Incidence, Associated Risk Factors, and Outcomes of Postoperative Anxiety in Elderly: A Retrospective Observational Study

ABSTRACT

Background and Aims

Postoperative anxiety (POA) is a frequently underrecognized complication in elderly surgical patients, with potential impacts on recovery and long‐term mental health. This study aimed to investigate the incidence of POA, identify associated risk factors, and evaluate its short‐term consequences in elderly individuals undergoing surgery.

Methods

A retrospective observational study was conducted among patients aged 65 years and older who underwent elective surgery under general anesthesia between May 2020 and March 2021. Anxiety was measured using the Generalized Anxiety Disorder‐7 (GAD‐7) scale at baseline and during the first 7 days postoperatively. Postoperative pain, sleep quality, and other clinical outcomes were evaluated. Multivariable logistic regression identified independent risk factors for POA, and subgroup analyses were performed based on gender, frailty status, and surgery type. Missing data were handled using multiple imputation.

Results

Among 308 eligible patients, 51.9% developed POA within 7 days post‐surgery. The highest incidence occurred after orthopedic (64.6%) and urologic (60%) procedures. POA was significantly associated with worse postoperative pain (higher NRS scores, increased use of rescue analgesia) and poorer sleep quality on postoperative days 1–3. Multivariable analysis revealed that preoperative anxiety (OR, 3.60; 95% CI, 1.76–7.40) and preoperative sleep disturbance (OR, 3.34; 95% CI, 1.82–6.13) were identified as independent risk factors of POA. Anxiety at 30 and 90 days after surgery was significantly increased compared with those without POA (26% vs 15% and 22% vs 12%, respectively).

Conclusion

POA is prevalent in elderly surgical patients and is associated with worse early postoperative outcomes. Screening for anxiety and sleep disturbances before surgery may help identify high‐risk individuals. Early psychological or sleep‐focused interventions could improve recovery and prevent persistent anxiety symptoms.

Abbreviations

ASA

American society of anesthesiologists

BMI

body mass index

CAM

confusion assessment method

GAD‐7

generalized anxiety disorder questionnaire

IQR

interquartile range

LASSO

least absolute shrinkage and selection operator

MMSE

minimum mental state examination

NRS

numeric rating scale

OR

odd ratio

POA

postoperative anxiety

POD

postoperative delirium

PSQ

postoperative sleep quality

STROBE

strengthening the reporting of observational studies in epidemiology

VAS

visual analog scale

1. Introduction

Anxiety disorders—characterized by excessive fear, worry, or avoidance of perceived threats—are among the most prevalent mental health conditions worldwide [1]. Clinically significant anxiety symptoms have been reported in 15% to 52% of individuals in community settings and up to 56% in clinical populations [2]. In older adults, however, the true prevalence may be underestimated due to age‐related diagnostic challenges and atypical symptom presentations.

The perioperative period introduces multiple psychological stressors for elderly patients, including concerns about surgical outcomes, anesthesia safety, postoperative pain, potential complications, loss of independence, caregiver burden, and fear of mortality [3, 4, 5, 6, 7]. The incidence of preoperative anxiety varies widely across clinical settings, with reported rates ranging from 11% to 80%, depending on surgical type, patient demographics, and assessment tools [5, 8, 9, 10, 11]. However, current research efforts have predominantly focused on preoperative anxiety, with limited studies examining anxiety that develops or persists in the postoperative period, particularly in older adults. Nevertheless, with the global rise in the aging population and a corresponding increase in surgical procedures among older adults, perioperative anxiety—particularly POA—represents a growing public health concern.

Perioperative anxiety is recognized as a negative factor that can adversely impact recovery. Preoperative anxiety has been linked to increased postoperative pain [12, 13, 14, 15], higher rates of delirium [16], elevated complication risks [17], longer hospital stays [18], reduced quality of life [12, 17, 19], and even increased mortality [20]. However, the impact of postoperative anxiety remains less well studied. Existing literature indicates that the incidence of POA in elderly surgical patients ranges from 5% to 38%, depending on the type of surgery and assessment tools used [21]. A retrospective study in patients with endometrial cancer found that POA was associated with increased postoperative pain and impaired immune function [22]. Although preoperative anxiety is considered a key predictor of POA [3], the direct influence of POA on recovery outcomes in elderly surgical patients warrants further investigation.

