Author + information
- Received March 11, 2002
- Revision received May 14, 2002
- Accepted June 7, 2002
- Published online November 6, 2002.
- Michael Billinger, MD*,
- Patrik Kloos, BS*,
- Franz R Eberli, MD*,
- Stephan Windecker, MD*,
- Bernhard Meier, MD, FACC* and
- Christian Seiler, MD, FACC*,* ()
- ↵*Reprint requests and correspondence:
Dr. Christian Seiler, Professor of Cardiology, University Hospital CH-3010 Bern, Switzerland.
Objectives We sought to evaluate whether coronary collateral flow is clinically relevant for future cardiac ischemic events.
Background The link between good collateral supply related to less myocardial damage and fewer cardiac events has not been established prospectively beyond doubt.
Methods In 403 patients with stable angina pectoris undergoing percutaneous transluminal coronary angioplasty (PTCA) and quantitative collateral assessment, the occurrence of major adverse cardiac events ([MACE] cardiac death, myocardial infarction, unstable angina pectoris) and stable angina pectoris was monitored during follow-up. Collateral flow index (CFI) was determined using intracoronary pressure or Doppler guidewires. Mean aortic ([Pao] mm Hg) and distal coronary artery occlusive pressure ([Poccl] mm Hg) during balloon angioplasty (PTCA), or distal coronary flow velocity time integral during ([Voccl] cm) and after ([Vø-occl] cm) PTCA were measured continuously. Pressure-derived CFI was calculated as follows: (Poccl− 5)/(Pao− 5). Doppler-derived CFI: Voccl/Vø-occl. Patients were subdivided into a group with well (CFI ≥ 0.25) and poorly developed collaterals (CFI < 0.25).
Results Average follow-up was 94 ± 56 (15 to 202) weeks. There were 134 patients with CFI ≥0.25 (61 ± 11 years) and 269 with CFI <0.25 (61 ± 10 years). The overall cardiac ischemic event rate (MACE and stable angina pectoris) during follow-up was 23% in patients with CFI ≥0.25 and 20% in patients with CFI <0.25 (p = NS). However, only 2.2% of patients with good collateral flow suffered a major cardiac ischemic event, compared with 9.0% among patients with poorly developed collaterals (p = 0.01). The incidence of stable angina pectoris was significantly higher in patients with well developed collaterals than in those with poorly developed collaterals (21% vs. 12%; p = 0.01).
Conclusions In this relatively large population with chronic stable coronary artery disease undergoing quantitative collateral measurement, the beneficial impact of well developed collateral vessels on the occurrence of future major cardiac ischemic events is clearly demonstrated.
In the situation of an acute coronary artery occlusion, well developed collateral vessels have been demonstrated to reduce myocardial infarct (MI) size, to protect against ventricular aneurysm formation, and to improve systolic ventricular function (1,2). Theoretically, those beneficial effects from a well developed collateral circulation should directly translate into a markedly lower incidence of future cardiac ischemic events among patients with jeopardized coronary vascular territories. The link between good collateral supply related to less myocardial damage and fewer cardiac events has not been established prospectively beyond doubt. This may be explained by the fact that previous clinical studies on the relationship between coronary collaterals and cardiac events lack quantitative means for collateral measurements (3–7)or sufficient patient numbers (8). The hypothesis of this study was that there is a beneficial impact of well developed collaterals on the occurrence of future major cardiac ischemic events. Thus, it was the purpose of our study to examine the influence of quantitatively determined collaterals on ischemic events in a cohort of more than 400 patients with stable coronary artery disease (CAD).
A total of 403 patients (age, 61 ± 11 years; 311 men, 92 women) with one- or two-vessel CAD were included in the study. All patients underwent percutaneous transluminal coronary angioplasty (PTCA) for clinical purposes (stable angina pectoris or positive exercise test with electrocardiogram [ECG] ST-segment depression).
The study population was subdivided into a group with well (i.e., high collateral flow index [CFI] ≥0.25) and poorly developed collaterals (i.e., low collateral flow index, CFI < 0.25) according to the pressure- or Doppler-derived CFI.
This investigation was approved by the local ethics committee, and all patients gave informed consent to participate in the investigation. The patients included in this study have been described, in part, elsewhere (9).
