Hypercapnia can impair cells' capacity to maintain energy status anerobically and enhances the risk of hypoxic injury when oxygen availability is reduced. The ability to maintain tissue oxygenation is determined by both bulk blood flow and the efficiency of oxygen extraction. Bulk blood flow is maintained during hypercapnia through increased sympathetic activity. The effect of hypercapnia on oxygen extraction, however, is unknown. This study evaluates the effect of hypercapnia on cells' capacity to adapt to reductions in oxygen availability by increasing oxygen extraction.


In three groups of paralyzed, mechanically ventilated dogs that were anesthetized with alpha-chloralose, the concentration of carbon dioxide in the inhaled gas mixture was adjusted to achieve normocapnia, moderate hypercapnia (Paco2 = 72 +/- 3 [SE] mmHg) or severe hypercapnia (Paco2 = 118 +/- 4 [SE] mmHg). Stepwise hemorrhage was induced until each dog's blood pressure was destabilized. At each stage in the hemorrhage protocol, the oxygen delivery, oxygen consumption, and oxygen extraction ratios (ratio of arteriovenous oxygen content difference to arterial oxygen content) were determined.


At the point of onset of delivery dependence of oxygen consumption, the oxygen delivery rate (critical oxygen delivery) was 7.8 +/- 1.5 (SE) and the oxygen extraction ratio (critical oxygen extraction ratio) was 0.72 +/- 0.04 (SE) in the normocapnic dogs. Moderate hypercapnia had no effect on these parameters. In the severely hypercapnic dogs, the critical values for oxygen delivery and extraction ratios were 12.5 +/- 1.8 (SE) and 0.54 +/- 0.035 (SE), respectively (P < 0.05 for differences from the normocapnic dogs).


The results identify a previously unrecognized threat to tissue oxygenation and emphasize the importance of ensuring adequate oxygen delivery when adopting mechanical ventilatory strategies that permit respiratory acidosis to develop.

IN the setting of respiratory failure, severe hypercapnia usually prompts emergency institution of mechanical ventilatory support to achieve normal arterial PCO2(PaCO2) and pH. Recently, the adverse effects of hypercapnia were shown to be overstated. [1]Similarly, the traditional mechanical ventilatory practice of trying to achieve normocapnia in patients with severely diseased lungs is increasingly recognized to aggravate lung injury. [2]These observations have popularized an alternative strategy in which greater priority is given to limiting pulmonary hyperinflation than to maintaining normal alveolar ventilation. [3,4]Preliminary experience has suggested that this approach may improve patient outcomes, [3,4]and thus physicians working in intensive care units have accepted it rapidly. As a result, however, what once would have been regarded as unacceptable derangements of PaCO2and acid-base balance have become increasingly prevalent among critically ill patients.

Although initial case reports and clinical trials involving small numbers of patients have been encouraging, insufficient data are available to determine the level of hypercapnia that can be safely tolerated or its influence on coexistent diseases in other organ systems. Of particular concern in critically ill patients is the interaction between hypercapnia and the adaptive changes in cellular metabolism and vascular regulation that defend tissues against the effects of hypoxia. Hypercapnia potently inhibits glycolytic metabolism through the inhibitory effect of intracellular acidosis on phosphofructokinase activity. [5]Energy status is preserved through a compensatory increase in oxidative deamination, with depletion of the intracellular amino acid pool. [5,6]As a result, tissues become dependent on oxidative processes for adenosine triphosphate production and vulnerable to dysfunction if oxygen availability is reduced. Not surprisingly, therefore, oxygen availability has been identified as the key variable determining survival from severe hypercapnia. [7].

Tissue oxygenation is determined by both bulk blood flow and the efficiency of oxygen extraction. Hypercapnia negatively influences myocardial contractility [8]and may threaten the ability to maintain bulk blood flow. Increased circulating catecholamine levels compensate for this effect, [8,9]however, and unless cardiac output is limited by hypovolemia or preexisting cardiac disease, bulk oxygen delivery is generally preserved. The effect of hypercapnia on oxygen extraction is unknown. The increase in catecholamine output would be expected to enhance oxygen extraction through alpha-adrenergic receptor stimulation. [10]Within the microcirculation, however, hypercapnia inhibits alpha-receptor activation [9]and may limit the efficacy of this compensatory mechanism. In patients in shock, therefore, efforts to limit the hemodynamic impact of mechanical ventilation through hypoventilation could contribute to tissue hypoxia. The current studies were performed to test the hypothesis that acute hypercapnic acidosis impairs oxygen extraction and to determine the level of PaCO2at which this effect may be anticipated.

