Jugular Bulb Oxygen Saturation During Propofol and Isoflurane/Nitrous Oxide Anesthesia in Patients Undergoing Brain Tumor Surgery

Abstract

We investigated, in brain tumor patients, the jugular bulb venous oxygen partial pressure (PjO₂) and hemoglobin saturation (SjO₂), the arterial to jugular bulb venous oxygen content difference (AJDO₂), and middle cerebral artery blood flow velocity (Vmca) during anesthesia, and the effect of hyperventilation on these variables.Twenty patients were randomized to receive either isoflurane/nitrous oxide/fentanyl (Group 1) or propofol/fentanyl (Group 2). At normoventilation (PaCO₂ 35 +/- 2 mm Hg in Group 1 and 33 +/- 3 mm Hg in Group 2), SjO₂ and PjO₂ were significantly higher in Group 1 than in Group 2 (SjO₂ 60% +/- 6% and 49% +/- 13%, respectively; P = 0.019) (PjO₂ 32 +/- 3 and 27 +/- 5 mm Hg, respectively; P = 0.027). In Group 2, 5 of 10 patients had SjO₂ <50%, and 3 of these patients had SjO₂ <40% and AJDO₂ >9 mL/dL. All patients in Group 1 had SjO₂ >50%. During hyperventilation, there were no differences in SjO₂, PjO₂, or AJDO₂ between the two groups. On hyperventilation, there was no correlation between the relative decreases of Vmca and 1/AJDO₂ (r = 0.21, P = 0.41). The results indicate during propofol anesthesia, half of the brain tumor patients showed signs of cerebral hypoperfusion, but not during isoflurane/nitrous oxide anesthesia. Furthermore, during PaCO₂ manipulations, shifts in Vmca are inadequate to evaluate brain oxygen delivery in these patients. Implications: During propofol anesthesia at normoventilation, 50% of brain tumor patients showed signs suggesting cerebral hypoperfusion, but this could not be demonstrated during isoflurane/nitrous oxide anesthesia. During PaCO (₂) manipulations, consecutive measurements of the cerebral blood flow velocity may be inadequate to assess cerebral oxygenation.

(Anesth Analg 1999;89:358-63)

Isoflurane and propofol are widely used for anesthesia during intracranial surgery. Both drugs reduce the cerebral metabolic rate (CMR) for oxygen (CMRO₂) ^[1-4], and the resulting decrease of the cerebral oxygen demand implies increased tolerance for a subsequently imposed ischemic insult. In contrast to their similar effect on cerebral metabolism, isoflurane and propofol have different effects on cerebral blood flow (CBF).

Studies have shown that, although isoflurane reduces CMRO₂ dose-dependently, it maintains or increases CBF ^[5,6]. In addition, several authors have reported that the CBF is greater when nitrous oxide is added to isoflurane ^[7,8], but this produces no effects on CMR ^[9]. This suggests that nitrous oxide produces cerebral vasodilation, which is unopposed by CMR-mediated vasoconstriction. Thus, isoflurane/nitrous oxide anesthesia maintains or increases CBF and reduces CMR, thereby possibly producing an increase of the CBF/CMR ratio.

In contrast, propofol has been reported to reduce CBF in humans ^[3,4,10]. Several studies demonstrated that the reduction of CBF was larger than the reduction of CMRO₂, which suggests that propofol may have direct cerebral vasoconstricting activity ^[3,11,12], which might lead to a decrease of cerebral perfusion and a decrease of the CBF/CMR ratio. The hypothesis of this study was that, during propofol anesthesia, the brain is relatively hypoperfused compared with isoflurane/nitrous oxide anesthesia, resulting in a decreased brain oxygen supply and demand ratio. Accordingly, the purpose of our study was to determine the jugular bulb venous blood oxygen saturation (SjO₂) as a measure of the flow metabolism ratio during normoventilation and hyperventilation under propofol or isoflurane/nitrous oxide anesthesia in patients with intracranial mass lesions. Middle cerebral artery blood flow velocity (Vmca), measured by using transcranial Doppler sonography, was used to correlate the changes of SjO₂.

