Critical Oxygen Delivery in Conscious Humans Is Less Than 7.3 ml O²· kg⁻¹· min⁻¹

With approval of the Institutional Review Board and with informed consent, we studied eight (five women, three men) healthy, paid volunteers. No volunteer smoked cigarettes or took prescription medications. All were without history of cardiovascular, pulmonary, or hepatic disease and had normal physical examination results. The data obtained at “baseline” and at a hemoglobin concentration of 5 g/dl for these eight volunteers were included as part of a larger group (n = 32) in a previous report. 7 

The methods for producing acute isovolemic hemodilution have been previously described. 7Briefly, peripheral venous and radial artery cannulae were inserted into each subject using 1% lidocaine local anesthesia. A flow-directed pulmonary artery cannula (Baxter Healthcare, Glendale, CA) was inserted via  the right internal jugular vein. Propofol (50–150 μg · kg⁻¹· min⁻¹intravenous) was infused briefly to provide sedation during placement of the pulmonary artery cannula. After all cannulae were inserted and the propofol infusion was discontinued, the subjects rested for 30 min before measurement of any variables. Cardiac output was measured by thermodilution (duplicate, or triplicate if duplicates differed by > 10%; A/S3 Datex Medical Instrumentation, Tewksbury, MA). As blood was removed, isovolemia was maintained, as judged by constant central venous and pulmonary capillary wedge pressures, by infusion of 5% human serum albumin (Baxter Healthcare) and the subject’s platelet-rich plasma as it became available after separation from the erythrocytes of the removed blood. At the times of cardiovascular measurements, arterial and mixed venous blood was sampled for measurement of p  H, oxygen content, and oxyhemoglobin saturation (OSM3 Hemoximeter; Radiometer, Copenhagen, Denmark) and arterial plasma lactate concentration (YSI No 0.2700; Yellow Springs Instrument Co., Yellow Springs, OH). Cardiac index, stroke volume index, systemic vascular resistance index, DO², and oxygen consumption (VO²) were calculated using standard formulae. The subject’s pulmonary artery temperature was maintained at 37°C by body surface warming with heated air (Bair Hugger model 1200; Augustine Medical Inc., Eden Prairie, MN) and by warming of the infused fluids.

Measurements were made 30 min after insertion of invasive cannulae and before removal of blood (baseline); after isovolemic hemodilution to a hemoglobin concentration of 5 g/dl; and again after 15 min steady-state reduction of the heart rate (HR) to approximately 85% of that at the end of hemodilution, achieved by an intravenous infusion of esmolol (50 μg · kg⁻¹· min⁻¹to an a priori  maximum of 150 μg · kg⁻¹· min⁻¹) without infusion of additional fluid. All erythrocytes were transfused after the conclusion of the experiment (no measurements made) and a physician examined all subjects the following morning before discharge.

Electrocardiography (ECG; five-lead) and a three-channel Holter ECG was monitored in all subjects (Del Mar model 459; Del Mar Avionics, Irvine, CA). The Holter monitor recorded continuously from 1 h before start of the study until completion of the study. The frequency response of the Holter recorder met the American Heart Association specification for ST changes, the cutoff limit being 0.05 Hz for low frequency and 100 Hz for high frequency. For Holter monitoring, three bipolar leads—CC5, modified CM5, and ML—were used. Each ECG recording on Holter tapes was scanned visually using an ECG analysis system (Del Mar model 750). All normal QRS complexes were identified, and all abnormal QRS complexes (e.g. , ventricular ectopic beats and conduction abnormalities) were excluded from ST-segment analysis. Continuous ST-segment trends were generated for the entire tape. All possible ischemic episodes were reviewed and verified by an investigator who was blinded to patient identity, hemoglobin concentration, and administration of esmolol. Prospective criteria for an ischemic episode were defined as a reversible ST-segment shift from baseline of 0.1 mV or greater depression at J+ 60 ms or 0.2 mV or greater elevation at the J point lasting for at least 1 min. The time after the J point chosen to measure ST-segment depression was adjusted to exclude T wave during tachycardia.

All data are expressed as the mean ± SD. Data obtained during esmolol infusion were compared with data obtained at the end of hemodilution (5 g/dl) and at baseline by analysis of variance with repeated measures and a post hoc  Newman-Keuls test. Statistical significance was accepted at P < 0.05.

Volunteers were 21.9 ± 2.2 yr old (mean ± SD), weighed 68 ± 13 kg, and had an estimated body surface area of 1.81 ± 0.23 m². The duration of the experiment was 190 ± 39 min.

