In-Hospital Defibrillation

This issue has important implications for defibrillation protocols. For example, it makes sense to me (and I think is consistent with the new Guidelines) to recommend an initial empiric shock for apparent asystole before taking the time to rule out occult VF--particularly with AEDs, with which there still is not a practical way to check in another lead.

The complete worksheet may be viewed or downloaded here.

Recommendation to change ACLS Guideline "strongly discourag[ing]" shocking apparent asystole

Our recommendation for the 2000 Guidelines is to delete the current guideline "strongly discourag[ing]" shocking asystole; and replace it with an explicit statement that there is no evidence of harm from the practice.

We also recommend a change to call for at least one empiric shock for apparent asystole when using adhesive dual-function electrodes.

Since the proposal is for deletion of a negative recommendation which first appeared in the 1992 ACLS Guidelines, this discussion will consider the rationale and evidence supporting the alleged danger of shocking apparent asystole as well as at the strength of evidence supporting the intervention. Our working premise is that no intervention should be explicitly prohibited or "strongly discourage[d]" without evidence of harm or at least a reasonable rationale for possible harm, regardless of whether strong evidence exists favoring the intervention. If a negative recommendation is unsupportable, that alone should be sufficient reason to drop it.

The argument for the negative recommendation is quoted below in its entirety from the 1992 ACLS Guidelines (reference numbers in original text replaced with corresponding references for this discussion):

Clinicians administer atropine routinely to all asystolic patients. On rare occasions high levels of parasympathetic tone may lead to cessation of both ventricular and supraventricular pacemaker activity.[Brown 1979] Electric shocks also produce parasympathetic discharge.[Brown 1979, Vassalle 1985] For this reason routine shocking of asystole because "it cannot make the rhythm any worse" should be strongly discouraged. Such shocks to asystole could eliminate any possibility for return of spontaneous cardiac activity. Studies of shocks for asystole have detected no improvement in survival.[Thompson 1984, Stults 1987, Stults 1984] Rescuers should confirm asystole as the rhythm when faced with a flat line on the monitor by changing to another lead on the lead-select switch or by changing placement of the defibrillation paddles by 90^o. Operator errors that lead to "false asystole" are much more common than VF that masquerades as false asystole.[Cummins 1988]

The  argument begins with the statement that atropine is routinely administered in asystolic arrests because ``on rare occasions high levels of parasympathetic tone may lead to cessation of both ventricular and supraventricular activity.'' The only article cited to support this statement [Brown 1979] is speculative in tone, offering possible explanations for preliminary positive results with a small (n=8) series of patients; these results have not been confirmed by larger studies in the 20 years since the article was published--and in any case, showing that atropine is effective for asystole would still be far from showing that countershocks are harmful. The text of the Guidelines continues, ``Electric shocks also produce parasympathetic discharge,'' but the two references cited to support this statement [Brown 1979, Vassalle 1985] actually are silent about the effect of electric shock on parasympathetic tone. Also, the principal role that Vassalle postulates for increased parasympathetic tone in ventricular standstill does not apply to cases of pre-existing asystole (as a reading of the article indicates, and as JAS confirmed in a phone conversation with the author). The argument ends with the statement that ``studies of shocks for asystole have detected no improvement in survival.'' However, prohibiting a treatment on that basis, if applied consistently, would call into question the entire ACLS algorithm for asystole, including the use of atropine for asystole with which the argument started.

The ACLS Guidelines call for positive confirmation of VF on the monitor in order to avoid shocking asystole. The Guidelines acknowledge the occurrence of occult VF and mandate checks in two leads to rule it out, also calling (at least implicitly) for a check for flat line artifact as well. If these maneuvers produce a shockable pattern, countershock is to be delivered. With this protocol, a potentially effective and highly time-dependent treatment may be delayed in order to eliminate the possibility of delivering a hypothetically harmful treatment to a pulseless patient in a nonshockable rhythm, whose chance of neurologically intact survival is vanishingly small in any case. In addition, confirmation of VF may not be possible even with switching leads: in a recent case report [Amaya 1999], after checks in two leads failed to confirm VF, it was confirmed by ultrasound and was converted to a perfusing rhythm by a single countershock.

The current ACLS guideline does not address the problem of occult VF in relation to adhesive hands-free electrodes and appears to rule out shocks to patients with occult VF or fine VF in one lead if hands-free electrodes are used. The practical result is to prevent a proven treatment for some patients in VF in order to preserve a hypothetical "last chance'' for patients in nonshockable pulseless rhythms.