The objectives of this study are threefold: (1) to determine the incidence of postoperative anxiety in elderly surgical patients through retrospective analysis; (2) to identify potential risk factors associated with POA; and (3) to evaluate the short‐term postoperative outcomes related to POA.

2. Methods

2.1. Study Design and Patient Selection

The study protocol (ethical number: 2019 KEY‐E‐115) obtained approval from the Ethics Committee of the First Affiliated Hospital of Guangxi Medical University and adhered to the STROBE (Strengthening the Reporting of Observational Studies in Epidemiology) guidelines for retrospective cohort studies [23]. The research, conducted in accordance with the Helsinki Declaration, ensured patient data confidentiality.

The study included records of elderly patients aged 65 to 90 years who underwent elective non‐cardiac and non‐neurosurgical procedures, under general anesthesia at the First Affiliated Hospital of Guangxi Medical University between May 1, 2020, and March 31, 2021. Patient selection criteria followed the American Society of Anesthesiologists (ASA) classification I–IV. Patients were excluded if they met any of the following criteria: emergency surgery, history of mental illness, communication disorder, current substance abuse, or receipt of sedatives within 1 week before surgery. Throughout the entirety of the study period, patients with missing preoperative or intraoperative relevant information were also excluded from the analysis. Surgeries were categorized into general, orthopedic, and urologic urological procedures based on the surgical site.

2.2. Data Collection Protocol

As our team is dedicated to researching the perioperative recovery of elderly patients, routine evaluations are conducted in our hospital to assess frailty, anxiety, sleep quality, and cognitive function etc., among elderly surgical patients. The study was approved by the Institutional Review Board of the First Affiliated Hospital of Guangxi Medical University (No. 2019 KEY‐E‐115). As this was a retrospective analysis, the requirement for informed consent was waived.

2.3. Preoperative and Intraoperative Data

All preoperative data were obtained through the hospital's anesthesia visit system, a standardized digital platform that records preoperative evaluations, perioperative clinical variables, and anesthesia‐related parameters.

Cognitive impairment was defined based on education level: MMSE score ≤ 19 for patients with ≤ 6 years of education, and ≤ 23 for those with > 6 years, based on published norms in Chinese elderly populations [24]. Frailty was defined as the FRAIL Score ＞ 2, which has been validated in elderly populations [25]. Sleep quality was also evaluated using a 0–10 Numeric Rating Scale (NRS), with higher scores indicating poor sleep. The same sleep scale was used preoperatively and postoperatively for consistency [26].

Intraoperative data, such as surgery duration, estimated blood loss, episodes of hypotension, administration of opioid drugs (fentanyl equivalent doses for various opioids), and peripheral nerve blockade, were documented. Intraoperative hypotension is defined as a systolic blood pressure lower than 90 mmHg or a decrease of more than 20% from baseline. Given the potential impact of dexmedetomidine on anxiety [27, 28], we also recorded the intraoperative utilization of dexmedetomidine.

2.4. Postoperative Follow‐Up Data

Postoperative sleep quality (PSQ) was assessed using the Numeric Rating Scale (NRS) on the first three postoperative nights, with scores ranging from 0 (best possible sleep) to 10 (worst). Postoperative pain intensity was evaluated 24 h post‐surgery using the Visual Analog Scale (VAS), where scores ranged from 0 (no pain) to 10 (unbearable pain). Patients with VAS scores exceeding 4 received rescue analgesia with flurbiprofen, and the need for rescue analgesia was recorded.

