Impact of Manual Sustained Inflation vs Stepwise PEEP on Pulmonary and Cerebral Outcomes in Carotid Endarterectomy Patients

Introduction

Carotid endarterectomy (CEA) is a surgical method for preventing stroke in patients with symptomatic or asymptomatic carotid artery stenosis in recent decades, and previous studies have shown that it can reduce the risk of ischemic stroke [1,2]. With the improvement of surgical methods and comprehensive monitoring measures, the perioperative safety of carotid endarterectomy has been improved. A recent retrospective study showed that CEA patients receiving general anesthesia had worse outcomes compared to the regional anesthesia group [3]. Various monitoring methods have been used to detect cerebral ischemia events during CEA, including cerebral hemodynamic monitoring and cerebral oxygen metabolism detection. Intraoperative transcranial Doppler (TCD) monitoring provides continuous flow changes and cerebral microemboli in the middle cerebral artery (MCA) on the surgical side, which can reduce the incidence of intraoperative and early postoperative stroke in CEA patients [4]. Near-infrared spectroscopy (NIRS) can measure cerebral oxygenation noninvasively and continuously and has been used to monitor intraoperative cerebral ischemia and reduce the incidence of perioperative complications [5].

There is an increased risk of pneumonia after neck surgery, peripheral vascular surgery, and neurosurgery, possibly due to impairment of normal swallowing and respiratory clearance mechanism after surgery [6]. Although the operation time of CEA is not long, most patients are complicated with high-risk factors of postoperative respiratory complications (PPCs) such as smoking history, hypertension, and cerebrovascular history [7]. Poor oxygenation during the operation will damage cerebral oxygenation and increase the risk of cerebral hypoxia events. Research has shown that lung recruitment maneuver (RM) combined with positive end-expiratory pressure (PEEP) can effectively improve systemic oxygenation in patients undergoing mechanical ventilation [8]. Various methods of recruitment can reduce intraoperative atelectasis and improve arterial oxygenation, including manual sustained inflation (SI) to high airway pressure and slow recruitment with a stepwise increase in PEEP (IP). SI uses a continuous high airway pressure to expand the alveoli. PEEP is the pressure in the alveoli at the end of expiration, which can avoid complete collapse of the alveoli. IP gradually increases PEEP, and a higher PEEP is used to increase the pressure in the alveoli and recruit the collapsed alveoli [9]. An animal study comparing IP and SI found that IP was more effective in improving atelectasis caused by anesthesia [10].

The difference between the 2 RMs has been compared for lung compliance and OI, but the effects on recruitment of collapsed alveoli have been difficult to determine. Electrical impedance tomography (EIT) has been used as a non-invasive imaging modality to measure pulmonary ventilation distribution as EIT images calculated from functional measures [11,12]. Therefore, the difference between the 2 RMs on the pulmonary aeration can be determined using EIT.

A recent randomized controlled study confirmed that IP has a better hemodynamic tolerance than SI [13]. Hemodynamic fluctuations during lung recruitment can affect cerebral blood flow, and secondary changes in oxygen partial pressure and end-tidal carbon dioxide partial pressure can also affect cerebral hemodynamics. Our previous study found that cerebral hemodynamics can be significantly reduced during SI [14]. At present, there have been no studies to explore the effects of these 2 kinds of RM on intraoperative cerebral hemodynamics. We applied TCD and NIRS for continuous intraoperative monitoring to observe the changes of cerebral blood flow and oxygenation during RMs.

The objective of our study was to compare the differences in the impact of SI and IP on lung aeration, cerebral hemodynamics, and hemodynamics in 30 patients during carotid endarterectomy.

Material and Methods

STUDY POPULATION:

The study was conducted from February 2022 to December 2022 at the Xuanwu Hospital, Beijing, China, and was approved by the Clinical Research Ethics Board of Xuanwu Hospital of Capital Medical University in January 2022 (approval No. LYS 2021-233). Written informed consent was obtained from all participants. The trial was registered at the Chinese Clinical Trial Registry (Registration number: ChiCTR2100043603, principal investigator, Lei Zhao, date of registration, February 23, 2021). Thirty-four patients with a grade II–III American Society of Anesthesiologists (ASA) physical status, body mass index (BMI) 18–35 kg/m², aged over 40 years, and scheduled for CEA were enrolled in this study. Exclusion criteria were pulmonary bullae, cardiac insufficiency, hemodynamic instability, previous history of lung surgery, severe chronic obstructive or restrictive pulmonary disease, asthma or thoracic diseases such as mediastinal tumor, or current participation in other interventional studies. Randomization was based on a computer-generated random number list performed by a statistician who was not involved in the trial. Patients were randomly assigned to SI or IP groups in a 1: 1 ratio. The treatment allocated to a patient was disclosed to the investigators only after the patient was enrolled in the study.

