Author + information
- Received June 23, 2014
- Revision received August 11, 2014
- Accepted September 11, 2014
- Published online February 1, 2015.
- Liam M. McCormick, MBBS∗,
- Stephen P. Hoole, MA, DM†,
- Paul A. White, PhD‡,
- Philip A. Read, MA, MD†,
- Richard G. Axell, MSc‡,
- Sophie J. Clarke, BSc∗,
- Michael O’Sullivan, PhD†,
- Nick E.J. West, MA, MD† and
- David P. Dutka, MA, DM∗∗ ()
- ∗Department of Cardiovascular Medicine, University of Cambridge, Cambridge, United Kingdom
- †Department of Interventional Cardiology, Papworth Hospital, Cambridge, United Kingdom
- ‡Department of Medical Physics and Clinical Engineering, Cambridge University Hospital National Health Service Foundation Trust, Cambridge, United Kingdom
- ↵∗Reprint requests and correspondence:
Dr. David P. Dutka, Department of Cardiovascular Medicine, ACCI Level 6, Box 110, Addenbrooke’s Hospital, Hills Rd, Cambridge CB2 0QQ, United Kingdom.
Objectives This study sought to determine whether pre-treatment with intravenous glucagon-like peptide-1 (GLP-1)(7-36) amide could alter myocardial glucose use and protect the heart against ischemic left ventricular (LV) dysfunction during percutaneous coronary intervention.
Background GLP-1 has been shown to have favorable cardioprotective effects, but its mechanisms of action remain unclear.
Methods Twenty patients with preserved LV function and single-vessel left anterior descending coronary artery disease undergoing elective percutaneous coronary intervention were studied. A conductance catheter was placed into the LV, and pressure-volume loops were recorded at baseline, during 1-min low-pressure balloon occlusion (BO), and at 30-min recovery. Patients were randomized to receive an infusion of either GLP-1(7-36) amide at 1.2 pmol/kg/min or saline immediately after baseline measurements. Simultaneous coronary artery and coronary sinus blood sampling was performed at baseline and after BO to assess transmyocardial glucose concentration gradients.
Results BO caused both ischemic LV dysfunction and stunning in the control group but not in the GLP-1 group. Compared with control subjects, the GLP-1 group had a smaller reduction in LV performance during BO (delta dP/dTmax, –4.3 vs. –19.0%, p = 0.02; delta stroke volume, –7.8 vs. –26.4%, p = 0.05), and improved LV performance at 30-min recovery. There was no difference in transmyocardial glucose concentration gradients between the 2 groups.
Conclusions Pre-treatment with GLP-1(7-36) amide protects the heart against ischemic LV dysfunction and improves the recovery of function during reperfusion. This occurs without a detected change in myocardial glucose extraction and may indicate a mechanism of action independent of an effect on cardiac substrate use. (Effect of Glucgon-Like-Peptide-1 [GLP-1] on Left Ventricular Function During Percutaneous Coronary Intervention [PCI]; ISRCTN77442023)
- glucagon-like peptide-1
- myocardial ischemia
- myocardial stunning
- myocardial substrate use
Glucagon-like peptide-1 (GLP-1) is an incretin hormone, which regulates carbohydrate metabolism (1). As well as its effects on glucose homeostasis, the identification of GLP-1 binding sites in the heart has generated strong interest in a potential role on cardiovascular function (2). In animal models, GLP-1 attenuates ischemia-reperfusion injury (3–6), and clinical studies investigating the potential effects of GLP-1 modulation therapy are starting to emerge (7–12). However, there is uncertainty regarding its mechanisms of action.
