Long-term outcome of survivors of cardiac arrest whose therapy is guided by electrophysiologic testing

JACC Vol . 19, No . 4 Merck 15, 1992:780-8

780

ELE CTRO PHYSIOLOGIC STUDIES Long-Term Odtcome of Survivors of Cardiac Arrest Whose Therapy is Guided by Electrophysioiogic Testing RICHARD N . FOGOROS, MD . FACC, JAMES J . ELSON . MD, PHD, CHRISTOPHER A . BONNET, MD, FACC . SUSAN B . FIEDLER, RN, JOHN C- CHENARIDES, MD P3rr.h,, 0r0,

Pr,,,,, y trr,r,rr,

The long-term outcome of 217 consecutive survivors of cardiac arrest whose therapy was guided by at irephyaiulugiclustingwas analyzed. After eleclrophysiolopir testing, St patients (37%) were classified as having no Inducible arrhythmia and were treated without aatierrhythmic drugs; 23 received an implantable de . llhrillator. Of the 136 patients with inducible arrhythmia, the 51 (38%) who responded to serial drug testing were treated with the saecessful drug and the 35 (62%) with unssovessful drug testing were treated with an hnptantable defibrillator (47 patients), amiodaroue (36 patients) or drops that were unsuccessful during testing (2 patients) . The mean fallow-up interval For all patients was 35 ± 23 mouths. The actuarial Incidence of sudden death and overall death was similar for patients whose arehytlmias were not inducible, drag responders and nonresponders. The actuarial incidence rate of recurrent arrhythmic events In nonresponders was 35 ± 5% and 53 ± 7% at 2 and 5 years, respectively . These values were signlhcaully lower (and statistically similar to each other) to the other two patient groups : patients with nnnindueihle arrhythmia

(19 ± 5S, and 31 ± 7%, respectively, p c 0.05) and drug responders (13 t 5% and 23 t 5%, eospa tivdy, p a 0.01) . t'alients with an implantable defibrillator who had recurrent arrhythmic events were sianlgcandy less likely to the suddunly than were patients without a defibrillator who had recurrent events (p < 0.001). The eteclrophystelsgic study was useful in subgrouping survivurs of cardiac aresstinm rnl 5ively high risk (nourespordoes) and low risk (patients with no inducible arrhythmins and drug responders) groups, but even the low risk groups had a substantial cumulative risk of recurrent arrhythmic events, With an implant . able degbrlgator, mortality in the high risk group was similar to that in the low risk groups despite a significantly higher incidence of recurrent mAythmic events- The use of the implanlahle defibrillator should be studied prospectively in survivors of car-

Although electeophysierlogic sealing is often considered a necessary step in determining therapy for survivors of car . diac arrest, there are few actuarial data on the outcome of therapy guided by such testing in these patients . Consequently . the long-term prognosis is not well characterized for

center between July L982 and January 1990 whose therapy subsequently was guided by invasive electeophysiologic testing . Each patient had presented with sustained ventricular taehyeardia or ventricular fibrillation that occurred in the absence of acute myocardial infarction and caused a loss of consciousness that required cardiopulmonary resuscitalion. The clinical characteristics of these patients are shown in Table I .

some subsets of these patients whose therapy is based on

eltorophysiologic testing . The present retrospective analysis attempts to estimate the lung-term efficacy of therapy guided by clectrophysiologic testing in a large group of survivors of cardiac arrest-

diac arrest who respond to drug testing or have no Inducible

arrhythmia . (J Am Co41 Carsliol J99J ;19;780-8)

Elecfropkysiofogfc Testing Methods Study patients (Table Ii . The study group consisted of 217 consecutive survivors of cardiac arrest referred to our From the "run or Cardiclogy, Allegheny General Hospital, Medical Coney of N-rlvuoia . Pttlcbargb . nr.ns,slvur~a. Manuscript received July 1,1991 : revised nlaaucdpl received September 30, 1991, axepmd Oetobee 6, 1991. Add,,forreedatu: Richard N . Fogoras, Mn, Alleaheny General

Hospital-Eleclraphysiolagy Lehantory,120 East North Avenue, Hhsburet, INnnnylousie 15212, 51992

by Ihr American College orC.ud :mossy

The stimulation protocols used in electrophysiologie test-

ing and the definitions used for arrhythmia inducibility and noninducibility and drug response remained the same throughout this series . Before baseline (drug-free) electro . physiologic testing, all 217 patients went through a period of stabilization in which all class I antiarrhythmic medications were withdrawn, metabolic disturbances were corrected and therapy for isehem is heart disease (including coronary artery bypass grafting, if necessary) and congestive heart failure was optimized . All but 7 of the 217 patients underwent their initial electrophysiologic study in the drug-free state ; these 7 0735-1097192155.00



1ACC Vol. 19. No. 4 March 15 . 1992'.780-8

FOGOROS ET AL. ELEC1ROPHYSIOLOGIC TES ING A : TErt CARDIAC ARREST

Touts 1 .

