Effect of optimizing cerebral oxygen saturation on postoperative delirium in older patients undergoing one-lung ventilation for thoracoscopic surgery

Abstract

Objectives

This randomized controlled trial investigated whether the regional cerebral oxygenation saturation (rScO₂)-guided lung-protective ventilation strategy could improve brain oxygen and reduce the incidence of postoperative delirium (POD) in patients older than 65 years.

Methods

This randomized controlled trial enrolled 120 patients undergoing thoracic surgery who received one-lung ventilation (OLV). Patients were randomly assigned to the lung-protective ventilation group (PV group) or rScO₂-oriented lung-protective ventilation group (TPV group). rScO₂ was recorded during the surgery, and the occurrence of POD was assessed.

Results

The incidence of POD 3 days after surgery—the primary outcome—was significantly lower in the TPV group (23.3% versus 8.5%). Meanwhile, the levels of POD-related biological indicators (S100β, neuron-specific enolase, tumor necrosis factor-α) were lower in the TPV group. Considering the secondary outcomes, both groups exhibited a lower oxygenation index after OLV, whereas partial pressure of carbon dioxide and mean arterial pressure were significantly increased in the TPV group. In addition, minimum rScO₂ during surgery and mean rScO₂ were higher in the TPV group than in the PV group.

Conclusion

Continuous intraoperative monitoring of brain tissue oxygenation and active intervention measures guided by cerebral oxygen saturation are critical for improving brain metabolism and reducing the risk of POD.

Introduction

With the development of minimally invasive technology, increasing numbers of techniques of radical thoracoscopic resection of lung cancer have been developed, and lung and brain injuries caused by prolonged intraoperative one-lung ventilation (OLV) in older patients have attracted increasing clinical attention. ¹ Hypoxemia is one of the most common complications during OLV, occurring with an incidence of 1% to 24%. ² In cases of low peripheral oxygenation saturation, brain cell function might be impaired or even damaged because of reduced oxygen transport. ³ A study found that prolonged OLV during thoracic surgery caused cerebral deoxygenation saturation in 70% of patients, and cerebral deoxygenation saturation is believed to be associated with major cerebral dysfunction. ⁴ Postoperative delirium (POD) refers to the acute state of mental disorder that occurs after surgery, mainly manifesting as cognitive dysfunction, consciousness disorder, a lack of concentration, and sleep–wake cycle disorder. ⁶ Cerebral hypoxia is an important risk factor for POD. ⁵ Therefore, avoiding cerebral hypoxia during OLV is crucial for preventing POD.

Regional cerebral oxygen saturation (rScO₂) is a key indicator of cerebral oxygen metabolism. POD can reflect the balance in brain oxygen supply and demand and changes in cerebral blood flow.^7,8 Previous studies revealed a decrease in rScO₂ during thoracic surgery using OLV.^9,10 Therefore, monitoring rScO₂ during thoracic surgery and addressing its deterioration during OLV are beneficial. In a randomized study, Wang et al. applied a measure to maintain rScO₂ at ±20% of the baseline level during OLV and found that this strategy could reduce perioperative inflammatory factor levels and the incidence of POD in elderly patients. ¹¹ In this study, the effects of rScO₂-guided perioperative management combined with lung-protective ventilation (LPV) on the risk of POD was evaluated in older patients during OLV to clarify its mechanism and provide clinical reference.

Materials and methods

Trial design

In this prospective, single blind, randomized, controlled trial of an intervention to improve the incidence of POD in older patients, the CONSORT guideline ¹² was followed. The CONSORT checklist is presented in Supplementary Material 1. The study was conducted in accordance with the Declaration of Helsinki (World Medical Association, 2013). The protocol was approved by the Ethics Committee of Ganyu District People’s Hospital in Lianyungang City ([2022] No. 16, approval date: 2022-12-20). The patients or their immediate families signed the informed consent form. The study was registered in the Chinese Clinical Trial Registry (ChiCTR2300071966).

