Associations between polygenic risk scores for Alzheimer's disease and postoperative delirium risk

Abstract

Background

Postoperative delirium (POD) affects up to a third of older surgical patients, leading to significant morbidity, mortality, and potential progression to Alzheimer's disease (AD). Polygenic risk scores (PRS) capture inherited susceptibility to complex diseases, but their relevance to POD is unclear.

Objective

We examined whether higher AD-PRS predict increased POD risk in patients without dementia and whether sleep burden modifies this relationship.

Methods

This study included 345 414 UK Biobank participants (mean [SD] age: 70.1 [7.9], range: 40.4–87.6 years; 54.0% women) to identify new-onset POD, using the International Classification of Disease-10 coding within three days of surgery. Participants with mild cognitive impairment, dementia, or dementia diagnosed within one year of POD were excluded. AD-PRS was calculated as a weighted sum of genetic variants, with scores divided into quartiles due to the absence of standardized thresholds. Covariates included demographics, comorbidities, and lifestyle factors. Cox proportional hazard models were used to evaluate the relationship between AD-PRS and POD risk.

Results

A total of 1610 POD cases were identified. Compared to Q1, individuals in Q3 (HR = 1.23, 95% CI [1.07–1.42], p < 0.01) and Q4 (1.35, [1.18–1.56], p < 0.001) had progressively higher POD risk. Findings were consistent across alternate POD definitions and subgroups defined by sleep burden, sex, age, cardiovascular risk, and inflammatory markers.

Conclusions

Higher AD-PRS is independently associated with greater POD risk in adults without dementia. AD genetic susceptibility may help identify high-risk surgical patients and warrants validation in prospective perioperative cohorts.

Keywords

Alzheimer's disease perioperative period polygenic risk score postoperative delirium

Introduction

Postoperative delirium (POD) is a common and serious complication affecting up to a third of older surgical patients.¹ Characterized by acute disturbances in attention and cognition, POD is associated with prolonged hospitalization,² long-term cognitive decline,³ and increased mortality.⁴ Evidence also suggests POD may accelerate the trajectory toward Alzheimer's disease (AD),^5,6 highlighting the possibility of shared neurobiological pathways. However, mechanisms linking AD risk and POD remain poorly understood.

Genetic susceptibility to AD is well-established, with the apolipoprotein E ε4 (APOE ε4) allele accounting for up to half of genetically attributable risk.^7,8 However, its role in POD risk has been inconsistently demonstrated. Some studies have found associations between APOE ε4 and POD risk,^9,10 whereas others—particularly among surgical patients without dementia—have not.^11,12 This raises questions about whether broader genetic factors beyond a single locus would provide more insight into delirium vulnerability. Polygenic risk scores (PRS) aggregate the effects of numerous genetic variants identified through genome-wide association studies (GWAS), thereby capturing cumulative inherited risk.¹³ Scores have been calculated for a wide range of complex disease phenotypes, including AD.¹⁴ An AD-PRS therefore provides a more comprehensive measure of genetic susceptibility, yet its relationship with POD has not been systematically examined.

Beyond genetic predisposition, other characteristics may modify delirium risk. For example, sleep disturbance has been independently linked to both AD and POD and may represent a modifiable vulnerability factor.^15,16 Likewise, demographic and clinical factors such as age, sex, cardiovascular risk burden, and systemic inflammation (e.g., C-reactive protein [CRP]) may influence susceptibility.¹⁰ Understanding these interactions with AD-PRS could help identify high-risk subgroups for targeted prevention, with prehabilitation (e.g., sleep optimization),¹⁷ potentially reducing POD incidence.

In this study, we investigated whether a higher AD-PRS, reflecting broader genetic liability to AD, is associated with increased risk of POD in a large cohort of middle-aged to older adults without known dementia from the UK Biobank. Participants were free of mild cognitive impairment (MCI) or dementia at the time of surgery and during the subsequent year. We also examined whether the association between AD-PRS and POD differed across subgroups defined by sleep burden, age (<75 versus ≥75 years), sex, cardiovascular risk, and levels of CRP. We hypothesized that higher AD-PRS would be associated with greater POD risk and that this association would be amplified among individuals with poor sleep.

