Clinical Relevance of Preoperative Lymphocyte-to-Monocyte Ratio in Predicting Postoperative Delirium Across Frailty Status in Older Surgical Patients

Introduction

Postoperative delirium (POD) is an acute and fluctuating neurocognitive disorder commonly observed in older surgical patients, characterized by impairments in attention, cognition, and awareness.^1,2 The reported incidence of POD among elderly surgical patients ranges from 5% to 50%.^3–5 POD not only prolongs hospital stay and increases postoperative complications, but also contributes to higher short- and long-term mortality, making it a significant challenge in perioperative care for the elderly.⁶ The pathogenesis of POD is multifactorial; however, growing evidence pinpoints surgery-induced neuroinflammation as a central mechanism.^4,7 Systemic inflammatory responses can disrupt the blood–brain barrier and activate microglia, triggering neuronal dysfunction, which highlights the potential utility of biomarkers reflecting these processes for early risk stratification.^8–10

In this context, the preoperative lymphocyte-to-monocyte ratio (LMR) has emerged as a readily accessible indicator of systemic inflammation and immune balance, and has been associated with a range of adverse clinical outcomes, including postoperative infections,¹¹ postoperative acute kidney injury,¹² and ischemic stroke.¹³ Lower preoperative LMR indicates a shift toward a pro-inflammatory state, characterized by heightened monocyte activity and impaired lymphocyte-mediated immune regulation, which may contribute to the development of POD. This pro-inflammatory shift may exacerbate perioperative neuroinflammation, disrupting the blood–brain barrier, and promoting cognitive dysfunction. Evidence supporting this mechanism has been reported in patients undergoing major abdominal surgery.¹⁴ Although direct evidence of LMR predicting POD in older non-cardiac, non-neurosurgical patients is lacking, studies in cardiac surgery and intensive care unit (ICU) contexts suggest that a lower LMR is associated with increased POD risk.¹⁵

Frailty is a common syndrome in older adults characterized by diminished physiological reserves.¹⁶ It is increasingly recognized as a major predisposing factor for POD.^17,18 Frail individuals often have chronic low-grade inflammation and immune dysregulation.¹⁹ Systemic inflammatory markers are involved in the pathophysiological processes of frailty and have also been identified as important predictors of POD, suggesting a potential biological link between frailty, inflammation, and delirium.^20,21 We hypothesized that the predictive value of inflammatory markers for POD may exhibit heterogeneity among older surgical patients. Specifically, frailty may act as an effect modifier that influences the association between the lymphocyte-to-monocyte ratio (LMR) and POD. As frailty severity increases, the dose–response relationship between LMR and POD may change. Clarifying this relationship may provide important insights for achieving precise perioperative risk stratification in older adults. Previous guidelines have emphasized that comprehensive geriatric assessment is essential for individualized POD prevention strategies.²² Despite these insights, previous studies have largely treated frailty merely as a confounding variable to be adjusted for, rather than a crucial effect modifier. Consequently, the specific subgroup of patients who would most benefit from LMR-based risk stratification remains undefined.

Therefore, this study investigated the association between LMR and POD within a frailty-stratified framework. By examining the dose–response relationship across different frailty strata, we aimed to identify the subgroup in which LMR provides the greatest predictive value. These findings may help improve perioperative risk stratification and support the development of personalized preventive strategies for this high-risk population.

Methods

Study Design and Data Source

This study was a retrospective analysis based on the Perioperative Database of Chinese Elderly Patients (PDCEP), a prospectively established, multicenter database of older surgical patients in China. All investigators involved in the PDCEP were trained in standardized data collection procedures. The establishment of the PDCEP and the initial cohort enrollment were conducted in accordance with the Declaration of Helsinki and approved by the Ethics Committee of the Chinese People’s Liberation Army (PLA) General Hospital (Approval No. S2019-311-03). Written informed consent was obtained from all participants or their legally authorized representatives at the time of enrollment into the PDCEP cohort. The present secondary analysis was separately reviewed and approved by the Ethics Committee of the Chinese People’s Liberation Army (PLA) General Hospital (Approval No. S2025-1049-01). Permission to access and analyze the PDCEP data was formally obtained from the database owner and the hospital administration. All data used in this study were anonymized prior to analysis. The PDCEP study was registered at ClinicalTrials.gov (NCT04911530). Reporting of this manuscript followed the Strengthening the Reporting of Observational Studies in Epidemiology (STROBE) guidelines.