The cardioprotective properties of GLP-1 are attractive as a potential therapeutic adjunct to support the heart during the multiple episodes of supply ischemia that occur with balloon inflations during percutaneous coronary intervention (PCI). We have previously used conductance catheter-derived pressure-volume loops to quantify left ventricular (LV) function during PCI, and we have demonstrated that 1-min coronary balloon occlusion (BO) results in late post-ischemic myocardial dysfunction (stunning), with further BO after 30 min resulting in cumulative LV dysfunction (13). This may be particularly important in patients with pre-existing LV impairment and in those with proximal coronary lesions subtending large territories of myocardium.
In 1 of the few physiological studies investigating the cardioprotective efficacy of GLP-1 during supply ischemia in humans, we demonstrated that an intravenous infusion of GLP-1(7-36) amide, when commenced after an initial 1-min coronary BO, reduced ischemic LV dysfunction caused by subsequent BO, and mitigated myocardial stunning (9). However, it is not known whether these favorable cardiovascular effects are dependent on the activation of ischemic pre-conditioning pathways, or whether they are the result of a change in myocardial substrate use. This study was therefore undertaken to provide some mechanistic insights into the cardioprotective properties of GLP-1. Specifically, we sought to determine whether pre-treatment with intravenous GLP-1(7-36) amide, commenced before an ischemic insult, protected the heart during PCI and whether there were changes in the myocardial use of glucose and free fatty acids (FFA).
Patients with preserved LV function (defined as an ejection fraction ≥50% by transthoracic echocardiography) and single-vessel disease awaiting PCI to the left anterior descending artery were invited to participate. Exclusion criteria included a history of myocardial infarction (as defined by the Third Universal Definition of Myocardial Infarction) (14) within the preceding 3 months, and patients with diabetes receiving insulin, dipeptidyl-peptidase 4 inhibitors, or GLP-1 receptor agonists. The study was approved by the local ethics committee (REC number 09/H0311/17) and complied with the guidelines set out in the Declaration of Helsinki. All participants gave written informed consent. The trial number was ISRCTN77442023.
Patients were asked to abstain from consuming caffeine, alcohol, nicotine, as well as nicorandil and oral/sublingual nitrates in the 24 h leading up to the procedure. No other cardiovascular medications were omitted. All subjects fasted for 6 h and received aspirin 300 mg and clopidogrel 300 mg at least 6 h before PCI.
One 7-F sheath was placed in the right femoral artery, 1 6-F sheath in the right femoral vein, and another 6-F sheath in either the right radial or left femoral artery. All patients were anticoagulated with heparin (70 to 100 U/kg as an initial bolus) and the activated coagulation time maintained >250 s throughout the procedure (Figure 1). Simultaneous coronary artery (CA) and coronary sinus (CS) blood samples were taken via a 6-F guide catheter positioned at the ostium of the left main coronary artery and a 6-F multipurpose or Amplatz Left 1 catheter (Johnson & Johnson Medical, Diegum, Belgium) positioned inside the ostium of the CS using fluoroscopic imaging. In all cases, selective engagement of the main CS trunk was verified using retrograde contrast injection and confirmed by measuring oxygen saturations sampled from the catheter to ensure minimal contamination by right atrial blood. No hemodynamic-altering medication was administered during the study. The conductance catheter technique was used to determine pressure-volume relations and provide a beat-to-beat assessment of LV performance. A 7-F 8-electrode conductance catheter (Millar Instruments, Houston, Texas) was inserted through the 7-F arterial sheath, advanced across the aortic valve into the LV apex, and placed along the longitudinal axis of the ventricle. The catheter was connected to a MPVS Ultra signal-conditioning unit (Millar Instruments) in series with an ADInstruments PowerLab 16/30 Series sixteen channel amplifier (ADInstruments, New South Wales, Australia). After measuring blood resistivity, the catheter tip was submersed in a saline bath and the pressure transducer was zeroed.
