Adequate tissue oxygenation depends on the balance between tissue oxygen demand and tissue oxygen delivery (DO2). When systemic DO2 decreases, tissue oxygen uptake (VO2) remains constant, because the systemic oxygen extraction ratio (O2ER) progressively increases [1,2]. Below a critical value of DO2 (called critical DO2), the increase in O2ER is insufficient to compensate for the reduction in DO2, and VO2 becomes delivery dependent.
Cardiopulmonary bypass (CPB) is often used during cardiac surgery to maintain oxygen availability to the tissues. In the presence of stable arterial oxygen saturation and haemoglobin concentration, DO2 depends exclusively on pump flow. Values of pump flow between 2.0 and 2.4 mL min−1 m2 are currently used, governed by a combination of clinical experience and physiological research . However, critical DO2 and pump flow remain largely unknown during CPB , even though these variables have been well defined during experimental anaemia [5,6], haemorrhage [7,8], and septic shock [9,10].
The use of a pulsatile pump flow has been suggested to be associated with a better blood flow distribution [11,12]. However the precise effects of pulsatile pump flow on tissue oxygen extraction capabilities have not been defined. The present study was designed to define critical DO2 and pump flow during pulsatile and non-pulsatile 37°C CPB in anaesthetized dogs. In addition, the value of blood lactate levels in determining these critical parameters was evaluated.
Material and methods
All experimental procedures used in this investigation were performed according to the National Institutes of Health (NIH) Principles of Laboratory Animal Care guidelines and were approved by the institutional Animal Investigation Committee. The study included 18 mongrel dogs (28.2 ± 2.9 kg). After 12 h of fasting, animals were anaesthetized with an intravenous administration of 20 mg kg−1 of thiopental. Tracheal intubation was performed and mechanical ventilation was started with air (Elema 900 B; Siemens, Solna, Sweden). The respiratory rate was set at 12 min−1 and the tidal volume was adapted to obtain a PaCO2 between 4.5 and 5 kPa. Anaesthesia was maintained with fentanyl, given as a bolus dose of 50 μg kg−1 over 10 min, followed by a continuous infusion of 50 μg kg−1 h−1 until CPB and 25 μg kg−1 h−1 thereafter. Muscle relaxation was provided by pancuronium bromide, 0.1 mg kg−1 followed by repeated boluses of 1-2 mg h−1. A catheter was advanced into the abdominal aorta through a femoral artery for arterial pressure monitoring and blood gas sampling.
A median sternotomy was performed. After administration of 5 mg kg−1 of heparin, cannulas were inserted in the aortic root and the right atrial appendage. Blood circulation was maintained by a roller pump (Stöckert Instrumente GmbH, Munich, Germany) through a combined paediatric heat-exchanger-oxygenator (Midiflo D705; Sorin Dideco, Mirandola, Italy). Blood flow was estimated using the Stöckert device, from the number of rotations per minute and the internal diameter of the tubing. To be sure that this device gave adequate measures, we controlled the number of rotations per minute before the beginning of the experimental protocol with a stroboscopic device provided by the manufacturer (Stöckert Instruments GmbH, Munich, Germany) for the non-pulsatile flow mode. For the pulsatile flow mode, pump flow was controlled using an ultrasonic blood flowmeter (Transonic System, Ithaca, NY, USA, pulsatile blood flow). Correlation for both estimates was very good (Pearson coefficient: 0.997 and 0.993, respectively), so that the Stöckert device was used to measure blood flow during the protocol in both groups. The tubing in the pump race was 0.5 in. in diameter for all procedures and the same roller pump was used in all experiments.
