Preserved myocardial energetics in acute ischemic left ventricular failure – studies in an experimental pig model

Abstract

Objective: We hypothesised that acute ischemic left ventricular failure is characterised by depressed systolic and diastolic function combined with inefficiency in oxygen to mechanical work energy transfer. Methods: Eight anaesthetised pigs (32±3 kg) were employed in an in vivo open chest model. Intraventricular combined pressure and conductance catheters were used to generate continuous left ventricular pressure–volume relations. Myocardial oxygen consumption (MVO₂) was determined from coronary flow and coronary arteriovenous oxygen difference. After baseline measurements, ischemia was induced by repeated left coronary injections of 50 μm polystyrene microspheres until stroke volume was reduced by 30%. Haemodynamic and biochemical measurements were repeated 30, 90 and 150 min after microembolisation. Results: Coronary embolisation induced a significant reduction in stroke work (2749±504–1473±449 mmHg ml, P≪0.05) at 30 min compared to baseline. Post-embolic contractility was reduced measured by the slope of the preload recruitable stroke work index (66.2±12.8–50.0±5.8 mmHg, P≪0.05) and the slope of the curvilinearly fitted end-systolic pressure–volume relation in V₀ (7.1±2.2–4.9±2.2 mmHg/ml, P≪0.05). The dP/dt_min decreased (2076±291–1468±266 mmHg/s, P≪0.05), but there was no significant change in diastolic stiffness or Tau. Following the 30 min measurements, there were only small changes in most indices. We found no change in myocardial oxygen consumption for basal metabolic processes or excitation–contraction coupling (unloaded MVO₂), and there were no changes in conversion of oxygen to total mechanical work (MVO₂–PVA slope). However, decreased mechanical efficiency (SW/MVO₂) paralleled an increased ratio of arterial elastance to ventricular elastance. Conclusions: Coronary microembolisation in pigs induce a stable ischemic left ventricular failure characterised by reduced contractility and minimally impaired diastolic function. In this acute ischemic left ventricular failure, the main contributor to all over cardiovascular inefficiency is increased ratio of arterial- to ventricular elastance, a setting that impairs mechanical efficiency. However, efficiency of oxygen to total mechanical work transfer in the myocardium is unaltered. The mechanism behind this finding is elusive and warrants further investigation.

Myocardium, Failure, Ischemia, Contractility, Efficiency

1 Introduction

Models employing rapid ventricular pacing (RVP) and repeated microembolisations have been used to study mechanics and energetics in the failing left ventricle [1,2]. The results of these studies are conflicting with regard to both mechanical defects and the relationship between oxygen utilisation (MVO₂) and mechanical work. In heart failure induced by RVP, an early increase in unloaded myocardial oxygen consumption (MVO₂) has been found [3], indicating increased energy consumption in non-mechanical processes. However, in hearts with a more chronic failure after long-term RVP [2], the unloaded MVO₂ was unchanged, but the slope of the relation between MVO₂ and mechanical work was decreased, indicating an increased efficiency in mechanical work. Using chronic microembolisation in dogs, Todaka et al. [1], observed a small decrease in unloaded MVO₂, but concluded that an uncertainty exists as to whether this represents a true change in energetics. Thus, it seems likely that the energetics of cardiac failure depends on both the time span of heart failure and the model employed. Coronary microembolisation has previously been shown to induce a stable ischemic cardiac failure [4]. Furthermore, recent clinical observations indicate that microembolisation may be an important entity in human ischemic contractile failure after coronary interventions and in unstable coronary syndromes [5,6].

The aim of the present study was to assess the mechanoenergetic responses very early in acute ischemic left ventricular failure induced by microembolisation, and evaluate mechanoenergetic parameters in the pressure–volume framework i.e. differentiate energy consumption in basal metabolism, excitation–contraction coupling and mechanical work [7]. Based on previous studies in post-ischemic hearts [8,9], and other heart failure models [1–3], our hypothesis was that these hearts have an inefficient mechanoenergetic relationship after induction of ischemic left ventricular failure.

2 Methods

2.1 Animal preparation

The experimental protocol was approved by the local steering committee of the Norwegian Animal Experiments Authority (NARA). All animals received care in compliance with the European Convention on Animal Care.

