OUP user menu

Downstream resistance effects of intracoronary thrombosis in the stenosed canine coronary artery

M Mansaray, P.R Belcher, I Vergroesen, Z.M Wright, J.W Hynd, A.J Drake-Holland, M.I.M Noble
DOI: http://dx.doi.org/10.1016/S0008-6363(98)00270-3 193-200 First published online: 1 April 1999

Abstract

Objective: The presence is well established in unstable angina of intracoronary thrombosis in a stenosed epicardial coronary artery. The effects of the thrombus formation on the distal microcirculation are however still unclear. Methods: We adapted the Folts canine model of left circumflex coronary arterial stenosis and intracoronary thrombosis by the insertion of a pressure catheter distal to the stenosis and by the use of 15 μm radioactive microspheres for measurement of regional myocardial blood flow. This permitted measurement during circumflex artery occlusion of collateral flow, downstream vascular resistance and collateral resistance. Results: Distal circumflex resistance, obtained by dividing the distal circumflex coronary pressure gradient by the collateral flow, significantly increased with thrombosis (94.47±35.72 to 120.06±34.47; p=0.0018) mmHg/ml/min/g. Changes in collateral flow and resistance in the presence of thrombosis, during maximum ischaemic vasodilatation, were inconsistent. Conclusion: Thrombosis causes increased vascular resistance in the microcirculation distal to the site of injury. This may be of clinical relevance in unstable angina, characterised by episodes of thrombus growth and embolization, in which ischaemic episodes may be worsened by generalised downstream vascular changes.

Keywords
  • Regional blood flow
  • Platelets
  • Thrombosis

Time for primary review 25 days.

1 Introduction

The Folts model of platelet-rich intra-coronary thrombosis has been widely accepted as an accurate animal model of human unstable angina [1]. The combination of severe, concentric stenosis and intimal damage to a section of the canine coronary artery (in our case, the circumflex branch of the left coronary artery) in this model results in characteristic repeatable cyclic flow reductions (CFRs) [2]. These CFRs are attributed to the dynamic process of growth and embolization of platelet-rich thrombus, [3], the rate of flow decline being directly proportional to the rate at which platelets accumulate in the narrowed lumen.

In a previous study [4], we introduced a pressure line distal to the stenosis in order to measure cyclic changes in stenosis resistance. During these experiments, we noticed differences in the relationship of this distal pressure and its related flow between those animals with intra-luminal coronary thrombosis and those with undamaged arteries in which CFRs were simulated by coronary occlusion and release. These experiments suggested an increase in the distal coronary pressure/flow ratio in the presence of thrombosis, compared to the absence of thrombosis, and led us to postulate that the presence of thrombus caused a downstream increase in coronary vascular resistance. However, these data were flawed because of the failure to account for the collateral contribution to this distal pressure/flow relationship. The present study addresses this problem by measuring collateral flow.

2 Methods

2.1 Study design

Our aim was to measure the resistance of the peripheral vascular bed distal to the stenosis/thrombosis. This required measurement of the pressure gradient (in this case, left circumflex coronary (LCx) arterial pressure distal to the stenosis/thrombosis minus coronary venous pressure) divided by flow. Measurement of the latter presents a problem because, during cyclic flow reductions, the collateral flow entering the distal bed from the anterior descending branch of the left coronary artery (LAD) and from the right coronary artery (RCA) will be varying as the LCx flow varies.

In the presence of some direct flow through the stenosis, there is no method for measuring the collateral flow. The established method for measuring collateral flow is the radioactive microsphere method, but this demands a steady state and is only applicable during coronary arterial occlusion. Therefore, we interpolated LCx occlusions, waited until there was maximum vasodilatation, as indicated by distal pressure reaching a plateau, and then injected radioactive microspheres into the left atrium. Consequently, all the blood flow to the peripheral LCx vascular bed was the measured collateral flow and this could be divided into the pressure gradient.

A comparison was made during such LCx occlusion between a control period (in the presence of a stenosis only, in the proximal LCx segment) and a period with a stenosis plus thrombosis. The presence of the latter was indicated by the appearance of cyclic flow reductions (Fig. 1).