The CPB circuit was primed with a 6% hydroxyethyl starch solution with an average molecular weight 450 000 and a molar substitution ratio 0.7 (Plasmasteril, Fresenius, Bad Homburg, Germany), and fresh dog red packed cells to obtain a solution with a haemoglobin concentration between 90 and 100g L−1. Sodium bicarbonate was added to obtain a pH around 7.35. After checking that anticoagulation was adequate (activated coagulation time (ACT) > 450 s), CPB was started at a flow rate of 100 mL min−1 kg−1 in both groups. The aorta was then clamped and the heart stopped using 200 mL of cardioplegic solution (Ringer's lactate with 30 mmoL L−1 KCl). A left ventricular venting sump was then inserted. The dogs were randomly assigned to receive either continuous flow (Group C: n = 9) or pulsatile flow (Group P: n = 9). The Cobe-Stöckert pump system (München, Germany) was used in all cases with pulsatile and non-pulsatile perfusion being used according to the randomization. Pulsatile flow controls were set to provide a rate of 80 beats min−1, with a pump run-time of 50-55% of the total cycle length. CPB was performed at 37°C. Gas flow and inspired fraction of oxygen were adapted throughout the experiment to maintain PaCO2 between 4 and 5 kPa and PaO2 above 13.5 kPa. Arterial pH, PCO2 and PO2, and mixed venous oxygen saturation (SvO2) were continuously monitored (SW0200 and gas stat monitors, Baxter-Bentley, Irvine, CA, USA). To compensate for insensible losses during the experimental procedure, each dog received a saline infusion at a rate of 10 mL kg−1 during the first hour and 1 mL kg−1 h−1 thereafter. ACT was maintained above 450 s throughout the procedure, by additional doses of heparin (1-2 mg kg−1). Blood temperature was kept constant at 37°C.
After aortic cross clamping, a 15 min period was allowed to achieve a steady state, as defined by stable mean arterial pressure, arterial pH and SvO2. Baseline measurements were then recorded, including measurements of arterial pressure and pump flow. Immediately thereafter, arterial and venous blood samples were drawn from the arterial and venous lines for the measurement of blood gas tensions (ABL 2, Radiometer, Copenhagen, Denmark), haemoglobin concentration, and oxygen saturation (co-oximeter OSM3, Radiometer). Each sample was analysed at least twice, with less than 5% variability. Serum lactate concentration was assessed enzymatically (Kontron analyser, Basel, Switzerland).
Arterial oxygen content (CaO2), mixed venous oxygen content (CvO2), DO2, VO2 and O2ER, arterio-venous difference for pH (AV pH) and veno-arterial difference for PCO2 (VAPCO2) were calculated using the following formulas:
The pump flow was then progressively reduced by 5-10 mL min−1 kg−1 down to 20 mL min−1 kg−1. At each step, a 10 min period was allowed to achieve a new steady state before another set of measurements was obtained.
Haemodynamic and blood parameters obtained at baseline, at the experimental point nearest the critical point, and at the final stage of the experiment were compared between the two groups, using a two-way analysis of variance, followed by a Tukey's test when statistically significant. In each animal, the VO2/DO2 relationship was analysed from all experimental points. Critical DO2, defined as the DO2 value below which VO2 becomes supply dependent, was determined using the method described by Samsel and Schumacker .
Paired sets of linear regressions were calculated, after sorting the DO2 and VO2 paired values with increasing DO2, for all possible combinations of points separated into low (supply dependent) and high (supply independent) supply groups. Points were constrained to fall on either regression line, but never on both. The pair of regressions with the lowest sum of the standard errors of estimate was taken as the set that best fit the data. DO2and VO2 values at the intersection point were then calculated using the two regression equations, to obtain critical DO2 (DO2crit) and VO2 (VO2crit). The O2ER at the critical point (O2ERcrit) was calculated by dividing VO2crit by DO2crit. DO2crit was also determined in each animal from a plot of blood lactate vs. DO2, using the same computing method described above. This value was called DO2 (Lac). Critical pump flow was determined in each animal from a plot of blood lactate vs. pump flow using the same computing method.
Data obtained at the critical point were compared using a U-test. The evolution of the O2ER during the progressive decrease in pump flow was compared in the two groups using a two-way analysis of variance for repeated measurements. For all tests, a P-value <0.05 was considered as significant. All values are expressed as mean ± SD.
Haemodynamic and blood variables obtained at baseline, near critical point and at the end of the procedure are summarized in Table 1. The administration of a pulsatile flow resulted in a pulse pressure of 37 ± 18 mmHg. According to the protocol, haemoglobin and arterial oxygen saturation remained stable throughout the experimental procedure. The relationship between DO2 and VO2, and between DO2 and blood lactate in one dog from the non-pulsatile and one from the pulsatile group is shown in Figure 1. With the progressive decrease in pump flow, oxygen extraction increased in both groups (Fig. 2). Below critical DO2, blood lactate concentrations increased abruptly. There were no significant differences between the two groups in any measured parameters.
