During acute normovolemic hemodilution, the hematocrit is usually kept >25%, but values of 9% have been reported [1-4]. The decrease in hemoglobin concentration and the concomitant decrease in oxygen-carrying capacity are partially compensated for by increases in blood flow and oxygen extraction [5-8]. Anemia does, however, encroach on the margins for oxygen delivery, and maintenance of normovolemia is essential for the tolerance to acute hemodilution .
Tarnow et al. [10,11] studied the effect of a 15-mL/kg blood loss in dogs hemodiluted to 16% and found that the animals tolerated this degree of hypovolemia well. In fact, hemodiluted animals showed improved tolerance to hemorrhage compared with animals with normal hematocrit when judged by critical oxygen delivery because of an increase in oxygen extraction capability, probably related to the decreased blood viscosity . However, the effects of hypovolemia during more severe hemodilution have not been reported. We designed the current study to assess possible differences in tolerance to graded blood loss between anesthetized, extremely hemodiluted pigs and pigs with normal hematocrit. We hypothesized that during acute extreme hemodilution to a hematocrit of 11%, in contrast to during moderate hemodilution (hematocrit >15%), tolerance to blood loss would be reduced when judged by systemic oxygen uptake, arterial lactate concentration, and myocardial lactate uptake. In addition, we evaluated indicators of hypovolemia.
After approval of the local animal investigations committee, 12 Swedish landrace pigs (weighing 26.9 +/- 3.4 kg) were studied. The animals were fasted overnight but had free access to water. They were premedicated with 15 mg IM midazolam; anesthesia was induced with 10-15 mg/kg IV ketamine and maintained with an infusion of 20 [micro sign]g [center dot] kg-1 [center dot] h-1 fentanyl, 0.3 mg [center dot] kg-1 [center dot] h-1 midazolam, and 0.3 mg [center dot] kg-1 [center dot] h-1 pancuronium. Additional fentanyl (20 [micro sign]g/kg) and local lidocaine were administered before insertion of central catheters.
A cuffed endotracheal tube was placed through a tracheostomy, and the lungs were mechanically ventilated by using a Servo 900 ventilator (Siemens-Elema, Psolna, Sweden) delivering an inspired oxygen fraction of 0.35, a respiratory rate of 20 breaths/min, and 5 cm H2 O of positive end-expiratory pressure. Inspired oxygen fraction and end-tidal carbon dioxide were measured continuously. The ventilation was adjusted to give an end-tidal carbon dioxide concentration of 4.0%-4.5%. Core temperature was maintained in the normal range (38.5-39.5[degree sign]C) with blankets. Acetated Ringer's solution to which 20 g of glucose/L had been added was infused IV at a rate of 5 mL [center dot] kg-1 [center dot] h-1 throughout the experiment. A bladder catheter was inserted via a cystostomy.
Catheters were placed in the cranial caval vein for the administration of anesthetics and blood replacement, and in the right carotid artery and the pulmonary artery for blood sampling and measurements of blood pressure and cardiac output. Finally, a thermistor catheter (Webster Laboratories, Altadena, CA) was placed for measurements of cardiac venous flow and for sampling of blood. The catheter tip was positioned in the great cardiac vein 3 cm upstream of its confluence with the azygos vein.
Catheters were inserted through peripheral cut-downs, and their positions were confirmed by fluoroscopy, which was repeated regularly. In addition, correct catheter position in the great cardiac vein was verified by aspirating blood with a hemoglobin oxygen saturation of approximately 20%. Pressures were measured by fluid transmission and transducers.
Cardiac output was measured in triplicate by thermodilution, using 10-mL injections of room-temperature isotonic glucose in a separate, centrally placed catheter. Flow in the great cardiac vein, which mainly drains the left ventricle , was measured by continuous retrograde thermodilution as described by Ganz et al. . This technique has good reproducibility if the flow rate of the indicator is sufficiently high and if the catheter is not dislocated between measurements . We therefore used a constant rate infusion pump delivering 54 mL/min of isotonic saline at room temperature over approximately 20 s, and we fixed the catheter with a ligature at its entrance into the external jugular vein. The temperatures of the indicator and of blood in the great cardiac vein were used to calculate cardiac venous flow (vide infra). Corrections were not made for possible underestimation of flow caused by thermoconductivity within the catheter because the error at the flows measured was expected to be <3% .
