Introduction
Delirium is a common and severe neuropsychiatric syndrome that frequently occurs in critically ill patients admitted to the ICU.1 It is characterized by an acute onset of confusion, inattention, and fluctuating levels of consciousness. Delirium affects a substantial proportion of ICU patients, with reported incidence rates ranging from 20% to 80%, depending on the population and underlying comorbidities.2–4
The development of delirium in ICU patients has been associated with numerous adverse outcomes, including increased mortality, prolonged hospital and ICU stays, higher rates of long-term cognitive impairment, and greater healthcare costs.2,5,6 For example, patients who develop delirium are more likely to experience complications such as mechanical ventilation dependence and ICU-acquired weakness, which further exacerbate their prognosis.1,2 In a study involving elderly patients undergoing hip replacement, the follow-up mortality rate was significantly higher in those with both preoperative and postoperative delirium (80%) than in those with only postoperative delirium (38%) and those without delirium (24%), with a p value of 0.02.7 A study pooled data from prospective studies and randomized controlled trials (2015–2020) and demonstrated that ICU delirium was associated with a significant mean increase in length of stay of 4.77 days in the ICU and 6.67 days in the hospital.8 Despite these significant clinical implications, delirium often goes underdiagnosed or misdiagnosed due to its fluctuating and heterogeneous presentation.9
Regarding the pathogenesis of delirium, inflammatory responses triggered by factors such as surgery may promote the occurrence of delirium by affecting the central nervous system, eg, disrupting the blood-brain barrier and causing neurological dysfunction.10 Similarly, inflammatory responses are also an important driving factor in the pathogenesis of postoperative cognitive decline (POCD). The study by Glumac et al found that preoperative administration of dexamethasone can mitigate surgery-induced inflammation to reduce cognitive complications.11 Cotae et al reviewed preventive strategies for early POCD in emergency surgical patients, including anti-inflammatory interventions to mitigate neuroinflammation and reduce incidence.12 Taken together, these findings underscore that inflammatory responses may play a key role in the pathogenesis of delirium, as they do in POCD.
Efforts to predict the occurrence of delirium have identified a wide range of risk factors, including advanced age, pre-existing cognitive impairment, systemic inflammation, hypoalbuminemia, mechanical ventilation, and prolonged ICU stays.6,13,14 Various predictive tools and scoring systems have been developed to identify high-risk patients, but many lack precision and generalizability across diverse ICU populations.1,2,15 While machine learning models and biomarkers have shown promise in improving predictive accuracy, their clinical applicability remains limited due to their complexity and lack of validation in real-world settings.16
Nomograms, as graphical predictive tools, have gained popularity in recent years due to their ability to integrate multiple risk factors into a user-friendly and individualized prediction model.17,18 Several studies have applied nomograms to predict specific ICU complications, such as postoperative delirium following cardiac surgery or delirium in patients with specific conditions like sepsis or acute pancreatitis.2,3,13 However, most existing models are condition-specific and fail to account for the diverse etiologies and presentations of delirium in the general ICU population. Furthermore, few models have undergone external validation, limiting their practical utility.9,15
In this study, we aim to address the gaps in current research by developing and validating a comprehensive nomogram for predicting delirium risk in ICU patients. We hypothesized that integrating clinical, demographic, and laboratory parameters via LASSO and logistic regression would enable the development of a robust nomogram to predict delirium risk in a general ICU population, thereby facilitating timely interventions, improving patient outcomes, and providing a practical tool for clinicians to optimize care for high-risk patients.
Methods
Study Population
This retrospective cohort study enrolled critically ill patients admitted to the intensive care units (ICUs) of Ningbo Medical Center Lihuili Hospital between January 2020 and December 2023. The inclusion and exclusion criteria were designed to ensure the homogeneity and representativeness of the study population. Adult patients aged 18 years or older who were admitted to the ICU with medical, surgical, or trauma-related conditions and had an expected ICU stay exceeding 24 hours were eligible for inclusion. Only patients with available complete clinical and laboratory data and who were able to undergo delirium assessment were included.