Clinical Characteristics

of 217

781

Pelicans Patient Group

Ali

Tow .. Malclfemala nib. Mean ago (Y') Follow-up Imo) Cardiw diseoxe CAD •

Cardiomyopalhy Valvular None+ MsaeLVEF1017

J' 17 154)83

Arrkylhmm Not Indad'le

Drue Ret

81 t 12

51

15 - 23

33

004015

NenrcspoMkrc

5)

M 900 31

„5. -

19112

6ADI

hl _ !3

61 ± 11

35 '-0

3x ± 25

15217070!

42lihl

381754)

72185`/,1

331)541 2711274)

10117701

7)147) 7)1470)

121147)

201 7! 36 .14

14)11701 IIZ15

161 5110-7 111

1211 pla)

176 prsl

151 315)

14117701

6170) 11170) 32 0 13 184 past

-p < 0.011 . patiems with noninducible orthythmios sersus nonrcsponder„ p < 0.05. patients with nedind0dhte anhythmios versus drug responders: 7p < 0.001. mmrespnndcrs vtrs,n patients With nnnindueible arrhythmia ; p < 0.1 . norresponders versus drug responders : up < 0.05. patients with noninducible armymmla versus drug rspondere, p < 0.001 . palienl5 with nonindoc,ble 'arrhylhmias versus nonrespmMers ; p < 0.1 . drug responders nersua nunreapondcrs. CAD = cot- ..y artery disease : LVEF

were taking amiodarone because they had required it before testing to stabilize frequent recurrent ventricular arrhythmia. Baseline eleetrophysiologic testing . This was performed initially with the pacing lead positioned in the right ventricular apex. After eight paced heats at each of three cycle lengths (600, 500 and 400 ms) . first single, then double, then triple extrastimuli were added and each coupling interval was progressively reduced to ventricular refractoriness or to 200 ms. Burst pacing was then performed with 8 to 12 beats at cycle lengths from 350 ms to ventricular refractoriness or to 200 ms. If a ventricular tachyarrhythmia was not induced . the entire protocol was repeated front the right ventricular outflow tract . If a ventricular tachyarrhythmia was still not induced, an isopruturenol infusion of u5 yng/miu was begun (unless clinically contraindicated) to increase the rest heart rate to 110 to 130 beatsimin and stimulation was repeated . Inducibility war defined as the ability to reproducibly induce at least 10 beats of a ventricular tachyarrhythmia of any configuration . In general . the entire protocol was completed unless sustained ventricular tachycardia was induced . However, if several episodes of symptomatic. hemudynamieally compromising nonsustained ventricular tachycardia (that is, terminating spontaneously after 10 beats and before 30 s) were reproducibly induced. the baseline protocol was occasionally not carried to completion . Of the 217 patients . 136 (63%) were categorized as having inducible arrhythmia during the initial study (including all 7 who were receiving amiodarone) and the remaining 81 were considered to have no inducible arrhythmia . The mean cycle length of the initially induced arrhythmia was 235 ± 59 ms. Serial drug testing. For patients who had inducible ventricular tachyarrhythmia, serial drug testing was performed with the same stimulation protocol used during the baseline study. We tested the following drugs unless they had been

-

lea vemricular ejection fraction .

unsuccessful or had caused toxicity during empiric trials or were clinically contraindicated : procainamide, quinidine, phenytoin . disopyramide and mexiletine (usually in combination with a class la agent) . In patients whose clinical arrhythmia was associated with exercise or in whom an isoprolerenol infusion was required to induce the arrhythmia, we also tested beta-adrenergic blocking agents or verapamil . or both. We did not test class Ic agents in these survivors of cardiac arrest . A patient was considered a drug responder if no more than nine beats of ventricular tachycardia were seen during an electrophysiologlc drug study and if the patient was subsequently discharged on treatment with adequate doses of that drug . The 136 patients with inducible arrhythmias had a mean of 2.8 ± 1 .7 drug tests. Fifty-one (38%) of the 136 were drug responders (I8 responding to phenytoin, 13 to quinidine, I I to procainamide, 3 to a combination of mexiletine plus a class la drug . 2 to timolol and I each to mexiletine alone, disopyramide vesapatail and amiodarone plus procainamide). The other 85 patients (62%p) with inducible arrhythmia had unsuccessful serial drug testing and were classified as nonresponders . The clinical characteristics of all 217 patients in the total study group, are shown in Tahle I .

Therapy Patients without Inducible arrhythmim . For these

g1

pa-

tients the primary therapy was withholding of all class I and class Ill antiarrhythmic dregs . In addition, an attempt was made in each case to locate a reversible cause for the index arrhythmia and to advise the patient and the referring physician to take preventive measures . Finally, beginning in 1988, because of new reports (1,2) suggesting that survivors of cardiac arrest with noninducible arrhythmia might not



782

FOGOROS ET AL.