Animal Preparation

Approval for the study was obtained from the institutional animal ethics committee. Eighteen mongrel dogs (weighing 28.8 +/- 0.97 kg SD) were anesthetized with sodium thiopental (10 mg/kg) followed by alpha-chloralose (60 to 80 mg/kg). The animals were supine, tracheally intubated with cuffed endotracheal tubes, and their lungs were mechanically ventilated (initial tidal volume of 12 ml/kg and frequency of 15/min). Positive end-expiratory pressure (5 cm) was applied at the expiratory line. A catheter was placed in the aorta through the right carotid artery to monitor arterial pressure and another catheter was placed in the right femoral vein to administer fluids. A balloon-tipped pulmonary artery catheter (Edwards Laboratories, Santa Ana, CA) was placed into the pulmonary artery through the right internal jugular vein and used to measure cardiac output by thermodilution and to draw samples of mixed venous blood. Core body temperature was kept constant at approximately 37 degrees Celsius by a heating pad placed under the animal. Pancuronium (0.1 mg/kg) was administered intravenously to all animals. Paralysis was necessary to prevent increased respiratory muscle activation due to central chemoreceptor activation and to avoid possible confounding influences of hypercapnia on basal skeletal muscle tone.

Oxygen Delivery and Consumption Measurements

Systemic arterial and mixed venous blood oxygen contents were measured using a co-oximeter (Instrumentation Laboratory IL-482, Lexington, MA). The arterial oxygen content was multiplied by the cardiac output to determine the oxygen delivery rate. Systemic oxygen consumption (VO2) was measured by indirect calorimetry using a metabolic cart (Medgraphics CCM, Medical Graphics, St. Paul, MN) and by calculating the product of the cardiac output and the arteriovenous oxygen content difference.

Experimental Protocol

Three groups of six animals each were studied. In the first group, the ventilator settings were adjusted to maintain the arterial carbon dioxide tension (PaCO2) in the normocapneic range (38 to 42 mmHg). In the second group, the ventilator settings were adjusted to achieve normocapnia as in group 1. Using a gas cylinder and regulator, carbon dioxide was then added to the inspired gas mixture at a flow rate sufficient to maintain a target PaCO2rate of 70 mmHg (moderate hypercapnia). In the third group, the minute ventilation was set to achieve normocapnia, and then the flow of carbon dioxide into the inspiratory line was adjusted to maintain a target PaCO2rate of more than 100 mmHg (severe hypercapnia). Supplemental oxygen was supplied through the inspiratory line to maintain arterial PaO2at a rate greater than 100 mmHg. No effort was made to correct the pH change that accompanied increased PaCO2.

With the animal in a steady-state condition, cardiac output was measured and arterial and mixed venous blood samples were drawn to determine blood gas tensions and oxygen content. Stepwise hemorrhage was induced by withdrawing blood in 50- to 100-ml aliquots from the left femoral arterial line. After a steady-state condition was reached at each stage, oxygen uptake was recorded, cardiac output was measured, and arterial and mixed venous blood samples were drawn. Hemorrhage was continued until the dogs could no longer maintain stable blood pressures. To achieve this, 9 to 12 stages of hemorrhage were performed in each animal.

Data Analysis

The critical systemic oxygen delivery at which VO2became delivery dependent was calculated by plotting oxygen delivery on the x-axis against the corresponding oxygen consumption on the y-axis. For each animal, the data were sorted as oxygen delivery and oxygen consumption pairs with increasing oxygen delivery. All possible regression lines were then calculated as the data were grouped pair by pair into high- and low-oxygen delivery subsets. The pair of lines that resulted in the smallest sum of squared residuals was used in further calculations. The critical oxygen delivery rate was determined by solving the equations describing the regression lines for the value of this variable at the point of intersection.

The critical oxygen extraction ratio at the point of onset of delivery dependence of oxygen consumption was calculated as the ratio of oxygen consumption to oxygen delivery when the critical point corresponded to a direct measurement of oxygen delivery. When this was not the case, the critical extraction ratio was estimated by linear interpolation using values measured during the two stages bracketing the point of critical oxygen delivery. The maximum oxygen extraction ratio was taken as that during the final stage of each experiment.

Differences among mean values were evaluated using one-way analysis of variance. If that analysis revealed significant (P < 0.05) differences among the means, individual differences were evaluated post hoc using the Student-Newman-Keuls procedure. Probability values less than 0.05 were considered significant. Unless otherwise indicated, all values are reported as means +/- SEM.