At critical point, there was no significant difference between the two groups in DO2crit, VO2crit, or O2ERcrit (Table 2). In both groups, DO2crit obtained from blood lactate measurements (DO2crit (Lac)) were comparable to those obtained from VO2 measurements. Below the critical value of pump flow, blood lactate concentrations increased abruptly. The critical pump flow determined from blood lactate measurements was similar in the non-pulsatile and in the pulsatile groups (Table 2).
This study identified the critical DO2 and pump flow during pulsatile and non-pulsatile 37°C CPB in anaesthetized dogs. Under these experimental conditions, critical DO2 was found to be 7-8 mL min−1 kg−1 and critical pump flow 50-70 mL min−1 kg−1. The critical DO2 values obtained from blood lactate measurements were similar to those obtained from VO2 measurements. These values were not influenced by the pulsatile mode used.
In each animal, arterial oxygen content was maintained constant and DO2 was exclusively reduced by a progressive decrease in pump flow. In the presence of a stable tissue oxygen demand, critical DO2 and, therefore, critical pump flow closely depend on the ability of the tissues to extract oxygen; the higher the critical O2ER, the lower the critical DO2 and pump flow. Blood pressure (BP) was not maintained constant during the reduction of pump flow, therefore triggering the baroreflex and other neuro-humoral reflexes influencing the redistribution of blood flow among organs. This may result in a selective decrease in blood flow from organs such as the kidneys and the splanchnic viscera to the heart (not in our model) and the brain during critically low whole body DO2. Schlichtig and colleagues  studied the impact of flow redistribution on the critical DO2 in pentobarbital anaesthetized dogs submitted to progressive haemorrhage. They created a mathematical model wherein each organ-to-whole body DO2 ratio remained approximately constant. This model predicted that oxygen supply dependency without redistribution of blood flow would have begun at a higher value of whole body DO2. We could not study this in our model, as maintenance of BP during a progressive decrease in blood flow would have necessitated the use of vasopressor agents, which, themselves, would have influenced organ blood flow distribution and tissue oxygen demand. We have demonstrated this effect previously in studies on VO2/DO2relationships [14,15].
Values of systemic critical DO2 observed with our model were similar to those obtained in anaemic , haemorrhagic  or hypoxic models . Critical O2ER appeared somewhat lower (60%), which may be due to the exclusion of the heart and the lungs from the circulation during CPB. Indeed, the heart extracts most of the consumable oxygen available even under basal conditions and cannot increase oxygen extraction further to avoid anaerobic metabolism . Therefore, exclusion of this organ with such a high extraction rate (above 70%) may have influenced the systemic critical oxygen extraction value, which represents the sum of the oxygen extraction rates of all vascular beds. An alternative explanation may be an alteration in tissue oxygen extraction capabilities by the inflammatory reaction induced by the CPB .
As in other models of acute circulatory failure, blood lactate levels increased abruptly when DO2 and pump flow decreased below a critical value [19-21]. Although it is not absolute serum lactate levels but the lactate/pyruvate ratio that reflects an anaerobic state in the tissues , the fact that the abrupt increase in blood lactate was associated with a metabolic acidosis (decrease in pH and in PaCO2) suggests the development of oxygen imbalance in the tissues. Blood lactate levels during and after CPB may increase without compromise in total systemic or even splanchnic DO2, the most plausible explanation being a systemic inflammatory response to CPB . However, the fact that critical DO2values determined from VO2 calculation were similar to those determined from blood lactate values indicates that repeated measurements of this easily obtained parameter can be useful to detect insufficient blood flow during normothermic CPB.
Pulsatile blood flow, through the modified roller pump developed by Cobe-Stöckert, did not result in any improvement in oxygen extraction capabilities as demonstrated by a similar critical DO2 and pump flow. Failure to observe such an effect could be explained by the much lower pulsatile power generated by the system we used as compared to the human heart [23,24]. Whether other systems, able to generate power outputs similar to those of the human heart, could improve tissue oxygen capabilities resulting in a lower critical pump blood flow remains to be determined.