Isovolemic hemodilution reduced hemoglobin concentration from 12.5 ± 0.8 to 4.8 ± 0.2 g/dl (table 1;P < 0.001). Heart rate, stroke volume index, and cardiac index increased (cardiac index: from 3.1 ± 0.6 l · min⁻¹· m⁻²to 5.4 ± 0.9 l · min⁻¹· m⁻²;P < 0.001), but not sufficiently to prevent DO²from decreasing (from 14.0 ± 2.9 ml O²· kg⁻¹· min⁻¹to 9.9 ± 2.0 ml O²· kg⁻¹· min⁻¹;P < 0.001).

Infusion of esmolol after hemodilution, without additional hemodilution or infusion of fluid, did not change further the hemoglobin concentration (4.7 ± 0.2 g/dl;P > 0.05), but decreased cardiac index by 24 ± 6% to 4.1 ± 0.5 l · min⁻¹· m⁻²(P < 0.001), as a result of an 11 ± 5% decrease in HR (P < 0.01) and a 15 ± 6% decrease in stroke volume index (P < 0.001). Thus, esmolol reduced DO²to 7.3 ± 1.4 ml O²· kg⁻¹· min⁻¹(P < 0.001), which is equivalent to 274 ± 51 ml O²· min⁻¹· m⁻².

Acute isovolemic hemodilution increased VO²from the baseline value of 3.0 ± 0.5 ml O²· kg⁻¹· min⁻¹to 3.4 ± 0.6 ml O²· kg⁻¹· min⁻¹;P < 0.05). Further reduction of DO²by esmolol infusion did not change VO²(3.2 ± 0.6 ml O²· kg⁻¹· min⁻¹;P > 0.05;fig. 1). The relation of DO²and VO²for each subject is depicted in figure 2, and the change in VO²from baseline as a function of DO²is shown in figure 3. Only two of eight volunteers had values of VO²that were less than their individual baseline values. For one of these two volunteers, VO²subsequently increased at a lower DO².

One sample for lactate determination for one volunteer was lost; consequently, data for plasma lactate concentrations are reported for seven volunteers. Plasma lactate concentrations after hemodilution (0.62 ± 0.16 mM) and during esmolol infusion (0.66 ± 0.14 mM) were minimally, but statistically significantly, greater than those at baseline (0.53 ± 0.13 mM;P < 0.05), but did not differ from each other (P > 0.05). No plasma lactate concentration exceeded the upper limit of normal (fig. 4).

Arterial and mixed venous p  H and base-excess did not change from baseline (P > 0.05;table 1). Mixed venous oxygen saturation (Sv^O2) decreased from 77.8 ± 4.6% to 67.3 ± 8.0% with hemodilution (P < 0.001) and subsequently to 56.8 ± 7.6% (P < 0.001) as DO²decreased further with esmolol infusion. The oxygen extraction ratio (VO²/DO²) increased from 0.22 ± 0.04 to 0.35 ± 0.08 with hemodilution (P < 0.001) and to 0.44 ± 0.08 (P < 0.001) with infusion of esmolol.

No ST changes were observed in the real-time monitored ECG. The Holter tapes of all subjects were able to be analyzed. A single episode that met the criteria for significant ST changes occurred in a 25 yr old woman. A 0.11-mV ST depression occurred during the final stage of hemodilution, during the time that the hemoglobin was reduced from 5.3 to 4.6 g/dl. DO²was 10.9 ml O²· kg⁻¹· min⁻¹and VO²was 3.9 ml O²· kg⁻¹· min⁻¹, which was not less than the baseline value of 3.6 ml O²· kg⁻¹· min⁻¹. These ST changes resolved with initiation of the esmolol infusion, reduction of HR from 110 to 89 beats/min, and decrease of DO²from 10.9 to 8.1 ml O²· kg⁻¹· min⁻¹. No subject reported any cardiac symptoms. One subject complained of transient light-headedness and fatigue during the esmolol infusion; a second subject felt a “sense of dread.” No other symptoms related to decreased DO²or tissue hypoxia were reported, and all symptoms resolved promptly after discontinuation of the administration of esmolol and infusion of the subjects’ erythrocytes.

We reduced DO²from 14 to 10 ml O²· kg⁻¹· min⁻¹, with acute reduction of hemoglobin to 4.7 g/dl, and further to 7.3 ml O²· kg⁻¹· min⁻¹by infusion of a β-adrenergic antagonist, esmolol, in conscious, healthy, resting adults. The DO²at a hemoglobin concentration of 5 g/dl is similar to the value we reported from a larger group at this level of acute isovolemic anemia. 7The addition of an infusion of a β-adrenergic antagonist reduced DO²to a value substantially less than that achieved previously. Despite this nearly 50% reduction of DO²from baseline, we were unable to demonstrate inadequate systemic oxygenation as assessed by our two primary measures: VO²did not decrease, and the tiny increases in plasma lactate concentrations at a hemoglobin concentration of 5 g/dl with and without esmolol infusion are not physiologically important. All plasma lactate values were within normal limits (upper limit, 2 mM): the highest concentration was 0.87 mM, and the mean change was less than 0.2 mM.

