British Journal of Anaesthesia 1996; 76: 783–789 Oxygen consumption and delivery relationship in brain-dead organ donors O. LANGERON, P. COUTURE, J. MATEO, B. RIOU, J.-L. PANSARD AND P. CORIAT of hypovolaemia, left ventricular dysfunction, or Summary both. Moreover, an abnormal DO2/VO2 relationship The oxygen delivery (DO2) and consumption (VO2) could occur in brain-dead patients, as reported relationship in brain-dead organ donors is un- previously in critically ill patients [9–11]. Indeed, known. Therefore, in a prospective study, we several studies have supported the hypothesis of a determined the DO2/VO2 relationship in 21 con- decrease in oxidative metabolism in brain-dead secutive brain-dead patients. Patients were allo- patients [4, 12], that could be explained by a VO2 cated to one of two groups, according to plasma supply dependency. lactate concentration: normal (group NL, n : 11) The aim of this study was to measure systemic DO2 or high (92.5 mmol litre91) (group HL, n : 10). and VO2 in brain-dead patients before and after VO2 was measured independently, using indirect variation in DO2, to determine the DO2/VO2 re- calorimetry, under control conditions, during low lationship. To ensure an accurate and valid as- DO2 challenge with PEEP administration, and high sessment of this relationship, VO2 and DO2 were DO2 challenge with inflation of medical antishock measured independently. Moreover, plasma concen- trousers and volume expansion or blood trans- trations of lactate may indicate if metabolism is fusion, as required. Under control conditions, there predominantly aerobic or anaerobic, and has been were no significant differences between groups NL used to assess organ donors [3, 5]. Consequently, we and HL in haemodynamic or oxygenation variables, tested the hypothesis that the DO2/VO2 relationship both groups having a low VO2 (mean 114 could differ, according to plasma concentrations of (SD 21) ml min91 m92). In group HL there was a lactate. different DO2/VO2 relationship pattern, with a de- pendent VO2 only. The mean slope of the DO2/VO2 Patients and methods relationship was significantly higher in group HL than in group NL (0.12 (0.09) vs 0.04 (0.07), P : The study was approved by the Ethics Committee of 0.05). We conclude that brain death was associated Pitié-Salpêtrière Hospital and conducted according with a low VO2, and patients in group HL exhibited to French legislation concerning multiple organ DO2/VO2 dependency which was not observed in procurement. Brain death was certified by: a neuro- patients in group NL. (Br. J. Anaesth. 1996; 76: logical examination demonstrating the absence of 783–789) brainstem reflexes; an apnoea test performed with intratrachael continuous high flow (15 litre min91) of Key words oxygen and after 15 min of mechanical ventilation with an FI O2 of 100 %; absence of spontaneous Complications, brain death. Measurement techniques, calor- imetry. Oxygen, consumption. Oxygen, delivery. ventilation movement associated with an arterial PCO2 greater than 8 kPa; no electrical activity during a 20-min electroencephalographic recording; Transplantation of organs from brain-dead donors absence of hypothermia (35 C) and drugs known to has become a common therapy for patients suffering depress the central nervous system [1, 5]. from end-stage organ failure. Many physiological All patients had a stable haemodynamic state, disturbances are associated with brain death, and characterized by a mean arterial pressure greater management of the multiple organ donor is complex than 60 mm Hg without any change in requirements . Failure to maintain cardiovascular and body for fluids, inotropic support, or both, in the pre- homeostasis can induce sub-optimal organ perfusion ceding hour, and without hepatic or renal failure. and early graft dysfunction in the recipient . During the study, fluid therapy and inotropic However, in brain-dead patients, the optimal oxygen delivery (DO2) required to meet the metabolic demand and provide appropriate tissue oxygenation OLIVIER LANGERON*, MD, PIERRE COUTURE, MD, JOACHIM MATEO, is unknown. Oxygen consumption (VO2), calculated MD, BRUNO RIOU, MD, PHD, JEAN-LOUIS PANSARD, MD, PIERRE using the Fick method, has been suggested as being CORIAT, MD, Department of Anesthesiology and Critical Care, low in brain-dead patients [3–5]. This phenomenon Hôpital Pitié-Salpétrière, Paris VI University, Paris, France. Accepted for publication: January 16, 1996. could be related either to a decrease in oxygen *Address for correspondence: Département d’Anesthésie- requirements similar to that observed in anaes- Réanimation, Hôpital Pitié-Salpétrière, 47 Boulevard de thetized patients [6–8] or to a deceased DO2 because l’Hôpital, 75651 Paris Cedex 13, France. 784 British Journal of Anaesthesia support were guided by haemodynamic monitoring sensor, a carbon dioxide infrared analyser and a with a pulmonary artery catheter, urinary output and flowmeter. This system has been validated pre- laboratory estimation of plasma concentrations of viously for accuracy, sensitivity and reproducibility electrolytes . All patients were ventilated mechan- in the measurement of VO2 [17, 18]. The mean error ically with an FI O2 of 40 % to maintain PaO2 9 13 kPa, in the measurement of VO2 at baseline and after PaCO2 at 4.5–5.3 kPa and pH at 7.35–7.45. Oeso- increasing VO2 is 92 % (range 97 to 3 %) . phageal temperature was maintained greater than Calibration of the apparatus was performed with a 35.5 C using heated blankets and perfusion warm- mixture of 96 % oxygen and 4 % carbon dioxide ing. before each use. For each patient, measurements were performed minute-by-minute over a 20-min period in which MEASUREMENTS variation in minute-by-minute measurements of Haemodynamic measurements were obtained via an oxygen consumption was less than 5 %. In each indwelling radial artery catheter and a pulmonary patient, during each period of the study, VO2 was artery catheter (SP5507 S, Viggo-Spectramed, Mon- determined as the mean of 20 measurements over a tigny le Bretonneux, France) connected to a haemo- 20-min period. All patients were ventilated mech- dynamic monitor (HP 78354, Hewlett-Packard, anically with a constant FIO2 of 40 % throughout the Andover, MA, USA). The following haemodynamic study. Oxygen extraction was calculated by dividing variables were measured: heart rate (HR), mean VO2 by DO2: arterial pressure (MAP), mean pulmonary arterial ! − Oxygen extraction = V O 2 × D O 2 1. pressure (MPAP) and pulmonary capillary wedge pressure (PCWP). Cardiac output was measured in triplicate using the thermodilution method, and the EXPERIMENTAL DESIGN mean was calculated. Systemic vascular resistance The trachea was not suctioned and patients were not (SVR) and cardiac index (CI) were calculated turned within the 30 min before acquisition of according to standard formulae. Transoesophageal measurements and throughout the study. We per- echocardiography (HP Sonos 1500, Hewlett- formed a DO2 challenge to obtain changes in DO2 of Packard, Andover, MA, USA) was performed by a at least <20 % from control DO2. trained echocardiographer and left ventricular ejec- To decrease DO2, positive end-expiratory pressure tion fraction area (LVEFa) was calculated. (PEEP) was administered to decrease cardiac output. Plasma concentrations of lactate (normal values PEEP titration was performed to maintain MAP -2.5 mmol litre91) were measured in control con- greater than 40 mm Hg. The level of PEEP was ditions using an enzymatic method (Dimension increased up to 10 or 15 cm H2O (mean 12 apparatus, Dupont de Nemours, Wilmington, DL, (SD 4) cm H2O). After a 15-min stabilization period, USA). As an increase in plasma concentrations of VO2 was measured. lactate is thought to reflect a shift from oxidative to To increase DO2, inflation of medical antishock anaerobic metabolism with tissue hypoxia, brain- trousers (MAST, Jobst Ltd, OH, USA) was under- dead patients were divided into two groups based on taken to increase venous return and thus cardiac whether they had a normal (NL) or high (HL) output . The inflation pressure was approxi- plasma lactate concentration, above 2.5 mmol litre91, mately 40 cm H2O in the leg compartments and in control