Maintenance of an adequate oxygen delivery (DO2) to the tissues is one of the important therapeutic approaches in the treatment of critically ill patients [1,2]. Oxygen delivery is determined by arterial oxygen content and cardiac output, which in turn depends on preload, after load and myocardial contractility. Administration of colloids is routinely used to improve intravascular volume status and ventricular preload.
Colloids are high molecular weight substances that increase oncotic pressures, bind water and do not freely permeate across vascular membranes. Synthetic solutions with the same oncotic pressure as plasma, such as starch 6% or gelatine 4%, selectively increase the intravascular volume. However, not all colloids appear to have a similar effect on cardiac performance and the use of colloids for optimizing oxygen delivery has been an issue for many years [1,3]. For instance, gelatine has been reported to produce more favourable effects on oxygen delivery than starch in critically ill patients . In addition, it remains to be elucidated definitively how administration of the different colloids alters cardiac performance in the presence of myocardial ischaemia and reperfusion.
Based on the available literature, our hypothesis was that administration of gelatine would result in a better haemodynamic profile and DO2 when compared with starch. To test this hypothesis, the effects of gelatine and starch on global indices of oxygen transport and left ventricular function were evaluated in an acute open-chest rabbit model subjected to myocardial ischaemia and reperfusion.
Investigations conformed to the Institutional and National Institute of Health Guidelines (NIH Publication 85-23, revd 1985). The local Ethics Committee of Antwerp University approved the experimental procedures. Twenty-seven male New Zealand White rabbits (Oryctolagus cuniculus) were randomly assigned to three groups of nine subjects each. The experimental set-up has already been described in detail .
Anaesthesia and surgery
Anaesthesia was induced with ketamine (25 mg kg−1) and xylazine (15 mg kg−1) and maintained with propofol 0.6 mg kg−1 min−1 and fentanyl 0.5 μg kg−1 min−1; vecuronium (0.003 mg kg−1 min−1) was used to produce muscular paralysis. After tracheotomy and mechanical ventilation of the lungs, a bilateral cervical vagotomy was performed to eliminate parasympathetic influences on cardiac function.
A sternotomy was performed and a micromanometer was then inserted into the cavity of the left ventricle (LV) and the heart was paced at 240 beats min−1. External LV diameter (ExD) and wall thickness (WTH) were measured using ultrasound crystals on the wall (5 MHz, 2 mm diameter) and an arrow-shaped subendocardial crystal (5 MHz) which was connected to a sonomicrometer amplifier (Triton Electronics, San Diego, CA, USA). Facing the crystal on the LV wall, the point of the subendocardial crystal was inserted into the myocardium and its correct position was verified by sending out pulsed time signals of known duration to be received by the other crystal. The exact position was examined later at post-mortem.
The study design is shown in Figure 1. The experimental protocol consisted of two consecutive phases.
Phase 1: Fluid administration. According to the randomization, animals received one of the following colloidal fluids:
• Group 1: NaCl solution 0.9% + albumin 5% (Red Cross) contained Na+ 154 mmol L−1, Cl− 154 mmol L−1 and albumin 50 g L−1.
• Group 2: Hydroxyethylstarch (HES) 6%, mean MW 60 000, a substitution degree of 0.45-0.55 and a C2-to-C5 ratio of 3.4-5.2 (Hemohes 6%, B. Braun, Melsungen, Germany) contained Na+ 154 mmol L−1, Cl− 154 mmol L−1 and hydroxyethylstarch 60 g L−1.
• Group 3: Succinylated gelatine 4%, mean MW 30 000 (Gelofusin 4%, B. Braun) contained Na+ 154 mmol L−1, Cl− 120 mmol L−1, OH− 34 mmol L−1 and gelatine 40 g L−1.
Escalating aliquots of 5 mL solution were infused over 10-15 min to achieve an LV end-diastolic pressure (LVEDP) of between 8 and 10 mmHg. Haemodynamic measurements were performed before (T0), 1 (T1), 10 (T2) and 30 min (T3) after the colloids were infused. Blood sampling for oxygen transport variables were obtained at T0 and T3.
Phase 2: Ischaemia and reperfusion. A stabilization period of 30 min was first maintained. Then, at fixed LVEDP, ischaemia was induced by means of a reversible occlusion of the first branch of the left anterior descending coronary artery with a constricting ligature. Haemodynamic measurements and blood sampling were performed after 10 min of ischaemia (T4) and repeated 15 min after removing the ligature (T5).
Haemodynamic data. Signals were digitally recorded at a sampling rate of 500 Hz (WindaQ®; DataQ Instruments, Inc, Akron, OH, USA). The investigators were blind to the nature of the fluid being administered while recording and analysing the data. At each measurement, the pressure and dimension tracings of five consecutive beats were averaged. The first-order derivative of LV pressure provided the maximum rates of rise and fall of ventricular pressure (mmHg s−1) during the cardiac cycle, dP/dtmax and dP/dtmin respectively. dP/dtmax can be used as a measure of ventricular contractility during systole. However, it is dependent on preload and heart rate; dP/dtmin is an index of left ventricular pressure reduction.
