Friday, July 11, 2014

End-Tidal CO2 Monitoring Unhelpful in Sedation

Capnography during procedural sedation has rapidly become standard practice in the interests of “safety” – despite decades of Emergency Medicine experience safely performing sedation without.  The theory: earlier detection of inadequate ventilation allows for intervention and prevention of hypoxemic episodes.  In the most prominent randomized trial, there was a 17% absolute reduction in transient hypoxemic events.  Critics appropriately point out, however, that transient events are hardly a meaningful patient-oriented outcome.

This trial, in which sedation was performed by nurses in an outpatient gynecology setting in the Netherlands, describes 427 patients randomized either to capnography or “routine monitoring”.  Supplemental oxygen was not used in this setting, as these practioners considered it to obscure the additive value of pulse oximetry.  Essentially by definition, all patients were female, and the median age was 24 with few co-morbidities.

Overall, 25.7% of patients in the capnography group developed hypoxemic episodes versus 24.9% of patients in the control group.  There were also no differences in numbers of patients with profound hypoxemia (<81%) or prolonged hypoxemia (>60 seconds), although patients with hypoxemia in the capnography group had more frequent prolonged episodes than the control group (14.0% vs. 3.7%).  Likely as a result, patients in the capnography group underwent airway positioning maneuvers more frequently (49.5% vs. 32.1%).

It’s a stretch to say this is information is generalizable to Emergency Department sedation.  It is, at least, a useful window into what many skeptics have been saying all along – the additive value of capnography in sedation is low, and, rather, the extra information leads to additional interventions and procedural interruptions without measurable benefit.  Procedural success in this setting was not adversely impacted by the frequent interruptions, however.

“Capnography During Deep Sedation with Propofol by Nonanesthesiologists: A Randomized Controlled Trial”


  1. thankyou for highlighting this article

    I dont understand the practice of not providing supplemental oxygen when giving IV sedation. I dont think we should be emulating this in our ED practice!

    ETCO2 is useful to detect airway obstruction and apnoea during sedation. Laryngospasm does occur ,especially with ketamine sedation for kids in ED. I would rather detect this ASAP rather than wait for desaturation.

    ED sedation should be to the highest standard, as our patients often have full stomachs and multiple other acute problems

  2. Also unnerved by a 25% rate of hypoxaemia in both and groups


  3. If you want to monitor for apnea, you have to watch something - either the waveform, or the chest. Not sure why a waveform would be "better" than chest rise.

    As has been mentioned, it seems like a good idea to give supplemental oxygen!

  4. chest can be moving but no air moving= laryngospasm
    ETCO2 good

  5. Thank you Ryan for a thought provoking commentary on an important topic. I view this discussion not as a criticism of ETCO2 monitoring as a whole, but rather what occurs when it is employed in such a low risk population. The benefits of ETCO2 monitoring during procedural sedation are based on a number of assumptions. First using ETCO2 monitoring will detect apnea earlier than O2 sat and clinical judgment. Secondly, that the earlier detection of these apneic events is clinically relevant. And finally, if we intervene earlier in the event process, the patient outcomes will be markedly enhanced than if we wait to intervene when the patient begins to desaturate or the apneic episode becomes clinically obvious.

    The first of these three assumptions has proven to be true. ETCO2 detects apneic events far earlier and more frequently than either O2 saturation or clinical judgment. Unfortunately the remaining two assumptions do not seem to be supported by the literature. Deitch et al randomized patients undergoing procedural sedation in the Emergency Department to either standard monitoring or the addition of ETCO2 (1). Patients in the ETCO2 arm had 17% fewer hypoxic events (defineD as a O2 sat less than 93%). None of these events proved to be clinically relevant and 41% of them resolved without any intervention. In fact the number of hypoxic events severe enough to require any intervention were equal in both groups (13%). The use of ETCO2 monitoring did not prevent any intubations, hypoxic injuries or even use of BVM. More importantly, the ETCO2 monitoring detected a significant number of events that failed to result in any hypoxia (36%) and caused a significant increase in airway interventions and procedural interruptions. These results are consistent with the few other studies examining ETCO2 monitoring in procedural sedation as well as the results of this recent trial reviewed by Ryan (2,3).

    ETCO2 like any test is only as good as the population in which it is utilized. It is a fantastic tool that when used in the right circumstances is clinically valuable. For example in post-intubation tube confirmation, transport of intubated patients, and the management of cardiac arrest we have identified high acuity cohorts. In these settings with a high pre-test probability for clinically relevant disease, the interpretation of ETCO2 becomes relevant. Conversely, like with any highly sensitive test when applied to an extremely low-risk population, a large proportion of the results will inevitably be falsely positive and an even greater number can be categorized as overdiagnosis. In this case the data clearly shows that most of the apneic events ETCO2 detects are clinically irrelevant and do not result in hypoxic episodes. Even the ones that do go on to hypoxic episodes are generally self-limited, resolving without any intervention. The few that do require intervention are detected by standard monitoring in a timely fashion without clinically relevant consequences. The only clinically relevant effect ETCO2 monitoring demonstrated was a greater amount of interventions and procedural interruptions.

    I am not saying that ETCO2 should never be used in procedural sedation. There are certainly cases when the added monitoring would prove invaluable. Likewise it is equally unjust and somewhat dogmatic to say ETCO2 is required during every procedural sedation when the evidence does not support this. I think as physicians we should be capable of understanding the strengths and weaknesses of such a test, balance this with the patient before us, the procedure required, and use our own clinical judgment to decide when its use is required.

    Sometimes more monitoring provides us with clinically relevant information, sometimes it simply provides us with more monitoring…


  6. Yeah, I never need ETCO2 when I perform pelvic exams on healthy patients in the ED either. However, when I relocate hips on 80 y/o's with tons of co-morbidities, I find it helpful in spotting the apneas a little earlier.

  7. Lots of great comments all around! As usual, Rory puts my brief post to shame with multiple citations.

    ETCO2 probably has value – but so far, its value has only been measurable in terms of surrogates for patient-oriented outcomes. As I mention, we've been performing procedural sedation for decades with substantially less predictable agents than present day (methohexital, etomidate), and all we've needed to date is appropriate vigilance. That said, there are almost certainly high-risk patients for whom it may have greater than average utility.

    The oxygen-during-sedation debate is another interesting topic. It seems foolish not to use oxygen – but, some folks make a reasonable point that its use obscures hypoventilation. However, frankly, I make the same argument for oxygen as against ETCO2 – I don't really care about transient episodes of hypoventilation/hypoxia, and I don't want them interrupting my procedure.


Comments on posts > 10 days old will be moderated; blame spammers.