Changes of cerebral regional oxygen saturation during pneumoperitoneum and Trendelenburg position under propofol anesthesia: a prospective observational study

Background We evaluated the change of cerebral regional tissue oxygen saturation (rSO2) along with the pneumoperitoneum and the Trendelenburg position. We also assessed the relationship between the change of rSO2 and the changes of mean arterial blood pressure (MAP), heart rate (HR), arterial carbon dioxide tension (PaCO2), arterial oxygen tension (PaO2), or arterial oxygen saturation (SaO2). Methods Forty-one adult patients who underwent a robotic assisted endoscopic prostatic surgery under propofol and remifentanil anesthesia were involved in this study. During the surgery, a pneumoperitoneum was established using carbon dioxide. Measurements of rSO2, MAP, HR, PaCO2, PaO2, and SaO2 were performed before the pneumoperitoneum (baseline), every 5 min after the onset of pneumoperitoneum, before the Trendelenburg position. After the onset of the Trendelenburg position, rSO2, MAP, HR were recorded at 5, 10, 20, 30, 45, and 60 min, and PaCO2, PaO2, and SaO2 were measured at 10, 30, and 60 min. Results Before the pneumoperitoneum, left and right rSO2 were 67.9 ± 6.3% and 68.5 ± 7.0%. Ten minutes after the onset of pneumoperitoneum, significant increase in the rSO2 was observed (left: 69.6 ± 5.9%, right: 70.6 ± 7.4%). During the Trendelenburg position, the rSO2 increased initially and peaked at 5 min (left: 72.2 ± 6.5%, right: 73.1 ± 7.6%), then decreased. Multiple regression analysis showed that change of rSO2 correlated with MAP and PaCO2. Conclusions Pneumoperitoneum and the Trendelenburg position in robotic-assisted endoscopic prostatic surgery did not worsen cerebral oxygenation. Arterial blood pressure is the critical factor in cerebral oxygenation. Trial registration Japan Primary Registries Network (JPRN); UMIN-CTR ID; UMIN000026227 (retrospectively registered).


Background
In robotic-assisted endoscopic prostatic surgery, a carbon dioxide pneumoperitoneum and the Trendelenburg position are essential. In the pneumoperitoneum and the Trendelenburg position, intracranial pressure was reported to increase [1]. In addition, pneumoperitoneum increases intraperitoneal pressure that leads to increase in intrathoracic pressure. Furthermore, the Trendelenburg position increases intrathoracic pressure. Increase in intrathoracic pressure should result in increase in central venous pressure [2]. Both intracranial pressure and central venous pressure increase in the pneumoperitoneum and the Trendelenburg position. The cerebral perfusion pressure is regarded as the mean arterial blood pressure (MAP) minus the intracranial pressure (when intracranial pressure > central venous pressure) or the central venous pressure (when central venous pressure > intracranial pressure) [3]. Therefore, unless the MAP changes the cerebral perfusion pressure decreased and cerebral circulation might be impaired in the pneumoperitoneum combined with the Trendelenburg position. Nevertheless, cerebral blood flow is maintained constantly within a wide range of cerebral perfusion pressure that is known as cerebral autoregulation [4]. On the other hand, cerebral blood flow has been reported to fluctuate even within autoregulation [5]. Therefore, cerebral blood flow may be unchanged or reduced after the pneumoperitoneum and the Trendelenburg position.
Measurement of cerebral regional tissue oxygen saturation values (rSO 2 ) using near infrared spectroscopy can allow to assess cerebral circulation [6]. A previous study investigated that cerebral oxygenation during pneumoperitoneum in the Trendelenburg position [7]. However, the Trendelenburg position was firstly placed followed by pneumoperitoneum in that study. In robotic-assisted endoscopic prostatic surgery, pneumoperitoneum is performed before the Trendelenburg position.
We tested the following hypotheses. Firstly, rSO 2 does not change after the pneumoperitoneum. Secondary, the Trendelenburg position combined with pneumoperitoneum does not alter rSO 2 . The primary outcome was the change of rSO 2 along with the pneumoperitoneum and postural change. The secondary outcome was the relationship between the change of rSO 2 and MAP, heart rate (HR), arterial carbon dioxide tension (PaCO 2 ), arterial oxygen tension (PaO 2 ), or arterial oxygen saturation (SaO 2 ).

