Glycemic Variability and ICU Delirium After Abdominal Surgery in Older Adults: Nonlinear Associations in a Retrospective Single-Center Study

Glycemic Variability and ICU Delirium After Abdominal Surgery in Older Adults: Nonlinear Associations in a Retrospective Single-Center Study | Research Square window.SnipcartSettings = { analytics: { enabled: false } }; (function() { var accessVector = localStorage.getItem('access_vector') || ''; window.dataLayer = window.dataLayer || []; if (accessVector) { window.dataLayer.push({ user: { profile: { profileInfo: { snid: accessVector } } } }); } })(); (function(w,d,s,l,i){w[l]=w[l]||[];w[l].push({'gtm.start':new Date().getTime(),event:'gtm.js'});var f=d.getElementsByTagName(s)[0],j=d.createElement(s),dl=l!='dataLayer'?'&l='+l:'';j.async=true;j.src='https://www.googletagmanager.com/gtm.js?id='+i+dl;f.parentNode.insertBefore(j,f);})(window,document,'script','dataLayer','GTM-K279D39R'); Browse Preprints In Review Journals COVID-19 Preprints AJE Video Bytes Research Tools Research Promotion AJE Professional Editing AJE Rubriq About Preprint Platform In Review Editorial Policies Our Team Advisory Board Help Center Sign In Submit a Preprint Cite Share Download PDF Research Article Glycemic Variability and ICU Delirium After Abdominal Surgery in Older Adults: Nonlinear Associations in a Retrospective Single-Center Study Dandan Zhou, Jingjing Zhang, Lulu Gu, Chulin Chen This is a preprint; it has not been peer reviewed by a journal. https://doi.org/ 10.21203/rs.3.rs-9021880/v1 This work is licensed under a CC BY 4.0 License Status: Under Review Version 1 posted 5 You are reading this latest preprint version Abstract Background: ICU delirium is common after abdominal surgery in older adults. Glycemic variability (GV) may better capture dysglycemia-related physiologic stress than mean glucose, but evidence in postoperative ICU populations is limited. Methods: This retrospective single-center study included 473 patients aged ≥65 years admitted to a postoperative ICU after abdominal surgery (January 2022–December 2025). GV was quantified using four metrics: standard deviation (SD), coefficient of variation (GluCV), glycemic lability index (GLI), and mean amplitude of glycemic excursions (MAGE). Multivariable logistic regression and restricted cubic splines (RCS) assessed associations with ICU delirium, adjusting for confounders. Results: Delirium occurred in 122 patients (25.8%). All GV metrics were higher in delirium patients (all P<0.001). In adjusted models, each 1-SD increase in GV was associated with higher delirium risk: GluCV (aOR 3.34, 95%CI 2.39-4.66), SD (2.15, 1.59-2.90), GLI (3.16, 2.09-4.77), and MAGE (2.26, 1.68-3.02). RCS revealed nonlinear associations for SD, GluCV, and GLI (P for nonlinearity <0.05), with risk accelerating beyond moderate GV levels. Conclusions: Higher glycemic variability is independently associated with postoperative ICU delirium in older adults. The nonlinear patterns suggest potential thresholds for risk stratification. Prospective studies are needed to test whether GV-stabilizing strategies reduce delirium. Glycemic variability Delirium Postoperative care Intensive care unit Aged Abdominal surgery Figures Figure 1 Introduction Delirium is a common and consequential neuropsychiatric complication among older adults admitted to the intensive care unit (ICU) after major surgery. In older surgical patients, postoperative delirium has been linked to increased mortality, major complications, prolonged recovery, and sustained functional decline after discharge [1-3]. These adverse outcomes are particularly concerning after abdominal surgery, where older adults frequently experience a convergence of precipitating insults—systemic inflammation, pain and opioid exposure, sedatives, sleep fragmentation, infection, and hemodynamic perturbations—superimposed on age-related vulnerability [4, 5]. Despite widespread implementation of multicomponent prevention bundles, delirium remains common, underscoring the need to refine risk stratification and identify modifiable physiologic markers detectable during routine postoperative ICU care [6-8]. Dysglycemia is a biologically plausible and clinically actionable contributor to delirium risk. Traditional approaches have emphasized mean glucose levels and overt hypo-/hyperglycemia; however, mean values may fail to capture the neurobiologic stress imposed by rapid glucose excursions [9, 10]. Glycemic variability (GV) reflects the amplitude and frequency of glucose fluctuations and has been implicated in oxidative stress, neuroinflammation, cognitive dysfunction, and blood–brain barrier disruption—pathways aligned with contemporary mechanistic models of delirium [9]. From a translational perspective, GV is clinically attractive because it may increase even when average glucose appears acceptable and is influenced by ICU-related factors such as nutrition delivery, insulin titration practices, catecholamine use, infection, renal dysfunction, and corticosteroid exposure. Clinical evidence linking GV to delirium is emerging but remains heterogeneous. Prior studies have varied in patient case-mix (medical vs surgical ICU), timing of glucose assessment (intraoperative vs ICU), delirium ascertainment, and GV operationalization (e.g., SD, coefficient of variation [CV], or lability measures). For example, intraoperative GV assessed by glucose CV has been associated with postoperative delirium after cardiac surgery, suggesting that variability may carry prognostic information beyond mean glucose [11]. In contrast, other ICU-based investigations have observed more nuanced patterns—such as associations with hypoglycemia during delirium episodes in diabetic patients—highlighting that relationships may differ by context and patient phenotype [12]. More recently, studies using dynamic glucose patterns (e.g., trajectories) have further reinforced the concept that time-varying glycemic signals may relate to acute delirium in critical illness [13]. However, evidence specifically focused on older adults admitted to the ICU after abdominal surgery—a high-risk population with distinct perioperative exposures—remains limited. A further limitation in the existing literature is the frequent assumption of linearity between GV and delirium risk [11, 14]. Biologically, it is plausible that delirium risk accelerates beyond a clinically meaningful threshold of variability, which would have direct implications for risk stratification and glycemic management targets. Restricted cubic spline (RCS) methods provide a principled and widely used approach for characterizing dose–response patterns and formally evaluating departures from linearity for continuous exposures. Accordingly, we conducted a retrospective single-center study of older adults admitted to a postoperative ICU after abdominal surgery to: (1) evaluate the association between multiple prespecified GV metrics and ICU delirium; and (2) examine potential nonlinear dose–response patterns using RCS modeling after adjustment for relevant confounders. Methods Study design and setting We conducted a retrospective single-center observational study in a postoperative intensive care unit (ICU). Eligible admissions occurred between January 1, 2022 and December 31, 2025. Patients were identified through the electronic medical record system and ICU clinical documentation. Participants Patients were eligible if they were aged 65 years or older, were admitted to the ICU after abdominal surgery, stayed in the ICU for at least 48 hours, and had complete medical records required for analysis. Patients were excluded if they had a documented history of psychiatric illness or dementia, had experienced any cerebrovascular event within 3 months prior to ICU admission, died during the ICU stay, had long-term systemic corticosteroid use, or had clearly erroneous or internally inconsistent clinical records. Data collection and variables Two trained investigators independently extracted data from the electronic medical record system using a standardized case report form, and all entries were cross-validated; discrepancies were resolved by re-review of source documents. Prespecified variables included demographic characteristics (sex, age, and body mass index), lifestyle history (alcohol use), and baseline comorbidities (hypertension, diabetes mellitus, and respiratory system disease). Preoperative laboratory variables included albumin, hemoglobin, lymphocyte count, C-reactive protein, and D-dimer. Perioperative and ICU-related variables included surgical type, operative duration, intraoperative blood loss, intraoperative red blood cell transfusion, postoperative Acute Physiology and Chronic Health Evaluation II (APACHE II) score, and duration of mechanical ventilation. Glucose data and glycemic variability metrics To ensure consistency of glucose measurement, we included only capillary glucose values obtained using bedside finger-stick glucometers [15]. Routine glucose monitoring followed standard orders at 4-hour intervals; additional glucose tests performed for acute clinical indications were not included to reduce heterogeneity in sampling frequency. For patients whose ICU length of stay exceeded 7 days, only glucose data from the first 7 ICU days were analyzed. Glycemic variability (GV) was quantified using four prespecified metrics calculated from each patient’s included glucose series: glucose standard deviation (SD); glucose coefficient of variation (GluCV), defined as SD divided by mean glucose multiplied by 100% [16]; glycemic lability index (GLI), computed from successive glucose differences scaled by the time interval between adjacent measurements and normalized by the number of monitoring days [17]; and mean amplitude of glycemic excursions (MAGE), defined as the mean absolute amplitude of consecutive glucose excursions exceeding one SD of the individual’s glucose distribution [18]. GV metrics were analyzed as continuous variables; for interpretability across metrics, effect estimates were also expressed per 1-standard deviation increase. Outcome: ICU delirium The primary outcome was ICU delirium, assessed using the Confusion Assessment Method for the ICU (CAM-ICU) [19]. CAM-ICU evaluates acute onset or fluctuating course, inattention, disorganized thinking, and altered level of consciousness, and delirium was diagnosed when the first two features were present together with either disorganized thinking or altered level of consciousness. CAM-ICU assessments were performed by trained bedside nurses twice daily (approximately 08:00 and 18:00), starting on postoperative day 1 and continuing until ICU discharge or inter-hospital transfer. Assessment results were entered prospectively into the electronic medical record at the time of evaluation, and the present study retrospectively extracted these recorded outcomes. Statistical analysis All analyses were performed using SPSS 25.0 and R 4.3.1. Continuous variables were evaluated for normality. Normally distributed variables were summarized as mean ± standard deviation and compared using independent-samples t tests, whereas non-normally distributed variables were summarized as median (interquartile range) and compared using the Mann–Whitney U test. Categorical variables were summarized as counts (percentages) and compared using χ² tests or Fisher’s exact tests, as appropriate. Univariable analyses were performed to screen variables associated with delirium, and variables with P <0.10 in univariable analyses were entered into multivariable binary logistic regression to identify factors independently associated with delirium. To characterize potential nonlinear exposure–response relationships between GV metrics and delirium risk, restricted cubic spline models were fitted after adjustment for covariates, and nonlinearity was assessed using likelihood ratio tests comparing spline models with corresponding linear models. All tests were two-sided, and P <0.05 was considered statistically significant. Restricted cubic splines with knots at the 5th, 35th, 65th, and 95th percentiles were used to assess nonlinearity; the median was used as the reference and nonlinearity was tested by likelihood ratio tests. Sensitivity analyses To evaluate the robustness of the primary findings, several sensitivity analyses were performed. First, the primary multivariable models were additionally adjusted for insulin use during the ICU stay. Second, analyses were repeated after excluding patients with pre-existing diabetes. Third, analyses were repeated after excluding patients who experienced hypoglycemia events. All sensitivity models used the same covariate adjustment strategy as the primary analysis. Post hoc power analysis Because of the retrospective design, a post hoc power analysis was performed using G*Power 3.1.9.7 within a multivariable binary logistic regression framework [20, 21]. A medium effect size was specified as an odds ratio of 2.5, and the expected delirium incidence was set to approximately 25% based on prior literature [22]. With a two-sided α of 0.05 and power (1−β) of 0.90, the minimum required total sample