Critical haemoglobin concentration in anaesthetized dogs: comparison of two plasma substitutes

Transcription

1 British Journal of Anaesthesia 1998; 81: Critical haemoglobin concentration in anaesthetized dogs: comparison of two plasma substitutes P. VAN DER LINDEN, D. SCHMARTZ, F. DE GROOTE, N. MATHIEU, P. WILLAERT, I. RAUSIN AND J.-L. VINCENT Summary We have explored systemic and regional tolerance to haemodilution during anaesthesia with two different synthetic colloids. Eighteen dogs undergoing mechanical ventilation during anaesthesia with ketamine were submitted to progressive normovolaemic haemodilution with either gelatin (GEL; n 9) or hydroxyethylstarch (HES; n 9) administered on a 1:1 ratio. Systemic oxygen delivery was calculated from measurement of thermodilution cardiac output and arterial oxygen content, while systemic oxygen consumption was determined from expired gas analysis. Mesenteric oxygen delivery and consumption were determined using ultrasonic flow measurements, and arterial and mesenteric venous oxygen contents. The critical haemoglobin concentration (i.e. the haemoglobin value below which oxygen consumption becomes oxygen delivery dependent) was mean 3.6 (SD 0.8) g dl 1 in the GEL and 3.5 (1.5) g dl 1 in the HES group. The mesenteric critical oxygen extraction ratio (O ER) (GEL 50.1 (1.1) %; HES 48.5 (13.4) %) was significantly lower than the systemic critical O ER (GEL 66.1 (8.4) %; HES 67.7 (7.1) %). There were no significant differences between the GEL and HES groups for any of these variables, or in the amount of colloid administered. During the study, oxygen delivery decreased almost linearly with reduction in haemoglobin, indicating a lack of cardiac output response to anaemia during ketamine anaesthesia. (Br. J. Anaesth. 1998; 81: ) Keywords: anaesthetics i.v., ketamine; blood, haemodilution; blood, haemoglobin; oxygen, transport; blood, replacement; dog Increased awareness of the risks associated with allogenic blood has stimulated re-examination of the benefit risk relationship for transfusion therapy. A better knowledge of the physiology and effect of anaemia during and after surgery has allowed a re-evaluation of the magic value of 30% 1 during these periods. However, several important questions remain regarding patient tolerance to acute normovolaemic anaemia. First, the critical concentration of haemoglobin (i.e. the haemoglobin value below which oxygen delivery becomes insufficient to meet tissue oxygen demand in the anaesthetized individual) has not been well defined. Two major mechanisms allow maintenance of adequate oxygen transport to the tissues: an increase in cardiac output and an increase in tissue oxygen extraction. The increase in cardiac output during haemodilution has been well documented in awake and lightly anaesthetized individuals During mild haemodilution, the increase in cardiac output may overcompensate for the decrease in arterial content, so that oxygen transport may increase. 1 However, the ability of the organism to increase oxygen extraction under these conditions has been less well defined. Determination of the critical haemoglobin concentration during normovolaemic haemodilution allows participation of both mechanisms in the maintenance of adequate tissue oxygenation to be explored. 13 Second, redistribution of blood flow during normovolaemic haemodilution from the splanchnic area to the heart and brain has been well described, but the ability of the splanchnic area to extract oxygen during haemodilution has not been well defined. Third, the physicochemical properties of synthetic colloids used to maintain stable volume expansion during acute anaemia may influence the critical haemoglobin concentration in these conditions. Among the different available synthetic colloids, gelatin (GEL) and hydroxyethylstarch (HES) solutions are the most commonly used. Gelatins, obtained from degradation of collagen, have a limited volume expansion effect and a short intravascular half-life They have the advantages of excellent tolerance and low cost, but are associated with a relatively high incidence of anaphylactoid reactions, especially with urea-linked gelatins. 19 HES solutions are modified natural polysaccharides, which have a better volume expansion effect than gelatins. 0 1 Low molecular weight HES solutions are characterized by a shorter intravascular half-life and a smaller incidence of side effects than high molecular weight HES solutions. 0 In this study, we have explored systemic and regional tolerance to progressive normovolaemic haemodilution in anaesthetized dogs using either a PHILIPPE VAN DER LINDEN, MD, PHD, DENIS SCHMARTZ, MD, FRANÇOISE DE GROOTE, MD, NATHALIE MATHIEU, MD, PHILIPPE WILLAERT, MD, ISABELLE RAUSIN, MD (Department of Anaesthesia); JEAN-LOUIS VINCENT, MD, PHD (Department of Intensive Care); Erasme University Hospital, Free University of Brussels, 808 route de Lennik B-1070, Brussels, Belgium. Accepted for publication: April 17, 1998.

