- Email: [email protected]
Electrocardiographic diagnosis of reperfusion during thrombolytic therapy in acute myocardial infarction
Electrocardiographic Dia irosis of Reperfwsion During Thrombo f ytic Therapy in Acute Myocardial Infarction Pieter A. Doevendans, MD, Anton P. Gorgels, MD, Rien van der Zee, MD, Jean Partouns, h-its W. BC, MD, and Hein J.J. Wellens, MD Continuous 1 Nead electrocardiographic monitofing was performed in 61 patients receiving thrombotytic therapy for an acute myocardial infarction. Coronary angiography within 90 minutes revealed a patent vessel (Thrombolysis in Myocardial Infarction [TIM11trial 2 or 3) in 44 patients. Early signs of reperfusion were STsegment normalization (likelihood ratio 16.0), development of terminal T-wave inversion (likelihood ratio 10.6), accelerated idioventricular rhythm (likelihood ra-
tio 6.0), and a twofold increase in ventricular premature complexes (likelihood ratio 2.5). Relief of chest pain after 60 minutes was reported by 96%. During reperfusion of the infarct-related vessel, an increase in STsegment deviation was recorded in 61% of the patients, whereas 69% had an increase in chest pain preceding the eventual decline. (Am J Cardiol 1995;75: 1206-l 2 10)
eperfusion of the infarct-related vesselduring thromR bolytic therapy can be assessedby invasive or noninvasive means. Noninvasive assessmentof repetfusion
with anisoylatedplasminogenstreptokinaseactivatorcomplex (Eminase@,Beecham,Rjiswjik, The Netherlands). Electrocardiographic monitoring: During treatment,patients were connected to a Marquette CASE 12 electrocardiograph (Marquette Electronics, Milwaukee, Wisconsin), which allowed continuous 1Zlead arrhythmia monitoring, and simultaneously to a Marquette MAC 15 to record a 1Zlead electrocardiogram approximately every 30 seconds. ST segment: Every 5 minutes the ST segment was measured60 ms after the J point by 2 of the investigators, and this scorewas usedfor determining the changes in ST segmentthat occurred. These measurementswere used for statistical analysis. In addition, an ST-segment trend was plotted using electrocardiograms recorded every 30 secondsby a MAC 15. No measurementwas obtained during an arrhythmia. A change of 25% compared with the maximal measuredST-segmentdeviation was considered significant and scored as either a decreaseor increaseof ST-segmentdeviation. For this score, 2 leads showing the most pronounced ST-segmentelevation or depressionwere selected.Terminal T-wave inversion was consideredpresentif the terminal part of the T wave changed direction from positive to negative or vice versa. Arrhythmia monitoring: The automatic arrhythmia detection mode of the Marquette CASE 12 was used. This computer algorithm can detect ventricular premature complexes,arrhythmias, and suddenchangesin heart rate. An additional review mode allowed a 12-channel, beat-to-beat analysis of the full-hour recording. For statistical analysis, only data obtained during reviewing of the complete registration by 2 investigators were used. The number of ventricular premature complexes was counted during every 5-minute interval. The number of ventricular prematurecomplexesduring the iirst 15 minutes divided by 3 was used as a baseline count. When after 15 minutes the number of ventricular premature complexesincreasedtwofold during a S-minute interval, this was considered a positive result. When the baseline count was 51, an increaseto 22 was scoredpositive. The
may make an emergency coronary angiography redundant. Noninvasive signs of reperfusion are rapid ST-segment normalization,’ relief of chest pain,2 and arrhythmias.3Peakcreatine kinase activity within 12hours after the start of thrombolysis is a well-known sign of reperfusion, but this parameteris not available during the first hours of admissi0n.l In patients receiving thrombolytic therapy we observed considerable fluctuations in STsegment deviations and chest pain during reperfusion. These changes preceded an eventual normalization of both parametersin a subset of patients. Becauseof the possible clinical and pathophysiologic implications, we prospectively studied these phenomena during the first hour of thrombolytic therapy using continuous recording of the 12-lead electrocardiogram. METHODS Patients: Based on previous studies,j we included patients who fulfilled the following criteria: admission within 4 hours after onset of chest pain; ST-segmentelevation of >2 mm in 2 extremity leads, or >3 mm in 2 precordial leads; no contraindications for thrombolytic therapy; and age ~76 years. Patients were excluded if they had continuous atria1 fibrillation or an increase in serum glutamic oxaloacetic transaminase activity <2 times the maximal normal value (35 U/L). Treatment: Conventional antianginal treatment was used as needed.Patients were treated with prourokinase intravenously (Saruplase@,Grunenthal, Aachen, Germany). A bolus injection of 20 mg was followed by an infusion of 60 mg in 1 hour. Three patients were treated From the Department of Cardiology, Academic Hospitol Maastricht, Cardiovascular Research Institute Maastricht, Universiiy of Limburg, Maastricht, The Netherlands. Manuscript received September 7, 1994; revised manuscript received and accepted March 27, 1995. Address for reprints: Anton P. Gorgels, MD, Deportment of Cardiology, Academic Hospital Maastricht, P.O. Box 5800, 6202 AZ Maastricht, The Netherlands.
1206
THE AMERICAN
JOURNAL
OF CARDIOLOGY”
VOI.
