Abstract
Background
Postoperative delirium (POD) is a common and serious complication in older adults. Prior studies have validated serum amyloid A (SAA) as a potential biomarker for various inflammatory conditions; however, its role in POD is poorly characterized. This study aimed to examine the association between preoperative SAA levels and the occurrence of POD.
Methods
This study comprised a retrospective review of the electronic medical records of patients aged 60 and older who underwent hip surgery between April 2022 and January 2024. Cognitive function was assessed using the Nursing Delirium Screening Scale, while diagnosis was confirmed by psychiatrists using the Confusion Assessment Method. The associations between POD occurrence and preoperative SAA levels and other patient, anesthesia, and surgical factors were analyzed using logistic regression models.
Results
Of 731 patients, delirium occurred in 121 patients (16.6%) within the first five postoperative days (five-day POD). Preoperative SAA levels were significantly higher in the POD group (96.6 mg/L) than the non-POD group (6.6 mg/L) (P < 0.001). Logistic regression showed that preoperative SAA levels were independently associated with POD occurrence (odds ratio [OR] = 1.004, 95% CI [1.002‒1.007], P < 0.001). Age (P < 0.001) and preoperative albumin levels (P = 0.008) were also identified as significant factors influencing POD risk.
Conclusions
Elevated preoperative SAA levels were associated with an increased risk of five-day POD in older adults undergoing hip surgery. Further research is required to explore the clinical utility of SAA as a biomarker for identifying patients at risk of POD.
Keywords: Biomarker, Cognition, Delirium, Neuroinflammation, Older adult, Serum amyloid A
Introduction
Postoperative delirium (POD) is a common and serious complication in older adults after hip surgery [1], and is associated with adverse outcomes including prolonged hospital stays, increased morbidity, and long-term cognitive decline [2]. Older patients, who often have multiple comorbidities and altered physiological responses to surgery, are particularly vulnerable to POD. The identification of reliable risk factors and predictive markers for POD is essential to facilitate early intervention, reduce complications, and improve patient outcomes.
POD is believed to be primarily driven by neuroinflammation, a process in which systemic inflammatory responses trigger alterations in brain function [3]. Inflammation, induced by surgical trauma, anesthesia, and other stressors, can lead to the activation of microglial cells and the release of proinflammatory cytokines that influence neuronal signaling and contribute to delirium development [3]. Serum amyloid A (SAA), a protein whose expression increases in response to acute-phase systemic inflammation, has been shown to play a role in regulating inflammatory pathways [4]. Given the inflammatory nature of POD, we hypothesized that SAA may serve as a key biomarker for identifying patients at higher risk of developing POD.
This study aimed to assess whether preoperative SAA levels are associated with the incidence of POD in older adults undergoing hip surgery. Clarifying this association may provide valuable insights for developing clinical strategies to prevent or reduce POD and to promote better recovery after surgery. In addition, a deeper understanding of neuroinflammation and other clinical factors involved in delirium could enhance our ability to identify high-risk patients, enabling more targeted monitoring and timely interventions to mitigate the impact of POD and ultimately improve postoperative outcomes in this vulnerable population.
Materials and Methods
Study setting and patients
This retrospective study was approved by the Institutional Review Board of Seoul National University Bundang Hospital (approval number B-2403-888-102, February 20, 2024) and adhered to the institution’s research protocols. Given the retrospective study design, informed consent from participants was not required. The study was reported in accordance with the Strengthening the Reporting of Observational Studies in Epidemiology (STROBE) guidelines.
Data from the electronic medical records were extracted by a trained medical record specialist blinded to the study’s objectives. The study focused on older adults aged 60 years and older who underwent hip surgery between April 2022 and January 2024. ‘Older adults’ were defined as patients aged 60 years and above, a criterion consistent with the World Health Organization’s (WHO) definitions for global health and demographic studies [5]. This approach allowed a broader assessment of age-related risk factors, encompassing the ‘young-old’ population. Patients with comorbid cognitive disorders (e.g., dementia, Parkinson’s disease, or preoperative delirium) or incomplete medical records were excluded from the analysis.
