Influence of Collateral Circulation on In-Hospital Death From Anterior Acute Myocardial Infarction

1.1 Study Patients

We selected patients who fulfilled all the following criteria: 1) admission to the hospital because of suspected anterior wall AMI (continuous ischemic chest pain lasting >30 min and a minimum of 0.2 mV of ST segment elevation in two contiguous anterior precordial leads); 2) diagnostic coronary angiography showing complete occlusion of the left anterior descending coronary artery (LAD) (Thrombolysis in Myocardial Infarction [TIMI] flow grade 0 or 1); 3) treatment by primary PTCA within the first 6 h of the onset of symptoms; and 4) left and right coronary angiograms obtained before the attempted angioplasty of sufficient quality to assess the filling of the LAD and side branches by collateral circulation. Patients with AMI secondary to acute or subacute coronary occlusion after PTCA or coronary artery bypass graft surgery (CABG) and patients treated by PTCA after failed thrombolysis (rescue PTCA) were not considered.

1.2 Coronary Angiography and Primary Angioplasty

Patients received 300 mg of aspirin intravenously as soon as possible unless there was a known allergy. Coronary angiography and primary PTCA were performed in the first 6 h from the onset of symptoms of AMI. Left heart catheterization and left ventriculography were not performed at this time. Once the femoral artery was catheterized, 10,000 IU of heparin was administered with additional boluses of 3,000 IU if needed to achieve an activated clotting time of ≥350 s. Coronary arteriograms were obtained in at least two projections. Intracoronary nitroglycerin was not administered before diagnostic angiograms were obtained. Flow in the LAD was determined before and after primary PTCA and graded as described in the TIMI trial [12]. The LAD occlusion was considered proximal if it was located before the origin of the first well developed septal branch, distal if it was located after the origin of the third diagonal branch and mid if located between these limits. Coronary stenoses were measured by an on-line automatic edge detection system (Phillips DCI). Angiographic images were recorded on 35-mm film at 12.5 frames/s. Primary PTCA was performed with conventional catheter balloon technique. Stents were used only for suboptimal balloon angioplasty result. The policy was to perform angioplasty only to the infarct-related artery, but in exceptional cases additional angioplasty to another vessel was performed because of compromised hemodynamic status. Intraaortic balloon pumping was used in patients with hemodynamic instability or established cardiogenic shock. The primary PTCA was considered successful when flow in the infarct-related artery at the end of the procedure was TIMI grade 2 or 3 and the residual stenosis was <50%.

1.3 Coronary Collateral Circulation

We analyzed the collateral circulation to the occluded LAD existent in the first 6 h of AMI before the angioplasty attempt. It was assessed in coronary cineangiograms by two experienced coronary angiographers who had no knowledge of patient data. We excluded from the present study patients whose angiograms were inadequate for assessing the collateral circulation to the LAD because the injection period was too short, filming of the LAD territory in contralateral injections was incomplete or right coronary angiograms were not obtained before angioplasty. Collateral circulation was graded by using a semiquantitative scale from 0 to 3 (collateral index) depending on the degree of angiographic opacification of the occluded LAD in the best injection [13]: 0 = no collateral circulation; 1 = collateral filling of LAD side branches without visualization of any epicardial segment; 2 = collateral partial filling of the epicardial segment, and 3 = collateral filling of the complete epicardial segment. In case of noncoincidence, the angiographers reached a consensus.

1.4 Groups

In the prospective design of this study we decided to classify the patients into two groups: group A, patients without angiographic collateral filling of the LAD or side branches (collateral index 0), and group B, patients with some angiographic collateral filling of the LAD or side branches (collateral index 1, 2 or 3).