Blood samples were drawn simultaneously from the carotid artery, the pulmonary artery, and the great cardiac vein, on which analysis of blood gases were performed. Hemoglobin concentration and oxygen saturation were obtained spectrophotometrically, and the values were adjusted to porcine blood. The arterial hematocrit was obtained with a microhematocrit centrifuge. Lactate was determined in arterial and cardiac venous blood samples by using a spectrophotometric method .
Blood volume was measured by labeled albumin using the125 I-albumin technique : a blood sample (8 mL) was drawn for determination of background activity, and the calculated amount of125 I-albumin was injected (0.0015 mBq/kg) in 10 mL of sterile water through a separate IV catheter. To reduce the amount of (125) I-albumin that adhered to the syringe and the injection port, the syringe and the catheter were carefully flushed with saline. Residual activity in the syringe was measured. Blood samples for determination of125 I-albumin activity were obtained after 10, 20, and 30 min. Determination of the blood volume was performed during three stages: at the outset, after hemodilution (after a comparable period of time in the control animals), and after a blood loss of 20 mL/kg. To reduce the influence of activity from previously injected125 I-albumin during the repeated measurements of blood volume, the injected amount of125 I-albumin was increased by a factor of 3 for each new determination of blood volume (to 0.0045 and 0.0135 mBq/kg, respectively). Activity in the obtained samples was counted to calculate blood volume. To determine whether the leakage of125 I-albumin from the intravascular to the extravascular compartment influenced blood volume determinations, the counts obtained 10, 20, and 30 min after125 I-albumin injection were plotted for comparison. The obtained curves indicated stable conditions. Blood volume was calculated as the relation between total amount of125 I-albumin injected and activity in the obtained sample. Corrections were made for the measured baseline activity at each blood volume determination and for the residual activity in the syringe used for injection.
Flow in the great cardiac vein (FGCV) was calculated assuming that heat lost from the indicator was gained by blood  as: Equation 1 where F (I) = indicator flow and T = temperature of indicator (I), blood (B), and the indicator and blood mixture (M). The value 1.08 is the relation between the density (S) and the specific heat (C) of blood and indicator ([SI [center dot] CI]:[SB [center dot] CB]).
The oxygen content of blood, systemic and myocardial (left ventricular) oxygen delivery and uptake, and oxygen extraction ratio were calculated .
After the preparation, which lasted 60-90 min, the animals were undisturbed for at least 30 min before baseline measurements were made. The animals were then randomly assigned to either of two time-matched groups of six animals each: one was hemodiluted, the other served as control.
Hemodilution was performed during a 30-min period by removing blood from the arterial catheter during simultaneous substitution with the same volume of a warmed (38[degree sign]C) 1:1 mixture of 6% dextran-70 and acetated Ringer's solution . A similar mixture gives isovolemic plasma expansion in humans . The mean (+/- SD) volume exchanged to achieve a hematocrit of 11% +/- 1% was 61 +/- 7 mL/kg (51-72 mL/kg). Fifteen minutes after hemodilution, new measurements were made.
Both groups were then exposed to an uncompensated blood loss of 40 mL/kg. The decrease in blood volume was accomplished in steps of 10 mL/kg. Measurements were made at each level after 10 min.
Two-way (group and stage) analysis of variance with repeated measures was applied for continuous variables (saturation, FGCV, etc) to determine whether there was any significant overall difference within or between the groups. If a difference was found, Fisher's protected least significant difference test was used to assess whether this was significant. The relationship between filling pressures and blood volume was studied by using linear regression. Probability values <0.05 were considered significant. Data are reported as means +/- SD when not otherwise indicated.