75
JUNE
15,
1995
following criteria were used: acceleratedidioventricular chest pain, and extent of coronary artery disease.Maxrhythm (AIVR)-a run of >3 ventricular beats, with a imal serum glutamic oxaloacetic transaminaseincrease rate between 60 and 120 beats/mm; nonsustained ven- occurred earlier in the reperfusion group (12.9 vs 16.9 tricular tachycardia-ventricular rhythm >120 beats/mm hours, respectively, p = 0.027). lasting ~30 seconds; atrial tachycardia->3 consecutive ectopic P waves, with a rate of >lOO beats&n; bradycardia-atrioventricular block; sinus bradycardia-50 TABLE I Characteristics of Patients (n = 61) in Relation to Status of the Infarct-Related Coronary Artery as Seen on Angiography beats/mm. ST-segmentmeasurementsand evaluation of Performed Within 90 Minutes After Start of Thrombolytic Therapy the arrhythmias were performed by 2 experienced cardiologists who were not informed about the clinical outVessel Vessel Status of the Infarct-Related Patent Occluded come of the patient. p Value Vessel IW IV Monitoring of chest pain: A 5-point scale to monitor severity of chest pain was used and adjusted in score 17 (28) 44 (72) No. of patients 13 (76) NS Men 37 (84) every 5 minutes by a trained nurse. The scoring system NS Mean age (yr) 59.7 zt 10.0 63.5 zt 7.8 used was: 0 = no chest pain; 1 = minor chest pain; 2 = 4 (24) NS slight chest pain; 3 = moderate chest pain; 4 = severe History of 9 (201 angina pectoris >l mo chest pain; and 5 = severe chest pain accompanied by Myocardial infarction 1 (6) NS 4 (9) restlessness. Infarct localization Coronary angiography: Coronary angiography was 10 (59) NS Anterior 24 (55) performed between 60 and 90 minutes after the start of Inferior 14 (32) 4 (24) NS thrombolytic therapy in all patients irrespective of the Posterior 2 (12) NS 4 (9) clinical findings. Patency of the infarct-related coronary lateral 1 (6) NS 2 (51 artery was assessedby 2 angiographers who were not Mean duration of pain before 2.0 * 0.98 2.2 * 1.05 NS informed about the electrocardiographic changes that start of therapy (hr) occurred before arrival in the cardiac catheterization labEnzyme release oratory. The classification of the Thrombolysis in MyoMean maximal SGOT (U/l) 325 * 194 408 * 186 NS Time to peak level (hr) 12.9 f 6.0 16.9 * 6.8 0.027 cardial Infarction (TIMI) trial group was used.6Patency was defined as TIM1 grade 2 or 3. Coronary angiography 1-vessel disease 22 (50) 9 (53) NS Venous blood samples: Blood sampleswere obtained 2-vessel disease 14 (32) 3 (18) NS on admission and every 8 hours thereafter until the max3-vessel disease 5 (29) NS 5 (11) imal level of serum glutamic oxaloacetic transaminase Left mainstem 0 NS 3 (7) was reached. Values are expressedas number (9) or mean * SD. Statistical analyses: Group data are expressedas mean SGOT = serumglutamic ox&acetic tronsaminase. + SD for continuous variables or as rates (percentage) for categorical variables. For continuous variables, the unpaired r test was TABLE II Changes Before Coronary Angiography and Within 90 Minutes After Start performed. This test provided the 95% of Thrombolytic Therapy confidence interval of the difference. Vessel Vessel For categoricalvariables,the chi-square Patent Occluded test or, when appropriate,Fisher’s exact (56) p Value LR 95% Cl (“4 test for 2 X 2 tables was performed. ST segment The likelihood ratio was defined as No. of patients 17 sensitivity/l - specificity. The 95% Decreased deviation 1; (95) CO.01 16.0 0.63/l .O* 1 (6) confidence interval was calculated by -0.91/-0.45 12 (71)
CORONARY
Arrhythmia No. of patients Conduction disturbance Atrial tachycardia Increased number of ventricular premature complexes AIVR Nonsustained ventricular tachycardia Ventricular fibrillation Pain No. of patients No change Increase Decrease at 60 min
38 7 (18) 5 (13) 28 (74)
14 2 (14) 1 (7) 4 (29)
co.01
16 (42)
6 (16)
1 (7) 0
2 (51 26 0
1 14) 25 (96)
1.3 1.8 2.5
-0.19/0.27 -0.12/0.26 0.15/0.75
6.0
0.07/0.65 -0.04/0.36
1 (7)
NS
0.7
-0.16/O. 12
8 4 (50) 1 113) 3 (38)
c0.5+ NS+ <0.5+
0.3 2.6
-0.76/-0.25 -0.28/0.1 0.29/0.89
NS NS
*Approximate value; Wisher’sexact test. AIVR 7 accelerated idioventricular rhythm; 95% Cl = 95 % confidence interval of the difference; iR = likelihood ratio.
ARTERY DlSEASEjREPERFUSlON
DURING
THROMBOLYSIS
IN MYOCARDIAL
INFARCTION
1207
FIGURE 1. Three-dimensional presentation of 20 sequential tial increase in ST deviation before the eventual decrease.
A
III Av2
(12-lead) e Iec tr ocardiograms
B
--L
--P
v3-L-If
FIGURE 2. Exam les of 3 different patients with acute m ocardial infarction o P the inferior, anterior, and posterior wa rI. A, during acute infarction; 6, after reperfusion, which was documented by coronary angiography afterward. Lead Ill = example of T-wave inversion in inferior wall myocardial infarction; note the decrease in ST-segment elevation; lead V, = example of terminal T-wave inversion in anterior wall infarction; lead V, = example of reciprocal T-wave inversion in posterior wall infarction.
1208
THE Ah’\ERICAN JOURNAL OF CARDIOLOGY~
\jOL. 75
recorded within
14 minutes. Note the ini-
ST-segment analysis (Table II): After 60 minutes, 95% of the patients (42 of 44) with a patent vessel had a decrease in the ST segment of 225%, compared with the maximal deviation measured in the first hour after the start of thrombolytic therapy. In 27 patients (61%j, an increase in ST-segment deviation preceded the eventual decrease (Figure 1). The development of a terminal Twave inversion was documented in 64% of patients (28 of 44) (Figure 2). In the nonreperfusion group, 1 of 17 patients had both a decrease in ST-segment deviation and terminal negative T-wave development. No significant change in ST segment was seen in 12 of 17 patients (71%) (likelihood ratio 0.03). Repetfusion arrhythmias: Nine patients (15%) had an incomplete registration or incomplete review because of technical problems. A twofold increase in the number of ventricular premature complexes after 15 minutes occurred in 28 of the 38 patients (74%) with a patent vessel versus 4 of 14 (29%) (likelihood ratio 2.5); 16 of 38 patients had AIVR versus 1 of 1-l (likelihood ratio 6.0). Transient conduction disturbances were observed in 7 patients with reperfusion, and 6 patients had an inferior wall myocardial infarction. Sinus bradycardia was seen most often (n = 4) (likelihood ratio 1.3). Pain score: A complete pain score was obtained in a subset of 34 patients (56%j, in which 26 had a patent JUNE 15, 1995
r