Methods

The study protocol was approved by our medical ethical committee. Twenty patients scheduled for elective brain tumor surgery gave their informed written consent to participate. The patients were randomly allocated into two groups: Group 1 (n = 10, 6 men/4 women), receiving isoflurane/nitrous oxide/fentanyl, and Group 2 (n = 10, 4 men/6 women), receiving propofol/fentanyl, using a nonblinded protocol design. None of the patients had signs of hepatic, renal, cardiac, pulmonary, or endocrine impairment, and they were without clinical signs of increased intracranial pressure. Preoperatively, with the patients lying in the supine position and in a slight head-up tilt, Vmca was measured on the tumor side (T side), as well as on the nontumor side (NT side) using a 2-MHz, pulsed-wave, range-gated transcranial Doppler sonography (TCD) probe (TC 2-64B; EME, Ueberlingen, Germany). Doppler signals were identified through the temporal window with a hand-held probe and measured at a depth of 45-55 mm, corresponding to the proximal segment of the middle cerebral artery. In each patient, a constant depth range was maintained throughout the study. Vmca values, as calculated by the TCD apparatus over four to five heart cycles, were registered. For every individual measurement, four sequential registrations were averaged. All patients received diazepam 10 mg orally as premedication. In Group 1, anesthesia was induced with IV thiopental 5 mg/kg and fentanyl 2-3 [micro sign]g/kg. Neuromuscular blockade was achieved with pancuronium bromide 0.1 mg/kg IV, followed by endotracheal intubation. End-tidal isoflurane concentration was maintained at 0.4%-0.6%, and an oxygen/nitrous oxide mixture was given to maintain the fraction of inspired oxygen FIO₂ at 0.4. In Group 2, anesthesia was induced with IV propofol 2 mg/kg and fentanyl 2-3 [micro sign]g/kg, and neuromuscular blockade was achieved with pancuronium bromide 0.1 mg/kg IV. After intubation of the trachea, the lungs were ventilated with an oxygen/air mixture to maintain FIO₂ at 0.4. Anesthesia was maintained with a continuous IV infusion of propofol 10 mg [middle dot] kg^-1 [middle dot] h^-1 for the first 10 min, 8 mg [middle dot] kg^-1 [middle dot] h^-1 for the next 10 min, and 6 mg [middle dot] kg^-1 [middle dot] h^-1 for maintenance during the whole anesthesia procedure. In all patients, ventilation was adjusted to achieve PaCO₂ of 35 mm Hg (normoventilation). Fentanyl, in doses of 50-100 [micro sign]g IV, was added to supplement anesthesia. After the induction of anesthesia, the radial artery was cannulated for monitoring mean arterial blood pressure (MAP) and for repeated blood sampling. An IV phenylephrine infusion was started, if necessary, to maintain MAP >70 mm Hg. A thermocouple was inserted into the nasopharynx to monitor body temperature, which was maintained at the preinduction temperature. A central venous line was instituted via the left cubital vein in the arm. Using Seldinger's technique, a 14-gauge Hydrocath[trade mark sign] (Ohmeda, Swindon, UK) percutaneous sheath introducer was inserted retrogradely into the right internal jugular vein in all patients independent of the side of the brain pathology. A 4F Opticath[registered sign] catheter (Abbott Laboratories, Chicago, IL) was inserted via the sheath with the tip of the catheter positioned in the right jugular bulb. Correct placement and positioning of this catheter were verified by lateral radiographic inspection. Forty-five minutes after the induction of anesthesia, a first set of measurements was performed for Vmca on the left and right side, MAP, heart rate, and nasopharyngeal temperature. Simultaneous arterial and jugular bulb venous blood samples were drawn for hemoglobin (Hb), hematocrit, and blood gas analysis and were immediately analyzed using an automated blood gas analyzer (Ciba Corning, Medfield, MA). Minute ventilation was then increased to achieve a PaCO₂ of 25 mm Hg (hyperventilation). After a 25-min period of stabilization, all measurements were repeated.