Methods

Study participants

This cohort study utilized data from over 500,000 participants from the UK Biobank, a large community-based cohort recruited between 2006 and 2010. Participants ranged in age from 37 to 70 years (mean age 57 ± 8 years, 54% female) and completed extensive questionnaires on demographics, lifestyle, and medical history, alongside the collection of blood/serum samples for biochemical analysis. Genotyping was performed for most participants, enabling the calculation of the AD-PRS. The analysis included 345,414 participants with available AD-PRS data who were hospitalized at least once after baseline assessment and had no prior diagnosis of any delirium, MCI, or dementia at baseline (Figure 1). This cohort constituted the population at risk for postoperative delirium. Participants were followed longitudinally through linked hospitalization records to identify incident POD, defined as delirium occurring within three days of surgery. Participants with a diagnosis of MCI or dementia within one year following POD were excluded.

Figure 1.

Flowchart of participant selection from the UK Biobank Cohort. Illustration of the inclusion criteria and stepwise exclusions used to identify participants with incident postoperative delirium for the present analysis.

Standard protocol approvals, registrations, and patient consents

The UK Biobank received National Research Ethics Approval, and all participants provided written informed consent. This study was conducted under UK Biobank access number 40556 and received institutional review board approval from Mass General Brigham (#2020P002097).

Derivation of polygenic risk score for Alzheimer's disease

PRS are derived as a weighted sum of risk-associated genetic variants identified from GWAS for complex traits, with higher scores reflecting greater genetic susceptibility.¹⁸ As a PRS specific to AD, AD-PRS quantifies an individual's genetic risk for AD based on relevant variants. In our study, AD-PRS was calculated with PRS-CS, a Bayesian regression method that applies continuous shrinkage priors to GWAS summary statistics to account for linkage disequilibrium (Supplemental Figure 1).¹⁹ Summary statistics were obtained from a recent large AD GWAS¹⁸ using a 1KG-EUR LD-reference panel. Given the absence of standardized classification thresholds, our study categorized AD-PRS into quartiles for analysis.

Assessment of delirium diagnosis

The UK Biobank provided hospitalization records from the National Health Service for participant follow-up. Delirium diagnoses were identified through hospitalization records using International Classification of Disease (ICD-10) code F05, consistent with previous studies^15,20–23. Cases of POD were defined as delirium occurring within three days following a surgical procedure. This was determined by linking UK Biobank records to a surgical database spanning two decades to match surgery dates with the onset of delirium.

Assessment of covariates

We evaluated participants’ medical history using a combination of self-reported data from nurse-led interviews, electronic health records, and baseline medication usage. Covariates were categorized into three main groups: (a) demographics, (b) lifestyle factors including sleep burden, and (c) significant cardiovascular risk factors/comorbidities.

Demographic variables included age, sex, ethnicity, education level, material deprivation, and body mass index (BMI, weight [kg] divided by the height squared [m²]). Age at POD was calculated in years using participants’ date of birth. Sex and ethnicity were self-reported. Since most participants identified as British or White European (94%), ethnicity was classified as European or non-European. Education level was recorded as college-level or not (yes/no). Material deprivation was determined using the Townsend Deprivation Index (TDI), a median score derived from national geographic census data and categorized as high or low deprivation.

Lifestyle factors encompassed alcohol consumption (<4 versus ≥4 drinks per week), depression diagnosis/treatment (“any”, from seeing a psychiatrist, or a self-reported/ICD-10 diagnosis), and sleep burden (score ranging from 0–9), which was previously determined in the same cohort.²⁴ Participants’ sleep duration, excessive daytime sleepiness, insomnia, napping, and chronotype were assessed and assigned a score. These were then summed, with higher total scores indicating greater cumulative sleep burden.

Cardiovascular risk was quantified using a separate composite score (0–5) based on hypertension, high cholesterol, smoking, diabetes, and ischemic heart disease, identified through self-report and ICD-10 codes. Additional comorbidities were assessed using a previously described morbidity burden score,^22,25,26 categorizing individuals as having none (0), modest (1–3), or high (≥4) chronic conditions. This included cancer, neurodegenerative, gastrointestinal, renal, hematological, endocrine, musculoskeletal, connective tissue, and infectious diseases. Cognitive performance was evaluated at baseline using a raw processing speed test, measuring mean reaction time for accurate card matching.²⁷ Serum 25-hydroxyvitamin D (25[OH]D) levels, a proxy for vitamin D status, were categorized as sufficient (>50 nmol/L), low (25–50 nmol/L), or deficient (<25 nmol/L) due to its established link to delirium risk within the same cohort.^20,21 Preoperative serum CRP was included as a marker of systemic inflammation, given that prior work demonstrating its independent association with increased delirium incidence.²⁸