Participants

Patients were enrolled from June 2023 to January 2024 in Ganyu District People’s Hospital Affiliated to Nanjing Medical University. Older patients with lung cancer undergoing selective radical thoracoscopic surgery under general anesthesia were selected. Patients aged 65 to 80 years with American Society of Anesthesiologists (ASA) class I to III physical function were eligible. The exclusion criteria were as follows: preoperative lung infection or radiotherapy, chemotherapy, and immunotherapy; preoperative hypoproteinemia and hypoxemia; previous history of chronic lung disease or asthma; and previous history of lung surgery, mental disorder, or neurological disease.

Treatment allocation and randomization

After meeting the inclusion criteria, patients were randomized into the lung-protective ventilation (PV group) or rScO₂-oriented lung-protective ventilation group (TPV group) based on an Excel-generated randomization table. To ensure allocation concealment, allocation was hidden in an opaque envelope. Randomization was conducted by an independent investigator who was not involved in the administration of anesthesia. Blinding was only performed for participants in this study. Data were collected and recorded by anesthesiologists.

General anesthesia

None of the patients was administered premedication, and intravenous access was routinely opened. Electrocardiography, blood pressure (noninvasive), SpO₂, and body temperature were monitored after entering the operating room. Intraoperatively, an inflatable heating blanket was used to keep the patient warm, and the liquid was warmed for infusion to maintain the nasopharyngeal temperature at 36.0 to 37.5°C. rScO₂ was monitored continuously using a near-infrared electrode placed in front of the patient’s forehead before anesthesia induction until 10 minutes after removing the tracheal tube (FORE-SIGHT monitor, CAS Medical Systems, Branford, CA, USA), and radial artery puncture catheterization under local anesthesia was used to monitor mean arterial pressure (MAP). Moreover, dexmedetomidine 0.5 µg/kg was given intravenously for 10 minutes. Anesthesia was induced intravenously using midazolam 0.05 mg/kg, sufentanil 0.5 µg/kg, cisatracurium 0.3 mg/kg, and etomidate 0.2 mg/kg. Endobronchial intubation was performed after analgesia, and muscle relaxation was achieved. Then, the patient was connected to the anesthesia machine for pure oxygen inhalation, and volume-controlled ventilation was performed using fiberoptic bronchoscopy to confirm the position of double-lumen endobronchial tube. Central venous pressure (CVP) was measured via right internal jugular vein puncture. The anesthesia maintenance protocol consisted of propofol 2–4 mg kg⁻¹ · hour⁻¹, remifentanil 0.1–0.3 µg kg⁻¹ · minute⁻¹, and cisatracurium 0.1–0.25 mg kg⁻¹ · hour⁻¹ delivered by intravenous pumping, and the anesthetic dosage was adjusted to maintain the bispectral index (BIS) within 40 to 60. At the end of OLV, the lungs were suctioned.

Interventions

The ventilation strategies for the PV group included tidal volume (VT) of 7 mL/kg, positive end-expiratory pressure (PEEP) of 5 cmH₂O, and lung recruitment during two-lung ventilation and VT of 5 mL/kg, PEEP of 5 cmH₂O, and lung recruitment during OLV. The lung recruitment strategy was as follows: maintain the inspiratory pressure at 15, 20, and 25 cmH₂O for three breaths every 45 minutes and keeping PEEP at 5 to 8 cmH₂O. The ventilation mode for the TPV group was the same as that for the PV group, and the following measures were implemented in the TPV group when the rScO₂ decline exceeded 10% of the baseline value: (1) determine whether the electrode position is accurate, (2) adjust the depth of anesthesia, (3) adjust the respiratory parameters and increase end-tidal carbon dioxide tension to slightly higher than 40 mmHg, (4) keep the fluctuation range of MAP to ±20% of baseline value, and (5) when the decline in MAP was >20% of the baseline value, initiate volume therapy according to CVP or the urine output, and use vasoactive drugs such as norepinephrine as appropriate. The baseline MAP was defined as the value measured with the patient in a calm state before anesthesia induction.