Assessment of POD
Cognitive function was assessed by nurses during each nursing shift (conducted thrice daily) using the Korean-validated version of the Nursing Delirium Screening Scale (Nu-DESC) [6], specifically adapted to our clinical context. The Nu-DESC evaluates five symptoms —disorientation, inappropriate behavior, inappropriate communication, hallucinations/illusions, and psychomotor retardation — scoring each as 0 or 1 based on the absence or presence of the symptom. For screening, delirium was defined by a Nu-DESC score of 2 or higher. If any signs of delirium, such as agitation, irritability, fear, or hallucinations, were observed, the orthopedic surgeons referred the patient for psychiatric evaluation; a psychiatrist subsequently assessed the patient’s cognitive function using the Confusion Assessment Method [7], and, if delirium was diagnosed, appropriate treatment for POD was initiated. Cases were still categorized as POD, even if symptoms resolved spontaneously prior to psychiatric consultation. Since the majority of POD cases develop within the first five days post-surgery [8], we reviewed the medical records up to five days postoperatively (five-day POD). Any patient who tested positive on the Nu-DESC at least once during this period was classed into the POD group.
Endpoints and data collection
The primary aim of this study was to assess the association between preoperative SAA levels (normal reference range: < 10 mg/L) and the occurrence of five-day POD in older adults undergoing hip surgery. Further, we investigated other factors influencing the onset of POD. Relevant variables were divided into three categories for extraction: 1) patient-related factors, including sex, age, weight, height, body mass index (BMI), American Society of Anesthesiologists physical status (ASA-PS), underlying comorbidities (such as hypertension, diabetes, chronic kidney, liver, or lung disease, and a history of stroke), and preoperative laboratory results; 2) anesthesia-related factors, including anesthesia duration, anesthesia type (regional or general), and red blood cell (RBC) transfusion; and 3) surgical-related factors, including surgery duration and intensive care unit (ICU) admission.
Statistical analysis
The baseline characteristics, anesthesia-related variables, and surgical variables of the patients were summarized as the median [interquartile range (Q1, Q3)] or mean (standard deviation) for continuous variables, or as counts (%) for categorical variables. All continuous data were evaluated for normality using the Shapiro–Wilk test. Comparisons of variables between the non-POD and POD groups were conducted using the Mann–Whitney U and chi-squared tests for continuous variables and categorical data, respectively. Variables were first assessed for multicollinearity using variance inflation factors (VIF). In this preliminary assessment, weight (VIF 4.37) and operation time (VIF 10.67) showed the highest collinearity — coinciding with BMI and anesthesia time, respectively — and were therefore excluded from further analysis. After exclusion, VIF values for the remaining variables ranged from 1.06 to 2.22, indicating no significant multicollinearity. These remaining variables were then entered into univariate logistic regression to screen for potential risk factors, and those with P < 0.2 in univariate analysis were considered for inclusion in the final multivariable model. Odds ratios (ORs) were calculated, and a multivariable model was applied to adjust for potential confounders, including age, height, BMI, underlying health conditions, anesthesia method, type of surgery, duration of anesthesia, estimated blood loss, RBC transfusion, preoperative hemoglobin, albumin, SAA levels, and ICU admission. Because preoperative SAA levels were markedly right-skewed, with a wide separation in the interquartile range between the non-POD and POD groups, we performed a sensitivity analysis to evaluate the robustness of the association between preoperative SAA and POD to this skewed distribution, in which the univariate and multivariable logistic regression analyses were repeated after log2-transformation of preoperative SAA. In addition, to evaluate the robustness of the final multivariable model and address potential separation problems arising from sparsely observed categories (such as the 'Others' category of surgery type and ASA class III/IV) with rare observed counts, we additionally performed Firth’s penalized (bias-reduction) logistic regression as a sensitivity analysis.
As this study was designed to explore factors associated with POD rather than to develop or externally validate a predictive model, the receiver operating characteristic (ROC) curve analysis was additionally performed to evaluate the discriminatory ability of preoperative SAA, that is, its ability to differentiate patients who developed POD from those who did not, rather than to estimate predicted probabilities of POD.