The arterial to jugular bulb venous oxygen content differences (AJDO (₂)) were calculated from the arterial and jugular bulb venous oxygen partial pressure and saturation using the equation: (Equation 1) where SaO₂ is arterial oxygen saturation, SjO₂ is jugular bulb venous oxygen saturation, and PjO₂ is the jugular bulb venous oxygen partial pressure. Global hypoperfusion was defined as SjO₂ <50%; ischemia was defined as SjO₂ <40% and AJDO₂ >9 mL/dL ^[13,14].

Because CMRO₂ is not influenced by hyperventilation during steady-state anesthesia, any change in cerebral oxygen consumption is considered to be reciprocally related to the change in CBF; thus, changes in 1/AJDO₂ will reflect changes in CBF. Because we assumed that, on hyperventilation, the relative changes of Vmca parallel relative changes of CBF, the relationship between the relative change of Vmca on the T side and the NT side, and the relative change of 1/AJDO₂ was accordingly subjected to linear correlation analysis using Pearson's correlation coefficient.

Comparisons between values were assessed by using analysis of variance. When significance was found, a post hoc test (Bonferroni-Dunn) was performed to delineate where differences lay. A P value < 0.05 was considered statistically significant.

Results

The two groups were similar with respect to demographic data and preoperative measurements (Table 1). In Group 2, 2 of 10 patients presented with a midline process. In these patients, the T side was considered the side on which the peritumoral edema was the most prominent on computed tomography scan. To maintain MAP >70 mm Hg, a phenylephrine infusion was required in three patients in Group 1 (at rates of 0.11, 0.13, and 0.38 [micro sign]g [middle dot] kg^-1 [middle dot] min^-1) and in two patients in Group 2 (0.15 and 0.18 [micro sign]g [middle dot] kg^-1 [middle dot] min^-1).

Table 1: Demographic Data and Preoperative Hemodynamic Values

The physiologic variables of both groups during anesthesia at normoventilation are shown in Table 2. PaCO₂ in Group 1 was slightly higher compared with Group 2, but the difference was not significant (35 +/- 2 and 33 +/- 3 mm Hg, respectively; P = 0.064). SjO₂ in Group 1 was significantly higher compared with Group 2 (60% +/- 6% and 49% +/- 13%, respectively; P = 0.019) (Table 2, Figure 1). All patients in Group 1 had a SjO₂ >50%, whereas SjO₂ <50% was demonstrated in five patients in Group 2. Three of these patients showed SjO₂ <40%. In the five patients in Group 2 who presented with SjO₂ < 50% during normoventilation, mean MAP (88 mm Hg) was not significantly different from the mean MAP (81 mm Hg) of the five patients who had SjO₂ >50%. PjO₂ in Group 1 was significantly higher compared with Group 2 (32 +/- 3 and 27 +/- 5 mm Hg, respectively; P = 0.027) (Table 2, Figure 1). AJDO₂ was not significantly different between the two groups. However, four patients in Group 2 presented with AJDO₂ >9 mL/dL, and three of these patients had SjO₂ <40%. All patients in Group 1 showed AJDO₂ <9 mL/dL.

Table 2: Physiologic Variables During Anesthesia

Figure 1: A, Jugular bulb venous oxygen saturation (SjO₂). B, Jugular bulb venous oxygen partial pressure (PjO₂) in the two groups at two different levels of PaCO₂ for each individual patient. Five patients from Group 2 showed SjO (₂) <50% at relative normoventilation. Two patients from Group 2 showed, on hyperventilation, an increase of SjO₂ at almost unchanged PjO_2-consistent with the Bohr effect.

Measurements during hyperventilation are shown in Table 2. In Group 1, four patients had SjO₂ <40% and AJDO₂ >9 mL/dL. In Group 2, five patients had SjO₂ <40% and AJDO₂ >9 mL/dL, and one patient presented with AJDO₂ >9 mL/dL.

Vmca on the T side, as well as on the NT side, was not significantly different between Group 1 and Group 2 in the awake state (Table 1) and during anesthesia at normoventilation and at hyperventilation (Table 2). The relative decrease of Vmca on the T side and NT side during anesthesia on hyperventilation was not significantly different between the two groups. There was no correlation between the relative decreases of Vmca on the T side and the NT side and the relative decrease of 1/AJDO₂ (r = 0.21, P = 0.41, n = 38) (Figure 2).