Study Cohort

Patients aged ≥65 years who underwent elective non-cardiac and non-neurosurgical procedures and recieved the preoperative health and functional status questionnaire between April 2020 and April 2022 were eligible for inclusion. Exclusion criteria were as follows: (1) underwent local anesthesia or monitored anesthesia care; (2) missing lymphocyte or monocyte count data; (3) missing POD assessment; (4) incomplete preoperative FRAIL scale evaluation; and (5) missing data for more than 20% of covariates. The flowchart of patient selection is presented in Figure 1.

Figure 1 Flowchart of the study. Abbreviations: LMR, lymphocyte-to-monocyte ratio; POD, postoperative delirium; FRAIL, a scale used to assess frailty status.

Variable and Outcome Assessment

The primary variable of interest was preoperative LMR, calculated as the ratio of the absolute lymphocyte count to the absolute monocyte count. Lymphocyte and monocyte values were obtained from the most recent preoperative laboratory test before surgery. Frailty was assessed one day before surgery using the FRAIL scale, which comprises five domains: Fatigue, Resistance, Ambulation, Illnesses, and Weight loss.²³ Each item is scored as 0 or 1, yielding a total score ranging from 0 to 5, with higher scores indicating greater frailty. Based on the total score, patients were categorized as robust (0 point), pre-frail (1–2 points), or frail (3–5 points).²⁴

The primary outcome was the occurrence of POD, evaluated by trained researchers using the validated Chinese version of the 3-Minute Diagnostic Interview for Confusion Assessment Method (3D-CAM).⁷ Delirium assessments were performed twice daily for up to seven consecutive postoperative days or until hospital discharge, whichever occurred first.

Covariates and Data Collection

All study data were sourced from the Perioperative Database of Chinese Elderly Patients (PDCEP). To ensure data reliability, research assistants at each center underwent standardized training on variable extraction and questionnaire administration prior to data collection. Demographic characteristics (age, sex, body mass index, and American Society of Anesthesiologists physical status) and preexisting comorbidities (hypertension, coronary artery disease, stroke, and diabetes mellitus) were retrieved from preoperative clinical assessments and electronic medical records integrated within the PDCEP. Laboratory variables, including fasting blood glucose, serum creatinine, alanine aminotransferase (ALT), and aspartate aminotransferase (AST), were derived from routine preoperative blood tests recorded in the database. Perioperative details, such as anesthesia method, surgical approach, surgical duration, and intraoperative blood transfusion, were extracted from standardized anesthesia and surgical records housed in the PDCEP system.

Sample Size Calculation

The sample size was determined based on the number of eligible patients in the PDCEP cohort. Previous studies have reported that the incidence of POD is approximately 20% among older adults undergoing non-cardiac surgery.²⁵ Based on the observed association in our study, we assumed an odds ratio of 0.90 per one–standard deviation increase in the preoperative LMR. Using the sample-size formula for logistic regression with a continuous predictor,²⁶ we determined that a minimum of 4,420 participants would be required to achieve 80% power at a two-sided α of 0.05., After accounting for up to 20% missing data, the target sample size was increased to 5,525. The final analytic cohort included 6,475 patients, thereby exceeding the prespecified sample-size requirement.

Statistical Analysis

Missing covariate data (<20%) were imputed using the random forest multiple imputation method. Continuous variables were expressed as mean (standard deviation, SD) for normally distributed data and compared using one-way analysis of variance (ANOVA). Non-normally distributed variables were expressed as median (interquartile range, IQR) and compared using the Kruskal–Wallis test. Categorical variables were summarized as counts (percentages) and compared using the Chi-square test or Fisher’s exact test, as appropriate.

Both continuous and categorical analyses were performed. LMR was categorized into quartiles (Q1–Q4) based on its distribution in the study population, with Q1 representing the lowest and Q4 the highest quartile. To evaluate the association between preoperative LMR and POD, logistic regression analyses were conducted. Four models were sequentially constructed based on established clinical knowledge and prior literature regarding risk factors for postoperative delirium: Model 1, unadjusted; Model 2, adjusted for patient demographic and clinical characteristics including age, sex, BMI, frailty status, ASA physical status classification, hypertension, coronary artery disease, stroke, diabetes mellitus, fasting blood glucose, serum creatinine, ALT, and AST; Model 3, adjusted for recognized surgery-related variables including anesthesia method, surgical approach, surgical duration, and intraoperative blood transfusion; and Model 4, fully adjusted for all covariates. These models were designed to minimize potential confounding effects when estimating the independent association between preoperative LMR and POD.