size was 158. The final analytic cohort included 473 patients, exceeding this requirement and supporting the stability of the fitted models. Ethics Approval Statement Ethics approval was obtained from the Ethics Committee of Jinling Hospital, Nanjing University (Approval No. 2022DZKY-036-01). Results Study population During the study period, 473 older adults (≥65 years) admitted to the surgical ICU after abdominal surgery met the eligibility criteria (ICU length of stay ≥48 h and complete records). ICU delirium occurred in 122 patients (25.8%), and 351 (74.2%) did not develop delirium. Baseline characteristics Baseline and perioperative characteristics stratified by delirium status are summarized in Table 1. In general, patients who developed delirium tended to present with a higher-risk clinical profile, including older age, a greater burden of comorbidity (notably diabetes), higher postoperative illness severity, and more frequent exposure to ICU sedative/analgesic medications. Several preoperative laboratory markers and perioperative factors also differed between groups. Detailed distributions and between-group comparisons are provided in Table 1. Table 1. Baseline characteristics of older adults after abdominal surgery according to ICU delirium status (N = 473). Variable No delirium (n=351) Delirium (n=122) P value Male sex 226 (64.4%) 76 (62.3%) 0.760 Age, years 72.00 (69.00–78.00) 75.00 (71.00–81.00) 0.005 Body mass index, kg/m² 22.27 (20.00–25.08) 22.00 (18.99–24.73) 0.133 Alcohol use 114 (32.5%) 30 (24.6%) 0.068 Diabetes mellitus 43 (12.3%) 33 (27.0%) <0.001 Respiratory disease 25 (7.1%) 5 (4.1%) 0.335 Albumin, g/L 36.30 (32.30–40.20) 34.10 (29.32–38.27) <0.001 Hemoglobin, g/L 117.00 (100.00–131.50) 112.00 (88.25–128.75) 0.018 C-reactive protein, mg/L 2.45 (0.60–14.90) 7.90 (1.23–81.25) <0.001 Surgery type * 73 (20.8%) 54 (44.3%) <0.001 Operative duration, h 2.60 (2.00–3.50) 3.00 (2.00–4.00) 0.479 Intraoperative blood loss, mL 50.00 (0.00–150.00) 50.00 (0.00–187.50) 0.559 Intraoperative RBC transfusion† 0.00 (0.00–100.00) 0.00 (0.00–218.75) 0.277 APACHE II Score 0.00 (0.00–0.00) 0.00 (0.00–7.00) <0.001 Mechanical ventilation, h 0.00 (0.00–0.00) 0.00 (0.00–2.00) <0.001 Sedative exposure (any vs none) 99 (28.2%) 96 (78.7%) <0.001 Analgesic exposure (any vs none) 176 (50.1%) 73 (59.8%) 0.082 Insulin use 37 (10.5%) 41 (33.6%) <0.001 Hypoglycemia event 13 (3.7%) 9 (7.4%) 0.158 SD (glycemic variability) 1.05 (0.82–1.35) 1.46 (1.12–2.03) <0.001 GluCV (glycemic variability) 14.43 (11.65–17.56) 20.57 (17.01–24.81) <0.001 GLI (glycemic variability) 0.80 (0.46–1.32) 1.74 (1.00–4.01) <0.001 MAGE (glycemic variability) 1.73 (1.38–2.25) 2.35 (1.86–3.53) <0.001 Notes. Values are presented as mean (SD) for normally distributed continuous variables, median (IQR) for non-normally distributed continuous variables, and n (%) for categorical variables. Between-group comparisons were performed using the independent-samples t test, the Mann–Whitney U test, and the χ² test or Fisher’s exact test, as appropriate. All tests were two-sided, and P < 0.05 was considered statistically significant. Abbreviations. ICU, intensive care unit; BMI, body mass index; IQR, interquartile range; SD, standard deviation; GV, glycemic variability; GluCV, coefficient of variation of glucose; GLI, glycemic lability index; MAGE, mean amplitude of glycemic excursions; CRP, C-reactive protein; APACHE II, Acute Physiology and Chronic Health Evaluation II. Glycemic variability profiles All four prespecified glycemic variability (GV) indices—glucose SD, GluCV, GLI, and MAGE—were higher among patients who developed ICU delirium (all P < 0.001, Table 1), supporting a consistent unadjusted association between greater glucose fluctuation and delirium. Univariable screening and multivariable logistic regression Variables associated with ICU delirium at P < 0.10 in univariable analyses were entered into multivariable logistic regression. In the primary multivariable model with GluCV as the GV exposure, GluCV remained independently associated with ICU delirium after adjustment for prespecified confounders. Specifically, per 1-SD increase in GluCV, the odds of delirium increased by approximately approximately three-fold (adjusted OR [aOR] 3.34, 95% CI 2.50–5.16; P < 0.001; Table 2). Several clinical covariates reflecting illness severity and ICU exposures (e.g., sedation and severity score) also retained independent associations with delirium in the adjusted model (Table 2). Table 2. Multivariable logistic regression model for ICU delirium (primary model including GluCV). Predictor aOR 95% CI P value GluCV (per 1-SD) 3.34 2.39–4.66 <0.001 Age (per 1 year) 1.03 0.99–1.08 0.134 Body mass index (per 1 kg/m²) 0.99 0.95–1.03 0.693 Albumin (per 1 g/L) 0.98 0.94–1.02 0.222 Severity score (per 1 point) 1.07 1.02–1.12 0.003 Mechanical ventilation (per 1 h) 1.04 0.99–1.09 0.082 Female (vs male) 1.00 0.56–1.79 0.999 Diabetes mellitus (yes vs no) 1.42 0.68–2.97 0.348 Sedative exposure (yes vs no) 12.26 6.60–22.78 <0.001 Analgesic exposure (yes vs no) 0.90 0.51–1.59 0.716 Notes. The analysis included 473 patients with complete data on all covariates (122 delirium events). GluCV was modeled per 1-SD increase. Odds ratios are adjusted for all variables listed in the table. All tests were two-sided. Abbreviations. aOR, adjusted odds ratio; CI, confidence interval; GluCV, coefficient of variation of glucose. To assess robustness across different operationalizations of GV, we repeated the adjusted analysis by substituting GluCV with each of the other GV indices (each scaled per 1-SD increase) while keeping the same adjustment strategy. GV remained significantly associated with delirium across all indices: SD (aOR 2.15; P < 0.001), GLI (aOR 2.92; P < 0.001), and MAGE (aOR 2.22; P < 0.001) (Table 3). Taken together, these findings indicate that the GV–delirium association was consistent across complementary GV constructs. Table 3. Adjusted associations of glycemic variability indices with ICU delirium (per 1–SD increase) GV metric (per 1-SD) N Events aOR 95% CI P value SD 473 122 2.15 1.59–2.90 <0.001 GluCV 473 122 3.34 2.39–4.66 <0.001 GLI 473 122 3.16 2.09–4.77 <0.001 MAGE 473 122 2.26 1.68–3.02 <0.001 Notes. Adjusted odds ratios (aORs) and 95% CIs were estimated from separate multivariable logistic regression models in which each GV index (SD, GluCV, GLI, and MAGE) was entered individually (per 1-SD increase), using the same covariate set as the primary model. The analysis included 473 complete cases (122 delirium events). All tests were two-sided. Abbreviations. GV, glycemic variability; SD, standard deviation; GluCV, coefficient of variation of glucose; GLI, glycemic lability index; MAGE, mean amplitude of glycemic excursions; aOR, adjusted odds ratio; CI, confidence interval. Nonlinear dose–response associations To characterize the exposure–response shape, we fitted restricted cubic spline models for each GV index, with knots placed at the 5th, 35th, 65th, and 95th percentiles and the median as the reference. Nonlinearity was evaluated using likelihood ratio tests comparing spline models with corresponding linear-term models. For GluCV, the spline indicated a nonlinear association with delirium risk (P for nonlinearity = 0.044; Figure 2a). Using the median GluCV (15.84%) as the reference, the curve was relatively flat across the lower–mid range and showed a clearer upward gradient beyond approximately 17–18%, corresponding roughly to the upper-middle distribution of observed values. Similarly, nonlinear associations were observed for SD (P = 0.015; Figure 2b) and GLI (P = 0.0016; Figure 2c). For SD, the adjusted curve suggested that risk increased more visibly once SD exceeded approximately 1.2–1.3. For GLI, a comparable pattern was observed, with a steeper risk gradient emerging beyond values around 1.2–1.4. In contrast, MAGE did not demonstrate a statistically significant departure from linearity (P = 0.123; Figure 2d). Nevertheless, the spline visually suggested a monotonic increase in delirium risk at higher MAGE values, while uncertainty widened at distributional extremes where fewer observations contributed to estimation. Sensitivity analyses Sensitivity analyses yielded findings consistent with the primary models (Supplementary Table S1). Associations between each glycemic variability metric and ICU delirium remained materially unchanged after additional adjustment for insulin use, after exclusion of patients with diabetes, and after exclusion of patients experiencing hypoglycemia events. Discussion In this retrospective single-center observational cohort of older adults admitted to a postoperative ICU after abdominal surgery, higher glycemic variability (GV) was consistently associated with ICU delirium across four prespecified metrics (SD, GluCV, GLI, and MAGE). The associations remained robust after adjustment for key clinical covariates, including illness severity, mechanical ventilation, and sedative exposure. Restricted cubic spline analyses further suggested that the relationship between GV and delirium may be nonlinear for some indices, with risk appearing to accelerate beyond moderate levels of variability. The concordance of findings across multiple GV constructs strengthens the evidence that glycemic instability captures a reproducible risk signal relevant to acute brain dysfunction in this high-risk surgical population. The prognostic relevance of GV has been increasingly recognized in critically ill patients, where higher variability predicts mortality independently of mean glucose [23-26]. In perioperative settings, emerging data link glycemic instability to postoperative neurocognitive complications. For instance, intraoperative GV assessed by coefficient of variation has been associated with delirium after cardiac surgery [11, 13], and dynamic glucose trajectories derived from the MIMIC-IV database were related to delirium in mixed ICU populations [7]. However, these studies differ in case mix (cardiac surgery vs. general ICU), timing of glucose assessment (intraoperative vs. ICU stay), and GV metrics. Our study extends this evidence by focusing specifically on older adults after abdominal surgery—a population with distinct perioperative exposures such as prolonged fasting, systemic inflammation, and opioid-based analgesia—and by demonstrating consistent associations across four well-established GV indices. The nonlinear patterns observed for SD, GluCV, and GLI warrant cautious interpretation. While the spline curves suggested a steeper increase in delirium risk beyond certain GV values, these apparent inflection points should not be misconstrued as clinical thresholds. The exploratory nature of the nonlinearity tests, coupled with widening confidence intervals at the tails of the distributions, precludes definitive cut-point identification. Nevertheless, the findings raise the hypothesis that the biologic impact of glucose fluctuations may be amplified once variability surpasses compensatory capacity, potentially through cumulative oxidative stress and endothelial dysfunction [27, 28]. In contrast, the linear pattern for MAGE may reflect its design focus on larger excursions, which could capture a more uniformly distributed risk gradient. Future studies with larger sample sizes and continuous glucose monitoring are needed to validate these shape differences. Several biologic mechanisms may explain the link between glycemic instability and delirium. Rapid glucose fluctuations can provoke oxidative stress, mitochondrial dysfunction, and endothelial activation, all of which are implicated in neuroinflammation and blood–brain barrier disruption [9, 29-32]. In older adults with age-related cerebral vulnerability, these insults may synergize with other perioperative stressors—such as systemic inflammation, sedative exposure, and pain—to trigger acute brain dysfunction. Moreover, GV may serve as a composite marker of physiologic volatility, reflecting underlying processes such as fluctuating insulin resistance, intermittent catecholamine stimulation, or variable nutrition delivery. Thus, GV could integrate both patient-specific pathophysiology and ICU care dynamics, making it a clinically accessible signal of global instability. From a clinical perspective, these findings support the use of GV as a risk stratification tool rather than a direct therapeutic target. Current critical care guidelines recommend moderate glucose control (typically 7.8–10.0 mmol/L) with emphasis on hypoglycemia avoidance, given the harms of intensive insulin therapy in ICU trials [33, 34]. Within this framework, elevated GV may identify patients at heightened delirium risk even when mean glucose is acceptable, prompting earlier deployment of non-pharmacologic delirium-prevention bundles and review of modifiable contributors to glucose swings (e.g., nutrition interruptions, sedative dosing, insulin titration practices). Whether interventions aimed at stabilizing GV—without increasing hypoglycemia—can improve delirium outcomes remains to be tested in prospective studies. Strengths and limitations In interpreting these findings, several