Exclusion criteria included patients with pre-existing neurological diseases, such as dementia, Parkinson’s disease, or traumatic brain injury, as well as those with psychiatric disorders, such as schizophrenia or major depressive disorder, that could interfere with the evaluation of delirium. Patients with severe sensory deficits, including blindness or deafness, or significant language barriers were also excluded. Additionally, patients under palliative care or with an expected survival of less than 48 hours, as well as those who declined consent or whose legal representatives refused participation, were excluded from the study.
Data Collection
Comprehensive clinical data were collected for all enrolled patients through electronic medical records and bedside assessments conducted by trained researchers. Variables were selected based on their clinical relevance to delirium pathogenesis or outcomes in ICU settings and data availability from routine electronic medical records to ensure feasibility and generalizability. Demographic information, including age, sex and weight, was recorded to account for basic patient characteristics. Medical history was meticulously documented, including pre-existing comorbidities such as congestive heart failure, hypertension, diabetes, and chronic kidney disease. Laboratory parameters included creatinine, chloride, sodium, glucose, inflammatory markers, electrolyte levels, blood urea nitrogen (BUN), hematocrit, hemoglobin, Potassium, and platelet counts, etc. For biochemical indicators with multiple measurements, both the minimum and maximum values were retained for subsequent analyses. Sequential Organ Failure Assessment (SOFA) scores were also were also gathered to assess disease severity and physiological status.
ICU length of stay, total hospital stay, the presence and severity of delirium and survival status were also recorded. Data collection occurred at predefined time points, including within 24 hours of ICU admission, daily during the ICU stay, and at the onset of delirium when applicable. A dedicated research team ensured the completeness and accuracy of collected data through regular quality checks and cross-referencing medical records.
Delirium Assessment
Delirium was assessed using the Confusion Assessment Method for the ICU (CAM-ICU), a widely recognized and validated tool for diagnosing delirium in critically ill patients. Assessments were conducted twice daily (morning and evening) by trained ICU nurses who had undergone standardized training to ensure consistency and reliability. Delirium was defined as the presence of acute changes or fluctuations in mental status, inattention, and either disorganized thinking or an altered level of consciousness.
Statistical Analysis
Statistical analyses were performed using R software (version 4.2.2) and SPSS (version 22.0). Descriptive statistics were used to summarize the data, with continuous variables expressed as means with standard deviations or medians with interquartile ranges, and categorical variables presented as frequencies and percentages. Univariate analyses were conducted to compare variables between patients in the delirium and non-delirium groups. Continuous variables were analyzed using t-tests or Mann–Whitney U-tests depending on distribution, while categorical variables were compared using chi-square tests or Fisher’s exact tests.
To identify potential predictors of delirium, Least Absolute Shrinkage and Selection Operator (LASSO) regression was applied to reduce dimensionality and select the most relevant variables. The penalty parameter (λ) was determined through cross-validation to minimize binomial deviance, and only significant predictors were retained for further analysis. The selected variables were then included in a multivariate logistic regression model using the Enter method to identify independent risk factors for delirium. Adjusted odds ratios (ORs) and 95% confidence intervals (CIs) were calculated to quantify the strength of associations. In all tests, a two-tailed p < 0.05 was considered to indicate a significant difference.
Based on the results of the logistic regression analysis, a nomogram was constructed to provide individualized risk estimates for delirium. The predictive performance of the nomogram was evaluated using calibration curves, which assessed the agreement between predicted and observed probabilities, and receiver operating characteristic (ROC) curves, with the area under the curve (AUC) quantifying discrimination ability. Decision curve analysis (DCA) was conducted to evaluate the clinical utility of the nomogram by comparing net benefits across different risk thresholds.