Patients underwent left heart catheterization and coronary angiography for diagnostic purposes. Aortic pressure was measured using the 6F guiding catheter. Biplane left ventricular angiography was performed followed by diagnostic coronary angiography. Coronary artery lesions were estimated quantitatively as percent diameter reduction. Angiographic collateral degree (0 to 3) was determined according to the extent of epicardial coronary artery filling via collaterals with contrast medium from the contralateral side before PTCA (10). Patients were divided into two groups: 1) no or low angiographic collateral degree (0 to 1); or 2) high angiographic collateral degree (2 to 3).
Coronary collateral assessment
A fiberoptic pressure-sensored 0.014-in. PTCA guidewire (WaveWire, Jomed, Switzerland) was used to determine pressure-derived CFI (CFIP) by simultaneous measurement of mean aortic pressure ([Pao], mm Hg, via the angioplasty guiding catheter) and the distal coronary artery perfusion pressure during balloon occlusion (i.e., coronary wedge pressure [Poccl] mm Hg) (Fig. 1). Central venous pressure (CVP) was estimated to be equal to 5 mm Hg; CFIPwas calculated as (Poccl− CVP) ÷ (Pao− CVP).
Doppler flow velocity measurements were performed using a 0.014-in. PTCA Doppler guidewire with a 12-MHz piezoelectric crystal at its tip (FloWire, Jomed, Switzerland). The velocity-derived index of collateral flow (CFIV) was determined as the ratio of flow velocity time integral distal to the occluded stenosis ([Voccl] cm) ÷ flow velocity time integral during vessel patency ([Vø-occl] cm) obtained at identical wire location 1 to 4 min after PTCA when the hyperemic response was abolished and the flow velocity signal was again constant over 30 s: Voccl/Vø-occl(11). In patients with bidirectional flow velocity signals, they were added to obtain total collateral flow velocity integral.
The validation of the pressure- and Doppler-derived CFI has been described previously (11).
After diagnostic coronary angiography, an interval of at least 10 min was allowed for dissipation of the effect of the nonionic contrast medium (iopamidol 755 mg/mL) on coronary flow velocity and vasomotion. An intracoronary (i.c.) bolus of 0.2 mg of nitroglycerin was given in order to maintain epicardial coronary artery caliber constant and, thus, to eliminate dimensional changes of the epicardial vessels that may influence collateral flow indexes. The pressure or Doppler guidewire was positioned distal to the stenosis to be dilated. Coronary and aortic pressure or flow velocity were measured continuously. During the entire protocol, an i.c. ECG obtained from the guidewire, a three-lead surface ECG, and blood pressure were recorded.
After a follow-up period of 94 ± 56 weeks (minimum 12 weeks), all patients or the attending physician were contacted about cardiac ischemic events (i.e., cardiac death, MI, unstable and stable angina pectoris).
Between-group comparison of continuous angiographic, hemodynamic, and collateral data was performed by an unpaired two-sided Student ttest (independent sample ttest). A chi-square test was used for comparison of categorical variables among the two study groups, and event analysis was performed by Kaplan-Meier analysis. Statistical significance was defined as a p value of <0.05.
There were no statistically significant differences among the two groups with well and poorly developed collateral vessels with regard to patient age, gender, hemodynamic variables during cardiac catheterization (heart rate and blood pressure), frequency of cardiovascular risk factors, and vasoactive drugs used (Table 1).
Angiographic and coronary collateral data
In the group with CFI ≥0.25, the severity of the dilated stenosis was higher and more vessels were affected by CAD than in the group with CFI <0.25. No statistical difference was determined between the groups regarding left ventricular ejection fraction (Table 2). In addition, the occurrence of angina pectoris and ST-segment changes on i.c. ECG during balloon occlusion was significantly lower in the group with CFI ≥0.25 than in the group with CFI <0.25.
Follow-up data and event rate
Follow-up ranged between 15 to 202 weeks (mean: 94 ± 56). Follow-up duration was significantly longer in the group with CFI ≥0.25 than with CFI <0.25 (Table 3). Only 2.2% of the patients with good collateral flow suffered a major cardiac ischemic event as compared with 9.0% among patients with poor collaterals (p = 0.01, Fig. 2). All major cardiac events in both groups occurred within the first year after PTCA (Fig. 3). Regarding the overall event rate, there was no statistically significant difference between the two groups: 23% and 20% (Table 3, Fig. 4). However, the incidence of minor ischemic events defined as chronic stable angina was significantly higher in patients with well developed compared with patients with poorly developed collaterals (21% vs. 12%; p = 0.01; Table 3, Fig. 4).