Hemodilution reduced the oxygen-carrying capacity and, despite a 31% increase in cardiac output, systemic oxygen delivery decreased, whereas systemic oxygen uptake was unchanged (Table 1 and Table 2). The systemic oxygen extraction ratio increased, which was reflected in a decrease in mixed venous oxygen saturation (Table 2). Arterial lactate increased, but arterial pH was unchanged. Systemic vascular resistance decreased, causing a decrease in mean arterial pressure (Table 1).
Cardiac venous blood flow was 3 times higher in hemodiluted pigs compared with control animals (Table 3). Myocardial oxygen delivery and uptake and myocardial lactate uptake did not differ between the groups.
After a blood loss of 10 mL/kg, systemic oxygen uptake was maintained by a further increase in systemic oxygen extraction, but arterial lactate increased and arterial base excess decreased despite this (Table 2). The high extraction rate was also reflected in a significant decrease in mixed venous oxygen saturation (Table 2).
Heart rate was unchanged at a blood loss of 10 mL/kg, but it increased in hemodiluted animals when the blood loss reached 20 mL/kg (Table 1). Cardiac output decreased in both groups but remained greater in the hemodiluted animals until blood loss exceeded 30 mg/kg. Systemic and pulmonary arterial blood pressure decreased in both groups, and systemic arterial blood pressure and vascular resistance were continuously lower in hemodiluted animals (Table 1). Central venous pressure and pulmonary capillary wedge pressure were 3 and 2 mm Hg lower at a blood loss of 20 mg/kg than after hemodilution, whereas these measures were unchanged in the control group (Table 1). There was no correlation between blood volume and central venous pressure or pulmonary capillary wedge pressure (Figure 1). Likewise, no correlation was found between changes in blood volume and changes in central venous pressure or pulmonary capillary wedge pressure.
Cardiac venous blood flow continued to be greater in the hemodiluted animals (Table 3), and myocardial oxygen delivery and uptake were maintained. Coronary venous oxygen saturation did not differ from the controls, and lactate was not produced.
When the blood loss reached 30 mg/kg, two hemodiluted animals died of circulatory failure preceded by a period of progressive hypotension. Arterial lactate in the hemodiluted animals was greater than that in the controls (Table 2), and myocardial lactate production was observed (Table 3). Mean arterial blood pressure and mixed venous oxygen saturation were similar in both groups (Table 1 and Table 2).
When the blood loss reached 40 mg/kg, one animal in the hemodiluted and one animal in the control group died of circulatory failure. The data obtained during this stage were therefore obtained in only three and five animals, respectively. In the hemodiluted group, systemic oxygen delivery and uptake decreased (Table 2). Arterial lactate increased simultaneously and arterial base excess decreased (Table 2). Despite a significant increase in systemic vascular resistance in the hemodiluted animals, mean arterial pressure decreased further, and cardiac output became lower than in the control animals (Table 1). Central venous pressure and pulmonary capillary wedge pressure were similar in both groups and were unchanged compared with previous stages.
No hemodiluted animal survived a blood loss of 40 mg/kg, although five animals with normal hematocrit survived. To elucidate the response to a further decrease in systemic oxygen delivery, these animals were exposed to a blood loss of 50 mg/kg. Only one animal survived this blood loss.
Coronary venous flow in the hemodiluted animals was greater than in the control animals until the blood loss reached 40 mg/kg (Table 3), but no difference was observed at this point. In the hemodiluted animals, myocardial lactate production started at a blood loss of 30 mg/kg and increased at 40 mg/kg (P < 0.01).
During uncompensated progressive blood loss in acute, extremely hemodiluted, anesthetized pigs, we found that: 1) mixed venous oxygen saturation decreased to a critical value and arterial lactate concentration increased significantly at a blood loss of 10 mL/kg; 2) myocardial lactate uptake was maintained until the blood loss exceeded 20 mL/kg; 3) an unchanged systemic vascular resistance (up to a blood loss of 30 mg/kg) and a decreased cardiac output lowered the arterial blood pressure; 4) heart rate did not increase until the blood loss exceeded 10 mg/kg; and 5) neither central venous pressure nor pulmonary capillary wedge pressure changed markedly.
