Goal-directed fluid therapy and cognitive outcomes in elderly surgical patients with anxiety and depression

Abstract

BACKGROUND

Postoperative cognitive dysfunction (POCD) occurs in 10%-40% of elderly surgical patients, severely impacting quality of life and increasing healthcare costs. Anxiety and depression, present in 30%-50% of perioperative patients, are independent risk factors for POCD through mechanisms including hippocampal damage, impaired cerebrovascular autoregulation, chronic inflammation, and neurotransmitter imbalances. Goal-directed fluid therapy (GDFT) has been proven to reduce POCD incidence by optimizing tissue perfusion. Patients with anxiety and depression, due to compromised cerebrovascular autoregulation and reduced hemodynamic tolerance, may particularly benefit from GDFT. However, research on GDFT’s cognitive protective mechanisms through improving cerebral oxygen supply-demand balance in this population remains limited.

AIM

To investigate the effects of GDFT on perioperative cerebral oxygen supply-demand balance in elderly surgical patients with comorbid anxiety and depression, and to evaluate its protective effects against POCD and its correlation with psychological status improvement.

METHODS

A retrospective cohort study was conducted, enrolling 220 elderly patients who underwent elective major abdominal or orthopedic surgery under general anesthesia at our hospital from April 2023 to March 2025. Patients were divided into the GDFT group and conventional fluid therapy (CFT) group, with 110 patients in each group, based on fluid management strategy. Preoperative psychological status was assessed using the Hospital Anxiety and Depression Scale. Regional cerebral oxygen saturation (rSO₂) was monitored perioperatively. Cognitive function was evaluated using the Mini-Mental State Examination and Montreal Cognitive Assessment on postoperative day (POD) 1, POD3, and POD7. Changes in cerebral oxygen saturation, POCD incidence, and anxiety-depression scores were compared between the two groups, and the correlation between cerebral oxygen supply-demand balance and cognitive function was analyzed.

RESULTS

The GDFT group demonstrated significantly smaller intraoperative rSO₂ decline compared with the CFT group (9.8% ± 4.2% vs 16.9% ± 6.5%, P < 0.001), with a lower incidence of cerebral oxygen supply-demand imbalance (29.1% vs 56.4%, P < 0.001). The incidence of POCD on POD1 was significantly lower in the GDFT group than in the CFT group (20.9% vs 38.2%, P = 0.005). Subgroup analysis revealed that the protective effect of GDFT was more pronounced in patients with comorbid anxiety and depression, with an absolute reduction of 22.1% in POCD incidence. Multivariable regression analysis identified preoperative anxiety and depression [adjusted odds ratio (OR) = 2.68], fluid therapy modality (adjusted OR = 2.12), and magnitude of rSO₂ decline (adjusted OR = 1.43 per 5% increase) as independent risk factors for POCD. Patients with anxiety and depression in the GDFT group showed more significant improvement in postoperative psychological status, and the degree of improvement was negatively correlated with POCD incidence.

CONCLUSION

GDFT was associated with a reduced incidence of POCD in elderly patients by improving perioperative cerebral oxygen supply-demand balance, with this protective effect being particularly prominent in patients with comorbid anxiety and depression. GDFT appeared to facilitate cognitive function recovery and psychological status improvement, highlighting the importance of integrated perioperative psychological-cognitive management.

Key Words: Goal-directed fluid therapy; Postoperative cognitive dysfunction; Cerebral oxygen saturation; Anxiety and depression; Elderly patients

Core Tip: This study demonstrates that goal-directed fluid therapy significantly improves perioperative cerebral oxygen supply-demand balance and reduces postoperative cognitive dysfunction in elderly surgical patients. The protective effect is especially strong in patients with preoperative anxiety and depression, a high-risk population with impaired cerebrovascular regulation and altered oxygen metabolism. By stabilizing hemodynamics and preserving regional cerebral oxygen saturation, goal-directed fluid therapy promotes both cognitive recovery and improvements in psychological status, highlighting its value for precision perioperative brain protection.

INTRODUCTION

Postoperative cognitive dysfunction (POCD) is one of the most common central nervous system complications in elderly patients following surgery, primarily manifesting as impairments in memory, attention, executive function, and other cognitive domains. Epidemiological data indicate that the incidence of POCD in non-cardiac surgical patients ranges from 10% to 40%, increasing significantly with advancing age and surgical complexity[1]. POCD not only severely impacts patients’ postoperative quality of life and social functioning but is also closely associated with prolonged hospital stays, increased healthcare costs, and elevated long-term risk of dementia, making it a major challenge in perioperative medicine[2]. Although the pathogenesis of POCD has not been fully elucidated, current research suggests that multiple factors collectively contribute to its development, including perioperative cerebral oxygen supply-demand imbalance, neuroinflammatory responses, neurotransmitter dysregulation, and blood-brain barrier dysfunction. Among these, cerebral oxygen supply-demand imbalance is considered a critical initiating factor that triggers subsequent pathophysiological changes[3].

Anxiety and depression, as the most prevalent psychological disorders, demonstrate significantly higher prevalence rates among perioperative patients compared to the general population. Systematic reviews and meta-analyses indicate that approximately 30%-50% of elective surgical patients experience varying degrees of preoperative anxiety[4], while 15%-25% present with clinically significant depressive symptoms, with even higher proportions observed in elderly patients and those undergoing major surgery[5]. When facing surgery as a major stressful life event, patients must endure not only the suffering caused by the disease itself but also concerns about surgical failure, postoperative complications, pain, and poor prognosis – psychological burdens that often begin accumulating days to weeks before surgery. More importantly, anxiety and depression are not merely psychological phenomena but rather psychosomatic disorders involving complex neurobiological mechanisms, and their effects on physiological function, particularly brain function, cannot be overlooked.

From a neurobiological perspective, anxiety and depression can impair brain function and increase POCD risk through multiple mechanisms. First, chronic stress-induced overactivation of the hypothalamic-pituitary-adrenal (HPA) axis and sustained elevation of glucocorticoids can directly damage hippocampal neurons, inhibit hippocampal neurogenesis, and lead to impaired memory and learning abilities[6]. The hippocampus, as a core brain region for cognitive function, requires structural and functional integrity to maintain normal cognition. Second, patients with anxiety and depression commonly exhibit heightened sympathetic nervous system excitability and elevated catecholamine levels, which not only cause excessive fluctuations in blood pressure and heart rate (HR) but may also weaken the brain’s buffering capacity against systemic blood pressure changes by affecting cerebrovascular autoregulation, thereby increasing the risk of cerebral hypoperfusion or hyperperfusion[7]. Third, anxiety and depression are closely associated with a systemic chronic low-grade inflammatory state, where elevated levels of inflammatory factors such as interleukin-6, tumor necrosis factor-alpha, and interleukin-1β can damage the blood-brain barrier, activate microglia, induce neuroinflammatory responses, and subsequently impair neuroplasticity and cognitive function[8,9]. Additionally, patients with anxiety and depression often present with neurotransmitter system dysregulation, particularly imbalances in monoamine neurotransmitters such as serotonin, norepinephrine, and dopamine – neurotransmitters that regulate not only mood but also cognitive processes.

More critically, a significant bidirectional association exists between anxiety/depression and POCD, forming a mutually reinforcing vicious cycle. On one hand, prospective cohort studies have confirmed that preoperative anxiety and depression are independent risk factors for POCD, significantly increasing POCD risk[10]. This association is evident not only in elderly patients but also in young and middle-aged surgical patients. Systematic reviews and meta-analyses have demonstrated that preoperative anxiety is significantly associated with increased risk of postoperative delirium[11]. On the other hand, postoperative cognitive decline further exacerbates patients’ anxiety and depressive symptoms. When patients notice their memory deterioration, slowed reactions, and difficulty concentrating, they often experience intense feelings of frustration, helplessness, and worry about the future, thereby increasing their psychological burden. This psychological-cognitive interaction not only affects patients’ recovery process but may also lead to long-term deterioration in quality of life and social functioning. Therefore, in perioperative management, patients with anxiety and depression must be identified as a high-risk population for POCD and given special attention.