Results
Baseline Characteristics of Patients and Univariate Analysis
This study analyzed a total of 964 ICU patients, of whom 186 (19.3%) developed postoperative delirium, while 778 (80.7%) did not (Figure 1). Baseline demographic, clinical, and laboratory characteristics were compared between the two groups, and significant differences were identified by univariate analysis (Table 1).
|
Table 1 Baseline Characteristics of the Patients
|
 |
Figure 1 The flowchart of this study.
|
Univariate analysis revealed several factors significantly associated with the occurrence of postoperative delirium. The two groups exhibited a comparable median age (73 years [IQR: 64–80.75] in the delirium group vs 74 years [IQR: 67–81] in the non-delirium group, p = 0.123). However, a significant difference in gender distribution was observed, with a higher proportion of males in the delirium group (68.8% vs 49.4%, p < 0.001), suggesting that male patients may be at a greater risk of developing delirium.
Among comorbidities, congestive heart failure (p = 0.011), peripheral vascular disease (p = 0.044), and other neurological disorders (p = 0.033) were more prevalent in patients who developed delirium. Fluid and electrolyte disorders (p < 0.001) were also significantly more common in the delirium group. Substance use disorders, including alcohol abuse (p < 0.001) and drug abuse (p < 0.001), were strongly associated with increased risk of delirium. Additionally, gender showed significant differences, with males being more likely to experience delirium (p < 0.001).
Laboratory parameters also demonstrated significant differences. Patients with delirium had lower minimum anion gap (p < 0.001), maximum anion gap (p < 0.001), and minimum creatinine (p < 0.001). Maximum creatinine levels (p < 0.001) and minimum chloride levels (p < 0.001) were significantly different between groups. Among hematologic markers, lower minimum hemoglobin (p = 0.003), minimum hematocrit (p = 0.003), and platelet counts (p = 0.011) were observed in the delirium group. Length of ICU stay (p = 0.006) and overall hospital stay (p < 0.001) were significantly longer in patients with delirium, indicating its impact on resource utilization and outcomes.
LASSO Regression Results
The LASSO regression was performed to select variables with the strongest association with delirium (Figure 2). As λ increased, coefficients of less relevant variables shrank to zero, leaving only significant predictors. The cross-validation curve identified the optimal λ value that minimized binomial deviance. At the optimal λ, 8 variables were selected as potential predictors to be included in the multivariate logistic regression analysis.
 |
Figure 2 Selection of delirium-associated variables using the LASSO regression model. (A) Coefficient path plot of the LASSO regression, illustrating the changes in coefficients of variables across different Log Lambda values. As the regularization strength increases, certain coefficients shrink to zero, enabling variable selection. (B) Binomial deviance of the LASSO regression model as a function of Log Lambda, based on 10-fold cross-validation. The optimal value of the parameter λ is determined using cross-validation.
|
Multivariate Logistic Regression Analysis
Multivariate logistic regression analysis was conducted to identify independent risk factors for delirium in ICU patients. The analysis retained 8 significant predictors: drug abuse, alcohol abuse, male gender, potassium_max, chloride_min, Length of hospital stay, BUN_max and hematocrit_min. The forest plot visually presents the odds ratios and 95% confidence intervals for the variables included in the logistic regression model (Figure 3). It highlights the relative contributions of each factor to the risk of delirium, with drug abuse and alcohol abuse showing the strongest associations.
 |
Figure 3 Forest plot of multivariate logistic regression for delirium-associated factors in ICU patients.
|
Nomogram Model Construction and Evaluation
A nomogram was constructed based on the results of the multivariate logistic regression analysis (Figure 4). The variables included in the model were drug abuse, alcohol abuse, gender, maximum potassium level, minimum chloride level, length of hospital stay, maximum BUN level, and minimum hematocrit level. These variables were selected for their significant contributions to the risk of delirium, as determined by their regression coefficients. Each variable was assigned a specific score, and the total score was calculated by summing these individual contributions. The total score was then mapped to a predicted probability of delirium using the nomogram.
 |
Figure 4 Nomogram constructed based on the results of multivariate logistic regression analysis. Predictors included drug abuse, alcohol abuse, gender, potassium_max, chloride_min, length of hospital stay, BUN_max, and hematocrit_min.