Haemoglobin concentration was maintained around 90-100g L−1, a value considered as an optimal compromise between the oxygen carrying capacity of the blood and blood flow [25,26]. In addition, acute haemodilution to these haematocrit levels has been associated with an increase in tissue oxygen extraction capabilities in anaesthetized dogs submitted to haemorrhage  or endotoxic shock . However, decreasing haemoglobin concentration in these otherwise stable conditions could be expected to be associated with an increase in critical pump flow. Liam and colleagues  reported that a haematocrit level greater than 18% was needed to maintain systemic DO2 and VO2 during normothermic CPB.
Critical DO2 and critical pump flow also depend on the level of tissue oxygen demand. Anaesthetic agents and myorelaxant drugs were administered continuously throughout the experiment to maintain a stable anaesthetic level. The temperature was kept constant at 37°C. Moderate to severe hypothermia is still used during CPB to decrease tissue oxygen demand, and one may expect that, for a given critical O2ER, critical DO2 and pump flow would be lower in these conditions. This has indeed been observed by some  but not all authors [31,32]. This apparent controversy might be explained by a deleterious effect of hypothermia on tissue oxygen extraction capabilities , through an increase in blood viscosity and a higher haemoglobin affinity for oxygen. We chose to maintain the dogs' temperature at 37°C, although real normothermia in dogs may correspond to 38°C, to avoid the alpha stat-pH stat controversy.
A possible limitation of our study was that DO2 and VO2 were obtained from the same pump flow measurement, which could have resulted in some mathematical coupling of the data . However, this phenomenon has been predominantly evoked in clinical studies where only two to three sets of measurements were performed and the cardiac output was measured by the thermodilution technique. In our study, 15 sets of measurements were performed in each animal, with pump flow varying from 100 to 20 mL min−1 kg−1. Pump flow was measured with an electronic system that had been controlled prior to the study either by a stroboscopic device or by an ultrasonic flowmeter and shown to correlate well. In addition, the same pump race tubing was used for each of the 18 experiments. Hence, we do not believe this limitation played a major role. It should also be noted that in the present study only systemic DO2 was determined. Since different organs and systems may reach supply dependency at different values of DO2, the threshold for regional anaerobic metabolism may be higher than the global critical value of DO2 .
In conclusion, a DO2 above 7-8 mL min−1 kg−1 was required to maintain systemic VO2 during 37°C CPB in anaesthetized dogs. The use of a pulsatile blood flow through the modified roller pump developed by Cobe-Stöckert did not influence this critical DO2 value. Elevation of blood lactate levels during bypass helps to identify inadequate tissue DO2 related to insufficient pump flow.
This work was conducted at Erasme University Hospital, Brussels, Belgium, and supported by a grant from the Erasme Foundation, Brussels, Belgium.
1. Schumacker PT, Cain SM. The concept of a critical oxygen delivery
. Intens Care Med
2. Cain SM. Acute lung injury. Assesment of tissue oxygenation. Crit Care Clin
3. Mangano CM, Hill L, Cartwright CR, Hindman BJ. Cardiopulmonary bypass
and the anesthesiologist. In: Kaplan JA, Reich DL, Konstadt SN, eds. Cardiac Anesthesia
. Philadelphia: WB Saunders, 1999: 1061-1110.
