Hansen, D. MD; Syben, R. MD; Vargas, O. MD; Spies, C. MD; Welte, M. MD
Klinik fur Anaesthesiologie und operative Intensivmedizin, Universitatsklinikum Benjamin Franklin, Freie Universitat Berlin, Berlin, Germany.
Accepted for publication December 1, 1998.
Address correspondence and reprint requests to Dr. D. Hansen, Klinik fur Anaesthesiologie und operative Intensivmedizin, Universitatsklinikum Benjamin Franklin, Freie Universitat Berlin, Hindenburgdamm 30, 12200 Berlin, Germany. Address e-mail to firstname.lastname@example.org.
Section Editor: Peter M. Suter.
Mild to moderate hypothermia occurs in anesthetic practice. Apart from unintended perioperative hypothermia  or hypothermia during massive transfusion , hypothermia is widely used during cardiopulmonary bypass . Furthermore, moderate hypothermia of 32-33[degree sign]C may be beneficial in patients with head injuries and increased intracranial pressure [4,5]. The accurate assessment of pulmonary gas exchange in head-injured patients could be important because they often experience respiratory failure because of additional chest trauma or sepsis [5,6]. In this context, it is still controversial whether arterial blood gases should be corrected to body temperature or whether uncorrected blood gases should be used for the management of hypothermic patients .
During hypothermia, the solubility of gases such as oxygen and carbon dioxide is increased. This results in a decrease of the tension of these gases for a given content of O2 and CO2 . Therefore, the Pa-CO2 corrected to actual body temperature should be used to calculate the arterial to end-tidal carbon dioxide gradient during hypothermia . The same applies for the alveolar-arterial oxygen tension difference (AaDO2). The AaDO2 is a widely used variable for assessing intrapulmonary oxygen exchange and can be calculated by subtracting PaO2 from alveolar oxygen tension (PAO2). In contrast to the arterial to end-tidal CO2 difference (Pa-ETCO2) the AaDO2 during hypothermia depends on more variables that change while the temperature decreases. The saturated water vapor pressure is reduced during hypothermia, and the respiration quotient may change.
It is surprising that these variables and their effect on pulmonary oxygen exchange have not yet been investigated in ventilated, moderately hypothermic patients. In a recent study, only the Pa-ETCO2 values at 34[degree sign]C and 32[degree sign]C were compared with 36[degree sign]C . No change in the Pa-ETCO2 was detected when the temperature-corrected arterial CO2 tension (cPaCO2) was used .
We designed a study to test the hypothesis that moderate hypothermia has no effect on the AaDO2 if all variables of the AaDO2 Equation arecorrected to actual body temperature.
After the study was approved by the hospital ethics committee and informed consent was obtained from all patients, we studied eight ASA physical status I patients during rewarming from hypothermia of 32[degree sign]C. These eight patients were a subgroup of patients scheduled for the resection of a malignant choroideal melanoma in profound arterial hypotension. For organ protection during profound hypotension, moderate hypothermia of 32[degree sign]C was induced before surgery. In the eight patients studied, hypotension was not induced because the tumor was found not to be resectable. All patients had normal preoperative spirometry values. Routine monitoring of these patients included direct arterial blood pressure monitoring and insertion of a catheter into the jugular bulb for continuous monitoring of cerebral venous oxyhemoglobin saturation. Anesthesia was maintained by continuous infusion of 6 mg [center dot] kg-1 [center dot] h-1 propofol and 1-3 [micro sign]g [center dot] kg-1 [center dot] h-1 fentanyl. The body temperature was taken as the blood temperature in the jugular bulb. Rewarming was induced by surface warming with a forced-air warming blanket. The patients were ventilated with a Servo 900 B ventilator (Siemens, Germany). The tidal volume was set to 10 mL/kg. The respiratory rate was adjusted to maintain normoventilation (temperature uncorrected PaCO2 38-42 mm Hg). At 32[degree sign]C, 33[degree sign]C, 34[degree sign]C, 35[degree sign]C, and 36[degree sign]C, the following measurements were performed: heart rate and mean arterial blood pressure; arterial blood samples drawn from the arterial catheter and analyzed for PaO2, cPaO2, PaCO2, and cPaCO2; inspired oxygen fraction (FIO2) and PETCO2 (side-stream gas analyzer was calibrated to the actual barometric pressure); respiratory quotient (RQ; the monitor was calibrated using a standard CO2-containing gas).