Eight castrated male pigs (sus scrofa domesticus, weight 32±3 kg) were kept in the animal department for 5–7 days and fasted overnight with free access to water. At the day of the experiment, the animals were premedicated with intramuscular injection of ketamine (20 mg/kg, Warner Lambert Nordic, Solna, Sweden) and atropine (1 mg, Nycomed Pharma, Oslo, Norway). Anaesthesia was induced by intravenous injection of pentobarbital-sodium (10 mg/kg, Nycomed Pharma) and fentanyl (0.01 mg/kg, Pharmlink, Spanga, Sweden). The animals were tracheostomised, intubated and ventilated with 60% oxygen in volume-controlled mode (Servo 900 respirator, Elma-Schonander, Stockholm, Sweden). Tidal volume was adjusted to keep arterial PaCO₂ in the range of 4.0–6.0 kPa. A central venous catheter (Secalon T, Beckton Dickinson, Singapore) was placed thorough the left internal jugular vein, and anaesthesia was maintained throughout the experiment using continuous infusion of pentobarbital-sodium (4.0 mg/kg/h), fentanyl (0.02 mg/kg/h) and midazolam (0.3 mg/kg/h, Alpharma, Oslo, Norway). Central venous pressure (CVP) was monitored through the right jugular vein. Mean arterial pressure (MAP) was measured in the descending thoracic aorta through a catheter inserted in the left femoral artery (Connecta, Beckton Dickinson, Helsinborg, Sweden). Arterial blood sampling was done through a catheter in the right femoral artery. After sternotomy, the left hemiazygos vein was ligated to avoid return of systemic blood to the coronary sinus. Transit time flow probes (CardioMed CM-4000, Medi-Stim AS, Horten, Norway) were placed on the left anterior descending, circumflex and right coronary arteries for determination of coronary blood flow. A flow probe was also placed on the pulmonary artery for assessment of cardiac output. A 7 Fr balloon catheter (Sorin Biomedical, Irvine, CA) was introduced in the inferior caval vein for preload reduction. A dual field, combined pressure-conductance catheter (7 Fr., 12 electrodes, Sentron, AC Roden, the Netherlands) was inserted to the left ventricular cavity via the left carotid artery. Assessment of individual segment pressure–volume loops confirmed the proper placement of the catheter. Myocardial venous blood was drawn from a catheter placed in the great cardiac vein via the coronary sinus (Blue Flex Tip, Arrow, Reading, PA). Finally, a catheter (Pediatric Jugular Vein Cat., Arrow, Reading, PA) was inserted into the main pulmonary trunk through the right ventricular wall for measurement of mean pulmonary artery pressure (MPAP) and injection of 10% NaCl for parallel conductance assessment. Blood volume was maintained throughout the experiment by continuos infusion of 0.9% NaCl supplemented with 1.25 g/l glucose. The bladder was catheterised via a cystostoma. After surgical preparation, the animal received 2500 IU heparin, and this was repeated once during the experiment.

2.2 Induction of left ventricular failure

After baseline measurements, microembolisations were done through a catheter in the left coronary main stem, placed under fluoroscopic guidance. Fifty micrometers of polystyrene microspheres (NEM-005, NEN Lifescience Products, Boston, MA) were dissolved in 0.9% NaCl and 0.01% Tween 80 (Sigma Chemical Co., St. Louis, MO) to a concentration of 1 mg microspheres per ml. The microspheres were kept under continuous stirring and thoroughly shaken before injection. Left ventricular failure was induced by repeated injections of 2.5–5.0 mg boluses every 5 min until a 30% decrease in stroke volume was obtained. An average of 30±5 mg microspheres given over approximately 40 min was needed to achieve this level of left ventricular failure.

2.3 Experimental protocol

The study was designed such that each animal served as its own control, i.e. all haemodynamic assessments were performed at baseline and then after induction of left ventricular failure. After surgery and instrumentation, the animals were stabilised for 30 min. Baseline measurements were then conducted. These measurements consisted of arterial blood samples for assessment of blood resistivity (ρ) and haemoglobin (Hb). Parallel conductance, or non-left ventricular blood conductance (V_p), was estimated by injection of 10% NaCl into the pulmonary artery with simultaneous left ventricular pressure–volume sampling [10]. Respiratory influence on heamodynamics was avoided by disconnecting the respirator during file sampling. In order to assess contractility, pressure–volume data were recorded during transient (12–15 s) preload reduction by vena cava occlusion. A series of pressure–volume assessments with concomitant MVO₂ determination were then performed at six different steady-state preloads in order to assess the MVO₂–PVA relationship. After stabilisation of cardiac output, arterial pressure and coronary flow, each preload step was kept for minimum 10 s before start of registrations in order to stabilise oxygen extraction. It has previously been shown that coronary oxygen extraction adjusts rapidly, even during large and swift changes in perfusion pressure [11], and that 10 s stabilisation is adequate to detect changes in the MVO₂–PVA relationship [12]. The lowest MAP during preload reduction was 45 mmHg. General haemodynamic values, such as CVP, MPAP, HR etc. were also noted. The same measurements were performed as described 30, 90 and 150 min after stable heart failure was obtained to assess heamodynamics, ventricular mechanics and the MVO₂–PVA relationship.