Conductance catheter calibration
A 20-kHz current was applied to the proximal and distal electrodes and time-varying conductance (Gt) was calculated by measuring the sum of the conductance between the intervening 5 adjacent segments of the 6 central electrodes. G measured by the conductance catheter reflects the sum of the conductance of both blood within the ventricular cavity and parallel conductance (i.e., conductance of the tissues and fluids surrounding the LV). Parallel conductance (GP) was determined using the hypertonic saline injection technique via a multipurpose catheter inserted into the pulmonary artery, as previously described (15). The formula by Baan et al. (15) for obtaining time-varying ventricular volume (V) is as follows: V = (1/α) × ρ × L2 × (Gt – Gp), where α refers to the ratio of the conductance-derived volume to true ventricular volume (calculated by calibration with the Fick-derived measure of cardiac output), ρ is the specific resistance of blood (measured directly using a calibrating cuvette), and L represents the known distance between each pair of electrodes.
After conductance catheter calibration, pressure-volume loops were recorded at baseline (BL). Consenting patients were randomized to receive either an intravenous infusion of normal saline (control group) or GLP-1(7-36) amide (Bachem, Bubendorf, Switzerland) at a dose of 1.2 pmol/kg/min (GLP-1 group) starting immediately after recording of BL pressure-volume loops. Randomization was undertaken by sealed envelope allocation prior to their PCI procedures. The infusions were prepared externally so that the patient and operator were blinded to the treatment strategy (GLP-1 vs. control). GLP-1 has a short half-life of only 1 to 2 min with a rapid onset and offset of action. Pressure-volume loops were recorded again during a 1-min low-pressure (<4 atm) balloon inflation at the site of the coronary stenosis, and once more in recovery after a period of 30 min (Figure 1). After the study measurements were completed, the lesions were treated conventionally by high-pressure balloon angioplasty and stenting, and the infusion of GLP-1 or saline was discontinued at the end of the procedure.
Coronary hemodynamic calibration
A 6-F guide catheter was positioned at the ostium of the left main CA. A 0.014-inch Combo-Wire XT 9500 (Volcano Therapeutics, Inc., San Diego, California) guidewire was used to enable intracoronary pressure measurements during PCI. The measurements were recorded digitally onto a ComboMap (Volcano Therapeutics) console for off-line analysis. Prior to insertion, the wire was calibrated to atmospheric pressure outside of the patient and then advanced via the guiding catheter to the ostium of the left main CA, where aortic and wire-tip pressures were equalized. The tip of the wire was then positioned 3 to 5 cm beyond the stenosis, in a segment of vessel that was straight and free from side branches.
Coronary hemodynamic measurements
The mean pressure distal to the stenosis (Pd) was recorded from an average of 5 beats and compared with the mean aortic pressure (Pa) simultaneously measured at the guiding catheter to derive the coronary stenosis pressure gradient (Pd/Pa). Distal coronary and aortic pressures were then acquired during a low-pressure BO at <4 atm for 1 min. Coronary occlusion was confirmed by contrast injection during balloon inflation. An average of 5 beats was used to calculate mean distal coronary wedge pressure (Pw) and Pa, just prior to balloon deflation. The pressure-derived collateral flow index (CFIp) was calculated as: CFIp = (Pw – Pv)/(Pa – Pv).
LV hemodynamic measurements
The conductance catheter data were analyzed off-line by a reviewer blinded to the treatment strategy using LabChart software (ADInstruments, New South Wales, Australia). Five cardiac cycles were recorded at BL, during ischemia (after 1-min BO just before balloon deflation) and at 30-min recovery. The parameters generated to measure systolic function were stroke volume (SV), ejection fraction (EF), cardiac output (CO), and dP/dTmax (maximum rate of isovolumic pressure increase). The parameters for diastolic function were dP/dTmin (maximum rate of isovolumic pressure decline) and tau (time constant of isovolumic relaxation). To calculate tau, the conductance catheter-derived Pt (time constant of pressure relaxation) is measured from the time of peak rate of pressure decline (dP/dTmin) to 5 mm Hg above end-diastolic pressure. Tau is derived from the monoexponential decay of the pressure waveform: Pt = K e-t/tau; where tau is the slope of the log Pt versus t relation (tau = –1/slope, assuming P∞ = 0).
