BET 1: Can the value of end tidal CO₂ prognosticate ROSC in patients coming into emergency department with an out-of-hospital cardiac arrest (OOHCA)?

Three-part question

Patient group—(In adults admitted to the ED with an out-of-hospital cardiac arrest)

Intervention—(does end tidal CO₂ measurement)

Outcome—(predict/prognosticate return of spontaneous circulation)?

Clinical scenario

A 60-year-old male is brought into the ED with an

out-of-hospital cardiac arrest (OOHCA). All monitoring is attached while ALS protocol is ongoing, including CO₂ monitoring. You want to assess whether the patient is going to survive and thereby achieve a return of spontaneous circulation (ROSC) and you wonder whether the patient’s end tidal CO₂ (ETCO₂) level can prognosticate this.

Search strategy

Medline, Cochrane and EMBASE databases (2006 to present).

[exp HEART ARREST/ or "Heart arrest".ti,ab or "cardiac arrest".ti,ab] AND [ETCO2.ti,ab or "end tidal co2".ti,ab or exp CAPNOGRAPHY/ or capnometry.ti,ab or {exp CARBON DIOXIDE/ AND exp TIDAL VOLUME/ AND 19 AND 20}] AND [{exp SURVIVAL/ OR exp SURVIVAL ANALYSIS/} or exp TREATMENT OUTCOME/ or {(rosc OR "return of spontaneous circulation").ti,ab} or exp "OUTCOME ASSESSMENT (HEALTH CARE)"/]

Conference abstracts removed. In-patient studies removed. Case reports, editorials and notes removed. Paediatric studies removed. Duplicate results removed. Limited to Human studies, English papers and solely papers between 2006 and 2016. Studies that did not answer research question removed manually.

Search outcome

232 articles obtained of which 4 were of sufficient quality (table 1). Results already reviewed in these meta-analyses were not presented below.

Table 1 Relevant papers

Author, year, country of publication	Patient group	Study type (level of evidence)	Outcomes	Key results	Study weaknesses
Hartmann et al, 2015,1 USA	7276 subjects from 27 studies used for qualitative analysis. 6565 subjects from 20 studies used for average EtCO2. 6550 subjects from 19 studies for meta-analysis	Systematic review and meta-analysis	Participants with ROSC after CPR have statistically higher levels of ETCO2	The overall mean ETCO2 value was significantly higher among participants with ROSC than those without ROSC (25.8/−9.8 mm Hg vs 13.1/−8.2 mm Hg, p=0.001)	(1) The overall level of evidence was characterised as very low by the GRADE criteria. (2) Mostly only cohort studies analysed (26/27 studies). (3) Big variance on time taken to initiate resuscitation, quality of compressions and use of different methods to deliver compressions between studies. (4) Presence of serious inconsistency, as measured by the degree of heterogeneity (p<0.001 and I2 value of 98.5%)
				The average ETCO2 level was 25 mm Hg in participants with ROSC
				The mean difference in ETCO2 was 12.7 mm Hg (95% CI 10.3 to 15.1) between participants with and without ROSC (p<0.001)
				The mean difference in ETCO2 was not modified by the receipt of sodium bicarbonate, uncontrolled minute ventilation or era of resuscitation guidelines
				The overall quality of data by Grades of Recommendations, Assessment, Development and Evaluation criteria is very low, but there are currently no prospective data
Poon et al,  2016,2 Hong Kong	319 patients	Prospective cohort study	A 3 min ETCO2 ≤10 mm Hg was associated with poor prognosis and low chance of ROSC	A 3 min ETCO2 >10 mm Hg was a predictor of ROSC with OR 18.16 (95% CI 4.79 to 51.32, p<0.001). In other words, when cardiac arrested, for a patient with a 3 min ETCO2 > 10 mm Hg the odds of ROSC was 18 times higher than those with ETCO2  ≤10 mm Hg	Large number of patients excluded due to improper documentation of the use of ETCO2 (approximately one-third). (2) Quality of chest compressions was not controlled or measured. (3) The decision to stop resuscitation may have been influenced by the ETCO2 value at the time, which could have potential bias on ROSC rate
Akinci et al,  2014,3 Turkey	80 patients	Prospective cohort study	PetCO2 values are higher in the ROSC group	ETCO2 levels of the ROSC group in the 5th, 10th, 15th and 20th min were significantly higher compared with the Exitus group (p<0.001)	(1) ETCO2 levels not measured on transport to hospital. ETCO2 value differences, which might be resulting from different arrest aetiologies (asphyxia and cardiac) could not be determined as a result of this. (2) Small sample size. (3) No clear indication or suggestion of what ETCO2 level can be used to prognosticate ROSC—however, does give an indication of when best to assess this. (4) Published in a low impact medical journal
			During the CPR, the most reliable time for ROSC estimation according to PetCO2 values is 20th min	In distinguishing ROSC and Exitus, ETCO2 measurements within 5–20 min intervals showed highest performance on the 20th (area below the ROC curve was determined to be 0.850 (95% CI 0.721 to 0.980)) and lowest on the 5th minute (area below the ROC curve was determined to be 0.730 (95% CI 0.610 to 0.849))
				None of the patients who had ETCO2 levels less than 14 mm Hg survived
Pantazopoulos et al,  2015,4 Greece	42 studies included in qualitative synthesis	Narrative review		Although changes and trends in ETCO2 values during CPR are more important than absolute ETCO2 levels, current data suggest that certain cut-off values may be targeted; an ETCO2 >10 mm Hg is correlated with increased possibility for ROSC	No systematic review or meta-analysis done
				Rescuers should target a 20 min ETCO2 of at least 20 mm Hg
				The value of a trend more than absolute ETCO2 values may be most important in the presence of a treatable cause
				An abrupt increase in ETCO2, under constant ventilation and CO2 production, provides the fastest indication of ROSC

