In patients undergoing emergent tracheal intubation, there is currently no universally accepted gold-standard test to confirm the location of the endotracheal tube (ETT).1 End-tidal carbon dioxide (CO2) detection is the best of the tests that are routinely utilized to confirm ETT placement, however, it has been shown to have an error rate as high as 1/10 for proper determination of ETT location in emergency intubations.2 As a result, multiple modalities are necessary to confirm ETT location, which can delay mechanical ventilation and other treatments. The lack of a single, reliable test to confirm ETT placement can potentially lead to confusion regarding the location of the tube. This confusion can result in both unrecognized esophageal intubations (“false positive”), as well as successful tracheal intubations that are subsequently removed (“false negative”), subjecting the patient to further unnecessary attempts at airway management. Both scenarios can lead to disastrous consequences.
The advent of point-of-care ultrasound (POCUS) has led to a potential solution to this problem. Over the past decade, emergency physicians, anesthesiologists, and others have studied the ability of POCUS to determine ETT location either during or immediately following laryngoscopy. Multiple approaches have been utilized including transtracheal at the suprasternal notch (Figure 1), transcricoid, and also assessments of lung sliding at the visceral-parietal pleural interface and the diaphragms.
- Studies evaluating the diagnostic accuracy of transtracheal ultrasound in confirming ETT location AND
- Results of the index test (POCUS) verified with that of a gold standard (capnography) AND
- Studies that enrolled living adult humans
Eleven studies encompassing 969 intubations were analyzed. The pooled sensitivity and specificity for the detection of proper ETT placement with US were:
- Sensitivity: 98% (95% C.I. 97-99%)
- Specificity: 98% (95% C.I. 95-99%)
- Positive Predictive Value: 99.5%
- Negative Predictive Value: 93.8%
- + Likelihood Ratio: 46
- – Likelihood Ratio: 0.0157
Of these 11 studies, 3 examined elective intubations and a sensitivity analysis was performed to exclude these trials. This resulted in an aggregate sensitivity and specificity of POCUS in emergency intubations of:
- Sensitivity: 98% (95% C.I. 97-99%)
- Specificity: 94% (95% C.I. 86-98%)
As with all meta-analyses, the robustness and applicability of the results are highly dependent on the quality of the included trials. Eight of the included trials were judged to have a low risk of bias (using the QUADAS-2 tool5) and the aggregate sensitivity and specificity was:
- Sensitivity: 98% (95% C.I. 97-99%)
- Specificity: 98% (95% C.I. 95-100%)
The heterogeneity between the included studies was determined using the inconsistency index (I2) and found to be in the mild-moderate range. However, further inspection into the variability between studies showed 9 different confirmatory findings on POCUS that were used to determine ETT location, including:
- One air-mucosal interface = tracheal intubation.
- Two air-mucosal interfaces = esophageal intubation.
- Two hyperechoic reverberation artifacts inside trachea = tracheal intubation.
- Two hyperechoic reverberation artifacts inside esophagus = esophageal intubation.
- “Bullet sign”= tracheal intubation (Figure 2)
- “Double track sign” = esophageal intubation (Figure 3).
- “Snowstorm sign” = tracheal intubation6
- Dynamic opening of the esophagus by the ETT seen on US performed during laryngoscopy = esophageal intubation. (Video 2)
- Brief flutter deep to the thyroid cartilage = tracheal intubation
Proof is in the pudding
Transducer Placement Location
While many of these sonographic findings are similar (and some are exactly the same), there is no consistent sonographic finding that is widely accepted amongst investigators. Another source of heterogeneity is in the transducer type and placement. The following locations were used in the included studies:
- Immediately cephalad to suprasternal notch: 8
- Cricothyroid membrane: 1
- Both suprasternal notch and cricothyroid membrane: 2
Six studies used high frequency linear probes, while 5 studies used low frequency curvilinear probes. The precise frequencies used in the various studies were:
- 3.75 MHz: 3
- 3-5 MHz: 1
- 5-10 MHz: 1
- 7-10 MHz: 2
- 9-12 MHz: 1
- 10 MHz: 2
- No frequency noted: 1
At this time, POCUS can (and should) be utilized as an adjunctive method as part of a multimodal approach to verify ETT location. It has an added benefit in cases where end-tidal CO2 detection is less reliable, such as cardiac arrest. Future studies of POCUS verification of ETT location should attempt to standardize the approach and type of equipment (transducer, settings, frequency, etc), and should utilize a technique that will not interfere with attempts at direct laryngoscopy.