JACC Vat . 19, No. 4 Mares 13, 1992:780-8

ELEcrearHTSIDLGGIC TESTING AFTER CARDIAC ARREST

have a benign prognosis, such patients in this series whG did net have an identifiable reversible cause for the index arrhythmia were offered treatment with an implantable defibrillator- Antiarrhythmic medications were withheld in the patients in this group who received the device, Drug responders. For these 51 patients, therapy was based on long-term use of the drug that was successful during etectrephysiologic study. Target drug levels far tongterra thcrEpy were defined as the drug levels measured during the successful electrophysiolagic study . Nonresponders. For these 85 patients the therapy of choice was the implantable defibrillator. The device was used in 47 nonrcspoedcrs ; in the other 39 it was refused by the patient or was unavailable (the implantable defibrillator was extremely difficult to obtain during some periods in 1985 and 198fi 131) or the patient's general medical condition was such that survival after surgery or [or an extended period after implantation was considered unlikely . Our assessment of the patient's general medical condition was subjective and was based largely on the patient's functional capacity and the presence and severity of underlying cardiac, peripheral vascular, cerebrovascular, renal, psychiatric and pulmonary disease . In the 38 nonresponders cviw did rim receive an implantable defibrillator, therapy with amiodarone was recommended . From 1982 to 1987 the use of this drug was entirely empiric. Beginning in 1987, electrophysiologic testing was offered after I to 2 weeks of therapy with amicdarone . 1' during the study of amiodarone the resulting arrhythmia caused hemodynamic collapse, an attempt was made to add a second drug that further slowed the induced anhythmia . For nomesponders who also refused therapy with amiedaone, an attempt was made to select the best ineffective drug tested, that is, the drug that resulted in the slowest induced arrhythmia or made the induced arrhythmia relatively difficult to induce . Follow-up . Patients who received the implantable defibrillator were seen in our center every 2 to 3 months . Other patients were generally returned to their referring physicians for follow-up care . Patients not actively followed up in our center were contacted by telephone approximately every year . The current status of 213 of the 217 patients in this study was ascertained between November 1990 and May 1991 . Four patients in this study were lost to follow-up study after 18, 25, 28 and 59 utonths of follow-up . respectively . Definitions. A recurrent arrhythmic event was defined as I) recurrent sustained veetrirulartachycardin or ventricular fibrillation that required intervention for ierminalion, 2) an appropriate shock delivered by an implantable defibrillator, or 3) sudden death . An appropriate shock was defined as delivery of a shock or a series of shocks by an implantable defibrillator in response to syncope, presyncope or severe light-headedness. Sudden death was defined as unexpected death that occurred during sleep or within l h of the onset of symp toms . Two patients who had sudden loss of consciousness

sudden c

pal Orndl 11-th !7!T! =I1

Years Follow.up SID RF

111012 579/2

1

e

nut

is/s

ws

Figure L Actuatialincidence of sudden death (SO), overall death (CID) and recanenl arrhythmic events (RE) in all 217 survivors of cardiac arrest . The table gives the number of patients (numerator) and the percent standard coon {denominator) for each actuarial curve at yearly intervals of follow-up .

due to ventricular tachyarrhyt h miss that required cardiopulmonary resuscitation but that resulted in ultimately terminal coma were tabulated as having sudden death . Statistical methods. Values are reported as mean values SD . Actuarial analysis of longitudinal follow-up data was calculated by the Kaptan-Meier method and actuarial values are reported as mean values ± SE. Actuarial curves were compared with use of the Mantel rank order test and mean values with use of the Student r test. Proportions were compared with the chi-square test. Statistical significance was defined as a p value < 0.05 . Results The mean follow-up interval for all 217 patients was 35 t 23 months . The actuarial incidence rates of recurrent arrhythmic events, sudden death and overall mortality for all 217 patients are shown in Figure 1 . Patients Without Inducible Arrhyrhmias Of the 81 patients classified as having no inducible arrhythmia after baseline testing, only 2 were inking a class I or class Ill antiarrhylhmic agent duriugfollow-up . Twentythree 128%) of the 81 patients received an implantable defibrillator. The mean follow-up interval for all 81 patients was 32 ± 21 months . The arruarial risk of recurrent arrhythmic events in this group was 19 t 5% at 2 years and 31 ± 7% at 5 years (Fig . 2) . Twenty-five patients with no inducible arrhythmia died during follow-up (Table 21. The overall actuarial mortality rate was 22 ± 5% at 2 years and 37 a 7% at 5 years (Fig. 3) . None of the 23 patients without irrdaci bte arrhythmia who received the implantable defibrillator died suddenly . In the



IACC Vol . 19. No. 4 March 1 5 . 1992 :750-8

too no m

FOGORDS ET AL. ELECTROPHYSIOLOGIC TESTING AFTER CARDIAC ARREST

Neehdudhle Draft r-pande. m . . ...... . . . ...... . . ......

70

783

Nonuduedb unrg r~mdan Names ..rides _

%With so Recurrent Arrhythmic 50 Events 4p w

w zv

1v 0~ 1

2

3 V- 1000w

Years r-ofow-op NI DR

NR

7512

4615

4913 55(5

4015 43(5

3216 3716 33(6

2016 19(8 2516

N5

712

6101 1517

DR

7014 4455

NR

69/4

tap

5215 5706 3616 2597 6114 4595

2566

two

717 7)E

25)7

7517

Figure 2. Actuarial incidence of recurrent arrhythmic events in 81 patients with noninducibte arrhythmia (N1) . 51 drug responders (DR) and 85 nonrespunders (NR). The table gives the number of patients (numerator) and the percent standard error (denominator) for each actuarial curve at yearly intervals of follow-up .