conditions. This value was not known 30 cm H2O in the abdominal compartment. These during the study, so that physicians performed the levels of inflation pressure induce venous compres- DO2 challenge blinded. sion without arterial compression, enabling an Arterial and mixed venous blood was obtained to increase in DO2 . MAST inflation was followed measure pH, partial pressures of oxygen (PO2) and by volume expansion of 500 ml of gelation solution carbon dioxide (PCO2), oxygen saturation (SO2) (BG (Plasmion, Laboratories Roger Bellon, France) or Electrolytes Apparatus, Instrumentation Labora- transfusion of packed red blood cells if packed cell tory, Milano, Italy) and haemoglobulin (Hb) con- volume was less than 30 %. This DO2 challenge was centration (Cobas Argo, Apparatus, Roche, Basle, monitored using a pulmonary artery catheter to Switzerland). DO2 was calculated using the following avoid PCWP values greater than 20 mm Hg. After a formulae: 15-min stabilization period, VO2 was measured. DO = CI ((1.34 × Hb × Sa O ) + (0.003 × Pa O )) × 10 2 2 2 At the end of each 20-min period, haemodynamic and metabolic variables were recorded, and arterial VO2 and DO2 were measured independently to and mixed venous blood was obtained for blood-gas obtain reliable results, avoiding mathematical coup- analysis. We determined VO2 in the same order: ling of shared variables [13–16]. As reported pre- control conditions, low DO2 and high DO2, with a viously [17–19], VO2 was measured using a portable recovery period between low and high DO2 charac- indirect calorimetry system (Deltatrac, Metabolic terized by a stable haemodynamic and metabolic Monitor, Datax Instrumentation Corp, Helsinki, state, as in control conditions. Finland). This metabolic monitor measures carbon dioxide and oxygen concentrations in inspired and STATISTICAL ANALYSIS expired gases per minute and calculates carbon dioxide production and oxygen consumption during Data are expressed as mean (SD or range). Com- that minute. These measurements were performed parison of control values between the two groups on the ventilator circuit with a paramagnetic oxygen (NL and HL) was performed using the Student’s t DO2/VO2 relationship in brain death 785 test and the Fisher’s exact method. Comparison of two groups in age, cause of brain death, mean time several means was performed using repeated- lapse between causal event and brain death, body measure analysis of variance and Newman–Keuls temperature, haemoglobin concentration, LVEFa test. The relationship between VO2 and DO2 was and the number of patients receiving dopamine fitted to a two-regression line model. The two best- throughout the study or dose of dopamine (table 1). fit regression lines were determined by the sum of Under control conditions, there were no significant least squares technique . The critical DO2 differences in haemodynamic variables between represented the threshold, below which VO2 started groups NL and HL, except for cardiac index (table to decrease with a further reduction in DO2. All P 2). There were no significant differences between values were two-tailed and P : 0.05 was considered groups for DO2 (group NL, 428 (131); range significant. Statistical analysis was performed on a 276–684 ml min91 m92 vs group HL, 450 (203); - computer using PCSM software (V3.1, Deltasoft, 190–733 ml min91 m92), S v O2 oxygen extraction, ar- Meylan, France). terial oxygen content and the arteriovenous differ- ence in oxygen content in control conditions (table 3). Moreover, a low VO2 was observed in the two groups Results (114 (21) ml min92 m92) and VO2 was not significantly We studied 21 brain-dead patients (15 men), mean different between groups in control conditions (table age 38 (range 16–53) yr. Eleven patients had a normal 3). (group NL) plasma lactate concentration During the DO2 challenge, we observed a sig- (1.9 (SD 0.4); range 1.25–2.5 mmol litre91), and 10 nificant increase in DO2 in both groups during MAST patients had an elevated (group HL) plasma lactate application (group NL, ;24 (17) %, P : 0.05; concentration (4.0 (0.9); 2.6–5.5 mmol litre91). group HL, ;48 (25) %, P : 0.05), and a significant There were no significant differences between the decrease in DO2 in both groups during PEEP Table 1 Patient characteristics (number of patients or mean (SD or range)). BD : Brain death, LVEFa : left ventricular ejection fraction area. No significant differences between groups Group NL Group HL (n : 11) (n : 10) Age (yr) 39 (21–53) 32 (16–51) Sex (M/F) 8/3 7/3 Causes of BD Head injury 6 6 Cerebral vascular disease 3 2 Cerebral anoxia related to cardiac arrest 2 2 Time lapse between causal event and BD (h) 41 (45) 35 (42) Dopamine No. of patients 8 10 Dose ( g kg91 min91) 4.9 (3.5) 5.0 (2.5) LVEFa (%) 60 (15) 58 (18) Haemoglobin (g dl91) 9.4 (2.3) 9.1 (2.5) PaO2 FIO2 (mm Hg) 341 (87) 360 (160) Body temperature (°C) 36.7 (0.9) 37.0 (0.6) Table 2 Haemodynamic variables in brain-dead patients with normal lactate (NL) and high lactate (HL) concentrations during oxygen delivery (DO2) challenge (mean (SD)). MAP : Mean arterial pressure; HR : heart rate; CI : cardiac index; PAP : mean pulmonary arterial pressure; PCWP : pulmonary capillary wedge pressure; SVR : systemic vascular resistance. *P : 0.05 compared with control condition Overall between- Group NL Group HL group (n : 11) (n : 10) comparison MAP Low DO2 58 (11)* 53 (10)* (mm Hg) Control 83 (8) 73 (12) ns High DO2 97 (16)* 102 (22)* HR Low DO2 111 (25) 114 (21) (beat min91) Control 107 (29) 110 (18) ns High DO2 106 (27) 114 (22) CI Low DO2 2.6 (0.7)* 2.5 (1.1)* (ml min91 m92) Control 3.8 (1.1) 3.9 (1.6) 0.05 High DO2 3.8 (1.0) 5.0 (1.9)* PAP Low DO2 22 (5)* 20 (5)* (mm Hg) Control 18 (7) 15 (5) ns High DO2 23 (10)* 23 (5)* PCWP Low DO2 15 (4)* 14 (4)* (mm Hg) Control 7 (5) 6 (3) ns High DO2 13 (7)* 13 (3)* SVR Low DO2 18.9 (8.2) 19.8 (12.1) (UI) Control 22.0 (7.3) 21.3 (10.7) ns High DO2 23.6 (6.2) 21.1 (11.8) 786 British Journal of Anaesthesia Table 3 Oxygenation variables in brain-dead patients with normal lactate (NL) and high lactate (HL) concentrations during oxygen delivery (DO2) challenge (mean (SD)). DO2 : Oxygen delivery; VO2 : oxygen consumption; S vO2 = mixed venous oxygen saturation; Ca O2 : arterial oxygen content; Ca O2 − Cv O2 : - arteriovenous oxygen content difference. *P : 0.05 compared with control condition Overall between- Group NL Group HL group (n : 11) (n : 10) comparison DO2 Low DO2 325 (108)* 277 (117)* (ml min91 m92) Control 428 (131) 450 (203) ns High DO2 525 (144)* 658 (293)* VO2 Low DO2 98 (17)* 100 (13)* (ml min91 m92) Control 109 (24) 119 (16) :0.001 High DO2 104 (24) 142 (24)* - S vO2 Low DO2 70 (12)* 67 (9)* (%) Control 79 (7) 77 (10) ns High DO2 83 (7) 79 (8) Oxygen extraction ratio Low DO2 33 (12)* 42 (18)* (%) Control 27 (8) 33 (20) ns High DO2 21 (6)* 26 (14) CaO2 Low DO2 12.3 (2.3) 12.0 (3.0) (ml 100 ml91) Control 12.4 (3.0) 12.0 (3.1) ns High DO2 13.3 (1.8) 12.7 (2.5) CaO2 − C VO2 Low DO2 3.5 (1.4)* 3.9 (1.1)* (ml 100 ml91) Control 2.5 (0.6) 2.7 (0.6) ns High DO2 2.0 (0.8) 2.3 (0.7) Figure 1 Individual relationship between oxygen consumption Figure 2 Individual relationship between mixed venous (VO2) and oxygen delivery (DO2) during DO2 challenge in - oxygen saturation ( S vO2 ) and oxygen delivery (DO2) during DO2 patients with normal plasma lactate concentrations (group NL, challenge in patients with normal plasma lactate concentrations n : 11) (A) and high plasma lactate concentrations (group HL, (group NL, n : 11) (A) and high plasma lactate concentrations n : 10 (B) under three conditions: low DO2 ( ), control DO2 (group HL, n : 10 (B) under three conditions: low DO2 ( ), ( ) and high DO2 (!). control DO2 ( ) and high DO2 (!). administration (group NL, –24 (10) %, P : 0.05; lationship in patients in group HL compared with group HL, –35 (12) %, P : 0.05). In group NL, patients in group NL (0.12 (0.09) vs 0.04 (0.07); P : VO2 did not increase significantly during high DO2 0.05). This DO2/VO2 relationship, either in NL or challenge (–4 (8) %) and was moderately decreased HL patients, impedes the use of a two-regression during low DO2 challenge (–9 (9) %, P : 0.05). In line model in each group and consequently the contrast, in group HL, VO2 increased