Pigs were chosen for experimental animals because their cardiovascular physiology is similar to that of humans . We used a continuous infusion of fentanyl, midazolam, and pancuronium in both groups throughout the experiment. This technique has previously been used by Woerkens and coworkers  in pigs, and no signs of cardiovascular depression were observed during hemodilution to a hematocrit of 9%. The hematocrit was reduced to 11%, i.e., to approximately one-third the normal porcine level. The compensatory increase in cardiac output, the reduced arterial pressure, and the reduced systemic vascular resistance after hemodilution confirm previous findings [8,21]. The increased arterial lactate concentration, the high oxygen extraction rate, and the low mixed venous oxygen saturation (Table 2) all indicate that systemic oxygen delivery was close to the critical value after hemodilution [22,23]. No animal showed myocardial lactate production after hemodilution, which indicates that myocardial oxygen demand was met.
The blood volume was unchanged after hemodilution. This confirms findings by Schott et al.  and indicates that dextran-acetated Ringer's solution provides isovolemic volume replacement (Table 1). The amount of maintenance fluid seemed to be adequate because neither the hematocrit nor hemoglobin levels changed in the control animals during the first and second stages (Table 1). The observed reduction in blood volume in control animals was probably caused by splenic accumulation of blood, which cannot be measured by conventional means because of its low exchange rate with the circulating blood . When the adrenergic system is stimulated, e.g., when oxygen delivery is compromised, the stored blood reenters the circulation . In hemodiluted animals, it is likely that the adrenergic system had already been activated during the hemodilution procedure, reducing the amount of blood stored in the spleen. The blood volume initially measured agreed with that previously reported , and after a blood loss of 20 mg/kg, the measured blood volumes were consistent with those expected.
We used lactate concentration as an indicator of insufficient oxygen delivery during hypovolemia in hemodiluted animals because this measure represents an independent monitor of tissue oxygenation. The maintained myocardial lactate uptake in hemodiluted animals when the blood loss reached 20 mg/kg (Table 3) indicated that myocardial blood flow was adequate to meet oxygen demand at this level of blood loss. Hence, the healthy heart may tolerate considerable blood loss during hemodilution possibly because the heart, in contrast to most other organs, is spared from redistributive vasoconstriction during hypovolemia [24,25]. Animals with normal hematocrits tolerated a blood loss of 30 mg/kg before arterial lactate concentration produced signs of tissue hypoxia [26,27]. A plot of individual values indicates that a close relation between low systemic oxygen delivery and increasing arterial lactate concentration, as well as decreasing systemic oxygen uptake, occurred at a systemic (critical) oxygen delivery of approximately 10 mg kg-1 [center dot] min-1 in hemodiluted animals (Figure 2). In a previous study, we found a critical systemic oxygen delivery of 10.4 mg [center dot] kg-1 [center dot] min-1 in hemodiluted animals, below which value arterial lactate concentration increased significantly . This value is in agreement with that observed by others during anemia, hypoxia, and low cardiac output [22,23].
There was a similar relation between low mixed venous oxygen saturation and increased arterial lactate concentration (Figure 3). In hemodiluted animals, arterial lactate increased and mixed venous oxygen saturation decreased below 30% already at a blood loss of 10 mg/kg (Table 2). This suggests that mixed venous oxygen saturation may be a useful indicator of tissue hypoxia during blood loss in hemodiluted animals. This is in contrast to the results obtained in hemodiluted animals during hypoxia, in which mixed venous oxygen saturation was a less reliable indicator of inadequate oxygen delivery .
When the blood loss exceeded 10 mg/kg, there was only a small increase in the systemic oxygen extraction ratio in the hemodiluted animals, which indicates that the oxygen extraction was already maximal in most tissue (Table 2). This was also reflected in the observation that only a small decrease in central venous oxygen saturation occurred, except as a terminal event when the blood loss reached 40 mg/kg. Thus, we could not confirm the findings of Van der Linden and coworkers  that hemodilution increases the tolerance to hemorrhage. It should be noted, however, that their experiment was performed at greater hematocrit values (20% and 30%).