Fig. 1

Representative trace from one experiment showing spontaneous cyclic flow reductions in the circumflex coronary artery during the thrombotic phase. Sudden increases in circumflex coronary flow corresponding to spontaneous embolisation of platelet-rich thrombus.

2.2 Surgical preparation

Beagles of both sexes (n=9), weighing between 14.0–15.5 kg, were used in a slight modification of the standard Folts’ model described below. The investigation was in accordance with the Guide for the Care and Use of Laboratory Animals by the US National Institutes of Health (NIH Publication No. 85-23, revised 1996).

The dogs were anaesthetised with intravenous sodium pentobarbitone (20 mg/kg) after premedication with acepromazine (0.2 mg/kg). Anaesthesia was maintained with half hourly additions of sodium pentobarbitone (3 mg/kg). Respiration was supported mechanically using a Manley volume cycled ventilator with a 1:2 NO2/O2 mixture, monitored and adjusted with reference to blood gas measurements and the end-tidal CO2.

A coronary sinus catheter (USCI, size 7.5) was radiographically screened into position via the jugular vein, its position verified by blood gas measurements. The right femoral artery was cannulated and, together with the coronary sinus catheter, connected via heparinised fluid-filled lines to Statham P23b transducers. These provided continuous records of systemic blood pressure and coronary venous pressure. A femoral vein was cannulated for drug infusion. A standard three-limb electrocardiogram was recorded continuously.

A left thoracotomy was performed through the fifth interspace and the pericardium opened and sutured to the wound edges, providing an effective cradle for the heart. The circumflex (LCx) was carefully dissected free from surrounding myocardium, and a 15–20 mm segment of each vessel was isolated, enabling a perivascular ultrasonic flow probe (Transonic Systems Inc., USA) to be proximally placed around the artery (Fig. 2). Proper ultrasonic contact between probe and vessel was achieved with the aid of an inert acoustic couplant (N1181 Superabsorbant, Nalco Chemical Company, IL, USA).

Fig. 2

Diagram of experimental preparation.

A 4-mm-long cylindrical polythene constrictor with an internal diameter of 1 mm was situated around the LCx, distal to the flow probe. Just distal to the constrictor, the artery was encircled by a mersilene thread snare for the purpose of obtaining transient coronary occlusions. The critical nature of the stenosis was confirmed by a blunted reactive hyperaemic response to a transient coronary artery occlusion. Fine adjustments to the stenosis could be further obtained by use of a tapered nylon fishing line (Fig. 2).

A 22-gauge teflon cannula (Abbocath, Abbott Ireland Ltd.) was inserted into the LCx about 10 mm beyond the mersilene thread snare and faced distally. It was connected by a heparinised fluid-filled line to a Statham P23b transducer, enabling continuous display and recording of mean and phasic distal coronary artery pressure (Fig. 2). All flow and pressure signals were passed through purpose-built identical low pass filters to ensure simultaneity of recording [4]. They were displayed and continuously recorded on a Gould ES2000 eight-channel electrostatic paper recorder.

2.3 Experimental protocol

Stable heart rate, blood pressure and coronary flow were observed in all animals for about 20 min. A brief 15 s occlusion of the LCx was then performed to verify the presence of a critical stenosis from the reactive hyperaemic response. After pressure and flow had returned to baseline values, a more prolonged occlusion (70 s) was performed; when distal LCx arterial pressure had fallen to a plateau value, maximum ‘ischaemic’ vasodilatation was assumed to be present, and regional and collateral blood flow measurements were obtained by the standard microsphere method [5]. The occlusion was released following completion of the reference flow collection. The whole procedure was repeated 1 h later, after crushing the artery beneath the stenosis to induce cyclic flow reductions (CFRs, Fig. 1). During this latter period, haemodynamic variables associated with the LCx were continuously varying due to the formation, growth and embolization of thrombus at the stenosis, which causes the CFRs. The overall status of the preparation, as judged by arterial pressure and heart rate, remained stable.