methodological features and constraints merit consideration. The cohort comprised a clearly defined high-risk postoperative surgical ICU population, with delirium assessed using a standardized CAM-ICU protocol performed by trained bedside nurses. Glycemic variability was characterized using a prespecified multi-index framework, and associations were directionally consistent across the examined GV metrics. Given the retrospective single-center design, residual confounding remains possible, particularly from time-varying factors such as evolving infection or inflammatory burden, changes in analgesia and sedation depth, and dynamic organ dysfunction. In addition, some ICU exposures included for adjustment (e.g., sedative/analgesic exposure and mechanical ventilation) may reflect both underlying risk and downstream care processes; accordingly, adjusted estimates are most appropriately interpreted as associations conditional on measured care patterns. Glucose values were derived from scheduled capillary testing, which improves measurement consistency but may miss short-lived excursions and could underestimate variability compared with continuous glucose monitoring, especially for indices sensitive to rapid fluctuations. CAM-ICU assessments were performed twice daily, so transient or hypoactive delirium may have been under-recognized; if present, non-differential misclassification would be expected to attenuate associations. Finally, reverse causality cannot be fully excluded: early subsyndromal delirium or prodromal physiologic deterioration may influence nutrition delivery, treatment adjustments, or monitoring intensity, thereby affecting observed GV. Sensitivity analyses recalculating GV using measurements obtained prior to delirium onset were conducted to reduce this concern, although it cannot be eliminated entirely, intermittent sampling may underestimate variability vs CGM Conclusion In older adults admitted to the ICU after abdominal surgery, higher glycemic variability was independently associated with delirium across multiple GV metrics. Nonlinear exposure-response patterns suggest that risk accelerates beyond moderate variability, supporting GV as a potential risk stratification tool. Prospective multicenter studies with continuous glucose monitoring are needed to validate these findings and test whether GV-stabilizing strategies improve delirium outcomes. Declarations Data Availability Statement : The data that support the findings of this study are available from the corresponding author upon reasonable request and subject to ethical approval. Funding Statement : Not applicable. Conflict of Interest Disclosure : The authors declare no conflicts of interest. Ethics Approval Statement: Ethics approval was obtained from the Ethics Committee of Jinling Hospital, Nanjing University (Approval No. 2022DZKY-036-01). Patient Consent Statement : The requirement for informed consent was waived because of the retrospective design and use of de-identified routinely collected data. Permission to Reproduce Material : Not applicable. Clinical trial number: Not applicable. References Shi Z, Mei X, Li C, Chen Y, Zheng H, Wu Y, Zheng H, Liu L, Marcantonio ER, Xie Z et al : Postoperative Delirium Is Associated with Long-term Decline in Activities of Daily Living . Anesthesiology 2019, 131 (3):492-500. Lander HL, Dick AW, Joynt Maddox KE, Oldham MA, Fleisher LA, Mazzeffi M, Lustik SJ, Shang J, Stone PW, Gloff MS et al : Postoperative Delirium in Older Adults Undergoing Noncardiac Surgery . JAMA Network Open 2025, 8 (7):e2519467-e2519467. Panayi AC, Friedrich S, Rühl J, Schaschinger T, Niederegger T, Knoedler L, Knoedler S, Rendenbach C, Treskatsch S, Harhaus L et al : Delirium after surgery: a retrospective study of predictors, complications, and screening patterns in the national surgical quality improvement program . EClinicalMedicine 2025, 90 :103629. Aldecoa C, Bettelli G, Bilotta F, Sanders RD, Aceto P, Audisio R, Cherubini A, Cunningham C, Dabrowski W, Forookhi A et al : Update of the European Society of Anaesthesiology and Intensive Care Medicine evidence-based and consensus-based guideline on postoperative delirium in adult patients . Eur J Anaesthesiol 2024, 41 (2):81-108. Hoogma DF, Milisen K, Rex S, Al Tmimi L: Postoperative delirium: identifying the patient at risk and altering the course: A narrative review . Eur J Anaesthesiol Intensive Care 2023, 2 (3):e0022. Devlin JW, Skrobik Y, Gélinas C, Needham DM, Slooter AJC, Pandharipande PP, Watson PL, Weinhouse GL, Nunnally ME, Rochwerg B et al : Executive Summary: Clinical Practice Guidelines for the Prevention and Management of Pain, Agitation/Sedation, Delirium, Immobility, and Sleep Disruption in Adult Patients in the ICU . Crit Care Med 2018, 46 (9):1532-1548. Zhang S, Han Y, Xiao Q, Li H, Wu Y: Effectiveness of Bundle Interventions on ICU Delirium: A Meta-Analysis . Crit Care Med 2021, 49 (2):335-346. van den Boogaard M, Pickkers P, Slooter AJ, Kuiper MA, Spronk PE, van der Voort PH, van der Hoeven JG, Donders R, van Achterberg T, Schoonhoven L: Development and validation of PRE-DELIRIC (PREdiction of DELIRium in ICu patients) delirium prediction model for intensive care patients: observational multicentre study . Bmj 2012, 344 :e420. Watt C, Sanchez-Rangel E, Hwang JJ: Glycemic Variability and CNS Inflammation: Reviewing the Connection . Nutrients 2020, 12 (12). Suh S, Kim JH: Glycemic Variability: How Do We Measure It and Why Is It Important? Diabetes Metab J 2015, 39 (4):273-282. Choi H, Park CS, Huh J, Koo J, Jeon J, Kim E, Jung S, Kim HW, Lim JY, Hwang W: Intraoperative Glycemic Variability and Mean Glucose are Predictors for Postoperative Delirium After Cardiac Surgery: A Retrospective Cohort Study . Clin Interv Aging 2022, 17 :79-95. van Keulen K, Knol W, Belitser SV, Zaal IJ, van der Linden PD, Heerdink ER, Egberts TCG, Slooter AJC: Glucose variability during delirium in diabetic and non-diabetic intensive care unit patients: A prospective cohort study . PLoS One 2018, 13 (11):e0205637. Guan B, Chen Z, Lin L, Shao F, Huang J, Liu L: Association between group-based trajectory modeling of blood glucose and acute delirium incidence in critically ill patients: a retrospective cohort study based on the MIMIC-IV database . Sci Rep 2025, 15 (1):44676. Lin YJ, Lin LY, Peng YC, Zhang HR, Chen LW, Huang XZ, Chen Q: Association between glucose variability and postoperative delirium in acute aortic dissection patients: an observational study . J Cardiothorac Surg 2021, 16 (1):82. Klonoff DC, Umpierrez GE, Rice MJ: A Milestone in Point of Care Capillary Blood Glucose Monitoring of Critically Ill Hospitalized Patients . J Diabetes Sci Technol 2018, 12 (6):1095-1100. Battelino T, Danne T, Bergenstal RM, Amiel SA, Beck R, Biester T, Bosi E, Buckingham BA, Cefalu WT, Close KL et al : Clinical Targets for Continuous Glucose Monitoring Data Interpretation: Recommendations From the International Consensus on Time in Range . Diabetes Care 2019, 42 (8):1593-1603. Ryan EA, Shandro T, Green K, Paty BW, Senior PA, Bigam D, Shapiro AMJ, Vantyghem M-C: Assessment of the Severity of Hypoglycemia and Glycemic Lability in Type 1 Diabetic Subjects Undergoing Islet Transplantation . Diabetes 2004, 53 (4):955-962. Service FJ, Molnar GD, Rosevear JW, Ackerman E, Gatewood LC, Taylor WF: Mean Amplitude of Glycemic Excursions, a Measure of Diabetic Instability . Diabetes 1970, 19 (9):644-655. Ely EW, Inouye SK, Bernard GR, Gordon S, Francis J, May L, Truman B, Speroff T, Gautam S, Margolin R et al : Delirium in Mechanically Ventilated PatientsValidity and Reliability of the Confusion Assessment Method for the Intensive Care Unit (CAM-ICU) . JAMA 2001, 286 (21):2703-2710. Faul F, Erdfelder E, Lang AG, Buchner A: G*Power 3: a flexible statistical power analysis program for the social, behavioral, and biomedical sciences . Behav Res Methods 2007, 39 (2):175-191. Faul F, Erdfelder E, Buchner A, Lang AG: Statistical power analyses using G*Power 3.1: tests for correlation and regression analyses . Behav Res Methods 2009, 41 (4):1149-1160. Marcantonio ER: Delirium in Hospitalized Older Adults . N Engl J Med 2017, 377 (15):1456-1466. Hryciw BN, Ghossein J, Rochwerg B, Meggison H, Fernando SM, Kyeremanteng K, Tran A, Seely AJE: Glycemic Variability As a Prognostic Factor for Mortality in Patients With Critical Illness: A Systematic Review and Meta-Analysis . Crit Care Explor 2024, 6 (1):e1025. Eslami S, Taherzadeh Z, Schultz MJ, Abu-Hanna A: Glucose variability measures and their effect on mortality: a systematic review . Intensive Care Med 2011, 37 (4):583-593. Krinsley JS: Glycemic variability: a strong independent predictor of mortality in critically ill patients . Crit Care Med 2008, 36 (11):3008-3013. Egi M, Bellomo R, Stachowski E, French CJ, Hart G: Variability of blood glucose concentration and short-term mortality in critically ill patients . Anesthesiology 2006, 105 (2):244-252. Monnier L, Mas E, Ginet C, Michel F, Villon L, Cristol JP, Colette C: Activation of oxidative stress by acute glucose fluctuations compared with sustained chronic hyperglycemia in patients with type 2 diabetes . Jama 2006, 295 (14):1681-1687. Taylor J, Parker M, Casey CP, Tanabe S, Kunkel D, Rivera C, Zetterberg H, Blennow K, Pearce RA, Lennertz RC et al : Postoperative delirium and changes in the blood-brain barrier, neuroinflammation, and cerebrospinal fluid lactate: a prospective cohort study . Br J Anaesth 2022, 129 (2):219-230. Saisho Y: Glycemic variability and oxidative stress: a link between diabetes and cardiovascular disease? Int J Mol Sci 2014, 15 (10):18381-18406. Ceriello A, Esposito K, Piconi L, Ihnat MA, Thorpe JE, Testa R, Boemi M, Giugliano D: Oscillating Glucose Is More Deleterious to Endothelial Function and Oxidative Stress Than Mean Glucose in Normal and Type 2 Diabetic Patients . Diabetes 2008, 57 (5):1349-1354. McKay TB, Khawaja ZQ, Freedman IG, Turco I, Wiredu K, Colecchi T, Akeju O: Exploring the Pathophysiology of Delirium: An Overview of Biomarker Studies, Animal Models, and Tissue-Engineered Models . Anesth Analg 2023, 137 (6):1186-1197. Mattison MLP: Delirium . Ann Intern Med 2020, 173 (7):Itc49-itc64. Duggan EW, Carlson K, Umpierrez GE: Perioperative Hyperglycemia Management: An Update . Anesthesiology 2017, 126 (3):547-560. Bohé J, Abidi H, Brunot V, Klich A, Klouche K, Sedillot N, Tchenio X, Quenot JP, Roudaut JB, Mottard N et al : Individualised versus conventional glucose control in critically-ill patients: the CONTROLING study-a randomized clinical trial . Intensive Care Med 2021, 47 (11):1271-1283. Additional Declarations No competing interests reported. Supplementary Files SupplementaryTableS1.docx Cite Share Download PDF Status: Under Review Version 1 posted Reviewers invited by journal 06 May, 2026 Editor invited by journal 07 Apr, 2026 Editor assigned by journal 08 Mar, 2026 Submission checks completed at journal 08 Mar, 2026 First submitted to journal 03 Mar, 2026 You are reading this latest preprint version Research Square lets you share your work early, gain feedback from the community, and start making changes to your manuscript prior to peer review in a journal. As a division of Research Square Company, we’re committed to making research communication faster, fairer, and more useful. We do this by developing innovative software and high quality services for the global research community. Our growing team is made up of researchers and industry professionals working together to solve the most critical problems facing scientific publishing. Also discoverable on Platform About Our Team In Review Editorial Policies Advisory Board Help Center Resources Author Services Accessibility API Access RSS feed Manage Cookie Preferences © Research Square 2026 | ISSN 2693-5015 (online) Privacy Policy Terms of Service Do Not Sell My Personal Information {"props":{"pageProps":{"initialData":{"identity":"rs-9021880","acceptedTermsAndConditions":true,"allowDirectSubmit":false,"archivedVersions":[],"articleType":"Research Article","associatedPublications":[],"authors":[{"id":640219305,"identity":"f93e73cc-7e36-4f46-8d3d-0bea152be857","order_by":0,"name":"Dandan Zhou","email":"","orcid":"","institution":"Jinling Hospital, Affiliated Hospital of Medical School, Nanjing University","correspondingAuthor":false,"prefix":"","firstName":"Dandan","middleName":"","lastName":"Zhou","suffix":""},{"id":640219306,"identity":"dfd15f89-7e69-437c-8f23-38d7769d053b","order_by":1,"name":"Jingjing Zhang","email":"","orcid":"","institution":"Jinling Hospital, Affiliated Hospital of Medical School, Nanjing University","correspondingAuthor":false,"prefix":"","firstName":"Jingjing","middleName":"","lastName":"Zhang","suffix":""},{"id":640219308,"identity":"3fd0cf06-068a-47fa-8ae0-01f01e2ca14f","order_by":2,"name":"Lulu Gu","email":"","orcid":"","institution":"Jinling Hospital, Affiliated Hospital of Medical School, Nanjing University","correspondingAuthor":false,"prefix":"","firstName":"Lulu","middleName":"","lastName":"Gu","suffix":""},{"id":640219310,"identity":"5fa65176-2d3d-45dd-92c1-01ad32d79f36","order_by":3,"name":"Chulin