Postoperative delirium (POD) was assessed for 7 days post‐surgery using the Confusion Assessment Method (CAM) through twice‐daily assessments [29]. The CAM is a structured interview that evaluates four clinical criteria: acute onset and fluctuating course, inattention, disorganized thinking, and altered level of consciousness. Delirium was diagnosed if the first two criteria and either the third or the fourth criterion were present. POD was diagnosed if CAM criteria were met on any of the assessments, with medical records and family interviews also reviewed for signs of delirium, such as confusion, agitation, hallucinations, delusions, and sedation.

2.5. Anxiety Status Evaluation

The primary outcome was the incidence of postoperative anxiety (POA). All participants were assessed for anxiety status at various time points: preoperative and postoperative days 1, 3, 7, 30, and 90 during the research. Anxiety was assessed with the Generalized Anxiety Disorder Questionnaire (GAD‐7), which evaluates the frequencies at which certain symptoms had been experienced over the last 2 weeks, ranging from 0 (‘not at all’) to 3 (‘nearly every day’). These questionnaires were self‐administered. The cutoff values of GAD‐7, scores range from 0 to 21, with scores over 4 indicating anxiety [30, 31]. During hospitalization, patients were assessed in person by a trained staff member. If discharged, follow‐up evaluations are conducted via telephone. POA was defined as meeting criteria on GAD‐7 greater than 4 at any timepoint assessment during the first 7 days post‐surgery.

2.6. Subgroup Analysis

To further explore differential risks for postoperative anxiety (POA), we conducted exploratory subgroup analyses stratified by gender (male vs. female), frailty status (frail vs. non‐frail), and surgical type (general, orthopedic, or urologic). For each subgroup, POA incidence was calculated, and odds ratios (ORs) with 95% confidence intervals (CIs) were estimated. P values were derived from logistic regression models. These subgroup analyses were not pre‐specified and should be interpreted as exploratory.

2.7. Statistical Analysis

The sample size for this retrospective study was calculated based on the estimated incidence of POA in elderly surgical patients. Considering a margin of error of 5%, a confidence level of 95%, and an anticipated POA incidence of 30% based on preliminary data, after accounting for potential missing data (20%), 405 patients were deemed sufficient to achieve the study objectives with adequate statistical power.

Categorical variables were compared using the chi‐square or Fisher's exact test, and continuous variables using the independent‐sample t‐test or Mann–Whitney U test, as appropriate. Multivariate logistic regression was subsequently performed to identify independent risk factors. Missing data for relevant covariates (including BMI, ASA classification, and preoperative sleep quality) were addressed using multiple imputation by chained equations (MICE) under the assumption of missing at random (MAR). Five imputed datasets were generated, and pooled estimates were reported in subsequent analyses. This approach was chosen to reduce bias and preserve statistical power.

The Least Absolute Shrinkage and Selection Operator (LASSO) regression was employed to execute the multivariate logistic regression. This method employs L1‐penalized least absolute shrinkage and selection regression for multivariable analyses, which is further augmented with 10‐fold cross‐validation for internal validation. The most predictive covariates were selected by the minimum (λ min). Subsequently, based on prior research findings and the results obtained from Lasso and univariate logistic regression analysis, eight risk factors, including gender, BMI, education level, preoperative anxiety, preoperative sleep disturbance, surgery type, use of peripheral nerve block, and use of dexmedetomidine, were incorporated into the multiple regression model.

We also explored whether postoperative anxiety had an impact on patients’ postoperative clinical outcomes. To minimize potential confounding factors, we adjusted the outcomes of binary variables based on the results of univariate logistic regression (p < 0.05). For continuous variables, we similarly adjusted for relevant covariates and compared using the method of least squares mean differences. All data were analyzed using SPSS 22.0 and R statistical software. A two‐tailed p value less than 0.05 was deemed statistically significant.

3. Results

A comprehensive review was conducted on a total of 14,409 patients who underwent elective surgery under general anesthesia at the First Affiliated Hospital of Guangxi Medical University from May 1, 2020, to March 31, 2021. Following the initial screening, 450 patients met the inclusion criteria. However, due to missing preoperative or intraoperative data, 142 patients were excluded, leaving 308 patients for the final analysis.

