Determinants of Cerebral Oxygenation During Cardiac Surgery
Background Neuropsychological deficits after cardiac surgery are attributed to the side effects of cardiopulmonary bypass (CPB). To protect the brain from ischemic damage, the influences of temperature, blood pressure, blood gases, acid-base status, and hemodilution on cerebral oxygenation have to be elucidated and quantified.
Methods Forty-one consecutive patients were investigated during cardiac surgery while on CPB. Operative management included moderate hypothermia of 26°C and the alpha-stat pH management. With near-infrared spectrophotometry, changes in oxygenated hemoglobin (Hbo2, representing oxygen delivery) and oxidized cytochrome a,a3 (Cto2, cellular oxygenation) in brain tissue were obtained noninvasively. In addition, venous saturation of the brain was measured via a catheter in the jugular bulb (SBJo2). The influence of operative management parameters on cerebral oxygenation was calculated by univariate and multiple regression analyses.
Results Before and after CPB there was no significant multivariate determinant of cerebral oxygenation. During CPB, Hbo2 depended solely on Pco2 (P<.01; r=.89). Cto2 was determined by pH (P<.01), esophageal temperature (P<.01), Pco2 (P<.01), and Hb (P<.01). These parameters explained nearly all changes of the cytochrome measurements during CPB (r=.99). Arterial Pco2 (P<.01) and pH (P<.01) influenced brain venous oxygen saturation (SBJo2; r=.98).
Conclusions Cerebral oxygenation is autoregulated during cardiac surgery before and after CPB. During CPB, Hb, temperature, pH, and Pco2 determine at least 85% of all changes in cerebral oxygenation. The main causes of impaired cerebral oxygenation are the decrease in Hb with hemodilution, vasoconstriction due to hypocapnia, and the leftward shift of the Hb binding curve in alkalosis and hypothermia.
Despite approximately 40 years of experience, there are many unresolved questions concerning the “ideal” CPB method. Neurological and neuropsychological deficits are not uncommon after cardiac surgery and have been attributed to the adverse effects of CPB on cerebral perfusion and oxygenation.1
Cerebral perfusion and oxygenation are regulated under physiological conditions, but it is well known that profound hypocarbia and hypotension2 may impair oxygen delivery to the brain. Other factors such as temperature, acid-base status, arterial Po2, hemodilution, and flow pattern and duration, and flow rate of the extracorporeal circulation may influence cerebral perfusion; the degree of the influences and the interactions of all these parameters are, however, not satisfactorily understood.3 4
During hypothermia, oxygen metabolism of the brain is reduced and oxygen transport to the tissue is affected by the leftward shift of the Hb binding curve. To obtain information on cerebral oxygenation, we therefore monitored three different parameters that represent oxygen delivery to the brain as well as cellular oxygenation and brain venous oxygen saturation. In recent years, NIRS has been shown to be useful in measuring cerebral oxygenation in both children and adults. NIRS obtains information on oxygen delivery (oxygenated and deoxygenated Hb) and cellular oxygenation (Cto2, the terminal enzyme of the respiratory chain).5 Cerebral venous oxygen saturation can be estimated by SBJo2. Information on cerebral oxygen uptake can be obtained by taking into consideration arterial oxygen saturation, flow, and Hb.
In this study, parameters of surgical management were analyzed to quantify their influence on cerebral oxygenation with and without CPB.
Forty-one consecutive patients who were undergoing cardiac surgery with CPB were investigated. The study was approved by the ethics committee of the Ludwig-Maximilians-University in Munich, and informed consent was obtained from all patients. The group comprised 9 women and 32 men with ages at surgery ranging from 32 to 76 years (mean age, 61.2 years). Twenty-five patients underwent coronary artery bypass graft (CABG) surgery; 14 patients had valve replacement (aortic valve, n=10; mitral valve, n=4); two patients had a combination of CABG surgery and valve replacement. The cardiological diagnosis was confirmed by a preoperative cardiac catheterization in all patients. In addition, all patients underwent Doppler sonography of the carotid arteries, and in the event of pathological findings, brain computed tomography (CT) scan and single photon emission CT after inhalation of xenon-133 (SPECT) were carried out. All patients were free of cerebral infarction preoperatively; however, in nine cases significant stenoses of one or both internal carotid arteries were detected. Two of the nine patients had a severe and five a moderate stenosis on the left side, where the NIRS measurements were obtained.