2.4 Measurement of collateral blood flow

The occlusion of the LCx, which always accompanied microsphere injection, allowed an estimation of collateral flow during maximum ‘ischaemic’ vasodilatation of the distal coronary bed. This provides a relatively ‘steady’ state necessary for the validity of this method. In this study, two microsphere measurements were made. The first set of radiolabelled microspheres was used to measure flow in the undamaged LCx, in the absence of thrombus (the non-thrombotic phase). The second set of radioactive spheres was used after the area of the LCx encircled by the constrictor was crushed by a haemostat, thus producing endothelial damage. This was confirmed by cyclic reductions in flow, indicating thrombus growth and embolization in the damaged artery, i.e. the thrombotic phase (Fig. 1).

Bolus injections of 15.5±0.1 μm NEN-TRAC™ radioactive microspheres (Du Pont De Nemours, Belgium) were made via a silastic catheter into the left atrium. Each injection was performed during withdrawal of a reference blood sample from the femoral artery at a constant measured flow rate. The microsphere protocol involved starting a timed collection of a reference sample of blood from the femoral artery at a flow rate of 13 ml/min, about 10 s before injection of between one and two million microspheres labelled with the isotopes 141Ce, 103Ru or 95Nb (chosen at random) into the left atrium. The atrial catheter was flushed with 15–20 ml of saline, to remove all residual radioactivity, and sample collection was completed over a total period of 70 s, during which the LCx was occluded.

On completion of the protocol, myocardium supplied by the LCx was defined. The LCx was occluded with the mersilene thread snare and about 15–20 ml of 15 μm yellow fluorescent zinc cadmium sulphide particles (Duke Scientific, Palo Alto, CA, USA), suspended in saline, were injected into the coronary artery cannula. Care was taken to ensure that no back flow occurred during the injection of the fluorescent spheres. The injection of the zinc cadmium particles allowed demarcation of the area of myocardium supplied by the occluded LCx post-mortem. The barbiturate anaesthesia was then deepened until the heart fibrillated. The heart was removed, the surface wiped clean of blood and the cavities flushed with saline. The atria were removed and the ventricles were fixed in a 10% formalin solution at room temperature for at least three–four days prior to analysis.

2.5 ‘Time control’ experiments

Since it was always necessary to damage the endothelium of the LCx artery in order to induce thrombosis, it was not possible to randomise the order of control and thrombotic phases in our paired studies. To exclude the possibility that any observed changes between the phases was time-dependent (i.e. due to deterioration of the preparation), time control experiments were performed in three dogs (mean weight, 15 kg) in which two microsphere injections were made during LCx occlusion. We followed the exact protocol just described, with two phases (1 and 2), but with two exceptions: (i) No stenosis on the LCx artery and (ii) no crushing/damage of the LCx artery. Phase 1 in the ‘time control’ experiments corresponded to the non-thrombotic phase of the other experiments, and phase 2 to the thrombotic phase. All measurements in these ‘time control’ experiments were thus made in animals with a patent, undamaged LCx artery prior to the occlusions.

2.6 Tissue preparation

The left ventricle was sliced into 4 mm transverse sections from apex to base, yielding between six and eight slices per heart. The slices of tissue were then placed between perspex sheets that were kept exactly 4 mm apart by spacers between the sheets. The anterior and posterior surfaces of each slice were then examined under ultra-violet light in order to accurately delineate the area supplied by the LCx shown by yellow fluorescence, from other areas of myocardium. All slices had areas of yellow fluorescence.

Precise heart maps made by tracings on acetate paper were obtained for each slice. The different areas of each slice were then separated, cut into pieces, weighed and placed in plastic tubes ready for radioactive counting. Each tube contained at least 0.5 g of tissue [6]. The radioactivity of the heart and reference blood samples was measured simultaneously in a Packard Autogamma Counter (model 5002/3). Window settings were optimised not only to include the main energy peaks of each isotope but to minimise overlap.

Myocardial blood flow values were calculated from the reference and myocardial sample counts after corrections for spillover as follows:-

Myocardial blood flow=myocardial counts×(reference flow/reference sample counts). Myocardial flow in each sample was measured in absolute terms and as ml/min/g of myocardium.