MONITORING AND MEASUREMENT PROCEDURES:

After the patients entered the operating room, standard monitoring was performed, including heart rate (HR), mean arterial blood pressure (MAP), arterial oxygen saturation by pulse oximetry (SpO₂), pulse pressure variation (PPV), bispectral index, tidal volume (VT), peak airway pressure (P_peak), end-tidal carbon dioxide, and urine volume. A radial intra-arterial catheter was inserted to monitor arterial blood pressure and obtain blood gas samples. P_peak was measured directly by the Datex-Ohmeda anesthesia machine. Dynamic lung compliance (C_dyn) was defined as VT/(P_peak – PEEP). An EIT belt, consisting of 16 electrodes, was placed at the level of the patient’s 4^th intercostal space, and EIT monitoring (Draeger, Lubeck, Germany) was initiated. Since the patient was supine, we defined 4 regions of interest (ROI 1–4) from ventral to dorsal position to characterize the distribution of regional ventilation and to measure atelectasis. ROI 1 and ROI 2 were located on the ventral side, representing non-dependent regions, while ROI 3 and ROI 4 were located on the dorsal side, reflecting dependent regions. Once connected, the EIT screen will display numbers, numerical values indicated the percentage of each area related to total ventilation. Mean velocity of blood flow in the middle cerebral artery (Vm) was assessed continuously with transcranial Doppler (TCD) by a professional vascular sonographer using a TCD-9PB machine (The Elica Company, Shenzhen, China). Regional cerebral oxygen saturation (rScO₂) was assessed continuously using near-infrared spectroscopy (NIRS) monitoring (CAS Medical Systems, Inc, Branford, CT) with optodes placed on the forehead.

ANESTHETIC MANAGEMENT AND INTERVENTIONS:

Both groups received standard general anesthesia induced using etomidate 0.2–0.5 mg/kg, sufentanil 0.3–0.5 μg/kg, and rocuronium 0.6–0.8 mg/kg. All patients were preoxygenated with pure oxygen, and endotracheal intubation was performed after 5 min of positive pressure ventilation by full face mask. Then, the ventilator (GE Datex-Ohmeda, Madison, USA) was connected to start mechanical ventilation. Patients were mechanically ventilated using the volume control mode with a tidal volume of 6–8 mL/kg, inspired fraction of oxygen (FiO₂) of 0.50, and an inspiration/expiration ratio of 0.5. An initial 5 cmH₂O PEEP was used in both groups. End-tidal carbon dioxide was maintained at 35–45 mmHg by adjusting the respiratory rate. Anesthesia was maintained with propofol 4–6 mg/kg/h, and remifentanil 0.3–0.5 μg/kg/min by continuous intravenous infusion. Depth of anesthesia was monitored according to a bispectral index maintained at 40–60. Fluid administration during anesthesia was guided by PPV maintained at below 13%.

After induction of anesthesia and hemodynamic stabilization, patients were randomly assigned to the SI group or IP group. Recruitment maneuver was not suitable for maintaining adequate cerebral blood flow after carotid artery occlusion, so recruitment was performed once 20 min after tracheal intubation in both groups. In the SI group, the anesthesiologist adjusted the ventilator to manual ventilation mode and adjusted the airway pressure limiting valve to 30 cmH2O. Then, patients in the SI group received sustained manual inflation to a Ppeak of 30 cmH2O for 30 s by squeezing the anesthesia reservoir bag, gradually increasing the peak inspiratory pressure for 3–5 s. In the IP group, alveolar recruitment was performed as follows: ventilation was set to pressure control mode and the driving pressure maintained at 10 cmH2O with an initial PEEP of 5 cmH2O, PEEP was then increased by 5 cmH2O every 30 s until the Ppeak reached 30 cmH2O, held for 30 s, then PEEP was gradually reduced at intervals of 5 cmH2O down to 5 cmH2O, when initial volume control ventilation was resumed. Data were recorded at awakening, baseline (immediately before RM), Ppeak=30 cmH2O, 1 min after RM, 10 min after RM, at the end of surgery, and 30 min after extubation. A schematic diagram of the research protocol is shown in Figure 1.