General Anesthesia and Surgical Management
Anesthesia was induced with etomidate 0.3 mg/kg and 3 μg/kg fentanyl. Each patient received 0.1 mg/kg pancuronium for muscle relaxation. After induction of anesthesia, controlled ventilation with an Fio2 of 0.5 was instituted. Anesthesia was maintained with isoflurane 0.6% to 1.0% and fentanyl in 0.5-mg increments as required. A Sarns 9000 heart-lung machine with nonpulsatile flow mode and membrane oxygenators (Maxima, Medtronic Inc) was used to provide extracorporeal circulation. The priming volume of the heart-lung machine was 1800 mL and consisted of a balanced electrolyte solution and 20% mannitol (3 mL/kg body wt). The average CPB time was 96.1±26.8 minutes. During CPB, a cardiac index of 2.4 L · min−1 · m−2 and alpha-stat blood gas values and systemic hypothermia to an esophageal temperature of 25.5±3.8°C were used. Alpha-stat blood gas values were later corrected for esophageal temperature6 to correlate the Pco2 and pH to Cto2 shifts.
Using an NIRS monitor (NIRO 500, Hamamatsu Corp), cerebral oxygenation was measured photometrically intraoperatively in all 41 patients. Two optodes were placed in a paramedian and a frontotemporal position on the left forehead 5 cm apart. The emitted laser light (wavelength 770 to 910 nm) is able to penetrate the skull and is dispersed in brain tissue, where light of specific wavelengths is absorbed by oxygenated and deoxygenated Hb (Hbo2, Hb) and Cto2. The optical path length in brain tissue can be estimated within 29.7 cm. The amount of reflected light of the specific wavelengths, the optical path length, and the brain specific gravity (1.05) allow the measurement of chromophores in brain tissue in concentration (μmol/L). The NIRS system also provides the parameter of total Hb (tHb) (using the formula tHb=Hb+Hbo2). The measurements were displayed and recorded every 2 seconds.
Oximetry in the Internal Jugular Vein
Only 11 of the 41 patients consented to the intraoperative measurement of jugular venous oxygen saturation. In three cases monitoring of oxygen saturation showed major artifact; therefore measurements in eight patients were analyzed. Correlation coefficients of SBJo2 to other data were calculated only for these cases. Four of these patients underwent CABG surgery, and four had heart-valve replacement. For the measurements, a fiberoptic catheter (Abbott Laboratories) was introduced into the left internal jugular vein and advanced into the jugular bulb under x-ray control. Oxygen saturation in the jugular bulb was measured and displayed continuously on an oxygen saturation monitor; this measurement reflects the venous oxygen saturation of the brain.
Standard Intraoperative Data
Mean arterial pressure, heart rate, and temperature (rectal and esophageal) were displayed and recorded continuously. For each measuring time (see Fig 1⇓), arterial blood-gas samples were analyzed (ie, Hb concentration, Po2, Pco2, oxygen saturation, pH, and base excess). Data from the heart-lung machine (ie, flow rate, gas flow, and Fio2) were recorded simultaneously.
All values are expressed as mean±SEM. For statistical analysis, data were obtained from 13 standardized points of measurement in each surgery. The values of point 1, obtained 10 minutes before thoracotomy, were used as baseline values. The points of measurements were as follows: Before CPB—point 1, 10 minutes before thoracotomy; point 2, during thoracotomy; point 3, 10 minutes after thoracotomy; and point 4, during aortic cannulation. Points during CPB were point 5, at onset of CPB; point 6, with aortic clamping on; point 7, 15 minutes after onset of CPB; point 8, at beginning of rewarming; point 9, 15 minutes after beginning of rewarming; point 10, with aortic clamping off; point 11, at end of CPB. After CPB, point 12 was 15 minutes after end of CPB and point 13, closure of thorax.
A value of P=.05 was considered significant. Differences between groups were tested with the Wilcoxon test for independent groups. Differences in time for one group were tested with the Wilcoxon test for related samples and corrected for multiple testing according to Bonferroni. Spearman correlation coefficients were used for statistical correlations. Parameters of surgical management were first tested in a linear univariate regression model. Every univariate significant parameter was included in a Cox multiple regression model and removed stepwise if no significant influence was proved. Statistical analysis was facilitated with the help of spss (SPSS Inc) software.