2 Critical haemoglobin concentration during anaesthesia 557 modified fluid gelatin or low molecular weight HES. Ketamine anaesthesia was chosen as this agent has well known cardiostimulant properties, 3 and has the least deleterious effects on tissue oxygen extraction capabilities. 4 Materials and methods All experimental procedures were performed according to the NIH guidelines and were approved by our Animal Ethics Care Committee. We studied 18 mongrel dogs (weight kg). After administration of thiopental 0 mg kg 1 i.v., tracheal intubation was performed and mechanical ventilation started on control mode (Elema 900 B, Siemens, Solna, Sweden) with air ( F I O = 0.1). Ventilatory frequency was set at 1 bpm and tidal volume was adapted to obtain Pa CO at kpa. Respiratory conditions remained unchanged until the end of the study. Anaesthesia was maintained with an i.v. bolus dose of ketamine 10 mg kg 1, followed by continuous infusion of 0.4 mg kg 1 min 1. Neuromuscular block was achieved with pancuronium at an initial dose of 0.1 mg kg 1, followed by hourly boluses of 1 mg. Electrodes were attached to the four limbs for monitoring heart rate. Two catheters (16G, Becton Dickinson and CO, Rutherford, NJ) were inserted, one into a peripheral vein for fluid and drug infusions, and one into the distal aorta via a femoral artery for monitoring arterial pressure and sampling arterial blood. A pulmonary artery catheter (Swan Ganz catheter model 93A-131 7F, Baxter Healthcare, Irvine, CA, USA) was inserted via a jugular vein into the pulmonary artery. A gas analyser (Capnomac AGM-103, Datex, Helsinki, Finland) was inserted in the respiratory breathing system for monitoring end-expired carbon dioxide concentration ( P E CO ). Exhaled gases were directed through a mixing chamber for sampling. The expired oxygen fraction ( F E CO ) was measured continuously by a semi-rapid gas analyser (model 500D PK, Morgan Co, Chatham, UK). A splenectomy was performed via a midline laparotomy to prevent blood autotranfusion. During laparotomy, ultrasonic flow probes (Transonic Systems Inc., Ithaca, NY, USA) were inserted, one around the mesenteric artery and one around the left renal artery. A third probe was inserted around the remaining femoral artery, through a small cutaneous incision. Mesenteric, renal and femoral blood flows were measured with a flowmeter (model T08, Transonic Systems Inc., Ithaca, NY, USA; calibration performed by the manufacturer). In seven dogs in each group, an 18-gauge catheter (Becton Dickinson and CO, Rutherford, NJ) was inserted in the mesenteric vein via a peripheral branch. To compensate for insensible losses, the dogs received i.v. infusion of 0.9% NaCl at a rate of 10 ml kg 1 during splenectomy and 1 ml kg 1 h 1 thereafter, until the end of the experiment. Heart rate, arterial pressure and pulmonary artery pressure were monitored continuously (Sirecust, Siemens AG, Erlangen, Germany) and recorded together with P E CO and F E CO (model 8000S, Gould Electronique, Ballainvilliers, France); zero pressure was set at the mid-chest level of the animal. Body temperature was maintained at C throughout the study, using warming blankets and humidified heated gases. Heart rate (HR) and intravascular pressures, including mean arterial pressure (MAP), mean pulmonary artery pressure (MPAP), right atrial pressure (RAP) and pulmonary artery occlusion pressure (PAOP) were measured at end-expiration from the paper trace. E CO P and F E CO were measured from the paper trace over 1 min. Expired volume was measured with a spirometer (model Magtrak II, London, UK) over the same period. After measurement of mesenteric, renal and femoral blood flows, cardiac output (CO) was measured by the thermodilution technique (module E 61/A, Siemens AG, Erlangen, Germany) by repeated injections of 10 ml of cold ( 5 C) 5% glucose in water. Each injection was started at the end of inspiration. Three to five measurements, with a variability of less than 10%, were averaged for each measurement of CO. Immediately