Blood samples were taken simultaneously from the CA and CS to measure transmyocardial gradients of glucose, insulin, FFA, and GLP-1(7-36) amide at BL and after ischemia (at the end of 1-min BO approximately 1 min after balloon deflation). The GLP-1 samples were drawn into syringes containing dipeptidyl peptidase-4 inhibitor (Millipore, Consett, Durham, United Kingdom) to prevent GLP-1 degradation. Plasma GLP-1(7-36) amide levels were measured using a commercially available assay (Meso Scale Discovery, Rockville, Maryland).
Data are expressed as mean ± SD, unless otherwise stated. LV hemodynamic data were converted to a percentage change from BL to facilitate data comparison. The primary outcome measures were dP/dTmax during BO and at 30-min recovery. The number of subjects had been calculated on the basis of our previous study in 20 patients undergoing PCI to the left anterior descending artery with concomitant LV conductance catheter measurements. In patients receiving an infusion of GLP-1(7-36) amide after initial BO, the delta dP/dTmax (from BL) at 30-min recovery improved from –12.2 ± 9.6% to –1.6 ± 8.5% (9). To detect a change in dP/dTmax of 6% with GLP-1 during BO, 20 patients were required (α = 0.05, β = 0.20). Comparisons between control and GLP-1 groups were made using the unpaired Student t test for continuous variables, or the Mann-Whitney U test where appropriate after testing for normality of distribution using the Shapiro-Wilk test. Categorical data are expressed as numbers (percentages) and compared by use of Fisher exact test. Comparisons within the groups were made using repeated-measures analysis of variance. Two-tailed tests were used on all occasions, and a probability value of ≤0.05 was considered statistically significant.
We recruited 20 patients (10 control and 10 GLP-1) into the study, although 1 of the patients in the GLP-1 group had extensive ventricular ectopy during BO such that the pressure-volume loops obtained were not of sufficient quality to enable accurate assessment of LV function. No other patients were excluded. Therefore, 19 patients were included in the final analysis (10 control and 9 GLP-1). Patient demographic data (Table 1) and BL hemodynamic data (Table 2) were not different between the 2 groups.
GLP-1 Group versus control group at baseline
At BL (before GLP-1 infusion), there were no differences in the plasma levels of glucose, insulin, FFA, or GLP-1(7-36) amide between the GLP-1 and control groups (Table 3).
Baseline versus balloon occlusion
In the control group, there was a reduction in FFA concentrations immediately following BO (compared with BL) (2,043 ± 788 μmol/l vs. 1,274 ± 684 μmol/l, p = 0.03), but there was no change in the plasma concentrations of insulin, glucose, or GLP-1(7-36) amide. In contrast, patients who received GLP-1 had an approximately 40-fold increase in the plasma concentration of GLP-1(7-36) amide after BO (3.04 ± 1.77 vs. 123 ± 34 pg/ml, p = 0.02) as well as a reduction in FFA (1,725 ± 1,022 μmol/l vs. 706 ± 268 μmol/l, p = 0.02) and glucose concentrations (6.22 ± 1.10 mmol/l vs. 5.27 ± 0.87 mmol/l, p = 0.02), but there was no change in the plasma concentration of insulin.
Control group versus GLP-1 group immediately after balloon occlusion
Compared with control subjects, patients receiving GLP-1 immediately after BO had increased plasma concentrations of insulin (47.1 ± 21.5 pmol/l vs. 130 ± 134 pmol/l, p = 0.04) and GLP-1(7-36) amide (1.53 ± 0.98 pmol/l vs. 123 ± 34.3 pg/ml, p < 0.001), reduced concentrations of FFA (1,273 ± 684 μmol/l vs. 706 ± 268 μmol/l, p = 0.02), and a nonsignificant reduction in plasma glucose concentration (5.94 ± 0.55 mmol/l vs. 5.27 ± 0.87 mmol/l, p = 0.08).