One patient died on the second postoperative day because of myocardial infarction due to early graft closure. No patient suffered from central neurological deficits postoperatively; one patient had a transient peripheral paresis of the left arm due to operative placement. Four patients showed transient neuropsychological deficits postoperatively diagnosed by a score ≤23 in the Mini-Mental State test obtained from the first through the third postoperative day. These patients had a lower minimum Cto2 compared with those without these deficits (−4.5 versus −0.7 μmol/L; P<.05). All other measured parameters showed no significant differences between the groups.
NIRS Monitoring and SBJo2
The parameters Hbo2 and Cto2 obtained by NIRS and oxygen saturation monitoring in the jugular bulb showed characteristic changes during cardiac surgery (Fig 1⇑). At the onset of CPB, the parameters Cto2 and Hbo2 decreased continuously (Cto2, P=.104; Hbo2, P<.001), reaching the minimum (Cto2, −1.07±0.32 μmol/L; Hbo2, −7.81±1.29 μmol/L) at the beginning of the rewarming phase. During rewarming the parameters rose constantly, reaching initial levels again at the end of the operation. SBJo2 increased at the onset of CPB and remained elevated (85±4%) during hypothermia. At the beginning of the rewarming phase, a decrease to minimum values of 73±5% (at end of aortic clamping) was seen; at thorax closure, the parameters had regained initial levels (80±3%).
Correlations to Parameters of Intraoperative Management
The correlations of NIRS and SBJo2 values to standard intraoperative data during CPB were dependent on whether CPB was “on” or “off.”
Influences on Cerebral Oxygenation During the Period Before and After CPB
During the period without CPB, temperature (ranging from 35.5°C to 36.7°C), mean arterial pressure (75 to 87 mm Hg), heart frequency (82 to 101 L/min), arterial oxygen pressure (310 to 390 mm Hg), Pco2 (33 to 40 mm Hg), and pH (7.39 to 7.44) as well as arterial Hb (96 to 128 g/L) showed no significant influence on cerebral oxygenation in univariate analysis. Thus, multiple regression models were not calculated.
Influences on Cerebral Oxygenation During the CPB Period
To calculate the influences of blood gas levels on brain oxygenation, pH, Po2, and Pco2 were corrected for temperature.8 A survey of the univariate correlation coefficients between the parameters of operative management and cerebral oxygenation is given in the Table⇓. Mean arterial pressure, arterial Po2, and blood flow did not affect oxygenation.
By means of multiple regression analysis of all probable influencing factors (with a univariate probability of <.05), equations were calculated to explain changes in cerebral oxygenation that resulted from changes in surgical management. Hbo2 (measured by NIRS) depends primarily on Pco2 (see Fig 2⇓). The following equation is estimated to explain 75% of all changes in Hbo2: Hbo2 (NIRS)=0.518×Pco2−19.9 (r=.88; P=.075).
Cto2, the measure for cellular oxygenation, was influenced by temperature (T), arterial Hb, Pco2, and pH (see Fig 3⇓). These parameters are estimated to determine 99% of the fluctuations of Cto2: Cto2 (NIRS)=0.0018×Hb−6.697×pH−0.025×Pco2+0.024×T (r=1.0; P<.001).
The oxygen saturation in the jugular bulb was affected by Pco2 and pH (Fig 4⇓). Ninety-three percent of the SBJo2 variations are explained by these parameters: SBJo2=288.815×pH+2.322×Pco2−2143.35 (r=.98; P<.01).
In Fig 5⇓ the measured values are plotted against values derived from the multiple regression analysis to visualize the fit of the multiple regression models.
Influences on Cerebral Oxygenation by Open Heart Surgery and Stenosis of the Left Carotid Artery
The duration of CPB and the time of aortic clamping lasted longer in valve surgery than in coronary surgery (103.2±3.2 versus 91.1±5.0 minutes and 55.7±3.4 versus 70.3±3.0 minutes, respectively). In valve surgery, minimum esophageal temperatures were also lower (26.5±0.5°C) than in coronary surgery (27.8±0.6°C). No significant differences were found between the groups with respect to brain tissue oxygenation. In seven patients a significant carotid artery stenosis of the left side was detected preoperatively; five of these patients underwent coronary revascularization; two had valve replacement. Patients with carotid artery stenosis differed only with respect to age from those without (70.9±1.7 versus 58.8±1.9; P<.05). The stenosis of the carotid artery showed no influence on the NIRS measurements.