A secondary analysis was conducted to examine the association between preoperative SAA levels and the maximum Nu-DESC score using Spearman’s rank correlation test.
Statistical analyses were conducted using Stata® SE version 19 (Stata Corp LLC) and IBM® SPSS® Statistics version 22.0 (IBM Corporation). Statistical significance for determining the association between POD and the variables under consideration was set at P < 0.05.
Results
Out of 766 eligible patients, 731 met the study inclusion criteria (Fig. 1). The characteristics of the enrolled patients, along with variables related to anesthesia and surgery, are presented in Table 1. Overall, five-day POD was observed in 121 patients (16.6%). When the cohort was divided into two groups (non-POD and POD), significant differences were found in age, height, weight, BMI, ASA classification, comorbidities, type of surgery/anesthesia, duration of surgery/anesthesia, estimated blood loss, amount of RBCs transfused, laboratory results, Nu-DESC score, ICU admission, and length of ICU stay.
Fig. 1.
Flow diagram of the study.
Table 1.
Baseline Characteristics of the Total Study Population
| Variable | Patients | ||||
|---|---|---|---|---|---|
| Total (n = 731) | non-POD (n = 610) | POD (n = 121) | Standardized mean difference (SMD, Hedges’s g) | P value | |
| Age (yr) | 74.0 (67.0, 81.0) | 72.0 (65.0, 79.0) | 84.0 (77.0, 88.5) | ‒1.09 | < 0.001 |
| Gender (F) | 478 (65.4) | 399 (65.4) | 79 (65.3) | 1.000 | |
| Height (cm) | 157.0 (151.0, 163.2) | 157.0 (151.5, 163.5) | 155.0 (150.0, 162.3) | 0.09 | 0.034 |
| Weight (kg) | 59.1 (52.4, 67.4) | 60.4 (53.4, 68.0) | 54.2 (49.3, 62.4) | 0.50 | < 0.001 |
| BMI (kg/m2) | 24.1 (21.6, 26.5) | 24.3 (22.0, 26.7) | 22.5 (20.1, 25.2) | 0.36 | < 0.001 |
| ASA PS | < 0.001 | ||||
| 1 | 41 (5.6) | 40 (6.6) | 1 (0.8) | ||
| 2 | 396 (54.2) | 365 (59.8) | 31 (25.6) | ||
| 3 | 269 (36.8) | 188 (30.8) | 81 (66.9) | ||
| 4 | 25 (3.4) | 17 (2.8) | 8 (6.6) | ||
| Underlying disease | |||||
| Hypertension | 462 (63.2) | 371 (60.8) | 91 (75.2) | 0.003 | |
| Diabetes mellitus | 183 (25.0) | 137 (22.5) | 46 (38.0) | 0.001 | |
| Chronic kidney disease | 95 (13.0) | 63 (10.3) | 32 (26.4) | < 0.001 | |
| Chronic liver disease | 33 (4.5) | 24 (3.9) | 9 (7.4) | 0.096 | |
| Chronic pulmonary disease | 83 (11.4) | 52 (8.5) | 31 (25.6) | < 0.001 | |
| History of stroke | 29 (4.0) | 22 (3.6) | 7 (5.8) | 0.304 | |
| Type of surgery | < 0.001 | ||||
| IM nailing | 134 (18.3) | 86 (14.1) | 48 (39.7) | ||
| CRIF | 51 (7.0) | 43 (7.0) | 8 (6.6) | ||
| BHA | 105 (14.4) | 66 (10.8) | 39 (32.2) | ||
| THA | 418 (57.2) | 395 (64.8) | 23 (19.0) | ||
| Incision and drainage (I&D) | 16 (2.2) | 13 (2.1) | 3 (2.5) | ||
| Others | 7 (1.0) | 7 (1.1) | 0 (0) | ||