2.7 Resistance calculations

During LCx occlusion, the pressure gradient across the LCx bed is given by the difference between the pressure distal to the occluded artery (POCC) and the coronary venous pressure (PV). ‘True’ distal resistance (RD) of the maximal ischaemically dilated LCx bed was obtained by dividing this distal circumflex coronary pressure gradient by the collateral flow (Qc) obtained during LCx occlusion. Embedded Image(1) Collateral resistance (RCOLL) was obtained by dividing the pressure gradient across the collateral vessels by the collateral flow. Embedded Image(2)

2.8 Statistical methods

The microsphere method of measurement of regional blood flow has consistently shown heterogeneity of blood flow normalised for sample mass between samples in any chosen region of ventricle [7]. This was quantified by determination of the mean, variance and standard deviation of flow, in ml/min/g tissue, for the stained samples supplied by the LCx and the unstained samples supplied by the LAD and the right coronary artery. The mean flow to the LAD region, in ml/min/g tissue (Table 1), was taken to be the normal blood flow. The distribution in control (no thrombus) to experimental (thrombus) phase for the regions of interest were compared by Student’s paired t-test (n=number of tissue samples per region per heart). Differences between flows and resistances measured in the control and thrombosis periods were analysed by Student’s paired t-test (n=9=number of experiments). Group data were described by the mean±SD of each variable per animal (n=9). All statistical calculations were carried out using In Stat software (Graph Pad Inc, San Diego, CA, USA) on a MacIntosh computer.

View this table:
Table 1

Paired measurements of key variables in thrombosis experiments (left) and time controls (right)

VariableStenosis and thrombosisTime controls
ControlThrombosisProbabilityPhase 1Phase 2Probability
RD94.47±35.72120.06±34.470.001877.03±28.2973.70±18.300.6700
RCOLL491.66±352.40495.57±338.140.8976377.93±242.80387.90±250.610.2842
Qc0.2731±0.19630.2152±0.11610.12100.2933±0.14360.2767±0.12700.4226
PA114.18±10.442108.42±13.3440.2427109.00±1.00102.67±6.130.1611
HR168.33±21.74168.56±16.910.9391154.33±13.65163.00±23.580.2892
POCC30.989±13.14133.044±11.8310.072827.067±4.87526.433±6.0010.4602
PV9.0656±3.54379.4244±3.97470.25226.9333±2.40077.0667±2.57940.3828
QLAD0.743±0.2410.628±0.0180.06130.787±0.0970.840±0.0520.5949
  • RD, vascular resistance in myocardium distal to the occluded left circumflex coronary artery in mmHg/ml/min/g of tissue; RCOLL, resistance of the collateral vessels, in mmHg/ml/min/g of tissue, in the left circumflex coronary artery region of supply; Qc, collateral blood flow, in ml/min/g of tissue, in the left circumflex coronary artery region of supply; PA, mean arterial pressure during radioactive microsphere injection, in mmHg; HR, heart rate, in beats/min; POCC, pressure distal to the occluded left circumflex coronary artery, in mmHg; Pv, coronary sinus pressure, in mmHg; QLAD, normal tissue blood flow to the myocardium supplied by the left anterior descending branch of the left coronary artery, in ml/min/g.

2.9 Limitations of the model

The experimental preparation is a compromise between what we wanted to measure, which was the collateral flow entering the distal bed, which varies as the LCx flow varies, and the practical reality of available collateral flow measuring techniques. In the presence of some direct flow through the stenosis, there is no method for measuring the collateral flow. The established method for measuring collateral flow is the microsphere method, but this demands a steady state and, thus, is only applicable during coronary arterial occlusion. The fact that thrombosis affects coronary flow dynamics means that these cannot be steady over the whole time course of the thrombosis experiment; their stability can only be judged in time control experiments.

There is some controversy with regards to what back pressure to use in calculating coronary vascular resistance [8–10]. Several investigators have suggested that left ventricular end-diastolic pressure determines myocardial perfusion in the maximally vasodilated heart when it exceeds values of coronary venous pressure [11, 12]. However, Spaan [10]effectively argued that this may apply only when the left ventricular end-diastolic pressure is higher than right atrial or pericardial pressure. These conditions are very unlikely to apply to our preparation and, thus, the coronary sinus pressure is, in our opinion, correctly taken [13]as the effective backflow pressure in our calculation of resistance, as indicated in Eq. (1).

3 Results

3.1 Haemodynamics

During the control (no thrombus) and thrombotic phases, there were no significant differences in the standard haemodynamic indices of mean arterial pressure and heart rate (Table 1). Prior to circumflex coronary occlusions, distal coronary pressure and coronary flow were constant in the pre-thrombotic phase and fluctuated with the cyclic flow variations in the thrombotic phase.