A problem during clinical hemodilution is how to detect early signs of inadequate oxygen delivery caused by hypovolemia. Arterial lactate concentration is difficult to monitor online. Monitoring of mixed venous oxygen saturation is appealing, but it may not always be appropriate to introduce a pulmonary artery catheter .
Central venous pressure and pulmonary capillary wedge pressure decreased somewhat during blood loss (Table 1). However, our results confirm previous studies in both humans and pigs with normal hematocrit that suggest that these measures are not sensitive enough to clearly indicate changes in blood volume during hemodilution [29,30]. In our study, a decrease in blood pressure, not an increase in heart rate, was the first sign of hypovolemia in hemodiluted animals (Table 1). When the heart rate increased at a blood loss of 20 mg/kg, tissue oxygenation was already compromised, as indicated by an increased arterial lactate concentration. This observation agrees with that obtained in anesthetized children and in pigs during severe hemorrhagic shock [29,31-33]. The findings differ from those of Vatner  and of Foex and coworkers  in adult humans, dogs, and primates, in which the first signs of hypovolemia were increases in both heart rate and systemic vascular resistance without any change in blood pressure. Thus, the primary circulatory response to hypovolemia seems species- and age-dependent and may be affected by the anesthetic technique.
1. Robertie PG, Gravlee GP. Safe limits of isovolemic hemodilution and recommendations for erythrocyte transfusion. Int Anesthesiol Clin 1990;28:197-204.
2. Perez de Sa V, Bekassy A, Schou H, et al. Hemodilution during bone marrow harvesting in children. Anesth Analg 1991;72:645-50.
3. Perez de Sa V, Bekassy A, Schou H, Werner O. Bone marrow harvesting in children managed without allogenic blood. Paediatr Anaesth 1994;4:375-81.
4. Fontana JL, Welborn L, Mongan PD, et al. Oxygen consumption and cardiovascular function in children during profound intraoperative normovolemic hemodilution. Anesth Analg 1995;80:219-25.
5. Jan K-M, Chien S. Effect of hematocrit variations on coronary hemodynamics and oxygen utilization. Am J Physiol 1977;233:H106-13.
6. Wilkerson DK, Rosen AL, Gould SA, et al. Oxygen extraction ratio: a valid indicator of myocardial metabolism in anemia. J Surg Res 1987;42:629-34.
7. Restorff W, Hofling B, Holtz J, Bassenge E. Effect of increased blood fluidity through hemodilution on coronary circulation at rest and during exercise in dogs. Pflugers Arch 1975;357:15-24.
8. Restorff W, Hofling B, Holtz J, Bassenge E. Effect of increased blood fluidity through hemodilution on general circulation at rest and during exercise in dogs. Pflugers Arch 1975;357:25-34.
9. Messmer K, Sunder-Plassmann L, Klovekorn WP, Holper K. Circulatory significance of hemodilution: rheological changes and limitations. Adv Microcirc 1972;4:1-77.
10. Tarnow J, Eberlein HJ, Hess W, et al. Hemodynamic interactions of hemodilution, anesthesia, propranolol pretreatment and hypovolaemia. I. Systemic circulation. Bas Res Cardiol 1979;74:109-22.
11. Tarnow J, Eberlein HJ, Hess W, et al. Hemodynamic interactions of hemodilution, anesthesia, propranolol pretreatment and hypovolaemia. II. Coronary circulation. Bas Res Cardiol 1979;74:123-30.
12. Van der Linden P, Gilbart E, Paques P, et al. Influence of hematocrit on tissue O2
extraction capabilities during acute hemorrhage. Am J Physiol 1993;264:1942-7.
13. Andersen FR, Sejersted OM, Ilebekk A. A model for quantitative sampling of myocardial venous blood in the pig. Acta Physiol Scand 1983;119:187-95.