| Type of anesthesia | < 0.001 | ||||
| Neuraxial | 609 (83.3) | 522 (85.6) | 87 (71.9) | ||
| General | 122 (16.7) | 88 (14.4) | 34 (28.1) | ||
| Operation time (min) | 75 (60, 95) | 75 (60, 100) | 65 (50, 80) | 0.35 | < 0.001 |
| Anesthesia time (min) | 115 (100, 140) | 120 (100, 145) | 105 (90, 125) | 0.32 | < 0.001 |
| Estimated blood loss | 300 (100, 500) | 300 (150, 500) | 100 (100, 400) | 0.22 | < 0.001 |
| RBCs (units, during surgery) | 0 (0, 0)/0.2 (0.8) | 0 (0, 0)/0.2 (0.8) | 0 (0, 0)/0.4 (0.8) | ‒0.19 | 0.001 |
| RBCs (units, after surgery) | 0 (0, 0)/0.3 (1.0) | 0 (0, 0)/0.3 (1.0) | 0 (0, 0)/0.5 (1.2) | ‒0.28 | < 0.001 |
| Laboratory findings—pre | |||||
| Hemoglobin (g/dL) | 12.3 (11.1, 13.5) | 12.6 (11.4, 13.7) | 11.2 (9.9, 12.3) | 0.30 | < 0.001 |
| Albumin (g/dL) | 4.1 (3.7, 4.4) | 4.2 (3.8, 4.5) | 3.6 (3.2, 4.0) | 0.84 | < 0.001 |
| SAA (mg/L) | 8.4 (4.4, 59.6) | 6.6 (4.0, 28.3) | 96.6 (22.9, 301.0) | ‒1.04 | < 0.001 |
| Laboratory findings—post | |||||
| Hemoglobin (g/dL) | 10.0 (9.0, 11.1) | 10.1 (9.1, 11.2) | 9.4 (8.6, 10.4) | 0.38 | < 0.001 |
| Albumin (g/dL) | 3.4 (3.1, 3.7) | 3.4 (3.2, 3.7) | 3.1 (2.9, 3.4) | 0.67 | < 0.001 |
| SAA | 301.0 (215.6, 451.5) | 301.5 (215.6, 452.8) | 301.0 (215.9, 425.8) | 0.03 | 0.636 |
| Postoperative ICU admittance | 50 (6.8) | 31 (5.1) | 19 (15.7) | < 0.001 | |
| ICU stay (d) | 0 (0, 0)/0.2 (1.8) | 0 (0, 0)/0.2 (1.4) | 0 (0, 0)/0.7 (2.8) | ‒0.32 | < 0.001 |
| Admission period (d) | 6.0 (5.0, 8.0) | 6.0 (5.0, 7.0) | 7.0 (6.0, 10.5) | ‒0.27 | < 0.001 |
| Nu-DESC score | 0 (0, 0)/0.4 (0.9) | 0 (0, 0)/0 (0.1) | 3 (2, 3)/2.5 (0.6) | ‒10.02 | < 0.001 |
| Delirium | 121 (16.6) | N/A | 121 (100) | N/A | |
| Delirium onset (d) | N/A | N/A | 1 (1, 1)/1.1 (0.4) | N/A | |
| Delirium duration (d) | N/A | N/A | 1 (1, 2)/1.7 (1.0) | N/A | |
Values are presented as median (Q1, Q3) with/without mean (SD), or number of the patients (%). POD: postoperative delirium, SMD: standardized mean difference, BMI: body mass index, ASA-PS: American Society of Anesthesiologists physical status, IM: intramedullary, CRIF: closed reduction and internal fixation, BHA: bipolar hemiarthroplasty, THA: total hip arthroplasty, I&D: incision and drainage, RBC: red blood cell, SAA: serum amyloid A, ICU: intensive care unit, Nu-DESC: Nursing Delirium Screening Scale, N/A: not applicable.
Association between SAA levels and POD developments
Preoperative SAA levels (median [Q1, Q3]) were 6.6 (4.0, 28.3) mg/L and 96.6 (22.9, 301.0) mg/L in the non-POD and POD groups, respectively, showing a significant difference (P < 0.001, Table 1). Furthermore, higher SAA levels were significantly associated with an increased risk of five-day POD (OR = 1.004, 95% CI [1.002‒1.007], P < 0.001; Table 2).
