Author + information
- Received September 29, 1998
- Revision received July 27, 1999
- Accepted September 13, 1999
- Published online January 1, 2000.
- Peter J Gheeraert, MD∗,* (, )
- José P.S Henriques, MD†,
- Marc L De Buyzere, MSc∗,
- Joeri Voet, MD∗,
- Pol Calle, MD, PhD†,
- Yves Taeymans, MD, PhD∗ and
- Felix Zijlstra, MD, PhD‡
- ↵*Reprint requests and correspondence: Dr. Peter Gheeraert, Department of Cardiology, University Hospital, De Pintelaan 185, B-9000 Ghent, Belgium
The study intended to compare the acute coronary anatomy of patients with acute myocardial infarction (AMI) complicated by out-of-hospital ventricular fibrillation (VF) versus patients with AMI without this complication.
More than half of the deaths associated with AMI occur out of the hospital and within 1 h of symptom onset. The angiographic determinants of out-of-hospital VF in patients with AMI have not been investigated in detail.
Acute coronary angiographic findings of 72 consecutive patients with AMI complicated by out-of-hospital VF were compared with findings from 144 matched patients with AMI without this complication.
Patients with an acute occlusion of the left anterior descending coronary artery (LAD) or left circumflex coronary artery (LCx) had a higher risk for out-of-hospital VF compared with patients with an acute occlusion of the right coronary artery (RCA) (odds ratio and 95% confidence interval, respectively, 4.82 [2.35 to 9.92] and 4.92 [2.34 to 10.39]). With regard to extent of coronary artery disease (CAD), the location of the culprit lesion in the coronary arteries (proximal vs. mid or distal), the flow in the infarct related artery (IRA), the presence or absence of collaterals to the IRA and chronic occlusions, there were no differences between the two groups.
Acute myocardial infarction due to occlusion in the left coronary artery (LCA) is associated with greater risk for out-of-hospital VF compared to the RCA. The location of occlusion within LCA (LAD, LCx, proximal or distal), amount of myocardium at risk for necrosis and extent of CAD are not related to out-of-hospital VF.
More than half of the deaths associated with acute myocardial infarction (AMI) occur out of the hospital and within 1 h of symptom onset (1,2). There is no doubt that early ventricular fibrillation (VF) is the major lethal complication of AMI (3,4). However, to our knowledge, the determinants of early VF or out-of-hospital VF in patients with AMI have not been investigated in depth (4–6). Nevertheless, a number of suggestions have been derived from coronary angiography in the setting of thrombolysis studies, at a late phase of AMI or from postmortem examinations. Unfortunately, previous studies reached conflicting conclusions for all the examined determinants: location of the infarction (7–11), identification of the infarct related artery (IRA) (7,10,12–14), infarct size (8–10)and extent of coronary artery disease (CAD) (7,10,13). Moreover, other determinants such as flow grade in the IRA before reperfusion therapy, collaterals to the IRA, chronic coronary occlusions and proximal or distal location of culprit lesions have not been studied yet. To our knowledge, no previous studies have compared coronary anatomy and coronary flow of patients with acute MI complicated by out-of-hospital VF with a control group of similar patients without out-of-hospital VF. Although coronary angioplasty is being increasingly used as the primary treatment for patients with AMI (15), acute coronary angiography is a unique method to study coronary anatomic and hemodynamic variables during AMI and before reperfusion therapy. Not only is it an accepted treatment option, but it also provides accurate information on infarct location, culprit lesion, presence of chronic occlusions, collateral circulation and initial flow grade in the IRA. It also provides information on the amount of myocardium at risk for necrosis (region at risk) and on the extent of CAD. The latter is evaluated by the number of diseased vessels or by the jeopardy score (JS, the amount of myocardium in “jeopardy”). Therefore the aim of the study is a comparison of coronary anatomy before reperfusion therapy of patients with acute coronary occlusion complicated by out-of-hospital VF versus patients with acute coronary occlusion without out-of-hospital VF. More specifically we studied the precise location of the culprit lesion, the extent of CAD (number of diseased vessels and JS), the region at risk, coronary flow grades in the IRA, presence or absence of collaterals and chronic occlusions in a noninfarct related artery.