Figure 2: Effect of hyperventilation on the relationship between the percent change of the reciprocal of the arterial to jugular bulb oxygen content difference (1/AJDO (₂), which is an estimate of the cerebral blood flow) and the percent change of the middle cerebral artery blood flow velocity (Vmca) on the tumor side and the nontumor side in 38 paired data from 20 brain tumor patients during either propofol or isoflurane/nitrous oxide anesthesia.

Discussion

The main finding of this study in patients with intracranial mass lesions is significantly lower SjO₂ and PjO₂ values during propofol anesthesia than during isoflurane/nitrous oxide anesthesia.

In awake healthy humans, SjO₂ ranges between 55% and 75% (mean 62%) ^[13]. SjO₂ reflects the balance between brain oxygen supply and demand and indicates whether CBF is sufficient to satisfy the oxygen demands of the brain tissues. Values of SjO₂ <50% indicate cerebral hypoperfusion, and readings <40% are associated with global cerebral ischemia ^[14,15]. In our study, half of the patients under propofol anesthesia showed SjO₂ <50%, compatible with relative cerebral hypoperfusion, and three of these patients showed SjO₂ <40% and AJDO₂ >9 mL/dL, compatible with cerebral ischemia. None of the patients had SjO₂ <50% during isoflurane/nitrous oxide anesthesia. SjO₂ monitoring provides a global cerebral measurement. However, regional ischemia cannot be detected. Thus, the normal SjO₂ in the patients under isoflurane/nitrous oxide anesthesia does not indicate the absence of regional ischemia, but the low SjO₂ in three patients in the propofol group may be indicative of global ischemia, focal ischemia, or both.

Several studies have reported SjO₂ values >50% in patients with intracranial pathology undergoing craniotomy during anesthesia with 0.5%-1% isoflurane in nitrous oxide/oxygen or air/oxygen at a mean PaCO₂ of 30 mm Hg ^[16,17]. In our study, SjO₂ was 60% +/- 6% at a PaCO₂ of 35 +/- 2 mm Hg during isoflurane/nitrous oxide anesthesia, and all patients had SjO₂ > 50%. Stephan et al. ^[3] found SjO₂ of 41% at a mean PaCO₂ of 30 mm Hg in patients during propofol anesthesia who received multiple cardiac drug therapy. Another study showed SjO₂ of 60% in patients undergoing cerebral aneurysm surgery under propofol anesthesia at PaCO (₂) values of 26-34 mm Hg; 23% of these had SjO₂ <50% ^[18]. In our study, SjO₂ was 49% in the patients under propofol anesthesia at a mean PaCO₂ of 33 mm Hg. The results of our study are in agreement with data from previous studies ^[3,16-18] and demonstrate that SjO₂ during propofol anesthesia is lower than that during isoflurane/nitrous oxide anesthesia. One explanation for the lower SjO (₂) in the propofol group is an imbalance between the oxygen demand of the brain and CBF, with CBF being inadequate to supply oxygen to the brain. Stephan et al. ^[3] showed a decrease of CBF of 54% and a reduction of CMRO₂ of 33% after the induction of propofol anesthesia. Vandesteene et al. ^[4] also showed a significant decrease of CBF by 28% (P < 0.02) and an insignificant decrease of CMRO₂ during propofol anesthesia. Van Hemelrijck et al. ^[11] found that propofol anesthesia reduced CBF by 40% but failed to find any concomitant decrease in CMRO₂. In contrast, CBF was maintained at low isoflurane concentrations and was increased with higher isoflurane concentrations ^[5,6].