thereafter, arterial, mixed venous and mesenteric blood samples were withdrawn for measurement of blood-gas tensions (ABL, Radiometer, Copenhagen, Denmark). Haemoglobin concentration and saturation were measured using a co-oximeter (OSM3, Radiometer, Copenhagen, Denmark). Each sample was analysed twice and the two consecutive values averaged. Blood lactate concentration was measured enzymatically using an automated analyser (Kontron, Basel, Switzerland). Cardiac index (CI), stroke index (SI), systemic and pulmonary vascular resistance (SVR, PVR), left and right ventricular stroke work index (LVSWI, RVSWI) and oxygen delivery (DO ) were calculated using the following formulae: CI (ml min 1 kg 1 ) CO/weight SI (ml kg 1 ) CI/HR SVR (dyn s cm 5 ) 79.9ε(MAP RAP)/CO PRV (dyn s cm 5 ) 79.9ε(MPAP RAOP)/CO LVSWI (g m kg 1 ) ε(MAP PAOP)εSI RVSWI (g m kg 1 ) ε(MPAP RAP)εSI 1 1 DO (ml min kg ) = CI Ca 10 O where C a O =oxygen content of arterial blood. Oxygen consumption VO (ml min 1 kg 1 ) was calculated from measured F E CO and minute ventilation, using the appropriate mass balance equation indexed for weight. 5 Oxygen extraction ratio (O ER) was obtained by dividing VO by DO. Mesenteric oxygen delivery (DO m) was determined from the product of mesenteric blood flow and arterial oxygen content, while mesenteric oxygen consumption (VO m) was obtained by the product of mesenteric blood flow and the difference between arterial and venous mesenteric oxygen contents. Mesenteric oxygen extraction (O ERm) was calculated as: ( CaO Cm O )/ C a O where Cm O oxygen content of mesenteric venous blood. Thirty minutes after completion of the surgical procedure, the dogs were allocated randomly to one of two groups of nine dogs each, according to the plasma substitute used in the subsequent haemodilution procedure. After baseline measurements, each animal was slowly haemodiluted by repeated withdrawals of 5-ml kg 1 alquots of arterial blood simul-

3 558 British Journal of Anaesthesia taneously replaced by the same volume of either 3% modified fluid gelatin (Geloplasma, Propharm, France) (GEL group; n 9) or 6% hydroxyethylstarch 00/0.5 (Haes-Stéril 6%, Fresenius AG, Bad Homburg, Germany) (HES group; n 9). After each exchange procedure, a 15-min period was allowed to reach a new steady state defined by stable mean arte- rial pressure, heart rate, E CO P and F E CO. Then a new set of measurements was performed. In an attempt to maintain normovolaemia during the progressive haemodilution, PAOP was measured after each exchange procedure and maintained at least at the baseline level by additional infusion (1 ml kg 1 ) of the allocated colloid. The exchange procedure was continued until the animal could no longer maintain a stable arterial pressure. Death followed shortly thereafter, with the animal still anaesthetized. STATISTICAL ANALYSIS In each animal, critical DO (DO crit) was determined from a plot of VO vs DO using the dual lines regression method. 6 DO crit was defined as the point of intersection of the two best-fit regression lines as determined by a least sum of squares technique. Critical O ER (O ERcrit) was calculated as the ratio of VO /DO at DO crit. DO crit was also determined in each animal using the same dual regression lines analysis from a plot of blood lactate vs DO. This value was called DO crit(lac). The critical value of haemoglobin (Hbcrit), defined as the value of haemoglobin below which VO began to decrease, was determined using the same mathematical analysis, from a plot of haemoglobin vs VO. In the seven animals in each group where a mesenteric venous catheter was inserted, the mesenteric DO /VO relationship was analysed similarly, from a plot of VO m vs DO m. In each animal, the systemic haemoglobin/do relationship was also analysed, by determining the polynomial that best fitted the data. From these polynomials, DO values corresponding to standardized haemoglobin values were obtained and compared between the two groups. In both groups, haemodynamic and respiratory variables, including regional