Furthermore, stratified logistic regression analyses were performed according to frailty status (robust, pre-frail, and frail) to determine whether the association between LMR and POD differed across frailty subgroups. The potential dose–response relationship between LMR and POD risk was further assessed using restricted cubic spline (RCS) regression. Subgroup analyses were also conducted based on age, sex, BMI, ASA classification, hypertension, diabetes mellitus, and surgical duration, and the results were presented as forest plots.

All statistical tests were two-tailed, with a P-value <0.05 considered statistically significant. Statistical analyses were performed using R version 4.5.2 (R Foundation for Statistical Computing, Vienna, Austria).

Results

Population Characteristics

Between April 2020 and April 2022, a total of 8,751 elderly surgical patients from 19 participating centers were screened, and 6,475 patients were included in the final analysis. Among them, 789 (12.2%) developed postoperative delirium (POD). The number and types of excluded cases are shown in Figure 1. In the overall study population, the LMR ranged from 0.11 to 30.70. Patients were stratified into four groups according to the quartiles of LMR: Q1 (0.11 < LMR ≤ 2.98, n = 1,619), Q2 (2.98 < LMR ≤ 4.00, n = 1,627), Q3 (4.00 < LMR ≤ 5.23, n = 1,610), and Q4 (5.23 < LMR, n = 1,619). The baseline characteristics of the study population across LMR quartiles are summarized in Table 1.

Table 1 Baseline Characteristics of the Study Population by Quartiles of LMR

The median age of the overall population was 71.0 years (IQR: 67.0–75.0), and 2,781 participants (42.9%) were male. Across LMR quartiles, the median age decreased from 71 years in Q1 to 70 years in Q4 (P < 0.001). The proportion of female patients increased markedly from 27.7% in Q1 to 61.8% in Q4 (P < 0.001). BMI showed a modest upward trend, with the median value rising from 23.8 in Q1 to 24.4 in Q4 (P < 0.001). Regarding frailty status, 221 patients (3.4%) were classified as frail, 2,697 (41.7%) as pre-frail, and 3,557 (54.9%) as robust. The prevalence of preoperative frailty declined significantly from 5.13% in Q1 to 2.16% in Q4 (P < 0.001). Patients in the lower LMR quartiles had a higher proportion of ASA classification III–IV (P < 0.001). In terms of laboratory findings, serum creatinine levels decreased progressively with increasing LMR quartiles (P < 0.001). Similarly, the frequency of intraoperative blood transfusion decreased from 16.1% in Q1 to 8.8% in Q4 (P < 0.001).

Association Between LMR and POD Risk

The incidence of POD showed a significant inverse trend across LMR quartiles (P < 0.001), decreasing from 16.5% (267/1619) in Q1 to 10.0% (162/1619) in Q4. Logistic regression analyses were conducted to assess the association between LMR and POD risk, and the results are presented in Table 2. After adjustment for potential patient-related and surgery-related variables (Model 4), each 1-unit increase in LMR was associated with a lower odds of POD (OR = 0.94, 95% CI: 0.90–0.98, P = 0.009). Multicollinearity diagnostics were performed to evaluate potential collinearity among the covariates included in the fully adjusted model. All variables showed adjusted GVIF values well below the conventional threshold of 2, indicating no evidence of significant multicollinearity (Table S1). Compared with the lowest quartile (Q1), higher LMR levels were associated with significantly lower odds of POD, with adjusted ORs of 0.73 (95% CI: 0.59–0.90) for Q2, 0.70 (95% CI: 0.56–0.86) for Q3, and 0.68 (95% CI: 0.54–0.85) for Q4. As LMR levels increased, the risk of POD showed a continuous downward trend (P for trend < 0.001), indicating a graded inverse association across LMR quartiles.

Table 2 Association Between LMR and POD Risk

Restricted cubic spline (RCS) analysis further revealed a significant nonlinear association between preoperative LMR and the risk of POD in the overall population (P-overall = 0.003; P-nonlinearity = 0.034; Figure 2A).