Statistical analysis

Differences between individuals who developed POD and those who remained POD-free during hospitalization were analyzed using chi-squared tests for categorical variables (e.g., sex, ethnicity, comorbidities, smoking) and t-tests or Kruskal–Wallis tests for continuous variables (e.g., age, deprivation index, physical activity, reaction time, CVD risk score). Cox proportional hazards models were used to assess the association between AD-PRS, sleep burden, and time to POD, reporting hazard ratios (HRs) with 95% confidence intervals (CIs). The proportional hazards assumption was evaluated using the global χ² test in R-package cox.zph (survival) incorporating methods described by Grambsch and Therneau.²⁹

The core model (A) adjusted for demographics, including age, sex, BMI, education, deprivation index, ethnicity, and hospitalization history. The lifestyle model (B) further accounted for alcohol use, sleep burden, and depression diagnosis. The final model (C) incorporated additional adjustments for CVD risk score, morbidity burden, cognitive performance, vitamin D levels, and baseline CRP. Sensitivity analysis explored how the relationship between AD-PRS and POD changed when redefining POD at 1, 5, and 7 days after surgery. Time-to-event was defined as the interval between baseline assessment and POD, with delirium-free participants censored as of September 2021, the last available follow-up. Statistical analyses were conducted using JMP Pro (v16, SAS Institute, Cary, NC, USA), with significance set at p < 0.05. Data are accessible through the UK Biobank upon application, and analysis syntax is available upon reasonable request.

Results

Participant characteristics

Approximately 500 000 participants aged 37–70 (57 ± 8 years, 54% women) were enrolled in the UK Biobank cohort. This study included 345 414 participants (mean [SD] age: 70.1 [7.9], range: 40.4–87.6 years; 54.0% women) who had all data available, were hospitalized at least once after the first assessment, and had no prior delirium (Figure 1). Participants were followed for a median of 12.5 years (IQR: 11.8–13.3) after genotyping. During follow-up, 1610 participants developed incident POD (mean age at POD: 73 ± 7 years; range: 32–84 years), with an average time to POD of approximately 10 years.

Table 1 presents the baseline characteristics of participants stratified by quartiles of AD-PRS. Demographic and clinical variables—including age, sex, body mass index, educational attainment, socioeconomic status (TDI), ethnicity, alcohol consumption, sleep burden, depression history, cardiovascular risk, comorbidity burden, cognitive performance, serum vitamin D, and C-reactive protein levels—were broadly comparable across quartiles. The covariate balance supports the internal validity of subsequent analyses, minimizing bias due to baseline differences between genetic risk quartiles.

Table 1.

Baseline characteristics of participants in each AD-PRS quartile.

Covariates	AD-PRS quartile 1Mean (SD), or n (%)^e	AD-PRS quartile 2Mean (SD), or n (%)	AD-PRS quartile 3Mean (SD), or n (%)	AD-PRS quartile 4Mean (SD), or n (%)
Demographics
Age at baseline	58.1 (7.9)	57.9 (7.9)	57.9 (7.9)	57.7 (8.0)
Males	45.8	46.2	46.3	46.0
College attendance	31.5	30.3	29.3	29.6
Ethnic background (European)	94.2	96.2	96.8	90.4
Townsend deprivation index^a	−1.37 (3.02)	−1.35 (3.05)	−1.31 (3.07)	−1.01 (3.28)
Body mass index (kg/m²)	27.5 (4.80)	27.6 (4.84)	27.7 (4.87)	27.8 (4.92)
Lifestyle
Alcohol (≥ 4 drinks/wk)	46.7	47.3	47.1	45.8
Sleep burden	2.52 (1.54)	2.54 (1.55)	2.54 (1.56)	2.58 (1.57)
Depression	7.04	7.66	7.79	7.60
Comorbidities
CVD risk^b (any)	45.2	45.9	46.7	46.4
Morbidity burden (high)	14.2	14.4	14.5	14.4
Cognition (reaction time)^c	562 (118)	562 (117)	562 (118)	568 (123)
Vitamin D^d (deficient)	13.3	12.9	12.9	14.4
C-reactive protein (mg/L)	2.79 (4.53)	2.79 (4.55)	2.73 (4.52)	2.64 (4.50)
POD cases	351 (21.8)	371 (23.0)	433 (26.9)	455 (28.2)

CVD: cardiovascular disease; SD: standard deviation; POD: postoperative delirium.

