Objective In patients with acute myocardial infarction (AMI), coronary vasomotor function is impaired in the myocardial territory supplied by the culprit artery and in remote myocardium supplied by angiographically normal vessels. The aim was to investigate the temporal evolution of coronary vasodilatory reserve in patients with AMI by use of [15O]H2O positron emission tomography, after successful percutaneous coronary intervention.
Methods 44 patients with AMI and successful revascularisation by percutaneous coronary intervention were included. Subjects were examined 1 week and 3 months after AMI with [15O]H2O positron emission tomography to assess the coronary flow reserve (CFR). CFR was defined as the ratio of myocardial blood flow (MBF) during hyperaemia and rest. Additionally, 45 age-matched and sex-matched subjects underwent similar scanning procedures and served as controls.
Results At baseline, CFR averaged 1.81±0.66 in infarcted myocardium versus 2.51±0.81 in remote myocardium (p<0.01). In comparison, CFR in the control group averaged 4.16±1.45 (p=0.001 vs both). During follow-up, the CFR increased to 2.74±0.85 in infarcted myocardium (p<0.01), and to 2.85±0.70 in remote myocardium (p<0.01). This was predominantly due to an increase in hyperaemic MBF, from 1.62±0.54 mL/min/g to 2.19±0.68 mL/min/g in infarcted myocardium (p<0.001), and 2.17±0.54 mL/min/g to 2.60±0.65 mL/min/g in remote myocardium (p<0.001).
Conclusions CFR in infarcted and remote myocardium is impaired 1 week after AMI. After 3 months vasomotor function partially recovers. However, as compared with control patients, MBF remains impaired in culprit and reference territories in patients with AMI.
Clinical trial registration NTR3164.
Statistics from Altmetric.com
In patients with acute myocardial infarction (AMI), the coronary flow reserve (CFR) in the infarct territory is reduced immediately after angiographically successful primary percutaneous coronary intervention (PCI). This is caused by an augmented vascular resistance due to vasomotor dysfunction of the coronary resistance vessels, embolisation of thrombus debris and destruction of the microvascular bed distal to the occlusion site. During AMI, non-affected reference coronary arteries are also characterised by a marked reduction in CFR.1 ,2 The vasodilatory capacity of remote myocardium shows potential for recovery in the months following recanalisation of the culprit artery. This points to the presence of additional systemic factors related to AMI that impede the vasomotor function of otherwise unaffected coronary resistance vessels.1 ,3 In addition, structural and functional adaptations of the non-infarcted, remote myocardium after AMI may affect oxidative metabolism, and thus myocardial perfusion as well.4 Cardiac positron emission tomography (PET) using oxygen-15-labelled water ([15O]H2O) is the established gold standard for quantitative myocardial blood flow (MBF) assessment.5 ,6 Recent advances have made it possible to measure perfusion in myocardial layers.7–9 Because myocardial ischaemia occurs primarily in the subendocardial layer of the myocardium before progressing to the subepicardial layer, subendocardial perfusion imaging could provide additional insights in the temporal evolution of vasomotor function across the LV wall.10 The purpose of the present study was to investigate the early temporal evolution of vasomotor function after successfully treated AMI by primary PCI, using [15O]H2O PET to asses MBF and coronary vascular resistance (CVR).
Forty-nine consecutive patients with an acute ST elevation myocardial infarction (STEMI) presenting at the catheterisation laboratory within 6 h after onset of symptoms and treated successfully by primary PCI (ie, Thrombolysis In Myocardial Infarction (TIMI) III flow after coronary stenting) were included in this study. Patients with three-vessel disease and those who were haemodynamically unstable were excluded. Other exclusion criteria were previous myocardial infarction or coronary revascularisation procedures. Patients were examined 1 week and 3 months after the cardiac event with [15O]H2O PET. No adverse events occurred between primary PCI and the follow-up imaging sessions and medication was not modified between scans. Five patients refused follow-up visits for PET and therefore, a total of 44 patients were available for the present analysis. Prior to intervention, all patients received 5000 units of intravenous heparin, 500 mg of intravenous aspirin and 600 mg of clopidogrel orally according to the protocol for acute coronary syndrome. All patients had ST elevation on 12-lead electrocardiography at presentation. Additionally, a control group of 45 patients was selected from a different study protocol11 within comparable age range and with a comparable proportion of men. These patients underwent [15O]H2O PET once at referral for evaluation of coronary artery disease (CAD). None of these patients displayed obstructive CAD at CT or invasive coronary angiography, nor did they display any wall motion abnormalities at echocardiography. For this reason there was no follow-up imaging in the controls. The protocol was in line with the principles of the Declaration of Helsinki.