LVEDP was timed at the peak R-wave on the electrocardiogram, while left ventricle end-systolic pressure (LVESP) was timed at dP/dtmin. Heart volumes were calculated using the cube of the internal diameter: (ExD − (2 × WTH))3, providing left ventricle end-diastolic volume (LVEDV) and stroke volume (SV) . Isovolaemic relaxation reflects the left ventricular pressure fall after aortic valve closure. The time constant of isovolaemic relaxation, τ, was calculated according to the logarithmic method of Weiss and colleagues, which assumes that an LV pressure fall follows a monoexponential course .
Blood-gas and oxygen transport data
Arterial and venous blood samples were drawn, respectively, from a femoral artery and the right jugular vein  using dedicated catheters. Samples were collected anaerobically for immediate analysis using a Blood-gas Analyzer System 865® and Co-Oximeter 2500® (Bayer N.V., Brussels, Belgium). The measured variables included blood-gas analysis, haemoglobin, and lactate and glucose concentrations. Oxygen transport variables and the arteriovenous carbon dioxide gradient were calculated using standard formulae .
With a power of 80%, a sample size of nine animals in each treatment group was calculated as appropriate. Data were expressed as the mean and 95% confidence interval (CI) using a t-test distribution for nine subjects and were presented as the mean ± 2.26 SEM in the text and tables and ±2.26 SEM error bars in the accompanying figures .
A two-way ANOVA for repeated measures was performed to take account of the expected time-related changes during the different experimental steps. Between groups and interaction analyses were used to detect differences between colloids at the different time points. Statistical significance was taken as P < 0.05. A one-way ANOVA was applied for isolated variables and baseline measurements. If appropriate, ANOVA procedures were followed by Newman-Keuls modified t-tests for post hoc comparisons of significant differences between groups or within groups over time using Bonferroni's correction for multiple comparisons. Sample size estimation and statistical analysis were performed using the STATISTICA® package for Windows® (StatSoft, Inc, Tulsa, OK, USA).
All animals were comparable in size. Total amounts of administered colloid, urine production and fluid balances throughout were comparable in all groups (Table 1).
Biochemical and oxygen transport variables at T0, T3, T4 and T5 are summarized in Table 2. Colloid administration was associated with a similar haemodilution and decrease in the arterial oxygen content in all groups. Glucose concentrations decreased similarly in all groups. Venous oxygen content was reduced during ischaemia and reperfusion in all three groups.
A slight fall in body temperature with a concomitant decrease in VO2 occurred over time in all groups.
DO2 and LVEDV increased similarly in all groups during volume loading (Fig. 2). During ischaemia (T4), DO2 was reduced in all groups, while LVEDV remained stable. After 15 min of reperfusion (T5), DO2 and LVEDV were similar to the values at T4. No differences between groups could be demonstrated.
As shown in Figure 3, the isovolaemic relaxation constant, τ, increased significantly in all groups during ischaemia and reperfusion. Differences between groups did not reach statistical significance.
Haemodynamic variables during fluid loading and with ischaemia are summarized, respectively, in Tables 3 and 4. All colloids equally increased preload as assessed by the similar changes in LVEDP and LVEDV over time for the three groups. Changes in positive and negative peak dP/dt are shown in Figures 4 and 5, respectively. With the fluid loading, LVESP, dP/dtmax, dP/dtmin and SV simultaneously increased significantly in all three groups. However, the observed declines of dP/dtmax, dP/dtmin and LVESP during ischaemia and reperfusion were more pronounced (P ≤ 0.05) in the hydroxyethylstarch 6% group.
The data indicated that administration of albumin 5%, hydroxyethylstarch 6% and succinylated gelatine 4% had similar effects on myocardial function and oxygen delivery (DO2) in the rabbit. However, during ischaemia and reperfusion, the decrease in myocardial function was more pronounced in the hydroxyethylstarch 6% group. It seems, therefore, that although all agents had similar effects at baseline conditions, the use of hydroxyethylstarch 6% may be less favourable in the presence of myocardial ischaemia and reperfusion.
These findings agree with previous reports. In pathological conditions, the response of cardiac performance to increases in circulating volume may vary depending on the type of fluid administered . Hydroxyethylstarch produced less favourable effects on cardiac performance and oxygen delivery in critically ill patients compared with gelatine . The precise mechanism of this phenomenon remains to be clarified. Interactions at the endothelial level with consequent modulation of microvascular permeability during reperfusion have been reported with hydroxyethylstarch [11,12].
Furthermore, cardiac endothelium has been shown to play a key role in the regulation of myocardial function [13,14]. Therefore, the hypothesis is that possible alterations in cardiac endothelial structure and function by hydroxyethylstarch may also influence cardiac performance during ischaemia and reperfusion.