Figure 2 Association between preoperative the lymphocyte-to-monocyte ratio (LMR) and the risk of postoperative delirium (POD) based on restricted cubic spline analysis. The solid line represents the adjusted odds ratios (ORs) for POD according to preoperative LMR, and the shaded area indicates the corresponding 95% confidence intervals (CIs). The bars represent the distribution of the study population across different LMR levels. The RCS model was adjusted for age, sex, body mass index (BMI), American Society of Anesthesiologists (ASA) class, preoperative fasting glucose, serum creatinine, alanine aminotransferase (ALT), aspartate aminotransferase (AST), comorbid hypertension, coronary heart disease, stroke, diabetes, surgical approach, operative duration, intraoperative blood transfusion, and anesthesia method. (A) Overall population; (B) Robust patients; (C) Pre-frail patients; (D) Frail patients. Abbreviations: LMR, lymphocyte-to-monocyte ratio; OR, odds ratio; CI, confidence interval.

Association Between LMR and POD Risk According to Frailty Status

As shown in Table 3, the incidence of postoperative delirium (POD) was 10.0% (357/3557) in the robust group, 13.9% (374/2697) in the pre-frail group, and 26.2% (58/221) in the frail group.

Table 3 Association Between the LMR and POD Risk According to Preoperative Frailty Status

In the robust group, no significant association was observed between the lymphocyte-to-monocyte ratio (LMR) and the risk of POD when LMR was modeled either as a continuous variable or categorized into quartiles. Consistent with these findings, restricted cubic spline (RCS) analysis did not demonstrate a significant association (P for overall = 0.335; P for nonlinearity = 0.736; Figure 2B).

In contrast, among pre-frail patients, a significant inverse association between LMR and POD emerged after adjusting for potential confounders. In Model 4, compared with the lowest LMR quartile (Q1), the adjusted odds ratios for POD were 0.59 (95% CI: 0.43 to 0.80, P = 0.001) for Q2, 0.64 (95% CI: 0.46 to 0.87, P = 0.005) for Q3, and 0.69 (95% CI: 0.50 to 0.95, P = 0.026) for Q4. This graded reduction in risk was further supported by restricted cubic spline (RCS) analysis, which revealed a significant nonlinear inverse relationship between LMR and POD risk (P for overall = 0.003; P for nonlinearity = 0.003; Figure 2C).

Among frail individuals, a weaker but still statistically significant inverse association was observed when LMR was treated as a continuous variable (OR = 0.79, 95% CI: 0.62 to 0.97, P = 0.035), with a significant linear trend across quartiles (P for trend = 0.032). However, only the comparison between Q3 and Q1 reached significance (P = 0.011), and RCS analysis did not indicate a nonlinear pattern (Figure 2D).

Taken together, these findings suggest that the association between LMR and POD may differ across frailty strata and appears to be most evident in pre-frail patients, modest and linear in frail individuals, and absent in those who are robust.

Subgroup Analysis

Subgroup analyses were further conducted to examine the robustness of the association between LMR and POD risk across clinically relevant strata (Figure 3). The inverse relationship between elevated LMR and reduced POD risk remained consistent across various subgroups, including different age, sex, BMI, ASA classification (I–II), preoperative hypertension, absence of preoperative diabetes mellitus, and surgery duration. No significant interaction was observed between LMR and any of these variables (all P for interaction > 0.05).

Figure 3 Subgroup analyses of the association between LMR and POD risk. Models were adjusted for age (<70 vs ≥70 years), sex, BMI (<24, 24–28, >28 kg/m2), ASA class (I–II vs III–IV), hypertension, coronary artery disease, stroke, diabetes mellitus, fasting blood glucose, serum creatinine, alanine aminotransferase (ALT), aspartate aminotransferase (AST), anesthesia method, surgical approach, surgical duration (<3 vs ≥3 hours), and intraoperative blood transfusion. Abbreviations: LMR, lymphocyte-to-monocyte ratio; BMI, body mass index; ASA, American Society of Anesthesiologists; OR, odds ratio; CI, confidence interval.

Sensitivity Analyses

To evaluate the robustness of the primary findings, several sensitivity analyses were conducted.

First, a parsimonious multivariable logistic regression model was constructed including key demographic variables (age, sex, BMI), major comorbidities (hypertension, coronary artery disease, prior stroke, diabetes mellitus), and selected perioperative factors (anesthesia method, surgical type, operative duration, and intraoperative blood transfusion). In this simplified model, preoperative LMR remained significantly associated with a reduced risk of POD (OR = 0.93, 95% CI 0.89–0.97, P = 0.002), consistent with the results of the fully adjusted model (Table S2).