The CBF changes observed with propofol and isoflurane/nitrous oxide are the net result of two interactions. First, because these drugs depress CMRO₂, they produce an indirect decrease of CBF if coupling between CMRO₂ and CBF is preserved. Second, they produce either direct vasoconstriction or vasodilation of the cerebral vessels. An explanation for the decrease of CBF during propofol could be attributed to a direct vasoconstricting effect on cerebral vessels. However, direct vasoconstricting effects of propofol on cerebral arteries in animals in vitro have not been demonstrated ^[19,20]. In those in vitro studies, only the larger basal arteries (the capacitance vessels) were investigated, but the smaller cerebral arterioles (the resistance vessels) were not examined. An alternative explanation for a decrease of CBF in our patients when propofol was administered is that the lower limit of the CBF autoregulation was shifted to the right and that CBF was pressure-dependent at the measured MAP values. For example, Moss et al. ^[18] observed that the increase in MAP with IV phenylephrine improved SjO₂ to values >54% in eight of nine patients who had SjO₂ <54% during subarachnoid hemorrhage surgery under propofol anesthesia. They defined a critical MAP, which was attained with phenylephrine, and which was between 80 and 110 mm Hg in their patients, as the MAP that resulted in SjO₂ improvement >54%. This increase in SjO₂ with increasing MAP could be the result of either a defective autoregulation after subarachnoid hemorrhage or an increase of the lower limit of autoregulation under propofol anesthesia ^[11]. In our study, the five patients with SjO₂ <50% had MAP values between 69 and 110 mm Hg. Thus, MAP values under the critical MAP as a cause for the SjO₂ <50% in these five patients could not be excluded. However, in our patients receiving isoflurane/nitrous oxide anesthesia, signs of a critical MAP could not be demonstrated because all patients had SjO₂ >50%.

In Group 2, SjO₂ increased in two patients in response to hyperventilation from 50% to 59%, and from 34% to 43%, despite a relatively constant PjO₂ (29-27 mm Hg and 22-22 mm Hg) (Figure 1). Because P₅₀ is not influenced during propofol or isoflurane anesthesia, we suggest that these increases in SjO₂ are compatible with a leftward shift of the oxygen dissociation curve in relation to hyperventilation (Bohr effect) ^[21], yielding decreases of jugular bulb venous PCO₂ of 14 and 13 mm Hg and arterial pH values of 7.69 and 7.59. Data from these two patients demonstrate that continuous monitoring of SjO₂ to assess the brain oxygenation on hyperventilation might give the wrong impression of an improvement in brain oxygenation. Determination of PjO₂ is probably a more reliable monitor of oxygenation of the brain.

TCD is a noninvasive means by which to measure red blood cell velocity. Although TCD does not provide a direct measure of CBF, relative changes of Vmca accurately reflect relative changes of CBF, provided the diameter of the insonated vessel remains unchanged ^[22,23]. Studies have shown that the diameter of the MCA does not change significantly with changes of PaCO₂ or with the systemic administration of phenylephrine ^[24-26]. Matta et al. ^[24] showed, in healthy patients during pharmacologically induced isoelectric electroencephalogram, that the addition of inhaled anesthetics produced an increase of Vmca that closely correlates with the increase of 1/AJDO (₂) (1/AJDO₂ being an equivalent of CBF). However, in our study, there was no correlation between the relative changes of Vmca and 1/AJDO₂ on hyperventilation (Figure 2). A similar lack of correlation has been reported in other studies ^[10,27,28]. For example, in subjects who underwent cerebrovascular reactivity tests for a variety of types of intracranial lesions, Brauer et al. ^[27] found no correlation between changes of CBF and Vmca. During isoflurane/nitrous oxide anesthesia, Shah et al. ^[28] found the CBF response to CO₂ to be unpredictable in subjects with edematous brain tumors. In patients with a brain tumor, Schregel et al. ^[10] showed reduced or even paradoxical effects of propofol and hyperventilation on Vmca. The results of our study show that, on hyperventilation in brain tumor patients, the adequacy of cerebral oxygenation cannot be obtained from consecutive Vmca measurements alone, and shifts in Vmca as an indication of CBF changes should be interpreted with caution in patients with brain tumors.

In conclusion, brain tumor patients under propofol anesthesia, even at clinically accepted PaCO₂ levels, showed SjO₂ and PjO₂ levels that were significantly lower than those in patients under isoflurane-nitrous oxide anesthesia. In these patients, consecutive Vmca measurements were inadequate to assess the cerebral oxygenation during PaCO₂ manipulations.