Patients and study protocol
In De Weezenlanden Hospital Zwolle and in the University Hospital Gent, primary angioplasty is the treatment of choice for complicated and uncomplicated acute MI as long as patients present within 6 h after symptom onset, have electrocardiographic (ECG) criteria for AMI (presence of ST segment elevation of >0.1 mV in at least two adjacent leads of a 12-lead ECG or presence of a presumed new left bundle branch block) and (for AMI cases complicated by out-of-hospital VF) have a reasonable chance to survive without major neurologic sequelae. Between January 1995 and December 1998, 75 survivors of out-of-hospital VF fulfilled these criteria and were candidates for primary angioplasty. Three patients with left bundle branch block had a normal coronary angiogram so that the suspected diagnosis of an acute coronary occlusion and AMI could not be confirmed. These patients were excluded for analysis. None of the patients received fibrinolytics. In the same period, a group of 1,000 patients with AMI without out-of-hospital VF fulfilled the same inclusion criteria and were treated with primary angioplasty. To form a matched control group, each patient with out-of-hospital VF (n = 72) was matched with two patients from the latter group. Patients were matched for age (± 5 years), gender, admission hospital (Gent or Zwolle) and primary or secondary admission (referred by a local community hospital after diagnosis of AMI).
All patients were treated intravenously with at least 10,000 U of heparin and at least 250 mg of aspirin and underwent immediate coronary angiography. Right and left coronary angiograms were obtained in multiple projections aiming to start with the non-IRA (based on ECG) in order to visualize collateral circulation to the IRA. Ventriculography was not performed in the acute phase. Angiography of the IRA was repeated after intracoronary administration of nitrates.
Two experienced cardiologists who were blinded to angiographic data interpreted all ECGs recorded upon admission. The principal angiographic and clinical data were prospectively entered in a database. Coronary stenoses were graded by visually estimating the reduction in luminal diameter (0% to 49%, 50% to 74%, 75% to 94%, 95% to 99% and 100%) in the angiographic projection in which the stenosis appeared most severe. Coronary lesions resulting in 75% reduction in luminal diameter or more by visual estimation were considered significant.
A lesion was considered the culprit lesion if it was a fresh occlusion at angiography. A coronary segment was considered occluded if it was subtotally or totally occluded with Thrombolysis in Myocardial Infarction (TIMI) flow grade <3 (16). An occlusion was considered acute if angiography revealed a thrombus at the site of the occlusion or if a guide wire passed easily through the occlusion if angioplasty was attempted. If there was no acute occlusion at angiography, the lesion with the most severe reduction of lumen diameter (at least 75%) was assigned as the culprit lesion if its localization corresponded with the location of ST segment elevations on the ECG.
The coronary arteries were divided into segments according to conventional terminology (17), and the number of significantly diseased coronary arteries was conventionally assigned from 1 to 3 (18). The extent of CAD was additionally scored using the coronary artery JS (19). This score provides more information on the amount of myocardium at risk for ischemia than the number of significantly diseased coronary arteries. Briefly, the coronary circulation is devided into six arterial segments: left anterior descending coronary artery (LAD), major anterolateral (diagonal) branch, first major septal perforator, left circumflex coronary artery (LCx), major circumflex marginal branch and the posterior descending artery. Each segment with a 75% or greater luminal diameter reduction is given a score of 1 point. Each segment distal to a segment with a 75% or greater stenosis is also given 1 point. The maximum number of points is 6. To determine the amount of myocardium at risk for necrosis (region at risk), we scored each angiogram by giving a point to segments that are distal to the infarct culprit segment. Thus, the maximum score is 5 (in case of left main occlusion).
Univariate analysis of categorical variables was carried out by a two-tailed Fisher’s exact test. Descriptive variables with a normal distribution were given by median and the 25th to 75th percentile. Multivariate analysis for prediction of out-of-hospital VF was carried out using logistic regression analysis (SPSS release 7.5).