Primary outcome

The primary outcome was the incidence of POD. Patients were assessed for the occurrence of delirium on postoperative day 3 using the confusion assessment method (CAM), with CAM scores > 22 indicating the development of POD. S100β, neuron-specific enolase (NSE), and tumor necrosis factor (TNF)-α levels in arterial blood were measured by enzyme-linked immunosorbent assay.

Secondary outcomes

The indicators recorded included rScO₂, BIS, MAP, partial pressure of carbon dioxide (PaCO₂), and partial pressure of oxygen (PaO₂) at five time points: before anesthesia induction (T0), 30 (T1) and 60 minutes (T2) after OLV, before the end of OLV (T3), and 10 minutes after tracheal extubation (T4). The operative time, OLV time, intraoperative blood loss, fluid administration, and vasoactive drug use were recorded. The lowest rScO₂ during surgery (rScO₂min), average rScO₂ (rScO₂mean), and maximum percent decrease in rScO₂ from baseline (rScO_2%max) were also recorded.

Sample size calculation

PASS version 2021 (NSCC, Kaysville, UT, USA) was used to estimate the sample size. According to a previous study ¹¹ and the investigations conducted in our center, the incidence of POD during OLV in elderly patients is 35%. With a power of 0.9 and a significance level of 5%, a minimum sample size of 54 patients was estimated for each group. Considering a dropout rate of 5%, the target enrollment was 60 patients in each group.

Statistical methods

IBM SPSS Statistics version 25.0 (IBM Corp., Armonk, NY, USA) was used to statistically analyze data. Normally distributed data were expressed as the mean ± standard deviation, nonnormally distributed data were expressed as the median (interquartile range), and comparisons between and within groups among different time points were performed by repeated-measures analysis of variance. Countable data were presented as proportion (%), and between-group comparisons were performed using the χ² test or Fisher’s exact test. P < 0.05 was considered statistically significant.

Results

This study included 125 patients. Two patients with a history of long-term diazepam use were excluded, and three patients were excluded before surgery because of severe liver and kidney dysfunction before surgery. Finally, 120 patients were enrolled (Figure 1). The clinical and demographic characteristics and a comparison of basic intraoperative information between the groups are presented in Table 1. No significant differences in sex, age, weight, ASA class, operative time, OLV time, blood loss, and the amount of intraoperative fluid replacement were found between the two groups. The intraoperative use of norepinephrine was significantly more frequent in the TPV group than in the PV group [39 (65%) versus 20 (33.3%), P < 0.05]. There was no difference in ICU occupancy rates between the two groups (5% versus 3.3%, P = 0.648), but the length of hospitalization stay was shorter in the TPV group than in the PV group [9.6 (3.0) versus 12.5 (3.0), P < 0.01].

Figure 1.

Flow diagram of the study. PV group, lung-protective ventilation group; TPV group: rScO₂-oriented lung-protective ventilation group.

Table 1.

Clinical and perioperative characteristics of the study population.

Characteristics	PV (n = 60)	TPV (n = 60)	P
Sex (male/female)	42/18	46/14	0.409
Age (years)	70.23 ± 3.9	71.28 ± 4.2	0.163
Weight (kg)	64.9 ± 6.9	63.6 ± 7.7	0.339
ASA II/III	24/36	21/39	0.572
Duration of surgery (min)	176.15 ± 24.3	175.38 ± 21	0.854
Duration of OLV (min)	128.8 ± 21.1	131.1 ± 20.7	0.545
Blood loss (mL)	150.30	165.50	0.059
Fluid replacement volume (mL)	1800 (500)	1800 (500)	0.561
Norepinephrine usage rate (%)	33.3	65	0.001
Admission to ICU (n, %)	3 (5)	2 (3.3)	0.648
Length of stay	12.5 (3.0)	9.6 (3.0)	<0.001

Data are presented as the mean ± standard deviation, median (interquartile range), or number (percentage %).