Figure 3 . Actuarial incidence of overall death in 81 patients with noninducibte arrhythmia (N), 51 drum responders (DR) and 85 nonresponders (NR) . The table gives the number of patients (numerate) and the percent standard error (denominator) for each actuarial curve at yearly intervals or fallow-up-

58 patients who did not receive the device, 9 of 10 recurrent arrhythmic events were manifested by sudden death . The actuarial risk of sudden death for all 81 patients was 8 . 3% at 2 years and 19 ± 7% at 5 years (Fig . 4). For patients without the implantable defibrillator, this risk was 14 -_ 5% at 2 years and 28 m 8% at 5 years . Influence of an identifiable reversible cause on outcome . In 42 (52%) of the 81 patients with no inducible arrhythmia, a possible reversible cause for the presenting arrhythmia was identified. These included cardiac ischemia in 20 (cardiac ischemia was identified as a potential reversible cause if 1) the clinical setting of the presenting arrhythmia was one in which an increased myocardial oxygen demand was likely in a patient with documented coronary artery disease, or 2) frank angina preceded the presenting arrhythmia), the use of class I antiarrhythmic drugs in 15 . acute alcohol toxicity in 4 and hypoxemia due to exacerbation of chronic lung disease in 3 . There was no significant difference between the 2-year actuarial risk of recurrent arrhythmic events in the 42 patients with a reversible cause for presenting arrhythmia (16 s 6%) and that in the 39 patients without an identifiable reversible cause (22 ± 7%).

However, only I of the 20 patients in whom ischemia was the presumed reversible cause for arrhythmias had a recurrent event. The actuarial risk of recurrent events in these 20 patients was 5 '- 5% at 2 years versus 24 t 9% in the 22 patients who had a reversible cause other than ischemia (p < 0. D. Drag ,csponders These 5 1 patients were followed up for a mean of 35 20 months . During this time drug intolerance developed in three patients who underwent further electrophysiologic testing to

Figure 4. Actuarial incidence of sudden death in 81 patients with noninducible arrhythmia INI), 51 drug responders (DR) and 85 nonresponders )NR(. The table gives the number of patients (numerator) and the percent standard error (denominator) for each actuarial curve at yearly intervals of follow-up . so 1 Nnnmbndble Drug-aepusdxs r4waerper5dn4

45{ b

~. Table 2 . Mode of Death in 217 Patients

en Sudden Death

t atiest Group

Tetaldac./o Sudden death Nonsuddcn cardiac death Noncardiae death

ee ls-

Arrhythmia Not Inducible (n = 81) .5 9 (369) 9(36%)

Drug Responders .a = 511

Nonresponders (n = 85)

17

72

665%) 605%)

10 (3191 17 (53%)

7 (26%)

5129%1

5(1617,)

1o

H, NI OR NR

7592 4613 6913

710005 Follow-up 52/3 3714 30/3 20/4 67/3 45(3

2595 1917 2516

712 7/7 ISIS



784

JACC Vol . 19, 90.4 March 15, 7992:780-8

FOlu000S ET AL . ELECTROPMYSIOLOGIC T£ST[NG AFTER CARDIAC ARREST

attempt to identify a second successful drug . In two of the three a second successful drug was found; 7n the third an implantable defibrillator was inserted . The actuarial risk of recurrent arrhythmic events was 13 t 5% at 2 years and 23 ± 8% at 5 years (Fig . 2) . Seventeen drug responders died during follow-up (Table 2) . The overall actuarial mortality rate was 24 ± 69% at 2 years and 45 - 8% A 5 years (Fig. 3). Sudden death was the first manifestation of arrhythmia recurrence in six of nine patients who had recurrent arrhythmias . Among the three patients who survived the recurrences, two were subsequently treated with an implantable defibrillator and one was treated empirically with aaaodarone . The actuarial risk of sudden death For all 51 drug responders was 6 3% at 2 years and ti ± 7% at 5 years (Fig. 4). Influence of specific drugs used. The risk of recurrent arrhythmic events was not significantly related to the specific successful drug used . Recurrent events were seen in 3 of IS patients treated whit phenytoin, t of 13 treated with quinidine, 4 of 11 treated with precainamide and I ai - 2 treated with rimolol. Nenrere.sponders The mean follow-up interval for these 85 patients was 38 25 months. Forty-seven (55%) of these patients received an implantable defibrillator. Twenty-nine (62%) of the 47 also received antiarrhythmic drug therapy to reduce the frequency of atria[ or ventricular aerhythmias (21 received amiadanme and 8 received a class Ia or lb drug]. Of the 38 patients (45%) who did not receive an implantable defibrillator, 36 were treated with amiodarouc and 2 refused amiadarone and were treated with a dnrg that had been ineffective during serial drug testing . The actuarial risk of recurrent arrhythmto eo'enrs in this group was 35 ± 5% at 2 years and 53 ± i% at 5 years (rig . 2). Thirty-two nonresponders died during follow-up (Table 2). The overall actuarial mortality rare was 17 a 4% at 2 years and 46 t 7% at 5 years (Fig. 31 . Sudden death. Ten nonresponders experienced sudden death . For patients without the implantable defibrillator, sudden death was the first manifestation of recurrent an rhythmias in R of I I patients who had recurrent events . Two patients with an implantable defibrillator died suddenly (in one of these cases the defibrillator battery was depletedl . The actuarial risk of sudden death for all nonresponders was 8 - 3% at 2 years and 155 '- 61 at 5 years (Fig . 4). Role of implantable defibrlllatar . Selection bias was applied in determining whether a neenresponder received an implantable defibrillator because this decision was often based an a subjective clinical impression regarding a patien s overall prognosis . Accordingly, patients who did not receive the defibrillator were significantly more likely to experience nonsudden death during follow-up (21 such deaths in 38 patients) than were patients who received a