|
Model Performance Evaluation
The receiver operating characteristic (ROC) curve (Figure 5) was used to evaluate the discrimination ability of the nomogram. The area under the ROC curve (AUC) was 0.732 (95% CI: 0.690–0.773), reflecting high predictive accuracy. At the optimal cutoff point of 0.207, the model achieved a sensitivity of 0.618 and a specificity of 0.735, demonstrating a reasonable balance between correctly identifying patients with delirium and avoiding false positives.
 |
Figure 5 Receiver operating characteristic (ROC) curve and area under the ROC curve (AUC) of the nomogram for predicting delirium in ICU patients. ROC: receiver operating characteristic; AUC: area under the ROC curve.
|
The calibration of the nomogram was evaluated using a calibration curve (Figure 6), which compared the predicted probabilities of delirium with the observed probabilities. The “Ideal” line represented perfect calibration, while the “Apparent” and “Bias-corrected” lines illustrated the actual performance of the model. The calibration curve demonstrated good agreement between predicted and observed probabilities, indicating that the nomogram was well-calibrated for predicting delirium in the training cohort.
 |
Figure 6 Calibration curve for assessing the consistency between the predicted probabilities and the actual probabilities of delirium occurrence based on the nomogram.
|
The clinical utility of the nomogram was assessed using decision curve analysis (DCA) (Figure 7). The DCA curve showed that the nomogram provided a higher net benefit across a wide range of risk thresholds compared to the “All” (treat all patients) and “None” (treat no patients) strategies. As shown in the DCA curve of nomogram, a net clinical benefit was achieved when the threshold probability was below 0.8.
 |
Figure 7 DCA curve for evaluating the clinical applicability of the nomogram. DCA: decision curve analysis.
|
Discussion
The main finding of this study is that we developed and validated a robust nomogram incorporating eight independent risk factors—drug abuse, alcohol abuse, male sex, maximum potassium level, minimum chloride level, length of hospital stay, maximum BUN, and minimum hematocrit—for predicting delirium risk in ICU patients, with satisfactory discrimination (AUC = 0.732), calibration, and clinical utility.
The Nomogram developed in this study demonstrated several strengths. By integrating multiple clinical, demographic, and laboratory variables, it provided a comprehensive and individualized risk prediction for ICU delirium. The model exhibited high predictive accuracy, as indicated by the ROC-AUC, and good calibration, as evidenced by the agreement between predicted and observed probabilities. These features underscore the model’s ability to offer precise and reliable predictions, making it a practical tool for predicting postoperative delirium in ICU patients, enabling early identification of high-risk individuals and informing preventive strategies. Moreover, its visual presentation simplifies interpretation for healthcare professionals, facilitating its adoption in routine clinical workflows. Such characteristics align with the growing recognition of nomograms as user-friendly and effective predictive tools in critical care settings.2,19,20
The findings of this study are consistent with previous research on ICU delirium while also offering unique insights. For example, Wang et al developed a nomogram to predict postoperative delirium in cardiac surgery patients, achieving an AUC of 0.852.20 While their model demonstrated higher predictive accuracy, it was specific to a surgical cohort, whereas our model targets a broader ICU population, enhancing its generalizability.
This study identified several independent risk factors for ICU delirium, as detailed above. These findings are consistent with prior research while also offering novel perspectives into the interplay of these factors in delirium pathophysiology, bridging the existing understanding and new observations to deepen our comprehension of this complex syndrome.