vessel. In the tist 12 patients no score TABLE III Reperfusion was obtained, and in 15 additional paTransient No tients the documentation was incomIncrease Increase p Value 95% Cl plete. An eventual decrease in chest Chest Pain Behavior During Reperfusion (n = 26) pain after 1 hour occurred in 25 patients (96%, likelihood ratio 2.6) (Table No. of patients 18 8 II). In the nonreperfusion group, 4 paMean age (yr) 57.6 f 10.4 64.4 * 8.5 NS -15.5/l .87 Women 3 (171 1 (121 NS -0.25/0.35 tients (50%) had no significant change Anterior myocardial infarction 10 (56) 3 (37) NS -0.22/0.61 in chest pain. 2(11) 0 NS -0.1 l/O.33 History of angina pectoris Reperfvsion: Table III summarizes >l mo data concerning ST-segment changes Duration of pain (hr) 2.1 i 0.8 2.1 * 1.64 NS -0.90/0.98 Time to SGOT peak (hr) 10.0 * 3.3 16.4 * 3.5 co.01 -9.35/-3.45 and chest pain behavior during reper146*99 Maximal SGOT level (U/L) Al3 zt 192 ~0.01 117.7/416.3 fusion. In 18 patients an increase in Coronary angiography chest pain preceded the eventual de7 139) 4 WI NS -0.52/0.3 1-vessel disease crease.In patients with an increase in 2-vessel disease 8 144) 2 (25) NS -0.26/0.6 pain during repetfusion,maximal serum 3-vessel disease 3 (17) 2 (25) NS -0.41/0.25 glutamic oxaloacetic transaminaseinST-Segment Changes During Reperfusion (n = 44) creasewas significantly higher (413 vs No. of patients 27 17 146)and registeredearlier (10.0 vs 16.4 Duration of pain (hr) 1.8 I? 0.8 2.3 zt 1.2 NS -1.11/0.01 hours). A similar trend was seenwhen Time to SGOT peak (hr) 12.4 f 4.9 15.6 f 7.9 NS -7.08/0.68 patients with an increase of ST-segMaximal SGOT level (U/L) 342*186 256*174 NS -27.4/l 99.4 ment deviation were compared with Values ore expressed as number (“i,) or mean * SD. patients with reperfusion without the Abbreviations CIS in Tables I and II. increase.These differences are not statistically significant. .~ r Positive predictive value: The posiTABLE IV Positive Predictive Value tive predictive value is high for evenPositive Negative tual ST-segmentnormalization (97%), Predictive Predictive terminal T-wave inversion developSensitivity Specificity Value Value ment (96%), AIVR (94%), and two(“Al (“4 (“4 (“4 fold increase in the number of ventriDecrease in ST-segment 95 94 97 a9 cular prematurecomplexesper 5minute deviotion (25%) intervals (87%) (Table IV). A high Terminal T-wave development 94 63 96 50 AIVR A2 93 94 37 negative predictive value is only calIncreased number of ventricular 74 71 a7 50 culated for the absenceof a 25% depremature complexes creasein ST-segmentdeviation (89%). Abbreviations
as in Table
II.
DISCUSSION To enable clinicians to select patients who may benefit from more aggressive treatment after thrombolytic therapy, early assessmentof reperfusion is obligatory. Therefore, we directed our attention to the iirst hour instead of after 24 hours.TsUse of a continuous 12-lead electrocardiogram instead of Holter registrations allows an accurate ST-segment and arrhythmia analysis.’ The main findings of the study are the high likelihood ratios for eventual normalization of the ST segment (ratio 16.0), the early development of terminal T-wave inversion (ratio 10.6), and recording of AIVR (ratio 6.0) in the reperfusion group. The absence of significant STsegmentchangesthroughout the first hour makes reperfusion highly unlikely (likelihood ratio 0.03). An important finding is the increase in ST-segment deviation (27 of 44 patients) and chest pain (18 of 26 patients) preceding normalization. The value of continuous ST-segment monitoring to assessreperfusion during thrombolytic therapy has been well established.1.2.7-9 Krucoff et al’ showed the importance of a steady state and ST-segment normalization occurring early (55 + 32 minutes) after reperfusion. According to Hohnloser et al,* the positive predictive value of a 50% reduction in ST-segmentelevation meaCORONARY
ARTERY DISEASE/REPERFUSION
sured after 90 minutes was 97%. Using a 12-lead electrocardiogram and by comparing the final electrocardiogram with the electrocardiogram showing maximal STsegmentdeviation, we found ST-segmentnormalization to be a very important indicator for reperfusion (sensitivity 95%, specificity 94%). When we used a 50% reduction as the cutoff point,“* the sensitivity decreased from 95% to 85%, without a changein specificity. In the nonreperfusion group, 70% of the patients did not have significant ST-segmentchanges.A similar study approach and comparablefindings were recently presentedby Shah et al.‘OThey found a rapid decreasein ST-segmentelevation and chest pain in all patients with TIMI grade 3 flow within 90 minutes. In 58% of the patients, this was precededby a transient decreaseand increasein ST-segment deviation. The developmentof terminal T-wave inversion during the tist hour is a less sensitive finding but a very specific sign of reperfusion (94%). The importance of arrhythmias during reperfusion has already been described in 1983 by Goldberg et al.3 They registered AIVR in 10 of 17 patients during intracoronary thrombolytic therapy at the time of recanalization. The value of AIVR as a marker of reperfusion was further established by Hohnloser et al8 and our DURING
THROMBOLYSIS
IN MYOCARDIAL
INFARCTION
1209
gro~p.~ Low overall specificity using 24-hour Holter recording was reportedby Gressin,l?Zehender,7and their associates.There was a clear increasein specificity (82%) when regarding early (~6 hours) occurrence of AIVR. Becauseof the low frequency of nonsustainedventricular tachycardia,no significant differencewas found. However, in severalpatients, we observedan increasein ventricular premature complexes followed by nonsustained ventricular tachycardia related to the supposedmoment of reperfusion. A limited value in the increased number of ventricular premature complexes (specificity 71%) was also described by Gressin et al.” Sinus bradycardia during recanalization of the vessel supplying the inferoposterior wall has already been described.3$8 Early pain resolution was mentioned by Barbash et al9 as a sign of reperfusion. Nicolau et al’ found a high positive predictive value (97.9) for suddenrelief of chest pain as a sign of reperfusion. Recently Christian et all3 reported on the value of the chest pain responseduring thrombolytic therapy as an indicator of infarct size. Eighteen of our 25 patients described a clear-cut increase in chestpain preceding a decreaseor disappearanceof chest pain. Four patients required additional opiates for this marked increase. Patients with a marked worsening in chest pain preceding subsequentrelief had a higher and earlier peak in serum glutamic oxaloacetic transaminase release. This early, high-enzyme peak and the increase in chest pain could be due to additional damagecaused by reperfusion. The increase in ST-segment elevation preceding normalization further supports this hyporhesis. Reperfusion injury induced through the formation of oxygen-derived free radicals has been established in experimental animal models.” Whether this mechanism is important in the setting of thrombolysis in acute coronary events in humans remains to be established. Another explanation of the dynamic changesoccurring in the ST segment could be differences in flow rate through the reperfused vessel.