(Table 1) shows the values for arterial blood gases, cardiac output, and mean arterial blood pressure before beginning the hemorrhage protocol and during the final stage of the experiments in the three groups. The critical values for oxygen delivery and oxygen extraction ratios determined using the indirect calorimetric measurements of oxygen consumption did not differ from those determined using values for oxygen consumption calculated from the thermodilution cardiac output and the arteriovenous oxygen content difference.

In Figure 1, mean values for oxygen delivery and oxygen consumption before hemorrhage, at the critical point, and during the final stage of the hemorrhage protocol are plotted for the normocapnic (group 1), moderately hypercapnic (group 2), and severely hypercapnic (group 3) animals. Critical oxygen delivery was significantly greater in the severely hypercapnic group than in the normocapnic group. The normocapnic and moderately hypercapnic animals did not differ.

(Figure 2) compares the critical and maximal oxygen extraction ratios (at the final stage of the hemorrhage protocol just before the animal became hemodynamically unstable) in the three groups of dogs. The critical and maximum oxygen extraction ratios were significantly less in the severely hypercapnic than in the normocapneic animals. In the moderately hypercapnic group, neither the critical nor the maximum oxygen extraction ratios differed from those recorded in the normocapnic group.

In Figure 3, the slope of the delivery-dependent portion of the oxygen delivery-consumption relationship is compared among the three groups. There was no difference between the slope of this relationship in the normocapnic and the moderately hypercapnic animals. The slope is less (P < 0.05) in the severely hypercapnic dogs (slope = 0.39 +/- 0.05, r2= 0.74) than that in the normocapnic animals (slope = 0.69 +/- 0.06, r2= 0.85).

The main findings of this study are that, during hemorrhagic shock in dogs, (1) moderate hypercapnia (PaCO2= 72 +/- 3 mmHg) had no effect on oxygen extraction or on the critical oxygen delivery; and (2) severe hypercapnia (PaCO2= 118 +/- 4 mmHg) increased the critical oxygen delivery and was associated with decreases in the critical and maximum oxygen extraction ratios as well as the slope of the delivery-dependent limb of the oxygen delivery-consumption relationship.

Effect of Hypercapnia on Oxygen Consumption

In the current study, hypercapnia did not alter baseline VO sub 2. Previous studies in which VO2was “preserved” have been advanced as evidence that hypercapnia does not produce clinically relevant, adverse metabolic effects. [9]These results, however, must be interpreted with care. Hypercapnia markedly increases circulating catecholamine levels, [11–13]and this, under other conditions, is associated with a large increase in the basal metabolic rate. [14]In addition, the efficiency of contractile processes is reduced by hypercapnic acidosis, and the oxygen cost of maintaining cardiac output [15]and ventilation [16]would increase accordingly. No increase in VO2in spontaneously breathing animals [17]is particularly striking because the excess oxygen demand associated with the increase in ventilatory work should be reflected in this measurement. Suppression of the expected increase in VO2, therefore, represents a negative influence of hypercapnia on oxygen use.

Hypercapnia could inhibit oxygen consumption by two mechanisms. A direct inhibitory action of carbon dioxide on cellular metabolism could reduce the need for oxygen uptake. Alternatively, hypercapnia could impair oxygen extraction to the extent that VO2becomes delivery dependent under basal conditions. The current results differentiate between these two possibilities because they show that the hypercapnic animals were not on the delivery-dependent limb of the oxygen delivery-consumption relationship before beginning the hemorrhage protocol. Suppression of the VO2response results, therefore, from an inhibitory action of carbon dioxide on cellular metabolic activity.

Effect of Hypercapnia on Oxygen Extraction

Little data concerning the effect of hypercapnia on the capacity to increase oxygen extraction are available for comparison with the present findings. In hypercapnic pigs (PaCO2= 125 mmHg), Wetterberg and colleagues [18]reported a premortal oxygen extraction ratio of 88% during progressive reductions in inspired oxygen concentration. Although these investigators made no attempt to identify the point of onset of delivery dependence of oxygen consumption, they noted a decline in oxygen consumption at a point when the oxygen extraction ratio was 63%. This is lower than the extraction ratio associated with delivery limitation of oxygen consumption in other species. [19].