blood flow, obtained at baseline and at the experimental point closest to critical DO were compared using two-way analysis of variance (time and group), followed by a Tukey test for pairwise comparisons when statistically significant. Critical values of DO and haemoglobin, volume of blood withdrawn and volume of colloid adminis- Table 1 Haemodynamic and oxygen derived data obtained at baseline and near the critical point in the gelatine haemodiluted dogs (group GEL, n 9) and in the hydroxyethylstarch haemodiluted dogs (group HES, n 9) (see text for details). *P 0.05, **P 0.01 vs baseline; P 0.05 HES vs GEL Group Baseline Near critical point Temperature ( C) GEL 37.5 (1.) 36.5 (1.7) HES 37.8 (1.0) 36.6 (1.8) HR (beat min 1 ) GEL 158 (19) 10 (8)** HES 164 (35) 134 (31)* MAP (mm Hg) GEL 105 (16) 5 (18)** HES 113 (31) 71 (16)** PAPm (mm Hg) GEL 1.7 (3.8) 10.7 (3.3) HES 13.9 (.3 ) 1.9 (.8) PAOP (mm Hg) GEL 3. (1.8) 3.9 (.0) HES 3.4 (1.3) 4.1 (1.4) RAP (mm Hg) GEL.1 (1.6).9 (.0) HES 1.7 (0.6).6 (1.8) CI (ml min 1 kg 1 ) GEL 111 (6) 111 (7) HES 134 (9) 19 (9) SI (ml kg 1 ) GEL 0.71 (0.16) 0.84 (0.35) HES 0.85 (0.4) 1.00 (0.30) SVR (dyn s cm 5 ) GEL 3498 (1155) 1579 (411)** HES 334 (955) 15 (805)** PVR (dyn s cm 5 ) GEL 411 (131) 340 (91) HES 41 (85) 413 (19) LVSWI (g m kg 1 ) GEL 0.98 (0.4) 0.61 (0.0)** HES 1.1 (0.35) 0.93 (0.40) RVSWI (g m kg 1 ) GEL 0.10 (0.06) 0.10 (0.04) HES 0.14 (0.03) 0.14 (0.04) Hb (g dl 1 ) GEL 14.5 (.6) 4.0 (0.7)** HES 14.7 (3.7) 3.8 (1.8)** S a O (%) GEL 94.4 (.6) 9.1 (6.7) HES 94.4 (1.5) 95.7 (1.6) S V O (%) GEL 7.8 (6.1) 35.6 (7.4)** HES 76.1 (4.1) 38.5 (7.)** DO (ml min 1 kg 1 ) GEL 1. (5.6) 5.9 (1.0)** HES 5.7 (7.) 6.5 (1.8)** VO (ml min 1 kg 1 ) GEL 4.5 (0.5) 3.7 (0.8)* HES 5.0 (1.1) 4. (1.) OER (%) GEL.3 (5.5) 63.8 (10.)** HES 0.1 (3.) 65.4 (1.0)** Lactate (mmol litre 1 ) GEL.4 (0.9) 5.1 (.5)** HES. (0.9).1 (1.0) P E CO (mm Hg) GEL 6.6 (4.4) 5.6 (7.7) HES 6.9 (3.5) 4.0 (5.3)

4 Critical haemoglobin concentration during anaesthesia 559 tered to reach the critical point were compared between groups using the Mann Whitney U test. Data obtained from the mesenteric DO /VO relationship were also compared between groups using the Mann Whitney U test. In both groups, O ERcrit and O ERm were compared using the Wilcoxon rank test. The systemic haemoglobin/do relationship was analysed using two-way analysis of variance for repeated measuremnts. In each group the DO value for a haemoglobin concentration below baseline was compared with the DO value obtained for baseline haemoglobin (i.e g dl 1 ) by a method of contrasts. All values are expressed as mean (SD). P 0.05 was considered significant. Results HAEMODYNAMIC AND RESPIRATORY EFFECTS OF HAEMODILUTION (TABLE 1) In both groups, progressive haemodilution was associated with a slight decrease in blood temperature (ns). HR and MAP decreased significantly. PAOP and RAP tended to increase (ns). CI did not change while SI tended to increase. SVR decreased, as did LVSWI. The decrease in arterial oxygen content resulted in a significant decrease in oxygen delivery. As CO did not increase during haemodilution, systemic oxygen delivery decreased in parallel to the decrease in haemoglobin (fig. 1). Despite a marked increase in oxygen extraction ratio, oxygen consumption also decreased, although this was significant only in the GEL group. The effects of haemodilution were similar in both groups, except for blood lactate, which increased significantly only in the GEL group. There was no group time interaction for all investigated variables. Table Data at the critical point in the GEL and HES groups GEL group (n 9) HES group. There was no significant difference in the critical haemoglobin concentration (table ). There was no significant difference between the two groups in amount of blood withdrawn (GEL 60.7 (19.5) ml kg 1 ; HES 74.6 (7.5) ml kg 1 ) or in the amount of colloid administered (GEL 68.3 (19.5) ml kg 1 ; HES 77.4 (9.4) ml kg 1 ) to reach the critical point. REGIONAL EFFECTS OF HAEMODILUTION (TABLE 3) HES group (n 9) DO crit (ml min 1 kg 1 ) 6.0 (1.) 