Only one woman had a single transient ST-segment change during the study period. It was not symptomatic and resolved despite a further reduction in DO². This change may have been secondary to myocardial ischemia or may have been an elevated HR-induced benign ECG change. The resolution of the ST depression during administration of esmolol may have been a result of a decrease in the HR from 110 to 89 beats/min, which should have reduced myocardial VO²by approximately 22%. 8In dogs, the critical myocardial DO²and the systemic critical DO²are reached at the same hemoglobin concentration. 9The absence of evidence for inadequate global DO²and the low specificity of ECG changes in the absence of other evidence of cardiac disease, 10,11especially in young women, 12,13are suggestive that this single ECG change was more likely benign and HR-induced rather than representative of myocardial ischemia.

Plasma lactate concentration is an established marker of inadequate systemic DO². 1,2In experiments of decreased DO², plasma lactate concentration increases simultaneously with decreases of VO², a primary marker of inadequate DO². The statistically significant, but physiologically unimportant increase in plasma lactate at the lowest levels of DO²that we achieved could indicate that DO²was beginning to approach the level at which plasma lactate concentration increases substantially with further decreases of DO². If that were true, VO²should have begun to decrease at that point to values below those at baseline. In a group of 32 healthy adults in whom DO²was decreased to 10.7 ± 2.0 ml O²· kg⁻¹· min⁻¹, plasma lactate concentration did not increase. 7In the current study, plasma lactate concentration increased minimally when DO²was reduced to 9.9 ± 2.0 ml O²· kg⁻¹· min⁻¹but did not increase further when DO²was decreased further to 7.3 ± 1.4 ml O²· kg⁻¹· min⁻¹. This suggests that the volunteers did not reach their critical DO². Alternatively, it is possible that these tiny increases reflect decreased hepatic clearance of lactate. However, hepatic clearance of lactate does not appear to decrease until systemic oxygenation is inadequate. In anesthetized dogs, hepatic blood flow increases, but not as much as does cardiac output, with acute anemia to a hematocrit of 17%, and the effect is not altered by β-adrenergic blockade. 14The hepatic extraction ratio of bromsulphalein decreases, but clearance increases. 14Hepatic clearance of lactate does not change in pigs anesthetized with ketamine and flunitrazepam and then made acutely anemic to hematocrit levels of 15%, despite a decrease in hepatic surface partial pressure of oxygen (P^O2). 15,16Hepatic uptake of lactate in those pigs did not decrease until systemic DO²was decreased, by the addition of isoflurane, from approximately 7–9 ml O²· kg⁻¹· min⁻¹to approximately 4 or 5 ml O²· kg⁻¹· min⁻¹, values below the critical DO². 15Furthermore, modeling of data obtained in humans suggests that, in the absence of increased lactate production, even large decreases of hepatic uptake of lactate will have only small effects on plasma lactate concentrations. 17The increase in base-excess with acute reduction of DO²to 9.9 ± 2.0 ml O²· kg⁻¹· min⁻¹was probably a result of hepatic metabolism of the citrate present in the transfused autologous plasma. This adds further support to our thought that hepatic clearance and metabolism continued during this period. Alternatively, it is possible that our method of measuring VO²was insufficiently sensitive to detect such a small change in VO².

The DO²we measured during acute anemia plus infusion of a β-adrenergic antagonist, 7.3 ml O²· kg⁻¹· min⁻¹(273 ml O²· min⁻¹· m⁻²), is lower than any value reported in healthy, conscious humans. We previously demonstrated that a DO²of 10.7 ml O²· kg⁻¹· min⁻¹in healthy conscious adults did not decrease VO²or increase plasma lactate concentration. 7Patients with substantial coronary artery disease, anesthetized for coronary artery surgery do not have anaerobic myocardial metabolism with hemoglobin concentrations as low as approximately 6 g/dl. 18A critical DO²reported by the authors to be 4.9 ml O²· kg⁻¹· min⁻¹, which we calculated to be 5.4 ml O²· kg⁻¹· min⁻¹, was found in an 84-yr-old man, who was anesthetized and had pharmacologically induced neuromuscular blockade and mechanically ventilated lungs. 6General anesthesia decreased his VO²by approximately 25%. Anesthesia, neuromuscular blockade, and mechanical ventilation of the lungs would have decreased the VO²and, thereby, the critical DO². In addition, modern inhaled halogenated anesthetics decrease myocardial contractility and myocardial VO², 19–21and function and VO²of other organs, such as the brain. 22Furthermore, the critical DO²in dogs is influenced by the type of anesthetic. 23Therefore, critical DO²determined during anesthesia cannot be applied to the awake condition.

