Aim: To determine the reliability of pulse palpation to diagnose paediatric cardiac arrest.
Materials and methods: With all cardiovascular information obscured, 209 doctors and nurses (rescuers) were requested once each to determine if a pulse was present in 1 of 16 infants and children (average age 1.8 years, range 1 week-13 years) provided with non-pulsatile circulation with veno-arterial extracorporeal membrane oxygenation or left ventricular assistance for cardiac arrest or failure. Rescuers did not know the stage of recovery of the heart and did not if a true pulse was present or absent. Rescuer decisions "pulse absent" or "pulse present" were compared with concurred decisions of investigators and bedside nurse who knew cardiovascular data and had unlimited time to palpate pulses.
Results: Rescuer pulse palpation accuracy was 78% (95% CI 70-82), sensitivity 0.86 (95% CI 0.77-0.90) and specificity 0.64 (95% CI 0.53-0.74). When investigators diagnosed cardiac arrest pulse pressure was 6+/-5mmHg (range 0-20) compared with 9+/-8mmHg (range 0-29) with rescuers (p=0.0004). With pulse pressure zero, rescuer accuracy was 89% and sensitivity 0.89. Sixty per cent of rescuers chose a brachial pulse, 33% a femoral pulse with respective accuracies of 78% and 77%, sensitivities 0.86 and 0.85 and specificities 0.67 and 0.56.
Conclusions: Pulse palpation is unreliable to diagnose paediatric cardiac arrest. Rescuers misdiagnose on 22% of occasions and which may lead them to withhold external cardiac compression on 14% of occasions when needed and on 36% to give it when not needed. Brachial palpation is slightly more reliable than femoral palpation.