Patients were instructed to refrain from intake of products containing caffeine or xanthine 24 h prior to the scan. All patients were scanned on a hybrid PET/CT device (Philips Gemini TF 64, Philips Healthcare, Best, The Netherlands). A dynamic PET perfusion scan was performed during resting conditions using 370 MBq of [15O]H2O, which was injected as a 5 mL bolus (0.8 mL/s), followed immediately by a 35 mL saline flush (2 mL/s). A 6 min emission scan was started simultaneously with the administration of [15O]H2O. This dynamic scan sequence was followed immediately by a respiration-averaged slow low dose CT scan to correct for attenuation (30 mAs; rotation time, 0.5 s; pitch, 0.829; collimation, 64×0.625) during normal breathing. After an interval of 10 min to allow for decay of radioactivity an identical PET sequence was performed during hyperaemia induced by an intravenous adenosine infusion (140 µg/kg/min). Adenosine was started 2 min prior to the dynamic PET sequence to achieve maximum vasodilatation and was terminated after the slow low dose CT. Images were reconstructed using the 3D row action maximum likelihood algorithm into 22 frames (1×10, 8×5, 4×10, 2×15, 3×20, 2×30 and 2×60 s), applying all appropriate corrections. Parametrical MBF images were generated and quantitatively analysed using inhouse developed software, Cardiac VUer.12 MBF was expressed in mL/min/g of perfusable tissue. The subendocardial and subepicardial layers were identified automatically by midline delineation between inner and outer myocardial contours, ensuring equal division between the two compartments. Transmural, subendocardial and subepicardial MBFs were calculated for each of the three vascular territories (LAD, left anterior descending; Cx, circumflex; and RCA, right coronary artery). Transmural perfusion gradient (TPG) was defined as the ratio of subendocardial MBF to subepicardial MBF.7 CFR was defined as the ratio between hyperaemic MBF and baseline MBF. Resting CVR was obtained by dividing mean arterial pressure with MBF, while minimal CVR was derived in a similar fashion, but only during infusion of adenosine. During all PET studies the rate-pressure product was monitored. Infarcted myocardium was defined as myocardial tissue perfused by the culprit coronary artery. In case of single vessel disease, remote myocardium was defined as myocardial tissue perfused by the non-culprit coronary artery with the least amount of angiographically detectable disease and in case of two-vessel disease, as myocardial tissue perfused by the non-culprit coronary artery without significant stenosis.
Continuous variables are presented as mean±SD, and categorical data are summarised as frequencies and percentages. Differences in haemodynamics, MBF, TPG and CVR between baseline and follow-up PET studies in the same study group were assessed using paired Student’s t test. Between-group differences in MBF, TPG and CVR, or intraindividual differences between myocardial regions, were assessed using linear regression analysis with correction for the covariates sex, age and smoking history. A p value of <0.05 was considered statistically significant. All statistical analyses were performed using the IBM SPSS software package (IBM SPSS Statistics V.21.0, Chicago, Illinois, USA).
Baseline characteristics for patients and controls are shown in table 1. The infarct-related artery was the LAD artery in 23 patients (52%), the RCA in 16 patients (36%) and the left circumflex (LCx) artery in 5 patients (11%). All patients underwent successful recanalisation of the infarct-related artery, and TIMI-III flow was observed in 41 patients (93%) after PCI.
Haemodynamic parameters during PET studies are listed in table 2 for both study groups. At 3 months after AMI, patients had a significantly lower heart rate at rest, but not during hyperaemia. Furthermore, there were significant increases in systolic blood pressure and diastolic blood pressure, and mean arterial pressure, during rest and hyperaemia. Nonetheless, rate pressure product (RPP) was not significantly altered at 3 months. In the control group, only heart rate and RPP increased between resting and hyperaemic PET scans. Compared with haemodynamic measures of patients with AMI at 1 week, control subjects had a significantly lower heart rate at rest and higher blood pressure and RPP during hyperaemia, whereas at 3 months there were no significant differences in haemodynamics between patients with AMI and control subjects.
