Clinical, Angiographic and Hemodynamic Predictors of Recruitable Collateral Flow Assessed During Balloon Angioplasty Coronary Occlusion

1.1 Cardiac catheterization.

Therapy with all antianginal medications was continued until cardiac catheterization. All patients received aspirin (100 mg) orally the night before coronary angioplasty. Lorazepam (1 mg) was orally administered before the procedure. At the beginning of the catheterization all patients received 5,000 to 7,500 IU of heparin intravenously as a bolus. Nitroglycerin (0.1 mg intracoronary) was only given for the occurrence of coronary artery spasm. Cardiac catheterization was performed in all patients by the percutaneous femoral approach. A 7F sheath was inserted in both femoral arteries. One guiding catheter was used for balloon angioplasty and another guiding catheter was used for angiography of the contralateral artery and insertion of the Doppler catheter or Doppler wire.

1.2 Study protocol.

1.3 Quantitative coronary angiography.

The severity of the coronary narrowings was determined using an automated contour detection algorithm (ARTREK, ADAC Laboratories) in two orthogonal projections, using the guiding catheter as a reference, to obtain the percent diameter stenosis and the minimal lumen diameter.

1.4 Quantitative coronary blood velocity analysis.

Collateral flow was determined by the transient increase of coronary blood flow velocity of the contralateral artery during balloon inflation. The resistance of the collateral vascular bed was expressed as a relative value of the resistance of the contralateral vascular bed, according to a method described previously ([6]): RcollRcla=Pao−PwPao×Vc,defVc,inf−Vc,def,where Rcoll = resistance collateral vascular bed; Rcla = resistance contralateral artery; Pao = mean aortic pressure; Pw = coronary wedge pressure; V_c,def= maximal diastolic blood flow velocity of the contralateral artery during balloon deflation; V_c,inf= maximal diastolic blood flow velocity of the contralateral artery during balloon inflation.

1.5 Statistics.

The relation between the continuous variables—age, duration of angina, mean aortic pressure, coronary wedge pressure, coronary lesion severity and relative collateral vascular resistance, expressed as mean ±SD—and the presence of collateral vessels during coronary occlusion was evaluated using the unpaired Student ttest. The exact test for 2 × 2 tables was used to compare dichotomous variables. The predictive value of the continuous variables for the absence or presence of collateral vessels during coronary occlusion was compared by means of the areas under the receiver operating characteristic curves. The best cutoff values of these curves were used as a substitute for the continuous variable in an univariate analysis. The variables with p < 0.1 after univariate analysis were entered in a multivariate analysis. This multivariate analysis was performed by means of forward stepwise logistic regression using an analytical software program (Statistix 4.1). The coefficients of the independent predictors were used to calculate the probability of collateral vessels being present during coronary occlusion. The 95% confidence interval (CI) was calculated by means of the variance–covariance matrix for coefficients. The coefficients of predictors for recruitable collateral vessels, multiplied by a factor of 10 and rounded to a whole number, were used in a scoring system. A p value <0.05 was considered statistically significant.

2.1 Angiography of the collateral vessels.

Before the first balloon inflation, collateral vessels were graded 0 in 72 patients; graded 1 in 24 patients; graded 2 in 6 patients; and graded 3 in 3 patients (Fig. 1). During balloon occlusion, collateral vessels remained absent (grade 0 or 1) in 54 patients and were present (grade 2 or 3) in 51 patients, 9 of whom had collateral vessels before occlusion. Thus, recruitable collateral vessels were present in 42 patients. Collateral vessels were absent in all patients after successful completion of the procedure.

• Download figure
• Open in new tab
• Download powerpoint

Fig. 1

Grading of collateral vessels according to the Rentrop classification before, during and after balloon inflation in 105 patients with single-vessel disease (grade 0 or 1 = absence of collateral vessels; grade 2 or 3 = presence of collateral vessels).

2.2 Clinical and hemodynamic correlates of the presence of collateral vessels during balloon coronary occlusion.

The clinical and hemodynamic factors in relation to the absence or presence of collateral vessels during coronary occlusion are summarized in Table 1. Collateral vessels during balloon coronary occlusion were related to the duration of angina and coronary lesion severity. Furthermore, a greater percentage of transient increase of coronary blood flow velocity of the contralateral artery was documented in the presence of collateral vessels during balloon coronary occlusion (19 ± 12% vs. 4 ± 6%, p < 0.001, Fig. 2). This resulted, in conjunction with an increased coronary wedge pressure, in a reduction of the relative collateral vascular resistance (p < 0.001) (Table 1). Finally, ECG signs of ischemia were significantly reduced when collateral vessels were present (p < 0.001, Table 1).