Chen","email":"data:image/png;base64,iVBORw0KGgoAAAANSUhEUgAAAZAAAAAyAQMAAABI0h/eAAAABlBMVEX///8AAABVwtN+AAAACXBIWXMAAA7EAAAOxAGVKw4bAAAA20lEQVRIiWNgGAWjYBACAwbmBoYPFWx2bOzNBw58+EGUFsYGxhln+JL5eY4lHpzZQ6QWZt42OcaZM3KMD3OwEaHFnP9g48cZZ8yYDW7kfDjMwMMgzy92AL8Wy4aDzRIfKtL4DM683XC4wILBcObsBAIOO9jYIDnjzDFmg+O5Gw7P4GFIMLhNSMthxubfvG3/GTccyHlwmIeNGC3HGNukedvYGGd25DAQqeUMY5vljDNsoEA2AAayBBF+OX/48A1oVD7+8OGHjTy/NAEt6ECCNOWjYBSMglEwCrADALruTgHPm5GNAAAAAElFTkSuQmCC","orcid":"","institution":"Jinling Hospital, Affiliated Hospital of Medical School, Nanjing University","correspondingAuthor":true,"prefix":"","firstName":"Chulin","middleName":"","lastName":"Chen","suffix":""}],"badges":[],"createdAt":"2026-03-03 15:38:05","currentVersionCode":1,"declarations":"","doi":"10.21203/rs.3.rs-9021880/v1","doiUrl":"https://doi.org/10.21203/rs.3.rs-9021880/v1","draftVersion":[],"editorialEvents":[],"editorialNote":"","failedWorkflow":false,"files":[{"id":109331574,"identity":"4ca35095-5bfc-4797-89e9-3b016c5ecba6","added_by":"auto","created_at":"2026-05-15 16:09:46","extension":"jpg","order_by":1,"title":"Figure 1","display":"","copyAsset":false,"role":"figure","size":125771,"visible":true,"origin":"","legend":"\u003cp\u003e\u003cstrong\u003eNonlinear association between GV and ICU delirium risk in adjusted restricted cubic spline models(a. b. c. d)\u003c/strong\u003e\u003c/p\u003e","description":"","filename":"Picture1.jpg","url":"https://assets-eu.researchsquare.com/files/rs-9021880/v1/24f3eac792efe4793f5bdd39.jpg"},{"id":109331733,"identity":"f17ff7d5-7731-4cda-a0f9-9ee34e3f7f29","added_by":"auto","created_at":"2026-05-15 16:10:07","extension":"pdf","order_by":0,"title":"","display":"","copyAsset":false,"role":"manuscript-pdf","size":409627,"visible":true,"origin":"","legend":"","description":"","filename":"manuscript.pdf","url":"https://assets-eu.researchsquare.com/files/rs-9021880/v1/48d9ebde-d575-4d24-aa60-df1fe1498068.pdf"},{"id":109331575,"identity":"05592649-15e3-4ec8-9ea7-f762fc5cc79a","added_by":"auto","created_at":"2026-05-15 16:09:47","extension":"docx","order_by":1,"title":"","display":"","copyAsset":false,"role":"supplement","size":15458,"visible":true,"origin":"","legend":"","description":"","filename":"SupplementaryTableS1.docx","url":"https://assets-eu.researchsquare.com/files/rs-9021880/v1/9ea2ecd0c1fc0cb31fc55830.docx"}],"financialInterests":"No competing interests reported.","formattedTitle":"Glycemic Variability and ICU Delirium After Abdominal Surgery in Older Adults: Nonlinear Associations in a Retrospective Single-Center Study","fulltext":[{"header":"Introduction","content":"\u003cp\u003eDelirium is a common and consequential neuropsychiatric complication among older adults admitted to the intensive care unit (ICU) after major surgery. In older surgical patients, postoperative delirium has been linked to increased mortality, major complications, prolonged recovery, and sustained functional decline after discharge [1-3]. These adverse outcomes are particularly concerning after abdominal surgery, where older adults frequently experience a convergence of precipitating insults\u0026mdash;systemic inflammation, pain and opioid exposure, sedatives, sleep fragmentation, infection, and hemodynamic perturbations\u0026mdash;superimposed on age-related vulnerability [4, 5]. Despite widespread implementation of multicomponent prevention bundles, delirium remains common, underscoring the need to refine risk stratification and identify modifiable physiologic markers detectable during routine postoperative ICU care [6-8].\u003c/p\u003e\n\u003cp\u003eDysglycemia is a biologically plausible and clinically actionable contributor to delirium risk. Traditional approaches have emphasized mean glucose levels and overt hypo-/hyperglycemia; however, mean values may fail to capture the neurobiologic stress imposed by rapid glucose excursions [9, 10]. Glycemic variability (GV) reflects the amplitude and frequency of glucose fluctuations and has been implicated in oxidative stress, neuroinflammation, cognitive dysfunction, and blood\u0026ndash;brain barrier disruption\u0026mdash;pathways aligned with contemporary mechanistic models of delirium [9]. From a translational perspective, GV is clinically attractive because it may increase even when average glucose appears acceptable and is influenced by ICU-related factors such as nutrition delivery, insulin titration practices, catecholamine use, infection, renal dysfunction, and corticosteroid exposure.\u003c/p\u003e\n\u003cp\u003eClinical evidence linking GV to delirium is emerging but remains heterogeneous. Prior studies have varied in patient case-mix (medical vs surgical ICU), timing of glucose assessment (intraoperative vs ICU), delirium ascertainment, and GV operationalization (e.g., SD, coefficient of variation [CV], or lability measures). For example, intraoperative GV assessed by glucose CV has been associated with postoperative delirium after cardiac surgery, suggesting that variability may carry prognostic information beyond mean glucose [11]. In contrast, other ICU-based investigations have observed more nuanced patterns\u0026mdash;such as associations with hypoglycemia during delirium episodes in diabetic patients\u0026mdash;highlighting that relationships may differ by context and patient phenotype [12]. More recently, studies using dynamic glucose patterns (e.g., trajectories) have further reinforced the concept that time-varying glycemic signals may relate to acute delirium in critical illness [13]. However, evidence specifically focused on older adults admitted to the ICU after abdominal surgery\u0026mdash;a high-risk population with distinct perioperative exposures\u0026mdash;remains limited.\u003c/p\u003e\n\u003cp\u003eA further limitation in the existing literature is the frequent assumption of linearity between GV and delirium risk [11, 14]. Biologically, it is plausible that delirium risk accelerates beyond a clinically meaningful threshold of variability, which would have direct implications for risk stratification and glycemic management targets. Restricted cubic spline (RCS) methods provide a principled and widely used approach for characterizing dose\u0026ndash;response patterns and formally evaluating departures from linearity for continuous exposures. Accordingly, we conducted a retrospective single-center study of older adults admitted to a postoperative ICU after abdominal surgery to: (1) evaluate the association between multiple prespecified GV metrics and ICU delirium; and (2) examine potential nonlinear dose\u0026ndash;response patterns using RCS modeling after adjustment for relevant confounders.\u003c/p\u003e"},{"header":"Methods","content":"\u003cp\u003e\u003cstrong\u003eStudy design and setting\u003c/strong\u003e\u003c/p\u003e\n\u003cp\u003eWe conducted a retrospective single-center observational study in a postoperative intensive care unit (ICU). Eligible admissions occurred between January 1, 2022 and December 31, 2025. Patients were identified through the electronic medical record system and ICU clinical documentation.\u003c/p\u003e\n\u003cp\u003e\u003cstrong\u003eParticipants\u003c/strong\u003e\u003c/p\u003e\n\u003cp\u003ePatients were eligible if they were aged 65 years or older, were admitted to the ICU after abdominal surgery, stayed in the ICU for at least 48 hours, and had complete medical records required for analysis. Patients were excluded if they had a documented history of psychiatric illness or dementia, had experienced any cerebrovascular event within 3 months prior to ICU admission, died during the ICU stay, had long-term systemic corticosteroid use, or had clearly erroneous or internally inconsistent clinical records.\u003c/p\u003e\n\u003cp\u003e\u003cstrong\u003eData collection and variables\u003c/strong\u003e\u003c/p\u003e\n\u003cp\u003eTwo trained investigators independently extracted data from the electronic medical record system using a standardized case report form, and all entries were cross-validated; discrepancies were resolved by re-review of source documents. Prespecified variables included demographic characteristics (sex, age, and body mass index), lifestyle history (alcohol use), and baseline comorbidities (hypertension, diabetes mellitus, and respiratory system disease). Preoperative laboratory variables included albumin, hemoglobin, lymphocyte count, C-reactive protein, and D-dimer. Perioperative and ICU-related variables included surgical type, operative duration, intraoperative blood loss, intraoperative red blood cell transfusion, postoperative Acute Physiology and Chronic Health Evaluation II (APACHE II) score, and duration of mechanical ventilation.\u003c/p\u003e\n\u003cp\u003e\u003cstrong\u003eGlucose data and glycemic variability metrics\u003c/strong\u003e\u003c/p\u003e\n\u003cp\u003eTo ensure consistency of glucose measurement, we included only capillary glucose values obtained using bedside finger-stick glucometers [15]. Routine glucose monitoring followed standard orders at 4-hour intervals; additional glucose tests performed for acute clinical indications were not included to reduce heterogeneity in sampling frequency. For patients whose ICU length of stay exceeded 7 days, only glucose data from the first 7 ICU days were analyzed. Glycemic variability (GV) was quantified using four prespecified metrics calculated from each patient\u0026rsquo;s included glucose series: glucose standard deviation (SD); glucose coefficient of variation (GluCV), defined as SD divided by mean glucose multiplied by 100% [16]; glycemic lability index (GLI), computed from successive glucose differences scaled by the time interval between adjacent measurements and normalized by the number of monitoring days [17]; and mean amplitude of glycemic excursions (MAGE), defined as the mean absolute amplitude of consecutive glucose excursions exceeding one SD of the individual\u0026rsquo;s glucose distribution [18]. GV metrics were analyzed as continuous variables; for interpretability across metrics, effect estimates were also expressed per 1-standard deviation increase.\u003c/p\u003e\n\u003cp\u003e\u003cstrong\u003eOutcome: ICU delirium\u003c/strong\u003e\u003c/p\u003e\n\u003cp\u003eThe primary outcome was ICU delirium, assessed using the Confusion Assessment Method for the ICU (CAM-ICU) [19]. CAM-ICU evaluates acute onset or fluctuating course, inattention, disorganized thinking, and altered level of consciousness, and delirium was diagnosed when the first two features were present together with either disorganized thinking or altered level of consciousness. CAM-ICU assessments were performed by trained bedside nurses twice daily (approximately 08:00 and 18:00), starting on postoperative day 1 and continuing until ICU discharge or inter-hospital transfer. Assessment results were entered prospectively into the electronic medical record at the time of evaluation, and the present study retrospectively extracted these recorded outcomes.\u003c/p\u003e\n\u003cp\u003e\u003cstrong\u003eStatistical analysis\u003c/strong\u003e\u003c/p\u003e\n\u003cp\u003eAll analyses were performed using SPSS 25.0 and R 4.3.1. Continuous variables were evaluated for normality. Normally distributed variables were summarized as mean \u0026plusmn; standard deviation and compared using independent-samples \u003cem\u003et\u003c/em\u003e tests, whereas non-normally distributed variables were summarized as median (interquartile range) and compared using the Mann\u0026ndash;Whitney \u003cem\u003eU\u003c/em\u003e test. Categorical variables were summarized as counts (percentages) and compared using \u0026chi;\u0026sup2; tests or Fisher\u0026rsquo;s exact tests, as appropriate. Univariable analyses were performed to screen variables associated with delirium, and variables with \u003cem\u003eP\u003c/em\u003e\u0026lt;0.10 in univariable analyses were entered into multivariable binary logistic regression to identify factors independently associated with delirium. To characterize potential nonlinear exposure\u0026ndash;response relationships between GV metrics and delirium risk, restricted cubic spline models were fitted after adjustment for covariates, and nonlinearity was assessed using likelihood ratio tests comparing spline models with corresponding linear models. All tests were two-sided, and \u003cem\u003eP\u003c/em\u003e\u0026lt;0.05 was considered statistically significant. Restricted cubic splines with knots at the 5th, 35th, 65th, and 95th percentiles were used to assess nonlinearity; the median was used as the reference and nonlinearity was tested by likelihood ratio tests.\u003c/p\u003e\n\u003cp\u003e\u003cstrong\u003eSensitivity analyses\u003c/strong\u003e\u003c/p\u003e\n\u003cp\u003eTo evaluate the robustness of the primary findings, several sensitivity analyses were performed. First, the primary multivariable models were additionally adjusted for insulin use during the ICU stay. Second, analyses were repeated after excluding patients with pre-existing diabetes. Third, analyses were repeated after excluding patients who experienced hypoglycemia events. All sensitivity models used the same covariate adjustment strategy as the primary analysis.