Coronary sinus versus coronary artery
In both the control and GLP-1 groups, there were no differences in the CS concentrations of GLP-1(7-36) amide, glucose, insulin, or FFA when compared with simultaneous samples taken from the CA at either BL or during BO. In particular, patients receiving GLP-1(7-36) amide, when compared with control subjects, demonstrated no difference in transmyocardial glucose concentration gradients during BO (Figure 2).
During low-pressure 1-min coronary BO, there were no differences in either Pw (24.4 ± 6.3 mm Hg vs. 20.3 ± 8.2 mm Hg, p = 0.36) or CFIp (0.18 ± 0.05 vs. 0.17 ± 0.07, p = 0.57) between the GLP-1 infusion and control groups.
Left ventricular function
In the control group, BO caused a reduction in both LV systolic (dP/dTmax 1,591 ± 244 mm Hg/s vs. 1,287 ± 282 mm Hg/s, p = 0.02; SV 83.1 ± 19.8 ml vs. 62.0 ± 22.4 ml, p=0.01; EF 62.4 ± 5.30% vs. 46.4 ± 10.7%, p = 0.01; CO 5.10 ± 0.92 l/min vs. 3.90 ± 1.02 l/min, p = 0.048) and diastolic (dP/dTmin –1,953 ± 375 mm Hg/s vs. –1,445 ± 370 mm Hg/s, p = 0.01; tau 53.5 ± 9.36 ms vs. 73.5 ± 13.5 ms, p = 0.001) function. At 30-min recovery (compared with BL), there remained a reduction in CO (5.10 ± 0.92 l/min vs. 4.21 ± 1.63 l/min, p = 0.02) and a borderline significant reduction in dP/dTmax (1,591 ± 244 mm Hg/s vs. 1,453 ± 306 mm Hg/s, p = 0.07) and EF (62.4 ± 5.30% vs. 51.7 ± 10.3%, p = 0.06) as well as an increase in tau (53.5 ± 9.36 ms vs. 59.5 ± 9.26 ms, p = 0.01) consistent with LV stunning (Figures 3 and 4⇓⇓).
In contrast, patients pre-treated with GLP-1 were protected against ischemic LV dysfunction during BO (dP/dTmax 1,417 ± 247 mm Hg/s vs. 1,346 ± 245 mm Hg/s, p = 0.56; SV 97.7 ± 24.8 ml vs. 90.3 ± 31.9 ml, p = 0.54; CO 5.93 ± 1.16 l/min vs. 5.77 ± 1.75 l/min, p = 0.95; dP/dTmin –1,944 ± 507 mm Hg/s vs. –1,757 ± 618 mm Hg/s, p = 0.40) (Figure 5). Furthermore, all parameters of LV performance remained at BL after 30-min recovery, indicating that no stunning had occurred. Compared with control subjects, the GLP-1 group had a smaller reduction in LV performance during BO (dP/dTmax, p = 0.02; SV, p = 0.05; CO, p = 0.049; dP/dTmin, p = 0.047; tau, p = 0.03) and improved LV performance at 30-min recovery (dP/dTmax, p = 0.04; SV, p = 0.03; CO, p = 0.02; dP/dTmin, p = 0.049; tau, p = 0.01). There was no difference in heart rate between the 2 groups during BO or after 30-min recovery.
No adverse events occurred during the study; in particular, no symptomatic or biochemical hypoglycemia was seen during or after GLP-1 infusion.
This study demonstrates that pre-treatment with an intravenous infusion of GLP-1(7-36) amide, commenced before 1-min coronary BO, protects against LV systolic and diastolic dysfunction during supply ischemia and also improves the recovery of function during reperfusion. This occurs with equivalent coronary collateral function and with no change in transmyocardial glucose concentration gradients and may suggest a mechanism of action independent of an effect on myocardial metabolism.