Substance abuse has long been recognized as a significant predictor of delirium due to its adverse effects on neurotransmitter regulation and cognitive reserve.21,22 Chronic alcohol use, for example, has been linked to alterations in gamma-aminobutyric acid (GABA) and glutamate pathways, which may predispose patients to delirium in the ICU setting.22 Similarly, drug abuse exacerbates systemic inflammation and compromises neurocognitive function, increasing the risk of delirium.2,19 A study aimed at evaluating the incidence of delirium in patients with blunt traumatic brain injury (TBI) associated with drug or alcohol abuse revealed that individuals testing positive for stimulants (OR: 1.340, P = 0.018), tricyclic antidepressants (OR: 3.107, P = 0.019), or cannabinoids (OR: 1.326, P ≤ 0.001) had a significantly higher risk of delirium.23
The association between male gender and delirium risk has been reported in multiple studies, with hormonal differences, higher rates of comorbidities, and differences in healthcare-seeking behavior proposed as contributing factors.24 Male sex has been identified as a significant risk factor for neuropsychiatric disorders in both humans and animal models, while estrogen may offer protective effects in individuals with potential cognitive impairment (20). A study investigating the influence of sex differences on postoperative delirium (POD) in adult patients undergoing cardiac valve surgery identified male sex as a significant risk factor for POD (OR: 2.213, P = 0.037).25 Almashari et al conducted a retrospective analysis of the incidence and risk factors for POD in elderly patients, including 108 patients (72 males). Their findings revealed significant associations between POD and advanced age, male sex, diabetes, hypertension, congestive heart failure, and chronic kidney disease.26 Our findings align with these observations, further supporting the inclusion of gender as a predictor in delirium risk models.
The brain operates in a highly complex environment, requiring precise regulation of electrolytes.27 Electrolyte imbalances, including elevated potassium and chloride levels, have been identified as significant predictors of delirium. These findings align with studies suggesting that electrolyte disturbances impair neuronal function, contributing to acute cognitive dysfunction.20,28 Moreover, correcting electrolyte imbalances has been shown to shorten the duration of delirium.29 Although these studies emphasize the role of electrolytes as a risk factor for postoperative delirium, the specific impact of different electrolytes remains controversial. Some studies suggest that disturbances in potassium or sodium levels may contribute to postoperative delirium.30 For instance, Wang et al studied 582 patients undergoing orthopedic surgery and found that imbalances in sodium and calcium were independent risk factors for POD.31 In our study, disruptions in potassium and chloride levels were associated with the occurrence of postoperative delirium. Therefore, while electrolyte imbalances are implicated, the exact mechanisms by which specific electrolytes contribute to postoperative delirium remain unclear.
Prolonged hospitalization was strongly associated with delirium, likely due to extended exposure to ICU stressors, sedatives, and invasive procedures.32 These findings are in line with previous research highlighting the cumulative burden of critical illness as a key driver of delirium.19,33 Štubljar et al reported that patients with delirium had significantly longer hospital stays (p = 0.002) and ICU stays (p = 0.032) compared to no n-delirium patients.32
Lower BUN and hematocrit levels were inversely associated with the risk of delirium. In contrast to our findings, some previous studies have demonstrated an association between higher BUN levels and delirium.34,35 For example, a study investigating the relationship between BUN levels and delirium risk in elderly critically ill patients without kidney disease found that the maximum BUN level exhibited the strongest non-linear positive association with the odds of delirium.34 This discrepancy may reflect differences in study populations or methodologies, underscoring the need for further research to elucidate the complex relationships between these variables and the pathophysiology of delirium.
Our findings on hematocrit levels are consistent with previous studies. Jang et al reported significantly lower hematocrit levels in patients with delirium, although hematocrit was not identified as an independent risk factor for delirium.35 In contrast, Krzych et al conducted a large-scale study investigating risk factors for postoperative delirium in cardiac surgery patients, and multivariate analysis identified lower hematocrit as an independent risk factor for delirium.36
Some previous studies have identified risk factors for predicting ICU delirium that differ from those in this study. Al-Hoodar et al identified sepsis and metabolic acidosis as key predictors of delirium in Omani ICU patients.19 In Jang’s study, depression, musculoskeletal disorders, traumatic brain injury, elevated levels of WBC, BUN, AST, and CRP, as well as decreased levels of potassium and phosphorus, were identified as independent risk factors for delirium.35 Similarly, findings from Pan et al indicated that the use of sedatives, length of ICU stay, and physical restraints were independent risk factors for delirium.37 Discrepancies between studies may be attributed to differences in study design, patient demographics, sample size, and data collection methods. These results underscore the importance of considering a wide range of clinical and laboratory variables when assessing delirium risk and provide a robust basis for developing predictive models.
