rapidly reversible muscle relaxants, but increases the likelihood of failure of all non-surgical airway techniques and the likelihood of needing to progress to an emergency surgical airway.
Airway Training: as an aside this study illustrates that even intensive training of staff in DL over a short period using mannequins alone is not effective. Mannequins aren’t good for teaching the subtleties of laryngoscopy and typically reward bad laryngoscopic technique in achieving a view of the larynx. The oesophageal intubation rate of 1 in 5 intubations is VERY high. The occurrence of oesophageal intubation results not only from difficulty getting a view but also failure to recognise the structures being visualised and thus persistence in passing the ETT into the oesophagus despite not being able to visualise the larynx. In experienced hands I would expect the rate of oesphageal intubation to be far lower than the rate of difficult intubation, as there would usually be recognition of the fact that the larynx could not be seen and thus no attempt to pass the ETT made. The fact that this kind of training yields clinicians who are unable to make that distinction rings alarm bells regarding the potential for simulation training to replace experience with intubation in real patients. Clearly this was not the intention of this study nor is any suggestion of this made by the authors – but it is an opportunity to recognise the level of competence that this type of training produces. Repeated practice on real patients over a sufficient period of time to allow reflection/improvement is needed to achieve competence at DL.
For inexperienced laryngoscopists, the superiority of video laryngoscopy over direct laryngoscopy cannot be questioned. For experienced operators, it is also clear that VL will often rescue failed DL. However we know that there are some instances when DL will rescue failed VL. This last fact, along with the fact that DL is universally available while VL isn’t and won’t be for some time if ever, and also breaks, means that DL remains an important skill. My feeling is that inexperienced operators should start with VL for most cases, while experienced operators should start with DL on patients not expected to be difficult. For expected easy cases, I will often perform DL myself, visualize the glottis, then step aside and let my trainee intubate with VL.
for these “easy” airway cases, I do either DL with a conventional laryngoscope or use a VL as DL. Once I’m comfortable with the view (usually < 10 seconds) I take the blade out and step aside. The trainee then does the entire procedure, start to finish, with as much or as little help from me as they need. this strategy allows
a. me to keep my laryngoscopy skills sharp
b. me to be confident that the glottis is viewable before the trainee mangles the technique
c. the trainee to see a nice example immediately before they do their own (I will explain what I’m doing as I go depending on the level/needs of the trainee)
d. the trainee to get a complete iteration of the procedure
Yen Chow, PHARM Co-Author
I see that Rueben had replied already but I have chosen not to read his response yet in order to provide my comments NEAT.
Thanks for the opportunity to chat more about airways and for sharing this paper. It is very thought provoking for sure and it adds more fuel to the fire of the great direct laryngoscopy (DL) vs indirect laryngoscopy (IL) debate. I think the findings are consistent with previous studies that the glidescope can work well in the hands of occasional and less experienced intubators.
First of all I would like to say that not all IL (or also known as video laryngoscopy (VL)) are created equal. Glidescope with the hyperangulation and freehand tube delivery is very different from channeled options, from more direct blade approaches like the CMAC direct, and different shaped blades like the Airtraq.
The rates of difficult airway for DL were high at 22% versus 7% for VL. Most airways are not anatomically difficult for DL if one is able to position properly and use first pass success techniques including but not limited to external laryngeal manipulation, head elevation and bougie. Note also that the rescue for VL/IL is usually DL. I love the simplicity and directness of DL and so that is my bias. Fluids are also the enemy of IL with its optics inside the airway. Ideally enhanced DL (which is the CMAC direct) will be the future: use DL or VL when desired, switch between either if needed, always have a bright light and always have DL as backup. Difficult airways are so far and few between that it is hard to be evidence based with the small numbers. In addition, different difficulties require different solutions so a hyperangulated blade might be required for some airway situations.
Although the study sounds like it put together a very extensive airway training program, it does bother me ALOT that they did not use bougie which is a basic standard of care tool these days. It also worries me a bit that there was no cadaver or operating room standardized training either for the fellows: there is a real difference between flesh and mannequins.
The authors do admit the weakness of this being an observational study of 2 cohorts (2008-2010 and 2010-2012) and feel that the fellows had the same skill or training level. One can argue however that as a group teaches something, they become better at it, also techniques improve with better emphasis on things that work better (like airway positioning ear to sternal notch etc etc). Thus the training in 2012 is likely better than the training in 2008. It is unlikely that the airway training made no advances over 4 years.
I do wonder about time to intubation. There was a statistically nonsignificant increase in hypoxemia in the VL group and could this be related to longer attempts at intubation? It has been shown and it is intuitive that glidescope tube delivery takes longer compared to DL.
Bougie first intubation would also have cut down on the esophageal intubation rate. I also wonder if they were using the best techniques at DL: optimal positioning, ensuring methodical progressive landmark exposure, finding the epiglottis, optimized laryngeal exposure with identification of the posterior cartilages, actually visualizing the tube going through the cords as opposed to tube obscuring view of the glottis and then resulting in esophageal intubation, straight to cuff styletting, passing the styletted tube below the line of sight etc etc. The results for DL reflects the occasional less experience intubators: DL is a hard skill to acquire and takes LOTS of practice and has many subtleties.
Note also that ultrasound visualizing of the tube or bougie passing into the trachea at the neck is also a game changer if applied to a study like this.
Finally, much to my chagrin, they did not use RSI which provides optimal intubating conditions in terms of muscle relaxation and open cords at the glottis.
All in all it is an interesting paper but I can’t draw any conclusions related to the experienced intubator or difficult airways in DL vs glidescope. Basically I think it reaffirms that glidescope allows safe intubation for novices after they have received an adequate training program whereas DL still remains a challenging skill for teachers to teach and learners to master.