Higher value = worse deprivation.

CVD risk score: summed hypertension, cholesterol, diabetes mellitus, smoking status, and ischemic heart disease.

Cognition reaction time in milliseconds: average timed tests of symbol matching.

Vitamin D levels: sufficient >50 nmol/L, low 25–50 nmol/L, and deficient <25 nmol/L.

AD-PRS quartiles were constructed using all participants meeting eligibility criteria, with each quartile containing one-quarter of the cohort.

AD-PRS and associations with POD

A stepwise increase in POD risk was observed with higher AD-PRS quartiles (Q) for the core model (Figure 2). Compared to Q1, individuals in Q3 (HR = 1.23, 95% CI [1.07–1.42], p < 0.01) and Q4 (1.35, [1.18–1.56], p < 0.001) had progressively higher POD risk in the fully adjusted model (Table 2).

Figure 2.

Hazard ratios (±95% CI) for incident postoperative delirium by Alzheimer's disease polygenic risk score groups, adjusted for age, sex, education, and ethnicity, using Cox proportional hazards regression models.

Table 2.

Alzheimer's disease polygenic risk score and associations with postoperative delirium.

Model	Hazard ratio [95% CI]	p
Model A
Quartile 2	1.07 [0.92–1.24]	0.36
Quartile 3	1.24 [1.08–1.43]	0.003
Quartile 4	1.36 [1.18–1.56]	<0.0001
Model B
Quartile 2	1.06 [0.92–1.23]	0.41
Quartile 3	1.24 [1.08–1.43]	0.003
Quartile 4	1.36 [1.18–1.56]	<0.0001
Model C
Quartile 2	1.06 [0.91–1.23]	0.43
Quartile 3	1.23 [1.07–1.42]	0.004
Quartile 4	1.35 [1.18–1.56]	<0.0001

95% CI: confidence intervals; HR: hazard ratio.

Cox proportional hazard models examining the association between AD-PRS quartiles (with quartile 1 as reference) and postoperative delirium.

Model A: core model adjusting for age, sex, education, ethnic background, deprivation, and BMI.

Model B: additionally includes alcohol consumption, sleep burden, and depression diagnosis.

Model C: adds on cardiovascular risk, morbidity burden, reaction time, vitamin D levels, and C-reactive protein levels.

The findings remained consistent when redefining POD as delirium occurring within 1, 5, or 7 days following surgery (Supplemental Figure 2). Notably, 91% of POD cases were identified within 1 day postoperatively, and 94% within 3 days, supporting the use of a 3-day postoperative window as the primary definition of POD in this study.

POD risk by subgroups

The risk of POD was further examined by minimal, mild, and moderate/severe sleep burden groups, as well as age (<75 years/≥75 years), sex, cardiovascular risk, and levels of C-reactive protein (Figure 3). Our study found AD-PRS to be equally predictive for POD risk across the three poor sleep burden groups. It also demonstrated similar predictive value across other subgroups, including by sex, age (<75 versus ≥75 years), presence or absence of cardiovascular risk, and levels of C-reactive protein (above- versus below-average).

Figure 3.

Forest plot of hazard ratios and 95% confidence intervals for higher Alzheimer's disease polygenic risk score (versus lower) predicting incident postoperative delirium based on subgroups of participants by age, sex, poor sleep burden, cardiovascular (CVD) risk and C-reactive protein (CRP) levels.

Discussion

In this cohort of 345,414 community-based UK Biobank participants, we found that those in the third and fourth quartiles of AD-PRS were at 23% and 35% higher risk of developing POD, compared with those in the lowest quartile. The association was stepwise across quartiles and remained significant after adjusting for sociodemographic characteristics, comorbidities, and biomarkers of inflammation and cognition. Sensitivity analyses using alternative POD definitions yielded consistent results. Subgroup analyses further showed that this relationship was not modified by age, sex, cardiovascular risk, CRP, or sleep burden. These findings underscore a potential role for AD genetic liability in delirium pathogenesis, even among individuals without clinically manifest MCI/dementia.