Nice overview of the statistics regarding this interesting and promising technique.
The article as it reads currently presumes some knowledge regarding use of US for verifying ETT, but many readers may be unfamiliar. I think an overview of the technique and the findings would be helpful before describing the test characteristics. Perhaps the \'signs\' could be clarified in that section.
At this time we have both still image and video of the ultrasound. I think the video provides more utility, but perhaps we can work on labeling the still image for reference.
Overall well put together, and potentially very helpful.
I like the idea of going over the various signs prior to delving into the research portion. My primary intent with this post was to put a scientific spin on what appears to be a promising US modality. Too often in the field of emergency US we are inundated with “look what I did this one time” anecdotes that then get promoted as “evidence”. I think it is very important that BEFORE we even begin to consider making an emergency US application a part of our practice of medicine that we familiarize ourselves with not only the technique, but also the performance characteristics of the test. My intent with this post was to primarily address the latter, as many folks before me have addressed the technique end. Also, by reviewing the primary articles included in the meta-analysis you will learn much about the technique of image acquisition.
I’ll update the still images with some labels and look to include a photo demonstrating probe placement.
Thanks for the feedback!
Overall, I think you do a good job presenting this POCUS technique as an adjunct rather than the definitive new way to confirm correct ETT placement. I agree with Dr. Shaikh that a discussion of actual technique (perhaps with video of you doing it on a patient?) would be beneficial to those who want to try it for themselves after reading your article. In addition, the videos are a necessary supplement to your text, but I honestly can\'t tell the difference between them. If there is a way to point out the actual \"sign\" you\'re referencing it would be helpful.
I think the post would benefit from clarification that your miss rate stats are based on traditional direct laryngoscopy (at least that\'s what I am assuming). I think direct visualization through cords with video laryngoscopy (ie glidescope, etc) would be sufficient to confirm correct tube placement without the need of POCUS. In the interest of bias, it must be said I\'m obsessed with the glidescope about as much as you\'re obsessed with ultrasound.
Your discussion of statistics from the prior meta-analysis by Das is also a nice touch and good way to back up your claims. But I think it gets wordy and confusing, and I found I had to reread the section a few times before I could move on.
Where I see a potential benefit in POCUS for ETT placement is in patients who have been transferred to your resus room already intubated -- whether the intubation occurred in the field or outlying facility. We know the ETT can be incorrectly placed by less experienced providers, or dislodged during transfer, and POCUS would offer a way to determine if the tube is correctly placed without having to remove the apparatus securing the tube and without having to again perform laryngoscopy.
Overall, a good job, and I think you will get other providers interested in trying the technique themselves.
Justin Bright, MD
Henry Ford Hospital
Insightful commentary. We are going to add some labels to the still images that should assist in your interpretation of the video clips. Basically, what you are looking for in an esophageal intubation video clip that is recorded during the process of tube placement, is the creation of a second lumen (the esophagus, which is a potential space, not seen on ultrasound unless there is something inside of it, or there is significant esophageal pathology…mass, etc…at the location you are imaging). This “second lumen” is located to the left of the trachea in approximately 70% of patients and to the right of the trachea in approximately 30%. Very rarely is the esophagus located immediately posterior to the trachea at the level of the sternal notch. In this video clip, the esophagus is located to the left of the subjects trachea.