Measurements of Cerebral Oxygenation: Limitations
In this study, cerebral oxygenation was assessed by three different parameters. NIRS measurements revealed information on oxygen delivery to brain (Hbo2) and cellular oxygenation (Cto2). Data on venous oxygen saturation of the brain were provided by monitoring of SBJo2. Obviously, it is difficult to compare NIRS and SBJo2 recordings since NIRS measurements give only local information on brain oxygenation,9 whereas SBJo2 indicates the saturation of venous blood of one half of the brain.10 The optodes for the NIRS measurements were placed on the left forehead. Since brain tissue scatters light to a great extent, the average path length of the infrared light in brain tissue is long and might be estimated at 29.7 cm11 if the distance between the optodes is 5 cm. One may thus assume that our NIRS measurements provide information about the left frontal brain. However, we found a correlation between the Cto2 values obtained by NIRS and the gross neuropsychological outcome of these patients. Therefore, the local measurements might be representative of the oxygenation of the whole brain.
In this study, cerebral blood flow was not measured, limiting the interpretation of the results since it is not possible to differentiate between changes of flow and oxygen consumption. Cerebral blood flow is considered to decrease during CPB as a function of time.10 The influences of these time-dependent changes on our multivariate models cannot be ruled out since our data are derived from a clinical study and other parameters (eg, temperature) also change with time. Therefore, some interrelations might be concealed by this effect.
Cerebral Oxygenation Before and After CPB
For the periods of the surgery before and after extracorporeal circulation was used, no significant influence on any parameter of cerebral oxygenation was found. These findings underscore the fact that cerebral blood flow and oxygenation are autoregulated, not only within a wide variety of mean arterial pressures but also with respect to changes in temperature, pH, Pco2, and Hb. It is important to mention that this autoregulation could be proved only for certain ranges of these parameters. We measured the autoregulation for a mean arterial pressure between 75 and 86 mm Hg, which is a small section of the pressure range of 60 to 125 mm Hg reported in the literature.11 Esophageal temperature during this time ranged from normal values to mild hypothermia of 35.2°C, which is reported not to affect the autoregulation.12 Temperature-corrected pH was always within the range of normal values. Mild hypocarbia with temperature-corrected Pco2 values ranging between 30 mm Hg and normal values (38 to 42 mm Hg) did not significantly affect cerebral oxygenation in our study. Data from the literature suggest that the metabolic needs of the brain are met with a Pco2 value of as low as 33 mm Hg.13 The Hb values of 100 to 120 g/dL approach the hematocrit level of 33%, which is optimal for oxygen delivery to the brain.14 Cerebral oxygenation seems to be autoregulated during the operation periods before and after extracorporeal circulation. The reason for this finding might be that the determinants of cerebral oxygenation are either close enough to normal values or physiological. With CPB established, oxygenation of the brain is dependent on parameters of the operative management, but it remains unknown whether CPB itself is the reason for this dependency or, the determinants of cerebral oxygenation, which are not physiological during CPB.
Cerebral Oxygenation During CPB Oxygen Delivery
During extracorporeal circulation (CPB), Hbo2, the parameter that measures oxygen delivery to the brain, was only significantly influenced by arterial Pco2. Using alpha-stat pH management, Pco2 values corrected for temperature ranged from 25 to 36 mm Hg and were thus much lower than those obtained when not on CPB. This moderate hypocarbia causes vasoconstriction in the brain and decreases cerebral blood flow.4 Arterial pressure had no influence on oxygen delivery in our patients; mean arterial pressure was on average always >55 mm Hg. Autoregulation of cerebral blood flow during alpha-stat management and mild hypothermia is described for a perfusion pressure range as great as 20 to 100 mm Hg.15
Cerebral Cto2, representing cellular oxygenation, was influenced by esophageal temperature, arterial pH, Pco2, and Hb values. Hypothermia and an alkalotic pH level cause a displacement of the blood oxygen dissociation curve to the left, leading to possible impaired tissue oxygenation. Similar observations were made by measuring Po2 in the brain.16 Vasoconstriction of brain vessels and as a consequence reduced cerebral blood flow as seen during hypocapnia is assumed to be mainly mediated by increasing the pH of the perivascular space.17 Surprisingly, in our multiple regression model, Cto2 was correlated inversely to Pco2, indicating slightly better cellular oxygenation during mild hypocarbia provided that pH, temperature, and Hb levels were still stable. In the literature, few attempts have been made to differentiate between the effects of alkalosis and those of hypocarbia on cerebral oxygenation; data showing inverse effects of these parameters, to the best of our knowledge, have not been published. It therefore remains unclear whether the differentiation between the closely related parameters Pco2 and pH and their effects on cerebral oxygenation has physiological significance or is simply an error of our mathematical multiple regression model.