Table 2.
Binary Logistic Regression for the Occurrence of POD
| Independent variables | Univariate | Multivariable | ||
|---|---|---|---|---|
| OR (95% CI) | P value | OR (95% CI) | P value | |
| SAA (preoperative) | 1.008 (1.006–1.010) | < 0.001 | 1.004 (1.002–1.007) | < 0.001 |
| Age (yr) | 1.131 (1.102–1.161) | < 0.001 | 1.084 (1.046–1.123) | < 0.001 |
| Sexa | ||||
| M | 1 (reference) | |||
| F | 0.995 (0.660–1.498) | 0.980 | ||
| Height (cm) | 0.974 (0.952–0.996) | 0.023 | 0.989 (0.962–1.018) | 0.457 |
| BMI (kg/m2) | 0.921 (0.879–0.964) | < 0.001 | 0.994 (0.937–1.054) | 0.834 |
| ASA-PS | < 0.001 | 0.172 | ||
| I | 1 (reference) | 1 (reference) | ||
| II | 3.397 (0.452–25.556) | 0.235 | 0.981 (0.120–8.060) | 0.986 |
| III | 17.234 (2.329–127.514) | 0.005 | 1.461 (0.170–12.543) | 0.730 |
| IV | 18.824 (2.182–162.393) | 0.008 | 0.441 (0.038–5.168) | 0.514 |
| Underlying disease | ||||
| Hypertension | 1.954 (1.254–3.045) | 0.003 | 1.152 (0.644–2.058) | 0.634 |
| Diabetes mellitus | 2.118 (1.401–3.202) | < 0.001 | 1.488 (0.855–2.590) | 0.160 |
| Chronic kidney disease | 3.116 (1.927–5.040) | < 0.001 | 1.083 (0.571–2.051) | 0.807 |
| Chronic liver disease | 1.962 (0.888–4.333) | 0.095 | 0.961 (0.321–2.882) | 0.944 |
| Chronic pulmonary disease | 3.696 (2.248–6.078) | < 0.001 | 1.828 (0.961–3.477) | 0.066 |
| History of strokea | 1.641 (0.685–3.392) | 0.267 | ||
| Hemoglobin (g/dL, preoperative) | 0.729 (0.657–0.808) | < 0.001 | 1.019 (0.971–1.070) | 0.443 |
| Albumin (g/dL, preoperative) | 0.213 (0.149–0.304) | < 0.001 | 0.470 (0.270–0.820) | 0.008 |
| Type of surgery | < 0.001 | 0.237 | ||
| IM nailing | 1 (reference) | 1 (reference) | ||
| CRIF | 0.333 (0.145–0.767) | 0.010 | 0.370 (0.134–1.023) | 0.055 |
| BHA | 1.059 (0.623–1.800) | 0.833 | 1.167 (0.614–2.218) | 0.638 |
| THA | 0.104 (0.060–0.181) | < 0.001 | 0.718 (0.307–1.679) | 0.444 |
| I&D | 0.413 (0.112–1.523) | 0.184 | 0.310 (0.047–2.061) | 0.226 |
| Others | 0.000 (0.000–N/A) | 0.999 | 0.000 (0.000–N/A) | 0.999 |
| Type of anesthesia | ||||
| Neuraxial | 1 (reference) | 1 (reference) | ||
| General | 2.318 (1.469–3.658) | < 0.001 | 1.716 (0.929–3.171) | 0.085 |
| Anesthesia time | 0.992 (0.987–0.997) | 0.003 | 1.000 (0.994–1.007) | 0.964 |
| Estimated blood loss | 0.999 (0.998–1.000) | 0.014 | 1.000 (0.999–1.001) | 0.865 |
| RBCs (units, during surgery) | 1.196 (0.984–1.454) | 0.073 | 0.955 (0.663–1.376) | 0.805 |
| ICU admittance | 3.479 (1.893–6.394) | < 0.001 | 1.357 (0.561–3.285) | 0.498 |
POD: postoperative delirium, OR: odds ratio, SAA: serum amyloid A, BMI: body mass index, ASA-PS: American Society of Anesthesiologists physical status, IM: intramedullary, CRIF: closed reduction and internal fixation, BHA: bipolar hemiarthroplasty, THA: total hip arthroplasty, I&D: incision and drainage, RBC: red blood cell, ICU: intensive care unit, N/A: not applicable. aP > 0.2 in univariate analysis was excluded from the multivariable regression model. Note: The OR for the ‘Others’ category of surgery type could not be reliably estimated because of the very small number of patients in this category, resulting in (quasi-)complete separation; this estimate, and those for ASA-PS class in the univariate model, should therefore be interpreted with caution. A sensitivity analysis using Firth’s penalized logistic regression, reported in Supplementary Table 2, showed that the estimate for preoperative SAA was not materially affected by this issue.