Additionally, some studies have utilized imaging data to predict the risk of delirium, such as white matter changes (WMC) and atrophy.38 However, the feasibility of incorporating such approaches into routine clinical practice may be limited. By contrast, our model focuses on readily available clinical variables, making it more practical for widespread use.
Several limitations of this study must be acknowledged. First, the relatively small sample size and single-center design may limit the generalizability of the findings. Second, the inclusion of ICU patients with highly diverse admission diagnoses (eg, medical, surgical, and trauma-related conditions) introduces heterogeneity, as different underlying diseases are associated with varying baseline risks of delirium. This diversity may affect the model’s performance across specific subgroups, as the pathophysiological mechanisms and risk profiles for delirium can differ substantially between, for example, post-surgical patients and those with acute medical illnesses. Third, the exclusion of certain potential predictors, such as biomarkers of neuroinflammation or genetic predispositions, may have impacted the model’s predictive power. Fourth, the reliance on internal validation without external testing restricts the model’s applicability to broader populations. Finally, the use of clinical and laboratory data at specific time points may not fully capture the dynamic nature of delirium risk in critically ill patients. Addressing these limitations in future studies will be critical to optimizing the nomogram’s utility and effectiveness.
Conclusion
This study developed and validated a nomogram model to predict the risk of delirium in ICU patients. By integrating key clinical, demographic, and laboratory variables, the model demonstrated high predictive accuracy and good calibration, providing a practical tool for individualized risk assessment. The Nomogram is applicable in ICU settings for a broad population of critically ill patients. The findings highlight the nomogram’s value in identifying high-risk patients early, thereby enabling timely preventive interventions and improving patient outcomes.
Data Sharing Statement
The datasets used or analysed during the current study are available from the corresponding authors on reasonable request.
Ethics Approval and Consent to Participate
The present study followed the Declaration of Helsinki. This study was approved by the Ethics Committee of Ningbo Medical Center Lihuili Hospital (No. KY2025SL008-01), and written informed consent was obtained from all subjects participating in the trial, and their information was stored and used for research anonymously. This study was in accordance with the Declaration of Helsinki.
Author Contributions
All authors made a significant contribution to the work reported, whether that is in the conception, study design, execution, acquisition of data, analysis and interpretation, or in all these areas; took part in drafting, revising or critically reviewing the article; gave final approval of the version to be published; have agreed on the journal to which the article has been submitted; and agree to be accountable for all aspects of the work.
Funding
There is no funding to report.
Disclosure
The authors declare no conflicts of interest in this work.
References
1. Xu F, Zhang S, Zhang Y. High level of systemic immune inflammation index elevates delirium risk among patients in intensive care unit. Sci Rep. 2024;14(1):30265. doi:10.1038/s41598-024-81559-9
2. Schreiber N, Eichlseder M, Orlob S, et al. Sex specific differences in short-term mortality after ICU-delirium. Crit Care. 2024;28(1):413. doi:10.1186/s13054-024-05204-7
3. Chen M, Yu M, Zhang D, et al. Effect of prophylactic perphenazine on delirium after extubation in severe acute pancreatitis. Eur J Med Res. 2024;29(1):572. doi:10.1186/s40001-024-02158-y