The Hb values during extracorporeal circulation decreased below 90 g/L. Oxygen transport to tissue is impaired when Hb values fall below 100 g/dL because decreasing oxygen transport capacity cannot be balanced by more improved rheological attributes.14
Oxygen Saturation in the Jugular Bulb
An increase in SBJo2 is, according to the multiple regression model, associated with alkalosis and increasing Pco2. As discussed above, the calculated inverse effects of hypocarbia and alkalosis remain unclear, especially because univariate regression showed an opposite result with respect to Pco2 (see Fig 4⇑); increased cerebral blood flow and impaired oxygen uptake might be the reason for increased oxygen saturation in the jugular bulb; Cto2 and SBJo2 had an inverse correlation during CPB (see the Table⇑). Therefore, increased SBJo2 during hypothermia must not be interpreted as an improved tissue oxygenation,18 since oxygen delivery to the brain at lower temperatures depends on dissolved oxygen and is a function of Po2 and flow and not of actual saturation.19 On the contrary, during CPB a high SBJo2 may give a hint of impaired oxygen uptake. For these reasons, the postulated “luxury perfusion” of the brain during pH-stat management20 diagnosed by a high jugular bulb venous saturation21 is questionable.
In summary, the described multiple regression models are plausible and explain at least 85% of all measured changes in cerebral oxygenation. These models may give the surgical team practical hints to improve oxygen availability to the brain when no direct monitoring of cerebral oxygenation is available. Invasive measurements of venous saturation are only reliable about 80% of the time during nonpulsatile bypass22 and are difficult to interpret, because in hypothermia and alkalosis a high jugular venous saturation does not imply sufficient oxygenation of the brain. However, in a recent study22 the correlation between desaturation measured in the jugular bulb and impaired postoperative neuropsychological outcome was demonstrated, stressing the importance of cerebral oxygenation monitoring. NIRS measurements, on the contrary, are noninvasive, do not delay the operation, and provide detailed information on cerebral oxygenation; the presence of artifact is rare. Our superficial neuropsychological examinations also revealed a correlation between neuropsychological outcome and intraoperative brain hypoxia measured as a reduction in Cto2. This illustrates that this new method might have an important diagnostic value for postoperative neuropsychological dysfunction; further investigations are needed to prove this promising possibility.
Selected Abbreviations and Acronyms
|Cto2||=||oxidized cytochrome a,a3|
|SBJo2||=||oxygen saturation measurement in the jugular bulb|
- Copyright © 1995 by American Heart Association
Shaw PJ, Bates D, Cartlidge NEF, French JF, Heaviside D, Julian DG, Shaw DA. Neurologic and neuropsychological morbidity following major surgery: comparison of coronary artery bypass surgery and peripheral vascular surgery. Stroke. 1987;18:700-707.
Jöbsis FF. Noninvasive infrared monitoring of cerebral and myocardial oxygen sufficiency and circulatory parameters. Science. 1977;198:1264-1267.
Siggaard-Andersen O, Siggaard-Andersen M. The oxygen status algorithm: a computer program for calculating and displaying ph and blood gas data. Scand J Lab Invest. 1990;50(suppl 203):29-45.
Kern FH, Schell RM, Greeley WJ. Cerebral monitoring during cardiopulmonary bypass. J Neurosurg Anesthesiol. 1993;213:213-217.
Messmer K, Gornandt L, Jesch F, Sinagowitz E, Sunder-Plassmann L, Kessler M. Oxygen transport and tissue oxygenation during hemodilution with dextran. Adv Exp Med Biol. 1973;33:669-680.
Kuschinsky W, Wahl M. Local chemical and neurogenic regulation of cerebral vascular resistance. Physiol Rev. 1978;58:656-689.
Van der Linden J, Priddy R, Ekroth R, Lincoln C, Pugsley W, Scallan M, Tyden H. Cerebral perfusion and metabolism during profound hypothermia in children. J Thorac Cardiovasc Surg. 1991;114:102-103.