Methods
This study was approved by the institutional review board of University of Yamanashi (study No. 488), and and was registered in the University Hospital Medical Information Network Clinical Trials Registry (UMIN-CTR) under study number UMIN000026227. Written informed consent was obtained from all patients.

Anesthesia
Fifty-six adult patients (ASA physical status I or II) who underwent robotic-assisted endoscopic prostatic surgery were recruited. Patients having history of cerebral diseases such as cerebral infarction, cerebral hemorrhage, transient ischemic attack, or subarachnoid hemorrhage were excluded. No premedication was given. In the operating room, two sensors of near infrared spectroscopy (SAFB-SM, Covidien, Dublin, Ireland) were attached on the patient's forehead to measure the left and right rSO 2 . A pulse oximeter was used to monitor percutaneous arterial oxygen saturation (SpO 2 ). A bispectral index (BIS) sensor was also attached on the forehead (Model QUATRO, Covidien). An earphone-type infrared tympanic thermometer (CE Thermo, Nipro, Tokyo, Japan) was used to monitor body temperature. Non-invasive arterial blood pressure, HR, rSO 2 , SpO 2 , and body temperature were measured before induction of general anesthesia while the patient breathed room air (Pre values). Fluid was infused at 6 ml/kg/hr. Anesthesia was induced and maintained with propofol using target controlled infusion and remifentanil. Tracheal intubation was facilitated with rocuronium. After the induction of anesthesia, radial arterial catheter was placed to allow continuous monitoring of MAP and blood gas analysis (SaO 2 , PaO 2 , and PaCO 2 ). The patient's lungs were mechanically ventilated in a volume controlled mode (tidal volume: 6-8 ml/kg) with a positive end-expiratory pressure of 3 cm H 2 O. Peak airway pressure was controlled below 22 cmH 2 O. PaCO 2 was maintaind between 35 and 45 mmHg. Fraction of inspired oxygen was adjusted to maintain PaO 2 between 150 and 250 mmHg. Position of the blood pressure transducer was standardized to place at the level of the ear for every patient. Mean arterial blood pressure was controlled within 60-120 mmHg. If MAP fell below 60 mmHg, ephedrine or phenylephrine was given. If MAP went up over 120 mmHg, the infusion rate of remifentanil was increased. Bispectral index (BIS) was adjusted between 40 and 60 by controlling the target of propofol infusion. If rSO 2 went down below 50%, or by 20% of the preanesthetic value, inspired oxygen was increased.

Measurements
Before pneumoperitoneum, baseline measurements of rSO 2 , MAP, HR, SpO 2 , SaO 2 , PaO 2 , and PaCO 2 were made. Pneumoperitoneum with intra-abdominal pressure of 10-15 mmHg was established and measurements were repeated every 5 min. Approximately 15 min after the establishment of the pneumoperitoneum, the Trendelenburg position with 30°head-down tilt was started. Before the start of the Trendelenburg position, rSO 2 , MAP, HR, SpO 2 , SaO 2 , PaO 2 , and PaCO 2 were measured. Measurements of rSO 2 , MAP, HR, and SpO 2 were made at 5, 10, 20, 30, 45, 60 min after the start of Trendelenburg position. After the onset of the Trendelenburg position, blood gas analysis was performed at 10, 30, 60 min. The measurements were not blinded to the anesthesiologists.

Statistical analysis
We used Stat Flex version 6.0 (Artec, Osaka, Japan) for statistical analysis. Power analysis revealed that the sample size of 41 patients was sufficient to provided 80% power with an α level of 0.05 to detect mean differences of 5% in rSO 2 . Change in rSO 2 , MAP, HR, SpO 2 , SaO 2 , PaO 2 , and PaCO 2 were examined via analysis of variance and Tukey post hoc comparisons. Multiple regression analysis was performed to estimate the relationship between rSO 2 and MAP, HR, SpO 2 , SaO 2 , PaO 2 , or PaCO 2 . For multiple regression analysis, rSO 2 data were averaged. When statistical significances were obtained after the multiple regression analysis, we performed linear regression analysis. Values are represented as means ± SDs; a p value < 0.05 was considered statistically significant.