\u003c/p\u003e\n\u003cp\u003e\u003cstrong\u003ePost hoc power analysis\u003c/strong\u003e\u003c/p\u003e\n\u003cp\u003eBecause of the retrospective design, a post hoc power analysis was performed using G*Power 3.1.9.7 within a multivariable binary logistic regression framework [20, 21]. A medium effect size was specified as an odds ratio of 2.5, and the expected delirium incidence was set to approximately 25% based on prior literature [22]. With a two-sided \u0026alpha; of 0.05 and power (1\u0026minus;\u0026beta;) of 0.90, the minimum required total sample size was 158. The final analytic cohort included 473 patients, exceeding this requirement and supporting the stability of the fitted models.\u003c/p\u003e\n\u003cp\u003e\u003cstrong\u003eEthics Approval Statement\u003c/strong\u003e\u003c/p\u003e\n\u003cp\u003eEthics approval was obtained from the Ethics Committee of Jinling Hospital, Nanjing University (Approval No. 2022DZKY-036-01).\u003c/p\u003e"},{"header":"Results","content":"\u003cp\u003e\u003cstrong\u003eStudy population\u003c/strong\u003e\u003c/p\u003e\n\u003cp\u003eDuring the study period, 473 older adults (\u0026ge;65 years) admitted to the surgical ICU after abdominal surgery met the eligibility criteria (ICU length of stay \u0026ge;48 h and complete records). ICU delirium occurred in 122 patients (25.8%), and 351 (74.2%) did not develop delirium.\u003c/p\u003e\n\u003cp\u003e\u003cstrong\u003eBaseline characteristics\u003c/strong\u003e\u003c/p\u003e\n\u003cp\u003eBaseline and perioperative characteristics stratified by delirium status are summarized in Table 1. In general, patients who developed delirium tended to present with a higher-risk clinical profile, including older age, a greater burden of comorbidity (notably diabetes), higher postoperative illness severity, and more frequent exposure to ICU sedative/analgesic medications. Several preoperative laboratory markers and perioperative factors also differed between groups. Detailed distributions and between-group comparisons are provided in Table 1.\u003c/p\u003e\n\u003cp\u003e\u003cstrong\u003eTable 1. Baseline characteristics of older adults after abdominal surgery according to ICU delirium status (N = 473).\u003c/strong\u003e\u003c/p\u003e\n\u003ctable border=\"1\" cellspacing=\"3\" cellpadding=\"0\"\u003e\n \u003cthead\u003e\n \u003ctr\u003e\n \u003ctd\u003e\n \u003cp\u003e\u003cstrong\u003eVariable\u003c/strong\u003e\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd\u003e\n \u003cp\u003e\u003cstrong\u003eNo delirium (n=351)\u003c/strong\u003e\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd\u003e\n \u003cp\u003e\u003cstrong\u003eDelirium (n=122)\u003c/strong\u003e\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd\u003e\n \u003cp\u003e\u003cstrong\u003eP value\u003c/strong\u003e\u003c/p\u003e\n \u003c/td\u003e\n \u003c/tr\u003e\n \u003c/thead\u003e\n \u003ctbody\u003e\n \u003ctr\u003e\n \u003ctd\u003e\n \u003cp\u003eMale sex\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd\u003e\n \u003cp\u003e226 (64.4%)\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd\u003e\n \u003cp\u003e76 (62.3%)\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd\u003e\n \u003cp\u003e0.760\u003c/p\u003e\n \u003c/td\u003e\n \u003c/tr\u003e\n \u003ctr\u003e\n \u003ctd\u003e\n \u003cp\u003eAge, years\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd\u003e\n \u003cp\u003e72.00 (69.00\u0026ndash;78.00)\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd\u003e\n \u003cp\u003e75.00 (71.00\u0026ndash;81.00)\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd\u003e\n \u003cp\u003e0.005\u003c/p\u003e\n \u003c/td\u003e\n \u003c/tr\u003e\n \u003ctr\u003e\n \u003ctd\u003e\n \u003cp\u003eBody mass index, kg/m\u0026sup2;\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd\u003e\n \u003cp\u003e22.27 (20.00\u0026ndash;25.08)\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd\u003e\n \u003cp\u003e22.00 (18.99\u0026ndash;24.73)\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd\u003e\n \u003cp\u003e0.133\u003c/p\u003e\n \u003c/td\u003e\n \u003c/tr\u003e\n \u003ctr\u003e\n \u003ctd\u003e\n \u003cp\u003eAlcohol use\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd\u003e\n \u003cp\u003e114 (32.5%)\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd\u003e\n \u003cp\u003e30 (24.6%)\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd\u003e\n \u003cp\u003e0.068\u003c/p\u003e\n \u003c/td\u003e\n \u003c/tr\u003e\n \u003ctr\u003e\n \u003ctd\u003e\n \u003cp\u003eDiabetes mellitus\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd\u003e\n \u003cp\u003e43 (12.3%)\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd\u003e\n \u003cp\u003e33 (27.0%)\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd\u003e\n \u003cp\u003e\u0026lt;0.001\u003c/p\u003e\n \u003c/td\u003e\n \u003c/tr\u003e\n \u003ctr\u003e\n \u003ctd\u003e\n \u003cp\u003eRespiratory disease\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd\u003e\n \u003cp\u003e25 (7.1%)\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd\u003e\n \u003cp\u003e5 (4.1%)\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd\u003e\n \u003cp\u003e0.335\u003c/p\u003e\n \u003c/td\u003e\n \u003c/tr\u003e\n \u003ctr\u003e\n \u003ctd\u003e\n \u003cp\u003eAlbumin, g/L\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd\u003e\n \u003cp\u003e36.30 (32.30\u0026ndash;40.20)\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd\u003e\n \u003cp\u003e34.10 (29.32\u0026ndash;38.27)\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd\u003e\n \u003cp\u003e\u0026lt;0.001\u003c/p\u003e\n \u003c/td\u003e\n \u003c/tr\u003e\n \u003ctr\u003e\n \u003ctd\u003e\n \u003cp\u003eHemoglobin, g/L\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd\u003e\n \u003cp\u003e117.00 (100.00\u0026ndash;131.50)\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd\u003e\n \u003cp\u003e112.00 (88.25\u0026ndash;128.75)\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd\u003e\n \u003cp\u003e0.018\u003c/p\u003e\n \u003c/td\u003e\n \u003c/tr\u003e\n \u003ctr\u003e\n \u003ctd\u003e\n \u003cp\u003eC-reactive protein, mg/L\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd\u003e\n \u003cp\u003e2.45 (0.60\u0026ndash;14.90)\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd\u003e\n \u003cp\u003e7.90 (1.23\u0026ndash;81.25)\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd\u003e\n \u003cp\u003e\u0026lt;0.001\u003c/p\u003e\n \u003c/td\u003e\n \u003c/tr\u003e\n \u003ctr\u003e\n \u003ctd\u003e\n \u003cp\u003eSurgery type *\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd\u003e\n \u003cp\u003e73 (20.8%)\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd\u003e\n \u003cp\u003e54 (44.3%)\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd\u003e\n \u003cp\u003e\u0026lt;0.001\u003c/p\u003e\n \u003c/td\u003e\n \u003c/tr\u003e\n \u003ctr\u003e\n \u003ctd\u003e\n \u003cp\u003eOperative duration, h\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd\u003e\n \u003cp\u003e2.60 (2.00\u0026ndash;3.50)\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd\u003e\n \u003cp\u003e3.00 (2.00\u0026ndash;4.00)\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd\u003e\n \u003cp\u003e0.479\u003c/p\u003e\n \u003c/td\u003e\n \u003c/tr\u003e\n \u003ctr\u003e\n \u003ctd\u003e\n \u003cp\u003eIntraoperative blood loss, mL\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd\u003e\n \u003cp\u003e50.00 (0.00\u0026ndash;150.00)\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd\u003e\n \u003cp\u003e50.00 (0.00\u0026ndash;187.50)\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd\u003e\n \u003cp\u003e0.559\u003c/p\u003e\n \u003c/td\u003e\n \u003c/tr\u003e\n \u003ctr\u003e\n \u003ctd\u003e\n \u003cp\u003eIntraoperative RBC transfusion\u0026dagger;\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd\u003e\n \u003cp\u003e0.00 (0.00\u0026ndash;100.00)\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd\u003e\n \u003cp\u003e0.00 (0.00\u0026ndash;218.75)\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd\u003e\n \u003cp\u003e0.277\u003c/p\u003e\n \u003c/td\u003e\n \u003c/tr\u003e\n \u003ctr\u003e\n \u003ctd\u003e\n \u003cp\u003eAPACHE II Score\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd\u003e\n \u003cp\u003e0.00 (0.00\u0026ndash;0.00)\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd\u003e\n \u003cp\u003e0.00 (0.00\u0026ndash;7.00)\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd\u003e\n \u003cp\u003e\u0026lt;0.001\u003c/p\u003e\n \u003c/td\u003e\n \u003c/tr\u003e\n \u003ctr\u003e\n \u003ctd\u003e\n \u003cp\u003eMechanical ventilation, h\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd\u003e\n \u003cp\u003e0.00 (0.00\u0026ndash;0.00)\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd\u003e\n \u003cp\u003e0.00 (0.00\u0026ndash;2.00)\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd\u003e\n \u003cp\u003e\u0026lt;0.001\u003c/p\u003e\n \u003c/td\u003e\n \u003c/tr\u003e\n \u003ctr\u003e\n \u003ctd\u003e\n \u003cp\u003eSedative exposure (any vs none)\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd\u003e\n \u003cp\u003e99 (28.2%)\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd\u003e\n \u003cp\u003e96 (78.7%)\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd\u003e\n \u003cp\u003e\u0026lt;0.001\u003c/p\u003e\n \u003c/td\u003e\n \u003c/tr\u003e\n \u003ctr\u003e\n \u003ctd\u003e\n \u003cp\u003eAnalgesic exposure (any vs none)\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd\u003e\n \u003cp\u003e176 (50.1%)\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd\u003e\n \u003cp\u003e73 (59.8%)\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd\u003e\n \u003cp\u003e0.082\u003c/p\u003e\n \u003c/td\u003e\n \u003c/tr\u003e\n \u003ctr\u003e\n \u003ctd\u003e\n \u003cp\u003eInsulin use\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd\u003e\n \u003cp\u003e37 (10.5%)\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd\u003e\n \u003cp\u003e41 (33.6%)\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd\u003e\n \u003cp\u003e\u0026lt;0.001\u003c/p\u003e\n \u003c/td\u003e\n \u003c/tr\u003e\n \u003ctr\u003e\n \u003ctd\u003e\n \u003cp\u003eHypoglycemia event\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd\u003e\n \u003cp\u003e13 (3.7%)\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd\u003e\n \u003cp\u003e9 (7.4%)\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd\u003e\n \u003cp\u003e0.158\u003c/p\u003e\n \u003c/td\u003e\n \u003c/tr\u003e\n \u003ctr\u003e\n \u003ctd\u003e\n \u003cp\u003eSD (glycemic variability)\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd\u003e\n \u003cp\u003e1.05 (0.82\u0026ndash;1.35)\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd\u003e\n \u003cp\u003e1.46 (1.12\u0026ndash;2.03)\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd\u003e\n \u003cp\u003e\u0026lt;0.001\u003c/p\u003e\n \u003c/td\u003e\n \u003c/tr\u003e\n \u003ctr\u003e\n \u003ctd\u003e\n \u003cp\u003eGluCV (glycemic variability)\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd\u003e\n \u003cp\u003e14.43 (11.65\u0026ndash;17.56)\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd\u003e\n \u003cp\u003e20.57 (17.01\u0026ndash;24.81)\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd\u003e\n \u003cp\u003e\u0026lt;0.001\u003c/p\u003e\n \u003c/td\u003e\n \u003c/tr\u003e\n \u003ctr\u003e\n \u003ctd\u003e\n \u003cp\u003eGLI (glycemic variability)\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd\u003e\n \u003cp\u003e0.80 (0.46\u0026ndash;1.32)\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd\u003e\n \u003cp\u003e1.74 (1.00\u0026ndash;4.01)\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd\u003e\n \u003cp\u003e\u0026lt;0.001\u003c/p\u003e\n \u003c/td\u003e\n \u003c/tr\u003e\n \u003ctr\u003e\n \u003ctd\u003e\n \u003cp\u003eMAGE (glycemic variability)\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd\u003e\n \u003cp\u003e1.73 (1.38\u0026ndash;2.25)\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd\u003e\n \u003cp\u003e2.35 (1.86\u0026ndash;3.53)\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd\u003e\n \u003cp\u003e\u0026lt;0.001\u003c/p\u003e\n \u003c/td\u003e\n \u003c/tr\u003e\n \u003c/tbody\u003e\n\u003c/table\u003e\n\u003cp\u003e\u003cstrong\u003eNotes.\u0026nbsp;\u003c/strong\u003eValues are presented as mean (SD) for normally distributed continuous variables, median (IQR) for non-normally distributed continuous variables, and n (%) for categorical variables. Between-group comparisons were performed using the independent-samples t test, the Mann\u0026ndash;Whitney U test, and the \u0026chi;\u0026sup2; test or Fisher\u0026rsquo;s exact test, as appropriate. All tests were two-sided, and P \u0026lt; 0.05 was considered statistically significant.\u003c/p\u003e\n\u003cp\u003e\u003cstrong\u003eAbbreviations.\u003c/strong\u003e ICU, intensive care unit; BMI, body mass index; IQR, interquartile range; SD, standard deviation; GV, glycemic variability; GluCV, coefficient of variation of glucose; GLI, glycemic lability index; MAGE, mean amplitude of glycemic excursions; CRP, C-reactive protein; APACHE II, Acute Physiology and Chronic Health Evaluation II.