Despite the global impact of postoperative delirium, its pathophysiology remains poorly understood. Prior studies have primarily examined the role of APOE ε4, which is inconsistently associated with POD risk in surgical cohorts^9–12. By demonstrating a graded, stepwise increase in POD risk across AD-PRS quartiles, our results suggest that genetic liability to AD extends beyond APOE and may act as a broader marker of impaired neural resilience to perioperative stressors. This aligns with previous work indicating that higher AD-PRS is associated with accelerated cognitive decline and reduced brain reserve, even in cognitively normal individuals,^30,31 and supports the hypothesis that POD may, in part, reflect the unmasking of latent neurodegenerative processes.^32,33

While prior neuroimaging and fluid biomarker studies have found associations between AD-related pathology and POD,^34,35 our results offer genetic corroboration at the population level. This association persisted after adjusting for CRP, cognitive function, sleep burden, and comorbidity burden, suggesting an independent neurogenetic contribution. The lack of interaction by age, sex, or cardiovascular risk status suggests that AD-PRS exerts its effect independently of traditional risk stratifiers. This contrasts with some earlier studies which suggested that the effect of APOE ε4 on delirium risk is strongest among older adults or those with preexisting cognitive impairment.^36,37 The broader polygenic architecture captured by AD-PRS may explain these differences, as it encompasses both early-onset and late-onset genetic variants contributing to neuroinflammation, synaptic dysfunction, and neurovascular integrity.³⁸

Our findings help clarify the biological heterogeneity of POD, suggesting that genetic liability to neurodegeneration identifies a subset of patients prone to delirium via AD-related pathways. This aligns with prior cerebrospinal fluid and plasma biomarker studies demonstrating elevated tau and other AD-related markers in delirium, even in the absence of overt dementia.³³ This may offer a partial explanation for why POD is a strong predictor of long-term cognitive decline and dementia. The presence of high AD-PRS in affected individuals supports the interpretation that POD, for a subset of patients, reflects early neurodegenerative processes rather than a reversible complication. Integrating genetic classification with perioperative biomarker assays³³ (e.g., plasma tau, neurofilament light chain, and inflammatory cytokines) may refine prediction models, distinguish neurodegeneration-driven delirium from other subtypes, and support development of mechanism-based interventions. Our study also extends prior work linking AD genetic risk to perioperative neurocognitive disorders.³⁹ By leveraging the UK Biobank's large surgical cohort with refined exclusions for preexisting dementia/MCI, we strengthened causal inference and highlighted POD as a clinically meaningful phenotype. Unlike other delirium subtypes, POD often arises in elective surgery, a setting where proactive intervention is possible. Recognizing genetic and modifiable contributors may enable targeted prehabilitation strategies (e.g., sleep optimization, cardiovascular risk control, inflammation management) and supports viewing POD as an early biomarker of future neurocognitive trajectory rather than a transient syndrome.

This study leveraged data from the UK Biobank, a large and well-characterized population-based cohort with comprehensive genotyping, detailed clinical and lifestyle information, and long-term follow-up. The availability of over 500,000 participants, from which 1610 individuals with POD were identified, allowed us to investigate this rare yet clinically significant outcome with adequate statistical power. By excluding individuals with known MCI and dementia at baseline or those diagnosed within one year following delirium onset, we reduced the risk of misclassifying prodromal dementia as POD. Our use of a temporally defined POD window (within three days of surgery) and linkage to surgical records enhanced diagnostic precision, reducing misclassification and further clarifying the relationship between POD and AD genetic risk.

Limitations

Several limitations must be acknowledged. First, this was an observational study using retrospective linkage, limiting causal inference. POD diagnoses were based on ICD-10 codes, which, while specific, have limited sensitivity and likely underestimate incidence, particularly for hypoactive cases. Such misclassification may bias associations toward the null. Second, undiagnosed cognitive impairment at baseline may have contributed to associations despite our exclusions. Third, detailed perioperative, postoperative, and inpatient clinical data were not systematically captured in the UK Biobank. As a result, surgical characteristics, anesthetic factors, perioperative physiology, postoperative complications, and biochemical measures (e.g., vitamin B12 status, glycemic control, and thyroid function) could not be incorporated, leaving the possibility of residual confounding. Fourth, the UK Biobank is predominantly of European ancestry, and the AD-PRS was derived and validated in this population; generalizability to other ancestries is uncertain. Finally, as larger AD GWAS become available, refined polygenic models may alter effect size estimates and predictive performance.