Other factors associated with POD occurrence
Other factors associated with the occurrence of five-day POD included age (OR = 1.084, 95% CI [1.046‒1.123], P < 0.001) and preoperative albumin levels (OR = 0.470, 95% CI [0.270‒0.820], P = 0.008) (Table 2).
Correlation between SAA levels and Nu-DESC score
A significant positive correlation was identified between preoperative SAA levels and the Nu-DESC score (ρ = 0.404, P < 0.001), indicating that higher SAA levels are associated with more severe delirium symptoms.
Sensitivity analysis
In a sensitivity analysis accounting for the skewed distribution of preoperative SAA, log2-transformed preoperative SAA remained independently associated with POD in the multivariable model (OR = 1.329, 95% CI [1.171–1.509], P < 0.001; Supplementary Table 1), consistent with the primary analysis based on untransformed SAA and supporting the robustness of the observed association. Furthermore, in the sensitivity analysis using Firth’s penalized logistic regression to address separation problems from sparsely observed categories, effect estimates for preoperative SAA remained stable and consistent with those of the original model (Supplementary Table 2).
ROC curve analysis
The discriminatory ability of preoperative SAA levels to differentiate patients with POD from those without POD was assessed using ROC curve analysis (Fig. 2). The area under the curve was 0.812 (95% CI [0.781–0.841]), indicating good discrimination. Based on the Youden index, a cutoff level of 15.9 mg/L was identified, achieving a sensitivity of 81.9% and specificity of 69.0%.
Fig. 2.
ROC curve analysis of the discriminatory ability of preoperative SAA levels for five-day POD. The ROC curve analysis revealed an area under the curve of 0.812 (95% CI [0.781–0.841]), demonstrating good discrimination between patients with and without POD. A preoperative SAA threshold of 15.9 mg/L, identified by the Youden index, yielded a sensitivity of 81.9% and a specificity of 69.0%. ROC: receiver operating characteristic, SAA: serum amyloid A, POD: postoperative delirium.
Discussion
In the present study, we explored the relationship between preoperative SAA levels and the incidence of five-day POD in older adults undergoing hip surgery. Overall, our findings indicate that elevated preoperative SAA levels are significantly associated with an increased risk of developing POD. Additionally, we identified age and preoperative albumin levels as other important factors contributing to POD. These results highlight the potential role of neuroinflammation in POD and indicate that SAA could serve as a valuable biomarker for identifying high-risk patients.
The association between preoperative SAA levels and POD observed in the present study is consistent with the results of the existing literature suggesting that inflammatory processes play a critical role in delirium development. Neuroinflammation, triggered by surgical stress and compounded by the reduced capacity of the aging brain to cope with such stress, is a key mechanism in the pathogenesis of POD [3,9]. SAA, an acute-phase reactant, increases in response to systemic inflammation, and its elevation in the preoperative period may indicate an underlying pro-inflammatory state that could predispose individuals to POD [10]. Although the elevation of preoperative SAA levels observed in the POD group likely reflects a pre-existing inflammatory state, rather than a direct causative factor, it nevertheless serves as a valuable biomarker of vulnerability to neuroinflammation. Prior studies have shown that systemic inflammation and the activation of neuroinflammatory pathways can disrupt normal brain functioning, thereby contributing to delirium onset [3,9]. Consequently, our demonstration of an association between elevated preoperative SAA levels and POD further supports the hypothesis that inflammation plays a pivotal role in the development of this condition.