The actual barometric pressure (BP) was obtained from the meteorological institute, which is 1 km away from the hospital. The saturated water pressures (PH (2) O) at different temperatures were taken from physical standard tables as follows: 47 mm Hg at 37[degree sign]C, 44.3 mm Hg at 36[degree sign]C, 42 mm Hg at 35[degree sign]C, 39.5 mm Hg at 34[degree sign]C, 37.5 mm Hg at 33[degree sign]C, and 35.5 mm Hg at 32[degree sign]C.
The following variables were calculated: AaDO2: (Equation 1) cAaDO (2): (Equation 2), (Equation 3) and (Equation 4)
All data are given as mean +/- SEM. Data were analyzed for normal distribution (Kolmogorov-Schmirnov test). After analysis of variance was performed, changes of variables with body temperature were analyzed for significance using Student's t-test for paired samples. Differences were considered significant at P < 0.05.
Five male and three female patients were included in this study. The age of the patients was 40 +/- 6 yr, their body weight was 79 +/- 4 kg, and their height was 175 +/- 3 cm. The hemodynamic data, the RQ, and the values for cPaO2 and PaO2 during rewarming from 32[degree sign]C to 36[degree sign]C are given in Table 1. The mean arterial blood pressure was unaffected by moderate hypothermia. The heart rate was significantly slower at temperatures <36[degree sign]C. The RQ did not change at temperatures of 32[degree sign]C, 33[degree sign]C, 34[degree sign]C, or 35[degree sign]C compared with 36[degree sign]C (Table 1).
The temperature-uncorrected AaDO2 was not affected by hypothermia. However, when PH2 O, PaCO2, and PaO2 were corrected to actual body temperature, the cAaDO2 was higher at temperatures <36[degree sign]C compared with those at 36[degree sign]C. At 33[degree sign]C and 32[degree sign]C, these differences reached statistical significance (56 +/- 13 and 64 +/- 14 vs 39 +/- 10) (Figure 1). The cPa-ETCO2 values were not affected by hypothermia, but the Pa-ETCO (2) value was significantly higher at 33[degree sign]C and 32[degree sign]C compared with 36[degree sign]C (Table 2).
The most important finding of this study is that intrapulmonary oxygen exchange deteriorates during moderate hypothermia. The mean cAaDO2 was 64% higher at 32[degree sign]C compared with 36[degree sign]C when the following physical aspects were considered while assessing the AaDO2 during hypothermia. The PH2 O at the actual body temperature decreases during the reduction of body temperature. The PaCO2 must be corrected to the actual body temperature because the cPaCO (2) is the ambient PaCO2 in the arterial blood of the patient. Rewarming the blood to 37[degree sign]C in the blood gas analyzer results in a higher PaCO2 because the solubility of CO2 decreases with increasing temperature. This results in a higher PaCO2 for a given CO2 content in the blood sample . The same applies for PaO2, which must also be corrected to actual body temperature.
The increase in AaDO2 during hypothermia in our patients confirms experimental data on intrapulmonary gas exchange . In mechanically ventilated dogs, the AaDO2 increased from 20 mm Hg at 38[degree sign]C to 73 mm Hg at 25 [degree sign]C . In that study, the PaO2 decreased from 78 mm Hg at 38[degree sign]C to 53 mm Hg at 25[degree sign]C. The blood gases were measured at actual body temperature by tonometry and are comparable with the temperature correction of blood gases we used in our study.
The cause of the increase in AaDO2 during hypothermia is unknown. Hemodynamic changes may result in changes in the ventilation-perfusion ratio in the lung. The cardiac index decreases during hypothermia, which is reflected by a decrease in heart rate [11,12]. The systemic vascular resistance increases and the mean arterial blood pressure remains unchanged during moderate hypothermia . The decrease in heart rate and the constant mean arterial blood pressure in our study confirm these findings. Beside these global hemodynamic changes, hypothermia of 21[degree sign]C has been shown to change blood flow distribution in dogs. The bronchial arterial blood flow decreased to 20% of baseline during hypothermia . Additionally, the oxygen diffusion capacity decreased by 52% at 25[degree sign]C compared with 38[degree sign]C in dogs .