Predefined criteria for protocol interruption during RM were systolic blood pressure (SBP) <80 mmHg, diastolic blood pressure (DBP) <50 mmHg, reduction in Vm to <50% of baseline or reduction in rScO₂ to <20% of baseline, without return to baseline values within 2 min.

The primary outcome was the difference in ventilation of dependent regions (ROI 3–4) between the 2 groups. Secondary outcomes included changes in Vm, rScO₂, C_dyn, HR, MAP, and arterial blood gas analysis. The incidence of postoperative pulmonary complications within 3 days following surgery were recorded, including persistent hypoxemia, respiratory failure, atelectasis, adult respiratory distress syndrome (ARDS) and pulmonary infection. New cerebral infarction and cerebral hyperperfusion within 3 days after surgery were followed up according to brain magnetic resonance imaging (MRI) and clinical symptoms.

STATISTICAL ANALYSIS:

Because the sample size of previous related studies varies greatly, we conducted a pilot study to determine the sample size. Based on a pilot study performed on 10 patients (5 in each group), the changes in aeration in dependent regions (ROI 3–4) for the SI and IP groups were 16.2±6.5% and 24.2±9.0%, respectively. Calculation therefore suggested that a sample size of 28 patients was required to achieve an α error level of 0.05 and a power of 0.9. Taking into account a dropout rate of 20%, a total of 34 patients (17 in each group) would be needed. Because of the restricted number of patients sent for surgery, the patients in the pilot study were included in the total number of patients.

Continuous variables are presented as mean±SD and categorical variables are reported as the number and percentage. A normal distribution for continuous variables was determined using the Shapiro–Wilk test and homogeneity of variances was accessed by the Levene test. Demographic data and perioperative data were examined with independent sample t tests or Mann-Whitney U tests for continuous variables and Fisher’s exact test for categorical variables. Repeated measurement analysis of variance (RM-ANOVA) was used to analyze intraoperative repeated measurement data to compare the differences within and between groups. Paired sample t tests and Wilcoxon signed-rank tests were used to analyze the data before and after surgery. All statistical calculations and analyses were performed using SPSS Statistics for Windows (version 28.0, IBM Corporation, Armonk, New York, USA). All reported P values were 2-sided and P<0.05 was considered statistically significant.

Results

PERIOPERATIVE VENTILATION AND DYNAMIC LUNG COMPLIANCE:

Compared to baseline, ventilation in dependent regions (ROI 3–4) improved significantly 1 min after RM in both groups (both P<0.001, Figure 3). At Ppeak of 30 cmH2O, IP was significantly more effective in improving ventilation in dependent regions than SI (58.2±8.4% vs 46.0±8.3%, P=0.001). At the end of surgery, dependent ventilation was significantly higher than baseline in the IP group (34.9±9.7% vs 31.6±8.9%, P=0.016), but not in the SI group. Cdyn was significantly improved in both groups 1 min after RM compared to baseline and was significantly higher in the IP group than in the SI group (53.2±13.7 vs 41.1±6.2 mL/cmH2O, P=0.007). In the IP group, Cdyn was significantly higher than baseline at the end of surgery (39.3±5.1 vs 37.0±4.8 mL/cmH2O, P=0.007), but there was no significant difference in the SI group (Figure 3). For postoperative pulmonary ventilation distribution, there was no significant difference between awakening and 30 min after extubation in the IP group. However, dorsal ventilation (ROI 4) 30 min after extubation in the SI group was significantly reduced and did not return to preoperative level (from 9.9±3.8% to 7.7±2.6%, P=0.003) (Figure 4).

INTRAOPERATIVE CHANGES IN CEREBRAL HEMODYNAMICS:

Vm and rScO2 decreased significantly in both groups at Ppeak of 30 cmH2O, with no significant difference between the 2 groups. Vm and rScO2 returned to baseline immediately after RM in the SI group (51.3±18.6 vs 54.1±21.0 cm/s, P=0.189 and 69.2±3.5% vs 70.2±2.3%, P=0.718, respectively), but there were significant differences 1 min after RM compared to baseline in the IP group (42.5±12.6 vs 50.9±18.8 cm/s, P<0.001 and 68.1±3.5% vs 70.6±3.7%, P=0.001, respectively). Vm was lower than baseline in both groups at 10 min after RM, while rScO2 had no difference from baseline (Figure 3). There was no significant difference in Vm and rScO2 at the end of surgery compared with baseline in the 2 groups.