4.0 (0.9) VO crit (ml min 1 kg 1 ) 6.3 (1.4) 4.3 (1.1) O ERcit (ml min 1 kg 1 ) 66.1 (8.4) 7.6 (1.3) DO crit (Lac) (ml min 1 kg 1 ) 67.7 (7.1) 3.6 (0.8) Hbcrit (g dl 1 ) 6.6 (1.7) 3.5 (1.5) Progressive haemodilution was associated with an increase in mesenteric blood flow but not in renal and femoral blood flows in both groups. When these regional blood flows were expressed as a fraction of the CO, mesenteric blood flow increased, while renal blood flow tended to decrease and femoral blood flow remained unchanged. EFFECTS OF HAEMODILUTION ON CRITICAL MESENTERIC DO (TABLE 4, FIG. 3) There was no significant difference in critical mesenteric blood flow, venous oxygen saturation, DO, VO or O ER between the two groups. In both groups, mesenteric O ERcrit was significantly lower than global O ERcrit (P 0.0). EFFECTS OF HAEMODILUTION ON CRITICAL DO AND CRITICAL HAEMOGLOBIN CONCENTRATION (TABLE ) Figure shows the DO /VO and haemoglobin/vo relationships in one representative dog from each group. There was no significant difference in DO crit, VO crit or O Ercrit between the two groups. DO crit obtained from blood lactate measurements was also similar in both groups, although blood lactate at the critical point was significantly higher in the GEL group than in the Figure 1 Haemoglobin (Hb)/DO relationship in the GEL and HES groups. *Significantly different from baseline DO (DO for a haemoglobin concentration of 15.0 g dl kg 1 ). Discussion We have examined the systemic and regional effects of progressive normovolaemic haemodilution in anaesthetized animals. Although not often used in clinical practice, ketamine was chosen as the anaesthetic agent for its cardiovascular properties and its effects on tissue oxygen extraction. 4 We found that the critical haemoglobin concentration in anaesthetized dogs undergoing ventilation with air was g dl 1. Interestingly, data obtained from analysis of either the haemoglobin/vo relationship or the DO /VO relationship were identical. The critical haemoglobin concentration we observed is in agreement with two other animal studies. In pentobarbital-anaesthetized dogs, 9 Cain estimated the critical haematocrit concentration to be approximately 10%. In pentobarbital-anaesthetized pigs undergoing progressive haemodilution with dextran 40, Räsänen 30 found a critical haemoglobin value of 3.9 g dl kg 1. Obviously, corresponding values are difficult to obtain in humans. Nevertheless, van Woerkens, Trouwborst and van Lanschot 31 studied a sedated Jehovah Witness patient who died from extreme haemodilution, and observed a critical haemoglobin concentration of 4 g dl kg 1. At the critical point, whole body oxygen extraction ratio was approximately 60%, a value which is very

5 560 British Journal of Anaesthesia Figure DO /VO (top) and haemoglobin (Hb)/ VO (bottom) relationships in one representative animal from the GEL group (left) and the HES group (right). similar to that observed during haemorrhagic shock in dogs using the same anaesthetic technique. 4 Critical oxygen delivery and consumption tended to be lower in the haemodilution than in the haemorrhagic protocol, but this could be attributed to the small decrease in body temperature which was unavoidable during massive administration of i.v. fluids. Progressive haemodilution was associated with redistribution of blood flow to the mesenteric area, but not to the kidneys or muscles. These data are in agreement with other animal studies demonstrating that the kidneys are the organs at risk during haemodilution Using radioactive microspheres, Crystal and Salem 15 3 demonstrated, in pentobarbitalanaesthetized dogs, that haemodilution to a haemoglobin value of 8 g dl kg 1 was not associated with an increase in renal blood flow, resulting in a decrease in renal oxygen delivery. Although mesenteric blood flow tended to increase during haemodilution, mesenteric oxygen extraction was significantly lower than systemic oxygen extraction at the critical point. This can be related to alteration of mesenteric