significantly assessment of critical DO2 could not be achieved in during high DO2 challenge (;19 (17) %, P : 0.05) either group (fig. 1). and decreased significantly during low DO2 challenge As shown in table 3 and figure 2, we observed a (–16 (6) %, P : 0.05) (table 3). - different pattern in the S v O2 /DO2 relationship in There was a different pattern in the DO2/VO2 groups NL and HL, with a significant correlation relationship in groups NL and HL. VO2 increased - between S v O2 and DO2 in group NL (r : 0.64, P : with DO2 throughout DO2 challenge in patients in 0.05), and no significant correlation in group HL group HL, whereas patients in group NL exhibited - (r : 0.34). The mean slope of the S v O2 /DO2 relation- small variations in VO2 ( VO2) during control and ship (fig. 2) was not significantly different between high DO2 conditions. Nevertheless, it should be patients in groups NL and HL (0.07 (0.05) vs 0.05 pointed out that variations in DO2 ( DO2) were lower (0.05) % ml91 min m2). in group NL compared with group HL (48 (25) % vs 83 (26) %; P : 0.05). However, the magnitude of VO2/ DO2 was significantly higher in group HL than in group NL (0.13 (0.09) vs 0.05 (0.04); Discussion P : 0.05). In this study, we observed a low VO2 in brain-dead Individual analysis of the DO2/VO2 relationship is patients, both in group NL and group HL (table 3), depicted in figure 1. We observed a significant and a different pattern in the DO2/VO2 relationship increase in the mean slope of the DO2/VO2 re- between these two groups. Pathological VO2 de- DO2/VO2 relationship in brain death 787 pendency was noted in patients with elevated plasma pattern compared with patients in group NL (fig. 1). concentrations of lactate and a physiological In group HL, the increased plasma lactate con- DO2/VO2 relationship with a plateau was observed in centration suggested inadequate DO2 and that tissue patients with normal plasma lactate concentration. oxygen requirements were no longer satisfied, and Brain-dead patients had low VO2 values, as DO2/VO2 dependency was observed. Inadequate DO2 reported previously in normal humans at rest  usually implies either hypoperfusion or arterial (range of 110–150 ml min91 m92) or in critically ill desaturation, whereas impaired oxygen utilization surgical patients undergoing mechanical ventilation by the tissues reflects abnormal oxygen metabolism at rest (114 (18) ml min91 m92) or at a low level of or dysoxia . However, the lack of significant activity (131 (21) ml min91 m92) . Moreover, in differences in haemodynamic and echocardiographic our study, VO2 values were close to those found in data, and blood-gas analysis between the two anaesthetized patients, where VO2 reached a plateau groups suggest dysoxia in group HL, with a weak at 109 (16) ml min91 m92 or at 105 (13) ml min91 m92, oxygen extraction and utilization by the tissues in before and after cardiopulmonary bypass, respect- group HL, as suggested by the non-significant ively [6, 7]. During general anaesthesia, oxygen difference in arteriovenous oxygen content difference consumption decreased by 25–30 % [8, 23], cor- between groups HL and NL. Furthermore, in group responding to the values found in our study during HL, the inability to obtain a critical DO2 despite a brain death. Mechanical ventilation decreases oxy- wide DO2 challenge with VO2 dependency would gen consumption and minimizes oxygen uptake. suggest a pathological DO2/VO2 dependency . In Consequently, similar VO2 values were observed group NL, we probably did not decrease DO2 as during brain death and general anaesthesia . much as necessary to obtain critical DO2 from which The DO2/VO2 relationship has been studied under VO2 begins to decrease, with a further reduction in many circumstances [24–27], but is often a source of DO2. Thus our results suggest that critical DO2 in controversy, mainly because an artefactual corre- brain-dead patients with normal plasma lactate lation between DO2 and VO2 can occur by sharing concentration was less than 330 ml min91 m92 which common data for cardiac output and arterial oxygen is the critical DO2 