Other factors identified as being significantly associated with POD in this study included age and preoperative albumin levels. Age has long been recognized as a major risk factor for POD, with older patients exhibiting more pronounced inflammatory responses to surgery as well as a higher likelihood of cognitive decline [11]. Preoperative albumin that reflects an individual’s nutritional status and the body’s ability to manage stress was also found to exert a protective effect against POD. Low albumin levels are often indicative of poor nutritional status that has been associated with an increased risk of postoperative complications, including delirium [12]. Together, these factors emphasize the multifaceted nature of POD risk, and highlight the need for a comprehensive preoperative assessment that considers both inflammatory and non-inflammatory variables.
Despite the strength of our findings, this study has some limitations. Firstly, the retrospective design means that our analysis is subject to limitations inherent to the use of existing medical records, including potential biases related to data accuracy and completeness. Additionally, due to this retrospective design, we were unable to consistently collect data on other inflammatory markers, such as C-reactive protein and erythrocyte sedimentation rate that could have provided a more comprehensive overview of the patients’ inflammatory status. Second, although we controlled for several key variables, this retrospective study was unable to account for other important factors known to influence POD risk, such as preoperative cognitive status, medication history (e.g., anticholinergic use), and perioperative pain and analgesic use that required a more detailed assessment. Additionally, another limitation of this study is that other unmeasured factors, such as the level of C-reactive protein and a patient’s preoperative cognitive function status, may also influence POD risk. Future prospective research should aim to capture this missing information to overcome these limitations. Third, our study only included patients undergoing hip surgery that may limit the generalizability of the findings to other surgical populations. Further prospective studies are therefore needed to validate the role of SAA and other inflammatory markers in POD, as well as to explore the mechanisms by which preoperative inflammation influences delirium onset. Fourth, while this retrospective study that utilized a large existing dataset of 731 patients made an a priori calculation unfeasible, we conducted a post hoc power analysis to evaluate the study’s power to detect the association between preoperative SAA and POD. Based on the logistic regression analysis with our sample size of 731 and a significance level of 0.05, the statistical power to detect the observed OR of 1.004 was > 0.99. This demonstrates that our sample size was sufficient to identify a significant association if one existed; however, this does not address the limitations inherent to the retrospective design. Fifth, in the sensitivity analysis using Firth's penalized logistic regression to address the separation problem, the association between preoperative Hb and POD reached nominal significance (P = 0.047), whereas it was not significant in the original model (P = 0.443), although the OR was similar in magnitude between the two models (1.027 vs. 1.019). This borderline discrepancy likely reflects sensitivity to the estimation method near the conventional significance threshold rather than a substantive change in effect size, and it did not affect the primary finding regarding preoperative SAA. Finally, our study has only investigated the incidence of POD within the first five postoperative days that may not capture late-onset delirium. Consequently, our results cannot be generalized to cover the entire postoperative period.
In conclusion, the present study provides evidence that increased preoperative SAA levels are associated with a higher risk of POD in older adults undergoing hip surgery. This finding indicates that SAA may be a useful biomarker for identifying patients at higher risk of POD, potentially allowing for earlier interventions to mitigate this debilitating condition. Further prospective research into the role of neuroinflammation in POD, together with external validation of biomarkers such as SAA, is required to clarify their potential clinical utility for identifying patients at risk of POD, ultimately improving outcomes for older surgical populations.
Footnotes
Funding
None.
Conflicts of Interest
Bon-Wook Koo has been an editor for the Korean Journal of Anesthesiology since 2023. However, he was not involved in any process of review for this article, including peer reviewer selection, evaluation, or decision-making. There were no other potential conflicts of interest relevant to this article.
Data Availability
The datasets generated during and/or analyzed during the current study are available from the corresponding author on reasonable request.