INTRAOPERATIVE CHANGES IN HEMODYNAMICS AND ARTERIAL BLOOD GAS:

In the IP group, there was no significant change in heart rate during RM. However, in the SI group, RM caused a significant reduction in heart rate (from 55.9±6.6 bpm to 52.2±6.9 bpm, P=0.008), which recovered immediately after RM. MAP decreased significantly during recruitment in 2 groups (both P<0.001, Figure 3). MAP gradually recovered after RM, and there was no significant difference from baseline 10 min after RM. There were no significant differences in MAP between the 2 groups. There was no significant difference in PaCO2 between groups at any time point. Arterial pH at the end of surgery was lower than the baseline (7.37±0.04 vs 7.39±0.03, P=0.001) only in the IP group. Lactate was significantly lower at the end of surgery than at baseline in the IP group (1.08±0.38 vs 1.31±0.38 mmol/L, P=0.038) but not in the SI group.

The MRI of the patients on the first day after surgery showed that 2 patients in the IP group had new punctate infarcts, but Fisher’s exact probability test showed that there was no significant difference between the 2 groups in the incidence of postoperative cerebral infarction (P=0.484). There were no cases of cerebral hyperperfusion or respiratory complications in either group.

Discussion

In patients undergoing CEA, IP improved intraoperative lung aeration significantly more than SI, with less effect on heart rate. Furthermore, this advantageous improved lung aeration may persist postoperatively. Both kinds of RM can lead to a significant decrease of cerebral blood flow. The recovery of cerebral blood flow in IP is slow and the impact of lung recruitment on cerebral blood flow is temporary.

Under general anesthesia, with patients in the supine position, the distribution of ventilation is heterogeneous over different regions of the lungs [15]. Non-dependent regions located ventrally are preferentially ventilated, while dependent regions located dorsally are less ventilated, contributing to intraoperative atelectasis, leading to ventilation/blood flow mismatching and intrapulmonary shunt, affecting intraoperative oxygenation and postoperative ventilatory function. RM and intraoperative PEEP are used to alleviate pulmonary atelectasis. Published studies have evaluated the effect of RM on intraoperative PaO2, OI, and lung compliance [8,16]. This study added EIT to determine the ventilation in each region of the lung, and compared the differences between the 2 RM. We found that ventilation in dependent regions (expressed as percentage of total ventilation) was significantly increased during both RMs, indicating recruitment of the collapsed alveoli of the dorsal lung, in turn reducing intraoperative atelectasis. Cdyn was also significantly increased as alveoli in the atelectatic regions were recruited.

An animal study found that the IP group had a higher percentage of alveolar area than the SI group [10]. A study in a pig model of ARDS found that IP significantly improved inhomogeneity of the lungs compared to SI [17]. Our study comparing the 2 methods in anesthetized patients with healthy lungs also confirmed that IP recruited the collapsed alveoli of dependent regions to a greater extent. There was no difference in the duration of surgery between the 2 groups, but the effect of IP was maintained longer, until the end of the operation. IP showed a greater effect with a longer duration than SI, as determined by both EIT and Cdyn.

Some studies have reported that RM does not improve postoperative respiratory outcomes [18]. A randomized controlled study showed that RM could reduce the incidence of early postoperative atelectasis, but could not persist after 24 h after surgery [19]. This study showed that postoperative lung aeration returned to the preoperative awake level 30 min after extubation in the IP group, but dorsal ventilation 30 min after extubation in the SI group was still significantly lower than before surgery, suggesting that IP is more conducive to recovery of respiratory function after general anesthesia. However, lung aeration monitoring did not continue when the patient returned to the ward.

Many patients undergoing CEA commonly also have hypertension, diabetes, coronary heart disease, and other comorbidities. Adverse events such as stroke, myocardial infarction, and death may occur during the perioperative period [20,21]; therefore, it is important to maintain stable hemodynamics and adequate cerebral blood flow. In senile patients with carotid stenosis, dynamic cerebral autoregulation is impaired and stable cerebral blood flow cannot be maintained when blood pressure fluctuates.