Clinical data of 72 patients with out-of-hospital VF and 144 patients without out-of-hospital VF are summarized in Table 1. Time from symptom onset to collapse is unknown. The median time interval between collapse and angiogram was 135 min (25th to 75th percentile: 90 to 180 min). The median time interval between onset of pain and angiogram of patients without out-of-hospital VF was 202 min (25th to 75th percentile: 141 to 285 min). Both groups (matched as previously described) were comparable for history of CAD (including angina, coronary angioplasty and history of a previous MI). In patients with ST segment elevations (n = 211, excluding five patients with left bundle branch block), anterior localization was statistically more frequent in patients with out-of-hospital VF (62.5% vs. 37.5%; p = 0.0005). Left ventricular ejection fraction (LVEF) was documented in 43 patients with, and in 122 patients without, out-of-hospital VF. The LVEFs were, respectively, 42.4% (SD ± 13.8) and 47.3% (SD ± 11.9) (p = 0.03). The LVEFs of patients with acute occlusion of left coronary artery (LCA) with or without out-of-hospital VF were, respectively, 41.6% (SD ± 14.1) and 44.2% (SD ± 12.8) (p = 0.3). The LVEFs of patients with acute occlusion of the right coronary artery (RCA) with or without out-of-hospital VF were, respectively, 48.6% (SD ± 11.5) and 51.0% (SD ± 9.5) (p = 0.7).
Angiographic data of the total population
Angiographic data are summarized in Table 2. The most frequent (48.6%) location of culprit lesion was on the LAD followed by the RCA (36.1%). In most of these patients mid or proximal right coronary segments were involved. In all but one of these patients the culprit lesion was located proximal to the posterior descending artery. Nineteen of the 216 patients (8.8%) had chronic occlusions in a non-IRA. Flow in the IRA was absent or severely decreased in 90.3% of the patients; however, in 9.7%, a TIMI 3 flow grade was observed. Angiographic visible collaterals to the IRA were present in 29.2% of the patients. Although 58.3% of the patients presented with one-vessel disease, there was a large variation in the extent of CAD when expressed by JS (Table 2). Forty-three percent of the patients presented with a moderate (=2) to large (=3) region at risk.
Univariate analysis of angiographic data
The infarct artery in patients with out-of-hospital VF was LAD, LCx or RCA in, respectively, 65.3%, 19.4% and 15.3% (Table 3). This distribution was significantly different from patients without out-of-hospital VF (respectively, 40.3%, 13.2% and 46.5%; p < 0.001). Patients with an acute occlusion on the LAD or LCx had a higher risk for out-of-hospital VF compared to patients with an acute occlusion on the RCA (odds ratio and 95% confidence interval, respectively, 4.82 [2.35 to 9.92] and 4.92 [2.34 to 10.39] [Table 4]). Patients with an acute occlusion of the LAD had comparable risk for out-of-hospital VF when compared to patients with occlusion of LCx (odds ratio and 95% confidence interval: 1.10 [0.50 to 2.42]). The distribution of regions at risk differs between both groups (p = 0.01) (Table 3). However, in patients with occlusion on the LCA, the region at risk (≤1, 2 and 3) was not associated with out-of-hospital VF (Table 5).
With regard to the extent of CAD (expressed by number of diseased vessels or by JS), location of the culprit along the coronary arteries (proximal vs. mid or distal), flow in the IRA, presence or absence of collaterals to the IRA and chronic occlusions, there were no differences between the two groups (Table 3).
Logistic regression analysis was performed to detect those that are independently associated with out-of-hospital VF. Forward analysis retained four significant variables (Table 6). Acute occlusion of both LAD and LCx is significantly associated with out-of-hospital VF. A TIMI flow of 0 in the IRA or the presence of an additional (chronic) occlusion in a non-IRA are both of borderline significance as predictors of out-of-hospital VF. Stepwise addition of any other angiographic variable, such as region at risk, JS, and proximal versus distal occlusion was performed, and none of them showed an association with out-of-hospital VF, independent of the former four variables.