3.2 Heterogeneity

Examples of the spread of values of tissue flow from counted piece to piece, in ml/min/g, are shown in (Fig. 3) for the circumflex region during arterial occlusion. In this example, the control coefficient of variation was 62.5% in the occluded LCx region. However, the comparison between the two microsphere measurements in each piece was reasonably consistent, i.e. high or low flows in both measurements of a particular tissue sample (Fig. 3).

Fig. 3

In one experiment, the distribution of myocardial blood flow among individual myocardial samples in the occluded left circumflex coronary artery region of supply (yellow fluorescent stain) is illustrated in the control period (left) and the thrombosis period (right). The columns are means with standard deviation error bars. The lines connect the pairs of blood flow measurements of each individual sample. p=probability of no difference in flow for the population of samples by Student’s paired t-test.

3.3 Effects of thrombosis on collateral flow and resistance

Collateral flow (Qc) obtained with radioactive microspheres during a 70-s occlusion of the circumflex artery, i.e. maximal ischaemic vasodilatation of the distal circumflex bed, showed no significant (paired t-test) changes in the presence of thrombus (Table 1). Collateral resistance (mmHg/ml/min/g), calculated from Eq. (2), also showed no consistent change with thrombosis (Fig. 4).

Fig. 4

For all experiments (n=9), the resistance of the collateral vessels was calculated from the pressure gradient from the aorta to the left circumflex coronary artery distal to its occlusion, divided by collateral flow. Left, control period; right, thrombosis period. The columns are means with standard deviation error bars. The lines connect the pairs of resistance measurements of each individual experiment. p=probability of no difference in resistance for all experiments by Student’s paired t-test.

3.4 Effects of thrombosis on downstream pressure and resistance

A positive distal pressure at zero flow (POCC) was always obtained during occlusion of the circumflex artery. This pressure was not significantly altered by the presence of thrombosis in a paired statistical test (Table 1). The major aim at the outset of this study was a measurement of ‘true’ distal resistance, incorporating coronary venous pressure measurements made by insertion of a catheter in the coronary sinus. During occlusion of the circumflex artery, this value can be obtained from Eq. (1). Resistance in the distal circumflex bed during maximal ischaemic vasodilatation (mmHg/ml/min/g) showed a consistent rise with thrombus in all nine experiments (Fig. 5 and Table 1). However, the components of the resistance ratio (pressure gradient and flow) did not always change in the same direction and were therefore not statistically significantly affected in paired t-tests.

Fig. 5

For all experiments (n=9), the resistance of the vascular bed in the myocardium distal to the occluded left circumflex coronary artery is shown. This was calculated from the pressure gradient between the left circumflex coronary artery distal to its occlusion and coronary sinus, divided by collateral flow. Left, control period; right, thrombosis period. The columns are means with standard deviation error bars. The lines connect the pairs of resistance measurements of each individual experiment. p=probability of no difference in resistance for all experiments by Student’s paired t-test.

3.5 Time control experiments

No significant change was observed in mean arterial pressure and heart rate in all three time control experiments (Table 1). Coronary haemodynamic variables were also constant, except during the coronary occlusions. No changes between phases one and two, were seen in collateral flow obtained with radioactive microspheres during maximum ischaemic vasodilation of the distal bed (Table 1). Distal resistance and collateral resistance, calculated as previously mentioned, showed no changes between phases one and two (Table 1).

A power calculation showed that any increase in the value of n, from n=3 to any practical value, would not lead to a statistically significant difference in distal circumflex resistance between the simulated thrombus and non-thrombus phases. For ethical reasons, we therefore refrained from further experiments.

4 Discussion

Our study shows that the presence of thrombosis in the critically stenosed left circumflex coronary artery caused an increased resistance in the distal circumflex bed during occlusion of the circumflex coronary artery (maximal ischaemic vasodilation).

