Expired carbon dioxide during newborn resuscitation as predictor of outcome

K Holte, H Ersdal, C Klingenberg, J Eilevstjønn

Aim

To compare end-tidal carbon dioxide(ETCO2) and heart rate (HR) as predictors of survival in neonatal resuscitation.

Rationale

HR is an indirect measurement of ventilation and a lower HR is associated with poor neonatal outcomes. If ETCO2 is measured it could potentially more accurately be associated with mortality than simply the HR alone.  

Design

Observational study of electronic data of the resuscitation which was collected automatically.

Midwives/resuscitators unable to see end tidal CO2 readings. 

Resuscitation undertaken as per “helping babies breath” training. 

Setting

Rural Tanzanian referral hospital (Approximately 4000 deliveries per year)

Population

434 live born babies with HR <120

Excluded those with incomplete data, APGAR score of 0, <28 weeks and those ventilated with PEEP. 

Analysis

HR, End tidal C02 and expired tidal volume associated with death were compared using logistic regression analysis. 

Primary outcome

Mortality at 24 hours.

Results

Increase in HR and CO2 both independent predictors of survival

The time to reach the AOC (area under the curve) determined sensitivity limit for predictor survival was sooner with CO2 than with HR .

Author's Discussion

This study confirms previous smaller studies in resource rich settings (and on preterm infants) that end tidal CO2 correlated with mortality and provided this indication earlier. 

Strengths

Simple patient orientated outcome, appropriately short follow-up in this age group (if survival to 24 hours significantly more like to survive longer).

Interesting and effective use of remote data collection

Pragmatic look at realistic low-resources resuscitation efforts using widely available equipment (other than end tidal CO2)

No oximetry available, realistic in a low-resource setting. 

Limited number of variables compared (limits data-dredging)

Weakness

No description of patients with incomplete data, potentially causing selection.

Retrospective descriptive study. 

Some funding from makers of resuscitation equipment 

Unable to distinguish between patient factors and clinician factors for cause of low ETC02; e.g. if improved ventilation could improve survival or if the ETCO2 is just predictive or an inevitable process. 

 My humble opinion

It was interesting that CO2 output was associated with mortality even when the midwife could not see this reading. If indeed colorimetric CO2 monitors are easily available compared to HR monitoring this would be an interesting cost-effective intervention. 

If CO2 monitoring is available I would of course use it to guide ventilation efforts. I wonder if a study comparing blinded midwives/resuscitators to those who could see the CO2 reading would demonstrate the effectiveness of this device in improving ventilation techniques in this setting.