‘j In ischemic tissue, adenosine is formed from adenosine monophosphate. The suddenincreasein adenosinereleasefrom reperfusedtissue could explain the described increase in chest pain. Different studies have shown that adenosineadministration can give rise to angina-like chest pain.16J In the group of patients with reperfusion, we comparedpatients with a transient increase in ST-segment deviation with patients who did not show this sign. We found a trend toward an early, high-enzyme peak in the group with a transient increase; however, this difference was not statistically significant. Study limitations: The number of patients studied is limited. Only patients with considerable ST-segmentdeviation were included. Our data are relevant for similar patient groups, but may not be applicable to all patients with myocardial infarction. The patients were connected to 2 electrocardiographs and required much attention and preparation. Becauseof the complicated setup in frequently very ill patients, it was not possible to collect complete data. Another limitation of this and many other
1210
THE AMERICAN JOURNAL OF CARDIOLOGY’
VOL. 75
studiesrelatedto this subjectis the useof coronary angiography as the gold standardfor assessingreperfusion. The dynamic changesin the ST segmentsand changesin pain experienced by the patient cannot be related to the exact moment of reperfusion. In contrast to other studies, ure therefore performed coronary angiography as soon as possible after the effect of thrombolytic therapy.Because of the short period of ST-segment recording (the first hour), we are not informed about reocclusion. As shown by Krucoff et a1,18reocclusion may be asymptomaticand can be detected by continuous 12-lead electrocardiographic monitoring. 1. Kmcoff MW, Green CE, Satler LF, Miller FC. Pallas RS, Kent KM, Del Negro AA, Peale DL. Fletcher RD. Rackley CE. Noninvasive detection of coronary artery patency using continuous ST-segment monitoring. 4m JCmdiol 1986;57:916-922. 2. Nicolau JC, Loga AM, Garzon SAC, Jacob SLC, Machado NCS, Bellini AJ. Greco OT, Marques LAF, Bmile DM. Clinical and laboratory signs of repetfusion: are they reliable? brr J Cardiol 1989:25:313-320. 3. Goldberg S, Greenspan AJ. Urban PL, Muza B. Beger B, Walinsky P, Maroko PR. Reperfusion arrhythmia: a marker of restoration of antegrade flow during intracoronary thrombolysis for acute myocardial infarction. Am Henrff 1983; 105:2&32. 4. Van der Laarse A. Vetmeer F, Hermens WT, Willems GM, De Neef K. Simoons ML, Sermys PW, Res I. Verheugt FWA. Krauss XH, Blr F, de Zwaan C. Lubsen I. Effects of early inuacoronary streptokinase on infarct size estimated from cumulative enzyme &ease and enzyme release rare: a randomized trial of 533 patients with acute myocardial infarction. Am Heart J 1986;112:672+83. 5. Bjir FW. Vermeer F, de Zwaan C, Ramentol M, Braat S, Simoons ML. Hermens WT. van der Lawse A, Verheugt FWA. Krauss XH. Wellens HJJ. Value of the admission electrocardiogram in predicting outcome of thrombolytic therapy in acute myocardial infarction. .4m J Cardiol 1987:59:&13. 6. Chesebro JH, Knatterud G. Roberts R, Borer J, Cohen LS, Dale” I. Dodge HT, Francis CK, for the TIM1 study group. mombolysis in Myocardial Infarction (TIMI) trial. phase I: a cqmparison between intravenous tissue plarminogen activator and intravenous streptokinase. Circzrhrkm 1987;76:142-154. 7. Zehender M, Utzolino S, Futtwangler A, Kasper W, Meinertz T, Just H. Time course and intemelation of repetfuion-induced ST changes and ventricular arrhythmias in acute myocardial infarction. .4,x J Cardiol 1991;68:1138-1142. 8. Hohnloser SH, Zabel M. Kasper W, Meinertz T, Just H. Assessment of coronary artery patency after thrombolytic therapy: accurate prediction utilizing the combined analysis of three noninvasive markers. J Am Coil Cardiol 1991;18:44-!9. 9. Barbash GI. Roth A, Hod H. Miller HI, Rath S, Ha-Zahav Y, Modan M, Seligsohn U, Battler A, Kaplinsky E, Rabinowitz E, Laniado S. Rapid resolution of ST elevation and prediction of clinical outcome in patients undergoing thrombolysis with alteplase (recombinant tissue-type plasminogen activator): results of the Israeli Study of Earl! Intervention in Myocardial Infarction. Br Heart J 1990;64:241-247. 10. Shah PK. Cercek B, Lew AS, Ganz W. Angiographic validation of bedside markers of reperfusion. J Am Co// Cardiol 1993;21:5561. 11. Gorgels A. Vos M. Letsch I. Verschuren F, Blr F. Janssen J, Wellens HJJ. Usefulness of [tie accelerated idioventricular rhythm as a marker for myocardial necrosis and reperfusion during thrombolytic therapy in acute myocardial infarction. Am J Cardiof 1988;61.1,31-235. 12. Gressin V. Louvard Y, Pezrano M, Lardoux H. Holter recording of ventricular arrhythmias during intravenous thrombolysis for acute myocardial infarction. ANI J Cardiol 1992;69:152-159. 13. Christian TF, Gibbons RJ. Hopfenspirger MR, Gersh BJ. Severity and response of chest pain during thrombolytic therapy: a useful indicator of myocardial salvage and infarct size. J.4m Co// CardioI 1993;2?:1311-1316. 14. McCord I. Oxygen-derived free radicals in postischemic tissue injury N Rtg/ JMrd 1985;312:159-163. 15. Sb;rensen SG. Hackworthy RA. Fitzpatrick PG. Menlove R. Andersen IL. Variability of thmmbolytic coronary reperfosion: an angiographic study of streptokinax and anistreplase. C/in Cardiol 1990;13: 15-19. 16. Sylven C. Beermann B, Jonzon B, Brandt R. Anginri pectoris-like chest pain provoked by intravenous adenosine in healthy volunteers. Br Med J 1986:293: 227-230. 17. Crea F, Pupita G, Galassi A. El-Tamini H, Davies G, Maseri A. Role of adenosine in pathogenesis of angina1 pain. Cirruhiorr 199O;XI: l&172. 18. Kracoff h0V. Wagner NB, Pope JE, Mortara DM, Jackson YR, Bottner RK, Wagner GS, Kent KM. The portable programmable microprocessor-driven realtime L-lead electrocardiographic monitor: a preliminary report of a new device for the noninvasive detection of successful repafusion or silent coronary reocclusion. Am J Cardiol 1990;65: 143-148.
jUNE 15, 1995
tio 6.0), and a twofold increase in ventricular premature complexes (likelihood ratio 2.5). Relief of chest pain after 60 minutes was reported by 96%. During reperfusion of the infarct-related vessel, an increase in STsegment deviation was recorded in 61% of the patients, whereas 69% had an increase in chest pain preceding the eventual decline. (Am J Cardiol 1995;75: 1206-l 2 10)
eperfusion of the infarct-related vesselduring thromR bolytic therapy can be assessedby invasive or noninvasive means. Noninvasive assessmentof repetfusion
with anisoylatedplasminogenstreptokinaseactivatorcomplex (Eminase@,Beecham,Rjiswjik, The Netherlands). Electrocardiographic monitoring: During treatment,patients were connected to a Marquette CASE 12 electrocardiograph (Marquette Electronics, Milwaukee, Wisconsin), which allowed continuous 1Zlead arrhythmia monitoring, and simultaneously to a Marquette MAC 15 to record a 1Zlead electrocardiogram approximately every 30 seconds. ST segment: Every 5 minutes the ST segment was measured60 ms after the J point by 2 of the investigators, and this scorewas usedfor determining the changes in ST segmentthat occurred. These measurementswere used for statistical analysis. In addition, an ST-segment trend was plotted using electrocardiograms recorded every 30 secondsby a MAC 15. No measurementwas obtained during an arrhythmia. A change of 25% compared with the maximal measuredST-segmentdeviation was considered significant and scored as either a decreaseor increaseof ST-segmentdeviation. For this score, 2 leads showing the most pronounced ST-segmentelevation or depressionwere selected.Terminal T-wave inversion was consideredpresentif the terminal part of the T wave changed direction from positive to negative or vice versa. Arrhythmia monitoring: The automatic arrhythmia detection mode of the Marquette CASE 12 was used. This computer algorithm can detect ventricular premature complexes,arrhythmias, and suddenchangesin heart rate. An additional review mode allowed a 12-channel, beat-to-beat analysis of the full-hour recording. For statistical analysis, only data obtained during reviewing of the complete registration by 2 investigators were used. The number of ventricular premature complexes was counted during every 5-minute interval. The number of ventricular prematurecomplexesduring the iirst 15 minutes divided by 3 was used as a baseline count. When after 15 minutes the number of ventricular premature complexesincreasedtwofold during a S-minute interval, this was considered a positive result. When the baseline count was 51, an increaseto 22 was scoredpositive. The
may make an emergency coronary angiography redundant. Noninvasive signs of reperfusion are rapid ST-segment normalization,’ relief of chest pain,2 and arrhythmias.3Peakcreatine kinase activity within 12hours after the start of thrombolysis is a well-known sign of reperfusion, but this parameteris not available during the first hours of admissi0n.l In patients receiving thrombolytic therapy we observed considerable fluctuations in STsegment deviations and chest pain during reperfusion. These changes preceded an eventual normalization of both parametersin a subset of patients. Becauseof the possible clinical and pathophysiologic implications, we prospectively studied these phenomena during the first hour of thrombolytic therapy using continuous recording of the 12-lead electrocardiogram. METHODS Patients: Based on previous studies,j we included patients who fulfilled the following criteria: admission within 4 hours after onset of chest pain; ST-segmentelevation of >2 mm in 2 extremity leads, or >3 mm in 2 precordial leads; no contraindications for thrombolytic therapy; and age ~76 years. Patients were excluded if they had continuous atria1 fibrillation or an increase in serum glutamic oxaloacetic transaminase activity <2 times the maximal normal value (35 U/L). Treatment: Conventional antianginal treatment was used as needed.Patients were treated with prourokinase intravenously (Saruplase@,Grunenthal, Aachen, Germany). A bolus injection of 20 mg was followed by an infusion of 60 mg in 1 hour. Three patients were treated From the Department of Cardiology, Academic Hospitol Maastricht, Cardiovascular Research Institute Maastricht, Universiiy of Limburg, Maastricht, The Netherlands. Manuscript received September 7, 1994; revised manuscript received and accepted March 27, 1995. Address for reprints: Anton P. Gorgels, MD, Deportment of Cardiology, Academic Hospital Maastricht, P.O. Box 5800, 6202 AZ Maastricht, The Netherlands.
1206
THE AMERICAN
JOURNAL
OF CARDIOLOGY”
VOI.
75
JUNE
15,
1995
following criteria were used: acceleratedidioventricular chest pain, and extent of coronary artery disease.Maxrhythm (AIVR)-a run of >3 ventricular beats, with a imal serum glutamic oxaloacetic transaminaseincrease rate between 60 and 120 beats/mm; nonsustained ven- occurred earlier in the reperfusion group (12.9 vs 16.9 tricular tachycardia-ventricular rhythm >120 beats/mm hours, respectively, p = 0.027). lasting ~30 seconds; atrial tachycardia->3 consecutive ectopic P waves, with a rate of >lOO beats&n; bradycardia-atrioventricular block; sinus bradycardia-50 TABLE I Characteristics of Patients (n = 61) in Relation to Status of the Infarct-Related Coronary Artery as Seen on Angiography beats/mm. ST-segmentmeasurementsand evaluation of Performed Within 90 Minutes After Start of Thrombolytic Therapy the arrhythmias were performed by 2 experienced cardiologists who were not informed about the clinical outVessel Vessel Status of the Infarct-Related Patent Occluded come of the patient. p Value Vessel IW IV Monitoring of chest pain: A 5-point scale to monitor severity of chest pain was used and adjusted in score 17 (28) 44 (72) No. of patients 13 (76) NS Men 37 (84) every 5 minutes by a trained nurse. The scoring system NS Mean age (yr) 59.7 zt 10.0 63.5 zt 7.8 used was: 0 = no chest pain; 1 = minor chest pain; 2 = 4 (24) NS slight chest pain; 3 = moderate chest pain; 4 = severe History of 9 (201 angina pectoris >l mo chest pain; and 5 = severe chest pain accompanied by Myocardial infarction 1 (6) NS 4 (9) restlessness. Infarct localization Coronary angiography: Coronary angiography was 10 (59) NS Anterior 24 (55) performed between 60 and 90 minutes after the start of Inferior 14 (32) 4 (24) NS thrombolytic therapy in all patients irrespective of the Posterior 2 (12) NS 4 (9) clinical findings. Patency of the infarct-related coronary lateral 1 (6) NS 2 (51 artery was assessedby 2 angiographers who were not Mean duration of pain before 2.0 * 0.98 2.2 * 1.05 NS informed about the electrocardiographic changes that start of therapy (hr) occurred before arrival in the cardiac catheterization labEnzyme release oratory. The classification of the Thrombolysis in MyoMean maximal SGOT (U/l) 325 * 194 408 * 186 NS Time to peak level (hr) 12.9 f 6.0 16.9 * 6.8 0.027 cardial Infarction (TIMI) trial group was used.6Patency was defined as TIM1 grade 2 or 3. Coronary angiography 1-vessel disease 22 (50) 9 (53) NS Venous blood samples: Blood sampleswere obtained 2-vessel disease 14 (32) 3 (18) NS on admission and every 8 hours thereafter until the max3-vessel disease 5 (29) NS 5 (11) imal level of serum glutamic oxaloacetic transaminase Left mainstem 0 NS 3 (7) was reached. Values are expressedas number (9) or mean * SD. Statistical analyses: Group data are expressedas mean SGOT = serumglutamic ox&acetic tronsaminase. + SD for continuous variables or as rates (percentage) for categorical variables. For continuous variables, the unpaired r test was TABLE II Changes Before Coronary Angiography and Within 90 Minutes After Start performed. This test provided the 95% of Thrombolytic Therapy confidence interval of the difference. Vessel Vessel For categoricalvariables,the chi-square Patent Occluded test or, when appropriate,Fisher’s exact (56) p Value LR 95% Cl (“4 test for 2 X 2 tables was performed. ST segment The likelihood ratio was defined as No. of patients 17 sensitivity/l - specificity. The 95% Decreased deviation 1; (95) CO.01 16.0 0.63/l .O* 1 (6) confidence interval was calculated by -0.91/-0.45 12 (71)
CORONARY
Arrhythmia No. of patients Conduction disturbance Atrial tachycardia Increased number of ventricular premature complexes AIVR Nonsustained ventricular tachycardia Ventricular fibrillation Pain No. of patients No change Increase Decrease at 60 min
38 7 (18) 5 (13) 28 (74)
14 2 (14) 1 (7) 4 (29)
co.01
16 (42)
6 (16)
1 (7) 0
2 (51 26 0
1 14) 25 (96)
1.3 1.8 2.5
-0.19/0.27 -0.12/0.26 0.15/0.75
6.0
0.07/0.65 -0.04/0.36
1 (7)
NS
0.7
-0.16/O. 12
8 4 (50) 1 113) 3 (38)
c0.5+ NS+ <0.5+
0.3 2.6
-0.76/-0.25 -0.28/0.1 0.29/0.89
NS NS
*Approximate value; Wisher’sexact test. AIVR 7 accelerated idioventricular rhythm; 95% Cl = 95 % confidence interval of the difference; iR = likelihood ratio.
ARTERY DlSEASEjREPERFUSlON
DURING
THROMBOLYSIS
IN MYOCARDIAL
INFARCTION
1207
FIGURE 1. Three-dimensional presentation of 20 sequential tial increase in ST deviation before the eventual decrease.
A
III Av2
(12-lead) e Iec tr ocardiograms
B
--L
--P
v3-L-If
FIGURE 2. Exam les of 3 different patients with acute m ocardial infarction o P the inferior, anterior, and posterior wa rI. A, during acute infarction; 6, after reperfusion, which was documented by coronary angiography afterward. Lead Ill = example of T-wave inversion in inferior wall myocardial infarction; note the decrease in ST-segment elevation; lead V, = example of terminal T-wave inversion in anterior wall infarction; lead V, = example of reciprocal T-wave inversion in posterior wall infarction.
1208
THE Ah’\ERICAN JOURNAL OF CARDIOLOGY~
\jOL. 75
recorded within
14 minutes. Note the ini-
ST-segment analysis (Table II): After 60 minutes, 95% of the patients (42 of 44) with a patent vessel had a decrease in the ST segment of 225%, compared with the maximal deviation measured in the first hour after the start of thrombolytic therapy. In 27 patients (61%j, an increase in ST-segment deviation preceded the eventual decrease (Figure 1). The development of a terminal Twave inversion was documented in 64% of patients (28 of 44) (Figure 2). In the nonreperfusion group, 1 of 17 patients had both a decrease in ST-segment deviation and terminal negative T-wave development. No significant change in ST segment was seen in 12 of 17 patients (71%) (likelihood ratio 0.03). Repetfusion arrhythmias: Nine patients (15%) had an incomplete registration or incomplete review because of technical problems. A twofold increase in the number of ventricular premature complexes after 15 minutes occurred in 28 of the 38 patients (74%) with a patent vessel versus 4 of 14 (29%) (likelihood ratio 2.5); 16 of 38 patients had AIVR versus 1 of 1-l (likelihood ratio 6.0). Transient conduction disturbances were observed in 7 patients with reperfusion, and 6 patients had an inferior wall myocardial infarction. Sinus bradycardia was seen most often (n = 4) (likelihood ratio 1.3). Pain score: A complete pain score was obtained in a subset of 34 patients (56%j, in which 26 had a patent JUNE 15, 1995
r
vessel. In the tist 12 patients no score TABLE III Reperfusion was obtained, and in 15 additional paTransient No tients the documentation was incomIncrease Increase p Value 95% Cl plete. An eventual decrease in chest Chest Pain Behavior During Reperfusion (n = 26) pain after 1 hour occurred in 25 patients (96%, likelihood ratio 2.6) (Table No. of patients 18 8 II). In the nonreperfusion group, 4 paMean age (yr) 57.6 f 10.4 64.4 * 8.5 NS -15.5/l .87 Women 3 (171 1 (121 NS -0.25/0.35 tients (50%) had no significant change Anterior myocardial infarction 10 (56) 3 (37) NS -0.22/0.61 in chest pain. 2(11) 0 NS -0.1 l/O.33 History of angina pectoris Reperfvsion: Table III summarizes >l mo data concerning ST-segment changes Duration of pain (hr) 2.1 i 0.8 2.1 * 1.64 NS -0.90/0.98 Time to SGOT peak (hr) 10.0 * 3.3 16.4 * 3.5 co.01 -9.35/-3.45 and chest pain behavior during reper146*99 Maximal SGOT level (U/L) Al3 zt 192 ~0.01 117.7/416.3 fusion. In 18 patients an increase in Coronary angiography chest pain preceded the eventual de7 139) 4 WI NS -0.52/0.3 1-vessel disease crease.In patients with an increase in 2-vessel disease 8 144) 2 (25) NS -0.26/0.6 pain during repetfusion,maximal serum 3-vessel disease 3 (17) 2 (25) NS -0.41/0.25 glutamic oxaloacetic transaminaseinST-Segment Changes During Reperfusion (n = 44) creasewas significantly higher (413 vs No. of patients 27 17 146)and registeredearlier (10.0 vs 16.4 Duration of pain (hr) 1.8 I? 0.8 2.3 zt 1.2 NS -1.11/0.01 hours). A similar trend was seenwhen Time to SGOT peak (hr) 12.4 f 4.9 15.6 f 7.9 NS -7.08/0.68 patients with an increase of ST-segMaximal SGOT level (U/L) 342*186 256*174 NS -27.4/l 99.4 ment deviation were compared with Values ore expressed as number (“i,) or mean * SD. patients with reperfusion without the Abbreviations CIS in Tables I and II. increase.These differences are not statistically significant. .~ r Positive predictive value: The posiTABLE IV Positive Predictive Value tive predictive value is high for evenPositive Negative tual ST-segmentnormalization (97%), Predictive Predictive terminal T-wave inversion developSensitivity Specificity Value Value ment (96%), AIVR (94%), and two(“Al (“4 (“4 (“4 fold increase in the number of ventriDecrease in ST-segment 95 94 97 a9 cular prematurecomplexesper 5minute deviotion (25%) intervals (87%) (Table IV). A high Terminal T-wave development 94 63 96 50 AIVR A2 93 94 37 negative predictive value is only calIncreased number of ventricular 74 71 a7 50 culated for the absenceof a 25% depremature complexes creasein ST-segmentdeviation (89%). Abbreviations
as in Table
II.
DISCUSSION To enable clinicians to select patients who may benefit from more aggressive treatment after thrombolytic therapy, early assessmentof reperfusion is obligatory. Therefore, we directed our attention to the iirst hour instead of after 24 hours.TsUse of a continuous 12-lead electrocardiogram instead of Holter registrations allows an accurate ST-segment and arrhythmia analysis.’ The main findings of the study are the high likelihood ratios for eventual normalization of the ST segment (ratio 16.0), the early development of terminal T-wave inversion (ratio 10.6), and recording of AIVR (ratio 6.0) in the reperfusion group. The absence of significant STsegmentchangesthroughout the first hour makes reperfusion highly unlikely (likelihood ratio 0.03). An important finding is the increase in ST-segment deviation (27 of 44 patients) and chest pain (18 of 26 patients) preceding normalization. The value of continuous ST-segment monitoring to assessreperfusion during thrombolytic therapy has been well established.1.2.7-9 Krucoff et al’ showed the importance of a steady state and ST-segment normalization occurring early (55 + 32 minutes) after reperfusion. According to Hohnloser et al,* the positive predictive value of a 50% reduction in ST-segmentelevation meaCORONARY
ARTERY DISEASE/REPERFUSION
sured after 90 minutes was 97%. Using a 12-lead electrocardiogram and by comparing the final electrocardiogram with the electrocardiogram showing maximal STsegmentdeviation, we found ST-segmentnormalization to be a very important indicator for reperfusion (sensitivity 95%, specificity 94%). When we used a 50% reduction as the cutoff point,“* the sensitivity decreased from 95% to 85%, without a changein specificity. In the nonreperfusion group, 70% of the patients did not have significant ST-segmentchanges.A similar study approach and comparablefindings were recently presentedby Shah et al.‘OThey found a rapid decreasein ST-segmentelevation and chest pain in all patients with TIMI grade 3 flow within 90 minutes. In 58% of the patients, this was precededby a transient decreaseand increasein ST-segment deviation. The developmentof terminal T-wave inversion during the tist hour is a less sensitive finding but a very specific sign of reperfusion (94%). The importance of arrhythmias during reperfusion has already been described in 1983 by Goldberg et al.3 They registered AIVR in 10 of 17 patients during intracoronary thrombolytic therapy at the time of recanalization. The value of AIVR as a marker of reperfusion was further established by Hohnloser et al8 and our DURING
THROMBOLYSIS
IN MYOCARDIAL
INFARCTION
1209
gro~p.~ Low overall specificity using 24-hour Holter recording was reportedby Gressin,l?Zehender,7and their associates.There was a clear increasein specificity (82%) when regarding early (~6 hours) occurrence of AIVR. Becauseof the low frequency of nonsustainedventricular tachycardia,no significant differencewas found. However, in severalpatients, we observedan increasein ventricular premature complexes followed by nonsustained ventricular tachycardia related to the supposedmoment of reperfusion. A limited value in the increased number of ventricular premature complexes (specificity 71%) was also described by Gressin et al.” Sinus bradycardia during recanalization of the vessel supplying the inferoposterior wall has already been described.3$8 Early pain resolution was mentioned by Barbash et al9 as a sign of reperfusion. Nicolau et al’ found a high positive predictive value (97.9) for suddenrelief of chest pain as a sign of reperfusion. Recently Christian et all3 reported on the value of the chest pain responseduring thrombolytic therapy as an indicator of infarct size. Eighteen of our 25 patients described a clear-cut increase in chestpain preceding a decreaseor disappearanceof chest pain. Four patients required additional opiates for this marked increase. Patients with a marked worsening in chest pain preceding subsequentrelief had a higher and earlier peak in serum glutamic oxaloacetic transaminase release. This early, high-enzyme peak and the increase in chest pain could be due to additional damagecaused by reperfusion. The increase in ST-segment elevation preceding normalization further supports this hyporhesis. Reperfusion injury induced through the formation of oxygen-derived free radicals has been established in experimental animal models.” Whether this mechanism is important in the setting of thrombolysis in acute coronary events in humans remains to be established. Another explanation of the dynamic changesoccurring in the ST segment could be differences in flow rate through the reperfused vessel.‘j In ischemic tissue, adenosine is formed from adenosine monophosphate. The suddenincreasein adenosinereleasefrom reperfusedtissue could explain the described increase in chest pain. Different studies have shown that adenosineadministration can give rise to angina-like chest pain.16J In the group of patients with reperfusion, we comparedpatients with a transient increase in ST-segment deviation with patients who did not show this sign. We found a trend toward an early, high-enzyme peak in the group with a transient increase; however, this difference was not statistically significant. Study limitations: The number of patients studied is limited. Only patients with considerable ST-segmentdeviation were included. Our data are relevant for similar patient groups, but may not be applicable to all patients with myocardial infarction. The patients were connected to 2 electrocardiographs and required much attention and preparation. Becauseof the complicated setup in frequently very ill patients, it was not possible to collect complete data. Another limitation of this and many other
1210
THE AMERICAN JOURNAL OF CARDIOLOGY’
VOL. 75
studiesrelatedto this subjectis the useof coronary angiography as the gold standardfor assessingreperfusion. The dynamic changesin the ST segmentsand changesin pain experienced by the patient cannot be related to the exact moment of reperfusion. In contrast to other studies, ure therefore performed coronary angiography as soon as possible after the effect of thrombolytic therapy.Because of the short period of ST-segment recording (the first hour), we are not informed about reocclusion. As shown by Krucoff et a1,18reocclusion may be asymptomaticand can be detected by continuous 12-lead electrocardiographic monitoring. 1. Kmcoff MW, Green CE, Satler LF, Miller FC. Pallas RS, Kent KM, Del Negro AA, Peale DL. Fletcher RD. Rackley CE. Noninvasive detection of coronary artery patency using continuous ST-segment monitoring. 4m JCmdiol 1986;57:916-922. 2. Nicolau JC, Loga AM, Garzon SAC, Jacob SLC, Machado NCS, Bellini AJ. Greco OT, Marques LAF, Bmile DM. Clinical and laboratory signs of repetfusion: are they reliable? brr J Cardiol 1989:25:313-320. 3. Goldberg S, Greenspan AJ. Urban PL, Muza B. Beger B, Walinsky P, Maroko PR. Reperfusion arrhythmia: a marker of restoration of antegrade flow during intracoronary thrombolysis for acute myocardial infarction. Am Henrff 1983; 105:2&32. 4. Van der Laarse A. Vetmeer F, Hermens WT, Willems GM, De Neef K. Simoons ML, Sermys PW, Res I. Verheugt FWA. Krauss XH, Blr F, de Zwaan C. Lubsen I. Effects of early inuacoronary streptokinase on infarct size estimated from cumulative enzyme &ease and enzyme release rare: a randomized trial of 533 patients with acute myocardial infarction. Am Heart J 1986;112:672+83. 5. Bjir FW. Vermeer F, de Zwaan C, Ramentol M, Braat S, Simoons ML. Hermens WT. van der Lawse A, Verheugt FWA. Krauss XH. Wellens HJJ. Value of the admission electrocardiogram in predicting outcome of thrombolytic therapy in acute myocardial infarction. .4m J Cardiol 1987:59:&13. 6. Chesebro JH, Knatterud G. Roberts R, Borer J, Cohen LS, Dale” I. Dodge HT, Francis CK, for the TIM1 study group. mombolysis in Myocardial Infarction (TIMI) trial. phase I: a cqmparison between intravenous tissue plarminogen activator and intravenous streptokinase. Circzrhrkm 1987;76:142-154. 7. Zehender M, Utzolino S, Futtwangler A, Kasper W, Meinertz T, Just H. Time course and intemelation of repetfuion-induced ST changes and ventricular arrhythmias in acute myocardial infarction. .4,x J Cardiol 1991;68:1138-1142. 8. Hohnloser SH, Zabel M. Kasper W, Meinertz T, Just H. Assessment of coronary artery patency after thrombolytic therapy: accurate prediction utilizing the combined analysis of three noninvasive markers. J Am Coil Cardiol 1991;18:44-!9. 9. Barbash GI. Roth A, Hod H. Miller HI, Rath S, Ha-Zahav Y, Modan M, Seligsohn U, Battler A, Kaplinsky E, Rabinowitz E, Laniado S. Rapid resolution of ST elevation and prediction of clinical outcome in patients undergoing thrombolysis with alteplase (recombinant tissue-type plasminogen activator): results of the Israeli Study of Earl! Intervention in Myocardial Infarction. Br Heart J 1990;64:241-247. 10. Shah PK. Cercek B, Lew AS, Ganz W. Angiographic validation of bedside markers of reperfusion. J Am Co// Cardiol 1993;21:5561. 11. Gorgels A. Vos M. Letsch I. Verschuren F, Blr F. Janssen J, Wellens HJJ. Usefulness of [tie accelerated idioventricular rhythm as a marker for myocardial necrosis and reperfusion during thrombolytic therapy in acute myocardial infarction. Am J Cardiof 1988;61.1,31-235. 12. Gressin V. Louvard Y, Pezrano M, Lardoux H. Holter recording of ventricular arrhythmias during intravenous thrombolysis for acute myocardial infarction. ANI J Cardiol 1992;69:152-159. 13. Christian TF, Gibbons RJ. Hopfenspirger MR, Gersh BJ. Severity and response of chest pain during thrombolytic therapy: a useful indicator of myocardial salvage and infarct size. J.4m Co// CardioI 1993;2?:1311-1316. 14. McCord I. Oxygen-derived free radicals in postischemic tissue injury N Rtg/ JMrd 1985;312:159-163. 15. Sb;rensen SG. Hackworthy RA. Fitzpatrick PG. Menlove R. Andersen IL. Variability of thmmbolytic coronary reperfosion: an angiographic study of streptokinax and anistreplase. C/in Cardiol 1990;13: 15-19. 16. Sylven C. Beermann B, Jonzon B, Brandt R. Anginri pectoris-like chest pain provoked by intravenous adenosine in healthy volunteers. Br Med J 1986:293: 227-230. 17. Crea F, Pupita G, Galassi A. El-Tamini H, Davies G, Maseri A. Role of adenosine in pathogenesis of angina1 pain. Cirruhiorr 199O;XI: l&172. 18. Kracoff h0V. Wagner NB, Pope JE, Mortara DM, Jackson YR, Bottner RK, Wagner GS, Kent KM. The portable programmable microprocessor-driven realtime L-lead electrocardiographic monitor: a preliminary report of a new device for the noninvasive detection of successful repafusion or silent coronary reocclusion. Am J Cardiol 1990;65: 143-148.
jUNE 15, 1995