Second, a bootstrap resampling procedure with 1000 iterations was performed within the frail subgroup. The bootstrap estimates were consistent with the primary regression results, and the confidence intervals remained stable, suggesting that the observed association between LMR and POD was unlikely to be driven by sampling variability in the relatively small frail subgroup (Table S3).

Third, we conducted an alternative analysis in which the pre-frail and frail groups were combined into a single vulnerable group. In this model, higher LMR remained significantly associated with a reduced risk of POD after adjustment for covariates (adjusted OR 0.936, 95% CI 0.879–0.996, P = 0.041), indicating that the inverse association between LMR and POD was not limited to the small frail subgroup but remained evident across a broader population of physiologically vulnerable patients (Table S4).

Predictive Performance of LMR

To further evaluate the potential clinical utility of LMR, we assessed whether adding LMR to a baseline model including frailty and clinical covariates improved prediction of POD.

The addition of LMR resulted in a small increase in the area under the receiver operating characteristic curve from 0.669 to 0.670 (P = 0.590) (Table S5). Although the improvement in overall discrimination was minimal, the IDI analysis suggested a small but statistically significant improvement in risk discrimination (IDI = 0.002, P = 0.008). (Table S5). The net reclassification improvement was not statistically significant (NRI = 0.026, P = 0.330) (Table S6).

Receiver operating characteristic analysis identified an exploratory cutoff value of 3.09 for LMR, corresponding to a sensitivity of 37.1% and a specificity of 73.7% (Table S7).

Discussion

In this multicenter cohort study of elderly non-cardiac, non-neurosurgical patients, we observed a significant association between the lymphocyte-to-monocyte ratio (LMR) and the incidence of postoperative delirium (POD). Further stratified analyses according to frailty status revealed that, among pre-frail individuals, lower preoperative LMR levels were significantly associated with a higher risk of POD in categorical analyses, showing a clear dose–response trend, while a nonlinear pattern was observed in restricted cubic spline (RCS) modeling. In frail patients, LMR also showed a significant inverse association with POD when modeled as a continuous variable, although no clear nonlinear pattern was detected. In contrast, no significant associations were found in the robust group. Previous work from the same research team using the same institutional database has primarily focused on the clinical effects of anesthetic agents, such as midazolam, on postoperative outcomes.²⁷ In contrast, the present study addresses a distinct research question by examining systemic inflammatory biomarkers and their association with postoperative delirium, with additional stratification by frailty status.

Postoperative delirium is one of the most common postoperative complications in older adults, yet its pathophysiology remains incompletely understood. Growing evidence supports a pivotal role of neuroinflammation, with surgery-induced systemic inflammation shown to activate microglia, disrupt the blood–brain barrier, and promote neuronal dysfunction in both human and animal studies.^28–31 In this context, preoperative inflammatory status, reflected by the LMR, may indicate an individual’s immune-inflammatory resilience to surgical stress.

The different patterns observed across frailty strata likely reflect distinct pathophysiological backgrounds. In robust patients, low baseline systemic inflammation and preserved homeostatic capacity may reduce the relevance of inflammatory biomarkers such as LMR. In frail individuals, the high baseline risk of POD and widespread immune-inflammatory dysregulation, along with restricted LMR variability, may limit the ability of LMR to further stratify risk beyond the already high baseline risk in this subgroup. In contrast, the pre-frail state represents a transitional phase of vulnerability in which physiological reserve is reduced but not yet exhausted, and inflammatory activity is increased but remains modifiable. In this context, inflammatory biomarkers such as LMR may still capture clinically meaningful variations in immune-inflammatory balance that predispose patients to delirium.

Overall, these findings indicate that the prognostic value of LMR for POD may depend on the degree of frailty and underlying physiological reserve. The prominent inverse and nonlinear association observed in pre-frail patients, together with the residual linear effect in frail individuals, suggests that LMR may serve as a useful biomarker for identifying patients who could benefit from perioperative strategies aimed at controlling systemic inflammation to prevent POD.