\u003c/p\u003e\n\u003cp\u003e\u003cstrong\u003eGlycemic variability profiles\u003c/strong\u003e\u003c/p\u003e\n\u003cp\u003eAll four prespecified glycemic variability (GV) indices\u0026mdash;glucose SD, GluCV, GLI, and MAGE\u0026mdash;were higher among patients who developed ICU delirium (all P \u0026lt; 0.001, Table 1), supporting a consistent unadjusted association between greater glucose fluctuation and delirium.\u003c/p\u003e\n\u003cp\u003e\u003cstrong\u003eUnivariable screening and multivariable logistic regression\u003c/strong\u003e\u003c/p\u003e\n\u003cp\u003eVariables associated with ICU delirium at P \u0026lt; 0.10 in univariable analyses were entered into multivariable logistic regression. In the primary multivariable model with GluCV as the GV exposure, GluCV remained independently associated with ICU delirium after adjustment for prespecified confounders. Specifically, per 1-SD increase in GluCV, the odds of delirium increased by approximately approximately three-fold (adjusted OR [aOR] 3.34, 95% CI 2.50\u0026ndash;5.16; P \u0026lt; 0.001; Table 2). Several clinical covariates reflecting illness severity and ICU exposures (e.g., sedation and severity score) also retained independent associations with delirium in the adjusted model (Table 2).\u003c/p\u003e\n\u003cp\u003e\u003cstrong\u003eTable 2. Multivariable logistic regression model for ICU delirium (primary model including GluCV).\u003c/strong\u003e\u003c/p\u003e\n\u003ctable border=\"1\" cellspacing=\"3\" cellpadding=\"0\" width=\"579\"\u003e\n \u003cthead\u003e\n \u003ctr\u003e\n \u003ctd\u003e\n \u003cp\u003e\u003cstrong\u003ePredictor\u003c/strong\u003e\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd\u003e\n \u003cp\u003e\u003cstrong\u003eaOR\u003c/strong\u003e\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd\u003e\n \u003cp\u003e\u003cstrong\u003e95% CI\u003c/strong\u003e\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd\u003e\n \u003cp\u003e\u003cstrong\u003eP value\u003c/strong\u003e\u003c/p\u003e\n \u003c/td\u003e\n \u003c/tr\u003e\n \u003c/thead\u003e\n \u003ctbody\u003e\n \u003ctr\u003e\n \u003ctd\u003e\n \u003cp\u003eGluCV (per 1-SD)\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd\u003e\n \u003cp\u003e3.34\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd\u003e\n \u003cp\u003e2.39\u0026ndash;4.66\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd\u003e\n \u003cp\u003e\u0026lt;0.001\u003c/p\u003e\n \u003c/td\u003e\n \u003c/tr\u003e\n \u003ctr\u003e\n \u003ctd\u003e\n \u003cp\u003eAge (per 1 year)\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd\u003e\n \u003cp\u003e1.03\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd\u003e\n \u003cp\u003e0.99\u0026ndash;1.08\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd\u003e\n \u003cp\u003e0.134\u003c/p\u003e\n \u003c/td\u003e\n \u003c/tr\u003e\n \u003ctr\u003e\n \u003ctd\u003e\n \u003cp\u003eBody mass index (per 1 kg/m\u0026sup2;)\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd\u003e\n \u003cp\u003e0.99\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd\u003e\n \u003cp\u003e0.95\u0026ndash;1.03\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd\u003e\n \u003cp\u003e0.693\u003c/p\u003e\n \u003c/td\u003e\n \u003c/tr\u003e\n \u003ctr\u003e\n \u003ctd\u003e\n \u003cp\u003eAlbumin (per 1 g/L)\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd\u003e\n \u003cp\u003e0.98\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd\u003e\n \u003cp\u003e0.94\u0026ndash;1.02\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd\u003e\n \u003cp\u003e0.222\u003c/p\u003e\n \u003c/td\u003e\n \u003c/tr\u003e\n \u003ctr\u003e\n \u003ctd\u003e\n \u003cp\u003eSeverity score (per 1 point)\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd\u003e\n \u003cp\u003e1.07\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd\u003e\n \u003cp\u003e1.02\u0026ndash;1.12\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd\u003e\n \u003cp\u003e0.003\u003c/p\u003e\n \u003c/td\u003e\n \u003c/tr\u003e\n \u003ctr\u003e\n \u003ctd\u003e\n \u003cp\u003eMechanical ventilation (per 1 h)\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd\u003e\n \u003cp\u003e1.04\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd\u003e\n \u003cp\u003e0.99\u0026ndash;1.09\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd\u003e\n \u003cp\u003e0.082\u003c/p\u003e\n \u003c/td\u003e\n \u003c/tr\u003e\n \u003ctr\u003e\n \u003ctd\u003e\n \u003cp\u003eFemale (vs male)\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd\u003e\n \u003cp\u003e1.00\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd\u003e\n \u003cp\u003e0.56\u0026ndash;1.79\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd\u003e\n \u003cp\u003e0.999\u003c/p\u003e\n \u003c/td\u003e\n \u003c/tr\u003e\n \u003ctr\u003e\n \u003ctd\u003e\n \u003cp\u003eDiabetes mellitus (yes vs no)\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd\u003e\n \u003cp\u003e1.42\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd\u003e\n \u003cp\u003e0.68\u0026ndash;2.97\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd\u003e\n \u003cp\u003e0.348\u003c/p\u003e\n \u003c/td\u003e\n \u003c/tr\u003e\n \u003ctr\u003e\n \u003ctd\u003e\n \u003cp\u003eSedative exposure (yes vs no)\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd\u003e\n \u003cp\u003e12.26\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd\u003e\n \u003cp\u003e6.60\u0026ndash;22.78\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd\u003e\n \u003cp\u003e\u0026lt;0.001\u003c/p\u003e\n \u003c/td\u003e\n \u003c/tr\u003e\n \u003ctr\u003e\n \u003ctd\u003e\n \u003cp\u003eAnalgesic exposure (yes vs no)\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd\u003e\n \u003cp\u003e0.90\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd\u003e\n \u003cp\u003e0.51\u0026ndash;1.59\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd\u003e\n \u003cp\u003e0.716\u003c/p\u003e\n \u003c/td\u003e\n \u003c/tr\u003e\n \u003c/tbody\u003e\n\u003c/table\u003e\n\u003cp\u003e\u003cstrong\u003eNotes.\u003c/strong\u003e The analysis included 473 patients with complete data on all covariates (122 delirium events). GluCV was modeled per 1-SD increase. Odds ratios are adjusted for all variables listed in the table. All tests were two-sided.\u003c/p\u003e\n\u003cp\u003e\u003cstrong\u003eAbbreviations.\u003c/strong\u003e aOR, adjusted odds ratio; CI, confidence interval; GluCV, coefficient of variation of glucose.\u003c/p\u003e\n\u003cp\u003eTo assess robustness across different operationalizations of GV, we repeated the adjusted analysis by substituting GluCV with each of the other GV indices (each scaled per 1-SD increase) while keeping the same adjustment strategy. GV remained significantly associated with delirium across all indices: SD (aOR 2.15; P \u0026lt; 0.001), GLI (aOR 2.92; P \u0026lt; 0.001), and MAGE (aOR 2.22; P \u0026lt; 0.001) (Table 3). Taken together, these findings indicate that the GV\u0026ndash;delirium association was consistent across complementary GV constructs.\u003c/p\u003e\n\u003cp\u003e\u003cstrong\u003eTable 3. Adjusted associations of glycemic variability indices with ICU delirium (per 1\u0026ndash;SD increase)\u003c/strong\u003e\u003c/p\u003e\n\u003ctable border=\"1\" cellspacing=\"3\" cellpadding=\"0\" width=\"570\"\u003e\n \u003cthead\u003e\n \u003ctr\u003e\n \u003ctd\u003e\n \u003cp\u003e\u003cstrong\u003eGV metric (per 1-SD)\u003c/strong\u003e\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd\u003e\n \u003cp\u003e\u003cstrong\u003eN\u003c/strong\u003e\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd\u003e\n \u003cp\u003e\u003cstrong\u003eEvents\u003c/strong\u003e\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd\u003e\n \u003cp\u003e\u003cstrong\u003eaOR\u003c/strong\u003e\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd\u003e\n \u003cp\u003e\u003cstrong\u003e95% CI\u003c/strong\u003e\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd\u003e\n \u003cp\u003e\u003cstrong\u003eP value\u003c/strong\u003e\u003c/p\u003e\n \u003c/td\u003e\n \u003c/tr\u003e\n \u003c/thead\u003e\n \u003ctbody\u003e\n \u003ctr\u003e\n \u003ctd\u003e\n \u003cp\u003eSD\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd\u003e\n \u003cp\u003e473\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd\u003e\n \u003cp\u003e122\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd\u003e\n \u003cp\u003e2.15\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd\u003e\n \u003cp\u003e1.59\u0026ndash;2.90\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd\u003e\n \u003cp\u003e\u0026lt;0.001\u003c/p\u003e\n \u003c/td\u003e\n \u003c/tr\u003e\n \u003ctr\u003e\n \u003ctd\u003e\n \u003cp\u003eGluCV\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd\u003e\n \u003cp\u003e473\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd\u003e\n \u003cp\u003e122\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd\u003e\n \u003cp\u003e3.34\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd\u003e\n \u003cp\u003e2.39\u0026ndash;4.66\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd\u003e\n \u003cp\u003e\u0026lt;0.001\u003c/p\u003e\n \u003c/td\u003e\n \u003c/tr\u003e\n \u003ctr\u003e\n \u003ctd\u003e\n \u003cp\u003eGLI\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd\u003e\n \u003cp\u003e473\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd\u003e\n \u003cp\u003e122\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd\u003e\n \u003cp\u003e3.16\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd\u003e\n \u003cp\u003e2.09\u0026ndash;4.77\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd\u003e\n \u003cp\u003e\u0026lt;0.001\u003c/p\u003e\n \u003c/td\u003e\n \u003c/tr\u003e\n \u003ctr\u003e\n \u003ctd\u003e\n \u003cp\u003eMAGE\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd\u003e\n \u003cp\u003e473\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd\u003e\n \u003cp\u003e122\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd\u003e\n \u003cp\u003e2.26\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd\u003e\n \u003cp\u003e1.68\u0026ndash;3.02\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd\u003e\n \u003cp\u003e\u0026lt;0.001\u003c/p\u003e\n \u003c/td\u003e\n \u003c/tr\u003e\n \u003c/tbody\u003e\n\u003c/table\u003e\n\u003cp\u003e\u003cstrong\u003eNotes.\u003c/strong\u003e Adjusted odds ratios (aORs) and 95% CIs were estimated from separate multivariable logistic regression models in which each GV index (SD, GluCV, GLI, and MAGE) was entered individually (per 1-SD increase), using the same covariate set as the primary model. The analysis included 473 complete cases (122 delirium events). All tests were two-sided.\u003c/p\u003e\n\u003cp\u003e\u003cstrong\u003eAbbreviations.\u0026nbsp;\u003c/strong\u003eGV, glycemic variability; SD, standard deviation; GluCV, coefficient of variation of glucose; GLI, glycemic lability index; MAGE, mean amplitude of glycemic excursions; aOR, adjusted odds ratio; CI, confidence interval.\u003c/p\u003e\n\u003cp\u003e\u003cstrong\u003eNonlinear dose\u0026ndash;response associations\u0026nbsp;\u003c/strong\u003e\u003c/p\u003e\n\u003cp\u003eTo characterize the exposure\u0026ndash;response shape, we fitted restricted cubic spline models for each GV index, with knots placed at the 5th, 35th, 65th, and 95th percentiles and the median as the reference. Nonlinearity was evaluated using likelihood ratio tests comparing spline models with corresponding linear-term models.\u003c/p\u003e\n\u003cp\u003eFor GluCV, the spline indicated a nonlinear association with delirium risk (P for nonlinearity = 0.044; Figure 2a). Using the median GluCV (15.84%) as the reference, the curve was relatively flat across the lower\u0026ndash;mid range and showed a clearer upward gradient beyond approximately 17\u0026ndash;18%, corresponding roughly to the upper-middle distribution of observed values.\u003c/p\u003e\n\u003cp\u003eSimilarly, nonlinear associations were observed for SD (P = 0.015; Figure 2b) and GLI (P = 0.0016; Figure 2c). For SD, the adjusted curve suggested that risk increased more visibly once SD exceeded approximately 1.2\u0026ndash;1.3. For GLI, a comparable pattern was observed, with a steeper risk gradient emerging beyond values around 1.2\u0026ndash;1.4.\u003c/p\u003e\n\u003cp\u003eIn contrast, MAGE did not demonstrate a statistically significant departure from linearity (P = 0.123; Figure 2d). Nevertheless, the spline visually suggested a monotonic increase in delirium risk at higher MAGE values, while uncertainty widened at distributional extremes where fewer observations contributed to estimation.\u003c/p\u003e\n\u003cp\u003e\u003cstrong\u003eSensitivity analyses\u003c/strong\u003e\u003c/p\u003e\n\u003cp\u003eSensitivity analyses yielded findings consistent with the primary models (Supplementary Table S1). Associations between each glycemic variability metric and ICU delirium remained materially unchanged after additional adjustment for insulin use, after exclusion of patients with diabetes, and after exclusion of patients experiencing hypoglycemia events.