MBF values for both study groups are shown in figure 1. In the culprit territory, MBF at rest was significantly reduced from baseline to follow-up (from 0.93±0.22 mL/min/g to 0.83±0.24 mL/min/g, p=0.04), whereas hyperaemic MBF was significantly increased from baseline (1.62±0.54/mL/min/g) to follow-up (2.19±0.68 mL/min/g, p<0.001). In the remote myocardium, there was no significant difference in MBF at rest between baseline (0.91±0.27 mL/min/g) and follow-up (0.94±0.28 mL/min/g, p=0.69), while hyperaemic MBF was significantly increased (from 2.17±0.54 mL/min/g to 2.60±0.65 mL/min/g, p<0.001). As a result, CFR in culprit (from 1.81±0.66 to 2.74±0.85) and remote (from 2.51±0.81 to 2.85±0.70) territories increased significantly 3 months after AMI (p<0.01 for both), but remained significantly reduced compared with controls (p<0.001 for both, figure 2A). The net increase in CFR was larger in culprit territory than in remote territory (0.93±0.85 vs 0.42±0.78 in remote myocardium, p=0.007, figure 2B). In patients with single-vessel disease, there were no regional differences in resting and hyperaemic MBF and CFR in non-infarcted myocardium at baseline, or during follow-up.
Transmural myocardial perfusion gradient
Subendocardial and subepicardial flow values, together with TPG, for both study groups are summarised in table 3. Subendocardial MBF during follow-up was significantly increased compared with baseline. However, this was accompanied by a similar increase in subepicardial MBF, thereby maintaining a TPG of greater than unity. A similar TPG pattern was observed in the remote myocardium and in control subjects. During hyperaemia, subepicardial MBF slightly exceeded subendocardial MBF, resulting in a TPG below unity, except for remote myocardium at follow-up. There was a significant increase in hyperaemic TPG in remote myocardium from baseline to follow-up, which was comparable with controls. Nonetheless, even though the transmural distribution of hyperaemic perfusion returned to normal in remote myocardium during follow-up, the maximal MBF values in subendocardial and subepicardial layers of the culprit territory remained severely depressed.
Coronary vascular resistance
CVR during PET studies for both study groups is listed in table 4. At rest, CVR in the culprit territory was significantly increased from baseline to follow-up, which then was significantly higher than in controls. In contrast, CVR in the remote myocardium remained unchanged between studies and was comparable to controls. During hyperaemia, CVR was significantly higher in the culprit territory compared with remote myocardium. During follow-up, CVR was significantly reduced in culprit and remote myocardium, although it still remained higher in both territories compared with controls.
The main findings of this study are that (1) after AMI, there is a severe reduction in CFR of the infarct-related artery as well as in coronary resistance vessels of remote, non-infarcted myocardium; (2) following successful recanalisation of the culprit artery, CFR of infarcted and remote myocardium increases over time, although no full recovery is observed 3 months after AMI; (3) coronary revascularisation did not affect resting or hyperaemic TPG in the infarcted myocardium, whereas during follow-up hyperaemic TPG in remote myocardium was significantly increased and comparable with that in control subjects.
Myocardial perfusion in infarcted versus remote myocardium
Although resting MBF in the culprit territory was preserved shortly after AMI, a significant reduction was observed during follow-up. This presumably reflects reduced oxygen consumption in the infarcted tissue due to reduced contractility. Metabolic demand and myocardial perfusion during physiological conditions are closely matched by autoregulatory mechanisms of the vascular bed.13 Consequently, CVR was increased in infarcted myocardium, indicating that coronary flow was not limited by coronary perfusion pressure.14 In the remote myocardium, resting MBF and CVR were not altered between baseline and follow-up, suggesting that oxidative metabolism in remote myocardium is not affected by AMI, although substrate metabolism may be altered.4
The severe reduction in CFR in the culprit territory directly after AMI has been documented previously, using a variety of invasive and non-invasive imaging techniques.1 ,2 ,15–17 Although results of early thrombolysis studies are less reliable due to an undefined degree of residual epicardial stenosis,1 ,15 subsequent angioplasty studies revealed that CFR early after successful recanalisation may be severely depressed, despite visual patency of the culprit artery.2 ,4 In the absence of an epicardial coronary stenosis, CVR is primarily determined by vasomotor function of resistance vessels (ie, mainly arterioles) combined with the patency of the capillary bed. Elective angioplasty for obstructive, but stable CAD is also associated with a delayed recovery of CFR despite minimal residual stenosis.3 ,6 Although in the present study the severity of coronary stenosis prior to plaque rupture was unknown, vasomotor function distal from culprit lesion may have already been exhausted in order to maintain coronary flow. Second, coronary intervention by itself leads to release of vasoactive agents that may cause downstream vasoconstriction and augment CVR. Third, placement of a rigid metallic stent alters vessel geometry and may interrupt physiological cellular signalling pathways that impair its response to pharmacological vasodilators.7 ,9 ,18 At the microvascular level, prolonged ischaemia results in inflammation and/or destruction of the capillary bed, and manipulation of the epicardial obstruction may lead to dislodgement of microthrombi and subsequently to occlusion of the capillary bed.5 ,10
The finding of an impaired CFR in remote, non-infarcted myocardium indicates presence of additional factors related to AMI that impede vasomotor function. During AMI, the rise in circulating levels of catecholamines and other vasoactive agents such as angiotensin and vasopressin lead to vasoconstriction.8 ,11 ,19 In addition, an increase in neuronal sympathetic activity due to physical stress and pain may increase α-adrenergic vasoconstriction of the coronary vasculature.