Results
Of 56 eligible patients, two patients failed to meet the inclusion criteria, and 13 patients developed the protocol fault such as data acquisition failure (n = 8), position of blood pressure transducer error (n = 2), and blood pressure measurement failure (bending of the arterial catheter) (n = 3). Therefore, we enrolled 41 patients ( Fig.  1). Patients' age, height, weight, body mass index (BMI) were 67 ± 6 yr., 164.8 ± 6.0 cm, 64.4 ± 9.2 kg, and 23.7 ± 2.9 kg/m 2 , respectively. One patient has BMI over 30 kg/m 2 (30.6 kg/m 2 ). No patients have comorbidities such as chronic obstructive pulmonary disease, heart failure, or uncontrolled hypertension. BIS before anesthesia was 93 ± 6. Propofol was infused at 2.8 ± 0.5 μg/ ml and remifentanil was infused at 0.37 ± 0.10 μg/kg/min. BIS was maintained at 44 ± 6 during the study period. Body temperature before anesthesia was 36.3 ± 0.4°C, and was maintained at 36.5 ± 0.5°C during the study period.
Mean arterial blood pressure before anesthesia was 95 ± 9 mmHg. There were 8 patients who developed hypotension (MAP below 60 mmHg) after the induction of anesthesia. They were treated with ephedrine 5 mg and phenylephrine 0.05 mg. As shown in Fig. 2a, MAP decreased to 67 ± 10 mmHg before the pneumoperitoneum, and increased significantly after the pneumoperitoneum (81 ± 13 mmHg). After the Trendelenburg position, MAP slightly increased by 3 mmHg but the change was not statistically significant. Then it decreased significantly at 30, 45, and 60 min after the Trendelenburg position  Pre-T  T5  T10  T20  T30  T45  T60   100   150   200  PaO 2 Pre Pre-P P-5 P10 Pre-T  T5  T10  T20  T30  T45  T60   35   40 45 † Pre Pre-P P-5 P10 Pre-T  T5  T10  T20  T30  T45  T60   50   60   70   80   90   100   110  MAP  HR   Pre  Pre-P  P-5  P10  Pre-T  T5  T10  T20  T30  T45  T60   97   98   99 100 Fig. 2 a Changes in mean arterial blood pressure (MAP) and heart rate (HR). b Changes in percutaneous (SpO 2 ) and arterial (SaO 2 ) oxygen saturation. c Changes in arterial oxygen tension (PaO 2 ). d Changes in carbon dioxide tension (PaCO 2 ). e Changes in left cerebral regional oxygen saturation (rSO 2 ). f Changes in right cerebral regional oxygen saturation (rSO 2 ). Pre: before the induction of anesthesia, Pre-P: just before the pneumoperitoneum, P-5, compared with that before the Trendelenburg position. Heart rate before anesthesia was 75 ± 12 beats/min. It decreased to 63 ± 10 beats/min before the pneumoperitoneum. Heart rate from 10 min to 60 min after the Trendelenburg position significantly decreased compared with that at before pneumoperitoneum (Fig. 2a). SpO 2 before anesthesia (Pre) was 98.0 ± 1.4%, and increased to 99.0 ± 1.0% throughout the study period (Fig. 2b). SaO 2 did not change in this study (Fig. 2b). PaO 2 before the pneumoperitoneum was 167.4 ± 43.9 mmHg. PaO 2 decreased after the pneumoperitoneum, and increased after the Trendelenburg position. PaCO 2 before the pneumoperitoneum was 40.3 ± 3.5 mmHg (Fig.2c). PaCO 2 slightly but significantly increased after the pneumoperitoneum. PaCO 2 remained high level during the pneumoperitoneum combined with the Trendelenburg position (Fig.2d).
Before general anesthesia, left (Fig.2e) and right (Fig.2f ) rSO 2 were 70.0 ± 6.2% and 70.3 ± 5.6%, respectively. Before the pneumoperitoneum, left and right rSO 2 decreased to 67.9 ± 6.3% and 68.5 ± 7.0%. Ten min after the pneumoperitoneum, left and right rSO 2 significantly increased to 69.6 ± 5.9% and 70.6 ± 7.4%. While patients were in the Trendelenburg position, left and right rSO 2 significantly increased temporarily (5 min after the Trendelenburg position), and decreased to the baseline value afterwards but the change was not statistically significant.