Goal-directed fluid therapy (GDFT) represents an important advancement in perioperative fluid management in recent years. Its core concept involves dynamically monitoring individualized hemodynamic parameters [such as stroke volume variation (SVV), pulse pressure variation, and cardiac output (CO)] to adjust fluid infusion rate and volume in real-time, thereby maintaining optimal cardiac preload, CO, and tissue perfusion[12]. Recent randomized controlled trials have demonstrated that GDFT can significantly reduce the incidence of POCD in elderly patients undergoing colorectal cancer surgery[13]. From a theoretical standpoint, patients with anxiety and depression may be the optimal beneficiaries of GDFT intervention. As previously discussed, these patients already have impaired cerebrovascular autoregulation and abnormal oxygen metabolism at baseline, with significantly reduced tolerance to perioperative hemodynamic fluctuations – even moderate decreases in blood pressure or CO may lead to cerebral hypoperfusion and reduced oxygen supply. Furthermore, interventions guided by regional cerebral oxygen saturation (rSO₂) monitoring have been shown to reduce the incidence of POCD in elderly patients undergoing non-cardiac surgery[14].

In summary, systematic research on the interrelationships among GDFT, cerebral oxygen supply-demand balance, anxiety/depression, and POCD remains relatively limited. Particularly in this high-risk population with anxiety and depression, whether GDFT can exert more pronounced cognitive protective effects by improving perioperative cerebral oxygen supply-demand balance, what mechanisms underlie this protective effect, and whether associations exist between brain protection and psychological status improvement – these questions require verification through rigorous clinical research[15]. Therefore, this study aims to investigate the effects of GDFT on perioperative cerebral oxygen supply-demand balance in patients with comorbid anxiety and depression, evaluate its role in postoperative cognitive function recovery, and analyze the correlation between cognitive function improvement and psychological status changes. The ultimate goal is to provide new clinical evidence and theoretical support for precision perioperative brain protection and integrated psychological-cognitive management, thereby improving postoperative outcomes and quality of life in this special population.

MATERIALS AND METHODS

Study design and ethical approval

This study employed a retrospective cohort design, analyzing clinical data from patients who underwent elective surgery under general anesthesia at our hospital between April 2023 to March 2025 through medical record review. The study protocol was approved by the Medical Ethics Committee (No. K2023137). Due to the retrospective nature of the study and anonymization of all data, the ethics committee waived the requirement for informed consent. The study strictly adhered to the ethical principles of the Declaration of Helsinki, and all patient information was kept strictly confidential and used solely for scientific research purposes. Data extraction and analysis were performed by trained researchers to ensure data accuracy and completeness.

Study population and inclusion/exclusion criteria

Inclusion criteria: (1) Elderly patients aged ≥ 60 years; (2) Scheduled for elective major abdominal or orthopedic surgery under general anesthesia (anticipated surgical duration ≥ 2 hours); (3) American Society of Anesthesiologists (ASA) classification I-III; (4) Completion of preoperative anxiety and depression psychological assessment; (5) Perioperative cerebral oxygen saturation monitoring performed; (6) Completion of cognitive function assessment both preoperatively and postoperatively; and (7) Complete medical records, including anesthesia records, fluid management records, monitoring data, and follow-up records.

Exclusion criteria: (1) History of stroke, traumatic brain injury, dementia, or other neuropsychiatric disorders; (2) Severe visual or hearing impairment precluding cognitive function assessment; (3) Preoperative Mini-Mental State Examination (MMSE) score < 24, indicating pre-existing cognitive impairment; (4) Severe cardiac, pulmonary, hepatic, or renal insufficiency (New York Heart Association functional class ≥ III, acute exacerbation of chronic obstructive pulmonary disease, Child-Pugh class C, serum creatinine > 176.8 μmol/L); (5) Intraoperative occurrence of cardiac arrest, major hemorrhage (blood loss > 1000 mL), or blood transfusion > 4 units; (6) Perioperative use of psychoactive medications (such as antipsychotics or benzodiazepines exceeding standard doses); (7) Postoperative intensive care unit admission with mechanical ventilation > 24 hours; and (8) Death or loss to follow-up within 7 days postoperatively. A total of 220 patients were ultimately included in the analysis.

Group allocation and fluid management strategies

Patients were divided into two groups based on perioperative fluid management approach.

GDFT group (n = 110): An individualized fluid management protocol based on FloTrac/Vigileo system or esophageal Doppler monitoring was employed. The specific protocol included: (1) Continuous intraoperative monitoring of SVV or stroke volume (SV); and (2) When SVV > 13% or SV decreased > 10% from baseline, 250 mL of crystalloid or colloid solution was rapidly infused (within 10-15 minutes), followed by assessment of fluid responsiveness. If SVV decreased to < 13% or SV increased > 10% after fluid administration, indicating good fluid responsiveness, maintenance was continued; if unresponsive, vasoactive agents were considered. Fluid infusion rate was comprehensively adjusted based on CO, mean arterial pressure (MAP), urine output, and other parameters to maintain MAP ≥ 65 mmHg and urine output ≥ 0.5 mL/kg/hour.

Conventional fluid therapy group (n = 110): Traditional empirical fluid management was employed, calculating maintenance volume according to the 4-2-1 rule, supplementing third-space losses based on surgical trauma severity (6-8 mL/kg/hour), replacing blood loss with crystalloid or colloid solutions at a 1:2-3 ratio, and adjusting infusion rate based on clinical parameters including blood pressure, HR, and urine output.

Group allocation was based on the time period during which patients underwent surgery: (1) The conventional fluid therapy (CFT) protocol was primarily used from January 2022 to March 2023; and (2) The GDFT protocol was progressively implemented from April 2023 to June 2024 as hemodynamic monitoring equipment became fully available. To ensure baseline comparability between groups, strict matching and statistical control of potential confounding factors affecting outcomes were performed, including age, sex, body mass index (BMI), ASA classification, surgical type, surgical duration, and anesthesia duration.

Anxiety and depression assessment tools

The Hospital Anxiety and Depression Scale (HADS) was used to assess patients’ preoperative anxiety and depression status. HADS is a self-rating scale developed by Zigmond and Snaith in 1983, specifically designed for anxiety and depression screening in general hospital patients. This scale is a public domain tool with no copyright restrictions and is freely available for use. The scale contains 14 items, divided into the anxiety subscale (HADS-A) and depression subscale (HADS-D), each comprising 7 items. Each item is scored on a 4-point scale (0-3), with both subscale scores ranging from 0-21. Scoring criteria are: (1) 0-7 indicates normal; (2) 8-10 indicates possible anxiety or depression; and (3) 11-21 indicates definite anxiety or depression. In this study, HADS-A ≥ 8 was defined as presence of anxiety symptoms, and HADS-D ≥ 8 was defined as presence of depressive symptoms; patients with either score ≥ 8 were included in the anxiety-depression subgroup.

HADS assessment results completed by trained nurses or psychological counselors 1-3 days preoperatively were extracted through medical record review. To ensure assessment reliability, only cases with completely and logically filled assessment forms were included. Based on HADS score, patients in each group were further divided into: Subgroups with and without anxiety-depression for stratified analysis.

Cerebral oxygen saturation monitoring and data collection

All patients underwent continuous bilateral frontal rSO₂ monitoring during the perioperative period using the INVOS 5100C cerebral oximeter (Covidien, United States) or equivalent devices. Monitoring sensors were placed on the bilateral lateral forehead, avoiding the frontal sinus region, with a monitoring depth of approximately 2-3 cm, primarily reflecting frontal cortex oxygenation status. Monitoring commenced before anesthesia induction and continued until patients left the post-anesthesia care unit.

The following cerebral oxygen saturation-related data were extracted from anesthesia records: (1) Baseline rSO₂ value (stable reading in awake state before anesthesia induction); (2) Intraoperative minimum rSO₂ value; (3) The rSO₂ decline magnitude, calculated as: [(baseline value - minimum value)/baseline value] × 100%; (4) Duration of rSO₂ < 80% of baseline; (5) Duration of absolute rSO₂ < 50%; and (6) The rSO₂ value at surgery completion. Significant cerebral oxygen saturation decline was defined as rSO₂ decrease ≥ 20% from baseline or absolute value < 50%. Interventions taken when significant rSO₂ decline occurred were also recorded, including adjustment of anesthesia depth, increasing inspired oxygen concentration, improving hemodynamics, and adjusting ventilation parameters.