oxygen extraction capabilities related to haemodilution. Kiel, Riedel and Shepherd 33 observed in pentobarbitalanaesthetized dogs a reduction in intestinal oxygen extraction during haemodilution which might be attributed to the vascular organization of the intestinal mucosa. In young pigs anaesthetized with ketamine and flunitrazepam, Nöldge and colleagues 34 also observed that haemodilution to a packed cell volume of 14% was associated with a decrease in mean surface PO of the liver and small intestine, indicating that the increases in blood flow and oxygen extraction were no longer able to compensate for the decrease in splanchnic oxygen delivery. At this level of haemodilution, small intestine oxygen extraction was only 33%, while liver oxygen extraction was 49%. We have also explored the possibility that the physiological modifications associated with normovolaemic haemodilution might be influenced by the type of fluid used for the exchange. In view of the shorter intravascular persistence of gelatin, its volume could have been greater than the volume of HES, but this was not the case, probably because of the short duration of the experiments. Under the conditions of our study, differences between the two solutions were small. These observations are in agreement with a recent study, 35 showing that the in vivo rheological effects of these two colloids are similar and less favourable than those of 3.5% dextran 40. The GEL group had a higher blood lactate concentration at the critical point, but this was likely to be Table 3 Regional blood flows at baseline and near the critical point in the GEL (n 9) and HES (n 9) groups. *P vs baseline Group Baseline Near critical point Mesenteric blood flow (ml min 1 ) GEL 07 (19) 306 (09) HES 19 (101) 77 (10) Mesenteric blood flow (% of CO) GEL 7.3 (4.3) 11.1 (4.5)* HES 7. (4.4) 10.8 (3.6)* Renal blood flow (ml min 1 ) GEL 197 (103) 1 (67) HES 180 (80) 163 (16) Renal blood flow (% of CO) GEL 7.5 (3.1) 5.1 (.9) HES 7. (3.3) 6.0 (3.3) Femoral blood flow (ml min 1 ) GEL 36 (6) 4 (34) HES 56 (30) 51 (1) Femoral blood flow (% of CO) GEL 1.3 (0.5) 1.5 (1.0) HES.0 (0.9).0 (0.8)

6 Critical haemoglobin concentration during anaesthesia 561 Table 4 Mesenteric DO /VO relationship in the GEL and HES groups GEL group (n=7) HES group (n=7) DO crit (ml min 1 kg -1 ) (0.395) (0.94) VO (ml min -1 kg 1 ) 0.34 (0.0) 0.81 (0.076) O ERcrit (%) 50.1 (1.1) 48.5 (13.4) Mesenteric blood flow (ml min 1 ) 40 (168) 95 (147) Mesenteric SvO (%) 45.9 (10.8) 56.0 (10.0) related to the presence of lactate in the gelatin solution. Interestingly, progressive haemodilution was not associated with an increase in CO in our anaesthetized animals. Similar results were reported in dogs and in humans during anaesthesia with halogenated anaesthetics, which have known cardiodepressant and sympatholytic effects. 38 Our observations were somewhat surprising as ketamine has sympathomimetic actions related to direct stimulation of central nervous system structures. 7 8 The absence of an increase in CO, despite the decrease in systemic vascular resistance and maintenance of constant cardiac filling pressures, could be related to the negative inotropic properties of ketamine at the dose used in our study, 7 39 as suggested by the decrease in left ventricular stroke work in the presence of a constant PAOP. Previous administration of thiopental at induction of anaesthesia might also have blunted the cardiovascular stimulant properties of ketamine, 40 although we used very small doses of thiopental and the actual procedure started at least h after administration of thiopental. Even more surprising was the significant decrease in HR during haemodilution in our animals. The reason for this decrease is unclear, as ketamine did not seem to increase parasympathetic tone 41 or depress baroreceptor reflex activity. 