This case was submitted by my friend Dr. Victoria Stephens. She is a third year Emergency Medicine Registrar from at the University of the Witwatersrand in Johannesburg, South Africa, and a great asset to FOAMed. Follow her on Twitter: @EMcardiac.CaseA 71 year old man was admitted to the ICU with neutropenic sepsis complicated by septic shock. He was intubated and ventilated and was started on an adrenaline infusion to maintain his blood pressure. The admission ECG was normal. Thirty-six hours into his ICU stay he went into a cardiac arrest. The monitor showed a wide complex tachycardia. CPR was commenced while the defibrillator was brought to the bedside. A doctor was called from the ED to assist. The pads were attached to the patient and the defib was placed in AED mode by the nurse. The following rhythm strip was recorded (on a separate monitor from the AED, of course):
|A rhythm strip recorded from lead II. A wide complex tachycardia is present with a rate of approximately 170 BPM. The QRS duration is very wide. It is regular and monomorphic and all but diagnostic of VT.|
The code blue team recognized that the rhythm was ventricular tachycardia and that immediate defibrillation was required. They waited for the AED function on the defib to recommend a shock. Instead, it kept saying “…analysing…analysing”. No shock was advised. The doctor could not remember how to operate the manual mode of the AED defibrillator (this was an defibrillator that has an AED mode; not all defibrillators have that). Since the AED was not advising a shock, he assumed that the defibrillator was faulty, and called for another defibrillator to be fetched from the ED. CPR was continued, adrenaline and amiodarone were given. By the time the second defibrillator arrived, return of spontaneous circulation (ROSC) had occurred. The patient converted into a normal sinus rhythm a short while later.
What happened? Why did the AED not recognize such an obvious case of VT? Was the defibrillator faulty?
No, the defibrillator was not faulty. The technology is, rather, imperfect.
How does an AED work?
The automatic external defibrillator (AED) was initially designed to be used by laypeople or first responders with little or no experience in defibrillation, in order to improve survival from out-of-hospital cardiac arrest (OHCA) (1). The AED uses a Rhythm Analysis Algorithm (RAA) which essentially is software that is programmed to discriminate between shockable and non-shockable rhythms. The RAA then prompts the AED to advise or not advise a shock. The RAA uses up to 18 internal algorithms to determine if a rhythm is shockable or not; the most important of these are: 1) heart rate, 2) QRS width and 3) QRS amplitude.
With regards to VT, the RAA is programmed to recognize VT as shockable only at certain heart rates. For most AEDs, this heart rate is above 150 BPM (2). The rationale for this is twofold; 1) to prevent lay people from potentially defibrillating a perfusing VT in a patient who may still have a pulse, and 2) that patients are more likely to arrest from VT at heart rates greater than 150.
How does the manual defib work?
There is no RAA. The healthcare provider decides if the rhythm is shockable and whether a shock is advised. The number of joules for each shock is also set by the operator.
When can the RAA in the AED fail?
Multiple studies have demonstrated the safety and efficacy of AEDs in Cardiac Arrest.(1) They have demonstrated not only improved survival but also improved neurological outcomes. However, as with any device, error does occur. The RAA in the AED can in certain circumstances fail to recognize a shockable rhythm or incorrectly advise a non-shockable rhythm to be shocked. These circumstances are:
1) Interference from artifact
a. Pacemaker spikes and internal cardioverter defibrillators can cause artifact, interfering with the RAA function.
b. Motion artifacts caused by chest compressions, handling of the patient, movement during ambulance transportation, breathing and seizures may also interfere with the RAA.