CPR, cardiopulmonary resuscitation; EtCO₂, end tidal CO₂; PetCO₂, end tidal CO₂ tension; ROSC, return of spontaneous circulation.

Comments

Overall the recent paper by Hartmann et al is well-written and the most up to date and most pertinent study on this research topic. While the quality of the study may have been deemed low by the GRADE criteria, it is important to remember that ETCO2 in participants (with and without ROSC) can be compiled, despite the variety of interventions in the included studies, as ETCO₂ is a proxy measurement for cardiac output—a physiological outcome that can be achieved under many differing circumstances. In a meta-analysis such as this one where homogenous study design is not necessary to evaluate a physiological state, the level of evidence appears falsely poor. This has been clearly addressed by the authors. This paper gives a clear idea of what ETCO₂ level should be aimed for when resuscitating patients and thereby can prognosticate between a positive and a negative outcome. It also emphasises that current guidelines may need to be updated to acknowledge that an aim of 10–20 mm Hg may be too low. The paper by Poon et al is important in deciding whether a 3 min ETCO₂ level of ≤10 mm Hg can help clinicians decide whether to discontinue resuscitation on the basis that there is a much greater chance of morbidity and mortality. The paper by Akinci et al highlights that a 20 min ETCO₂ check has a greater performance in predicting ROSC than earlier times, although the data itself may not be robust enough to go by from a resuscitation guideline perspective. Having said this, the data are important and as such more studies in this research topic would definitely help.

Clinical bottom line.

Current literature suggests that: (1) Our current ETCO₂ aim of 10–20 mm Hg may be inadequate and should be modified to 25 mm Hg. (2) A 3–5 min ETCO₂ level of ≤10 mm Hg is associated with bad prognosis and as such, it may be beneficial to consider stopping patient resuscitation should this be the clinical case. (3) It is important to see the trend of ETCO₂ rather than making a decision solely on one specific value, as sometimes an abrupt increase in ETCO₂ could be a sign of impending ROSC. (4) More robust prospective data on the optimal ETCO₂ value that is associated with ROSC would be helpful in defining a more accurate future target for intervention.