gable 3. Comparison of Overall Actuarial Outcomes in 217 palings aecunen7 Arrhythmia Sudden Death Overall Dealt Events NIvs .DR p=Ne p--NS p-NS N Ivs.NR p
defibrillator (9 such deaths in 47 patients) (p ° 0,01) . The overall actuarial mortality rate in nonresponders who did not receive an implantable defibrillator was 34 ± 8% at 2 years and 61 r 9% at 5 years. Comparison of nutenmes for patients without inducible arrhythmias, drug responders and nonrsponders. Comparisons of the actuarial incidences of recurrent arrhythmic events, sudden death and overall mortality among these three groups of patients arc shown in Table 3 . The three groups had almost identical actuarial sudden death mortality and total mortality. Patients who had no inducible arrhythmias at baseline and those whose arrhythmia was rendered noninducihle during drug testing had statistically similar incidence rates of recurrent arrhythmic events that were significantly lower than rates for nonresponders . Influence of ejection fraction en recurrent arrhythmic events (Table 4) . Patients without inducible arrhythmia and nonresponders with a left ventricular ejection fraction x30% had significantly more recurrent arrhythmic events during follow-up than did patients with a higher ejection fraction . The same trend was present for drug responders but did not reach statistical significance . However, the range of ejection fractions in patients who had recurrent arrhythmic events was wide in all three groups.

Table 4. Influence of Ejection Fraction on Recurrent Arrhythmie Events paunia creep Arr hylhmins

NotIn&eiate

orug Responders

Nonrespnnden

EFS30u,

22

22

43

EF531rd

54

29

41

2." arhnvial

rid, of

reemeani arrhythmia taunts EF 530% EF >3LCh RangcofU1 7 For

50 x 127t

22 0 10%

x 4%c

7 ! 4%

19%to7n%

18%to51%

49 ± 8%c 20 0 6%

13%to53%

patients wish rrem anhylheic events eEF C 30 versus EF > 30, p e 0A I: vEF s 30 versus EF > 30,p<0 .001 . EF = left ventricular ejection fraction .



FOGOROS ET AL. FLECTROPHYSIOLINAC TESTING ACrER CARDIAC ARREST

IACC VOL 19, No. 4 March 15, 1992:790-a Table 5. Influence of the Presence of Coronary Anery Di.cane on Recurrent Arrhythmic Events P'alirri Soup

No. with CAD No, without CAD' 2-yr ucmorial risk of at arrhythmic events Patients with CAD NO,- without CAD

Arrhythmia Notlnducible

Drag Responders

Nonrespanderc

42 39

39 13

72 13

19 5 6A 19 x 17,

12 a 501 is - 1190

35 ! 65/ 32 - 1391

785

tan so sudden deed, m n aeA m araymnae - .-.. % '50 a 30 rn 1. Years Follow-up

'Rupunlon of patients without inducible v+,nrri-uta : tachycantia, p < 0 .001 . comparing patients with versus those without coronary artery disease . CAD = wronary artery 6i,...

Influence of the presence of coronary artery disease on recurrent arrhythmic events (Table 5). Patients without coronary artery disease were significantly less likely to have inducible ventricular tachycardia during baseline testing than were patients with coronary disease (p < 0 .001) . However, once patients were classified by electrophysiologic testing, the actuarial ineidence off rec=ent arrhythmic events was similar for patients with and without underlying coronary artery disease for all three sub troops . Influence of the induced baseline arrhythmia on recurrent arrhythmic events . Table 6 shows the outcome of patients whose baseline study induced arrhythmia other *.ban sustained monomorphic ventricular tachycardia . Although a response to serial drug testing tended to predict a more favorable outcome even in these patients . that tendency did not reach significance (probably because of the small number of patients in this subgroup) .