Author Contributions
Hyun-Jung Shin (Conceptualization; Data curation; Formal analysis; Investigation; Methodology; Resources; Software; Validation; Writing – original draft)
Bon-Wook Koo (Conceptualization; Data curation; Investigation; Methodology; Resources)
Hyo-Seok Na (Conceptualization; Data curation; Formal analysis; Investigation; Supervision; Validation; Writing – review & editing)
Supplementary Materials
Binary logistic regression for the occurrence of POD using log2-transformed SAA.
Multivariable Firth's penalized logistic regression analysis for the occurrence of POD.
References
- 1.Rizk P, Morris W, Oladeji P, Huo M. Review of postoperative delirium in geriatric patients undergoing hip surgery. Geriatr Orthop Surg Rehabil. 2016;7:100–5. doi: 10.1177/2151458516641162. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 2.Kunicki ZJ, Ngo LH, Marcantonio ER, Tommet D, Feng Y, Fong TG, et al. Six-year cognitive trajectory in older adults following major surgery and delirium. JAMA Intern Med. 2023;183:442–50. doi: 10.1001/jamainternmed.2023.0144. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 3.Liu Y, Fu H, Wang T. Neuroinflammation in perioperative neurocognitive disorders: From bench to the bedside. CNS Neurosci Ther. 2022;28:484–96. doi: 10.1111/cns.13794. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 4.Zhang Y, Zhang J, Sheng H, Li H, Wang R. Acute phase reactant serum amyloid A in inflammation and other diseases. Adv Clin Chem. 2019;90:25–80. doi: 10.1016/bs.acc.2019.01.002. [DOI] [PubMed] [Google Scholar]
- 5.World Health Organization Decade of Healthy Ageing: Baseline Report [Internet] Geneva: World Health Organization; 2021 [cited 2025 Jun 25]. Available from https://www.who.int/publications/i/item/9789240017900.
- 6.Kim KN, Kim CH, Kim KI, Yoo HJ, Park SY, Park YH. Development and validation of the Korean Nursing Delirium Scale. J Korean Acad Nurs. 2012;42:414–23. doi: 10.4040/jkan.2012.42.3.414. [DOI] [PubMed] [Google Scholar]
- 7.Inouye SK, van Dyck CH, Alessi CA, Balkin S, Siegal AP, Horwitz RI. Clarifying confusion: the confusion assessment method. A new method for detection of delirium. Ann Intern Med. 1990;113:941–8. doi: 10.7326/0003-4819-113-12-941. [DOI] [PubMed] [Google Scholar]
- 8.Oh ST, Park JY. Postoperative delirium. Korean J Anesthesiol. 2019;72:4–12. doi: 10.4097/kja.d.18.00073.1. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 9.Jia S, Yang H, Huang F, Fan W. Systemic inflammation, neuroinflammation and perioperative neurocognitive disorders. Inflamm Res. 2023;72:1895–907. doi: 10.1007/s00011-023-01792-2. [DOI] [PubMed] [Google Scholar]
- 10.Sack GH., Jr Serum amyloid A - a review. Mol Med. 2018;24:46. doi: 10.1186/s10020-018-0047-0. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 11.Sadeghirad B, Dodsworth BT, Schmutz Gelsomino N, Goettel N, Spence J, Buchan TA, et al. Perioperative factors associated with postoperative delirium in patients undergoing noncardiac surgery: an individual patient data meta-analysis. JAMA Netw Open. 2023;6:e2337239. doi: 10.1001/jamanetworkopen.2023.37239. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 12.Wang W, Yao W, Tang W, Li Y, Lv Q, Ding W. Association between preoperative albumin levels and postoperative delirium in geriatric hip fracture patients. Front Med (Lausanne) 2024;11:1344904. doi: 10.3389/fmed.2024.1344904. [DOI] [PMC free article] [PubMed] [Google Scholar]
Associated Data
This section collects any data citations, data availability statements, or supplementary materials included in this article.
Supplementary Materials
Binary logistic regression for the occurrence of POD using log2-transformed SAA.
Multivariable Firth's penalized logistic regression analysis for the occurrence of POD.