3.1. Demographic and Intro‐Operative Data

The cohort comprised 308 patients (median [IQR] age, 70 [67–75] years; 168 [54.5%] men and 140 [45.5%] women) who underwent elective surgery. The surgical procedures encompassed 157 (51.0%) elective general, 96 (31.2%) orthopedics, and 55 (17.8%) urologic surgery. In the preoperative health assessment of all patients, 88 patients (28.6%) were identified with preoperative anxiety, 62 (20.1%) exhibited frailty, and 83 patients (26.9%) had sleep disturbance. Approximately 38.7% of patients received PNB for multimodal analgesia, and 228 (74.0%) experienced hypotension during anesthesia and surgery. Table 1 presents all demographic and clinical data of the enrolled patients.

Table 1.

Patient demographics and baseline characteristics.

Variable	Overall, N = 308	Male, N = 168	Female, N = 140
Age, years	70 (67, 75)	70 (67, 75)	71. (67, 76)
BMI, kg/m2	22.6 (20.5, 25.0)	22.3 (20.5, 24.4)	22.9 (20.3, 25.8)
ASA
Ⅱ	218 (70.8%)	123 (73.2%)	95 (67.9%)
Ⅲ	85 (27.6%)	40 (23.8%)	45 (32.1%)
Ⅳ	5 (1.6%)	5 (3.0%)	0 (0.0%)
Smoke	238 (77.3%)	99 (58.9%)	139 (99.3%)
Drink	66 (21.4%)	63 (37.5%)	3 (2.1%)
Education levels (years)
≤ 6	152 (49.4%)	61 (36.3%)	91 (65.0%)
6～12	100 (32.5%)	66 (39.3%)	34 (24.3%)
＞12	56 (18.2%)	41 (24.4%)	15 (10.7%)
Cognitive impairment	40 (13.0%)	19 (11.3%)	21 (15.0%)
Hypertension	107 (34.7%)	49 (29.2%)	58 (41.4%)
Coronary	20 (6.5%)	13 (7.7%)	7 (5.0%)
Cerebrovascular	11 (3.6%)	8 (4.8%)	3 (2.1%)
Diabetes	38 (12.3%)	17 (10.1%)	21 (15.0%)
COPD	15 (4.9%)	10 (6.0%)	5 (3.6%)
Preoperative anxiety	88 (28.6%)	31 (18.5%)	57 (40.7%)
Frailty	62 (20.1%)	35 (20.8%)	27 (19.3%)
Sleep disturbance	83 (26.9%)	44 (26.2%)	39 (27.9%)
Surgery type
General	157 (51.0%)	100 (59.5%)	57 (40.7%)
Orthopedics	96 (31.2%)	30 (17.9%)	66 (47.1%)
Urologic	55 (17.9%)	38 (22.6%)	17 (12.1%)
Dexmedetomidine	115 (37.3%)	55 (32.7%)	60 (42.9%)
PNB	113 (36.7%)	61 (36.3%)	52 (37.1%)
Hypotension	228 (74.0%)	119 (70.8%)	109 (77.9%)
Surgery time, min	140 (91, 196)	155 (97, 219)	125 (90, 176)
Estimated bleeding, ml	50 (30, 150)	55 (30, 150)	50 (30, 163)
Fentanyl consumption, mg	1.20 (0.30, 1.30)	1.20 (0.30, 1.60)	1.15 (0.30, 1.20)

Note: Data were presented as median (IQR) or number (%). Cognitive impairment was defined as the MMSE score of ≤ 20 with education≤ 6, MMSE score of ≤ 24 with education ＞ 6. Anxiety was defined as the GAD‐7 score ＞ 4. Frailty was defined as the FRAIL Score ＞ 2. Sleep Disturbance defined as the sleep‐NRS score ≥ 6.