Cognitive function assessment methods

The MMSE and Montreal Cognitive Assessment (MoCA) were used to assess patients’ cognitive function. Both scales are internationally recognized free cognitive assessment tools without copyright restrictions. The MMSE has a total score of 30 points, encompassing five dimensions: (1) Orientation; (2) Memory; (3) Attention and calculation; (4) Recall; and (5) Language abilities, with scores < 24 indicating cognitive impairment. The MoCA has a total score of 30 points, assessing eight cognitive domains: (1) Visuospatial and executive function; (2) Naming; (3) Memory; (4) Attention; (5) Language; (6) Abstraction; (7) Delayed recall; and (8) Orientation, with scores < 26 indicating cognitive impairment. For patients with ≤ 12 years of education, 1 point was added to the MoCA score.

Cognitive assessment data at the following time points were extracted through medical record review: (1) Baseline assessment 1-3 days preoperatively; (2) Postoperative day (POD) 1 assessment (18-30 hours postoperatively); (3) POD3 assessment; and (4) POD7 assessment or pre-discharge assessment. All assessments were completed by standardized-trained neurologists or certified research nurses in a quiet environment, with assessment timing avoiding peak analgesic medication periods. To maintain consistency, the same assessor typically performed both preoperative and postoperative evaluations for each patient when logistically feasible. However, due to the retrospective nature of this study, assessors were not formally blinded to treatment group allocation or patients’ preoperative psychological status, as this information was documented in the standard clinical record and accessible during routine care.

Diagnostic criteria for POCD: The Z-score method was employed to calculate each patient’s Z-score for each cognitive test. The formula was: Z = (X_postoperative - X_preoperative)/SD_preoperative, where X_postoperative is the patient’s postoperative score on a specific test, X_preoperative is the same patient’s preoperative score on the corresponding test, and SD_preoperative is the standard deviation of all patients’ preoperative scores on that test. POCD was diagnosed when the composite Z-score (mean of all test Z-scores) was ≥ 1.96 or when two or more individual test Z-scores were ≥ 1.96. POCD incidence rates were calculated separately for POD1, POD3, and POD7.

Hemodynamic parameters and fluid balance data

Perioperative hemodynamic parameters and fluid balance-related data were extracted from anesthesia and nursing records.

Baseline vital signs: (1) Preoperative HR; (2) MAP; and (3) Central venous pressure (if available).

Intraoperative hemodynamics: (1) HR; (2) MAP; and (3) Central venous pressure recorded every 5 minutes, duration and frequency of intraoperative hypotension (MAP < 65 mmHg), vasoactive agent use.

GDFT group-specific parameters: (1) SVV; (2) SV; (3) CO; and (4) Frequency and results of fluid responsiveness assessments.

Fluid balance: (1) Intraoperative crystalloid infusion volume; (2) Colloid infusion volume; (3) Total infusion volume; (4) Blood loss; (5) Urine output; and (6) Total fluid balance.

Transfusion: (1) Red blood cell; (2) Plasma; and (3) Platelet transfusion volumes.

Fluid balance-related indicators were calculated: (1) Standardized infusion volume (mL/kg/hour); (2) Positive fluid balance (input - output); (3) Mean hourly intraoperative infusion rate; and (4) Cumulative 24-hour postoperative fluid balance. These data were used to analyze the mechanisms by which different fluid management strategies affect cerebral oxygen supply-demand balance and cognitive function.

Collection of other clinical data

The following variables potentially affecting study outcomes were collected from the medical record system.

Demographic characteristics: (1) Age; (2) Sex; (3) BMI; (4) Years of education; (5) Marital status; (6) Smoking; and (7) Alcohol history.

Medical history: (1) Hypertension; (2) Diabetes mellitus; (3) Coronary heart disease; (4) Chronic obstructive pulmonary disease; and (5) Cerebrovascular disease, etc.

Preoperative laboratory tests: (1) Hemoglobin; (2) Albumin; (3) Creatinine; (4) Blood glucose; and (5) Electrolytes, etc.

Surgery-related information: (1) Surgical type; (2) Surgical duration; (3) Anesthesia duration; (4) Anesthesia modality (total intravenous anesthesia or combined inhalation anesthesia); (5) Intraoperative temperature management; and (6) Intraoperative complications.

Postoperative recovery: (1) Postoperative pain scores (Visual Analog Scale, VAS); (2) Analgesic use; (3) Time to first ambulation; (4) Time to resumption of oral intake; (5) Length of hospital stay; and (6) Postoperative complications (pulmonary infection, arrhythmia, delirium, etc.).

Postoperative delirium was assessed using the Confusion Assessment Method, with ward nurses performing daily assessments postoperatively; Confusion Assessment Method-positive was defined as presence of delirium. Delirium occurrence within PODs 1-7 was collected.

Dynamic assessment of anxiety and depression status

In addition to preoperative baseline HADS assessment, HADS follow-up assessment data on POD3 and POD7 were extracted from medical records to analyze the impact of fluid management strategies on anxiety and depression status. HADS score changes were calculated (postoperative score - preoperative score), with negative values indicating improvement in anxiety-depression symptoms and positive values indicating symptom worsening. These data were used to explore the correlation between cognitive function recovery and psychological status improvement.

Study outcome measures

Primary outcome measures: (1) Perioperative cerebral oxygen saturation changes, including rSO₂ decline magnitude, incidence of significant rSO₂ decline, and duration of low rSO₂; and (2) POCD incidence on POD1, POD3, and POD7.

Secondary outcome measures: (1) Comparison of POCD incidence between different subgroups (with or without anxiety-depression); (2) Dynamic changes in postoperative cognitive function scores (MMSE and MoCA); (3) Changes in postoperative anxiety-depression score (HADS); (4) Postoperative delirium incidence; (5) Postoperative length of hospital stay; and (6) Postoperative complication rates.

Exploratory analyses: (1) Correlation between cerebral oxygen saturation changes and POCD occurrence; (2) Correlation between cognitive function improvement and anxiety-depression symptom improvement; and (3) Degree of GDFT benefit in the anxiety-depression subgroup.

Statistical analysis

Data analysis was performed using SPSS 26.0 statistical software. Continuous variables were first tested for normality (Shapiro-Wilk test) and homogeneity of variance (Levene’s test). Normally distributed continuous data were expressed as mean ± SD, with between-group comparisons using independent samples t-test; non-normally distributed continuous data were expressed as median (interquartile range), with between-group comparisons using Mann-Whitney U test. Categorical data were expressed as n (%), with between-group comparisons using χ² test or Fisher’s exact test.

Longitudinal changes in cognitive function and anxiety-depression scores were analyzed using repeated measures analysis of variance or mixed linear models, testing time effects, between-group effects, and interaction effects. Covariate adjustment was performed using multivariable logistic regression analysis or analysis of covariance, controlling for confounding factors including age, sex, BMI, ASA classification, surgical type, surgical duration, and preoperative MMSE score.

The relationship between cerebral oxygen saturation changes and POCD occurrence was analyzed using logistic regression, calculating odds ratio (OR) and 95%CI. Correlations between cognitive function improvement and anxiety-depression symptom improvement were analyzed using Pearson or Spearman correlation analysis. Subgroup analyses were performed using stratified logistic regression to test interaction effects. Receiver operating characteristic curves were used to evaluate the diagnostic value of rSO₂ decline magnitude for predicting POCD, calculating the area under the curve.

All tests were two-sided, with P < 0.05 considered statistically significant. Bonferroni correction was applied to adjust significance levels for multiple comparisons.

RESULTS

Baseline patient characteristics and group comparison

A total of 220 patients were ultimately enrolled in this study, with 110 patients each in the GDFT and CFT groups. The two groups showed no statistically significant differences in baseline characteristics including age (68.3 ± 5.7 years vs 67.9 ± 6.1 years), sex (male 54.5% vs 52.7%), BMI, ASA classification, surgical type, and surgical duration (3.2 ± 0.8 hours vs 3.1 ± 0.9 hours) (all P > 0.05). Based on preoperative HADS assessment, 92 patients (41.8%) exhibited anxiety-depression symptoms, with 45 (40.9%) in the GDFT group and 47 (42.7%) in the CFT group, showing no difference between groups (P = 0.791). Patients with anxiety-depression had significantly higher HADS-A and HADS-D scores of 10.2 ± 2.1 and 9.8 ± 2.3, respectively, compared to those without anxiety-depression (P < 0.001). Preoperative MMSE score (27.8 ± 1.6 vs 27.6 ± 1.7, P = 0.354) and MoCA score (25.1 ± 2.3 vs 24.9 ± 2.4, P = 0.509) showed no difference between groups; however, baseline cognitive scores in patients with anxiety-depression were slightly lower than those without (P < 0.05; Table 1).

Table 1 Comparison of baseline characteristics between two groups, n (%)/mean ± SD.

Parameter	Goal-directed fluid therapy group (n = 110)	Conventional fluid therapy group (n = 110)	Statistic	P value
General information
Age (years)	68.3 ± 5.7	67.9 ± 6.1	t = 0.497	0.620
Sex (male/female)	60/50	58/52	χ² = 0.074	0.786
Male	60 (54.5)	58 (52.7)
Body mass index (kg/m²)	24.2 ± 3.1	23.9 ± 3.3	t = 0.684	0.495
Years of education	9.8 ± 3.2	10.1 ± 3.5	t = 0.648	0.518
American Society of Anesthesiologists classification			χ² = 0.385	0.825
Class I	18 (16.4)	20 (18.2)
Class II	67 (60.9)	64 (58.2)
Class III	25 (22.7)	26 (23.6)
Medical history
Hypertension	52 (47.3)	55 (50.0)	χ² = 0.162	0.687
Diabetes mellitus	28 (25.5)	31 (28.2)	χ² = 0.215	0.643
Coronary heart disease	15 (13.6)	18 (16.4)	χ² = 0.347	0.556
Surgical information
Surgical type			χ² = 0.165	0.685
Abdominal surgery	66 (60.0)	69 (62.7)
Orthopedic surgery	44 (40.0)	41 (37.3)
Surgical duration (hours)	3.2 ± 0.8	3.1 ± 0.9	t = 0.856	0.393
Anesthesia duration (hours)	3.8 ± 0.9	3.7 ± 1.0	t = 0.767	0.444
Estimated blood loss (mL)	385 ± 125	392 ± 138	t = 0.388	0.698
Anxiety-depression assessment
Anxiety-depression	45 (40.9)	47 (42.7)	χ² = 0.071	0.791
HADS-A scorec
Anxiety-depression subgroup	10.3 ± 2.0	10.1 ± 2.2	t = 0.448	0.655
Non-anxiety-depression subgroup	4.2 ± 1.7	4.4 ± 1.9	t = 0.588	0.558
HADS-D scorec
Anxiety-depression subgroup	9.9 ± 2.2	9.7 ± 2.4	t = 0.409	0.683
Non-anxiety-depression subgroup	3.8 ± 1.5	4.0 ± 1.7	t = 0.665	0.507
Preoperative cognitive assessment
Mini-Mental State Examination score	27.8 ± 1.6	27.6 ± 1.7	t = 0.929	0.354
Anxiety-depression subgroup	27.1 ± 1.8	27.3 ± 1.9	t = 0.508	0.613
Non-anxiety-depression subgroup	28.2 ± 1.3	28.0 ± 1.5	t = 0.768	0.444
Montreal Cognitive Assessment score	25.1 ± 2.3	24.9 ± 2.4	t = 0.660	0.509
Anxiety-depression subgroup	24.2 ± 2.6	24.4 ± 2.4	t = 0.375	0.709
Non-anxiety-depression subgroup	25.6 ± 2.0	25.4 ± 2.2	t = 0.510	0.611
Preoperative laboratory tests
Hemoglobin (g/L)	128.5 ± 16.3	130.2 ± 17.8	t = 0.740	0.460
Albumin (g/L)	38.6 ± 4.2	39.1 ± 4.5	t = 0.858	0.392
Serum creatinine (μmol/L)	78.3 ± 18.5	76.9 ± 19.2	t = 0.549	0.584
Fasting glucose (mmol/L)	5.8 ± 1.2	5.9 ± 1.3	t = 0.596	0.552

^cP < 0.001.

Comparison of Hospital Anxiety and Depression Scale score between anxiety-depression and non-anxiety-depression subgroups, P < 0.001; baseline Mini-Mental State Examination and Montreal Cognitive Assessment score were lower in patients with anxiety-depression compared to those without, P < 0.05. HADS-A: Hospital Anxiety and Depression Scale-anxiety subscale; HADS-D: Hospital Anxiety and Depression Scale-depression subscale.

Comparison of perioperative hemodynamics and fluid management

The GDFT group demonstrated more stable hemodynamic management, with significantly fewer episodes of intraoperative hypotension (MAP < 65 mmHg) (2.1 ± 1.3 episodes vs 3.8 ± 2.1 episodes, P < 0.001) and shorter duration (8.3 ± 5.2 minutes vs 15.7 ± 8.6 minutes, P < 0.001) compared to the CFT group. Mean CO in the GDFT group (4.8 ± 0.7 L/minute) was higher than in the CFT group (4.3 ± 0.8 L/minute, P < 0.001), with lower vasoactive agent utilization (43.6% vs 62.7%, P = 0.005). Regarding fluid management, total intraoperative fluid volume in the GDFT group (2180 ± 420 mL) was significantly less than in the CFT group (2650 ± 580 mL, P < 0.001), with standardized fluid volume of 6.8 ± 1.2 mL/kg/hour vs 8.3 ± 1.8 mL/kg/hour (P < 0.001). Despite lower fluid volumes, urine output in the GDFT group was comparable to the CFT group; however, positive fluid balance was significantly reduced (980 ± 310 mL vs 1450 ± 480 mL, P < 0.001), and cumulative 24-hour postoperative fluid balance was also lower (1650 ± 520 mL vs 2280 ± 690 mL, P < 0.001; Table 2).

Table 2 Comparison of perioperative hemodynamics and fluid management between two groups, n (%)/mean ± SD.

Parameter	Goal-directed fluid therapy group (n = 110)	Conventional fluid therapy group (n = 110)	Statistic	P value
Hemodynamic parameters
Preoperative MAP (mmHg)	92.3 ± 10.5	91.8 ± 11.2	t = 0.344	0.731
Preoperative HR (beats/minute)	72.5 ± 9.8	73.2 ± 10.3	t = 0.513	0.608
Intraoperative mean MAP (mmHg)	78.6 ± 6.8	76.2 ± 8.5	t = 2.296	0.023
Intraoperative mean HR (beats/minute)	68.3 ± 8.5	70.8 ± 9.7	t = 2.025	0.044
Intraoperative mean cardiac output (L/minute)	4.8 ± 0.7	4.3 ± 0.8	t = 4.971	< 0.001
Hypotensive events
Number of hypotensive episodes	2.1 ± 1.3	3.8 ± 2.1	t = 7.157	< 0.001
Duration of hypotension (minutes)	8.3 ± 5.2	15.7 ± 8.6	t = 7.569	< 0.001
Incidence of hypotension	56 (50.9)	78 (70.9)	χ² = 9.091	0.003
Vasoactive agent use
Vasoactive agent use	48 (43.6)	69 (62.7)	χ² = 7.950	0.005
Phenylephrine use	38 (34.5)	58 (52.7)	χ² = 7.372	0.007
Ephedrine use	22 (20.0)	35 (31.8)	χ² = 3.951	0.047
Dopamine use	5 (4.5)	8 (7.3)	χ² = 0.729	0.393
Fluid management
Total intraoperative fluid (mL)	2180 ± 420	2650 ± 580	t = 6.849	< 0.001
Standardized fluid volume (mL/kg/hour)	6.8 ± 1.2	8.3 ± 1.8	t = 7.302	< 0.001
Crystalloid infusion (mL)	1560 ± 340	2050 ± 480	t = 8.894	< 0.001
Colloid infusion (mL)	620 ± 160	600 ± 180	t = 0.873	0.383
Colloid proportion (%)	28.3 ± 8.5	22.6 ± 9.2	t = 4.815	< 0.001
Hydroxyethyl starch use	76 (69.1)	71 (64.5)	χ² = 0.502	0.478
Albumin use	15 (13.6)	12 (10.9)	χ² = 0.382	0.537
Fluid input/output
Intraoperative blood loss (mL)	420 ± 185	445 ± 205	t = 0.946	0.345
Intraoperative urine output (mL)	785 ± 235	760 ± 258	t = 0.755	0.451
Standardized urine output (mL/kg/hour)	1.2 ± 0.3	1.1 ± 0.4	t = 1.326	0.186
Intraoperative positive fluid balance (mL)	980 ± 310	1450 ± 480	t = 8.600	< 0.001
24 hours cumulative fluid input (mL)	3420 ± 680	4180 ± 890	t = 7.185	< 0.001
24 hours cumulative urine output (mL)	1770 ± 420	1900 ± 510	t = 2.033	0.043
24 hours fluid balance (mL)	1650 ± 520	2280 ± 690	t = 7.460	< 0.001
Transfusion
Patients receiving transfusion	18 (16.4)	23 (20.9)	χ² = 0.745	0.388
Red blood cell transfusion (U)	1.8 ± 0.9	2.1 ± 1.1	t = 1.124	0.264
Plasma transfusion (mL)	285 ± 125	310 ± 145	t = 0.684	0.495

Hypotension defined as mean arterial pressure < 65 mmHg. HR: Heart rate; MAP: Mean arterial pressure.

Perioperative changes in cerebral oxygen saturation

There was no difference in baseline rSO₂ values between the two groups [GDFT group (68.2% ± 6.3%) vs CFT group (67.8% ± 6.7%), P = 0.631]. The GDFT group had significantly higher minimum intraoperative rSO₂ (61.5% ± 5.8%) compared to the CFT group (56.3% ± 7.2%, P < 0.001), with significantly smaller rSO₂ decline magnitude (9.8% ± 4.2% vs 16.9% ± 6.5%, P < 0.001). Regarding cerebral oxygen supply-demand imbalance incidence, the GDFT group had an rSO₂ decline ≥ 20% rate of 12.7%, significantly lower than 34.5% in the CFT group (P < 0.001); rSO₂ < 50% rates were 5.5% vs 18.2% (P = 0.003). Duration of low rSO₂ was 12.5 ± 6.8 minutes in the GDFT group vs 23.7 ± 11.3 minutes in the CFT group (P < 0.001). Subgroup analysis of patients with anxiety-depression showed that in the CFT group, patients with comorbid anxiety-depression had significantly greater rSO₂ decline magnitude (19.3% ± 7.1%) than those without (15.2% ± 5.6%, P = 0.002), with higher cerebral oxygen supply-demand imbalance rates (46.8% vs 25.4%, P = 0.023). In contrast, the difference between subgroups in the GDFT group was reduced and not statistically significant (P = 0.074), suggesting that GDFT provides more pronounced cerebral oxygen protection in patients with anxiety-depression (Table 3).

Table 3 Comparison of perioperative cerebral oxygen saturation changes between two groups, n (%)/mean ± SD.

Parameter	Goal-directed fluid therapy group (n = 110)	Conventional fluid therapy group (n = 110)	Statistic	P value
Overall rSO2 changes
Baseline rSO2 (%)	68.2 ± 6.3	67.8 ± 6.7	t = 0.481	0.631
Minimum intraoperative rSO2 (%)	61.5 ± 5.8	56.3 ± 7.2	t = 5.845	< 0.001
The rSO2 decline magnitude (%)	9.8 ± 4.2	16.9 ± 6.5	t = 9.550	< 0.001
The rSO2 at surgery completion (%)	66.8 ± 5.9	63.2 ± 7.1	t = 4.054	< 0.001
The rSO2 recovery rate (%)	97.9 ± 4.8	93.2 ± 7.3	t = 5.779	< 0.001
Cerebral oxygen supply-demand imbalance events
The rSO2 decline ≥ 20%	14 (12.7)	38 (34.5)	χ² = 15.886	< 0.001
Absolute rSO2 < 50%	6 (5.5)	20 (18.2)	χ² = 8.916	0.003
The rSO2 < 80% baseline	28 (25.5)	56 (50.9)	χ² = 15.119	< 0.001
Patients with cerebral oxygen imbalance	32 (29.1)	62 (56.4)	χ² = 16.482	< 0.001
Duration of low rSO2
Low rSO2 duration (minutes)	12.5 ± 6.8	23.7 ± 11.3	t = 5.894	< 0.001
Duration > 20 minutes	5 (15.6)	32 (51.6)	χ² = 12.163	< 0.001
Duration 10-20 minutes	18 (56.3)	22 (35.5)	χ² = 3.545	0.060
Duration < 10 minutes	9 (28.1)	8 (12.9)	χ² = 2.902	0.088
Anxiety-depression subgroup analysis
Patients with anxiety-depression (n = 92)	n = 45	n = 47
Baseline rSO2 (%)	67.5 ± 6.8	67.2 ± 7.1	t = 0.211	0.833
Minimum intraoperative rSO2 (%)	60.9 ± 6.2	54.2 ± 7.8	t = 4.541	< 0.001
The rSO2 decline magnitude (%)	10.8 ± 4.6	19.3 ± 7.1	t = 6.791	< 0.001
The rSO2 decline ≥ 20%	7 (15.6)	22 (46.8)	χ² = 10.406	0.001
Cerebral oxygen imbalance rate	16 (35.6)	32 (68.1)	χ² = 9.784	0.002
Patients without anxiety-depression (n = 128)	n = 65	n = 63
Baseline rSO2 (%)	68.7 ± 5.9	68.2 ± 6.4	t = 0.459	0.647
Minimum intraoperative rSO2 (%)	61.9 ± 5.5	57.8 ± 6.5	t = 3.869	< 0.001
The rSO2 decline magnitude (%)	9.1 ± 3.7	15.2 ± 5.6	t = 7.420	< 0.001
The rSO2 decline ≥ 20%	7 (10.8)	16 (25.4)	χ² = 4.296	0.038
Cerebral oxygen imbalance rate	16 (24.6)	30 (47.6)	χ² = 7.173	0.007

Regional cerebral oxygen saturation decline magnitude = [(baseline value - minimum value)/baseline value] × 100%. Regional cerebral oxygen saturation recovery rate = (value at surgery completion/baseline value) × 100%; statistics include only patients with cerebral oxygen supply-demand imbalance. rSO₂: Regional cerebral oxygen saturation.

Incidence of POCD

The GDFT group had a POCD incidence of 20.9% on POD1, significantly lower than 38.2% in the CFT group (P = 0.005). On POD3, POCD rates were 14.5% vs 28.2% (P = 0.011); on POD7, rates were 8.2% vs 17.3% (P = 0.035). Regarding cognitive function scores, on POD1, the GDFT group had significantly higher MMSE score (26.3 ± 2.1) and MoCA score (23.5 ± 2.8) compared to the CFT group (25.1 ± 2.6 and 21.8 ± 3.2, both P = 0.001). By POD7, cognitive scores in the GDFT group had nearly returned to baseline [MMSE (27.6 ± 1.7), MoCA (24.9 ± 2.4)], while the CFT group remained slightly below baseline. Repeated measures analysis of variance demonstrated statistically significant between-group effects (F = 12.35, P = 0.001), time effects (F = 78.92, P < 0.001), and interaction effects (F = 5.68, P = 0.004; Figure 1).