4 As CO did not increase during progressive haemodilution, oxygen delivery decreased in parallel with the reduction in arterial oxygen content. Therefore, we did not observe the relationship between haemoglobin and oxygen delivery as demonstrated by Messmer and colleagues, 1 showing a maximum oxygen delivery for haemoglobin concentration close to 10 g dl kg 1. Our observations are in agreement with those reported by Cain in pentobarbital-anaesthetized dogs. 9 Therefore, ketamine, in common with halogenated anaesthetics, may limit the ability of the organism to tolerate severe haemodilution. In summary, our data demonstrated that ketamine anaesthesia prevented the increase in CO, which is expected during normovolaemic haemodilution. Under these conditions, the only mechanism allowing maintenance of adequate tissue oxygen delivery is the oxygen extraction reserve. This mechanism allows maintenance of tissue oxygenation to a critical haemoglobin concentration of g dl kg 1. The physiological response to progressive haemodilution we observed was not influenced by the type of synthetic colloid used (gelatin or hydroxyethylstarch). At the regional level, despite an increase in splanchnic blood flow, oxygen availability to the gut may be threatened by weaker tissue extraction capabilities and to the kidneys by decreased blood flow as a result of redistribution of CO to other organs. Acknowledgement Supported by grants from the Erasme Foundation. Figure 3 Mesenteric DO /VO relationship in one representative dog from the GEL group (A) and from the HES group (B). References 1. Janvier G, Annat G. Are there limits to haemodilution? Annales Francaises d Anesthesie et de Réanimation 1995; 14: Welch GH, Meehan KR, Goodnough LT. Prudent strategies for elective red blood cell transfusion. Annals of International Medicine 199; 116: Spahn DR, Leone BJ, Reves JG, Pasch T. Cardiovascular and coronary physiology of acute isovolemic hemodilution: a review of nonoxygen-carrying and oxygen-carrying solutions. Anesthesia and Analgesia 1994; 78: Biboulet P, Capdevila X, Benetreau D, Aubas P, D Athis F, Du Cailar J. Haemodynamic effects of moderate normovolaemic haemodilution in conscious and anaesthetized patients. British Journal of Anaesthesia 1996; 76: Lacks H, Pilon RN, Klovekorn WP, Anderson W, Mac Callum JR, O Connor NE. Acute hemodilution: its effects on hemodynamics and oxygen transport in anesthetized man. Annals of Surgery 1974; 180: Vincent JL. Determination of oxygen delivery and consumption vs cardiac index and oxygen extraction ratio. Critical Care Clinics 1996; 1: Fontana JL, Welborn L, Mongan PD, Sturm P, Martin G, Bünger R. Oxygen consumption and cardiovascular function in children during profound intraoperative normovolemic hemodilution. Anesthesia and Analgesia 1995; 80: Spahn DR, Zollinger A, Schlumpf RB, Stöhr S, Seifert B, Schmid ER, Pasch T. Hemodilution in elderly patients without known cardiac disease. Anesthesia and Analgesia 1996; 8: Gisselsson L, Rosberg B, Ericsson M. Myocardial blood flow oxygen uptake and carbon dioxide release of the human heart during hemodilution. Acta Anaesthesiologica Scandinavica 198; 6: Kreimer U, Messmer K. Hemodilution in clinical surgery: State of the art World Journal of Surgery 1996; 0:

7 56 British Journal of Anaesthesia 11. Habler OP, Kleen MS, Podtschaske AH, Hutter JW, Tiede M, Kemming GI, Welte MV, Corso CO, Messmer K. The effect of acute normovolemic hemodilution (ANH) on myocardial contractility in anesthetized dogs. Anesthesia and Analgesia 1996; 83: Messmer K, Sunder-Plasmann L, Klövekorn WP, Holper K. Circulatory significance of hemodilution: rheological changes and limitations. Advances in Microcirculation 197; 4: Schou H, Perez de Sa V, Sigurdardottir M, Roscher R, Jonmarker C, Werner O. Circulatory effects of hypoxia, acute normovelemic hemodilution and their combination in anesthetized pigs. Anesthesiology 1996; 84: Chapler CK, Cain CM. The physiologic reserve in oxygen carrving capacity: studies in experimental hemodilution. Canadian Journal of Physiology and Pharmacology 1986; 64: Crystal GJ, Salem MR. Myocardial and systemic hemodynamics during isovolemic hemodilution alone or combined with nitroprusside-induced controlled