Conclusion

In summary, higher AD-PRS is associated with an increased risk of POD in adults without known dementia, with a stepwise rise in risk across AD-PRS quartiles. This association remained robust after adjusting for sociodemographic, cardiovascular risk, morbidity, and cognitive function, and was consistent across subgroups defined by sleep burden, age, sex, cardiovascular risk, and levels of CRP. These findings support the role of polygenic AD risk in shaping delirium vulnerability and point to the potential utility of AD-PRS as a preoperative risk stratification tool. Future research should validate these findings in prospective surgical cohorts with standardized delirium assessments and explore integration of polygenic risk scores with established clinical risk models, incorporating detailed perioperative exposures (e.g., surgical and anesthetic factors, ICU admission, postoperative complications) and relevant biochemical markers (e.g., vitamin B12, glycemic control, thyroid function)^40–44 to inform targeted prevention strategies.

Supplemental Material

sj-docx-1-alz-10.1177_13872877261430955 - Supplemental material for Associations between polygenic risk scores for Alzheimer's disease and postoperative delirium risk

Supplemental material, sj-docx-1-alz-10.1177_13872877261430955 for Associations between polygenic risk scores for Alzheimer's disease and postoperative delirium risk by Yun Jin Chen, Arlen Gaba, Hui-Wen Yang, Matthew Maher, Richa Saxena, Peng Li, Kun Hu and Lei Gao in Journal of Alzheimer's Disease

Footnotes

Acknowledgements

The authors acknowledge the UK Biobank participants and investigators for their valuable contributions.

ORCID iDs

Yun Jin Chen

Arlen Gaba

Hui-Wen Yang

Richa Saxena

Peng Li

Kun Hu

Lei Gao

Ethical considerations

This study was conducted under UK Biobank access number 40556 and received institutional review board approval from Mass General Brigham (#2020P002097). All patient information was de-identified and patient consent was not required. Patient data will not be shared with third parties.

Consent to participate

Not applicable

Consent for publication

Not applicable

Author contribution(s)

Yun Jin Chen: Conceptualization; Data curation; Formal analysis; Investigation; Methodology; Project administration; Writing – original draft; Writing – review & editing.

Arlen Gaba: Formal analysis; Writing – review & editing.

Hui-Wen Yang: Data curation; Writing – review & editing.

Matthew Maher: Data curation; Writing – review & editing.

Richa Saxena: Writing – review & editing.

Peng Li: Writing – review & editing.

Kun Hu: Writing – review & editing.

Lei Gao: Conceptualization; Data curation; Methodology; Supervision; Writing – review & editing.

Funding

The authors disclosed receipt of the following financial support for the research, authorship, and/or publication of this article: This work was supported by the Queen's University Ruth Taylor Research Fund in Brain Disorders to Y.C., the National Institutes of Health [R35GM160254 and R03AG087439 to L.G.], the Patient-Centered Outcomes Research Institute [DE-2023C1–31327 to L.G.], and the Alzheimer's Association Clinician Scientist Fellowship [AACSF-23-1148490 to L.G.].

Declaration of conflicting interests

The authors declared no potential conflicts of interest with respect to the research, authorship, and/or publication of this article.

Data availability statement

Data are available through the UK Biobank upon application (). The syntax for conducting the analysis will be provided to the UK Biobank and will be available upon reasonable request.

Supplemental material

Supplemental material for this article is available online.

References

American Geriatrics Society Expert Panel on Postoperative Delirium in Older Adults. Postoperative delirium in older adults: best practice statement from the American Geriatrics Society. J Am Coll Surg 2015; 220: 136–148.e1.

Kirfel

Guttenthaler

Mayr

, et al. Postoperative delirium is an independent factor influencing the length of stay of elderly patients in the intensive care unit and in hospital. J Anesth 2022; 36: 341–348.

Goldberg

Chen

Wang

, et al. Association of delirium with long-term cognitive decline: a meta-analysis. JAMA Neurol 2020; 77: 1373–1381.