4. Kim MS, Kim SH. Risk factors for postoperative delirium in patients with cardiac surgery. Sci Prog. 2024;107(3):368504241266362. doi:10.1177/00368504241266362
5. Flaws D, Tronstad O, Fraser JF, et al. Tracking outcomes post intensive care: findings of a longitudinal observational study. Aust Crit Care. 2025;38(3):101164. doi:10.1016/j.aucc.2024.101164
6. Chi R, Perkins AJ, Khalifeh Y, et al. Serum albumin level at intensive care unit admission and delirium duration and severity in critically ill adults. Am J Crit Care. 2024;33(6):412–420. doi:10.4037/ajcc2024650
7. Wu P, Yang Y, Yuan A, Wang Y, Zhang Y. Postoperative delirium increases follow-up mortality following Hip arthroplasty in older patients with femoral neck fracture. Australas J Ageing. 2024;43(4):715–724. doi:10.1111/ajag.13366
8. Dziegielewski C, Skead C, Canturk T, et al. Delirium and associated length of stay and costs in critically ill patients. Crit Care Res Pract. 2021;2021:6612187. doi:10.1155/2021/6612187
9. Tuomisto A, Kennedy P. Improving the recognition and assessment of ICU delirium. J Contin Educ Nurs. 2024;55(11):530–534. doi:10.3928/00220124-20240829-03
10. Fan YY, Luo RY, Wang MT, Yuan CY, Sun YY, Jing JY. Mechanisms underlying delirium in patients with critical illness. Front Aging Neurosci. 2024;16:1446523. doi:10.3389/fnagi.2024.1446523
11. Glumac S, Kardum G, Sodic L, Supe-Domic D, Karanovic N. Effects of dexamethasone on early cognitive decline after cardiac surgery: a randomised controlled trial. Eur J Anaesthesiol. 2017;34(11):776–784. doi:10.1097/EJA.0000000000000647
12. Cotae A-M, Mirea L, Cobilinschi C, Ungureanu R, Grinţescu IM. Early postoperative cognitive decline—are there any preventive strategies for surgical patients in the emergency setting? Signa Vitae. 2024;20(1):1–7. doi:10.22514/sv.2024.001
13. Liao J, Xie C, Shen X, Miao L. Predicting delirium in older adults with community-acquired pneumonia: a retrospective analysis of stress hyperglycemia ratio and its interactions with nutrition and inflammation. Arch Gerontol Geriatr. 2025;129:105658. doi:10.1016/j.archger.2024.105658
14. Pei W, Tan H, Dai T, Liu J, Tang Y, Liu J. Risk factors for postoperative delirium in adult patients undergoing cardiopulmonary bypass in cardiac surgery. Am J Transl Res. 2024;16(9):4751–4760. doi:10.62347/TXAC6999
15. Ma X, Cheng H, Zhao Y, Zhu Y. Prevalence and risk factors of subsyndromal delirium in ICU: a systematic review and meta-analysis. Intensive Crit Care Nurs. 2025;86:103834. doi:10.1016/j.iccn.2024.103834
16. Zhang G, Luo H, Lu X, et al. Machine learning-based identification and validation of amino acid metabolism related genes as novel biomarkers in chronic kidney disease. Heliyon. 2025;11(2):e41872. doi:10.1016/j.heliyon.2025.e41872
17. Lei Y, Zhang N, Liu Y, Du X. A prediction nomogram for residual after negative pressure aspiration for endogenic cesarean scar ectopic pregnancy: a retrospective study. BMC Pregnancy Childbirth. 2025;25(1):107. doi:10.1186/s12884-025-07255-2
18. Wang Y, Wu X, Wang Y, et al. Development and validation of the systemic inflammatory response index-based nomogram for predicting short-term adverse events in patients with acute uncomplicated type b aortic intramural hematoma. J Inflamm Res. 2025;18:1303–1316. doi:10.2147/JIR.S496007
19. Al-Hoodar RK, Lazarus ER, Alomari O, Alzaabi O. Development of a delirium risk predication model among ICU Patients in Oman. Anesthesiol Res Pract. 2022;2022:1449277. doi:10.1155/2022/1449277
20. Guo Y, Li C, Mu Y, Wu T, Lin X. Incidence and Associated factors of postoperative delirium in adults undergoing cardiac surgery with cardiopulmonary bypass: a prospective cohort study. J Clin Nurs. doi:10.1111/jocn.17596