A lower preoperative LMR indicates a shift toward a pro-inflammatory state, characterized by increased monocyte activity and impaired lymphocyte-mediated immune regulation. Both acute and chronic inflammation can lead to systemic monocytosis.³² Monocytes are recruited from the circulation to sites of inflammation, where they participate directly in inflammatory processes or differentiate into macrophages that sustain inflammatory responses, even in sterile conditions such as atherosclerosis and cardiovascular diseases.^33,34 However, excessive monocytosis during inflammation may hinder the resolution of inflammation and promote tissue damage through persistent immune activation.³² In addition, lymphopenia increases the risk of infection and infection-related mortality in hospitalized patients.³⁵

The LMR has emerged as a promising biomarker reflecting systemic inflammation and immune status, and has been significantly associated with POD in various clinical contexts. For example, a low LMR has been identified as an independent predictor of delirium in older adults with sepsis in ICU.³⁶ The MLR, the inverse of LMR, has been reported as a risk factor for POD in patients undergoing cardiac surgery.¹⁵ Similarly, a study in spinal surgery patients found that a higher MLR was associated with an increased risk of POD.³⁷ Consistent with these observations, our study demonstrates that a lower preoperative LMR is significantly associated with a higher risk of POD, whether analyzed as a continuous or categorical variable.

LMR has also been associated with outcomes in other inflammatory or chronic disease contexts. For example, in non-oncologic conditions, a higher LMR has been independently associated with reduced all-cause mortality among individuals with obese hypertension, irrespective of diabetes comorbidity, highlighting its complex role in systemic inflammation and disease outcomes.³⁸ Collectively, these findings suggest that a low LMR may reflect an elevated inflammatory state and poorer prognosis across multiple disease contexts.

Preoperative low LMR reflects a higher systemic inflammatory burden, which may render individuals more susceptible to the physiological stress and inflammatory stimuli associated with surgery. Surgical trauma activates the innate immune system, leading to the release of pro-inflammatory mediators and immune cell activation, which can disrupt the blood–brain barrier, allowing peripheral inflammatory factors and immune cells to infiltrate the central nervous system. This systemic inflammation triggers neuroinflammatory responses, including tau phosphorylation, microglial and astrocytic activation, and subsequent neuronal dysfunction and cognitive decline, which are pathophysiological processes that contribute to perioperative neurocognitive disorders.^30,31 Systemic inflammation may also accelerate neurodegenerative changes.³⁹ Notably, aging is accompanied by elevated baseline inflammation and microglial priming, predisposing the brain to an exaggerated neuroinflammatory response upon secondary immune challenges.⁴⁰ Animal studies have shown that even minor surgery can elicit an amplified neuroinflammatory response in aged mice.⁴¹ Thus, a low preoperative LMR, as a marker of heightened inflammatory vulnerability, may identify patients at increased risk of exaggerated neuroinflammatory responses and subsequent POD development.

It is essential to acknowledge several limitations in this study. First, although we adjusted for multiple patient- and surgery-related covariates, the influence of unmeasured or unavailable confounding factors cannot be completely excluded. Residual confounding may therefore persist, potentially affecting the validity of our findings. Second, although the PDCEP database was prospectively established, the present study represents a retrospective analysis of prospectively collected data, and therefore potential bias and residual confounding inherent to observational studies cannot be completely excluded. Third, the LMR was calculated based on a single blood test obtained at hospital admission. Given that LMR may fluctuate over time due to disease progression or treatment effects, repeated measurements would likely provide a more dynamic and reliable assessment of its prognostic value. Fourth, categorizing frailty into three groups may have resulted in limited sample sizes within certain population, which could have reduced statistical power and affected the robustness of the stratified analyses. Consequently, while our bootstrap analyses support the stability of the association within the frail subgroup, we caution against overinterpreting this specific finding given the limited sample size. Finally, the study period overlapped with the COVID-19 pandemic. However, information on SARS-CoV-2 infection status and vaccination history was not systematically recorded in the database, which prevented further evaluation of their potential impact on postoperative outcomes. Therefore, these findings should be interpreted cautiously and confirmed in future studies with larger cohorts. If validated, LMR may provide a simple and accessible biomarker to support perioperative risk stratification for postoperative delirium in older patients.

Conclusion

In conclusion, our study demonstrates that preoperative LMR may serve as an accessible biomarker to identify older patients at heightened risk of POD, particularly among those with intermediate frailty. Future studies are warranted to validate these results and to investigate whether interventions targeting systemic inflammation could mitigate POD risk in vulnerable populations.