65/i

4011

29N

1412

9/2

6512

4613

2914

1417

917

200

1517

55/5

907

Influence of the implantable defibrilillator on survival- Seveiaty 325/) of the 217 survivors of cardiac arrest received an implantable defibrillator as primary therapy for prevention of sudden death . Only 2 of the 30 patients with an implantable defibrillator who had recurrent arrhythmic events died suddenly (7%) compared with 23 (77%) of 30 patients without a defibrillator who had recurrent arrhythmic events (p < 0 .001) . The actuarial incidence rate of sudden death in recipients clan implantable defibrillator was I ± 1% at 2 years and 3 ± 2% at 5 years (Fig. 5). The overall actuarial mortality rate for recipients of a defibrillator was 6 ± 3% at 2 years and 21 s

Table 6 . Influence of Baseline Arrhythmia on Outcome in 136 Patients Patients

Classified as Having Inducible VT DNB

All patients Inducible VT Sustained, monomorphic Polymorphic VT/VFNonsustained Polymorphic and/or nonsustained 2.yr actuarial risk of recurrent anhythmte events All patients Inducible VT Sustained, monomorphic Polymorphic VT/PT .N .stained Palymorphic and/or nonsustained

(J7

Figure 5 . Actuarial outcome for 70 recipients of the implantable defibrillator. The table beneath the figure gives the number of patients (numeratost and the percent standard error (denominator) for each actuarial curve at yearly intervals of follow-up . Abbreviations as in Figure I .

p Value

Patients

Responders

Nonrespwtders

136

51

05

95 29 24 41

31 6 9 20

64 13 15 21

13±591

3505%

<0.01

15-77, 13 n VA 0 10±7%

a-717 225 ± 5390 14 0 9% 20!9%

<0.05 NS N5 NS

'Proportion of patients responding to serial drug testing . polymorphic ventricular tachycardia or fibrillation versus sustained moromorphic ventricular tachycardia . p < 0.05 . NSVT = nonsustained ventricular tachycardia ; VT = ventricular tachycardia : VTiV F = polymorphic ventricular tachycardia or fibrillation .

786

FOGOROS ET AL . ELECTROPHVSIOLOGIC TESTING AFTER CARDIAC ARREST

7% at 5 years (no Patient died within 341 days of inmplants . Lion), The actuarial incidence of recurrent arrhythmic events in defibrillator recipients was 40 ± 6% at 2 years and 50 ± 7% at 5 years . Assuming that these recurrent arrhythmic events would have resulted in sudden death had the defibriltutor not been implanted, the yresumed sudden death mortality vithnut the implantable defibrillator -old have hesignificantly higher than both the measured sudden death mortality (p < 0 .001) and the measured overall mortality if,

< .011 0 in defibrillator recipients . Discussion This report applies actuarial methods to esandne the long-term outcome of survivors or cardiac arrest whuse therapy was guided by electrophysioiogic testior ;. It confirms and extends several previous observations . Our results regarding the proportion of cardiac arrest survivors who have inducible arrhythmia and who respond to drug testing are quite typical (1,2,4-7) . Our series also confirms previous reports (2,4 .7) that the electraphysiologic study is useful in subgrouping survivors of cardiac arrest into categories of patients who have a relatively high risk (nonresponders) and a relatively low risk (patients without inducible arrhythmias and drug responders) of recurrent arrhythmic events . NonrTsponders. A major challenge to electrophysiologists caring for survivors of cardiac arrest has been to reduce the risk of sudden death in the high risk subgroup (the nonresponders) . In the present study the selective use of an implantable defibrillator reduced the sudden death mortality in nonresponders to the extent that their actuarial incidence rates of sudden death and overall mortality were virtually identical with those of the two lower risk groups (Fig . 3 and 4) despite a significantly higher incidence of recurrent arrhythmic events . Onrdecisian I^ [rear es -v-pander nigh as brrplarrveble defibrillator was partly bused on a subjective assessment of whether the patient was likely to survive surgery and remain alive for an extended period after implantation . Our results suggest that the decision to withhold the implantahle defibrillator in some nonresponders was reasonable. Although 8 of the 311 non-ponders who did not receive an imnlamahle defibrillator died suddenly, "I (70%) of the remaining 30 experienced nonsudden death . Because none of our patients who received an implantable defibrillator died in the pcrioperative period, it is not at all clear that the more aggressive use of the implantable defibrillator in our "sicker nonresponders would 'Tave further reduced overall mortality . Prospective studies are indicated to develop criteria for deciding which high risk patients are most likely to benefit from defibrillator implantation . Patients withoul inducible arrhythmias . Although many early reports (7-9) suggested that survivors of cardiac arrest who had no inducible arrhythmia had an excellent prognosis, later reports (1 2,41 tended to suggest a less benign prognosis . The present data are in agreement with these later