This sizeable prospective study in patients with stable CAD undergoing collateral flow measurement during PTCA documents a fourfold reduced rate of major adverse cardiac events during follow-up in individuals with well versus poorly grown collateral vessels. This appears to challenge other investigations of the predictive value of the collateral circulation.
Controversy in the literature about the protective effect of the collateral circulation
The ongoing controversy over whether coronary collateral vessels are protective with respect to future cardiac events is substantially related to common methodological problems such as relatively low numbers of individuals included in the studies (3–6,12,13), short observation periods (13–15), and use of surrogate end points for cardiac events such as systolic left ventricular function (4,12,13,15). More specifically, it probably has to do with the blunt instrument for assessing the collateral circulation used in most of the approximately 10 recent studies on the topic, namely coronary angiography (3–7,13,14). Also, among the majority of the studies just mentioned, angiographic collateral qualification has been performed to look for spontaneously visible instead of recruitable collaterals, the latter of which appear in response to occlusion of the collateral receiving vessel and, thus, reflect collateral supply more comprehensively than the former. These problems have been overcome by the present investigation, that is, in more than 400 patients with an average follow-up of almost two years, quantitative assessment of collateral flow during coronary balloon occlusion was performed before documenting major as well as minor (i.e., stable angina pectoris) adverse cardiac events.
Aside from the mentioned statistical and technical factors that are likely to have contributed to the fact that three of the 10 cited studies have described no effect or an adverse effect of collaterals on outcome (5–7), biological situations at variance to those chosen in our study design have probably caused contradicting results. Those conditions include selection of the study baseline during acute MI (7,13–15), of patients with chronic (3–6)or without infarction (8,12), focus on collaterals present at the time of (7,13–15)or developing after (3–6)acute MI, or investigation of patients with chronic total occlusions (3–5,12), balloon occluded vessels (8,15), or residual stenoses after thrombolysis (7).
In the setting of acute MI, the rate of subsequent ischemic events during the first months of follow-up is much higher (cardiac mortality of 4% to 9% within six months after MI ), as compared with chronic stable CAD (mortality of 0.5% to 1%/1.8 years of follow-up in our study). Thus, our study design is very conservative with regard to the zero hypothesis that collaterals do not have a protective influence on future cardiac events. Accordingly, the finding that overall major cardiac events and the rate of MI were significantly lower in patients with good versus poor collaterals weighs substantially.
The collateral circulation as indicator of the severity of CAD
Using the model of early infarct artery collateral flow after thrombolysis, confusing study results can be found even within a single investigation. In their recently published study, Nicolau et al. (7)found antegrade flow to be directly, but collateral flow to the same acutely infarcted area to be inversely, associated with survival. With regard to myocardial perfusion, this finding does not make sense, because it must be entirely irrelevant whether an ischemic territory is supplied via native or via collateral vessels as long as it is adequately subtended. The investigation by Nicolau et al. (7)and by Gohlke et al. (6)were both designed in a way that angiographic collateral degree was a marker for the culprit residual lesion severity rather than an independent variable for “bypass flow.” The degree of a coronary stenosis is the strongest (although still quite weak) predictor for the status of the human collateral circulation (9). Thus, a tight residual coronary stenosis after thrombolytic therapy is related to enhanced collateral flow, but also to impaired antegrade flow, the latter of which has been documented to be the main determinant of outcome in the mentioned setting (6).