4. Cavaliere F, Gennari A, Martinelli L, Zamparelli R, Schiavello R. The relationship between systemic oxygen uptake and delivery during moderate hypothermic cardiopulmonary bypass
: critical values and effects of vasodilation by hydralazine. Perfusion
5. Van der Linden P, Schmartz D, De Groote F et al
. Critical haemoglobin concentration in anaesthetized dogs: comparison of two plasma substitutes. Br J Anaesthesiol
6. Cain SM. Oxygen delivery
and uptake in dogs during anemic and hypoxic hypoxia. J Appl Physiol
7. Van der Linden P, Gilbart E, Engelman E, Schmartz D, Vincent JL. Effects of anesthetic agents on systemic critical O2
delivery. J Appl Physiol
8. Schlichtig R, Kramer DJ, Pinsky MR. Flow redistribution during progressive hemorrhage is a determinant of critical O2
delivery. J Appl Physiol
9. Nelson DP, Beyer C, Samsel RW, Wood LDH, Schumacker PT. Pathological supply dependence of O2
uptake during bacteremia in dogs. J Appl Physiol
10. Bredle DL, Samsel RW, Schumacker PT, Cain SM. Critical O2
delivery to skeletal muscle at high and low PO2
in endotoxemic dogs. J Appl Physiol
11. Hickey PR, Buckley MJ, Philbin DM. Pulsatile and nonpulsatile cardiopulmonary bypass
: review of counterproductive controversy. Ann Thorac Surg
12. Nakamura K, Harasaki H, Fukumura F, Fukamachi K, Whalen R. Comparison of pulsatile and non-pulsatile cardiopulmonary bypass
on regional renal blood flow in sheep. Scand Cardiovasc J
13. Samsel RW, Schumacker PT. Determination of the critical O2
delivery from experimental data: sensitivity to error. J Appl Physiol
14. Zhang H, Smail N, Cabral A, Rogiers P, Vincent JL. Effects of norepinephrine on regional blood flow and oxygen extraction capabilities during endotoxic shock. Am J Respir Crit Care Med
15. Zhang H, DeJongh R, De Backer D, Cherkaoui S, Vray B, Vincent JL. Effects of alpha and beta adrenergic stimulation on hepatosplanchnic perfusion and oxygen extraction in endotoxic shock. Crit Care Med
16. Cilley RE, Scharenberg AM, Bongiorno PF, Guire KE, Barlett RH. Low oxygen delivery
produced by anemia, hypoxia and low cardiac output. J Surg Res
17. Jan KM, Chien S. Effect of hematocrit variations on coronary hemodynamics and oxygen utilization. Am J Physiol
18. Miller BE, Levy JH. The inflammatory response to cardiopulmonary bypass
. J Cardiothorac Vasc Anesth
19. Simmons DH, Alpas AP, Tashkin DP, Coulson A. Hyperlactatemia due to arterial hypoxemia, or reduced cardiac output, or both. J Appl Physiol
20. Cain SM. Appearance of excess lactate in anesthetized dogs during anemic and hypoxic hypoxia. Am J Physiol
21. Moss M, Moreau G, Lister G. Oxygen transport and metabolism in the conscious lamb. The effects of hypoxemia. Pediatr Res
22. Landow L. Splanchnic lactate production in cardiac surgery patients. Crit Care Med
23. Wright G. Hemodynamic analysis could resolve the pulsatile blood flow controversy. Ann Thorac Surg
24. Wright G. Factors affecting the pulsatile hydraulic power output of the Stockert roller pump. Perfusion
25. Messmer K, Kreimeier U, Intaglietta M. Present state of intentional haemodilution. Eur Surg Res
26. Mirhashemi S, Messmer K, Intaglietta M. Tissue perfusion during normovolemic hemodilution investigated by a hydraulic model of the cardiovascular system. Int J Microcirc Clin Exp
27. Van der Linden P, Gilbart E, Paques P, Simon C, Vincent J-L. Influence of hematocrit on tissue O2
extraction capabilities in anesthetized dogs during acute hemorrhage. Am J Physiol
28. Creteur J, Sun Q, Abid O, De Backer D, Van Der Linden P, Vincent JL. Normovolemic hemodilution improves oxygen extraction capabilities in endotoxic shock. J Appl Physiol
29. Liam BL, Plöchl W, Cook DJ, Orszulak TA, Daly RC. Hemodilution and whole oxygen balance during normothermic cardiopulmonary bypass
in dogs. J Thorac Cardiovasc Surg
30. Sinard JM, Vyas D, Hultquist K, Harb J, Bartlett RH. Effects of moderate hypothermia on O2
consumption at various O2
deliveries in a sheep model. J Appl Physiol
31. Cain SM, Bradley WE. Critical O2
transport values at lowered body temperatures in rats. J Appl Physiol
32. Willford DC, Hill EP, White FC, Moores WY. Decreased critical mixed venous oxygen tension and critical oxygen transport during induced hypothermia in pigs. J Clin Monit
33. Van der Linden P, De Groote F, Bélisle S, Mathieu N, Willaert P. Effects of hypothermia on tissue O2
extraction capabilities in dogs. Anesth Analg
34. Moreno LF, Stratton HH, Newell JC, Feustel PJ. Mathematical coupling of data: correction of a common error for linear calculations. J Appl Physiol
35. Boston US, Slater JM, Orszulak TA, Cook DJ. Hierarchy of regional oxygen delivery
during cardiopulmonary bypass
. Ann Thorac Surg