At the end of the experimental protocol, the animals were sacrificed by intracardiac injection of KCl and lethal dose injection of pentobarbital. The heart was then excised and the myocardium was weighed to assess blood flow per. weight of myocardium and normalise values to 100 g left ventricular wet weight.

2.4 Data acquisition and analysis

The conductance catheter technique has been extensively described earlier [10]. Briefly, the intraventricular volume was calculated from segments of intraventricular blood conductances, employing the formula:  

where V(t) is total intraventricular volume, α the slope factor relating conductance volume to an independent volume estimation, L the interelectrode distance, ρ the blood resistivity, G(t) the sum of segmental conductances and G_p the parallel conductance. The pressure- and conductance signals were processed using a conductance conditioner (Leycom, Sigma 5DF, Cardiodynamics, the Netherlands). The sampling rate of the signals was set to 200 Hz, and real time signals were displayed on a personal computer using the software Conduct PC, CPC V3.15 (Cardiodynamics) Blood resistivity (ρ) was estimated using a cuvette designed for the conductance equipment. Parallel conductance was estimated in the software from the saline dilution technique, using at least four consecutive beats with increasing conductance in the ventricular phase of the saline wash-through. The slope factor α was estimated from the ratio between conductance and transit time derived cardiac outputs. The pressure–volume calculations were performed using a customised software package (Conduct PC, CD Leycom, the Netherlands) after thorough examination of the pressure–volume loops for proper electrocardiac marking.

Pressure–volume area (PVA, in mmHg ml) represents the total mechanical work as described by Suga [7]. Briefly, PVA consists of the area bounded by the pressure–volume loop (external work, SW) and the area limited by the line of the end-systolic- and end-diastolic pressure–volume relationships (PE, potential energy) (see Fig. 1 ). PVA was calculated by the formula:  

where SW is Stroke work or the area of the pressure–volume loop calculated from integrated pressure-volume data; ESP and ESV are end-systolic pressure and volume, respectively. V₀ is the extrapolated x-intercept of quadratic fitted end-systolic pressure–volume relationship during vena cava occlusion, and EDP is end-diastolic pressure.

Fig. 1

(A) The pressure volume area (PVA) or total mechanical work. The area bounded by the loop conforms to stroke work (SW), and the area bounded by the end-systolic- and end-diastolic pressure–volume relationship (ESPVR and EDPVR, respectively) is potential energy (PE). Total mechanical work (PVA) is the sum of stroke-work and potential energy. (B) The linear relationship between mechanical work (PVA) and myocardial oxygen consumption (MVO₂). Myocardial oxygen consumption consists of unloaded MVO₂ (oxygen used for basal metabolic processes and excitation–contraction coupling) and PVA-dependent MVO₂ (oxygen used to generate mechanical work). Inotropic stimulation (arrows) works by increasing the amount of calcium involved in excitation–contraction coupling, and thus increases unloaded MVO₂ (dashed line). The inverse of the slope defines efficiency in the generation of total mechanical work from oxygen.

Haemoglobin oxygen saturation was measured using a standard blood gas analyser (Ciron Diagnostics, Ciba, Medfield, MA). Left ventricular coronary blood flow was estimated by the formula:  

where LVCBF and CBF are left ventricular and total coronary blood flow, respectively. W is total and LVW is left ventricular myocardial weight, respectively. Left ventricular oxygen consumption was calculated by:  

where MVO₂ is left ventricular myocardial oxygen consumption, avdO₂ the difference between aortic and myocardial venous oxygen saturation, Hb haemoglobin in g/ml, 1.39 is a constant (in ml O₂/g Hb) and HR the heart rate. In order to convert MVO₂ to energy equivalent to mechanical energy, the factor 20.2 J/ml O₂ was used. MVO₂ relates to PVA in a linear fashion (see Fig. 1).