\u003c/p\u003e"},{"header":"Discussion","content":"\u003cp\u003eIn this retrospective single-center observational cohort of older adults admitted to a postoperative ICU after abdominal surgery, higher glycemic variability (GV) was consistently associated with ICU delirium across four prespecified metrics (SD, GluCV, GLI, and MAGE). The associations remained robust after adjustment for key clinical covariates, including illness severity, mechanical ventilation, and sedative exposure. Restricted cubic spline analyses further suggested that the relationship between GV and delirium may be nonlinear for some indices, with risk appearing to accelerate beyond moderate levels of variability. The concordance of findings across multiple GV constructs strengthens the evidence that glycemic instability captures a reproducible risk signal relevant to acute brain dysfunction in this high-risk surgical population.\u003c/p\u003e\n\u003cp\u003eThe prognostic relevance of GV has been increasingly recognized in critically ill patients, where higher variability predicts mortality independently of mean glucose [23-26]. In perioperative settings, emerging data link glycemic instability to postoperative neurocognitive complications. For instance, intraoperative GV assessed by coefficient of variation has been associated with delirium after cardiac surgery \u0026nbsp;[11, 13], and dynamic glucose trajectories derived from the MIMIC-IV database were related to delirium in mixed ICU populations [7]. However, these studies differ in case mix (cardiac surgery vs. general ICU), timing of glucose assessment (intraoperative vs. ICU stay), and GV metrics. Our study extends this evidence by focusing specifically on older adults after abdominal surgery\u0026mdash;a population with distinct perioperative exposures such as prolonged fasting, systemic inflammation, and opioid-based analgesia\u0026mdash;and by demonstrating consistent associations across four well-established GV indices.\u003c/p\u003e\n\u003cp\u003eThe nonlinear patterns observed for SD, GluCV, and GLI warrant cautious interpretation. While the spline curves suggested a steeper increase in delirium risk beyond certain GV values, these apparent inflection points should not be misconstrued as clinical thresholds. The exploratory nature of the nonlinearity tests, coupled with widening confidence intervals at the tails of the distributions, precludes definitive cut-point identification. Nevertheless, the findings raise the hypothesis that the biologic impact of glucose fluctuations may be amplified once variability surpasses compensatory capacity, potentially through cumulative oxidative stress and endothelial dysfunction [27, 28]. In contrast, the linear pattern for MAGE may reflect its design focus on larger excursions, which could capture a more uniformly distributed risk gradient. Future studies with larger sample sizes and continuous glucose monitoring are needed to validate these shape differences.\u003c/p\u003e\n\u003cp\u003eSeveral biologic mechanisms may explain the link between glycemic instability and delirium. Rapid glucose fluctuations can provoke oxidative stress, mitochondrial dysfunction, and endothelial activation, all of which are implicated in neuroinflammation and blood\u0026ndash;brain barrier disruption [9, 29-32]. In older adults with age-related cerebral vulnerability, these insults may synergize with other perioperative stressors\u0026mdash;such as systemic inflammation, sedative exposure, and pain\u0026mdash;to trigger acute brain dysfunction. Moreover, GV may serve as a composite marker of physiologic volatility, reflecting underlying processes such as fluctuating insulin resistance, intermittent catecholamine stimulation, or variable nutrition delivery. Thus, GV could integrate both patient-specific pathophysiology and ICU care dynamics, making it a clinically accessible signal of global instability.\u003c/p\u003e\n\u003cp\u003eFrom a clinical perspective, these findings support the use of GV as a risk stratification tool rather than a direct therapeutic target. Current critical care guidelines recommend moderate glucose control (typically 7.8\u0026ndash;10.0 mmol/L) with emphasis on hypoglycemia avoidance, given the harms of intensive insulin therapy in ICU trials [33, 34]. Within this framework, elevated GV may identify patients at heightened delirium risk even when mean glucose is acceptable, prompting earlier deployment of non-pharmacologic delirium-prevention bundles and review of modifiable contributors to glucose swings (e.g., nutrition interruptions, sedative dosing, insulin titration practices). Whether interventions aimed at stabilizing GV\u0026mdash;without increasing hypoglycemia\u0026mdash;can improve delirium outcomes remains to be tested in prospective studies.\u003c/p\u003e\n\u003cp\u003e\u003cstrong\u003eStrengths and limitations\u003c/strong\u003e\u003c/p\u003e\n\u003cp\u003eIn interpreting these findings, several methodological features and constraints merit consideration. The cohort comprised a clearly defined high-risk postoperative surgical ICU population, with delirium assessed using a standardized CAM-ICU protocol performed by trained bedside nurses. Glycemic variability was characterized using a prespecified multi-index framework, and associations were directionally consistent across the examined GV metrics.\u003c/p\u003e\n\u003cp\u003eGiven the retrospective single-center design, residual confounding remains possible, particularly from time-varying factors such as evolving infection or inflammatory burden, changes in analgesia and sedation depth, and dynamic organ dysfunction. In addition, some ICU exposures included for adjustment (e.g., sedative/analgesic exposure and mechanical ventilation) may reflect both underlying risk and downstream care processes; accordingly, adjusted estimates are most appropriately interpreted as associations conditional on measured care patterns. Glucose values were derived from scheduled capillary testing, which improves measurement consistency but may miss short-lived excursions and could underestimate variability compared with continuous glucose monitoring, especially for indices sensitive to rapid fluctuations. CAM-ICU assessments were performed twice daily, so transient or hypoactive delirium may have been under-recognized; if present, non-differential misclassification would be expected to attenuate associations. Finally, reverse causality cannot be fully excluded: early subsyndromal delirium or prodromal physiologic deterioration may influence nutrition delivery, treatment adjustments, or monitoring intensity, thereby affecting observed GV. Sensitivity analyses recalculating GV using measurements obtained prior to delirium onset were conducted to reduce this concern, although it cannot be eliminated entirely, intermittent sampling may underestimate variability vs CGM\u003c/p\u003e"},{"header":"Conclusion","content":"\u003cp\u003eIn older adults admitted to the ICU after abdominal surgery, higher glycemic variability was independently associated with delirium across multiple GV metrics. Nonlinear exposure-response patterns suggest that risk accelerates beyond moderate variability, supporting GV as a potential risk stratification tool. Prospective multicenter studies with continuous glucose monitoring are needed to validate these findings and test whether GV-stabilizing strategies improve delirium outcomes.\u003c/p\u003e"},{"header":"Declarations","content":"\u003cp\u003e\u003cstrong\u003eData Availability Statement\u003c/strong\u003e:\u003c/p\u003e\n\u003cp\u003eThe data that support the findings of this study are available from the corresponding author upon reasonable request and subject to ethical approval.\u003c/p\u003e\n\u003cp\u003e\u003cstrong\u003eFunding Statement\u003c/strong\u003e:\u003c/p\u003e\n\u003cp\u003eNot applicable.\u003c/p\u003e\n\u003cp\u003e\u003cstrong\u003eConflict of Interest Disclosure\u003c/strong\u003e:\u003c/p\u003e\n\u003cp\u003eThe authors declare no conflicts of interest.\u003c/p\u003e\n\u003cp\u003e\u003cstrong\u003eEthics Approval Statement:\u003c/strong\u003e\u003c/p\u003e\n\u003cp\u003eEthics approval was obtained from the Ethics Committee of Jinling Hospital, Nanjing University (Approval No. 2022DZKY-036-01).\u003c/p\u003e\n\u003cp\u003e\u003cstrong\u003ePatient Consent Statement\u003c/strong\u003e:\u003c/p\u003e\n\u003cp\u003eThe requirement for informed consent was waived because of the retrospective design and use of de-identified routinely collected data.\u003c/p\u003e\n\u003cp\u003e\u003cstrong\u003ePermission to Reproduce Material\u003c/strong\u003e:\u003c/p\u003e\n\u003cp\u003eNot applicable.\u003c/p\u003e\n\u003cp\u003e\u003cstrong\u003eClinical trial number:\u003c/strong\u003e\u003c/p\u003e\n\u003cp\u003eNot applicable.\u003c/p\u003e"},{"header":"References","content":"\u003col\u003e\n\u003cli\u003eShi Z, Mei X, Li C, Chen Y, Zheng H, Wu Y, Zheng H, Liu L, Marcantonio ER, Xie Z\u003cem\u003e et al\u003c/em\u003e: \u003cstrong\u003ePostoperative Delirium Is Associated with Long-term Decline in Activities of Daily Living\u003c/strong\u003e. \u003cem\u003eAnesthesiology \u003c/em\u003e2019, \u003cstrong\u003e131\u003c/strong\u003e(3):492-500.\u003c/li\u003e\n\u003cli\u003eLander HL, Dick AW, Joynt Maddox KE, Oldham MA, Fleisher LA, Mazzeffi M, Lustik SJ, Shang J, Stone PW, Gloff MS\u003cem\u003e et al\u003c/em\u003e: \u003cstrong\u003ePostoperative Delirium in Older Adults Undergoing Noncardiac Surgery\u003c/strong\u003e. \u003cem\u003eJAMA Network Open \u003c/em\u003e2025, \u003cstrong\u003e8\u003c/strong\u003e(7):e2519467-e2519467.\u003c/li\u003e\n\u003cli\u003ePanayi AC, Friedrich S, R\u0026uuml;hl J, Schaschinger T, Niederegger T, Knoedler L, Knoedler S, Rendenbach C, Treskatsch S, Harhaus L\u003cem\u003e et al\u003c/em\u003e: \u003cstrong\u003eDelirium after surgery: a retrospective study of predictors, complications, and screening patterns in the national surgical quality improvement program\u003c/strong\u003e. \u003cem\u003eEClinicalMedicine \u003c/em\u003e2025, \u003cstrong\u003e90\u003c/strong\u003e:103629.\u003c/li\u003e\n\u003cli\u003eAldecoa C, Bettelli G, Bilotta F, Sanders RD, Aceto P, Audisio R, Cherubini A, Cunningham C, Dabrowski W, Forookhi A\u003cem\u003e et al\u003c/em\u003e: \u003cstrong\u003eUpdate of the European Society of Anaesthesiology and Intensive Care Medicine evidence-based and consensus-based guideline on postoperative delirium in adult patients\u003c/strong\u003e. \u003cem\u003eEur J Anaesthesiol \u003c/em\u003e2024, \u003cstrong\u003e41\u003c/strong\u003e(2):81-108.\u003c/li\u003e\n\u003cli\u003eHoogma DF, Milisen K, Rex S, Al Tmimi L: \u003cstrong\u003ePostoperative delirium: identifying the patient at risk and altering the course: A narrative review\u003c/strong\u003e. \u003cem\u003eEur J Anaesthesiol Intensive Care \u003c/em\u003e2023, \u003cstrong\u003e2\u003c/strong\u003e(3):e0022.\u003c/li\u003e\n\u003cli\u003eDevlin JW, Skrobik Y, G\u0026eacute;linas C, Needham DM, Slooter AJC, Pandharipande PP, Watson PL, Weinhouse GL, Nunnally ME, Rochwerg B\u003cem\u003e et al\u003c/em\u003e: \u003cstrong\u003eExecutive Summary: Clinical Practice Guidelines for the Prevention and Management of Pain, Agitation/Sedation, Delirium, Immobility, and Sleep Disruption in Adult Patients in the ICU\u003c/strong\u003e. \u003cem\u003eCrit Care Med \u003c/em\u003e2018, \u003cstrong\u003e46\u003c/strong\u003e(9):1532-1548.\u003c/li\u003e\n\u003cli\u003eZhang S, Han Y, Xiao Q, Li H, Wu Y: \u003cstrong\u003eEffectiveness of Bundle Interventions on ICU Delirium: A Meta-Analysis\u003c/strong\u003e. \u003cem\u003eCrit Care Med \u003c/em\u003e2021, \u003cstrong\u003e49\u003c/strong\u003e(2):335-346.\u003c/li\u003e\n\u003cli\u003evan den Boogaard M, Pickkers P, Slooter AJ, Kuiper MA, Spronk PE, van der Voort PH, van der Hoeven JG, Donders R, van Achterberg T, Schoonhoven L: \u003cstrong\u003eDevelopment and validation of PRE-DELIRIC (PREdiction of DELIRium in ICu patients) delirium prediction model for intensive care patients: observational multicentre study\u003c/strong\u003e. \u003cem\u003eBmj \u003c/em\u003e2012, \u003cstrong\u003e344\u003c/strong\u003e:e420.\u003c/li\u003e\n\u003cli\u003eWatt C, Sanchez-Rangel E, Hwang JJ: \u003cstrong\u003eGlycemic Variability and CNS Inflammation: Reviewing the Connection\u003c/strong\u003e. \u003cem\u003eNutrients \u003c/em\u003e2020, \u003cstrong\u003e12\u003c/strong\u003e(12).\u003c/li\u003e\n\u003cli\u003eSuh S, Kim JH: \u003cstrong\u003eGlycemic Variability: How Do We Measure It and Why Is It Important?\u003c/strong\u003e \u003cem\u003eDiabetes Metab J \u003c/em\u003e2015, \u003cstrong\u003e39\u003c/strong\u003e(4):273-282.\u003c/li\u003e\n\u003cli\u003eChoi H, Park CS, Huh J, Koo J, Jeon J, Kim E, Jung S, Kim HW, Lim JY, Hwang W: \u003cstrong\u003eIntraoperative Glycemic Variability and Mean Glucose are Predictors for Postoperative Delirium After Cardiac Surgery: A Retrospective Cohort Study\u003c/strong\u003e. \u003cem\u003eClin Interv Aging \u003c/em\u003e2022, \u003cstrong\u003e17\u003c/strong\u003e:79-95.