The rationale for the current guideline prominently mentions the possibility of operator error and flat-line artifact, raising the possibility of a patient not in asystole or VF suffering significant harm from receiving an inappropriate shock. Though of course the probability of such an occurrence cannot be precisely determined, some idea of the likelihood can be gleaned from existing data. Let's assume the probability of a caregiver missing a perfusing pulse and proceeding with treatment for an arrest is one in ten. Figures from Cummins 1986 for the incidence of flat-line artifact are 10 in 118, or 7.9%. The probability of inducing VF from an unsynchronized countershock has been calculated at 2%--let's say 5%. That gives us 0.10 x 0.079 x 0.05=0.0004 (or 0.04%) incidence among patient presenting in apparent asystole, and that is disregarding the probability that a prompt second shock could convert the patient back to a perfusing rhythm.

In addition to its import for individual cases, a prohibition on shocking asystole emphasizes specificity beyond its true clinical importance and strengthens the mindset that defibrillation depends on a sophisticated differential diagnosis of the monitor pattern, either by a caregiver with advanced training or by a computer algorithm. In so doing, it presents at least a small additional obstacle to achieving widespread rapid defibrillation. The greatest public health threat associated with defibrillation is simply delaying it. The major purported danger of shocking asystole lacks a scientific basis, and the chance of an inappropriate shock significantly injuring a patient is miniscule when compared to the dangers of defibrillation delayed or denied to patients in VF.

The available evidence supports at least a neutral stance on empiric shocks for apparent asystole. Though there is no compelling evidence that this approach will significantly improve survival, the question will remain unresolved if future research is blocked by the current negative recommendation. Rescinding that recommendation requires an explicit statement: merely dropping it without comment will keep it alive in the minds of caregivers and researchers for years to come (The "I remember reading somewhere that it's bad" effect). An explicit statement is also necessary to address the problem of detecting occult VF when using self-adhesive dual-function electrodes, which effectively prevent checks of alternate leads.

Recommendations

Explicitly rescind negative recommendation on shocking asystole.

Proposed treatment guideline: Shocks for apparent asystole

When using adhesive dual-function electrodes: IIA or IIB (acceptable and useful,  good or fair evidence) except immediately post-shock after conversion of VF or VT or in monitored arrests in which asystole was preceded by a bradyarrhythmia.

Other situations: IIB (acceptable and useful, fair evidence) or Indeterminate, except immediately post-shock after conversion of VF or VT or in monitored arrests in which asystole was preceded by a bradyarrhythmia.

References

References from current ACLS guideline:

[Brown 1979] Brown DC, Lewis AJ, Criley JM. Asystole and its treatment: the possible role of the parasympathetic nervous system in cardiac arrest. JACEP. 1979;8:448-452.

Only indirectly related to the current guideline's rationale; used a small, nonrepresentative sample; irrelevant to the second point for which it was cited (increased parasympathetic tone caused by countershocks).

[Cummins 1988] Cummins RO, Austin D Jr. The frequency of 'occult' VF masquerading as a flat line in prehospital cardiac arrest. Ann Emerg Med. 1988;17:813-817.

Some may differ with the authors' conclusion that a clearly treatable condition occurring with an upper confidence limit of 8.8% should be termed "rare" in a patient group where survival is generally conceded to be less than 1%. Beyond that, the study (like the Martin study) considered only treatment of initial asystole (Ornato suggested a greater incidence of late occult VF). The study design called for a flat-line artifact check (leads and cable, calibration on screen and paper printout, and battery status) before checking alternate leads. No information was given about the time taken for these artifact checks. Since it is well-accepted that VF decreases in amplitude over time, finally becoming indistinguishable from asystole, it is quite possible that some instances of occult VF were missed due to these delays. Also, the point is made that operator error was the cause of apparent asystole more often than true occult VF. What significance this has for treatment is not explained; delay of defibrillation due to operator error would seem to be as much a treatment error as delay due to monitoring in a suboptimal lead. Finally, two occurrences of VF after apparent asystole were excluded from the statistical analysis because they were determined to have occurred several seconds after the lead changes. It seems at least plausible that some small error in synchronizing the audio and ECG records might be responsible for this discrepancy, rather than some unknown mechanism producing VF after a period of true electrical inactivity. If these two cases had been included as occult VF, the confidence limits for frequency would have been higher. The discussion includes an argument that parallels the parasympathetic stimulation argument in the ACLS Guidelines under review here. In addition, references are cited to support the possibility of countershocks otherwise interfering with the spontaneous recovery of the myocardium and/or natural cardiac pacemakers. These arguments appear to refer to shocks given for transient post-defibrillation asystole, for which no author to our knowledge has ever advocated countershock.

Apart from the authors' discussion and conclusions, which parallel the rationale supporting the current negative recommendation (see discussion above), the results of the study can reasonably be taken as neutral or even as supporting the proposed revision: the authors acknowledge that occult VF does occur and that ACLS protocols should deal with it by checking a second lead. The problem of doing this with adhesive dual-function electrodes is not addressed.