Below is an example where external artifacts occurred at the beginning of the AED analysis.(3) The AED incorrectly advised no shock for this case of coarse VF.
|This is an image of coarse VF. The AED incorrectly made a "no shock advised" decision.|
[Image used with permission from: Calle PA et al. Inaccurate treatment decisions of automated external defibrillators used by emergency medical services personnel: Incidence, cause and impact on outcome. Resuscitation 2015;88:68-74]
2) The type of shockable rhythm: they detect VF better than VT
Several studies have examined the accuracy of the RAA by downloading the ECG strips and responses advised from the AED memory module. (3-5) These studies showed that the AED is more accurate in correctly detecting VF than VT; the AED being 100% specific and 95% sensitive for coarse VF. AEDs have also demonstrated similar accuracy in correctly detecting non-shockable rhythms such as PEA, normal sinus rhythm, supraventricular arrhythmias and asystole.
The AED is much less reliable however, in correctly detecting VT: the sensitivity for VT ranged from only 63% to 83% in these studies, indicating that in several instances no shock was advised when VT was actually present.
Inconsistent shock advisories have also been demonstrated for polymorphic VT, including the subtype Torsades de Pointes. One AED advised did not advise shock for any of the Torsades rhythms it was subjected to(!)(2).
Several authors have recommended that manufacturers improve their algorithms and that physicians should be aware of the potential pitfalls in their use.
3) The VT rate
As mentioned above, the RAA of several AEDs is set to recommend a shock if the VT rate is more than 150. Slower forms of VT will not get a shock advisory, even if the patient is in arrest and requires it. Though higher VT heart rates are more likely to cause cardiac arrest, patients with poor systolic function may not tolerate a sustained VT of 130-150 BPM and may indeed be in arrest or near-arrest. (2). Several AEDS were shown in a recent study to only advise a shock if the VT rate exceeded 180, and others only if higher than 250 (!).(6)
Multiple studies have shown that the AED is very accurate at detecting VF and tends to advise a shock nearly 100% of the time. AEDs have been shown to reduce mortality in cardiac arrest, especially in OHCA.
The AED similarly recognizes sinus rhythms, supraventricular rhythms and asystole reliably.
The AED is far less accurate at determining VT; with regards to both the monomorphic and polymorphic forms. If the AED fails to recognize the VT, the VT will eventually degrade to VF which subsequently the AED is more likely to recognize. However, this delay may result in significant harm to the patient’s outcome as time to defibrillation is crucial to survival.
With VT arrests, the trained healthcare provider is superior to the AED. For those staff who have a defibrillator with both manual and AED modes, they should know how to recognize VT, or probable VT, use the AED, and, if using the defibrillator in AED mode, know how to switch to manual mode.
1. Part 6: Electrical therapies: automated external defibrillators, defibrillation, cardioversion and pacing. 2010 American Heart Association Guidelines for Cardiopulmonary Resuscitation and Emergency Cardiovascular Care. Link MS, Atkins DL, Passman RS, Halperin HR, Samson RA, White RD, Cudnik MT et al. Circulation 2010;122(suppl 3):s706-s719
2. Inconsistent shock advisories for monomorphic VT and Torsade de Pointes – a prospective experimental study on AEDs and defibrillators. Fitzgerald A, Johnson M, Hirsh J, Rich M-A, Fidler R. Resuscitation 2015 (Article in press)
3. Inaccurate treatment decisions of automated external defibrillators used by emergency medical services personnel: Incidence, cause and impact on outcome. Calle, PA, Mpotos N, Calle SP, Monsieurs KG. Resuscitation 2015;88:68-74
4. Performance and error analysis of automated external defibrillator use in the out-of-hospital setting. JL, Weinstein C. Ann Emerg Med 2001;38:262-267.
5. Machine and operator performance analysis of automated external defibrillator utilization. Ko PC1, Lin CH, Lu TC, Ma MH, Chen WJ, Lin FY J Formos Med Assoc. 2005 Jul;104(7):476-81
6. What is ventricular tachycardia for automated external defibrillators? Kette F, Bozzola M, Locatelli A, Zoli A. J Clin Exp Cardiolog 2014;5:285 doi:10,4172/2155-9880.1000285