We measured the ‘true’ distal resistance, downstream of a growing thrombus. By definition, resistance across a vascular bed is calculated from the pressure drop across the bed, divided by the flow through, which is affected by several factors [14]. We decided to reduce these influences to an absolute minimum by making measurements during maximum vasodilatation. The consistent and, therefore, statistically significant increase in distal resistance was sometimes due to a dominant increase in pressure gradient and sometimes due to a dominant decrease in flow. Thus, the changes in pressure gradient and flow were not consistent and, therefore, not statistically significant.

The pathophysiological processes implicated in the formation of a dynamic mural thrombus at the site of a critical stenosis have been described extensively [15–18]and are known to involve platelet activation and aggregation with associated release of platelet products that mediate an increase in vessel tone [17–20]. It has been shown previously that intracoronary platelet deposition can cause vasoconstriction of large epicardial coronary arteries, and that this is largely mediated by release of thromboxane A and serotonin from activated platelets [21, 22].

Our objective was to determine if these processes in the proximal epicardial artery caused any downstream effect beyond the thrombus which, extrapolated to human unstable angina, would imply an additional pro-ischaemic effect over and above that of the stenosis itself; this does appear to be the case.

There is also the potential for a thrombus to embolise and pass downstream into the segment of myocardium supplied by that artery [23–25]. The literature provides considerable evidence in favour of this embolic hypothesis. There exists evidence for a platelet-rich thrombus at the site of a stenosis, embolising, forming microthrombi that can be detected downstream [26]. Indeed, the cyclic flow reductions first described by Folts et al. [2]are due to cyclical, sometimes spontaneous, episodes of thrombus growth and embolization [3]. Haerem [27]observed platelet aggregates in intramyocardial vessels of patients who had died suddenly of coronary disease and postulated a plugging of the distal microvasculature, resulting in a hampering of intramyocardial blood flow and myocardial injury. Falk [23]showed a significant increase in the incidence of intramyocardial platelet emboli in patients with unstable angina who died suddenly, compared to patients without unstable angina. It must be stated however that these histological studies do not provide conclusive proof for plugging of the distal microcirculation causing the observed increase in distal resistance in these experiments.

Platelet emboli downstream of a mural thrombus may also release vasoconstrictor substances, including serotonin and thromboxane A2 (TXA2). Lam et al. [28]have previously shown a rough correlation between the occurrence and extent of vasoconstriction with the intensity of platelet deposition in porcine carotid arteries after balloon angioplasty. In addition, endothelial injury and subsequent platelet-rich thrombosis in a canine model have recently been shown to be associated with increased local production of the vasoconstrictor peptide endothelin-1 (ET-1) [29]. In humans and other species, ET-1, at pathological concentrations, is a potent constrictor of conductance and resistance coronary arteries [30–32]. Measurements of levels of vasoactive substances in plasma during thrombosis indeed provide supportive evidence that elevated levels of serotonin, TXA2 and other platelet substances may promote a vasoconstrictive milieu. Experimental administration of antiplatelet or antithombotic agents eliminate these phenomena [1], but cannot differentiate between mechanical obstruction of the vessels and the vasoactive effects of the platelet-rich thrombi.

In order to eliminate the possibility of deterioration of the experimental preparation introducing error between the first and second set of measurements, we performed three time control experiments. Continuous circumflex flow and pressure were recorded in the absence of stenosis or endothelial damage to document the stability of the preparation. These studies demonstrated no cyclic flow reductions, allowing us to conclude that our experimental observations represent true flow variations induced by platelet-rich thrombi. Our study provides data on the acute effects of thrombosis on collateral flow and resistance. Collateral flow measured by 15 μm radioactive microspheres and collateral resistance showed no change from the control to the thrombotic period. This is the expected result as the collateral vessels from the LAD and RCA to the LCx are presumably not in a stream of emboli or a flux of potentially vasoconstrictive substances, as are the vessels of the distal LCx bed. The fact that collateral flow does not change in the presence of downstream vasoconstriction suggests that collateral vessels are independently controlled [33, 34].

Our findings suggest that platelet-rich thrombus growth and embolisation may contribute to distal flow disturbances in unstable coronary syndromes, thereby adding to the effects of epicardial vessel narrowing. Targeting these effects may improve the treatment of these syndromes.

Acknowledgements

This study was supported by a grant from the British Heart Foundation.

Footnotes

  • 1 See pages 6–8.

References

View Abstract