21. Maldonado JR. Delirium pathophysiology: an updated hypothesis of the etiology of acute brain failure. Int J Geriatr Psychiatry. 2018;33(11):1428–1457. doi:10.1002/gps.4823
22. Inouye SK, Westendorp RG, Saczynski JS. Delirium in elderly people. Lancet. 2014;383(9920):911–922. doi:10.1016/S0140-6736(13)60688-1
23. Safdar M, Colosimo C, Khurshid MH, et al. Delirium, and trauma: substance use and incidence of delirium after traumatic brain injury. J Surg Res. 2024;301:45–53. doi:10.1016/j.jss.2024.05.042
24. Klein SL, Flanagan KL. Sex differences in immune responses. Nat Rev Immunol. 2016;16(10):626–638. PMID: 27546235. doi:10.1038/nri.2016.90
25. Wang H, Guo X, Zhu X, et al. Gender differences and postoperative delirium in adult patients undergoing cardiac valve surgery. Front Cardiovasc Med. 2021;8:751421. doi:10.3389/fcvm.2021.751421
26. Almashari Y, Alshaya RA, Alenazi RR, et al. Incidence and risk factors of developing post-operative delirium among elderly patients in a tertiary care hospital: a retrospective chart review. Cureus. 2024;16(7):e65188. doi:10.7759/cureus.65188
27. Diringer M. Neurologic manifestations of major electrolyte abnormalities. Handb Clin Neurol. 2017;141:705–713. doi:10.1016/B978-0-444-63599-0.00038-7
28. Jacoby N. Electrolyte disorders and the nervous system. continuum. Continuum. 2020;26(3):632–658. doi:10.1212/CON.0000000000000872
29. Mei X, Liu YH, Han YQ, Zheng CY. Risk factors, preventive interventions, overlapping symptoms, and clinical measures of delirium in elderly patients. World J Psychiatry. 2023;13(12):973–984. doi:10.5498/wjp.v13.i12.973
30. Koster S, Oosterveld FG, Hensens AG, Wijma A, van der Palen J. Delirium after cardiac surgery and predictive validity of a risk checklist. Ann Thorac Surg. 2008;86(6):1883–1887. doi:10.1016/j.athoracsur.2008.08.020
31. Wang LH, Xu DJ, Wei XJ, Chang HT, Xu GH. Electrolyte disorders and aging: risk factors for delirium in patients undergoing orthopedic surgeries. BMC Psychiatry. 2016;16(1):418. doi:10.1186/s12888-016-1130-0
32. Štubljar D, Štefin M, Tacar MP, Cerović O, Grosek Š. Prolonged hospitalization is a risk factor for delirium onset: one-day prevalence study in Slovenian Intensive Care Units. Acta Clin Croat. 2019;58(2):265–273. doi:10.20471/acc.2019.58.02.09
33. Friol A, Dumas G, Pène F, et al. A multivariable prediction model for invasive pulmonary aspergillosis in immunocompromised patients with acute respiratory failure (IPA-GRRR-OH score). Intensive Care Med. 2025;51(1):72–81. doi:10.1007/s00134-024-07767-z
34. Fang Y, Tang X, Gao Y, et al. Association between blood urea nitrogen and delirium in critically ill elderly patients without kidney diseases: a retrospective study and mendelian randomization analysis. CNS Neurosci Ther. 2025;31(1):e70201. doi:10.1111/cns.70201
35. Jang S, Jung KI, Yoo WK, Jung MH, Ohn SH. Risk factors for delirium during acute and subacute stages of various disorders in patients admitted to rehabilitation units. Ann Rehabil Med. 2016;40(6):1082–1091. doi:10.5535/arm.2016.40.6.1082
36. Krzych LJ, Wybraniec MT, Krupka-Matuszczyk I, et al. Complex assessment of the incidence and risk factors of delirium in a large cohort of cardiac surgery patients: a single-center 6-year experience. Biomed Res Int. 2013;2013:835850. doi:10.1155/2013/835850
37. Pan Y, Yan J, Jiang Z, Luo J, Zhang J, Yang K. Incidence, risk factors, and cumulative risk of delirium among ICU patients: a case-control study. Int J Nurs Sci. 2019;6(3):247–251. doi:10.1016/j.ijnss.2019.05.008
38. Pendlebury ST, Thomson RJ, Welch SJV, Kuker W, Rothwell PM. Oxford Vascular Study. Utility of white matter disease and atrophy on routinely acquired brain imaging for prediction of long-term delirium risk: population-based cohort study. Age Ageing. 2022;51(1):afab200. doi:10.1093/ageing/afab200