TCD is cheap and relatively simple to operate, which make it suitable for real-time monitoring of cerebrovascular blood flow velocity. The principle of TCD is to use low-frequency ultrasound to penetrate the appropriate acoustic window [22], transmit the ultrasound to the cerebral vessels, and reflect it back to the receiver after encountering the flowing red blood cells. Affected by Doppler effect, the reflected ultrasonic frequency can calculate the speed of bleeding flow [23]. TCD can help to identify the real-time changes of cerebral blood flow and microemboli during suegery, and provides a guarantee for safe implementation of carotid endarterectomy [24,25]. Some previous studies have shown that RMs caused a reduction in cerebral perfusion as well as cerebral oxygen saturation and cerebral metabolism [14,26]. In this study, we observed that both SI and IP caused a significant decrease in cerebral blood flow during lung recruitments, but did not lead to intraoperative cerebral ischemic events. The cerebral blood flow of IP group did not return to baseline immediately after RM, possibly because of the slow reduction in airway pressure from its Ppeak.

Near-infrared spectroscopy is another simple technique that can measure the regional cerebral oxygen saturation in the frontal lobe of patients to reflect the state of cerebral oxygenation [27]. It does not require a professional operator or a specific acoustic window. Previous studies reported that near-infrared spectroscopy can detected and prevent cerebral ischemia [28,29]. Our study showed that RMs lead to a decrease in cerebral oxygenation in patients with CEA. Although there was no immediate recovery following IP, there was no difference between the 2 groups 10 min after RM, suggesting there was no significant difference in cerebral oxygenation.

Transient hemodynamic fluctuation is the most common adverse event related to RM. Several studies comparing SI and IP suggest that SI causes a greater decrease in stroke volume and more serious adverse hemodynamic effects at the same Ppeak [13,30]. Our study showed that both RMs caused a marked decrease in MAP, with no significant difference between the 2 groups. In agreement with a previous study [31], we observed that SI caused a small but statistically significant decrease in HR. One possible reason for this decrease in HR is the elevated intrathoracic pressure with enhancement of vagal-mediated reflexes by the transient high tidal volume. In addition, 2 patients in the SI group were excluded because of a SBP lower than 80 mmHg during RM, requiring vasopressor treatment. Although the results of the chi-square test for adverse events were not statistically significant, this suggests that IP may be more suitable for patients with circulatory instability.

RM is an important part of a lung-protective ventilation strategy, which has been shown to improve outcomes in patients undergoing general anesthesia [9]. An optimal RM should reopen collapsed alveoli but avoid hyperinflation of non-atelectatic lung structures, as demonstrated in ARDS patients [32]. Another study showed that pressure of 30 cmH2O did not lead to hyperinflation in healthy people [33]. Therefore, we chose a Ppeak of 30 cmH2O to avoid alveolar hyperinflation and alveolar damage.

Commonly employed measures to clinically monitor the effectiveness of RM include PaO2, OI, SpO2, and lung compliance. Other monitoring methods have included measurement of respiratory resistance, pulmonary ultrasound, and EIT. However, respiratory resistance has the limitation that it cannot actually be determined when adding PEEP or raising airway pressure too high [34]. Although pulmonary ultrasound can be repeated, it is time-consuming and technically demanding for anesthesiologists, unsuitable for continuous monitoring, and inappropriate for detecting hyperinflation [35,36]. EIT can observe RM in real time and serve as a continuous, non-invasive, bedside monitoring technique [37]. Our study mainly used EIT to compare the effects of the 2 RM, while also determining Cdyn.

A limitation of this study is that we only performed RM once before clamping. Since the carotid artery on the operated side is clamped during surgery, the anesthesiologist needs to increase the blood pressure beforehand to ensure adequate cerebral blood flow. Therefore, we chose to compare the 2 groups at various time points after performing one of the RMs. The second limitation is that the sample size of this study was too small to determine the impact of the 2 types of RM on PPCs such as atelectasis and pneumonia. Further, most patients in this study had normal lung function, and the research results cannot be generalized to all patients. It is necessary to expand the sample size for further research. Additionally, we were unable to stratify by BMI to compare the response of normal and obese subjects to different RMs because of the small sample size. Obese patients are more prone to develop intraoperative atelectasis [38], and may respond differently to RM, but a much larger study would be required to detect any difference in respiratory complications.

Conclusions

The finding from this study is that IP has a better effect on improving intraoperative lung ventilation distribution and has less impact on heart rate than SI. As confirmed by EIT, IP is more conducive to the rapid recovery of postoperative respiratory function. Both types of RM can cause a transient decline in cerebral hemodynamics in patients with CEA, but they recover quickly. The recovery of cerebral hemodynamics in IP group was slower.