Expirations: Does End-Tidal CO2 Monitoring Predict Adverse Respiratory Events In Sedation?
In the growing backlog of articles and journal club presentations I’d like to write up, I came across this publication which was presented a few months back — Does End-tidal Carbon Dioxide Monitoring Detect Respiratory Events Prior to Current Sedation Monitoring Practices? from Burton, Harrah, Germann and Dillon in Academic Emergency Medicine 2006; 13:500-504. My notes on the presentation have long since disappeared (I believe it was given by… Tim?) but my interest in the topic was rekindled after a recent M+M.
The authors note that, in the OR, end-tidal monitoring is standard practice, but no EM organizations have advocated its routine use in ED procedural sedation. Thus, the goal of this study was to see if ETCO2 monitoring could detect acute respiratory events before current methods. They used a prospective observational case series with a convenience sample of ED patients undergoing PSA at a 50,000-pt/year ED among consenting patients when investigators were present. Controls got SpO2, RR, HR and BP monitoring, and an investigator not involved with pt care “assured” the quality of waveform data for SpO2 and ETCO2, as well as proper positioning of monitoring equipment, and was the only person present who saw the ETCO2 waveform and numerical CO2 display. Measurements were recorded every 30 seconds, as well as during med administration, observed acute respiratory events, interventions, and return to baseline.
What was a respiratory event? The authors define it as a desat to 92% or less, an increase above 2LNC in administered O2, hypoventilation, or use of bag-valve mask or ventilatory assistance, use of physical or verbal stimulation, or administering reversal meds. Changes in ETCO2 of more than 10 mmHg from baseline, or an intrasedation ETCO2 of less than 30 mmHg or more than 50 mmHg, were considered investigational acute respiratory events.
They planned to enroll 250 patients but, in dramatic fashion, had to stop the study early, after just 59 patients (60 sedations). Abnormal ETCO2 findings (as described above) were found in 36 of the 60 encounters — 32 low and five high (yeah, one guy had both low and high).
Of course, 44% of these “abnormal” ETCO2 readings were not associated with any adverse respiratory event or intervention. Some were just single, abnormal blips in recordings. But the investigators did log 20 acute respiratory events in their 60 encounters (33%) and 17 of them (85%) were associated with abnormal ETCO2 recordings. Stay with me here — of these 17 encounters, 14 (70%) had the abnormal ETCO2 recording before changes in SpO2 or RR.
So, what to make of this? Well, just browse through the tables and see what kind of events we’re talking about. First, 18 of the 20 acute respiratory events occured on propofol, which we can’t use in the ED. Ketamine and Etomidate account for the others, while midazolam was event-free (but they only used it three times). Only four of their twenty acute resp events required bag-valve mask ventilation, and two of those had no warning (ie, the ETCO2 didn’t help predict the event). Most of these abnormalities were low ETCO2, it only went high once (and you’d think low RR and CO2 retention would be a more common finding — as seen in previous studies). No one was hyperventilating, either.
In addition to not providing adequate warning in some of their worst acute respiratory events, ETCO2 provided warning 3-4 minutes ahead of some relatively minor events (requiring just some verbal stimulation). Is that really valuable? And don’t forget the fact that 44% of ETCO2 abnormalities were not accompanied by clinically observed respiratory events.
So, while ETCO2 monitoring is certainly becoming more widespread during sedation, I don’t think its advocates can really point to this study as demonstranting its relevance. I think you could argue that this study by Miner et al from Academic Emergency Medicine 2002 Apr;9(4):275-80 made the case for ETCO2 better. But in discussing that Miner study, and others, the 2004 ACEP policy guidelines for procedural sedation have this to say about capnometry:
Theoretically, early detection of hypoventilation with capnometry may be beneficial. However, there is no evidence that this benefit has an impact on patient outcome when used in procedural sedation and analgesia.
If you’re curious about monitoring sedated patients who aren’t undergoing procedural sedation (beause, say, they were agitated) I’ve only found one study that addressed this — it’s from Martel, Miner and Biros and was also printed in Academic Emergency Medicine 2005 Dec;12(12):1167-72. Unfortunately it wasn’t included in the recent ACEP agitation guidelines.
Emcrit.org is uncharacteristicly quiet on the issue of ETCO2 in procedural sedation, concluding it’s safer and summarizing the above paper in one line. But if you’re looking for a hands on guide to the different sedation agents we can give, their dosing and drawbacks, it’s a great source.