Discussion
We found in the present study that rSO 2 increased after the pneumoperitoneum and further increased temporarily after the steep Trendelenburg position, and decreased afterwards. These changes were along with the alteration of MAP and PaCO 2 . However, the changes did not correlate with the changes of HR, PaO 2 , or SaO 2 .
The cerebral perfusion pressure is regarded as MAP minus central venous pressure (or intracranial pressure) [3]. A previous study reported that central venous pressure increased by 2-5 mmHg during the pneumoperitoneum [8,9]. On the other hand, another previous study reported that MAP did not change after the pneumoperitoneum [9]. Thus, cerebral perfusion pressure should slightly decrease after the pneumoperitoneum. Contrary to the previous study [9], another study reported that pneumoperitoneum with a consequent increase in intracranial pressure produced systemic hypertension. [10]. Our study concurs with the latter study that MAP increased after the pneumoperitoneum. In patients in this study, cerebral autoregulation should be intact. Owing to the cerebral autoregulation, the cerebral blood flow is maintained constantly within a wide range of cerebral perfusion pressure. It is reasonable to assume that cerebral perfusion pressure remained normal level after the pneumoperitoneum. rSO 2 reflects cerebral perfusion [11]. Therefore, we assumed that rSO 2 would be unchanged after the pneumoperitoneum. However, rSO 2 increased after the pneumoperitoneum in this study. In steady state, cerebral blood flow is maintained constant with static cerebral autoregulation [4]. In acute change in blood pressure, cerebral blood flow is compensatory adjusted by dynamic cerebral autoregulation [12,13]. However, there is a time lag between the rise in blood pressure and the activation of dynamic cerebral autoregulation [14]. If blood pressure increased suddenly, cerebral blood flow may increase transiently. As a result, rSO 2 increased.
After the Trendelenburg position combined with CO 2 pneumoperitoneum, rSO 2 increased initially. Some studies reported that central venous pressure increased by 10-16 mmHg during the Trendelenburg position combined with CO 2 pneumoperitoneum [15][16][17]. On the other hand, blood pressure also increased by 10-15 mmHg during the Trendelenburg position combined with CO 2 pneumoperitoneum in the previous studies [2,15,17]. In agreement with those studies, we observed that MAP increased by 16-18 mmHg at 5-10 min after the Trendelenburg position combined with CO 2 pneumoperitoneum compared with that before the pneumoperitoneum. Change in cerebral perfusion pressure immediately after the Trendelenburg position was 16-18 mmHg (MAP change) minus 10-16 mmHg (assumed CVP change) [15][16][17]. The value was 0-8 mmHg. Whereas change in cerebral Change of rSO 2 was correlated with MAP (p < 0.05) and PaCO 2 (p < 0.0001) rSO 2 cerebral regional tissue oxygen saturation, MAP mean arterial blood pressure, HR heart rate, SpO 2 percutaneous arterial oxygen saturation, SaO 2 arterial oxygen saturation, B regression coefficient, SE standard error, β standardized partial regression coefficient perfusion pressure after pneumoperitoneum was 14 mmHg (MAP change) minus 2-5 mmHg (assumed CVP change) [8,9]. The value was 9-12 mmHg. Therefore, it is assumed that the cerebral perfusion pressure did not increase after the Trendelenburg position compared with that at the CO 2 pneumoperitoneum. Cerebral perfusion pressure was not involved in the transient increase in rSO 2 just after the Trendelenburg position combined with CO 2 pneumoperitoneum. Schramm et al. reported that cerebral autoregulation deteriorated with Trendelenburg position combined with pneumoperitoneum [18]. Garrett et al. [19] reported that the cerebral blood flow velocity decreased when the posture was changed from the supine to the seated position. Postural change influences cerebral blood flow. Based on those previous studies, cerebral blood flow increased temporarily when the posture was changed from the supine to the Trendelenburg position. Due to transient increase in cerebral blood flow, rSO 2 initially increased after the Trendelenburg position combined with CO 2 pneumoperitoneum. Cerebral blood flow varies with PaCO 2 . We recently reported that changes in rSO 2 significantly correlated with changes in PaCO 2 . [20]. Abdominally insufflated carbon dioxide is absorbed into systemic circulation and is exhaled with ventilation. We attempted to adjust tidal volume and respiratory rate to maintain PaCO 2 at 40 ± 5 mmHg. However, PaCO 2 increased significantly after the pneumoperitoneum. Thus, rSO 2 increased along with the rise in PaCO 2 . rSO 2 gradually decreased after the pneumoperitoneum combined with the Trendelenburg position. Nevertheless, PaCO 2 remained high level during the pneumoperitoneum combined with the Trendelenburg position. Although there was a correlation between rSO 2 and PaCO 2 , PaCO 2 may be less involved in the change of rSO 2 . On the other hand, MAP was observed highest at 5 min after the Trendelenburg position (T5 in the Fig. 2a) and it decreased thereafter. Head-down position in combination with a pneumoperitoneum impairs cerebral autoregulation over time [18]. It is likely that rSO 2 changes with alterations in mean blood pressure rather than change of PaCO 2 .