hypotension. Anesthesia and Analgesia 1991; 7: van Woerkens ECSM, Trouwborst A, Duncker DJGM, Koning MMG, Boosma F, Verdouw PD. Catecholamines and regional hemodynamics during isovolemic hemodilution in anesthetized pigs. Journal of Applied Physiology 199; 7: Saddler JM, Horsey PJ. The new generation gelatins. Anaesthesia 1987; 4: Van der Linden P, Schmartz D. Pharmacology of gelatins. In: Baron JF, ed. Plasma Volume Expansion. Paris: Arnette Blackwell, 199; Ring J, Messmer K. Incidence and severity of anaphylactoid reactions to colloid volume substitute. Lancet 1977; 1: Mishler IV JM. Synthetic plasma volume expanders their pharmacology, safety and clinical efficacy. Clinics in Haematology 1984; 13: Warren BB, Durieux ME. Hydroxyethyl starch. Safe or not? Anesthesia and Analgesia 1997; 84: Strauss RG, Stansfield C, Henriksen RA, Villhauer PJ. Pentastarch may cause fewer effects on coagulation than hetastarch. Transfusion 1988; 8: White PF, Way WL, Trevor AJ. Ketamine its pharmacology and therapeutic uses. Anesthesiology 198; 56: Van der Linden P, Gilbart E, Engelman D, Schmartz D, Vincent J. Effects of anesthetic agents on systemic critical O delivery. Journal of Applied Physiology 1991; 71: Hill RW. Determination of oxygen consumption by use of the paramagnetic oxygen analyzer. Journal of Applied Physiology 197; 33: Samsel RW, Schumacker PT. Determination of the critical O delivery from experimental data: sensitivity to error. Journal of Applied Physiology 1988; 64: Traber DL, Wilson RD, Priano LL. Differentiation of the cardiovascular effects of CI-581. Anesthesia and Analgesia 1968; 47: Chodoff P. Evidence for a central adrenergic action of ketamine. Anesthesia and Analgesia 197; 51: Cain SM. Oxygen delivery and uptake in dogs during anemic and hypoxic hypoxia. Journal of Applied Physiology 1977; 4: Räsänen J. Supply-dependent oxygen consumption and mixed venous oxyhemoglobin saturation during isovolemic hemodilution in pigs. Chest 199; 101: van Woerkens ECSM, Trouwborst A, van Lanschot JJB. Profound hemodilution: what is the critical level of hemodilution at which oxygen delivery-dependent oxygen consumption starts in an anesthetized human? Anesthesia and Analgesia 199; 75: Crystal GJ, Rooney MW, Ramez Salem M. Regional hemodynamics and oxygen supply during isovolemic hemodilution alone or in combination with adenosine-induced controlled hypotension. Anesthesia and Analgesia 1988; 67: Kiel JW, Riedel GL, Shepherd AP. Effects of hemodilution on gastric and intestinal oxygenation. American Journal of Physiology 1989; 56: H Nöldge GFE, Priebe HJ, Bohle W, Buttler KJ, Geiger K. Effects of acute normovolemic hemodilution on splanchnic oxygenation and on hepatic histology and metabolism in anesthetized pigs. Anesthesiology 1991; 7: Audibert G, Donner M, Lefèvre JC, Stolz JF, Laxenaire MC. Rheologic effects of plasma substitutes used for preoperative hemodilution. Anesthesia and Analgesia 1994; 78: Van der Linden P, Wathieu M, Gilbart E, Engelman E, Wautrecht JC, Lenaers A, Vincent JL. Cardiovascular effects of moderate normovolaemic haemodilution during enflurane nitrous oxide anaesthesia in man. Acta Anaesthesiologica Scandinavica 1994; 38: Van der Linden P, Gilbart E, Paques P, Simon C, Vincent JL. Influence of hematocrit on tissue O extraction capabilities in anesthetized dogs during acute hemorrhage. American Journal of Physiology 1993; 64: H Jones RM. Clinical comparison of inhalation anesthetic agents. British Journal of Anaesthesia 1984; 56: 57S 69S. 39. Schwartz DA, Horwitz LD. Effects of ketamine on left ventricular performance. Journal of Pharmacology and Experimental Therapeutics 1975; 194: Dobson MB. Anaesthesia with ketamine and thiopentone for short surgical procedures. Anaesthesia 1978; 33: Traber DL, Wilson RD, Priano LL. A detailed study of the cardiopulmonary responses to ketamine and its blockade by atropine. Southern Medical Journal 1970; 63: Slogoff S, Allen GW. The role of baroreceptors in the cardiovascular response to ketamine. Anesthesia and Analgesia 1974; 53: