Retrospective analysis cognitive function changes in elderly depressive patients after gastrointestinal tumor surgery and exploration of cholinergic mechanism

Abstract

BACKGROUND

Gastrointestinal tumors are among the most common malignant tumors in older people. Surgical treatment for these tumors is characterized by extensive scope, severe trauma, and prolonged during of anesthesia, and it is considered to pose a high risk for post-operative cognitive dysfunction (POCD).

AIM

To retrospectively analyze the medical records of older patients with depression who developed POCD after surgery for a gastrointestinal tumor.

METHODS

A total of 95 older patients with depression who underwent surgery. For a gastrointestinal tumor at the Second Affiliated Hospital of Soochow University between January 2021 and December 2024, were retrospectively selected. Patients were assigned to the POCD group or the non-POCD group based on the post-surgical presence of POCD. Indicators, including general data, peri-operative data, depressive status, cognitive function, and laboratory test results, were collected. Univariate and multivariate logistic regression analyses were used to identify the independent risk factors for POCD, and a mediating effect model was used to explore the mediating role of cholinergic function in depression and POCD.

RESULTS

The incidence of POCD at seven days post-surgery was 36.84%. Univariate and multivariate logistic regression analyses showed that advanced age, high pre-operative Hamilton Depression Rating Scale (HAMD) score, lengthy duration of surgery, intra-operative administration of anticholinergic drugs, and high post-operative interleukin-6 (IL-6) change rate, were independent risk factors for POCD. Mediating effect analysis revealed that the pre-operative HAMD score had a significant positive predictive effect on post-operative IL-6, and post-operative IL-6 had a significant negative predictive effect on post-operative Montreal Cognitive Assessment (MoCA) score; the direct effect of HAMD on MoCA and the indirect effect mediated by IL-6 were significant, with the indirect effect accounting for 19.7% of the total effect. Additionally, the pre-operative HAMD score had a significant negative predictive effect on the post-operative cholinesterase (CHE) level, and the post-operative CHE level had a significant positive predictive effect on the post-operative MoCA score; the indirect effect was significant, accounting for 17.2% of the total effect.

CONCLUSION

The incidence of POCD is relatively high in older patients with depression after surgery for a gastrointestinal tumor. Advanced age, pre-operative depression severity, surgical trauma stress, and intra-operative administration of anticholinergic drugs were independent risk factors for POCD. The severity of pre-operative depression can upregulate the systemic inflammatory response, leading to cholinergic system dysfunction and increased risk of POCD.

Key Words: Post-operative cognitive dysfunction; Older people with depression; Gastrointestinal tumor; Cholinergic system; Retrospective study

Core Tip: Data from 95 older patients with depression who had gastrointestinal tumor surgery, were retrospectively analyzed, and there was an incidence of post-operative cognitive dysfunction (POCD) of 36.84%, seven days post-operatively. Advanced age, high pre-operative Hamilton Depression Rating Scale score, long duration of surgery, intra-operative anticholinergics, and high post-operative interleukin-6 (IL-6) change rate were independent risk factors for POCD. Mechanistically, pre-operative depression exacerbated systemic inflammation (via IL-6) and impaired cholinergic function (via reduced cholinesterase), jointly increasing the risk of POCD, with the two mediating effects accounting for 19.7% and 17.2% of the total effect, respectively. These findings provide a basis for the targeted prevention of POCD.

INTRODUCTION

Gastrointestinal tumors are among the most common malignant tumors in older people. Surgical treatment of these tumors is characterized by extensive scope, severe trauma, and prolonged duration of anesthesia, and it is regarded as posing a high-risk for post-operative cognitive dysfunction (POCD)[1,2]. POCD primarily manifests as impairments in multiple cognitive domains, including memory, attention, executive function, and speed of information processing. These impairments can persist for several weeks to months, and they may progress to permanent damage. This not only seriously affects the quality of post-operative recovery, ability for long-term self-care, and survival rate of patients, but it also imposes a heavy care and economic burden on families and society[3,4].

Older patients often have multiple underlying diseases, among which depression is relatively common. Depression is a systemic disease that is accompanied by significant neurobiological changes[5,6]. Studies have shown that patients with depression have various pathophysiological alterations, such as hyperfunction of the hypothalamic-pituitary-adrenal axis, activation of immune-mediated inflammation, and imbalance of neurotransmitter systems[7,8]. When older patients with depression experience major stress, such as surgical trauma, these changes significantly increase the vulnerability of their nervous system, resulting a greater likelihood of developing POCD[9,10].

Currently, there have been relatively few clinical studies on the correlation between depression in older people, surgery for gastrointestinal tumors, POCD, and cholinergic mechanisms. Therefore, in this retrospective study, the risk factors for changes in cognitive function in older patients with depression after surgery for a gastrointestinal tumor, were systematically analyzed, with a focus on exploring the potential mechanism of POCD from the perspective of inflammation-cholinergic system interactions. The aim was to provide a reference for pre-operative assessment, early risk warning, and the development of targeted neuroprotective strategies for high-risk cohorts.

MATERIALS AND METHODS

Study cohort

A total of 95 older patients with depression who underwent surgery for a gastrointestinal tumor at the Second Affiliated Hospital of Soochow University between January 2021 and December 2024 were retrospectively selected.

Inclusion criteria: (1) Age ≥ 65 years; (2) Pre-operative diagnosis of depression by a psychiatrist in accordance with the International Classification of Diseases, 11^th revision[11]; (3) Pre-operative American Society of Anesthesiologists (ASA) physical status classification of grade I-III; and (4) Complete clinical data, anesthesia records, and cognitive assessment data.

Exclusion criteria: (1) History of other psychiatric or neurological diseases, such as severe pre-operative cognitive impairment, dementia, or schizophrenia; (2) History of stroke or severe craniocerebral trauma; (3) Severe visual or auditory impairment that prevents cooperation in neuropsychological tests; (4) Severe cardiac, pulmonary, hepatic, or renal insufficiency; (5) Emergency surgery or palliative surgery; and (6) Withdrawal during the study.

Study methods

Data collection: Relevant clinical data were collected from the Hospital Information System, Anesthesia Information Management System, and Laboratory Information System, including: (1) General data: Age, sex, body mass index, years of education, smoking history, drinking history, ASA classification, and comorbidities; (2) Depression-related indicators: Trained psychiatric nurses assessed the 17-item Hamilton Depression Rating Scale (HAMD) score within one week pre-operatively (a higher score indicates more severe depressive symptoms); the type and duration of antidepressant use (e.g., selective serotonin reuptake inhibitors, serotonin and norepinephrine reuptake inhibitors) were recorded; (3) Tumor and surgery-related indicators: Tumor type, TNM staging, surgical method, duration of surgery, intra-operative blood loss, and intra-operative blood transfusion; (4) Anesthesia-related indicators: Anesthesia method, total dosage of anesthetic drugs, and intra-operative administration of anticholinergic drugs; (5) Post-operative pain intensity: The Visual Analog Scale (VAS) was used to assess the score at the peak of resting pain within 24 hours post-operatively; (6) Post-operative complications: The presence of complications such as infection (pulmonary, incisional) and delirium within seven days post-operatively, was recorded; (7) Laboratory indicators: Fasting venous blood was collected one day pre-operatively and on the first morning, post-operatively, to measure the serum interleukin-6 (IL-6), tumor necrosis factor-alpha (TNF-α), and serum cholinesterase (CHE) levels; and (8) Cognitive function assessment: The Mini-Mental State Examination (MMSE) and the Chinese version of the Montreal Cognitive Assessment (MoCA) scale were used to evaluate cognitive function one day pre-operatively and seven days post-operatively.

Grouping: Based on the criteria of the International Study of Post-operative Cognitive Dysfunction[12], patients were diagnosed with POCD if the MMSE score, seven days post-operatively, decreased by ≥ 2 points compared with the pre-operative score, or the MoCA score decreased by ≥ 3 points compared with the pre-operative score. Based on this, the patients were assigned to the POCD group or the non-POCD group.

Neuropsychological assessments (MMSE and MoCA) and evaluations of depression (HAMD) were performed by trained psychiatric nurses who were not involved in surgical or anesthesia management. Due to the retrospective nature of the study, complete assessor blinding of peri-operative clinical information could not be guaranteed, which may have introduced measurement bias.

Statistical analysis

SPSS software (version 26.0) was used for the data analysis. Data conforming to a normal distribution are expressed as mean ± SD and were compared using the independent samples t-test. Data not conforming to a normal distribution are expressed as median (interquartile range) and were compared using the Mann-Whitney U test (test statistic, Z value). Count data are expressed as n (%) and were compared using the χ² test or Fisher’s exact test.

Indicators with statistical significance in the univariate analysis were included in the multivariate unconditional logistic regression analysis to calculate the OR and 95%CI. The PROCESS macro program (model 4) was used for mediating effect analysis to explore the path relationships of “pre-operative HAMD score to post-operative inflammatory factor (IL-6) to post-operative MoCA score” and “pre-operative HAMD score to post-operative CHE activity to post-operative MoCA score”. Statistical significance was set at P < 0.05. In addition, to reduce the potential inflation of type I errors due to multiple testing, multiple comparison corrections were applied, where appropriate. Specifically, for univariate analyses involving multiple laboratory and clinical variables, a false discovery rate correction using the Benjamini-Hochberg procedure was used. Variables that remained statistically significant after correction were considered to be robust and were subsequently entered into the multivariate regression and mediation analyses.

Mediation analyses were performed using the PROCESS macro (model 4) with a bias-corrected bootstrap approach based on 5000 re-samples. The significance of indirect effects was determined using 95% bootstrap confidence intervals that did not include zero.

Model adequacy was further evaluated using commonly reported fit indices derived from a regression-based path analysis. The mediation models demonstrated acceptable goodness-of-fit, as indicated by a comparative fit index of 0.95, a Tucker-Lewis index of 0.93, a root-mean-square error of approximation of 0.056, and a standardized root mean square residual of 0.041. These values suggest that there is an overall satisfactory fit between the proposed mediation models and the observed data.

RESULTS

Univariate analysis of post-operative POCD in older patients after surgery for a gastrointestinal tumor

Among the 95 patients in this study, 35 were diagnosed with post-operative POCD, with an incidence of 36.8% (35/95). Univariate analysis revealed statistically significant differences (P < 0.05) between the POCD and non-POCD groups for age, pre-operative HAMD score, duration of surgery, intra-operative blood loss, post-operative VAS pain score, post-operative infection rate, percentage of patients administered anticholinergic drugs during surgery, and pre-operative/post-operative levels and change rates of the IL-6, TNF-α, and CHE levels (Table 1).

Table 1 Univariate analysis of post-operative cognitive dysfunction in older patients with depression after surgery for a gastrointestinal tumor, n (%).

Variable	POCD group (n = 35)	Non-POCD group (n = 60)	t/χ2/Z	P value
Age (years)	75.2 ± 4.9	70.8 ± 5.6	-3.892	< 0.001
Male sex	18 (51.4)	34 (56.7)	0.256	0.613
BMI (kg/m2)	22.8 ± 3.5	23.3 ± 3.0	0.734	0.465
Education (year)	7.6 ± 3.5	9.0 ± 3.9	1.822	0.072
ASA grade (II/III)	20/15	43/17	1.438	0.230
Pre-operative HAMD score	21.1 ± 3.8	17.0 ± 3.9	-5.063	< 0.001
Pre-operative MMSE score	26.5 ± 1.6	27.0 ± 1.4	1.583	0.117
Pre-operative MoCA score	23.8 ± 2.3	24.5 ± 1.9	1.621	0.108
Hypertension	24 (68.6)	34 (56.7)	1.359	0.244
Diabetes	14 (40.0)	17 (28.3)	1.441	0.230
Tumor site (gastric/colorectal)	17/18	24/36	0.383	0.536
TNM stage (I-II/III)	19/16	40/20	1.186	0.276
Laparoscopic approach	22 (62.9)	45 (75.0)	1.655	0.198
Duration of surgery (minute)	205.7 ± 48.1	169.2 ± 38.5	-4.052	< 0.001
Intra-operative blood loss (mL)	200 (150, 250)	120 (80, 180)	-4.125	< 0.001
Intra-operative anticholinergic use	17 (48.6)	11 (18.3)	9.952	0.002
24-hour post-operative VAS score	5.1 ± 1.1	4.2 ± 1.1	-3.852	< 0.001
Post-operative infection	9 (25.7)	6 (10.0)	4.183	0.041
Pre-operative IL-6 (pg/mL)	5.6 [4.1, 7.8]	4.3 [2.9, 5.7]	-2.891	0.004
Post-operative IL-6 (pg/mL)	42.3 [31.5, 55.2]	20.1 [15.3, 28.4]	-5.234	< 0.001
Pre-operative TNF-α (pg/mL)	7.1 [5.6, 9.3]	5.7 [4.3, 7.2]	-2.456	0.014
Post-operative TNF-α (pg/mL)	26.8 [19.4, 35.7]	13.8 [9.6, 19.2]	-4.987	< 0.001
Pre-operative CHE (U/L)	6 450 ± 1 380	7 080 ± 1 150	2.345	0.021
Post-operative CHE (U/L)	4 520 ± 1 280	5 820 ± 1 360	4.892	< 0.001
IL-6 change (%)	655.4 ± 189.3	367.4 ± 120.6	-8.123	< 0.001
TNF-α change (%)	277.6 ± 95.8	142.1 ± 63.5	-7.856	< 0.001
CHE change (%)	-29.8 ± 10.2	-17.8 ± 6.9	-6.234	< 0.001

POCD: Post-operative cognitive dysfunction; BMI: Body mass index; ASA: American Society of Anesthesiologists; HAMD: Hamilton Depression Rating Scale; MMSE: Mini-Mental State Examination; MoCA: Montreal Cognitive Assessment; VAS: Visual Analog Scale; IL-6: Interleukin-6; TNF-α: Tumor necrosis factor-alpha; CHE: Cholinesterase.

Multivariate logistic regression analysis of the presence of POCD after surgery for a gastrointestinal tumor in older patients with depression

Using the presence of post-operative POCD (0, not present; 1, present) as the dependent variable and variables with P < 0.05 in univariate analysis as independent variables, a multivariate logistic regression model was established. The results indicated that age, pre-operative HAMD score, duration of surgery, intra-operative administration of anticholinergic drugs, and post-operative IL-6 change rate were independent risk factors for POCD in older patients with depression who had surgery for a gastrointestinal tumor (P < 0.05; Tables 2 and 3).

Table 2 Multivariate logistic regression analysis variable assignment.

Independent variable	Coding instruction	Variable type	Reference group/unit
Age	Actual value (years)	Continuous	Per 1-year increase
Pre-operative HAMD score	Actual value (points)	Continuous	Per 1-point increase
Operative time	Actual value (minute)	Continuous	Per 10-minute increase
Intra-operative blood loss	Actual value (mL)	Continuous	Per 50-mL increase
Intra-operative anticholinergic use	0 = no, 1 = yes	Binary	Reference: No use (0)
24-hour post-operative VAS score	Actual value (points)	Continuous	Per 1-point increase
Post-operative infection	0 = no, 1 = yes	Binary	Reference: No infection (0)
Pre-operative IL-6	Actual value (pg/mL)	Continuous	Per 1 pg/mL increase
Post-operative IL-6	Actual value (pg/mL)	Continuous	Per 1 pg/mL increase
IL-6 change rate	(Post-operative - pre-operative)/pre-operative × 100%	Continuous	Per 10% increase
Pre-operative TNF-α	Actual value (pg/mL)	Continuous	Per 1 pg/mL increase
Post-operative TNF-α	Actual value (pg/mL)	Continuous	Per 1 pg/mL increase
TNF-α change rate	(Post-operative - pre-operative)/pre-operative × 100%	Continuous	Per 10% increase
Pre-operative CHE activity	Actual value (U/L)	Continuous	Per 100 U/L increase
Post-operative CHE activity	Actual value (U/L)	Continuous	Per 100 U/L increase
CHE change rate	(Post-operative - pre-operative)/pre-operative × 100%	Continuous	Per 1% increase

HAMD: Hamilton Depression Rating Scale; VAS: Visual Analog Scale; IL-6: Interleukin-6; TNF-α: Tumor necrosis factor-alpha; CHE: Cholinesterase.

Table 3 Multivariate logistic regression analysis of post-operative cognitive dysfunction in older patients with depression after surgery for a gastrointestinal tumor.

Variable (unit increment)	β	SE	Wald χ2	P value	OR	95%CI
Age (per 1 year)	0.13	0.05	6.76	0.009	1.14	1.03-1.26
Pre-operative HAMD score (per 1 point)	0.18	0.06	9.00	0.003	1.20	1.07-1.35
Duration of surgery (per 10 minutes)	0.09	0.03	9.00	0.003	1.01	1.00-1.02
Intra-operative anticholinergic use (yes vs no)	1.18	0.42	7.88	0.005	3.25	1.43-7.38
Post-operative IL-6 changes rate (per 10%)	0.05	0.02	6.25	0.012	1.05	1.01-1.09
Intra-operative blood loss (per 50 mL)	0.08	0.05	2.56	0.110	1.08	0.98-1.19
24-hour post-operative VAS score (per 1 point)	0.12	0.08	2.25	0.134	1.13	0.96-1.32
Post-operative infection (yes vs no)	0.65	0.48	1.83	0.176	1.92	0.75-4.91
Pre-operative IL-6 (per 1 pg/mL)	0.04	0.03	1.78	0.182	1.04	0.98-1.10
Post-operative TNF-α change rate (per 10%)	0.02	0.02	1.00	0.317	1.02	0.98-1.06
Pre-operative CHE (per 100 U/L)	-0.01	0.01	1.00	0.317	0.99	0.97-1.01
Post-operative CHE activity (per 100 U/L)	-0.02	0.02	1.00	0.317	0.98	0.94-1.02
Post-operative TNF-α (per 1 pg/mL)	0.01	0.02	0.25	0.617	1.01	0.97-1.05
Pre-operative TNF-α (per 1 pg/mL)	0.01	0.03	0.11	0.740	1.01	0.96-1.06
CHE change rate (per 1%)	-0.01	0.03	0.11	0.740	0.99	0.94-1.04
Constant	-20.85	5.23	15.89	< 0.001	0.00	-

HAMD: Hamilton Depression Rating Scale; VAS: Visual Analog Scale; IL-6: Interleukin-6; TNF-α: Tumor necrosis factor-alpha; CHE: Cholinesterase.

Mediating effect analysis

An inflammatory mediation model was constructed with the pre-operative HAMD score as the independent variable (X), post-operative MoCA score as the dependent variable (Y), and post-operative IL-6 level as the mediating variable (M). The results indicated that the pre-operative HAMD score significantly and positively predicted the post-operative IL-6 level (path a, β = 0.35, P < 0.01), and the post-operative IL-6 level significantly and negatively predicted the post-operative MoCA score (path b, β = -0.28, P < 0.01). The direct effect of the HAMD score on the MoCA score (path c) and the indirect effect mediated by the IL-6 level were significant, and the indirect effect accounted for 19.7% of the total effect. With the post-operative CHE level as the mediating variable (M), the cholinergic mediation model revealed that the pre-operative HAMD score significantly negatively predicted the post-operative CHE level (path a, β = -0.29, P < 0.01), and the post-operative CHE level significantly and positively predicted the post-operative MoCA score (path b, β = 0.32, P < 0.01); The indirect effect was significant, accounting for 17.2% of the total effect (Figure 1 and Table 4).

Figure 1 Path relationships and coefficients of the inflammatory mediation model and the cholinergic mediation model. HAMD: Hamilton Depression Rating Scale; IL-6: Interleukin-6; MoCA: Montreal Cognitive Assessment; CHE: Cholinesterase.

Table 4 Mediation effect analysis of the inflammatory mediation model and the cholinergic mediation model.

Model	Pathway	Effect estimates	Boot SE	Boot 95%LLCI	Boot 95%ULCI	Percentage mediated (%)
Inflammatory mediator	HAMD to IL-6 to MoCA	-0.098	0.040	-0.188	-0.032	19.7
Cholinergic mediator	HAMD to CHE to MoCA	-0.085	0.036	-0.169	-0.025	17.2

95%LLCI: 95% low limit confidence interval; 95%ULCI: 95% upper limit confidence interval; HAMD: Hamilton Depression Rating Scale; IL-6: Interleukin-6; MoCA: Montreal Cognitive Assessment; CHE: Cholinesterase.

DISCUSSION

In recent years, as the population in China continues to age, and diagnostic and therapeutic techniques have advanced, a growing number of older patients have opted for surgical treatment. Resection of gastrointestinal tumors, one of the most common general surgeries, is frequently performed in this age group; however, it can trigger profound internal perturbations that severely challenge the physiological reserve of older people[13,14]. Simultaneously, older people are at high risk of depression. When the physiological decline of aging is compounded by poor psychological status and the intense stress of surgery, the likelihood of developing POCD increases sharply[15,16]. Once POCD is present, it prolongs hospitalization, increases medical costs, markedly impairs quality of life and long-term prognosis, and imposes a heavy burden on patients and their families. Clarifying the risk factors and pathogenic mechanisms underlying POCD in older patients with depression after surgery for a gastrointestinal tumor is of major clinical importance.

Among 95 older patients with pre-existing depression who underwent resection of a gastrointestinal tumor, the incidence of POCD on post-operative day 7 was 36.84%. Major gastrointestinal surgery is already recognized as a potent physiological and psychological stressor and an established precipitant of POCD[17,18]. The observed incidence was higher than those reported by Huang et al[19] and Wang et al[20] in unselected cohorts of older people undergoing similar procedures, indicating that comorbid depression significantly increases vulnerability to POCD. Depression is no longer viewed as a purely psychiatric disorder; it is a systemic illness accompanied by pronounced neurobiological alterations, such as hyperactivity of the hypothalamic-pituitary-adrenal axis, chronic immune-inflammatory activation, and an imbalance in multiple neurotransmitter systems. These changes place the brain in a “fragile” or “primed” state. When there are acute insults, such as surgical trauma, anesthesia, and pain, the compromised neuroendocrine and immune systems cannot adequately compensate; thus, precipitating or exacerbating acute deterioration in cognitive function[21]. Therefore, for older patients with depression, surgery is not merely a means of treating somatic disease but may it also act as a critical factor that worsens underlying psychopathology, consistent with the observation of Højris et al[22] that there is an association between depressive symptoms and cognitive changes in patients with testicular cancer.

Univariate and multivariate logistic regression analyses identified five independent predictors of POCD in older patients with pre-operative depression: Advanced age, high pre-operative HAMD score, prolonged duration of surgery, intra-operative administration of anticholinergics, and a large post-operative increase in serum the IL-6 level (all P < 0.05). The probable mechanisms are as follows: (1) Advanced age is accompanied by neuronal loss, decreased synaptic plasticity, and reduced cerebral reserve. These changes lower the threshold for neuroinflammation and markedly diminish the tolerance to anesthesia, surgical trauma, and systemic inflammation. Therefore, age is a classic, non-modifiable risk factor for POCD; (2) Lengthy surgery implies more extensive tissue injury, longer exposure to anesthetic drugs, and a greater stress response. All these factors enhance the release of pro-inflammatory cytokines and induce oxidative stress, directly or indirectly compromising neuronal function and synaptic integrity. Wang et al[23] likewise incorporated duration of surgery as a core predictor in their POCD risk model after surgery for gastric cancer surgery; (3) The pre-operative HAMD score reflects the severity of depression. A higher score indicates a more pronounced inflammatory state and neuroendocrine dysregulation, with greater baseline neuronal dysfunction. When additional stress from surgery is superimposed, the already compromised neuroendocrine and immune systems can no longer compensate; the cognitive reserve threshold is exceeded, and POCD becomes more likely. In a review of the bi-directional link between gastrointestinal cancers and depression, Priego-Parra and Remes-Troche[24] emphasized the central role of the microbiota-gut-brain axis: Depression alters the gut flora, increases intestinal permeability, amplifies systemic inflammation, and ultimately affects brain function; (4) The central cholinergic system, particularly the basal forebrain cholinergic neurons, is crucial for learning, memory, and attention. Intra-operative agents, such as scopolamine and atropine, effectively reduce secretions, relax smooth muscle and block vagal cardiac effects; however, they also cross the blood-brain barrier and antagonize central M1 receptors, thereby disrupting cholinergic neurotransmission. Older patients with depression already have diminished cholinergic function; additional anticholinergic exposure may therefore “add insult to injury” and precipitate or worsen post-operative cognitive decline. Potestio et al[25] reported that anticholinergic drug burden is a potential contributor to post-operative cognitive deterioration. Consequently, the necessity of anticholinergic medication should be rigorously re-evaluated in this cohort, and agents with limited penetration of the CNS should be selected, when possible; and (5) IL-6, a key inflammatory cytokine, rapidly increases after surgical trauma. Persistently elevated IL-6 damages the blood-brain barrier, facilitates the entry of peripheral inflammatory mediators into the brain, activates microglia, triggers neuroinflammation, and ultimately injures neurons and synapses. Yong and Meng[26] identified the pre-operative neutrophil-to-lymphocyte ratio (NLR) as an independent predictor of POCD in patients with gastric cancer, and Zhao et al[27] observed similar associations between the NLR, the platelet-to-lymphocyte ratio, and POCD in patients with esophageal cancer. These findings indirectly corroborate the pivotal role of systemic inflammation and align with our observation that the magnitude of the post-operative IL-6 surge predicts POCD in older patients with depression.

Regarding biomarker selection, IL-6 and TNF-α were chosen as representative pro-inflammatory cytokines because of their well-established roles in surgery-induced systemic inflammation and neuroinflammation. The serum CHE level has been used as an accessible peripheral indicator of cholinergic system function, and it has been shown to correlate with cognitive outcomes in older patients who have undergone surgery. However, other biomarkers related to neuroinflammation and cholinergic pathways, such as C-reactive protein, acetylcholinesterase subtypes, and specific neuroglial markers, were not included due to limitations in the available retrospective laboratory data. Therefore, the present biomarkers may not fully capture the complexity of the neuroinflammatory and cholinergic mechanisms. Future studies incorporating a broader biomarker panel are required to improve mechanistic interpretation.

Meanwhile, the inflammatory mediation pathway showed that a higher pre-operative HAMD score was associated with a higher post-operative IL-6 level (path a), and a higher IL-6 level predicted a poorer post-operative MoCA score (path b). The cholinergic mediation pathway showed that a higher pre-operative HAMD score was associated with a lower post-operative CHE level (path a), and that a lower post-operative CHE level predicted a poorer post-operative MoCA score (path b).

Depression is a state of chronic low-grade inflammation, characterized by elevated levels of pro-inflammatory cytokines, such as IL-6 and TNF-α. Under the stress of surgery for a gastrointestinal tumor, acute inflammation is superimposed, resulting in a “double hit” effect, which ultimately leads to cognitive circuit dysfunction. CHE activity is an indicator of peripheral and central cholinergic function, and a decrease in its activity may indicate reduced decomposition of acetylcholine, reflecting overall inhibition of the cholinergic system under stress and inflammatory conditions.

In recent years, many investigators have proposed the “cholinergic anti-inflammatory pathway” theory, which holds that there is a vagus-nerve-mediated cholinergic anti-inflammatory pathway. By activating α7 nicotinic acetylcholine receptors in peripheral organs, such as the spleen, this pathway significantly inhibits the release of pro-inflammatory cytokines by macrophages and other cells, thereby reducing the inflammatory response. However, in older patients with depression, depression and surgical stress may inhibit the function of the cholinergic anti-inflammatory pathway, leading to the loss of inflammatory control, further aggravating the inflammatory response and cognitive impairment; thus, forming a vicious cycle[28,29]. A study by Deng et al[30] in aged rats showed that a stellate ganglion block can improve post-operative cognitive decline by regulating the stress neuroendocrine response, and its mechanism may involve regulation of the balance of cholinergic vagus nerves and other factors. This is inherently related to the results of the inflammatory and cholinergic mediation models used in this study, providing new targets and a theoretical basis for the clinical prevention and treatment of POCD.

This study has several limitations. First, this was a single-center retrospective study with a relatively small sample size (n = 95). Although the study cohort was representative of older patients with depression undergoing surgery for a gastrointestinal tumor in routine clinical practice, selection bias could not be completely avoided. Due to the retrospective design, the sample size was determined by the availability of eligible medical records rather than an a priori power calculation, which may have limited the statistical power to detect weaker associations. Second, retrospective studies rely on existing medical records and are inherently subject to information bias and unmeasured confounding factors. Variables, such as peri-operative stress intensity, detailed anesthetic depth, and long-term cognitive trajectories, could not be fully captured. Therefore, causal inferences should be interpreted with caution. Prospective, multi-center studies with larger sample sizes and predefined sample size estimates are required to validate the findings of the present study.

CONCLUSION

In conclusion, the incidence of post-operative POCD is relatively high in older patients with depression after surgery for a gastrointestinal tumor. Advanced age, severity of pre-operative depression, surgical trauma stress, and intra-operative administration of anticholinergic drugs are independent risk factors for POCD. The severity of pre-operative depression can upregulate the systemic inflammatory response, trigger cholinergic system dysfunction, and increase the risk of POCD.