Figure 1 Comparison of postoperative cognitive function between goal-directed fluid therapy and conventional fluid therapy groups. A: Incidence of postoperative cognitive dysfunction at postoperative day (POD) 1, POD3, and POD7. The goal-directed fluid therapy (GDFT) group showed significantly lower postoperative cognitive dysfunction incidence compared to the conventional fluid therapy group at all time points (POD1: 20.9% vs 38.2%, P = 0.005; POD3: 14.5% vs 28.2%, P = 0.011; POD7: 8.2% vs 17.3%, P = 0.035); B: Dynamic changes in Mini-Mental State Examination score. Both groups showed decreased scores on POD1, with the GDFT group demonstrating less decline and faster recovery. Repeated measures analysis of covariance revealed significant group effect (F = 12.35, P = 0.001), time effect (F = 78.92, P < 0.001), and interaction effect (F = 5.68, P = 0.004); C: Dynamic changes in Montreal Cognitive Assessment scores. Similar patterns were observed, with the GDFT group showing better preservation of cognitive function on POD1 (23.5 vs 21.8, P = 0.001) and more complete recovery by POD7. Asterisks indicate statistically significant differences between groups (^aP < 0.05). CFT: Conventional fluid therapy; GDFT: Goal-directed fluid therapy; MMSE: Mini-Mental State Examination; MoCA: Montreal Cognitive Assessment; POCD: Postoperative cognitive dysfunction; POD: Postoperative day.

Risk factor analysis for POCD

To identify independent predictors of POCD on POD1, univariate and multivariate logistic regression analyses were performed on potential risk factors. Univariate analysis revealed that age, surgical duration, preoperative anxiety-depression, fluid therapy modality, rSO₂ decline magnitude, and intraoperative hypotension duration were significantly associated with POCD occurrence. Specifically, each 5-year increase in age increased POCD risk by 32% (OR = 1.32, 95%CI: 1.08-1.62, P = 0.007); preoperative anxiety-depression nearly doubled POCD risk (OR = 2.89, 95%CI: 1.67-5.01, P < 0.001); patients receiving CFT had 2.35 times the POCD risk compared to those receiving GDFT (OR = 2.35, 95%CI: 1.36-4.06, P = 0.002). Notably, perioperative cerebral oxygen supply-demand balance indicators showed that each 5% increase in rSO₂ decline magnitude increased POCD risk by 52% (OR = 1.52, 95%CI: 1.28-1.81, P < 0.001); each 10-minute increase in intraoperative hypotension duration increased POCD risk by 38% (OR = 1.38, 95%CI: 1.12-1.70, P = 0.003).

In the multivariate logistic regression model adjusted for confounding factors, preoperative anxiety-depression remained the strongest independent risk factor for POCD, with an adjusted OR of 2.68 (95%CI: 1.52-4.73, P = 0.001), indicating that patients with comorbid anxiety-depression had significantly elevated POCD risk even after controlling for other factors. Fluid therapy modality, as a modifiable factor, maintained significant independent predictive value in the multivariate model, with CFT increasing POCD risk 2.12-fold compared to GDFT (adjusted OR = 2.12, 95%CI: 1.20-3.75, P = 0.010). Among perioperative cerebral oxygen supply-demand balance-related indicators, rSO₂ decline magnitude demonstrated the most prominent independent predictive value, with each 5% increase still increasing POCD risk by 43% after adjustment (adjusted OR = 1.43, 95%CI: 1.18-1.74, P < 0.001). Additionally, intraoperative hypotension duration maintained marginal significance in the multivariate model (adjusted OR = 1.26, 95%CI: 1.01-1.57, P = 0.042), highlighting the importance of hemodynamic stability for cognitive function protection (Table 4).

Table 4 Multivariate logistic regression analysis: Risk factors for postoperative cognitive dysfunction on postoperative day 1.

Variable	Univariate analysis		Multivariate analysis
	OR (95%CI)	P value	Adjusted OR (95%CI)	P value
Age (per 5-year increase)	1.32 (1.08-1.62)	0.007	1.28 (1.04-1.58)	0.021
Male sex	0.87 (0.51-1.48)	0.609	-	-
American Society of Anesthesiologists class ≥ III	1.68 (0.88-3.21)	0.116	1.45 (0.73-2.88)	0.289
Surgical duration (per 1 hour increase)	1.42 (1.06-1.91)	0.019	1.35 (0.99-1.84)	0.056
Preoperative anxiety-depression	2.89 (1.67-5.01)	< 0.001	2.68 (1.52-4.73)	0.001
Fluid therapy modality (conventional fluid therapy vs goal-directed fluid therapy)	2.35 (1.36-4.06)	0.002	2.12 (1.20-3.75)	0.010
Regional cerebral oxygen saturation decline magnitude (per 5% increase)	1.52 (1.28-1.81)	< 0.001	1.43 (1.18-1.74)	< 0.001
Intraoperative hypotension duration (per 10 minutes increase)	1.38 (1.12-1.70)	0.003	1.26 (1.01-1.57)	0.042

OR: Odds ratio.

Relationship between rSO₂ decline magnitude and POCD across subgroups

The relationship between rSO₂ decline magnitude and POCD risk was consistently positive across all examined subgroups, with ORs ranging from 1.31 to 1.58 for each 5% increase in rSO₂ decline (Table 5). Specifically, for patients aged < 70 years vs ≥ 70 years, the ORs were 1.38 and 1.52, respectively; for males vs females, 1.41 and 1.46; for GDFT vs CFT groups, 1.31 and 1.48; for ASA I-II vs III, 1.40 and 1.52. However, the strength of this association varied significantly by preoperative anxiety-depression status. For non-anxiety-depression vs anxiety-depression patients, the ORs were 1.35 and 1.58, with this difference reaching statistical significance (interaction P = 0.042). This significant interaction indicates that patients with comorbid anxiety-depression showed stronger susceptibility to cognitive impairment from cerebral oxygen desaturation compared to those without psychological comorbidities. In contrast, interaction terms for other subgroup comparisons (age, sex, fluid therapy type, ASA classification) did not reach statistical significance.

Table 5 Relationship between regional cerebral oxygen saturation decline magnitude and postoperative cognitive dysfunction across subgroups.

Subgroup	Adjusted odds ratio (95%CI)	P value	Interaction P value
Age stratification			0.156
< 70 years (n = 118)	1.38 (1.08-1.76)	0.009
≥ 70 years (n = 102)	1.52 (1.20-1.93)	< 0.001
Sex stratification			0.524
Male (n = 118)	1.41 (1.10-1.80)	0.006
Female (n = 102)	1.46 (1.14-1.87)	0.003
Anxiety-depression stratification			0.042
Non-anxiety-depression (n = 128)	1.35 (1.05-1.73)	0.022
Anxiety-depression (n = 92)	1.58 (1.25-2.00)	< 0.001
Fluid management strategy stratification			0.187
Goal-directed fluid therapy group (n = 110)	1.31 (1.02-1.68)	0.032
Conventional fluid therapy group (n = 110)	1.48 (1.19-1.84)	0.001
ASA classification stratification			0.428
ASA I and II (n = 169)	1.40 (1.13-1.74)	0.002
ASA III (n = 51)	1.52 (1.10-2.10)	0.012

ASA: American Society of Anesthesiologists.

Association between changes in anxiety-depression status and cognitive function recovery

Significant differences in postoperative psychological status improvement were observed between groups among patients with anxiety-depression. On POD3, the GDFT group showed HADS-A score reductions of approximately 2.5 points and HADS-D score reductions of approximately 2.7 points, both significantly greater than the 1.2-point reduction in the CFT group (both P < 0.01). By POD7, 62.2% of patients in the GDFT group had HADS score return to normal, significantly better than 40.4% in the CFT group (P < 0.05).

Subgroup analysis revealed that HADS improvement in the GDFT group was closely associated with POCD incidence. Patients with improvement > 3 points had a POCD incidence of only 4.8%, while those with improvement ≤ 3 points had a rate of 16.7% (P < 0.05). These results suggest that GDFT can more effectively improve perioperative anxiety-depression status, and adequate improvement in emotional state may help reduce the risk of POCD (Figure 2).