1 ACC V .I. 19. Na . 4 starch 15, 1992 :780-R

reports. Although our patients with no inducible arrhythmia had significantly fewer recurrent arrhythmic events than did our nonresponders, they nonetheless had a cumulative incidence of recurrent events and also a cumulative incidence of sudden death (in thusu who did not receive an implantable defibrillator) of approximately 6%!year . Our patients without inducible arrhythmia in whom a possible reversible cause for the initial cardiac arrest could be idenhfed, fared no better as a group than the patients in whom a reversible cause was not found. However, consistent with the findings of others only I of the 20 patients in whom cardiac ischemia was believed to be the cause of the presenting arrhythmia had a recurrent event after the ischemia was adequately treated . As in the study of Wilber el al . (21, nor patients who had no inducible arrhythmia had a lower incidence of underlying cardiac disease than did patients with inducible arrhythmia, and patients without inducible arrhytania who had a higher ejection fraction had significantly fewer arrhythmic events than did those with a lower ejection fraction. Many elcctrophysiologists now consider the implantable defibrillator to be indicated for survivors of cardiac arrest who have no inducible arrhythmia during baseline testing 1131. and the Health Care Financing Administration no longer requires arrhythmia inducibility as a requirement for the implantable defibrillator under Medicare (14). Our data, which estimate a 5-year risk of sudden death approaching 30°x% in survivors of cardiac arrest who have no inducible arrhythmia and are not treated with an implantable defibrillator, support the use of this device in such patients . However. not all such patients may benefit . Patients whose in'''tial arrhythmia was precipitated by cardiac ischemia probably do not peed the device . Patients with a higher ejection fraction might appear to be less likely candidates than those with a lower ejection fraction : however, these may he the very patients who would receive the greatest long-term benefit from the device-a nearly 20% 5-year incidence rate of recurrent arrhythmic events may not be acceptable in patients with an otherwise high likelihood of surviving for many years . A prospective, randomized study would be helpful in evaluating the optimal use of the implantable dethrillamr in survivors of cardiac arrest who do not have inducible arrhythmia . Drug responders. Since the earliest reports (15,16) describing the efficacy, of serial drug testing in patients with inducible ventricular tachycardia . very little has been written about the long-term prognosis of drug responders . The actuarial incidence of recurrent arrhythmic events in our drug responders (approximately 5%lyear) was significantly lower than for nonresponders, was quite similar to that seen in patients who had no inducible arrhythmia during baseline testing and was consistent with the few previous studies (2.15) reporting actuarial data for drug responders . Drug responders in our series tended to have a higher ejection fraction than that of nonresponders and drug responders with an ejection fraction >30% tended to have fewer recur-



JACC Vol. 19 . No. 4 March 15, 1992:780-e rent arrhythmias than did those with a lower

CLECTROPHYSIOLOGIC TESTING

ejection frac-

tion . Whether the prognosis of drug responders would he improved by the implantahle defibrillator has seldom been addressed in published studies, even though these patients

have an outcome similar to that of patients without inducible arrhythmia (who frequently are treated with the device) . Because we did not implant the defibrillator in drug responders, our darn do not shed light on this issue . However . if the device is provided to patients without inducible arrhythmia . its use in drug responders at least deserves to be studied . Once again, a prospective . randomized study is called for. Limitations . The major limitation of this study is that it is a retrospective analysis . Thus, to a large extent the data cannot he used to make firm recommendations regarding therapy for survivors of cardiac arrest but instead should be used to add confirmatory evidence to previous retrospective reports and to suggest trends that can he subjected To prospective study . Because many of the patients in this study received the implantable defibrillator . esitnlating the incidence of recurrent arrhythmic events required making an assessment of whether shocks delivered by the device were appropriate . It is well recognized that making this assessment is prnhlematic . We have discussed this issue in detail 117) and have presented evidence that our definition of appropriate shocks may err on the side of underestimating the true incidence of arrhythmia recurrence . If so, our analysis would tend to underestimate the benefit conferred by the implantable defibrillator . Any report attempting to tabulate recurrent arrhythmic events in patients with the implantahle defibrillator will be subject to the same limitation until these devices are capable of storing relevant electrograms. We did not routinely use beta-blocking medication in patients who survived cardiac arrest and are not able to estimate any effect that the occasional prescription of these drugs by our referring physicians might have had on patient outcome . The appropriate Categorization of survivors of cardiac arrest in whom no sustained ntonomorphic tuclt_vcurdia is reduced during baseline testing is difficult . However. in

using the electrophysiologic study to manage these patients one must determine whether to use sulyptimal baseline arrhythmias as markers for serial drug teeing . Our decision to consider nonsustained and polymorphit tachycardias as appropriate markers for therapy was based oil 1) the Bayes theorem, which suggests that a "positive" study is relatively likely to be truly positive in patients who have had a spontaneous cardiac arrest ; and 2) the fact that without use of these markers, there were almost no other logaal approaches to therapy because the implantable defibrillator was generally not considered acceplab! •s therapy For patients without inducible arrhythmias du ing most of our study . Our results showed that serial drug testing aimed at these suboptimal arrhythmias tended to be useful, but the number of patients treated was too small to show statistical signifi-

FOGOROS ET AL. AFTER CARDIAC ARREST

787

canoe . Thus. although the data tend to validate our approach . they do not do so definitively . Now that the imp(anlable defibrillator is generally accepted as therapy for survivors of cardiac arrest with noninducibte arrhythmia, a prospective study (with defibrillator backup) of the predictive value of serial drug resting in such patients is feasibleConclusions. In 217 survivors of cardiac arrest etectrophysiologic testing was helpful in directing therapy by subgrouping patients into high and low risk prognostic categories . Patients who left the electrophysinlogy laboratory without inducible arrhythmia tatter either baseline or successful serial drug testing) had a significantly lower actuarial risk of recurrent arrhythmic events titan did patients whose arrhythmia remained inducible . Selective use of the implantable defibrillator significantly reduced sudden death mortality and overall mortality in high risk patients (ronresponders) and also seemed to reduce the risk of sudden death in patients with noninducibte arrhythmia . Prospective studies are indicated 1) to optimize the decision of which nonrcsponders and which patients witls roninducibte arrhythntta should receive the implantable defibrillator, and 2) to assess the potential benefit of the implantable defibrillator in survivors of cardiac arrest who respond to drugs during serial electrophysiologic testing.