1.5 In-Hospital Clinical Course

Adjunctive medical therapy followed the standards of the coronary care unit. The primary objective of this study was to examine in-hospital death. Secondary objectives were to examine the occurrence of reinfarction, malignant arrhythmias, mechanical complications and cardiogenic shock. Reinfarction was defined as the recurrence of ischemic chest pain of ≥30-min duration with ST segment elevation of ≥0.1 mV over the previous ST segment in two contiguous leads and an MB fraction of creatine kinase (CK-MB) profile consistent with the diagnosis. Malignant arrhythmias included sustained ventricular tachycardia, ventricular fibrillation or high degree atrioventricular (AV) block. Mechanical complications included free wall rupture, ventricular septal rupture and severe mitral regurgitation secondary to papillary muscle rupture. Cardiogenic shock was indicated by a systolic blood pressure <90 mm Hg in association with signs of tissue hypoperfusion (cold extremities, cyanosis, oliguria or altered mental status) not derived from extramyocardial causes (hypovolemia, arrhythmias, tamponade, drugs).

1.6 Statistical Analysis

Continuous variables that did not follow a normal distribution according to the Shapiro-Wilk test such as age, CK-MB peak serum level and residual stenosis, are expressed as median and the 25th and 75th percentiles. Time followed a normal distribution and is expressed as mean ±SD. For comparisons of continuous variables between both groups the Student ttest for unpaired observations was used. Unless otherwise specified, categoric variables are presented in the text and tables as number and percent of patients. We used the chi-square test to compare categoric variables between groups unless expected values were <5, in which case we used the Fisher exact test. The association between collateral circulation as a four-grade variable and mortality was analyzed by using the Mantel-Haenszel test; collateral indexes 2 and 3 were combined because their expected values were <5. Multivariate analyses for mortality and cardiogenic shock were performed by multiple logistic regression. Odds ratios were expressed together with 95% confidence intervals. All results were analyzed by using a two-sided significance level of 0.05, but all p values < 0.20 are reported. Data analyses were performed by using the statistical software JMP 3.0.1 from SAS Institute Inc, 1994, and SPSS 4.0 from SPSS Inc, 1990.

2.1 Patients and Study Groups

From July 1991 to September 1996, 238 patients with suspected anterior wall AMI were treated in our institution by primary PTCA within the first 6 h from the onset of symptoms. One hundred twenty-seven of these patients were treated by primary PTCA because they were enrolled in investigative trials; the majority of them were randomly selected in a trial comparing PTCA with thrombolytic treatment. These protocols required informed written consent and were approved by the review committee of our institution. The remaining patients were selected to be treated by primary PTCA because they had contraindications to thrombolytic treatment, cardiogenic shock, a large infarction or other high risk characteristics. The descriptive data and clinical in-hospital course of this group are summarized in Table 1. For the present analysis, we selected 180 of these patients who fulfilled the criteria just described. The 58 patients who did not fulfill the criteria had left coronary angiograms showing a patent LAD (25 patients) with TIMI flow grade ≥2 (33 patients) or had angiograms inadequate for assessing the collateral circulation to the LAD, mostly because right coronary angiography was not performed before angioplasty.

Of the 180 patients included, 115 patients (64%, group A) did not have angiographically visible collateral circulation to the LAD (collateral index 0). In the remaining 65 patients (36%, group B), we found some collateral filling of the LAD (collateral index 1 in 33 patients [18%], index 2 in 23 patients [13%] and index 3 in 9 patients [5%]).

2.2 Baseline Characteristics

The baseline clinical characteristics of groups A and B are summarized in Table 2. Significant differences between groups were not found in age, gender or prevalence of the main cardiovascular risk factors. However, group B had a greater prevalence of previous angina than did group A (22 patients [34%] vs. 17 [15%], p = 0.003) and a tendency to a greater prevalence of previous myocardial infarction (13 patients [20%] vs. 13 [11%], p = 0.11).