reported in patients during general content leading to mathematical coupling of these anaesthesia [6, 7]. Conversely, critical DO2 in brain- variables [13, 16, 28]. In our study, VO2 was mea- dead patients with elevated plasma concentrations of sured independently in order to avoid this pitfall. We lactate was not reached despite high DO2 levels, and used the gas exchange method to measure VO2 with remains unknown at present. a metabolic monitor which had been validated This pathological DO2/VO2 dependency has been previously [17, 18] and exhibits a small error in VO2 reported in sepsis , adult respiratory distress measurement when FIO2 is less than the critical level syndrome (ARDS) , or both . However, other of 75 % . In our study, FIO2 was maintained recent studies in sepsis [13, 19, 20] and ARDS constant at 40 %. The second criterion for a reliable [14, 32] did not confirm this DO2/VO2 dependency, DO2/VO2 relationship is to minimize spontaneous probably because VO2 was measured and not calcu- and provoked variations in metabolic demand with lated. Thus, this is the first time that a DO2/VO2 no muscle activity to keep VO2 constant. These dependency has been demonstrated using indepen- conditions are achieved in brain death. Moreover, dent measures of DO2 and VO2. the DO2/VO2 relationship may be influenced by body Our study indicated the occurrence of dysoxia in temperature . In our study we did not observe some brain-dead patients, but did not enable us to any significant difference in temperatures between determine precisely the mechanism(s) of these groups NL and HL which could explain the different abnormalities. However, several hypotheses might pattern of the DO2/VO2 relationship. Furthermore, be suggested: (1) the event causing brain death or we performed a DO2 challenge with PEEP adminis- brain death itself is able to release mediators or tration, fluid loading and MAST application, some- induce neurohumoral dysfunction altering the micro- times associated with blood transfusion. As reported circulation, with oxygen diffusion impairment or previously, VO2 is not affected by any of these microcirculatory mismatch, such as during sepsis, manoeuvres [16, 30]. Whereas the effects of catecho- causing dysoxia; (2) a switch from aerobic to lamines on cellular oxidative metabolism increase anaerobic metabolism caused by haemodynamic VO2 , dopamine is not thought to increase VO2 at deterioration occurring during brain death may the low doses used in our study [16, 31]. Moreover, induce an oxygen debt in some patients and this debt we did not observe any significant difference in may be paid off with increasing DO2. However, our dopamine doses between the two groups. Finally, to study was conducted under stable haemodynamic ascertain that a reliable DO2/VO2 relationship could conditions, several hours after resuscitation, and be obtained it is necessary to have at least two data therefore this hypothesis is unlikely; (3) a reduction pairs for each patient over a wide range of DO2 in circulating free triiodothyronine, as noted pre- values. We collected three data pairs for each patient viously in brain-dead patients [3, 12], may result in over a DO2 variation of at least <20 % of control decreased oxidative metabolism. Indeed, it has been values, which was our minimal end-point in the DO2 shown recently that high plasma lactate concen- challenge (table 3). As reported previously, these trations are associated with mitochondrial impair- different factors are essential to describe a reliable ment in oxygen metabolism in brain-dead patients DO2/VO2 relationship . . After DO2 challenge, we observed that patients in Several limitations should be discussed to assess group HL had a different DO2/VO2 relationship the clinical relevance of our study. First, systemic 788 British Journal of Anaesthesia DO2 does not assess the effectiveness of regional DO2 MR, eds. Oxygen Transport in the Critically Ill. Chicago: in different organs and vascular beds . Local and Year Book Medical Publishers, 1987; 16–21. 