References L Eldar M. Soave Ml . Scheinman MM. Elwlrophysidogic testing and

follow-up of patients with aborted sudden death. J Am Coll Caudiol 2.

1987 :10:291-8 . Wilber DJ . Gams

H. Finkelstein D, et al. (fin-of-hocpilal cardiac unrest : in the prediction ¢flag-teen Outcome,

use ofelectropliysialogic testing N End J Mid 1n8S18:19-24.

3, Fogoros RN. Fiedler SB . Elm 11. The automatic implantable cardio-

vener-defibnllalor in dmgrefraetory ventricular Iarhyanhylbmias . Mn latent Mud 1987:107:635-41 . Waxman HL. Kien le MG. Sutton AE. Maeehiinski FE . lasephson ME . Clinical cbarackoth cs and long-term folbwmp in 119 survivors ofcarmncamen :: relationloinducibilityateketrophysi ek0rop"lidark Rome . Am J Curdiol 198332 :969-74. 5 . Moody F. Scbeinnan MM. Hess D5 . Sung RJ. Shen E . Shapiro W .

4 . Ray D.

Flecteophysiobgie testing in the normgcmcnt of survivors of outofhospilal cardiac anent . Am J Candiul 1983;51:85-9. h Shale ST . Miles WM. HeerrJJ. byes DP, Prystowxky EN . 4uraiwrs on cardiac arena!, prevention orrecmrewe by drum Ihenay as eradiared by eledrophysioklgie lolling or ekclrocardiognphie monitoring . Am 1 Car. 7,

dial 1986:9:113-9 . Freedman RA . Swetdlaw CD . Sodenwlm-Difwe V. Mason JW . wannoul :c significarece of arrhylhmia indonibility or nmninducibiiity al initial

ekclrophysiobgic study in carnivals of cardiac areas . Am J Cardio] 1988 :61 :578-82_ . 8.

11-din 0G . We- DW Jr. Klein GJ. Plitzber MR. Krien 1M. Anderson RW. Prevention of recount cardiac anent : role of in-alive dear¢ phvoracolegie testing . J Am Coil Cardiol 19132 :418-25.

9. 2hcmhn TA. Steinmao RT, Nanioni TA . Kehoe RF. Lang- arrhyth.

mic outcome in curvitom of venlricalar Sbdllation wish absence of mducible veotneuinrtachyeardia. Am 1 Cordial lsna:2 :1213-7. set K o tl y P . Ration SN. Vluh a keorti . euo kioy MS Or, Fnnum an C& Dawn H. Surgical camaary mvascuiarizatian in survivors of ptehaspilal cardiac rest its effect an inducible vemaeular arrhythmias and long-term

Coll Cardiol 1990 :15:167-73. survival . J It . nsandun A . Lest C . Iluentni. Cannon S . Coronary revalcatanonlion in patients with sustained ventne iar lachycardia or cardiac arrest: erects



788

FOGOROS ET AL . ELECIROPHYSIOLCGIC TESTING AITER CARDIAC ARREST ou iudacibilily and loon- foliuw-up heRoin 1691 ;17:93A .

J Am Call Cordiol

12 . Kehoe R . Toonmoso C . Zheuldn T. et a, Focmre derermlnini pro-

gremmed sunulation responses and Inng-mrm arrhylhmn outcome in wrs of vemricalar l brilla9on whh Jscbemlc hear disease . Am Hean Ja1988 ;I IfiAsA-83 . 13 . Ru1Ty R. Palleel5 ar nigh risk od suudcn cardiac dC'clJ 3n' Horowila LN,

ed . Canrnt Mmwgemem or Arrhyrhmis . Philadelphia : B .C. Decker. 1991:191-7 . 14. Impinuta:inn atautammic defibrillators. Mediocre Coverage lseues Man-

JACC Vol. 19, No . 4 arch 15 . tt,. .7i0_8 vul. Dep+rtmem of Heuhh and Humun Scrvics, Trensmiuel N . .48 . Jun. 1991, Sedlun 7-7-85. 15 . Mama 1W, Winkle RA. Acwracy of the .curricular lachycardiainduennn andy far pmdicang I. . .. arm etpcacC and inefficacy efanttarrhyth i drugs . N Engl J Mod 1980 ;303:!075-7. Ib . HorourleLN,JosephsorME .KeatorJA .lolracaniiacelectrophysiologic curdles as a method for the oprtmizarion of drug therapy in chronic venldcolermrhylhmia. Prop Card.oseco Lls 1980:23 :81-98. 17 . Fogorns RN, Elsnn Jl . Barrel CA. Atlunrtnl incidence and pattern of e oP shocks following implanradon of rh . amomatic impianwbie cardimerterdrlbrillamr. PACT 1%9:12 1465-73.