Recovery of myocardial perfusion after PCI
Following percutaneous revascularisation, the increase in CFR was highest in the myocardial territory of the culprit artery, and it could be ascribed mainly to a significant improvement in hyperaemic MBF, and to a lesser degree to a limited decrease in resting MBF. This perfusion pattern is consistent with postischaemic vascular stunning that gradually resolves and is presumably related to regression of the aforementioned factors that augment CVR in the culprit territory, such as recovery of vasomotor function after relief of epicardial obstruction and partial healing of microvascular circulation in the viable myocardium. In the remote myocardium, the increase in CFR could be attributed to an improvement in hyperaemic MBF and was associated with a reduction in CVR as well, albeit to a lesser degree than in the culprit territory.
Nevertheless, follow-up CFR values remained significantly lower compared with control values. This can presumably, for the large part, be attributed to endothelial dysfunction as a result of generalised CAD. It has been previously shown that in patients with non-significant CAD, myocardial perfusion in response to flow-mediated or pharmacological vasodilation is significantly depressed. Indeed, coronary atherosclerosis is a diffuse process that causes a graded fall in perfusion pressure along the entire arterial length, mounting up to 10 mm Hg during pharmacologically induced hyperaemia. These findings imply that endothelial dysfunction alone hampers the physiological response to coronary vasodilatory stimuli, even in the absence of angiographically detectable coronary stenosis. Nonetheless, other factors that may affect MBF, albeit likely to a lesser degree, should not be disregarded. For instance, increased activity of the neurohumoural axis may persist beyond the subacute phase of AMI and attenuate recovery of MBF.20 Second, remodelling of the remote myocardium due to increased load may also induce structural alterations of the coronary vasculature, as seen in patients with arterial hypertension.21
Transmural perfusion gradient
Recent advances in PET technology have made it possible to distinguish between subendocardial and subepicardial MBF.7 Several studies have demonstrated the feasibility of this approach in an experimental setting9 and in human subjects,22 and under a wide range of flow values. In line with these studies, subendocardial perfusion was approximately 15–20% higher relative to the subepicardial layer under resting conditions. LV loading conditions (and consequently oxidative metabolism) are greater in the subendocardial layer of the myocardium, and resting MBF will be augmented relative to the subepicardial layer. Interestingly, this perfusion pattern was maintained in the culprit territory during follow-up, despite a significant reduction in transmural MBF, indicating that vasomotor autoregulation remains intact after AMI in order to match perfusion, and thus TPG, to metabolic demand.
During hyperaemia, TPG dropped below unity in culprit and remote myocardial territories. Contrary to resting conditions, autoregulation becomes exhausted during hyperaemia and myocardial perfusion is determined by the minimal CVR rather than metabolic demand. The more pronounced susceptibility of the subendocardium to myocardial ischaemia has been attributed primarily to local augmented extravascular forces (ie, diastolic perfusion time and end-diastolic wall stress), with an increased subendocardial vascular resistance and physiological drop in TPG as a result. Hence, the greater reduction in TPG and subendocardial flow during hyperaemia in patients with AMI as compared with controls is indicative of an elevated vascular resistance at the subendocardial layer, but affects culprit and remote myocardial territories in a similar fashion. Since TPG was also slightly but significantly decreased during resting conditions, these findings may be explained by an elevation in LV loading conditions and wall stress due to a transient depression in global myocardial contractility and increased LV end-diastolic volumes. During follow-up, hyperaemic TPG was increased only in remote myocardium, whereas TPG in the culprit territory remained unaltered. This apparent discrepancy between the different myocardial regions may be related to wall thinning of the infarcted tissue.