9. Bihari D, Smithies M, Gimson A, Tinker J. The effects of regional blood flow redistribution are important vasodilatation with prostacyclin on oxygen delivery and up- determinants of the adequacy of both regional and take in critical ill patients. New England Journal of Medicine systemic oxygen uptake, as reported previously . 1987; 317: 397–403. Brain death is associated with loss of peripheral 10. Gilbert EM, Haupt MT, Mandanas RY, Huaringa AJ, vasomotor tone and an inability to redistribute blood Carlson RW. The effect of fluid loading, blood transfusion, and catecholamine infusion on oxygen delivery and con- flow appropriately . Consequently, brain death sumption in patients with sepsis. American Review of may also modify the local or regional DO2/VO2 Respiratory Disease 1986; 134: 873–878. relationship, leading to potential organ impairment, 11. Danek SJ, Lynch JP, Weg JG, Dantzker DR. The de- despite appropriate systemic DO2. Second, elevated pendence of oxygen uptake on oxygen delivery in the adult respiratory distress syndrome. American Review of Respir- plasma lactate concentration might not always reflect atory Disease 1980; 122: 387–395. tissue hypoxia, as lactate concentration results from 12. Novitzky D, Cooper DKC, Morell D, Isaacs S. Change from the balance between lactate production and hepatic aerobic to anaerobic metabolism after brain death, and and renal clearance, the washout phenomenon after reversal following triiodothyronine therapy. Transplantation successful resuscitation, and from the activity of 1988; 45: 32–36. 13. Vermeij CG, Feenstra BWA, Bruining HA. Oxygen delivery pyruvate dehydrogenase which may be inhibited as and oxygen uptake in postoperative and septic patients. Chest in sepsis. Nevertheless, in our study, none of the 1990; 98: 415–420. patients with elevated plasma lactate concentrations 14. Ronco JJ, Phang PT, Walley KR, Wiggs B, Fenwick JC, had hepatic or renal failure. Lastly, some authors Russell JA. Oxygen consumption is independent of changes  have supported the theory of obtaining supra- in oxygen delivery in severe adult respiratory distress syndrome. American Review of Respiratory Disease 1991; 143: normal values of DO2 to prevent dysoxia and organ 1267–1273. system dysfunction in order to decrease mortality in 15. Pinsky MR. Beyond global oxygen supply–demand relations: critically ill patients. But a recent study failed to in search of measures of dysoxia. Intensive Care Medicine demonstrate any survival benefit from obtaining 1994; 20: 1–3. supranormal DO2 values . In the management of 16. Hanique G, Dugernier T, Laterre PF, Dougnac A, Roeseler J, Reynaert MS. Significance of pathologic oxygen supply organ donors, increasing DO2 in patients in group dependency in critically ill patients: comparison between HL seems reasonable, but increasing DO2 to supra- measured and calculated methods. Intensive Care Medicine normal values has yet to be demonstrated. However, 1994; 20: 12–18. our results do not provide evidence that increasing 17. Ronco JJ, Phang PT. Validation of an indirect calorimeter to measure oxygen consumption in critically ill patients. Journal DO2 may actually decrease plasma concentrations of of Critical Care 1991; 6: 36–41. lactate and that transplantation outcome could be 18. Takala J, Keinanen O, Vaisanen P, Kari A. Measurements of modified by decreasing plasma lactate concen- gas exchange in intensive care: laboratory and clinical trations. Further studies are required to confirm this validation of a new device. Critical Care Medicine 1989; 17: hypothesis. 1041–1047. 19. Ronco JJ, Fenwick JC, Tweeddale MG, Wiggs BR, Phang PT, Cooper DJ, Cunningham KF, Russell JA, Walley KR. Acknowledgements Identification of the critical oxygen delivery for anaerobic metabolism in critically ill septic and nonseptic humans. We thank G. Annat, Professor of Physiology, Hôpital E. Herriot, Journal of the American Medical Association 1993; 270: Lyons, France. 1724–1730. 20. Mira JP, Fabre JE, Baigorri F, Coste J, Annat G, Artigas A, Nitenberg G, Dhainaut JFA. Lack of oxygen supply de- References pendency in patients with severe sepsis. A study of oxygen 1. Robertson KM, Cook DR. Perioperative management of the delivery increased by military antishock trousers and dobu- multiorgan donor. Anesthesia and Analgesia 1990; 70: tamine. Chest 1994; 106: 1524–1531. 