In the present study, the critical oxygen extraction ratio in severely hypercapnic dogs was 54%, whereas in normocapnic dogs it averaged 72%. Assuming that the reduction in oxygen consumption observed by Wetterberg et al. [18]was attributable to limitation of the oxygen supply, the current findings are thus compatible with their results. The current results extend the previous findings by showing that the onset of supply dependence of oxygen uptake is hastened during severe hypercapnia due to limitation of the capacity to increase oxygen extraction. Maximal oxygen extraction was also reduced in the current study and supports this conclusion. It must be noted, however, that this value represents the premortal oxygen extraction ratio. Because hypercapnia may have impaired hemodynamic stability through effects on arterial tone or cardiac performance, the possibility that tissue oxygen extraction may have increased further had the dog survived longer cannot be excluded.

The onset of delivery dependence of oxygen consumption has been attributed to the development of tissue hypoxia as the limits of oxygen extraction are reached. [6]The current findings indicate that this limit is reduced during severe hypercapnia. This may arise because hypercapnia alters (1) the oxygen affinity of hemoglobin, (2) the degree of regional inhomogeneity in matching of oxygen delivery to metabolic demand, or (3) the intraregional distribution of nutritive compared with nonnutritive capillary perfusion.

Schumacker and associates [19]studied the effect of reducing hemoglobin P50with carbon monoxide on the systemic oxygen delivery-consumption relationship. They found no change in either the critical oxygen delivery or the extraction ratio at the point of onset of oxygen supply dependence. Other factors, therefore, appear to limit oxygen consumption before blood-tissue diffusion becomes important.

With regard to the second mechanism, Davidson and coworkers [20]studied moderate hypercapnia (average = 63 mmHg) in conscious sheep in which oxygen transport had been reduced by inhalation of hypoxic gas mixtures. In sheep that were both hypercapnic and hypoxic, the cardiac output was preferentially distributed to skeletal muscle. The change in blood flow was in proportion to skeletal muscle oxygen consumption, which was also increased. Because blood flow would have to be redistributed in disproportion to metabolic demand to impair oxygen extraction, no defect in whole-body oxygen extraction would be anticipated. This was, in fact, the result obtained in the current study in moderately hypercapnic dogs. The contribution of this mechanism to the extraction defect observed during more profound hypercapnia depends on the relative impact of carbon dioxide on metabolism compared with its effect on vascular tone, and further studies are indicated to evaluate this issue.

Interference with a normal increase in regional homogeneity of flow within tissues during reductions in bulk blood flow could, theoretically, contribute to impaired oxygen extraction. Hurtado and associates, [21]however, found that the muscle PO2histograms, an index of regional oxygenation, actually widened as blood flow declined during hypotension, indicating that this mechanism is not part of the normal adaptive response. A role in mediating the effect of hypercapnia on oxygen extraction is, therefore, unlikely.

With regard to the third mechanism, Whalen and Nair [22]found that myocyte, [22]as well as the oxidative state of skeletal muscle myoglobin and cyctochrome a,a3, [23]decline during hypercapnia, whereas venous PO2increases. [22]These results suggest shunting of blood away from metabolically active regions at the microcirculatory level. Hypercapnia relaxes preterminal arterioles through inhibition of alpha-adrenoreceptor activation. [10]Pharmacologic alpha -receptor antagonists reduce [24]whereas alpha-agonists increase [10]the critical oxygen extraction ratio. It is possible, therefore, that the effect of hypercapnia on oxygen extraction, observed in the current study, results from inhibition of adrenergically mediated optimization of the distribution of capillary flow.


The results presented in this report identify a previously unrecognized threat to tissue oxygenation that is relevant to patients experiencing acute hypercapnic acidosis. Hypercapnia handicaps cells' capacity to maintain energy status anaerobically and enhances the risk of hypoxic injury during anemia or other situations in which oxygen delivery is critically reduced. Consequently, adequate oxygen delivery must be maintained when implementing a management strategy that involves deliberate hypoventilation. In practice, the benefits of increasing pulmonary ventilation to reduce the PaCO2rate must be weighed against the potential detrimental effects on cardiac output and oxygen delivery. In this respect, the current results are reassuring because significant impairment of oxygen extraction occurred only during severe hypercapnia. Moderate elevations in PaCO2, which are more commonly encountered clinically, do not appear to cause this additional adverse effect.

The author thanks J. Petrella for technical assistance and Joan Longo for processing the manuscript. This study was funded by a grant from the Medical Research Council of Canada. Michael Ward is a scholar of the Medical Research Council of Canada.