At BL, there were no differences in FFA, insulin, glucose, or GLP-1(7-36) amide concentrations between groups. The elevated levels of FFA seen at baseline in both groups were likely a result of the heparin bolus (16). By the time of BO, the GLP-1 group had increased GLP-1(7-36) amide levels, associated with reduced FFA and glucose concentrations, and a nonsignificant rise in insulin levels. In the control group, the reduction in FFA was less marked and there was a modest fall in insulin with no change in glucose levels. Despite these systemic metabolic changes with GLP-1, we did not identify evidence of a change in myocardial substrate use. Transmyocardial concentration gradients of glucose and FFA after BO were virtually identical between the control and GLP-1 groups, suggesting that the mechanism for the observed reduction in ischemic LV dysfunction may be distinct from an alteration in myocardial metabolism.
Over the last decade, a large body of evidence has emerged highlighting the favorable cardiovascular properties of GLP-1, but the mechanisms remain unclear (17). Perhaps the most widely supported hypothesis is that GLP-1 may have a direct effect on myocardial substrate use (18). Myocardial energy is predominantly derived from the oxidation of FFA, but at times of ischemic stress, there is a natural transition toward greater carbohydrate use (19). GLP-1–related cardioprotection may therefore be mediated by a shift toward increased myocardial glucose use. In keeping with this concept, animal studies have shown that GLP-1 modulation is associated with increased myocardial glucose uptake (MGU) (20,21), reduced myocardial levels of lactate and pyruvate (22), and an increase in the relative oxidation of carbohydrate versus fat (21).
However, few studies have investigated the effect of GLP-1 modulation on myocardial energy metabolism in humans. Moberly et al. (23) demonstrated that GLP-1 (1.5 pmol/kg/min) significantly augmented MGU in lean humans, but not in obese subjects with type 2 diabetes. Gejl et al. (24) found that GLP-1 (1.2 pmol/kg/min) did not affect overall MGU during normo- and hypoglycemic clamps in healthy volunteers. The same group demonstrated that exenatide (a GLP-1 receptor agonist) increased myocardial blood flow but not MGU in diabetic patients during a hyperglycemic clamp (25). Finally, Witteles et al. (26) reported that therapy with sitagliptin (a dipeptidyl peptidase-4 inhibitor) for 4 weeks increased MGU in nondiabetic patients with nonischemic cardiomyopathy. These data suggest that GLP-1 may have variable effects on MGU depending on clinical circumstance.
To the best of our knowledge, our study is the first to investigate the effect of GLP-1 on human cardiac metabolic alterations during myocardial ischemia. GLP-1 protected the heart against ischemic and post-ischemic contractile dysfunction (stunning), but we did not detect a change in transmyocardial concentration gradients to attribute this to an increase in myocardial glucose extraction. In animal studies, myocardial glucose extraction has been demonstrated in low-flow (rather than no-flow) models, with significantly longer durations of ischemia (e.g., 50 min) (27). However, in humans, coronary balloon inflation for 70 s has been sufficient to cause detectable gradients in the transmyocardial concentrations of glucose and FFA 2 min after balloon deflation (28). A second potential explanation for the absence of a cardiometabolic effect observed with GLP-1 may relate to its glucose-dependent properties. The glucoregulatory actions of incretins are known to be dependent on the prevailing plasma glucose concentration, and there is some evidence in animal models to suggest a similar pattern with GLP-1–related cardioprotection (29).