Similarly, although the critical DO²is known for anesthetized, mechanically ventilated dogs (9 or 10 ml O²· kg⁻¹· min⁻¹), 1,2rats (23 ml O²· kg⁻¹· min⁻¹), 3pigs (8–12 ml O²· kg⁻¹· min⁻¹), 4,5and baboons (3–6 ml O²· kg⁻¹· min⁻¹), 24these values cannot be applied to conscious humans. The reduction of VO²by anesthetics and other drugs used in those experiments would have reduced the critical DO². In conscious, restrained, acutely instrumented baboons, a decrease in hematocrit to 15% and in DO²to 11 ml O²· kg⁻¹· min⁻¹did not decrease VO². 25The variation of critical DO²among the few species in which it has been determined makes application of the data to humans unreliable.

Our primary measures of whole-body VO²and blood lactate concentration are largely measures of adequacy of systemic oxygenation. Although VO²decreases in individual tissues and organs when DO²is inadequate, it is unlikely that we could have detected any decreased VO²in relatively small regions because of the error inherent in the measurement of VO². In addition, the use of cardiac output for the estimation of VO²has been criticized because of the potential problem of “mathematical coupling.”26However, if the data contain a range for DO²that is relatively large compared with the measurement error, as occurred with our data (by a factor of 5 to 10), the effect of mathematical coupling is small. 27In addition, the major influence of mathematical coupling is to erroneously indicate supply dependency of VO²when it is not truly present 27,28; however, we did not find supply dependency of VO²in subjects in the current study.

Volunteers in the current study were severely anemic for approximately 1 h. It is possible that a longer period would have resulted in inadequate systemic oxygenation. Because, in addition to the oxygen in the lungs and erythrocytes, the body contains little stores of oxygen, any later development of inadequate oxygenation would have to result from a degradation of compensatory mechanisms. Analysis of human kinetic data for lactate production and clearance indicates that increased production resulting from tissue hypoxia should be detected in blood within the period of time in which the volunteers had severely decreased DO². 17 

We observed the volunteers at rest only, and therefore cannot speculate as to the possible critical DO²during mild-to-moderate exercise. Within a relatively narrow range, hemoglobin concentration (9.7 ± 0.9 to 10.7 ± 0.9 g/dl) does not affect ability to function after surgical repair of femur fractures, 29nor does a somewhat larger range (approximately 8–12 g/dl) affect maximal duration of exercise after coronary artery surgery. 30Mild anemia (11.5 g/dl) in young healthy adults decreases DO²to the legs during maximal, but not submaximal, exercise. 31 

We did not achieve our goal of determining the critical level of systemic DO². We studied only eight volunteers, although we originally planned a larger study. When it became apparent that we would not define the critical DO², we were constrained not to enroll additional volunteers. We used β-adrenergic antagonism to reduce DO²below that achieved by severe anemia alone because of our concern for the safety of the volunteers at hemoglobin concentrations less than 5 g/dl. The methodology should not have influenced the results because reduction of DO²by either anemia or β-adrenergic blockade produces identical critical DO²in dogs. 2Only one subject had a final VO²less (by 17%) than his baseline VO². This occurred during the esmolol infusion, at a hemoglobin concentration of 4.5 g/dl and a DO²of 6.1 ml O²· kg⁻¹· min⁻¹(250 ml O²· min⁻¹· m⁻²). This subject did not have an abnormal blood lactate concentration and did not complain of any symptoms or have any abnormal ECG ST segments. It is possible that his DO²of 6.1 ml O²· kg⁻¹· min⁻¹was just below his critical value, but we could not confirm that by either plasma lactate concentration or Holter recording.

In summary, we found that reducing DO²to 7.3 ± 1.4 ml O²· kg⁻¹· min⁻¹by acute isovolemic anemia (hemoglobin = 4.7 ± 0.2 g/dl) plus an infusion of a β-adrenergic antagonist in resting healthy adults aged 19–25 yr does not produce evidence of inadequate systemic oxygenation. This extends our previous finding and suggests that the decreased DO²associated with a hemoglobin concentration of 4.5–5 g/dl is well-tolerated by conscious, healthy, young, resting adults.