Cerebral oxygenation can be monitored by rSO 2 [21]. Cerebral oxygenation may influence the change of rSO 2 in this study. According to the manufacturer, rSO 2 reflects 25% arterial and 75% venous portion of blood. If cerebral oxygen consumption decreased, venous blood oxygen could be increased. As a result, rSO 2 may have increased. However, BIS was unchanged after the induction of anesthesia. There were no factors that involved the decline in cerebral oxygen consumption after the pneumoperitoneum. Furthermore, PaO 2 was significantly decreased after pneumoperitoneum. Therefore, oxygenation status was unlikely to participate in the rise of rSO 2 .
In this study, data before anesthesia were obtained under room air breathing and other data were measured under oxygen inspiration. Therefore, arterial oxygen saturation before anesthesia was significantly lower than that after the induction of anesthesia. However, rSO 2 before anesthesia was significantly higher than that after the induction of anesthesia. Mean arterial blood pressure also declined after the induction of anesthesia. In addition, cerebral metabolic rate decreases after the induction of anesthesia that results in decrease in cerebral blood flow [22]. This phenomenon may indicate that rSO 2 is firmly affected by MAP and cerebral blood flow rather than arterial oxygenation status in this study situation.
Cerebral autoregulation, which is expected to keep cerebral blood flow constant, mainly depends on cerebral perfusion pressure [4]. If cerebral perfusion pressure were too low (below the lower limit of autoregulation), cerebral blood flow might depend on MAP [23]. When the cerebral perfusion pressure was below the lower limit of autoregulation, cerebral blood flow should have been changed directly with the fluctuation of MAP. In this situation, rSO 2 altered along with the change of MAP though decrease in blood pressure below the lower limit of autoregulation was not observed in this study. Nevertheless, individual variation in MAP while anesthetized and the lower limit of autoregulation might make the rSO 2 -MAP relationship different for different individuals.
In this study, PaO 2 decreased after pneumoperitoneum. A previous study suggested that pneumoperitoneum elevated diaphragm that can lead to basilar atelectasis, with resulting right to left shunt formation [24]. Atelectasis caused by pneumoperitoneum may contribute to the decrease of PaO 2 . After the Trendelenburg position, PaO 2 increased. Some studies reported that peak and mean airway pressure increased after the Trendelenburg position [25,26]. Elevated airway pressure may have contributed to reduce atelectasis. As a result, PaO 2 increased. This study has some limitations. First, we did not measure central venous pressure because central venous catheterization is an invasive method and not always necessary for robotic-assisted endoscopic prostatic surgery. In many studies, central venous pressure was increased after pneumoperitoneum [8,9] and the Trendelenburg position combined with pneumoperitoneum. [15][16][17] Therefore, the present study was based on an assumption that central venous pressure increased in this study. Second, cerebral blood flow was not measured. Transcranial Doppler can be utilized to assess the cerebral blood flow. However, due to the protection pads that supported the patient during the surgery, there was no space on the patient's head for the Doppler probe attachment. Third, extracranial contamination affects rSO 2 . Davie et al. reported that the extracranial contamination potentially affected rSO 2 [27]. They also indicated that forehead skin blood made an impact on rSO 2 [27]. Trendelenburg position may cause venous stasis, which could result in the increase of venous portion of blood. Therefore, Trendelenburg position may have affected the relative arterial and venous content of the forehead skin blood. Extracranial contamination might have influenced rSO 2 in this study.

Conclusions
In conclusion, pneumoperitoneum and the Trendelenburg position in robotic-assisted endoscopic prostatic surgery did not aggravate cerebral oxygenation. Changes in rSO2 were associated with the alteration of MAP and PaCO2, but did not correlate with the changes of HR, PaO2, or SaO2, indicating that arterial blood pressure is the critical factor in the cerebral oxygenation.