Prys-Roberts C, Smith WDA, Nunn JF: Accidental severe hypercapnia during anaesthesia. A case report and review of some physiological effects. Br J Anaesth 1967; 39:257-67.
Kolobow T, Moretti MP, Fumagalli R, Mascheroni D, Prato P: Severe impairment of lung function induced by high peak airway pressure during mechanical ventilation. Am Rev Respir Dis 1987; 135:312-15.
Darioli R, Perret C: Mechanically controlled hypoventilation in status asthmaticus. Am Rev Respir Dis 1984; 129:385-7.
Hickling KG, Henderson SJ, Jackson R: Low mortality associated with low volume pressure limited ventilation with permissive hypercapnia in severe adult respiratory distress syndrome. Intensive Care Med 1990; 16:372-7.
Cohen Y, Chang LH, Litt L, Kim F, Severinghaus JW, Weinstein PR: Stability of brain intracellular lactate and 31P-metabolite levels at reduced intracellular pH during prolonged hypercapnia in rats. J Cereb Blood Flow Metab 1990; 10:277-84.
Cinglani HE, Koretsune Y, Marban E: Recovery of contractility and pHi during respiratory acidosis in ferret hearts: The role of Sodium sup + -Hydrogen sup + exchange. Am J Physiol 1990; 259:H843-H849.
Potkin RT, Swenson ER: Resuscitation from severe acute hypercapnia. Determinants of tolerance and survival. Chest 1992; 102:1742-1745.
Walley KR, Lewis TH, Wood LDH: Acute respiratory acidosis decreases left ventricular contractility but increases cardiac output in dogs. Circ Res 1990; 67:628-35.
Mittelman A, Dos SJ, Barker HG, Hahas GG: Adrenocortical response during corrected and uncorrected hypercapnic acidosis. Am J Physiol 1962; 202:334-6.
Maginniss LA, Connolly H, Samsel RW, Schumacker PT: Adrenergic vasoconstriction augments tissue Oxygen sub 2 extraction during reductions in Oxygen sub 2 delivery. J Appl Physiol 1994; 76:1454-61.
Nahas GG, Ligou JC, Mehlman B: Effects of pH changes on Oxygen sub 2 uptake and plasma catecholamine levels in the dog. Am J Physiol 1960; 198:60-6.
Cantu RC, Nahas GG, Manger WM: Effect of hypercapnic acidosis and of hypoxia on adrenal catecholamine output of the spinal dog. Proc Soc Exp Biol Med 1966; 122:434-7.
Brofman JD, Leff AR, Munoz NM, Kirchoff C, White SR: Sympathetic secretory response to hypercapnic acidosis in swine. J Appl Physiol 1990; 69:710-17.
Steinberg D, Nestel PJ, Buskirk ER, Thompson RH: Calorigenic effect of norepinephrine correlated with plasma free fatty acid turnover and oxidation. J Clin Invest 1964; 43:167-75.
Hata K, Goto Y, Kawaguchi O, Takasago T, Saeki A, Nishioka T: Hypercapnic acidosis increases oxygen cost of contractility in the dog left ventricle. Am J Physiol 1994; 266:H730-H740.
Karetzky MS, Cain SM: Effect of carbon dioxide on oxygen uptake during hyperventilation in normal man. J Appl Physiol 1970; 28:8-12.
McLellan TM: The influence of a respiratory acidosis on the exercise blood lactate response. Eur J Appl Physiol 1991; 63:6-11.
Wetterberg T, Sjoberg T, Steen S: Effects of hypothermia in hypercapnia and hypercapnic hypoxemia. Acta Anaesthesiol Scand 1993; 37:296-302.
Schumacker PT, Long GR, Wood LDH: Tissue oxygen extraction during hypovolemia: Role of hemoglobin P sub 50. J Appl Physiol 1987; 62:1801-7.
Davidson D, Stalcup SA, Mellins RB: Systemic hemodynamics affecting cardiac output during hypocapnic and hypercapnic hypoxia. J Appl Physiol 1986; 60:1230-6.
Hurtado FJ, Gutierrez AM, Silva N, Fernandez E, Khan AE, Gutierrez G: Role of tissue hypoxia as the mechanism of lactic acidosis during E. coli endotoxemia. J Appl Physiol 1992; 72:1895-1901.
Whalen WJ, Nair P: Skeletal muscle PO sub 2 effect of inhaled and mtopically applied Oxygen sub 2 and CO sub 2. Am J Physiol 1970; 218:973-80.
Hampson NB, Piantadosi CA, Jobsis-VanderVliet FF: Near infrared optical monitoring of cat skeletal muscle during hypercapnia. Adv Exp Med Biol 1986; 200:523-30.
Stainsby WN, Sumners C, Eitzman PD: Effects of adrenergic agonists and antagonists on muscle Oxygen sub 2 uptake and lactate metabolism. J Appl Physiol 1987; 62:1845-51.