The maximal increase in AaDO2 in our study (i.e., 32[degree sign]C) was 25 mm Hg. This increase normally does not lead to any major changes in the clinical management of ventilated patients with moderate hypothermia. However, animal studies have shown that a body temperature <32[degree sign]C leads to further increases in AaDO2 and to clinically significant decreases in PaO2 , which may require an increase in FIO2. An important finding of our study is that the deterioration of intrapulmonary oxygen exchange during hypothermia is not detected when the variables of the AaDO2 Equation arenot corrected to the actual body temperature. When PH2 O and PaCO2 are not corrected, the alveolar PO2 will be falsely low. Together with an uncorrected, falsely high PaO2, AaDO2 will be assessed as being too low, and a deterioration of oxygen exchange will not be detected.
We found the RQ to be unchanged during moderate hypothermia. This is in accordance with experimental studies. The RQ was assessed to be 0.86 at 38[degree sign]C and 0.83 at 25[degree sign]C in dogs . Therefore, it does not seem to be necessary to measure RQ during moderate hypothermia. The widely used RQ of 0.8 applies for moderately hypothermic patients.
In our study the Pa-ETCO2 was not affected by moderate hypothermia when PaCO2 was corrected to the actual body temperature. This result is in accordance with the data of Sitzwohl et al. , who did not find any change in Pa-ETCO2 at 32[degree sign]C compared with 36[degree sign]C when PaCO2 was corrected to the actual body temperature. These data suggest that dead-space ventilation is not increasing during moderate hypothermia.
In summary, to accurately assess the AaDO2 and the Pa-ETCO2 during moderate hypothermia, blood gases and PH2 O should be corrected to the actual body temperature. Corrected values of AaDO2 increase with decreasing body temperature.
The authors thank Prof. Stein and Prof. Foerster for their excellent editorial assistance.
Formulas for temperature correction of PaO2 and PaCO2 used in the ABL 235 blood gas analyzer (Radiometer, Copenhagen, Denmark)
PaO2: (Equation 5) where cPaO2 = PaO2 corrected to actual body temperature, T = actual body temperature, and f = 0.058/(0.243 [PaO2/100] x 3.88 + 1) + 0.013.
PaCO2: (Equation 6) where cPaCO2 = PaCO2 corrected to actual body temperature and T = actual body temperature.
1. Frank StM, Higgins MS, Breslow MJ, et al. The catecholamine, cortisol, and hemodynamic response to mild perioperative hypothermia. Anesthesiology 1995;82:83-93.
2. Dunham CM, Belzberg H, Lyles R, et al. The rapid infusion system: a superior method for the resuscitation of hypovolemic trauma patients. Resuscitation 1991;21:207-27.
3. Baraka AS, Muallem MK, Baroody MA, et al. The temperature-corrected versus the uncorrected PO2
during hypothermic cardiopulmonary bypass-correlation with oxyhemoglobin saturation. J Extra Corp Technol 1993;24:130-4.
4. Marion DW, Penrod LE, Kelsy SF, et al. Treatment of traumatic brain injury with moderate hypothermia. N Engl J Med 1997;20:540-6.
5. Shiozaki T, Sugimoto H, Taneda M, et al. Effect of mild hypothermia on uncontrollable intracranial hypertension after severe head injury. J Neurosurg 1993;79:363-8.
6. Clifton GL, Allen S, Barrodale P, et al. A phase II study of moderate hypothermia in severe brain injury. J Neurotrauma 1993;10:263-73.
7. Prough DS, Stump DA, Troost BT. PaCO2
management during cardiopulmonary bypass: intriguing physiologic rationale, convincing clinical data, evolving hypothesis? Anesthesiology 1990;72:3-6.
8. Kofstad J. Blood gases and hypothermia: some theoretical and practical considerations. Scand J Clin Lab Invest 1996;56:21-6.
9. Sitzwohl Ch, Kettner StC, Reinprecht A, et al. The arterial to end-tidal carbon dioxide gradient increases with uncorrected but not with temperature-corrected PaCO2
determination during mild to moderate hypothermia. Anesth Analg 1998;86:1131-6.
10. Albers C, Brendel W, Hardewig A, Usinger W. Der intrapulmonale Gasaustausch in tiefer Hypothermie. Pflugers Arch Physiol 1958;266:394-407.
11. Bullard RW. Cardiac output of the hypothermic rat. Am J Physiol 1959;196:415-9.
12. Blair E, Montgomery V, Swan H. Posthypothermic circulatory failure. Circulation 1956;13:909-15.
13. Anzai T, Turner MD, Gibson WH, Neely WA. Blood flow distribution in dogs during hypothermia and posthypothermia. Am J Physiol 1978;234:H406-10.