Reproducibility of PET measurements
[15O]H2O PET measurements of resting and hyperaemic MBF have been shown to have good reproducibility (approximately 10–13%).23 To further minimise incertitude in measurements, a significant part of the analyses, in particular generation of quantitative parametrical perfusion images and placement of regions of interest, were performed in an automated fashion as described earlier.12 For transmural MBF measurements, Rimoldi et al have shown that PET subendocardial and subepicardial perfusion are in fairly good agreement with microsphere values, over a wide range of MBF (0.30–4.46 mL/min/g).9 Furthermore, intraobserver and interobserver readings show excellent intraclass correlation coefficients for TPG, with ICCs of 0.93 and 0.89, respectively.7 Hence, methodology and observer-dependency are expected to have limited effects on PET measurements.
The observation of a transient vasomotor impairment in culprit and remote myocardium may have several clinical implications. First, a satisfactory angiographic result (ie, TIMI III flow) does not indicate the immediate return of ‘normal’ coronary circulatory function, because it may take up to several months before vasomotor function has fully recovered. Microvascular dysfunction after AMI has been shown to be of prognostic importance after AMI.24 ,25 Therefore, secondary preventive medical therapy may be aimed at a normalisation of coronary microvascular function to potentially improve cardiovascular outcome. For statins and ACE inhibitors a beneficial effect on CFR has been shown to improve CFR.26 ,27 The effects of thienopyridines on microvascular function are currently being investigated,28 but at this moment, no substantiated statement can be made about this relationship.
Furthermore, an enhanced response of resistance vessels in remote myocardium due to neurohormonal vasoconstrictor stimuli could augment the extent of ischaemia at the border zone of the infarct by reducing collateral flow to the infarct-related arterial bed.29 Finally, although not covered by the scope of this article, it should be noted that the results of myocardial stress testing early after AMI may be affected by delayed recovery of vasomotor function in culprit and remote myocardium.
Several limitations should be taken into account. Although transmural MBF measurements using quantitative [15O]H2O PET have been validated extensively, elaborate model-based spillover and partial volume corrections are incorporated to achieve this goal. Nonetheless, estimated flow differences between myocardial layers are underestimated to a limited degree and reasonable scatter is present in comparison with microspheres to detect the TPG.9 In the patient group, the majority of subjects were smokers, whereas only a minority of control subjects had a smoking history. Since smoking significantly reduces CFR, a certain degree of covariate bias may have been introduced, even though statistical correction was performed.30
The CFR in infarcted and remote myocardium is impaired 1 week after AMI. After 3 months vasomotor function partially recovers. However, as compared with control patients, MBF remains impaired in culprit and reference territories in patients with AMI.
What is already known on this subject?
In patients with acute myocardial infarction (AMI), coronary vasomotor function is impaired in the myocardial territory supplied by the culprit artery and in remote myocardium supplied by angiographically normal vessels.
What might this study add?
Coronary flow reserve in infarcted and remote myocardium is impaired 1 week after AMI. After 3 months vasomotor function partially recovers. However, as compared with control patients, myocardial blood flow remains impaired in culprit and reference territories in patients with AMI.
How might this impact on clinical practice?
First, a satisfactory angiographic result does not indicate the immediate return of ‘normal’ coronary circulatory function, because it may take up to several weeks or months before vasomotor function has fully recovered. Second, an enhanced response of resistance vessels in remote myocardium could augment the extent of ischaemia at the border zone of the infarct by reducing collateral flow to the infarct-related arterial bed.
The authors thank Judith van Es, Robin Hemminga, Amina Elouahmani and Nghi Pham for performing the scans and Kevin Takkenkamp, Henri Greuter and Robert Schuit for producing the [15O]-labelled tracers.
Contributors PFT: conceived study protocol, collected and analysed data, drafted the and critically reviewed the manuscript. SAJT: analysed data, drafted and critically reviewed the manuscript. ID: collected data and critically reviewed the manuscript. GAdW: collected data and critically reviewed the manuscript. PMvdV: designed and revised statistical methods, critically reviewed the manuscript. PGR: served as scientific advisor, critically reviewed the manuscript. AAL: served as scientific advisor, critically reviewed the manuscript. ACVR: served as scientific advisor, critically reviewed the manuscript. NvR: conceived study protocol, served as scientific advisor and critically reviewed the manuscript. PK: conceived study protocol, served as scientific advisor and critically reviewed the manuscript.
Competing interests None declared.
Patient consent Obtained.
Ethics approval Medical Ethics Review Committee of the VU University Medical Center, Amsterdam, the Netherlands.
Provenance and peer review Not commissioned; externally peer reviewed.
If you wish to reuse any or all of this article please use the link below which will take you to the Copyright Clearance Center’s RightsLink service. You will be able to get a quick price and instant permission to reuse the content in many different ways.