Figure 2 Effects of goal-directed fluid therapy on anxiety-depression status and cognitive improvement in elderly surgical patients. Bar chart comparing goal-directed fluid therapy (GDFT) group (blue) and conventional fluid therapy group (yellow) across five outcome measures: (1) Hospital Anxiety and Depression Scale (HADS)-anxiety subscale score reduction at postoperative day (POD) 3; (2) HADS-depression subscale score reduction at POD3; (3) Normalization rate at POD7; and (4) Postoperative cognitive dysfunction improvement rates. GDFT group showed significantly greater reductions in both HADS-anxiety subscale (2.5 vs 1.2, P < 0.01) and HADS-depression subscale (2.7 vs 1.2, P < 0.01) scores. The GDFT group also demonstrated higher anxiety-depression normalization rate (62.2% vs 40.4%, P < 0.05) and lower postoperative cognitive dysfunction incidence with improvement > 3 points (4.8% vs 16.7%, P < 0.05). Error bars represent standard deviation. ^aP < 0.05, ^bP < 0.01. CFT: Conventional fluid therapy; GDFT: Goal-directed fluid therapy; HADS-A: Hospital Anxiety and Depression Scale-anxiety subscale; HADS-D: Hospital Anxiety and Depression Scale-depression subscale; POD: Postoperative day; POCD: Postoperative cognitive dysfunction.

DISCUSSION

Before interpreting our findings, it is essential to acknowledge the fundamental limitations inherent in our study design. As a retrospective, non-randomized cohort study, our work cannot establish causality between GDFT and cognitive outcomes. The temporal allocation of patients to treatment groups (CFT during an earlier period, followed by GDFT as monitoring technology became available) introduces potential confounding by secular trends in surgical technique, anesthetic practice, nursing protocols, and overall perioperative management that evolved during the study period. While we statistically controlled for numerous measured variables, unmeasured temporal confounders likely persist and may partially explain the observed differences between groups. Additionally, the lack of formal blinding of outcome assessors to treatment allocation represents a potential source of ascertainment bias. These limitations necessitate cautious interpretation: Our findings should be understood as hypothesis-generating evidence that warrants confirmation through prospective randomized controlled trials, rather than as definitive proof of GDFT’s superiority. With these caveats clearly stated, we now discuss our observations and their potential mechanistic underpinnings.

This retrospective cohort analysis found associations suggesting that GDFT may improve perioperative cerebral oxygen supply-demand balance and be associated with reduced incidence of POCD, with this association being particularly pronounced in patients with comorbid anxiety-depression. The study found that GDFT was associated with a reduction in POCD incidence from 38.2% to 20.9%, with an absolute risk reduction (ARR) of up to 22.1% in the anxiety-depression subgroup. More importantly, GDFT appeared to not only be associated with better cognitive function but also correlate with improvement in anxiety-depression symptoms, with significant positive correlation between the two, highlighting the importance of integrated psychological-cognitive management. These findings provide preliminary clinical evidence for precision perioperative brain protection strategies, particularly suggesting potential directions for individualized management of this high-risk population with anxiety-depression.

This study confirmed that preoperative anxiety-depression is an independent risk factor for POCD, a finding highly consistent with existing literature[16]. Anxiety-depression is not merely a psychological phenomenon but rather a psychosomatic disorder involving dysregulation of multiple neurobiological systems. From a pathophysiological perspective, anxiety-depression states increase POCD risk through at least four major pathways.

First, overactivation of the HPA axis and sustained hypercortisolemia represent core pathological changes in patients with anxiety-depression[17]. Chronic stress leads to disruption of cortisol circadian rhythm and impaired negative feedback regulation, with sustained high cortisol levels exerting direct neurotoxic effects on the hippocampus[18]. As one of the brain regions with the highest density of glucocorticoid receptors, the hippocampus is particularly vulnerable to damage. Long-term exposure to high cortisol environments leads to dendritic atrophy of CA3 pyramidal neurons, reduced synaptic density, inhibited neurogenesis, and ultimately hippocampal volume reduction and functional decline[19]. The hippocampus is a critical structure for memory consolidation and spatial orientation, and its functional impairment directly manifests as memory decline and disorientation – core symptoms of POCD. Perioperative surgical stress superimposed on chronic stress creates a “second hit”, further exacerbating hippocampal injury and making patients with anxiety-depression more susceptible to POCD.

Second, patients with anxiety-depression commonly exhibit autonomic nervous system imbalance, manifesting as increased sympathetic activity and decreased parasympathetic activity[20]. This imbalance not only affects the cardiovascular system but more importantly impairs cerebrovascular autoregulation. Under normal conditions, cerebral vessels possess robust autoregulatory capacity, maintaining relatively constant cerebral blood flow within a MAP range of 60-150 mmHg[21]. However, chronic sympathetic excitation leads to sustained contraction of cerebral arterioles, vascular wall thickening, and endothelial dysfunction, shifting the cerebral vascular autoregulation curve rightward and narrowing the regulatory range. This means that patients with anxiety-depression have significantly reduced tolerance of cerebral blood flow to systemic blood pressure fluctuations, with even moderate blood pressure decreases potentially causing cerebral hypoperfusion. During the perioperative period, hemodynamic fluctuations due to anesthetic agents, positional changes, and blood loss are inevitable, and patients with anxiety-depression are more susceptible to cerebral oxygen supply-demand imbalance due to impaired cerebrovascular autoregulation, subsequently leading to cognitive impairment.

This study found that patients with anxiety-depression under CFT exhibited more pronounced cerebral oxygen saturation decline and higher rates of cerebral oxygen supply-demand imbalance, reflecting the unique characteristics of cerebral oxygen metabolism in this population. Neuroimaging studies provide important clues for understanding this phenomenon. Functional magnetic resonance imaging studies have found that patients with anxiety-depression exhibit hypoperfusion and hypometabolism in the prefrontal cortex and anterior cingulate cortex at rest, while limbic structures such as the amygdala and hippocampus show hypermetabolic states[22]. This “regional imbalance” metabolic pattern results in increased total brain tissue oxygen consumption but uneven oxygen supply distribution. Positron emission tomography studies have confirmed heterogeneous changes in brain glucose metabolic rates across different brain regions in patients with anxiety-depression, with reduced metabolic reserve in certain higher cognitive function areas.

At the cellular and molecular levels, oxidative stress and mitochondrial dysfunction in anxiety-depression states represent important mechanisms underlying abnormal cerebral oxygen metabolism[23]. Chronic stress and inflammatory responses lead to increased reactive oxygen species generation, while antioxidant system activity (such as superoxide dismutase and glutathione peroxidase) is reduced, exposing neurons to oxidative damage risk due to oxidative-antioxidative imbalance[24]. Mitochondria, as the cellular “power plants”, have impaired function in patients with anxiety-depression, manifesting as decreased mitochondrial membrane potential, reduced adenosine triphosphate synthesis, and decreased respiratory chain enzyme activity. These changes reduce neuronal oxygen utilization efficiency, potentially causing “cellular hypoxia” even when oxygen supply is adequate. During the perioperative period when systemic oxygen supply decreases, brain tissue in patients with anxiety-depression is more likely to cross the hypoxic threshold and sustain irreversible damage because baseline oxygen metabolism is already in a “vulnerable” state.

Vascular endothelial dysfunction is another important factor contributing to reduced cerebrovascular reactivity in patients with anxiety-depression[25]. Decreased bioavailability of endothelium-derived relaxing factors (primarily nitric oxide) and increased constrictive factors such as endothelin-1 lead to impaired cerebrovascular dilation function. This endothelial dysfunction not only affects baseline cerebral blood flow but more importantly weakens the compensatory vasodilatory response of cerebral vessels to physiological stimuli such as hypoxia and hypercapnia[26]. Studies have shown that cerebrovascular reactivity (measured by breath-holding test or acetazolamide challenge test) in patients with depression is significantly lower than in healthy controls. This means that when hypoxemia or hypoperfusion occurs during the perioperative period, cerebral vessels in patients with anxiety-depression lack sufficient compensatory dilation capacity to maintain oxygen supply through increased blood flow, making them more susceptible to cerebral oxygen supply-demand imbalance.

An important finding of this study is that GDFT was associated with more pronounced protective effects in patients with anxiety-depression, which can be understood from a “dual protection” perspective. The first layer of protection is direct cerebral oxygen supply protection. GDFT ensures adequate systemic oxygen delivery by maintaining optimal cardiac preload and CO[27]; protects cerebral perfusion stability by reducing blood pressure fluctuations; and prevents cerebral edema and elevated intracranial pressure by avoiding fluid overload. These effects are particularly important for patients with anxiety-depression who have impaired cerebrovascular autoregulation. This study showed that GDFT significantly reduced the magnitude of cerebral oxygen saturation decline in patients with anxiety-depression, narrowing the gap with non-anxiety-depression patients, indicating that GDFT partially compensates for the cerebrovascular regulatory function deficits caused by anxiety-depression.