• Download figure
• Open in new tab
• Download powerpoint

Fig. 2

The topimages show contrast injection of the left anterior descending coronary artery (left anterior oblique view) before (left)and during (right)balloon occlusion of the right coronary artery, displaying opacification of the distal part of the right coronary artery by collateral vessels (grade 2 according to the Rentrop classification). The lower leftrecordings show aortic pressure and distal coronary lesion pressure recordings before, during and after balloon inflation, as well as simultaneously obtained blood flow velocity alterations of the left anterior descending coronary artery. The lower righttracings show the ECG before balloon inflation and at 1 min of coronary occlusion of the right coronary artery. Recruitability of collateral vessels coincides with a high coronary wedge/aortic pressure ratio (0.69), an 18% transient increase in the maximal diastolic blood flow velocity of the left anterior descending coronary artery and absence of ST segment changes at 1 min of coronary occlusion.

The predictive values of continuous variables, related to the presence of collateral vessels during coronary occlusion, were compared by the areas under the receiver operating characteristic curves (Table 2). The relative collateral vascular resistance was a better predictor of the presence of collateral vessels than the other continuous variables (p < 0.05).

The variables duration of angina and use of nitrates were the only independent clinical predictors of the presence of spontaneously visible and recruitable collateral vessels after stepwise logistic regression analysis (odds ratio [OR] 11.7, 95% CI 4.4 to 31, p < 0.001; and OR 0.29, 95% CI 0.1 to 0.8, p < 0.05, respectively). The overall predictive accuracy of the regression equation was 0.75. Table 3demonstrates the calculated probability of the presence of spontaneously visible and recruitable collateral vessels using these two variables.

The duration of angina, use of nitrates, coronary lesion location and coronary lesion severity were the only independent predictors of recruitable collateral vessels after stepwise logistic regression analysis (Table 4). The regression equation for the model using these variables is: logit (p) = −4.5 + 2.6 (duration ≥3 months) + 1.7 (no use of nitrates) + 1.5 (proximal lesion location) + 2.3 (diameter stenosis ≥75%). The overall accuracy of this regression equation was 0.80.

The coefficients of the independent predictors were used in a scoring system to calculate the probability of the presence of recruitable collateral vessels (Table 4). The relation between this calculated score, based on the aforementioned variables, and the probability of the presence of recruitable collateral vessels is shown in Fig. 3.

• Download figure
• Open in new tab
• Download powerpoint

Fig. 3

The relation between clinical and angiographic predictors, expressed in a scoring system, and the probability of the presence of recruitable collateral vessels. Dashed lines= 95% confidence interval.

3.1 Clinical and angiographic predictors of recruitable collateral vessels.

Coronary angioplasty serves as a model for studying the collateral circulation in a controlled fashion. In their elegant study, Rentrop et al. ([2]) demonstrated for the first time that the marked differences in angiographic appearance of collateral vessels before and during balloon coronary occlusion was related to the pressure gradient across the collateral vascular bed. The present study confirms their angiographic findings, showing that spontaneously visible collateral vessels are present in ∼10% of the patients with single-vessel disease; the recruitability of collateral vessels was documented in approximately 40% of these patients (Fig. 1). The functional significance of these recruitable collateral vessels has been demonstrated in several clinical studies ([3–6, 8]). This indicates that angiographic assessment of recruitable collateral vessels is important for appropriate classification of collateral vascular development.

3.2 Hemodynamic predictors of recruitable collateral vessels.

Previous clinical studies showed that the coronary wedge pressure is increased in the presence of collateral vessels, presumably as a consequence of a reduced collateral vascular resistance, resulting in a reduction of signs of ischemia during brief coronary occlusion ([3, 4, 6]). Our current understanding of the collateral circulation is limited, partly due to the lack of methods capable of expressing the development of the collateral vascular bed in terms of flow and resistance. The dynamic behavior of collateral vessels as seen angiographically before and during balloon coronary occlusion can be examined by blood flow velocity analysis in the contralateral donor coronary artery ([6, 7]). These studies demonstrated that a balloon coronary occlusion results in a transient 10% to 70% increase in coronary blood flow velocity in the contralateral artery when collateral vessels are present, although this phenomenon is marked in the absence of collateral vessels. Blood flow velocity changes in the contralateral artery during brief coronary occlusion may also be induced by alterations in preload due to akinesia of the occluded vascular bed ([15]). In the current investigation we did not assess alterations in preload during coronary occlusion. It can be expected, however, that an increase in preload is more pronounced when collateral vessels are absent ([4]), whereas in the present study, a transient increase of flow velocity during coronary occlusion was noted in those patients with collateral vessels. Furthermore, hyperkinesia of the nonischemic myocardium during coronary occlusion is a potential mechanism that may also explain the observed increase of blood flow velocity in the donor coronary artery. However, both experimental and human studies have indicated that this phenomenon is not operative during brief coronary occlusion ([16–19]). Consequently, these considerations support the contention that the observed flow velocity changes in the contralateral artery are related to collateral flow.