\u003c/li\u003e\n\u003cli\u003evan Keulen K, Knol W, Belitser SV, Zaal IJ, van der Linden PD, Heerdink ER, Egberts TCG, Slooter AJC: \u003cstrong\u003eGlucose variability during delirium in diabetic and non-diabetic intensive care unit patients: A prospective cohort study\u003c/strong\u003e. \u003cem\u003ePLoS One \u003c/em\u003e2018, \u003cstrong\u003e13\u003c/strong\u003e(11):e0205637.\u003c/li\u003e\n\u003cli\u003eGuan B, Chen Z, Lin L, Shao F, Huang J, Liu L: \u003cstrong\u003eAssociation between group-based trajectory modeling of blood glucose and acute delirium incidence in critically ill patients: a retrospective cohort study based on the MIMIC-IV database\u003c/strong\u003e. \u003cem\u003eSci Rep \u003c/em\u003e2025, \u003cstrong\u003e15\u003c/strong\u003e(1):44676.\u003c/li\u003e\n\u003cli\u003eLin YJ, Lin LY, Peng YC, Zhang HR, Chen LW, Huang XZ, Chen Q: \u003cstrong\u003eAssociation between glucose variability and postoperative delirium in acute aortic dissection patients: an observational study\u003c/strong\u003e. \u003cem\u003eJ Cardiothorac Surg \u003c/em\u003e2021, \u003cstrong\u003e16\u003c/strong\u003e(1):82.\u003c/li\u003e\n\u003cli\u003eKlonoff DC, Umpierrez GE, Rice MJ: \u003cstrong\u003eA Milestone in Point of Care Capillary Blood Glucose Monitoring of Critically Ill Hospitalized Patients\u003c/strong\u003e. \u003cem\u003eJ Diabetes Sci Technol \u003c/em\u003e2018, \u003cstrong\u003e12\u003c/strong\u003e(6):1095-1100.\u003c/li\u003e\n\u003cli\u003eBattelino T, Danne T, Bergenstal RM, Amiel SA, Beck R, Biester T, Bosi E, Buckingham BA, Cefalu WT, Close KL\u003cem\u003e et al\u003c/em\u003e: \u003cstrong\u003eClinical Targets for Continuous Glucose Monitoring Data Interpretation: Recommendations From the International Consensus on Time in Range\u003c/strong\u003e. \u003cem\u003eDiabetes Care \u003c/em\u003e2019, \u003cstrong\u003e42\u003c/strong\u003e(8):1593-1603.\u003c/li\u003e\n\u003cli\u003eRyan EA, Shandro T, Green K, Paty BW, Senior PA, Bigam D, Shapiro AMJ, Vantyghem M-C: \u003cstrong\u003eAssessment of the Severity of Hypoglycemia and Glycemic Lability in Type 1 Diabetic Subjects Undergoing Islet Transplantation\u003c/strong\u003e. \u003cem\u003eDiabetes \u003c/em\u003e2004, \u003cstrong\u003e53\u003c/strong\u003e(4):955-962.\u003c/li\u003e\n\u003cli\u003eService FJ, Molnar GD, Rosevear JW, Ackerman E, Gatewood LC, Taylor WF: \u003cstrong\u003eMean Amplitude of Glycemic Excursions, a Measure of Diabetic Instability\u003c/strong\u003e. \u003cem\u003eDiabetes \u003c/em\u003e1970, \u003cstrong\u003e19\u003c/strong\u003e(9):644-655.\u003c/li\u003e\n\u003cli\u003eEly EW, Inouye SK, Bernard GR, Gordon S, Francis J, May L, Truman B, Speroff T, Gautam S, Margolin R\u003cem\u003e et al\u003c/em\u003e: \u003cstrong\u003eDelirium in Mechanically Ventilated PatientsValidity and Reliability of the Confusion Assessment Method for the Intensive Care Unit (CAM-ICU)\u003c/strong\u003e. \u003cem\u003eJAMA \u003c/em\u003e2001, \u003cstrong\u003e286\u003c/strong\u003e(21):2703-2710.\u003c/li\u003e\n\u003cli\u003eFaul F, Erdfelder E, Lang AG, Buchner A: \u003cstrong\u003eG*Power 3: a flexible statistical power analysis program for the social, behavioral, and biomedical sciences\u003c/strong\u003e. \u003cem\u003eBehav Res Methods \u003c/em\u003e2007, \u003cstrong\u003e39\u003c/strong\u003e(2):175-191.\u003c/li\u003e\n\u003cli\u003eFaul F, Erdfelder E, Buchner A, Lang AG: \u003cstrong\u003eStatistical power analyses using G*Power 3.1: tests for correlation and regression analyses\u003c/strong\u003e. \u003cem\u003eBehav Res Methods \u003c/em\u003e2009, \u003cstrong\u003e41\u003c/strong\u003e(4):1149-1160.\u003c/li\u003e\n\u003cli\u003eMarcantonio ER: \u003cstrong\u003eDelirium in Hospitalized Older Adults\u003c/strong\u003e. \u003cem\u003eN Engl J Med \u003c/em\u003e2017, \u003cstrong\u003e377\u003c/strong\u003e(15):1456-1466.\u003c/li\u003e\n\u003cli\u003eHryciw BN, Ghossein J, Rochwerg B, Meggison H, Fernando SM, Kyeremanteng K, Tran A, Seely AJE: \u003cstrong\u003eGlycemic Variability As a Prognostic Factor for Mortality in Patients With Critical Illness: A Systematic Review and Meta-Analysis\u003c/strong\u003e. \u003cem\u003eCrit Care Explor \u003c/em\u003e2024, \u003cstrong\u003e6\u003c/strong\u003e(1):e1025.\u003c/li\u003e\n\u003cli\u003eEslami S, Taherzadeh Z, Schultz MJ, Abu-Hanna A: \u003cstrong\u003eGlucose variability measures and their effect on mortality: a systematic review\u003c/strong\u003e. \u003cem\u003eIntensive Care Med \u003c/em\u003e2011, \u003cstrong\u003e37\u003c/strong\u003e(4):583-593.\u003c/li\u003e\n\u003cli\u003eKrinsley JS: \u003cstrong\u003eGlycemic variability: a strong independent predictor of mortality in critically ill patients\u003c/strong\u003e. \u003cem\u003eCrit Care Med \u003c/em\u003e2008, \u003cstrong\u003e36\u003c/strong\u003e(11):3008-3013.\u003c/li\u003e\n\u003cli\u003eEgi M, Bellomo R, Stachowski E, French CJ, Hart G: \u003cstrong\u003eVariability of blood glucose concentration and short-term mortality in critically ill patients\u003c/strong\u003e. \u003cem\u003eAnesthesiology \u003c/em\u003e2006, \u003cstrong\u003e105\u003c/strong\u003e(2):244-252.\u003c/li\u003e\n\u003cli\u003eMonnier L, Mas E, Ginet C, Michel F, Villon L, Cristol JP, Colette C: \u003cstrong\u003eActivation of oxidative stress by acute glucose fluctuations compared with sustained chronic hyperglycemia in patients with type 2 diabetes\u003c/strong\u003e. \u003cem\u003eJama \u003c/em\u003e2006, \u003cstrong\u003e295\u003c/strong\u003e(14):1681-1687.\u003c/li\u003e\n\u003cli\u003eTaylor J, Parker M, Casey CP, Tanabe S, Kunkel D, Rivera C, Zetterberg H, Blennow K, Pearce RA, Lennertz RC\u003cem\u003e et al\u003c/em\u003e: \u003cstrong\u003ePostoperative delirium and changes in the blood-brain barrier, neuroinflammation, and cerebrospinal fluid lactate: a prospective cohort study\u003c/strong\u003e. \u003cem\u003eBr J Anaesth \u003c/em\u003e2022, \u003cstrong\u003e129\u003c/strong\u003e(2):219-230.\u003c/li\u003e\n\u003cli\u003eSaisho Y: \u003cstrong\u003eGlycemic variability and oxidative stress: a link between diabetes and cardiovascular disease?\u003c/strong\u003e \u003cem\u003eInt J Mol Sci \u003c/em\u003e2014, \u003cstrong\u003e15\u003c/strong\u003e(10):18381-18406.\u003c/li\u003e\n\u003cli\u003eCeriello A, Esposito K, Piconi L, Ihnat MA, Thorpe JE, Testa R, Boemi M, Giugliano D: \u003cstrong\u003eOscillating Glucose Is More Deleterious to Endothelial Function and Oxidative Stress Than Mean Glucose in Normal and Type 2 Diabetic Patients\u003c/strong\u003e. \u003cem\u003eDiabetes \u003c/em\u003e2008, \u003cstrong\u003e57\u003c/strong\u003e(5):1349-1354.\u003c/li\u003e\n\u003cli\u003eMcKay TB, Khawaja ZQ, Freedman IG, Turco I, Wiredu K, Colecchi T, Akeju O: \u003cstrong\u003eExploring the Pathophysiology of Delirium: An Overview of Biomarker Studies, Animal Models, and Tissue-Engineered Models\u003c/strong\u003e. \u003cem\u003eAnesth Analg \u003c/em\u003e2023, \u003cstrong\u003e137\u003c/strong\u003e(6):1186-1197.\u003c/li\u003e\n\u003cli\u003eMattison MLP: \u003cstrong\u003eDelirium\u003c/strong\u003e. \u003cem\u003eAnn Intern Med \u003c/em\u003e2020, \u003cstrong\u003e173\u003c/strong\u003e(7):Itc49-itc64.\u003c/li\u003e\n\u003cli\u003eDuggan EW, Carlson K, Umpierrez GE: \u003cstrong\u003ePerioperative Hyperglycemia Management: An Update\u003c/strong\u003e. \u003cem\u003eAnesthesiology \u003c/em\u003e2017, \u003cstrong\u003e126\u003c/strong\u003e(3):547-560.\u003c/li\u003e\n\u003cli\u003eBoh\u0026eacute; J, Abidi H, Brunot V, Klich A, Klouche K, Sedillot N, Tchenio X, Quenot JP, Roudaut JB, Mottard N\u003cem\u003e et al\u003c/em\u003e: \u003cstrong\u003eIndividualised versus conventional glucose control in critically-ill patients: the CONTROLING study-a randomized clinical trial\u003c/strong\u003e. \u003cem\u003eIntensive Care Med \u003c/em\u003e2021, \u003cstrong\u003e47\u003c/strong\u003e(11):1271-1283.\u003c/li\u003e\n\u003c/ol\u003e"}],"fulltextSource":"","fullText":"","funders":[],"hasAdminPriorityOnWorkflow":false,"hasManuscriptDocX":true,"hasOptedInToPreprint":true,"hasPassedJournalQc":"","hasAnyPriority":false,"hideJournal":false,"highlight":"","institution":"","isAcceptedByJournal":false,"isAuthorSuppliedPdf":false,"isDeskRejected":"","isHiddenFromSearch":false,"isInQc":false,"isInWorkflow":false,"isPdf":false,"isPdfUpToDate":true,"isWithdrawnOrRetracted":false,"journal":{"display":true,"email":"[email protected]","identity":"bmc-geriatrics","isNatureJournal":false,"hasQc":true,"allowDirectSubmit":false,"externalIdentity":"bgtc","sideBox":"Learn more about [BMC Geriatrics](http://bmcgeriatr.biomedcentral.com/)","snPcode":"","submissionUrl":"https://www.editorialmanager.com/bgtc/default.aspx","title":"BMC Geriatrics","twitterHandle":"BMC_series","acdcEnabled":true,"dfaEnabled":false,"editorialSystem":"em","reportingPortfolio":"BMC Series","inReviewEnabled":true,"inReviewRevisionsEnabled":true},"keywords":"Glycemic variability, Delirium, Postoperative care, Intensive care unit, Aged, Abdominal surgery","lastPublishedDoi":"10.21203/rs.3.rs-9021880/v1","lastPublishedDoiUrl":"https://doi.org/10.21203/rs.3.rs-9021880/v1","license":{"name":"CC BY 4.0","url":"https://creativecommons.org/licenses/by/4.0/"},"manuscriptAbstract":"\u003cp\u003e\u003cstrong\u003eBackground: \u003c/strong\u003eICU delirium is common after abdominal surgery in older adults. Glycemic variability (GV) may better capture dysglycemia-related physiologic stress than mean glucose, but evidence in postoperative ICU populations is limited.\u003c/p\u003e\n\u003cp\u003e\u003cstrong\u003eMethods: \u003c/strong\u003eThis retrospective single-center study included 473 patients aged ≥65 years admitted to a postoperative ICU after abdominal surgery (January 2022–December 2025). GV was quantified using four metrics: standard deviation (SD), coefficient of variation (GluCV), glycemic lability index (GLI), and mean amplitude of glycemic excursions (MAGE). Multivariable logistic regression and restricted cubic splines (RCS) assessed associations with ICU delirium, adjusting for confounders.\u003c/p\u003e\n\u003cp\u003e\u003cstrong\u003eResults: \u003c/strong\u003eDelirium occurred in 122 patients (25.8%). All GV metrics were higher in delirium patients (all P\u0026lt;0.001). In adjusted models, each 1-SD increase in GV was associated with higher delirium risk: GluCV (aOR 3.34, 95%CI 2.39-4.66), SD (2.15, 1.59-2.90), GLI (3.16, 2.09-4.77), and MAGE (2.26, 1.68-3.02). RCS revealed nonlinear associations for SD, GluCV, and GLI (P for nonlinearity \u0026lt;0.05), with risk accelerating beyond moderate GV levels.\u003c/p\u003e\n\u003cp\u003e\u003cstrong\u003eConclusions: \u003c/strong\u003eHigher glycemic variability is independently associated with postoperative ICU delirium in older adults. The nonlinear patterns suggest potential thresholds for risk stratification. Prospective studies are needed to test whether GV-stabilizing strategies reduce delirium.\u003c/p\u003e","manuscriptTitle":"Glycemic Variability and ICU Delirium After Abdominal Surgery in Older Adults: Nonlinear Associations in a Retrospective Single-Center Study","msid":"","msnumber":"","nonDraftVersions":[{"code":1,"date":"2026-05-15 16:08:58","doi":"10.21203/rs.3.rs-9021880/v1","editorialEvents":[{"type":"communityComments","content":0},{"type":"reviewersInvited","content":"","date":"2026-05-06T12:02:37+00:00","index":"","fulltext":""},{"type":"editorInvited","content":"","date":"2026-04-07T13:47:19+00:00","index":"","fulltext":""},{"type":"editorAssigned","content":"","date":"2026-03-09T00:57:34+00:00","index":"","fulltext":""},{"type":"checksComplete","content":"","date":"2026-03-09T00:57:11+00:00","index":"","fulltext":""},{"type":"submitted","content":"BMC Geriatrics","date":"2026-03-03T15:20:31+00:00","index":"","fulltext":""}],"status":"published","journal":{"display":true,"email":"[email protected]","identity":"bmc-geriatrics","isNatureJournal":false,"hasQc":true,"allowDirectSubmit":false,"externalIdentity":"bgtc","sideBox":"Learn more about [BMC Geriatrics](http://bmcgeriatr.biomedcentral.com/)","snPcode":"","submissionUrl":"https://www.editorialmanager.com/bgtc/default.aspx","title":"BMC Geriatrics","twitterHandle":"BMC_series","acdcEnabled":true,"dfaEnabled":false,"editorialSystem":"em","reportingPortfolio":"BMC Series","inReviewEnabled":true,"inReviewRevisionsEnabled":true}}],"origin":"","ownerIdentity":"6bcc297b-2d0c-4ecd-ad19-a2c014c6b054","owner":[],"postedDate":"May 15th, 2026","published":true,"recentEditorialEvents":[{"type":"reviewersInvited","content":"30","date":"2026-05-06T12:02:37+00:00","index":"","fulltext":""}],"rejectedJournal":[],"revision":"","amendment":"","status":"under-review","subjectAreas":[],"tags":[],"updatedAt":"2026-05-15T16:08:58+00:00","versionOfRecord":[],"versionCreatedAt":"2026-05-15 16:08:58","video":"","vorDoi":"","vorDoiUrl":"","workflowStages":[]},"version":"v1","identity":"rs-9021880","journalConfig":"researchsquare"},"__N_SSP":true},"page":"/article/[identity]/[[...version]]","query":{"redirect":"/article/rs-9021880","identity":"rs-9021880","version":["v1"]},"buildId":"XKTyCvWXoU3ODBz1xrDgd","isFallback":false,"isExperimentalCompile":false,"dynamicIds":[84888],"gssp":true,"scriptLoader":[]}