The miss rate statistics (all of the statistics actually) are based on a reference standard of end-tidal CO2 either through continuous waveform capnography, or colorimetric end-tidal CO2 detection. The utilization of end-tidal CO2 as a reference standard was part of the inclusion criteria for studies in the meta-analysis by Das. Although direct visualization of the ETT passing through the cords (when it is actually visualized going through the cords and not into the esophagus) is the textbook answer for tube confirmation, more novice intubators are prone to mis-identifying the esophageal lumen as the glottis (see the 9th video titled “Laryngoscopic Trachea versus Esophagus\" on this wonderful page from Life in the Fast Lane to illustrate the point. In academic institutions when direct laryngoscopy is utilized by a resident, the attending physician is unable to see the tube passing through the cords him or herself, and must rely on second-hand confirmation. Unsettling for some… and thus, many of us understand your glidescope obsession.
Your point regarding the already intubated patient hitting your door and needing to have ETT location confirmed is a great one. And although some work has already been done, this application would be well served by a randomized trial. Group 1: US confirmation, Group 2: “wild type” or “usual methods”… which at various institutions may include laryngoscopy, end-tidal CO2, chest radiography, pulse oximetry, etc. I’m game, who else is in?!?
Expert Peer Review
Great discussion on this topic. I am personally quite familiar with POCUS for verification of ETT placement and your post was easy to follow. I especially like the dynamic video of the esophageal intubation. Also for those of us not research inclined, the bullet points of the meta-analysis were very helpful.
For those not familiar with neck anatomy as visualized on US, I agree that a still image with labels would be quite helpful.
Although there is no universal approach to performing this US, perhaps you could briefly describe which method you have found best through your own experience.
Not sure if you ran into any studies looking at US identifying main stem intubations. Many times clinicians can tell an esophageal intubation has occurred just by looking at the clinical picture, such as noticeable hypoxia, no breath sounds on auscultation of the lungs, poor ETCO2, or breath sounds in the epigastrium. But mainstem intubations are sometimes more difficult to identify without imaging. Blaivas and Tsung looked at a case series on this. Kerrey et al. did a pediatric study on this as well. Might be nice to touch on this topic.
Overall great overview on this topic.
I tried to avoid flavoring the post with too much of “my own opinion” and wanted to let the evidence support my position that tracheal US offers a great adjunctive method to confirm ETT location…but since you asked…
I find the technique most useful in cardiac arrest patients (both those intubated in the field… a rare thing with our decimated EMS system here in Detroit… and those intubated in the ED) and also in patients with very difficult laryngoscopy and limited visualization of the glottic aperture. We had a case like this recently in someone with an unpredicted difficult airway (due to limited mouth opening… VERY limited) who was being intubated b/c of hypercarbic respiratory failure. A bougie was required b/c the glottis was unable to be visualized whatsoever… thus the ETT was never seen to pass through the vocal cords. The patient\'s pulse ox had begun to drop during the passage of the tube over the bougie, and (likely due to pulse ox lag b/c the patient was obese and in shock) did not recover after she began being bagged through the tube. The vent showed that the patient was not getting good tidal volumes (50-60 ml’s), the RT was convinced the tube was in the goose, however a tracheal US with the probe in a transverse orientation just cephalad to the suprasternal notch showed a single reverberation artifact in the trachea, with no evidence of a second lumen, rotation of the probe to a sagittal plane actually demonstrated the radiopaque stripe of the ETT clearly in the trachea. Eventually the pulse ox returned to 100% and the pt began getting good volumes. Premature removal of that ETT in a patient with such difficult laryngoscopy (no glidescope would ever fit in her mouth) would have been disasterous.
Regarding main stem intubations, the addition of lung sliding after tracheal US can assist in determining a mainstem intubation. Sim et al studied this and published their findings in Resus in 2012. They placed a low-frequency curvilinear transducer in the mid-axillary line of patients who had been intubated and compared this to chest radiography. 115 patients were enrolled, 31 of whom were intubated because of cardiac arrest. There were 9 one-lung intubations, and US was 91.5% sensitive (95% C.I. 84.5-96%) with a PPV of 96% (95% C.I. 90.2-98.9%). US had poor specificity and a poor NPV in this study. So although mechanistically promising, it appears based on this, the largest study of the topic to date, that lung sliding is not yet ready for prime-time to be a stand alone test for single lung or main stem intubation.