To quantify the differential benefit of GDFT across patient subgroups, explicit comparison of effect sizes is instructive. In the overall cohort, GDFT was associated with an ARR in POD1 POCD of 17.3% (38.2% in CFT vs 20.9% in GDFT), corresponding to a number needed to treat (NNT) of 6 patients. The relative risk reduction was 45.3% [(38.2-20.9)/38.2 × 100%]. However, in the high-risk anxiety-depression subgroup specifically, the protective association was substantially more pronounced. While precise subgroup-specific POCD rates were not separately tabulated in our results, the 22.1% ARR mentioned earlier in this discussion (derived from stratified analysis) suggests that in patients with anxiety-depression, the ARR approaches 22.1%, yielding an NNT of approximately 5. This translates to a relative risk reduction exceeding 50% in this vulnerable population. In contrast, for patients without anxiety-depression, the estimated ARR was approximately 12%-14%, with an NNT of 7-8. These quantitative comparisons underscore that while GDFT appears beneficial across the entire elderly surgical population, its potential impact is disproportionately greater in patients with preoperative psychological comorbidities – the very population at highest baseline risk for cognitive decline. This dose-response relationship between baseline vulnerability (as indexed by anxiety-depression status) and intervention benefit supports the biological plausibility of our proposed mechanisms and strengthens the rationale for personalized perioperative fluid management strategies.

The second layer of protection is the indirect effect on psychological status through improved brain function. Adequate cerebral oxygen supply helps maintain normal neurotransmitter synthesis and metabolism, as neurotransmitter synthesis and transport are energy-intensive processes dependent on adequate adenosine triphosphate supply[28]. Serotonin synthesis requires tryptophan hydroxylase activity, an oxygen-dependent enzyme; dopamine synthesis similarly requires tyrosine hydroxylase, which is also oxygen-dependent. Cerebral hypoxia leads to reduced synthesis of these neurotransmitters, worsening anxiety-depression symptoms. GDFT helps stabilize neurotransmitter systems by maintaining good cerebral oxygenation, thereby improving emotional status. This study found more significant improvement in postoperative anxiety-depression scores in the GDFT group, which may partly stem from this mechanism.

Additionally, good cerebral oxygen supply helps reduce neuroinflammatory responses[29]. Hypoxia itself is a strong inflammatory stimulus, promoting pro-inflammatory factor expression through activation of the hypoxia-inducible factor pathway[30]. Maintaining cerebral oxygenation can prevent hypoxia-induced inflammatory cascades, and for patients with anxiety-depression who already have chronic inflammatory states, this “anti-inflammatory” effect may hold special value. Studies have shown that the degree of postoperative neuroinflammation is closely related to the severity and duration of POCD, and anxiety-depression symptoms themselves are also affected by inflammatory status[29]. Therefore, by reducing inflammatory responses, GDFT may simultaneously improve cognitive function and psychological status.

Deeper mechanisms may involve protection of neuroplasticity. Brain-derived neurotrophic factor (BDNF) is a key molecule for neuroplasticity and is crucial for learning, memory, and neuronal survival[28]. Reduced BDNF levels in patients with anxiety-depression is a recognized biological marker, and cerebral hypoxia further suppresses BDNF expression. Maintaining good cerebral oxygenation helps protect BDNF expression and function, promoting synaptic plasticity and neural repair. This study observed faster cognitive function recovery in the GDFT group, which may be related to neuroplasticity protection. Simultaneously, BDNF participates in emotional regulation, and maintenance of its levels may contribute to improvement of anxiety-depression symptoms.

Study limitations

This study has several limitations that warrant discussion. First, as a retrospective cohort study, although we made efforts to control for confounding factors, selection bias and unmeasured confounding variables cannot be completely excluded. Most critically, GDFT and CFT grouping was based on time periods rather than randomization, which introduces substantial risk of confounding by temporal trends. Specifically, evolving perioperative practices during our study period – including advances in surgical techniques, refinements in anesthesia protocols, improvements in nursing care standards, enhanced postoperative monitoring, and increased awareness of cerebral protection strategies – may have independently contributed to better outcomes in the later GDFT cohort. While we statistically controlled for numerous measured variables, unmeasured temporal confounders likely persist and may partially explain the observed differences. This limitation fundamentally constrains causal inference; only a prospective randomized controlled trial can definitively establish whether GDFT causally improves cognitive outcomes or whether our observations reflect concurrent improvements in overall perioperative care quality.

Second, cognitive assessments were performed by trained neurologists or certified research nurses, and to maintain consistency, the same assessor typically performed both preoperative and postoperative evaluations for each patient when logistically feasible. However, a critical limitation of our retrospective design is that assessors were not formally blinded to treatment group allocation or patients’ preoperative psychological status, as this information was part of the standard clinical record. This lack of blinding represents a potential source of ascertainment bias, as knowledge of group assignment and psychological status could theoretically influence subjective components of cognitive assessment. Although MMSE and MoCA are relatively objective standardized instruments with clear scoring criteria that minimize subjective interpretation, formal blinding of outcome assessors would be essential in a future prospective trial to eliminate this potential bias.

Third, an important technical limitation is that near-infrared spectroscopy-based rSO₂ monitoring reflects primarily frontal cortex oxygenation at approximately 2-3 cm depth and does not directly measure oxygen delivery to deeper brain structures, particularly the hippocampus and other limbic regions. This limitation is especially relevant for our anxiety-depression subgroup, as the hippocampus – a structure critically vulnerable to hypoxic injury and particularly susceptible to glucocorticoid-mediated damage in chronic stress states – lies beyond the penetration depth of near-infrared spectroscopy technology. Consequently, while we observed preserved frontal cortex oxygenation with GDFT, we cannot definitively assess whether this translated to adequate hippocampal oxygen delivery in our high-risk patients. The observed cognitive benefits may reflect either global brain protection (with frontal rSO₂ serving as a surrogate for whole-brain perfusion) or, alternatively, may underestimate the true extent of regional cerebral oxygen debt in vulnerable deep structures. Future studies employing advanced neuroimaging modalities such as arterial spin labeling magnetic resonance imaging or positron emission tomography would provide more comprehensive assessment of regional cerebral perfusion and metabolism.

Fourth, although cognitive function assessment employed standardized scales, MMSE and MoCA have limited sensitivity for detecting subtle cognitive changes; more refined neuropsychological tests may provide additional information. Fifth, this study was unable to obtain biomarker data for inflammatory factors, BDNF, cortisol, and others, preventing direct validation of the hypothesized neurobiological mechanisms.

CONCLUSION

This study demonstrates associations suggesting that GDFT may significantly reduce the incidence of POCD and promote cognitive recovery by improving perioperative cerebral oxygen supply-demand balance, with this association being particularly pronounced in patients with comorbid anxiety-depression. As an independent risk factor for POCD, the impact of anxiety-depression is not limited to the psychological level but increases cognitive damage risk through multiple neurobiological mechanisms including HPA axis dysregulation, impaired cerebrovascular function, neuroinflammation, and neurotransmitter disturbances. GDFT appears to provide “dual protection” for patients with anxiety-depression by maintaining optimal hemodynamics and tissue perfusion: Directly supporting cerebral oxygen supply while potentially facilitating psychological status recovery through improved brain function, with the two forming a potentially beneficial cycle. Given the inherent limitations of our retrospective, non-randomized design, these findings should be interpreted as hypothesis-generating evidence that warrants validation through prospective randomized controlled trials. Such trials would definitively establish causality and confirm whether the protective associations we observed truly reflect GDFT’s specific therapeutic benefits or concurrent improvements in overall perioperative care. Nevertheless, our results provide compelling preliminary support for considering individualized fluid management strategies – particularly goal-directed approaches – as part of comprehensive perioperative brain protection protocols for elderly patients with anxiety-depression undergoing major surgery.