To our knowledge, this is the first study in patients with AMI complicated with out-of-hospital VF in which acute coronary anatomy is compared with acute coronary anatomy of a control group of patients with AMI not complicated by out-of-hospital VF. All patients, including those without out-of-hospital VF, were treated within the same time period and selected according to the same ECG criteria, and all underwent the same angiographic study protocol.
Acute occlusion of the LCA
Our major finding is the increased risk of out-of-hospital VF with acute occlusion on the LCA (LAD or LCx) when compared with acute occlusion of the RCA. This strong association persists even after multivariate adjustment for other anatomic variables such as JS or region at risk, which illustrates that this association is independent of infarct size or of amount of myocardium at risk for necrosis and that it is independent of the extent of CAD when it is expressed by number of diseased vessels or by the JS. For example, proximal LAD occlusion, which is associated with a large amount of myocardium at risk for necrosis, has no significantly larger risk for out-of-hospital VF than, for example, occlusion on the distal LCx (Table 4). Both occlusions, however, have a significantly higher risk of out-of-hospital VF than occlusion on the RCA. This finding contradicts earlier hypotheses derived on theoretical grounds, that acute occlusions of the RCA, which usually supplies the conduction system, are more prone to cause life-threatening arrhythmias (20). This theoretical prediction was in concert with observations of Davies and Thomas (21)in sudden ischemic death. Later studies, however, found associations of early VF and IRAs inconsistent (7–14), but some either included few or no patients with out-of-hospital VF or had no angiographic data (7,10–12). Other studies did not specifically address AMI complicated by out-of-hospital VF and possibly included a heterogeneous group of patients with cardiac arrest (13,14).
Protection of myocardium against orthosympathetic stimuli, for instance by beta-blockers, is known to increase the threshold for development of VF (22)or ventricular tachycardia (23)in the early phase of AMI. In experimental models of early VF by exercise-induced myocardial ischemia in conscious dogs with a healed MI, heart rate variability, autonomic function and baroreflex sensitivity have been studied (24). These studies associated decreased incidence of sudden death with high vagal tone and vagal reflexes. Moreover, vagal stimulation reduced incidence of early VF in the same model (25). Sinus bradycardia is particularly frequent in patients with inferior and posterior infarction (26,27). The cause of the vagotonia and resultant sinus bradycardia and hypotension appears to be stimulation of cardiac vagal afferent receptors (which are more common in the inferioposterior than the anterior or lateral portions of the left ventricle) with resulting efferent cholinergic stimulation of the heart. The phenomenon is a manifestation of the Bezold-Jarisch reflex (28). A hypothesis, generated by our data, is that vagal tone in patients with acute occlusion of RCA protects against early VF during AMI.
Amount of myocardium at risk of necrosis
A second finding is the relation between out-of-hospital VF and size of the region at risk (Table 3). In large studies of patients hospitalized for AMI, early VF is associated with final infarct size (8,9). In patients with out-of-hospital VF, no such relation to infarct size has been studied. In the present study, this association disappears after adjustment for the site of occlusion (RCA or LCA). Patients with occlusion on the LCA have, on average, a larger region at risk. Within this group, region at risk is no longer associated with out-of-hospital VF (Table 5). For example, patients with occlusion of the proximal LAD have an equivalent risk for out-of-hospital VF compared with patients with occlusion of the distal LCx (Table 4). Consequently, the association of region at risk with out-of-hospital VF in our study population is a confounding effect of occlusion of the LCA, and region at risk is not an independent risk factor for out-of-hospital VF.