A recent study by Pijls et al. ([20]) indicated that the fractional collateral flow reserve, that is, the coronary wedge pressure related to aortic pressure after correction for the central venous pressure, serves as an alternative for assessment of collateral flow. These investigators selected patients with stable angina of more than 3 months’ duration and did not document angiographic recruitability of collateral vessels. The patients in the present study had single-vessel disease and normal left ventricular function; hence, the aortic pressure minus the central venous pressure approximates the aortic pressure. Consequently, the fractional collateral flow reserve reflects the coronary wedge/aortic pressure ratio in our study. Our results indicate that the relative collateral vascular resistance is a better predictor of the angiographic presence of collateral vessels during coronary occlusion than the coronary wedge pressure/aortic pressure ratio of ischemia, as determined by the receiver operating characteristic curves (Table 3). This may be related to the fact that the blood flow velocity alterations, used to calculate the relative collateral vascular resistance, are measured in the contralateral donor coronary artery, whereas the coronary wedge pressure assessed in the recipient coronary artery is determined by collateral flow arising from both nondiseased coronary arteries and hemodynamic factors such as central venous pressure and left ventricular pressure ([11]).

Coronary blood flow analysis of the contralateral artery allows the expression of the development of collateral vascular bed in terms of the relative vascular resistance. This offers the possibility of studying the pharmacologic responsiveness of the collateral vascular resistance and provides insight into intrapatient variability of this resistance. The present study demonstrates that the angiographic presence of collateral vessels results in a threefold reduction of the relative collateral vascular resistance. The large interpatient variability of the collateral vascular resistance (Table 2) remains unclear and may be related to variations in stimulating factors such as the duration of angina, medication, coronary lesion severity, lesion location or other unknown contributing factors.

Finally, ECG signs of ischemia were reduced in the presence of collateral vessels, although the absence of ischemia (<0.1 mV ST segment shift) was noted in only 45% of the patients with collateral vessels during balloon coronary occlusion. This illustrates that collateral vessels are capable of reducing rather than abolishing myocardial ischemia during brief coronary occlusion.

3.3 Study limitations.

The collateral vascular development was related to the period of symptomatic disease, although the contribution of silent ischemia to the development of collateral vessels is unknown. This subjective factor may be an explanation for the overlap between patients with and without collateral vessels during coronary occlusion. The results of the study are only applicable to patients with single-vessel disease and normal left ventricular function, which represents only a small proportion of patients with coronary artery disease. The medical therapy of the patients studied was not uniform and this may have contributed to observed variations between patients. The angiographic grading of collateral vessels is sensitive to variations in the applied technique and is subject to intraobserver and interobserver viability. Furthermore, intracoronary blood flow velocity assessment is a sensitive technique for the detection of alterations in blood flow, but this method is also prone to technical failures. The functional significance of collateral vessels during coronary occlusion was determined by ECG monitoring and was not expanded to hemodynamic monitoring or assessment of global or regional ejection fraction. The study provides only information on the function of collateral vessels during brief coronary occlusion.

3.4 Clinical implications.

The establishment of the period required for maturation of preexisting collateral vessels in humans has important clinical implications. Experimental studies have indicated that myocardial infarct size is determined by the size of the myocardium at risk, the duration of coronary occlusion and collateral flow to the jeopardized myocardium ([21, 22]). Recent clinical studies have demonstrated that a short period of preceding angina (24 to 48 h) exerts a protective effect, as documented by an ∼25% to 30% reduction in myocardial infarct size ([23, 24]). The present study demonstrates that such a period is too short for effective collateral vascular development. The protective effect does not seem to be exerted by collateral flow but is most likely due to ischemic preconditioning, presumably related to stimulation of adenosine-A₁receptors and opening of adenosine triphosphate–dependent potassium channels ([25, 26]). The present study illustrates that a patient without preceding angina runs a high risk of developing a large myocardial infarction after abrupt coronary occlusion owing to the fact that collateral vessels are absent and a protective effect related to ischemic preconditioning is lacking. Furthermore, clinical studies have demonstrated that the time window for reperfusion therapy can be increased in the presence of collateral flow to the jeopardized myocardium ([27–29]). Finally, clinical studies using myocardial contrast echocardiography after the acute phase of myocardial infarction have demonstrated that collateral vascular supply to the occluded vascular bed is associated with improved left ventricular function after revascularization ([30, 31]).

The present study indicates that the clinical information on the duration of angina and the use of nitrates allows the prediction of spontaneously visible and recruitable vessels with a 75% overall accuracy. Furthermore, the clinical and angiographic variables predict recruitability of collateral vessels with an 80% overall accuracy. These findings are important for risk stratification of patients undergoing interventions for ischemic coronary syndromes.