The collateral circulation as predictor of future cardiac ischemic events
In order to control for covariables such as residual stenosis influencing the study results aside from the factor of interest (i.e., collateral flow), it is mandatory that a model with uniform stenosis severity is used. Such a condition is fulfilled either in the presence of a completely obstructed or entirely patent coronary artery. Using the model of chronic total occlusion after acute MI, Boehrer and coworkers (5)found no advantage of present versus absent angiographic collaterals with regard to future cardiac events (annual cardiac mortality, 4.3% and 5.4%, respectively, mean follow-up duration, 42 months). This result is most likely due to the fact that a majority of the 146 study patients with widely varying collaterals have been compared with a minority of 26 patients without collaterals. In the setting of total coronary occlusion without acute MI, Juilliere et al. (12)and Hansen (3)both found angiographically estimated, well developed collaterals to have a positive effect on future cardiac events. The advantage of the chronic occlusion model (i.e., the ability to investigate exclusively the effect of collateral flow on outcome) has certainly not been met with a study design like ours, whereby the influence on outcome of possibly developing restenosis after PTCA with collateral measurement cannot be excluded. In favor of restenosis as a covariable on adverse events is that practically all of them took place within the first seven months after PTCA (Fig. 3). Unfortunately, angiographic follow-up was not performed in our study to test such a possibility. The fact that angina pectoris occurred more often in the group with well versus poorly developed collaterals (Fig. 4) may also be in favor of restenosis as a cofactor of outcome aside from collateral flow; high collateral flow obtained during PTCA representing a risk factor for restenosis is in accordance with the mentioned considerations (16). However, despite all these arguments, it was the group with low collateral flow (linked to low degree stenotic lesion severity) that showed the higher frequency of adverse cardiac events and not vice versa. Thus, the increased frequency of angina pectoris in the group with CFI ≥0.25 may be interpreted as an indicator of enhanced collateral growth as it has been described previously (17).
Aside from the limitations of our study alluded to above, the completeness of follow-up is one that has to be considered. In 43 of 446 potentially eligible patients, follow-up was lost. These patients were not included into the study. Follow-up by direct interview of the patient was not possible in 77 of the 403 patients (19%; 8% in the group with CFI ≥ 0.25, respectively, 11% in the group with CFI < 0.25; p = NS). Consequently, the attending physician was contacted who may not have been correctly informed about the current status of the patient. However, he would have most likely known about a severe adverse cardiac event occurring to the patient.
Central venous pressure was not measured in most patients with pressure-derived CFI. However, in a population of 161 patients undergoing CFIPmeasurements, CVP was also obtained. In these patients, CVP was equal to 6.5 ± 3.8 mm Hg (range, 1 to 22 mm Hg). The standard error of estimate relative to average CFI was 7.7% (standard error of estimate, 0.016). Therefore, the error of calculating CFI by an assumed CVP appears to be rather small with regard to the present, large study population.
Follow-up duration differed significantly among the study groups. However, because it was shorter in the group with an increased rate of major adverse cardiac events, the weight of the finding of a negative effect of low collateral flow in the mentioned group is even more pronounced than it would be with similar observation periods.
☆ Supported by a grant from the Swiss National Science Foundation, grant #32-58945.99 (C.S.).
- coronary artery disease
- collateral flow index
- pressure-derived index of collateral flow
- velocity-derived index of collateral flow
- central venous pressure
- myocardial infarct, myocardial infarction
- mean aortic pressure
- coronary wedge pressure
- percutaneous transluminal coronary angioplasty
- ratio of flow velocity time integral distal to the occluded stenosis
- flow velocity time integral during vessel patency
- Received March 11, 2002.
- Revision received May 14, 2002.
- Accepted June 7, 2002.
- American College of Cardiology Foundation
- Habib G.B.,
- Heibig J.,
- Forman S.A.,
- et al.
- Hirai T.,
- Fujita M.,
- Nakajima H.,
- et al.
- Kodama K.,
- Kusuoka H.,
- Sakai A.,
- et al.
- Pijls N.H.,
- Bech G.J.,
- el Gamal M.I.,
- et al.
- Pohl T.,
- Seiler C.,
- Billinger M.,
- et al.
- Rentrop K.P.,
- Cohen M.,
- Blanke H.,
- Phillips R.A.
- Seiler C.,
- Fleisch M.,
- Garachemani A.,
- Meier B.
- Juilliere Y.,
- Marie P.Y.,
- Danchin N.,
- Karcher G.,
- Bertrand A.,
- Cherrier F.
- Lee C.W.,
- Park S.W.,
- Cho G.Y.,
- et al.
- Wahl A.,
- Billinger M.,
- Fleisch M.,
- Meier B.,
- Seiler C.
- Piek J.J.,
- van Liebergen R.A.,
- Koch K.T.,
- Peters R.J.,
- David G.K.