PV group, lung-protective ventilation group; TPV group, regional cerebral oxygenation saturation-oriented lung-protective ventilation group.

Primary outcomes

The 3-day POD rates in the PV and TPV groups were 23.3% (n = 14) and 8.5% (n = 5), respectively, and a statistical difference was found between the two groups (χ² = 5.065, P < 0.05). In the comparison of POD-related biological indicators, S100β, NSE, and TNF-α levels were significantly higher after surgery than at T0 in both groups. In the comparison of two groups at the same time point, the levels of these factors were all lower in the TPV group than in the PV group (P < 0.05, Table 2).

Table 2.

Comparison of postoperative delirium-related biological indicators between two groups at different time points.

Biological indicators	PV (n = 60)	TPV (n = 60)	P
S100β (pg/mL)
T0	98.11 ± 19.04	95.49 ± 21.64	0.483
Day 1 after surgery	228.30 ± 55.90	177.61 ± 52.76	<0.001
Day 2 after surgery	312.53 ± 76.86	234.61 ± 56.33	<0.001
Day 3 after surgery	254.61 ± 58.88	205.17 ± 52.66	<0.001
NSE (ng/mL)
T0	15.81 ± 3.63	14.71 ± 3.85	0.109
Day 1 after surgery	25.24 ± 3.85	18.48 ± 5.36	<0.001
Day 2 after surgery	23.88 ± 5.99	18.08 ± 5.17	<0.001
Day 3 after surgery	19.25 ± 5.10	15.23 ± 4.46	<0.001
TNF-α (pg/mL)
T0	310.26 ± 53.73	318.43 ± 52.00	0.399
Day 1 after surgery	644.34 ± 72.58	525.29 ± 78.14	<0.001
Day 2 after surgery	866.42 ± 67.27	609.22 ± 72.12	<0.001
Day 3 after surgery	554.84 ± 68.91	435.91 ± 82.14	<0.001

T0: before anesthesia induction. Data are presented as mean ± standard deviation.

NSE, neuron-specific enolase; TNF-α; tumor necrosis factor-alpha; PV group, lung-protective ventilation group; TPV group, regional cerebral oxygenation saturation-oriented lung-protective ventilation group.

Secondary outcomes

Compared with the findings in the PV group, rScO₂ was increased at T2 to T4 in the TPV group (P < 0.05, Table 3). In addition, rScO₂min and rScO₂mean were increased in the TPV group, whereas rScO₂%max was decreased (all P < 0.05, Table 4). In the TPV group, the oxygenation index (OI) was increased at T3 and T4 (P < 0.05), and PaCO₂ was increased significantly at T2 and T3 (P < 0.05). MAP was higher in the TPV group at T2 to T4 (all P < 0.05). Intragroup comparisons revealed that OI was lower (P < 0.05) at T1 to T4 than at T0, and PaCO₂ group was significantly increased in the TPV (P < 0.05). No significant difference was observed in BIS between the two groups at different time points. The effect sizes are summarized in Table 3.

Table 3.

Comparison of blood gas analysis and anesthesia monitoring parameters between the two groups of patients at different time points