Bai

Liang

Zhang

, et al. Association between postoperative delirium and mortality in elderly patients undergoing hip fractures surgery: a meta-analysis. Osteoporos Int 2020; 31: 317–326.

Fong

Vasunilashorn

Libermann

, et al. Delirium and Alzheimer’s disease: a proposed model for shared pathophysiology. Int J Geriatr Psychiatry 2019; 34: 781–789.

Fong

Inouye

. The inter-relationship between delirium and dementia: the importance of delirium prevention. Nat Rev Neurol 2022; 18: 579–596.

Farrer

Cupples

Haines

, et al. Effects of age, sex, and ethnicity on the association between apolipoprotein E genotype and Alzheimer disease: a meta-analysis. JAMA 1997; 278: 1349–1356.

Raber

Huang

Ashford

. Apoe genotype accounts for the vast majority of AD risk and AD pathology. Neurobiol Aging 2004; 25: 641–650.

Leung

Sands

Wang

, et al. Apolipoprotein E e4 allele increases the risk of early postoperative delirium in older patients undergoing noncardiac surgery. Anesthesiology 2007; 107: 406–411.

10.

Vasunilashorn

Ngo

Inouye

, et al. Apolipoprotein E genotype and the association between C-reactive protein and postoperative delirium: importance of gene-protein interactions. Alzheimers Dement 2020; 16: 572–580.

11.

Vasunilashorn

Ngo

Kosar

, et al.

Does apolipoprotein E genotype increase risk of postoperative delirium?

Am J Geriatr Psychiatry 2015; 23: 1029–1037.

12.

Figueiredo

Devezas

. Bridging the gaps in understanding POD and POCD: a thorough examination of genetic and clinical biomarkers. Perioper Care Oper Room Manag 2024; 35: 100401.

13.

Leonenko

Baker

Stevenson-Hoare

, et al. Identifying individuals with high risk of Alzheimer’s disease using polygenic risk scores. Nat Commun 2021; 12: 4506.

14.

Escott-Price

Shoai

Pither

, et al. Polygenic score prediction captures nearly all common genetic risk for Alzheimer’s disease. Neurobiol Aging 2017; 49: 214.e7–214.e11.

15.

Ulsa

, et al. Association of poor sleep burden in middle age and older adults with risk for delirium during hospitalization. J Gerontol A Biol Sci Med Sci 2022; 77: 507–516.

16.

Fadayomi

Ibala

Bilotta

, et al. A systematic review and meta-analysis examining the impact of sleep disturbance on postoperative delirium. Crit Care Med 2018; 46: e1204–e1212.

17.

Sugg

Gleeson

Baker

, et al. Sleep and circadian biomarkers of postoperative delirium (SLEEP-POD): protocol for a prospective and observational cohort study. BMJ Open 2024; 14: e080796.

18.

Wightman

Jansen

Savage

, et al. A genome-wide association study with 1,126,563 individuals identifies new risk loci for Alzheimer’s disease. Nat Genet 2021; 53: 1276–1282.

19.

Chen

C-Y

, et al. Polygenic prediction via Bayesian regression and continuous shrinkage priors. Nat Commun 2019; 10: 1776.

20.

Bowman

Jones

Pilling

, et al. Vitamin D levels and risk of delirium: a Mendelian randomization study in the UK biobank. Neurology 2019; 92: e1387–e1394.

21.

Pilling

Jones

Masoli

JAH

, et al. Low vitamin D levels and risk of incident delirium in 351,000 older UK Biobank participants. J Am Geriatr Soc 2021; 69: 365–372.

22.

Gao

Gaba

, et al. Heart rate response and recovery during exercise predict future delirium risk—A prospective cohort study in middle- to older-aged adults. J Sport Health Sci 2023; 12: 312–323.

23.

Gao

Gaba

Cui

, et al. Resting heartbeat complexity predicts all-cause and cardiorespiratory mortality in middle- to older-aged adults from the UK Biobank. J Am Heart Assoc 2021; 10: e018483.

24.

Ulsa

, et al. Association of poor sleep burden in middle age and older adults with risk for delirium during hospitalization. J Gerontol A Biol Sci Med Sci 2021; 77: 507–516.

25.

Gao

Gaykova

, et al. Circadian rest-activity rhythms, delirium risk, and progression to dementia. Ann Neurol 2023; 93: 1145–1157.