[Stults 1987] Stults K, Brown D, Kerber R. Should ventricular asystole be cardioverted? Circulation. 1987;76(suppl IV):IV-12. Abstract.

Abstract; no information on method of group assignment.

[Stults 1984] Stults K, Brown D. Converting asystole. J Emerg Med Serv. 1984;9:38-39.

Single case report of countershock conversion of apparent asystole after two previous countershocks for VF-could have been coincidental conversion of post-conversion asystole (occurred one minute after previous shock).

[Thompson 1984] Thompson BM, Brooks RC, Pionkowski RS, Aprahamian C, Mateer JR. Immediate countershock treatment of asystole. Ann Emerg Med. 1984;13(9, pt 2):827-829.

[Vassalle 1985] Vassalle M. On the mechanisms underlying cardiac standstill: factors determining success or failure of escape pacemakers in the heart. J Am Coll Cardiol. 1985;5:35B-42B.

Other references:

[Martin 1993] Martin DR, Gavin T, Bianco J, Brown CG, Stueven H, Pepe PE, Cummins RO, Gonzalez E, Jastremski M. Initial countershock in the treatment of asystole. Resuscitation. 1993;26:63-68.

This study is weakened by the fact that it is a retrospective analysis of a sample in which the only two survivors (no info about neurologic status) were known to be in the "no countershock" group-possibly that was a major impetus for the retrospective look at the records. The study might also be open to a charge of "data-dredging," given the variety of endpoints, subgroups, and possible variance factors that were looked at (though that is not a major issue in a study like this, which makes no claim to be definitive). The two groups were in different EMS services and were somewhat mismatched with regard to several arrest factors. Some differences favoring the no countershock group came close to the 0.05 level of significance in the bystander CPR subgroups, but none came close when the entire groups were considered. The study considered only treatment of initial asystole, though a previous study (Ornato 1985) had suggested that countershock is more effective for late asystole. Finally, the study excluded from the analysis any cases of conversion of asystole to a perfusing rhythm on the first shock. Apparently it is not known whether there actually were any such cases, and the authors argue that conversion on the first shock is highly improbable (based on the assertion that 'occult' VF is most likely fine VF). However, any such cases would obviously have improved the results of the countershock group, and there is one well-documented case (Amaya 1999) showing that such instances can actually occur. Considering all of these factors, and despite the fact that the authors conclude the results suggest a possible negative effect of countershock for asystole, the study results must be considered neutral at worst with regard to the proposed revision.

[Ornato 1985] Ornato JP, Gonzales ER, Morkunas AR, Coyne MR, Beck CL. Treatment of presumed asystole during pre-hospital cardiac arrest: superiority of electrical countershock. Am J Emerg Med. 1985;3(5):395-399.

This study set up comparison groups of interventions (i.e., countershocks and medication doses) but did not have a separate comparison group of cases. Results indicated that countershocks for apparent asystole may be more effective (and by inference, that 'occult' VF may be more common) late in an arrest than with initial asystole. Countershocks were found to significantly more likely to effect a rhythm change from apparent asystole than the standard ACLS drugs. The study design did not rule out confounding/synergistic effects between drugs and shocks.

[Pansegrau 1970] Pansegrau DG, Abboud FM. Hemodynamic effects of ventricular defibrillation. J Clin Investigation. 1970;49:282-297.

From the abstract: "[I]n anesthetized dogs…fibrillation was maintained for 15 to 30 sec and terminated with a 400 w sec capacitor discharge across the thoracic cage. Responses lasted 1-10 min after conversion and included a cholinergic and an adrenergic component. …These responses were related primarily to conversion of ventricular fibrillation rather than to the electrical discharge of countershock because countershock without ventricular fibrillation caused more transient and smaller responses than those observed with defibrillation; furthermore, the hemodynamic effects of defibrillation were augmented by prolongation of the duration of fibrillation."

[Amaya 1999] Amaya SC, Langsam A. Ultrasound detection of ventricular fibrillation disguised as asystole. Ann Emerg Med. 1999;33:344-346.

[McDonald 1982] McDonald JL. Coarse ventricular fibrillation presenting as asystole or very low amplitude ventricular fibrillation. Crit Care Med. 1982;10(11):790-791.

[Ewy 1984] Ventricular fibrillation masquerading as asystole. Ann Emerg Med. 1984 S;13(9 Pt 2):811-812.

[Ewy 1981] Ewy GA, Dahl CF, Zimmerman M, Otto C. Ventricular fibrillation masquerading as ventricular standstill. Crit Care Med 1981;9(5):392.

Web Pages Created With AOLpress