2.3 Coronary Angiography and Primary Angioplasty

The elapsed time between the onset of symptoms suggestive of AMI and the opening of the LAD was 195 ± 86 min in group A and 211 ± 90 min in group B (p = NS). The angiographic findings and the result of primary PTCA in both groups are presented in Table 3. There were no significant differences between groups in the number of diseased vessels, although in group A there was a tendency to present multivessel disease less frequently (51 patients [44%] vs. 36 [55%], p = 0.15). The proximal segment of the LAD was less frequently occluded in group A than in group B (43 patients [37%] vs. 36 [55%], p = 0.05). PTCA was considered successful in a similar proportion of patients from both groups. The proportion of patients who received stents during PTCA was also similar in both groups (22 patients [19%] in group A vs. 16 [25%] in group B, p = NS). An intraaortic balloon pump was implanted in 25 patients (22%) in group A and in 13 (20%) in group B (p = NS).

2.4 In-Hospital Course and Mortality Rate

The clinical complications that occurred in both study groups are shown in Table 4. Of the 31 patients (17%) who died during the hospital stay, 26 (23%) were from group A and 5 (8%) from group B (p = 0.01). A progressive decrease in the in-hospital mortality rate was observed with increasing grades of collateral circulation. The mortality risk in patients without collateral circulation was 2.5 times higher than in patients with collateral index 1 (p = 0.06) and 3.6 times higher than in patients with collateral index 2 or 3 (p = 0.03). This linear association was statistically significant by the Mantel-Haenszel test (p = 0.01, Fig. 1). Other factors associated with in-hospital mortality by univariate analyses are presented in Table 5.

• Download figure
• Open in new tab
• Download powerpoint

Fig. 1

Linear association between collateral index (CI) and in-hospital mortality rate (Mantel-Haenszel test, p = 0.01). ∗Collateral indexes 2 and 3 were grouped because their expected values were <5. Collateral circulation was index 2 in 23 patients and index 3 in 9 patients; 1 patient from each of these subgroups died.

The independent effect of collateral circulation on in-hospital death was confirmed by a multiple logistic regression analysis. After adjustment for age, previous infarction, diabetes, multivessel disease and PTCA result, the absence of collateral circulation remained an independent predictor of in-hospital death (p = 0.01). The final model, after excluding the variables that did not modify the independent weight of collateral circulation on in-hospital death, is shown in Table 6.

Cardiogenic shock was the most frequent cause of mortality, accounting for 74% of deaths. Cardiogenic shock developed more often in patients from group A than in patients from group B (30 patients [26%] vs. 4 [6%], p = 0.001). The timing of shock seemed different in each group: In group A, 24 (80%) of the 30 patients who experienced cardiogenic shock were already in shock before PTCA, whereas in group B only 2 of 4 patients had shock before PTCA. Multiple logistic regression analysis proved that the absence of collateral circulation is also an independent determinant of cardiogenic shock (odds ratio 5.6, 95% confidence interval 1.9 to 17, p = 0.002) after adjustment for age and previous infarction. This explains the higher in-hospital mortality rate of patients without collateral circulation.

3.1 Collateral Assessment

It has been established [14–17]that reperfusion can be achieved quickly and safely by emergency PTCA. Diagnostic angiography inherent in the primary PTCA procedure is a valuable opportunity for assessing coronary collateral circulation in the first hours of AMI. In recent years, alternative methods of evaluating coronary collateral circulation have been proposed [18–21]on the basis that angiography lacks sufficient sensitivity. It is certain that angiography cannot detect collateral vessels <100 μm in diameter and can miss some intramyocardial collateral vessels [9]. However, it is also certain that all normal human hearts have, among the territories of the major coronary arteries, small anastomotic vessels that range up to 200 μm in diameter [22]. Because collateral vessels that are too small to be visualized angiographically probably have no clinical impact, we believe that angiography offers sufficient sensitivity for clinical purposes.

Collateral vessels exhibit vasomotor responses to different agents. For example, nitroglycerin enhances collateral flow by marked vasodilation [23], whereas, various substances generated by platelet aggregation and thrombus formation in the setting of AMI (e.g., serotonin, thromboxane A₂) have been implicated in the vasoconstriction of small and collateral vessels of the area at risk [24]. Because we did not inject intracoronary nitroglycerin before performing the diagnostic angiograms, the collateral circulation documented better reflects the spontaneous previous status, without pharmacologic enhancing.