Alternatively, our observations may indicate that the cardioprotection seen with GLP-1 occurred independently of an effect on cardiac metabolism. Studies have suggested that GLP-1 may activate the same prosurvival kinase pathways recruited by ischemic conditioning phenomena, which in turn mediate their cardioprotective effects via the maintenance of mitochondrial membrane integrity (4–6). However, our data indicate that pre-treatment with GLP-1 protects against both ischemic LV contractile dysfunction and stunning, neither of which are believed to be attenuated by ischemic pre-conditioning when infarct size has been accounted for (30–32). This may therefore argue against the hypothesis that GLP-1 exerts its favorable cardiovascular effects solely as a pre-conditioning mimetic. Indeed, it has been suggested that GLP-1–related cardioprotection may be mediated by 2 separate physiological pathways—that the first depends on the GLP-1 receptor for ischemic pre-conditioning, and the second involves receptor-independent actions on post-ischemic reperfusion injury, via its principle breakdown product GLP-1(9-36) amide (3).
The invasive nature of our study protocol meant that we were only able to estimate glucose and FFA extraction (rather than uptake) as surrogates for myocardial substrate use. It is recognized that measurements of glucose metabolism in vivo are less precise than in isolated hearts (33). However, we believe that this is the first study of its kind to investigate the effect of GLP-1(7-36) amide on human myocardial metabolic alterations at the time of ischemia. In addition, we were unable to confirm equivalent baseline collateral flow in both groups because it is not possible to do so invasively without coronary balloon inflation. Finally, this study was not designed to assess the effect of GLP-1 on clinical endpoints. Further work, in the form of adequately powered randomized controlled trials, is needed to determine whether these favorable cardiovascular effects at the time of supply ischemia translate into improved clinical outcomes for patients undergoing PCI.
Pre-treatment with GLP-1(7-36) amide, given prior to the commencement of coronary balloon inflation, protects the heart against ischemic LV systolic and diastolic dysfunction and improves the recovery of function during reperfusion. This occurs without a detected change in myocardial glucose extraction and may indicate a mechanism of action independent of an effect on cardiac substrate use.
The authors thank the patients for their participation and the staff in the cardiac catheter laboratory at Papworth Hospital for their assistance throughout the study.
This study was funded by a grant from the Medical Research Council. The authors have reported that they have no relationships relevant to the contents of this paper to disclose.
- Abbreviations and Acronyms
- balloon occlusion
- coronary artery
- cardiac output
- coronary sinus
- ejection fraction
- free fatty acids
- glucagon-like peptide-1
- left ventricular
- myocardial glucose uptake
- percutaneous coronary intervention
- stroke volume
- Received June 23, 2014.
- Revision received August 11, 2014.
- Accepted September 11, 2014.
- 2015 American College of Cardiology Foundation
- Holst J.J.
- Ban K.,
- Noyan-Ashraf M.H.,
- Hoefer J.,
- Bolz S.S.,
- Drucker D.J.,
- Husain M.
- Bose A.K.,
- Mocanu M.M.,
- Carr R.D.,
- Brand C.L.,
- Yellon D.M.
- Nikolaidis L.A.,
- Mankad S.,
- Sokos G.G.,
- et al.
- Read P.A.,
- Hoole S.P.,
- White P.A.,
- et al.
- Read P.A.,
- Khan F.Z.,
- Heck P.M.,
- Hoole S.P.,
- Dutka D.P.
- Read P.A.,
- Khan F.Z.,
- Dutka D.P.
- McCormick L.M.,
- Kydd A.C.,
- Read P.A.,
- et al.
- Hoole S.P.,
- Heck P.M.,
- White P.A.,
- et al.
- Thygesen K.,
- Alpert J.S.,
- Jaffe A.S.,
- et al.
- Baan J.,
- van der Velde E.T.,
- de Bruin H.G.,
- et al.
- Nelson P.G.
- Nikolaidis L.A.,
- Elahi D.,
- Hentosz T.,
- et al.
- Stanley W.C.,
- Hall J.L.,
- Stone C.K.,
- Hacker T.A.
- Ovize M.,
- Przyklenk K.,
- Hale S.L.,
- Kloner R.A.
- Depre C.,
- Vanoverschelde J.L.,
- Taegtmeyer H.