Group	PV	TPV	P
OI
T0	421.95 ± 29.08	432.01 ± 31.90	0.073
T1	302.08 ± 23.15	307.20 ± 25.22	0.249
T2	280.31 ± 22.62	278.16 ± 23.05	0.609
T3	240.43 ± 22.98	272.81 ± 26.4	<0.001
T4	333.00 ± 6.61	356.40 ± 11.65	<0.001
PaCO2 (mmHg)
T0	37.80 ± 2.29	38.33 ± 2.25	0.202
T1	40.10 ± 3.44	45.60 ± 5.40	<0.001
T2	42.46 ± 4.18	50.63 ± 5.48	<0.001
T3	43.95 ± 5.07	51.78 ± 3.52	<0.001
T4	40.10 ± 3.43	42.10 ± 3.44	0.002
MAP (mmHg)
T0	89.78 ± 13.83	89.03 ± 6.95	0.708
T1	81.70 ± 10.91	85.3 ± 15.54	0.143
T2	84.68 ± 14.95	97.90 ± 13.07	<0.001
T3	78.38 ± 7.70	98.26 ± 11.67	<0.001
T4	94.60 ± 14.49	91.78 ± 13.28	0.269
rScO2
T0	71.30 ± 2.51	71.08 ± 2.33	0.626
T1	68.38 ± 4.98	68.76 ± 4.51	0.660
T2	65.78 ± 2.73	68 ± 4.49	0.001
T3	62.88 ± 6.27	67.5 ± 3.72	<0.001
T4	68.6 ± 68.6	70.2 ± 4.04	0.042
BIS
T0	91.63 ± 3.94	92.23 ± 2.52	0.233
T1	49.38 ± 5.16	49.90 ± 6.14	0.616
T2	48.68 ± 5.67	48.13 ± 5.91	0.626
T3	48.01 ± 3.98	48.48 ± 6.03	0.618
T4	89.71 ± 2.64	89.20 ± 2.51	0.274

Time points: before anesthesia induction (T0), 30  (T1) and 60 minutes (T2) after OLV, before the end of OLV (T3), and 10 minutes after tracheal extubation (T4). Data are presented as the mean ± standard deviation.

OI, oxygenation index; PaCO₂, partial pressure of carbon dioxide; MAP, mean arterial pressure; rScO₂, regional cerebral oxygen saturation; BIS, bispectral index; PV group, lung-protective ventilation group; TPV group, regional cerebral oxygenation saturation-oriented lung-protective ventilation group.

Table 4.

Comparison of rScO₂ between the two groups.

Group	PV	TPV	P
Baseline	71.30 ± 2.51	71.08 ± 2.33	0.505
rScO2mean	57.57 ± 2.41	60.43 ± 2.88	0.002
rScO2min	56.11 ± 2.17	62.76 ± 2.54	<0.001
rScO2%max	0.13 ± 0.03	0.09 ± 0.06	<0.001

Data are presented as the mean ± standard deviation.

rScO₂ mean, mean regional cerebral oxygen saturation; rScO₂ min, minimum regional cerebral oxygen saturation; rScO₂% max, maximum percent decrease in rScO₂ from baseline; PV group, lung-protective ventilation group; TPV group, regional cerebral oxygenation saturation-oriented lung-protective ventilation group.

Discussion

To ensure good surgical conditions, thoracoscopic surgery often requires OLV; however, OLV seriously influences the respiration and circulation, causing a series of pathophysiological changes, such as increased airway pressure, ventilation/blood flow ratio imbalance, hypoxic pulmonary vasoconstriction, and inflammatory response activation. ¹³ These changes affect oxygenation. Given that older patients have decreased cardiopulmonary reserve and cerebral circulation self-regulation, OLV is more likely to cause ventilation/blood flow ratio imbalance and increased intrapulmonary shunting, resulting in hypoxemia and decreased rScO₂. ¹⁴ Studies illustrated that POD is significantly related to the decline in brain oxygen metabolism.^7,8,15 Thus, rScO₂ decline is the key cause of POD. In addition, Monique et al. suggested that intraoperative cerebral oxygen desaturation occurred frequently during OLV, and it is associated with a high risk of POD. ¹⁶ POD seriously affects the quality of life of patients, prolongs hospitalization, and increases medical costs and the postoperative mortality rate. ¹⁷ Currently, no effective treatment has been identified; thus, early diagnosis and prevention of POD are important in older patients.