14. Ganz W, Tamura K, Marcus HS, et al. Measurement of coronary sinus blood flow by continuous thermodilution in man. Circulation 1971;44:181-95.
15. Reiz S, Haggmark S. Coronary sinus catheterisation and measurement of coronary sinus blood flow by the continuous thermodilution method. In: Prys-Roberts C, Vickers MD, eds. Cardiovascular measurement in anesthesiology. Berlin: Springer, 1982;174-82.
16. Ektachem Clinical Chemistry Slides (LAC). Colorimetric test (lactate). Publication C-212. Rochester, NY: Johnson & Johnson Clinical Diagnostics, 1995.
17. Hannon JP, Bossone CA, Rodkey WG. Splenic red cell sequestration and blood volume measurements in conscious pigs. Am J Physiol 1985;248:R293-301.
18. Schou H, Perez V, Sigurdardottir M, et al. Circulatory effects of hypoxia, acute anemia, and their combination in anesthetized pigs. Anesthesiology 1995;84:1443-54.
19. Schott U, Thoren T, Sjostrand U, et al. Three percent dextran 60 as a plasma substitute in blood component therapy. Acta Anaesthesiol Scand 1985;29:775-81.
20. Weaver ME, Pantely GA, Bristow JD, Ladley HD. A quantitative study of the anatomy and distribution of coronary arteries in swine in comparison with other animals and man. Cardiovasc Res 1986;20:907-17.
21. Woerkens ECSM, Trouwborst A, Duncker DJGM, et al. Catecholamines and regional hemodynamics during isovolemic hemodilution in anesthetized pigs. J Appl Physiol 1992;72:760-9.
22. Cain SM. Oxygen delivery and uptake in dogs during anemic and hypoxic hypoxia. J Appl Physiol 1977;42:228-34.
23. Cilley RE, Scharenberg AM, Bongiorno PF, et al. Low oxygen delivery produced by anemia, hypoxia, and low cardiac output. J Surg Res 1991;51:425-33.
24. Vatner SF. Effects of hemorrhage on regional blood flow distribution in dogs and primates. J Clin Invest 1974;54:225-35.
25. Slater GI, Vladeck BC, Bassin R, Kark AE, Shoemaker WC. Sequential changes in distribution of cardiac output in hemorrhagic shock. Surgery 1973;73:714-22.
26. Dunham CM, Siegel JH, Weireter L, et al. Oxygen debt and metabolic acidemia as quantitative predictors of mortality and severity of the ischemic insult in hemorrhagic shock. Crit Care Med 1991;19:231-43.
27. Simmons DH, Alpas AP, Tashkin DP, Coulson A. Hyperlactatemia due to arterial hypoxemia or reduced cardiac output, or both. J Appl Physiol 1978;45:195-202.
28. Connors AF, Speroff T, Dawson NV, et al. The effectiveness of right heart catheterization in the initial care of critically ill patients. JAMA 1996;276:889-97.
29. Jacobsen J, Sofelt S, Sheikh S, et al. Cardiovascular and endocrine response to haemorrhage in the pig. Acta Physiol Scand 1990;138:167-73.
30. Shippy CR, Appel PL, Shoemaker WC. Reliability of clinical monitoring to assess blood volume in critically ill patients. Crit Care Med 1984;12:107-12.
31. Jacobsen J, Secher NH. Heart rate during hemorrhagic shock. Clin Physiol 1992;12:659-66.
32. Sondeen JL, Gonzaludo GA, Loveday JA, et al. Renal responses to graded hemorrhage in conscious pig. Am J Physiol 1990;259:R119-25.
33. Battersby EF. Monitoring during paediatric anesthesia. In: Sumner E, Hatch DJ, eds. Clinics in anaesthesiology, paediatric anaesthesia. London: WB Saunders, 1985;597-617.
34. Foex BA, Kirkman E, Little RA. Hemorrhage and injury: cardiovascular and regional blood flow responses. In: Vincent JL, ed. Yearbook of intensive care and emergency medicine. Berlin: Springer, 1996:581-94.