In the present study, myocardial contractile function was assessed with dP/dtmax. Although dP/dtmax allows a fair estimation of cardiac performance, its values are preload-dependent. However, in the present experiments, preload - assessed by both LVEDPs and dimensions - did not differ between groups under the different experimental conditions. Changes in dP/dtmax can, therefore, be assumed to reflect changes in myocardial contractile function. Volume-based indices of myocardial contractile function, such as ventricular elastance, were not considered. Indeed, the use of these indices in this particular animal model can be questioned. The estimation and mathematical handling of such minuscule values such as the rabbit SV and its changes may implicit intrinsic errors resulting in a higher variability of volume-based variables .
The more pronounced decrease in LV dP/dtmax in the hydroxyethylstarch 6% group during ischaemia was associated with a more pronounced decrease in DO2, though it did not reach statistical significance. In the presence of decreased myocardial contractility, compensatory mechanisms are elicited to maintain cardiac output. As long as these mechanisms suffice to maintain an adequate cardiac output, the changes in DO2 will remain minimal. The present protocol was not designed to analyse potential differences under circumstances of critical DO2. Further studies will be needed to evaluate this issue.
It should be noted that rabbits are hyperthermic compared with human beings (39 versus 37°C). As temperature influences oxygen consumption and the oxygen affinity of haemoglobin, data based on blood-gas analyses may not be unequivocally extrapolated between the two species [16-18].
In conclusion, oxygen delivery and left ventricular function improved similarly after administration of albumin 5%, hydroxyethylstarch 6% and succinylated gelatine 4%. During ischaemia and reperfusion, the decline in left ventricular performance was more pronounced in the hydroxyethylstarch 6% group. Well-designed trials in larger animals and human beings are needed to see if the present experimental findings also apply to clinical practice for patients prone to cardiac ischaemia.
This work was supported by a grant from Messrs B. Braun, PO Box 1120, D-34209 Melsungen, Germany.
1. Shoemaker WC, Wo CC, Thangathurai D, et al.
Hemodynamic patterns of survivors and nonsurvivors during high risk elective surgical operations. World J Surg
2. Bishop MH, Shoemaker WC, Appel PL, et al.
Relationship between supranormal circulatory values, time delays, and outcome in severely traumatized patients. Crit Care Med
3. Hauser CJ, Shoemaker WC, Turpin I, Goldberg SJ. Oxygen transport responses to colloids
and crystalloids in critically ill surgical patients. Surg Gynecol Obstet
4. Beards SC, Watt T, Edwards JD, Nightingale P, Farragher EB. Comparison of the hemodynamic and oxygen transport responses to modified fluid gelatin and hetastarch in critically ill patients: a prospective, randomized trial. Crit Care Med
5. De Mulder PA, Van Kerckhoven RJ, Adriaensen HF, Gillebert TC, De Hert SG. Continuous total intravenous anesthesia, using propofol and fentanyl in an open-thorax rabbit model: evaluation of cardiac contractile function and biochemical assessment. Lab Anim Sci
6. Feigenbaum H, Popp R, Wolfe S. Ultrasound measurement of the left ventricle: a correlative study with angiocardiography. Arch Intern Med
7. Weiss JL, Fredriksen JW, Weisfeldt ML. Hemodynamic determinants of the time-course of fall in canine left ventricular pressure. J Clin Invest
8. Reinhart K, Rudolph T, Bredle DL, Hannemann L, Cain SM. Comparison of central-venous to mixed-venous oxygen saturation during changes in oxygen supply/demand. Chest
9. Ducey JP, Lamiell JM, Gueller GE. Arterial-venous carbon dioxide tension difference during severe hemorrhage and resuscitation. Crit Care Med
10. Gardner M, Altman G. Confidence intervals rather than P
-values. In: Altman G, ed. Statistics with Confidence.
London, UK: BMJ Publ., 1989: 83-100.
11. Oz MC, FitzPatrick MF, Zikria BA, Pinsky DJ, Duran WN. Attenuation of microvascular permeability dysfunction in postischemic striated muscle by hydroxyethyl starch. Microvasc Res
12. Vincent JL. Plugging the leaks? New insights into synthetic colloids
. Crit Care Med
13. De Hert SG, Gillebert TC, Andries LJ, Brutsaert DL. Role of the endocardial endothelium in the regulation of myocardial function: physiologic and pathophysiologic implications. Anesthesiology
14. Bettens KM, De Hert SG, Sys SU, Brutsaert D. Role of the endocardial endothelium in the negative inotropic effects of thiopental. Anesthesiology
15. Milnor WR. The normal hemodynamic state. In: Collins N, ed. Hemodynamics,
2nd edn. Baltimore, USA: Williams & Wilkins, 1989: 164-166.
16. Reeves RB. The effect of temperature on the oxygen equilibrium curve of human blood. Respir Physiol
17. Schmidt-Neilson K, Latimer JL. Oxygen dissociation curves of mammalian blood in relation to body size. Am J Physiol
18. Dexter F, Hindman BJ. Effect of haemoglobin concentration on brain oxygenation in focal stroke: a mathematical modelling study. Br J Anaesth