LPV is a treatment for ventilator-related lung injuries, such as barotrauma, volumetric injury, biological injury, and shear injury. It includes standard PEEP, low tidal volume ventilation, and manipulative recruitment strategies. LPV can reduce lung function damage during OLV, improve the efficiency of blood oxygen binding, effectively prevent alveolar collapse and atelectasis, and inhibit pulmonary edema. It is widely used in OLV^18,19 to increase functional residual gas volume and improve oxygenation status. However, PaO₂ and rScO₂ can decrease even after LPV during OLV. In this study, LPV guided by rScO₂ was used. When rScO₂ decreased by >10% versus the baseline value, VT was reduced to <5 mL/kg, or the respiratory rate was adjusted to 10 times/min so that PaCO₂ increased to 45 to 55 mmHg. ²⁰ When the MAP is <80% of the baseline value, the fluid infusion rate is appropriately increased according to CVP and urine output, and norepinephrine is given intravenously to maintain MAP at a normal level. In this study, intraoperative rScO₂min and rScO₂mean were significantly increased in the TPV group, and rScO₂%max was significantly reduced following the interventions. A certain degree of cerebral vasodilation can increase cerebral blood flow and cerebral perfusion, which is meaningful for the improvement of cerebral oxygen saturation, in line with the results of Deschamps et al. ²¹

Recently, a meta-analysis that included 22 randomized controlled trials found that the intraoperative use of electroencephalography and/or rScO₂ monitoring could decrease the risk of perioperative neurocognitive disorders. ²² In this study, we further explored the effect of rScO₂ improvement on POD in older patients. In the TPV group, the 3-day incidence of POD was significantly reduced, and the expression of delirium-related inflammatory factors was also reduced, possibly because POD is a metabolic disorder caused by cerebral ischemia/hypoxia. The improved measures in the TPV group can increase rScO₂ and reduce the risk of cerebral hypoxia. Age and rScO₂%max are independent risk factors that affect the occurrence of POD, and intraoperative rScO₂%max changes have a greater influence on POD. ²³ This result also indicated that intraoperative cerebral oximetry monitoring can reduce the risk of POD in thoracic surgery.

The brain has poor tolerance to hypoxia. Under hypoxic conditions, neutrophils can be induced to release large amounts of inflammatory factors, exacerbating brain tissue hypoxia and insufficient energy supply, leading to brain injury. The astroglial protein S100β is widely used as a parameter of death in several conditions of brain injury and glial activation. ²⁴ Surgical trauma and stress responses can increase the release of TNF-α, increase the activity of the hypothalamic–pituitary–adrenal cortex axis, and promote monoamine circulation, manifested as decreases in dopamine and acetylcholine levels. ²⁵ Serum NSE levels are relatively low under physiological conditions. ²⁶ During the perioperative period, various factors can induce a decrease in cerebral blood flow, brain cell damage, and neuronal damage in patients, leading to a large release of NSE into peripheral blood. The NSE level in peripheral blood can reflect the degree of brain tissue damage in patients. In this study, the expression of the aforementioned markers was lower in the TPV group than in the PV group within 3 days after surgery. This result suggested that intraoperative active intervention measures guided by cerebral oxygen saturation could reduce the levels of S100β, NSE and TNF-α and prevent surgery-induced brain injury and inflammatory responses. This might be related to increases in the uptake of oxygen by neurons and reductions in the inflammatory response caused by hypoxia, as well as improvement in cerebral tissue blood flow perfusion and the induction of protective effects on neurons.

This study had several limitations. In particular, whether permissive hypercapnia has brain-protective effects is controversial. Patients who have altered cerebrovascular structures and functions might have different reactivities of PaCO₂ to rScO₂. Therefore, more studies are needed to verify the effect of hypercapnia on cerebral oxygen. The evaluation time for POD in this study was only 3 days after surgery, and thus, selecting a more comprehensive period for the diagnosis of POD could be meaningful. In addition, although single-center research can ensure the consistency of the results, more studies are needed to validate our conclusions.

Conclusion

In addition to the use of LPV during thoracoscopic OLV surgery in older patients, continuous real-time intraoperative monitoring of brain tissue oxygenation and active intervention measures guided by cerebral oxygen saturation are of clinical significance for improving brain metabolism and reducing the occurrence of POD.

Data availability statement

The data are available from the corresponding author on reasonable request.

Supplementary material

Supplemental material for this article is available online.