3.2 Previous Coronary Heart Disease

In this study, patients with angiographically documented collateral vessels to the occluded LAD (group B) had a greater prevalence of previous angina. These findings are in accordance with previous works [25–29]. Other studies [25, 29]found a higher prevalence of myocardial infarction and multivessel disease in patients with collateral circulation. In our study, patients with collateral vessels only tended to have previous infarction and multivessel disease more frequently than did patients without collateral vessels (p = NS). The prevailing opinion is that patients with performed and matured collateral vessels probably have had a longer, although sometimes silent, course of ischemic disease [25, 26, 28]. The existence of a severe coronary stenosis, especially one with >90% diameter narrowing, or intermittent occlusions over a nonsignificant plaque, such as those in patients with vasospastic angina, can induce the development of collateral vessels from surrounding vascular beds that remain as potential conduits until a new total occlusion occurs [28]. Nevertheless, high interindividual variability seems to exist for collateral development in response to ischemic stimuli [30].

3.3 Mortality and Complications

In our patients with anterior AMI treated by primary PTCA, the absence of collateral blood flow to the ischemic myocardium in the early hours of infarction was a strong independent predictor of cardiogenic shock and mortality. The benefit of having collateral circulation was striking, especially as the patients with collateral vessels at the time of AMI, our group B, had a greater prevalence of some high risk characteristics, such as occlusion of the LAD in the proximal segment, multivessel disease and previous infarction. The presence of only mild degrees of collateral circulation (Rentrop index 1) might already exert a protective effect. With respect to other complications, the proportions of reinfarction and free wall rupture were similar in groups A and B, although the number of these events was low. Malignant arrhythmias such as sustained ventricular tachycardia, ventricular fibrillation or high degree AV block were more frequent in patients without collateral vessels, although there were no differences between groups after patients with cardiogenic shock were excluded.

Several studies [3–5]have demonstrated that collateral circulation at the time of AMI reduces infarct size, an effect that might explain, at least in part, how collateral vessels protect against cardiogenic shock and mortality. In this work, higher peak CK-MB levels were observed in patients without collateral circulation, but this finding has to be cautiously interpreted given the recognized limitations of CK-MB measurements in estimating infarct size [31–33], especially in patients with collateral circulation [34].

When total arterial flow is reduced by coronary stenosis or occlusion, coronary vasodilator reserve is exceeded and flow to the ischemic region is distributed along a gradient that is inversely related to the gradient of intramyocardial pressure. Therefore, the small amounts of blood flow provided through the collateral route are distributed in a nonuniform pattern, being preferentially shunted to the subepicardial zone where intramyocardial pressure is lowest [35]. In experimental models, nearly transmural infarcts develop in dogs with very little collateral flow, whereas nontransmural infarcts, with survival of myocardium in the subepicardial zone [36, 37], develop in dogs with substantial collateral flow, which is always greatest in this zone. The lower incidence of cardiogenic shock in patients with collateral vessels could be explained, at least in part, by the preservation of an epicardial rim of viable myocardium when operating collateral vessels are present at the time of infarction. This phenomenon could ameliorate the systolic dysfunction and promote a partial functional recovery of the involved wall segment early after AMI. This mechanism could also explain the observation of others [9, 38]of a low incidence of infarct expansion and aneurysm formation in patients with well developed collateral channels at the time of infarction.

3.4 Limitations of the Study

The population origin of the present study is a partly selected high risk group. Although the majority of patients were treated by primary PTCA in an unselected manner because of their participation in investigative trials, many patients in this study were selected to undergo primary PTCA because they had high risk characteristics, especially cardiogenic shock or extensive infarction. For this reason the rates of adverse events of the study population, particularly cardiogenic shock and death, are increased; consequently, the relative importance of collateral circulation in reducing cardiogenic shock and mortality might be less striking in unselected populations.