Left ventricular contractility was assessed using several indices. The slope of the end-systolic pressure–volume relationship (E_es) and slope of the stroke work–EDV relation (PRSW) have been described in detail elsewhere [13,14]. We assessed E_es both as the slope of the linear ESPVR and the slope of a fitted second order polynomial curve (y=ax²+bx+c) in V₀. This slope avoids the potential error of inadequate downloading during preload reduction, and yields a value in agreement with E_es of severely downloaded ESPVRs. Moreover, the method has been shown to detect changes in contractility over a wide range of interventions both in isolated hearts [15] and in vivo [16]. Therefore, we used this slope as an index of ventricular elastance (E_es) in calculations of arterial- to ventricular elastance ratios. Diastolic function was assessed by the derivative of pressure decay with respect to time and the time constant of relaxation, Tau, both calculated by the CPC software. In this program, Tau is calculated as the time from dP/dt_min until pressure reaches half the value at dP/dt_min. End-diastolic pressure–volume data were fitted to an exponential equation:  

where e is the constant of the natural logarithm, β describes the slope of the pressure–volume relationship and is an indicator of ventricular stiffness.

Beat to beat pressure–volume data were copied from the Conduct PC software to a spreadsheet (Excel 97, Microsoft) and processed further. Linear, curvilinear and exponential relationships were estimated by least squares fit regression in the spreadsheet.

2.5 Statistics

Statistical analysis was performed using a statistical software package (SPSS 9.0, SPSS inc. Chicago, IL). We used repeated measure analysis of variance (RANOVA) with Bonferroni correction for multiple comparisons. Difference from baseline and between each time point was assessed by pair wise comparison between each run by within subject contrasts. MVO₂–PVA relationships were also compared by multiple linear regression analysis and analysis of covariance with dummy variables coding for time-points. Significance testing of the curvilinearity of the ESPVR was done by regression of all three parameters of the second order polynomial equation. All values are reported as mean±SD unless stated otherwise. Values are significant if P≪0.05.

3 Results

The effects of microembolisation on haemodynamic parameters are summarised in Table 1 . There was only a modest, non-significant increase in end-diastolic pressure but end systolic- and mean arterial pressures were significantly decreased by the second post-embolisation measurement (90 min). There was no change in parallel conductance throughout the study. The slope factor alpha (relationship between CO measured by conductance catheter and transit time flow probes) changed minimally during the experiment (range 0.87–1.03), demonstrating a high degree of agreement between the two methods of measuring CO. Arterial elastance, depicting total arterial impedance faced by the left ventricle, was significantly increased after microembolisation. This increase combined with the reduction in ventricular contractility led to increased ratio of arterial-to-ventricular elastance (E_a/E_es).

Table 1

General haemodynamic variables before and 30, 90 and 150 min after induction of heart failure^a

Left ventricular contractility before and after microembolisation is presented in Table 2 . The reduced contractility was obvious from the classic load-dependent variables ejection fraction and dP/dt_max. However, there was also a significant reduction in the more load insensitive PRSW and the slope of the second order polynomial fitted ESPVR-data in V₀. ESPVR data were better fitted to a curvilinear equation for all ESPVRs at all time points, and all ESPVRs displayed significant curvilinearity concave to the volume axis. The slope of the linear ESPVR showed no significant change throughout the experiment, but there was a significant rightward shift of the linearly derived V₀.

Table 2

Indices of left ventricular contractility before and 30, 90 and 150 min after microembolisation^a

The impact of microembolisation on diastolic function assessments is summarised in Table 3 . Values of EDP and EDV are given in Table 1. There was only a small non-significant increase in beta, indicating no change in diastolic chamber stiffness. Tau showed a slight non-significant increase after microembolisation while dP/dt_min was reduced by 28% after induction of failure and remained low.

Table 3

Left ventricular diastolic properties before and 30, 90 and 150 min after induction of heart failure^a

Table 4 and Fig. 2 show the effect of microembolisation on left ventricular mechanical work and oxygen consumption. The highly significant 46% reduction in stroke work resulted from both reduced stroke volume and pressure generation. However, there were no significant changes in oxygen utilisation for non-contractile processes (unloaded MVO₂) or efficiency in the conversion of oxygen to total mechanical work (MVO₂–PVA slope). However, microembolisation led to a significant decline in both work efficiency (SW/PVA) and mechanical efficiency (SW/MVO₂). Left ventricular coronary flow did not alter throughout the experiment, but there was a significant increase in coronary venous oxygen saturation.

Table 4

Left ventricular energetics before and 30, 90 and 150 min after induction of heart failure^a

Fig. 2

MVO₂–PVA data for all animals at baseline (♦) and at first post-embolisation (○) measurement 30 min after induction of heart failure. Lines indicate mean values for all animals at baseline (—) or 30 min after heart failure (--). There was no significant change in slope or y-axis intercept (unloaded MVO₂) at this or subsequent measurements compared to baseline.

4 Discussion

This is the first study to examine the relationship between oxygen consumption and mechanical work in a model of acute ischemic left ventricular failure. The mechanoenergetic relationships observed after microembolisation revealed no alterations in the myofibrillar efficiency (the slope of the MVO₂–PVA relation), and no altered unloaded MVO₂ i.e. no apparent alteration in basal metabolism or excitation–contraction coupling. This is surprising, as a large number of studies have demonstrated an inefficiency in reperfused myocardium after ischemia [8,9] and incipient heart failure after RVP led to increased energy consumption for non-contractile work [3]. On the other hand, unaltered unloaded MVO₂ combined with reduced contractility, as in the present study, may indicate increased energy consumption in basal metabolism or excitation–contraction coupling [7,17]. However, even though the quantitative relationship between contractility and non-work-related oxygen consumption have been defined in isolated hearts [17], this has not been thoroughly described in a model similar to ours. Therefore, to ascribe the unchanged unloaded MVO₂ combined with reduced contractility to oxygen wasting in basal metabolism or excitation–contraction coupling, would be speculative. Further studies to address this are clearly needed. The unchanged slope of the MVO₂–PVA relationship indicates no alterations in oxygen-to-adenosine triphosphate (ATP) or ATP-to-PVA efficiencies. Although there is a theoretical possibility for reciprocal changes, this is also highly speculative. Probably no alterations do occur that can be attributed to distorted metabolism in myocytes.

After microembolisation, there was no significant change in coronary blood flow, but the cardiac venous oxygen level increased. This indicates an ‘end-artery disease’ or no-reflow phenomena, and shunting of blood away from metabolic feeding vessels. These observations also point to yet another possible mechanism for unaltered mechanoenergetics after microembolisation. Non-perfusion of the ischemic areas do lead to patch necrosis [4] and the possibility therefore exists that only normal myocardium contribute to energetics in these failing hearts, explaining the unaltered energetics. Seen in the context of previous studies in heart failure models [1–3] altered mechanoenergetic relations can be a dynamic process during failure with changing efficiencies in the different phases of heart failure.

In physiological conditions, ventricular elastance is approximately twice arterial elastance, E_a/E_es=0.5, a setting that optimises mechanical efficiency (SW/MVO₂). During heart failure, the ratio increases towards 1, a condition that optimises stroke work at the expense of efficiency [18]. The reduced contractility seen in the present study is coupled to increased arterial impedance. As predicted by others [18,19], the increased ratio of arterial to ventricular elastance is paralleled by a decrease in mechanical efficiency. A similar decrease is found in work efficiency (SW/PVA). Our observations of energetics thus show that even though the myocardial efficiency is preserved, the efficiency of the cardiovascular system as an integrated unit is severely impaired during acute ischemic left ventricular failure.

All commonly used indices of contractility show a variable degree of load- and frequency dependence. Ejection fraction is known to vary depending on both ventricular filling and afterload. While end-diastolic volume and pressure was not significantly altered, arterial input impedance (ESP/SV) increased in our study. The 25% decrease in EF must therefore be interpreted with caution. The same applies to approximately 30% reduction in dP/dt_max, which is known to be both preload- and frequency dependent. The slope of the SW–EDV relationship (Preload Recruitable Stroke Work, PRSW) is proposed as an index of contractility, independent of ventricular geometry, loading conditions and heart rate [14]. However, the index also incorporates diastolic function, as stroke work is dependent on diastolic pressure and volume, and PRSW slope is therefore probably an index of integrated ventricular function throughout the cardiac cycle. Because we found no significant change in diastolic stiffness, volume or pressure, we conclude that the 25% reduction in PRSW is mostly due to reduced contractility.

The slope of the linear ESPVR is often assumed to be independent of loading conditions [20], but more recent studies show that this is not always the case [21]. The non-significantly increased slope found in our study could theoretically be due to an increased heart rate and/or altered loading conditions during the experiment. However, in line with others, we found a significant rightward movement of V₀. Such a shift is almost a constant finding in ischemic and post-ischemic hearts, and seems more sensitive in detecting reduced contractility than the slope of the linear ESPVR [22]. Due to significant curvilinearity of the ESPVR, the slope of the linear ESPVR could be misleading as an index of contractility. It has been shown that the degree of curvilinearity depends on the contractility [15,16]. Therefore, the slope of a linear regression is very dependent on the degree of downloading. From our data, it was evident that due to the degree of curvilinearity, the linearly derived slope underestimated the contractility at baseline. Furthermore, after embolisation, when contractility and thus curvilinearity were lower, the linearly derived E_es did not underestimate contractility to the same degree.

Therefore, we also fitted end-systolic pressure–volume data to a second-order polynomial expression, and calculated the slope in V₀ (Fig. 3 ). Employing this model, we found a decrease in the slope in V₀ of 31% from baseline to 30 min. This decrease conforms to the reduction in other indices of contractility, indicating that the quadratic model describes the reduced contractile state better than a linear model. However, a problem with using the slope in V₀, is that it involves extrapolation of data beyond measured values. To assess whether this represented a problem in this study, we calculated the slope of the quadratic equation at ‘low load’ (lowest measured end-systolic-pressure and -volume at each loading intervention). These slopes displayed a similar trend as in V₀ with a significant reduction from baseline to first post-embolisation measurement (4.71±1.52 to 3.30±1.32, P≪0,01) and minor alterations thereafter. There were no significant differences in ESV or ESP defined as low load at the two time points (ESV=7.6±2.2 to 16.0±12.7, P=NS and ESP=65.4±16.6 to 62.3±12.4, P=NS). However, this method involves the potential error of evaluating the slope at different end-systolic pressure and volume points, a similar problem encountered with linear regression. Furthermore, this slope has not been as thoroughly investigated in different models as the slope in V₀. Another method of evaluating ESPVRs has been to evaluate ESP at a reference ESV or vice versa. However, when ESPVRs yield regression lines with different slopes, the results are highly dependent on the chosen reference value.

Fig. 3

End-systolic- and end-diastolic pressure–volume data for one representative animal before (♦) and after (○) microembolisation. A decreasing degree of curvilinearity of the ESPVR from baseline to the post-embolisation measurement was observed. Also, a decreased steepness of ESPVR in V₀ was found after embolisation. There was also a slightly steeper end-diastolic pressure–volume relationship after embolisation in this animal.

In normal hearts, increased heart rate leads to attenuated diastolic relaxation [23]. However, despite increased frequency, there was no increase in the time constant of isovolumic pressure decay, Tau. Nevertheless, dP/dt_min decreased 28% from baseline to the first post-embolisation measurement. The discrepancy between Tau and dP/dt_min is puzzling, but the calculation of Tau has some inherent problems. In the model employed by us, the pressure-asymptote is regarded as P_∞=0 mmHg. Even though this has been shown to be adequate during physiological conditions [24], this may not be true after ischemia. We found the diastolic stiffness-constant beta to increase slightly throughout the experiment, but this increase never reached statistical significance. These findings indicate only slightly impaired diastolic function and compliance during acute ischemia. However, all animals in the present study have open thorax and pericardium, and these findings may therefore not be directly transferable to less invasive models.

During the repeated bolus microembolisation, we found stroke volume to be the most sensitive indicator of heart failure progression. In line with studies in other species, we observed a rise in end-diastolic pressure. This increase was only modest with large inter-animal variations, and was therefore not specific as a predictor of pump failure. Pilot studies revealed that approximately 30% reduction in stroke volume led to moderate left ventricular failure, as measured by several indices of ventricular function. After adequate cardiac failure was established, there were only minor changes in performance parameters, indicating that the microembolisation itself and not the effect of time and anaesthesia caused the ventricular deterioration.

Care must be taken when extrapolating these findings to human ischemic left ventricular failure. However, microembolisation could be an important entity in several clinical syndromes, including unstable coronary syndromes and heart failure after coronary interventions. This study therefore provides a starting point for deciphering pathophysiology of acute ischemia and a substrate for further studies of energetics and cardiac mechanics during acute left ventricular failure.

This study was supported with a grant from the Norwegian Research Council and the Norwegian Council on Cardiovascular Diseases. We acknowledge the skilful technical assistance of Hanne Maehre, Ernst Rolf Albrigtsen, Hege Hagerup and Elinor Hareide.