546–556. 21. Samsel RW, Schumacker PT. Determination of the critical 2. Grundman R, Kammerer B, Franke E, Pichlmaier H. Effect O2 delivery from experimental data: sensitivity to error. of hypotension on the results of kidney storage and the use of Journal of Applied Physiology 1988; 64: 2074–2082. dopamine under these conditions. Transplantation 1981; 32: 22. Weissman C, Kemper M. The oxygen uptake–oxygen 184–187. delivery relationship during ICU interventions. Chest 1991; 3. Sztark F, Thicoïpe M, Masson F, Lassié P, Favarel- 99: 430–435. Garrigues JF, Petit-Jean ME. Metabolic status of brain-dead 23. Viale JP, Annat G, Lehot JJ, Quard S, Quintin L, Parlow J, patients managed for organ procurement. Transplantation Durand PG, Zabot JM, Villard J, Estanove S. Relationship Proceedings 1993; 25: 3171–3172. between oxygen uptake and mixed venous oxygen saturation 4. Depret J, Teboul JL, Bensadoun H, Benoit G, Richard C. in the immediate postoperative period. Anesthesiology 1994; Mise en évidence d’une hypoxie tissulaire par trouble de 80: 278–283. l’extraction périphérique de l’oxygène des sujets en état de 24. Schumaker PT, Cain SM. The concept of a critical oxygen mort cérébrale. La Presse Médicale 1991; 20: 1989–1990. delivery. Intensive Care Medicine 1987; 13: 223–229. 5. Riou B, Kalfon P, Arock M, Goarin JP, Saada M., Viars P. 25. Dantzker DR, Foresman B, Gutierrez G. Oxygen supply and Cardiovascular consequences of severe hypophosphataemia utilization relationships. American Review of Respiratory in brain-dead patients. British Journal of Anaesthesia 1995; Disease 1991; 143: 675–679. 74: 424–429. 26. Pinsky MR, Schlichtig R. Regional oxygen delivery in oxygen 6. Shibutani K, Komatsu T, Kubal K, Sanchala V, Kumar V, supply–dependent states. Intensive Care Medicine 1990; 16: Bizzari D. Critical level of oxygen delivery in anesthetized S169–S171. man. Critical Care Medicine 1983; 11: 640–643. 27. Vincent JL. The relationship between oxygen demand, 7. Komatsu T, Shibutani K, Kazufumi O, Kumar K, Kubal K, oxygen uptake, and oxygen supply. Intensive Care Medicine Sanchala V, Lees DE. Critical level of oxygen delivery after 1990; 16: S145–S148. cardiopulmonary bypass. Critical Care Medicine 1987; 15: 28. Archie JP. Mathematic coupling of data : a common source of 194–197. error. Annals of Surgery 1990; 193: 296–303. 8. Buran MJ. Oxygen consumption. In: Snyder JV, Pinsky 29. Schumaker PT, Rowland J, Saltz S, Nelson DP, Wood LDH. DO2/VO2 relationship in brain death 789 Effects of hyperthermia and hypothermia on oxygen ex- Garrigues JF, Petit-Jean ME. Modifications du métabolisme traction by tissues during hypovolemia. Journal of Applied oxydatif au cours de la mort cérébrale. Annales Françaises Physiology 1987; 63: 1246–1252. d’Anesthésie et de Réanimation 1994; 13: R266. 30. Pepe PE, Culver BH. Independently measured oxygen 34. Schlichtig R, Kramer DJ, Pinsky MR. Flow redistribution consumption during reduction of oxygen delivery by positive during progressive hemorrhage is a determinant of critical O2 end-expiratory pressure. American Review of Respiratory delivery. Journal of Applied Physiology 1991; 70: 169–178. Disease 1985; 132: 788–792. 35. Shoemaker WC, Appel PL, Kram HR, Waxman K, Lee TS. 31. Chiolero R, Flatt JP, Revelly JP, Jequier E. Effects of Prospective trial of supranormal values of survivors as catecholamines on oxygen consumption and oxygen delivery therapeutic goals in high risk surgical patients. Chest 1988; in critically ill patients. Chest 1991; 100: 1676–1684. 94: 1176–1186. 32. Annat G, Viale JP, Percival C, Froment M, Motin J. Oxygen 36. Hayes MA, Timmins AC, Yau EHS, Palazzo M, Hinds CJ, delivery and uptake in adult respiratory distress syndrome. Watson D. Elevation of systemic oxygen delivery in the American Review of Respiratory Disease 1986; 133: 999–1001. treatment of critically ill patients. New England Journal of 33. Sztark F, Thicoïpe M, Masson F, Lassié P, Favarel- Medicine 1994; 330: 1717–1722.