Extent of CAD
The third finding is the absence of association of out-of-hospital VF with the extent of CAD. Distribution of JSs is not significantly different in both groups. Our findings are in accordance with data found at autopsy in victims of out-of-hospital cardiac arrest in whom extent of CAD did not differ significantly from that found in patients with stable angina or previous infarction (29). These autopsy studies face the difficulty of identifying AMI in these victims. Presence of an AMI or of an unstable coronary artery in victims of sudden death varies between 20% and 90% (30). A possible explanation for these discrepancies lies in the differing autopsy techniques used to find an acutely occluded segment and to describe the extent of CAD (13). Kyriakidis et al. (10)reported an association of in-hospital primary VF with the extent of CAD, expressed by the Gensini score (31), in a small number of patients. In our study, we used two previously validated measurements of extent of CAD, the number of diseased vessels (18)and the JS (19). Neither was associated with out-of-hospital VF.
Although to our knowledge this is the largest comparative angiographic study of patients with AMI and out-of-hospital VF, the study is still somewhat limited by the moderate number of patients. As a consequence, relatively small effects of region at risk or extent of CAD cannot be excluded with certainty.
The second limitation is potential selection bias. Only successfully resuscitated victims who presented with ST segment elevation or left bundle branch block were studied. The coronary anatomy of the successfully resuscitated victims could be different from the victims in whom the resuscitation was not successful (for example, occlusion of left main coronary segment). Autopsy studies of nonsurvivors should be performed, but these studies face the difficulty of diagnosing AMI in its early phase as well as its arrhythmic complications. The major finding of this study (association of out-of-hospital VF with occlusion on the LCA) cannot be explained by selection bias, because patients with occlusion of RCA complicated by out-of-hospital VF have theoretically an equal chance to survive resuscitation as patients with occlusion of the LCA.
The third limitation is the lack of information on preexistent LVEF. This information is available only in a prospective study design. For obvious reasons, such studies are very difficult to perform. We determined LVEF during hospitalization in the majority of patients. After adjustment for occlusion site (RCA or LCA), LVEF was not related to out-of-hospital VF.
Acute myocardial infarction related to the LCA is associated with greater risk for out-of-hospital VF than MI related to the RCA. The location of occlusion within the LCA (LAD, LCx, proximal or distal), the amount of myocardium at risk for necrosis and the extent of CAD are not related to out-of-hospital VF. Presence of a chronic occlusion in a non-IRA or complete absence of antegrade coronary flow in the IRA are possibly additional independent determinants of out-of-hospital VF. These data, obtained by acute angiography, offer new insights into the mechanisms by which acute coronary artery occlusion induces sudden cardiac death and strongly support the hypothesis that vagal reflex during MI is protective against early VF.
- acute myocardial infarction
- coronary artery disease
- infarct-related coronary artery
- jeopardy score (= amount of myocardium in “jeopardy”)
- left anterior descending coronary artery
- left coronary artery
- left circumflex coronary artery
- left ventricular ejection fraction
- right coronary artery
- Thrombolysis in Myocardial Infarction
- ventricular fibrillation
- Received September 29, 1998.
- Revision received July 27, 1999.
- Accepted September 13, 1999.
- American College of Cardiology
- Pasternak R.C,
- Brauwnwald E,
- Sobel B.E
- Löwel H,
- Dobson A,
- Keil U,
- et al.
- O’Doferty M,
- Taylor D.I,
- Quinn E,
- et al.
- Campbell R.W.F
- Deshpande S,
- Akhtar M
- Myerburg R.J,
- Kessler K.M,
- Castellanos A
- TIMI Investigators,
- Berger P.B,
- Ruocco N.A,
- Ryan T.J,
- Frederick M.M,
- Podrid P.J
- Kyriakidis M,
- Petropoulakis P,
- Antonopoulos A,
- et al.
- Davies M.J
- Chesebro J.H,
- Knatterud G,
- Roberts R,
- et al.
- Brushke A.V.G,
- Proudfit W.L,
- Sones F.M
- Califf R.M,
- Phillips H.R,
- Hindman M.C,
- et al.
- Vanoli E,
- De Ferrari G.M,
- Stramba-Badiale M,
- et al.
- Mark A.L
- Sequeira R.F,
- Myerburg R.J