26.

Gao

Zheng

Baker

, et al. Associations of rest-activity rhythm disturbances with stroke risk and poststroke adverse outcomes. J Am Heart Assoc 2024; 13: e032086.

27.

Fawns-Ritchie

Deary

. Reliability and validity of the UK Biobank cognitive tests. PLoS One 2020; 15: e0231627.

28.

Vasunilashorn

Dillon

Inouye

, et al. High C-reactive protein predicts delirium incidence, duration, and feature severity after major non-cardiac surgery. J Am Geriatr Soc 2017; 65: e109–e116.

29.

Grambsch

Therneau

. Proportional hazards tests and diagnostics based on weighted residuals. Biometrika 1994; 81: 515–526.

30.

Wang

Chibnik

Choi

, et al. Association of genetic risk of Alzheimer’s disease and cognitive function in two European populations. Sci Rep 2025; 15: 6410.

31.

Pettigrew

Nazarovs

Soldan

, et al. Alzheimer’s disease genetic risk and cognitive reserve in relationship to long-term cognitive trajectories among cognitively normal individuals. Alzheimers Res Ther 2023; 15: 66.

32.

Inouye

Marcantonio

Kosar

, et al. The short-term and long-term relationship between delirium and cognitive trajectory in older surgical patients. Alzheimers Dement 2016; 12: 766–775.

33.

Fong

Davis

Growdon

, et al. The interface of delirium and dementia in older persons. Lancet Neurol 2015; 14: 823–832.

34.

Liang

Baldyga

Quan

, et al. Preoperative plasma tau-PT217 and tau-PT181 are associated with postoperative delirium. Ann Surg 2023; 277: e1232–e1238.

35.

Xie

McAuliffe

Swain

, et al. Cerebrospinal fluid Aβ to tau ratio and postoperative cognitive change. Ann Surg 2013; 258: 364–369.

36.

Adamis

Treloar

Martin

, et al. APOE And cytokines as biological markers for recovery of prevalent delirium in elderly medical inpatients. Int J Geriatr Psychiatry 2007; 22: 688–694.

37.

Kim

C-H

Lee

Kim

, et al. APOEε4, in-hospital delirium and long-term cognitive impairment: a longitudinal memory clinic study. Arch Gerontol Geriatr 2024; 116: 105204.

38.

Huq

Fulton-Howard

Riaz

, et al. Polygenic score modifies risk for Alzheimer’s disease in APOE ε4 homozygotes at phenotypic extremes. Alzheimers Dement 2021; 13: e12226.

39.

Thedim

Maher

, et al. Perioperative polygenic and APOE-based genetic risk assessment for neurocognitive disorders: a biobank study. Br J Anaesth Epub ahead of print 24 Jun 2025. DOI: 10.1016/j.bja.2025.05.014

40.

Sevuk

Baysal

, et al. Relationship between cobalamin deficiency and delirium in elderly patients undergoing cardiac surgery. Neuropsychiatr Dis Treat 2015; 11: 2033–2039.

41.

Park

Lee

J-H

, et al. Association of preoperative blood glucose level with delirium after non-cardiac surgery in diabetic patients. Korean J Anesthesiol 2024; 77: 226–235.

42.

Hermanides

Qeva

Preckel

, et al. Perioperative hyperglycemia and neurocognitive outcome after surgery: a systematic review. Minerva Anestesiol 2018; 84: 1178–1188.

43.

Suzuki

Nakanishi

Masuya

, et al. Risk factors for postoperative delirium in patients undergoing orthopedic procedures: a systematic review and meta-analysis. PLoS One 2025; 20: e0321025.

44.

Ling

Howe

Koh

JSB

, et al. Preoperative thyroid dysfunction predicts 30-day postoperative complications in elderly patients with hip fracture. Geriatr Orthop Surg Rehabil 2013; 4: 43–49.

Supplementary Material

Please find the following supplemental material available below.

For Open Access articles published under a Creative Commons License, all supplemental material carries the same license as the article it is associated with.

For non-Open Access articles published, all supplemental material carries a non-exclusive license, and permission requests for re-use of supplemental material or any part of supplemental material shall be sent directly to the copyright owner as specified in the copyright notice associated with the article.

0.00 MB

0.37 MB