Abbreviations: ASA, American Society of Anesthesiology; BMI, body mass index; COPD, chronic obstructive pulmonary disease; PNB, peripheral nerve block.

3.2. Incidence of Postoperative Anxiety

The overall incidence of POA was 51.9% (160 of 308 patients), with substantial variability observed among different surgical procedures. POA was most frequently observed after orthopedics surgery (64.6%), followed by urologic (60.0%) and general surgery (41.4%). Figure 1 illustrates the incidence of POA among all the enrolled patients. Notably, the prevalence of anxiety decreased over the initial 7 days post‐surgery, with occurrences on days 1, 3, and 7 at 38.0%, 35.1%, and 17.2%, respectively. The Kaplan‐Meier curve for the incidence of POA is shown in Figure 2.

Figure 1.

Incidence of postoperative anxiety after different types of surgery within the first 7 days after surgery.

Figure 2.

Kaplan–Meier curve for POA within 7 days following surgery.

3.3. Comparisons of Outcomes Between the POA and Non‐POA Groups

Based on the occurrence of POA, we divided the patients into POA group and non‐POA group. Clinical characteristics of the two groups showed significant differences in BMI, ASA score, preoperative anxiety, preoperative sleep disturbances, and surgery type (Table S1). Therefore, we adjusted for these variables when comparing postoperative outcomes between the two groups. Patients who developed POA exhibited significantly higher 24 h pain VAS scores and worsened PSQ scores on postoperative day 1, 2, and 3 than those patients without POA (Table 2). Moreover, rescue analgesic (OR, 1.84; 95% CI, 1.09–3.11) and anxiety on postoperative day 30 and 90 were elevated in patients with POA compared to those without POA. The odds ratio for the POA group was 2.22 (95% CI: 1.16 to 4.24) and 2.43 (95% CI: 1.23–4.80) compared to the non‐POA group on postoperative day 30 and 90 (Table 2). However, the POD was not different in the two groups.

Table 2.

Outcome analyses between postoperative anxiety and non‐postoperative anxiety.

Various	Unadjusted			Adjusted
	OR (95% CI)	Mean difference (95% CI)	P	OR (95% CI)	LS mean difference (95% CI)	P
24 h Pain‐VASa		1.03 (0.66, 1.40)	< 0.001		0.88 (0.51, 1.25)	< 0.001
PSQ‐1a		0.88 (0.50, 1.25)	< 0.001		0.68 (0.30, 1.06)	< 0.001
PSQ‐2a		0.77 (0.41, 1.13)	< 0.001		0.65 (0.28, 1.02)	< 0.001
PSQ‐3a		0.57 (0.22, 0.91)	0.002		0.50 (0.14, 0.85)	0.007
Rescue analgesiab	1.48 (0.93, 2.35)		0.010	1.84 (1.09, 3.11)		0.023
PODb	1.86 (0.83, 4.04)		0.130	1.72 (0.69, 4.29)		0.250
Axiety‐30b (n 1 = 287)	2.04 (1.15, 3.62)		0.015	2.22 (1.16, 4.24)		0.015
Anxiety‐90b (n 1 = 269)	2.02 (1.09, 3.76)		0.026	2.43 (1.23, 4.80)		0.011

Abbreviations: CI, confidential interval; LS, least squares; OR, odds ratio; POD, postoperative delirium; PSQ, postoperative sleep quality; VAS, visual analog scale.

^a Mean (Standard Deviation).

^b Number (%).

¹ There were 21 patients and 39 patients lost to follow‐up, and ultimately, 281 patients and 269 patients were included in the analysis for anxiety at postoperative days 30 and 90. Analyses were adjusted for BMI, ASA, preoperative anxiety, sleep disturbance, and surgery type.

3.4. Risk Factors for POA

According to the univariate logistic regression analysis, four distinct variables were linked to the occurrence of POA: BMI (OR, 1.08; 95% CI, 1.01–1.15), preoperative anxiety (OR, 4.36; 95% CI, 2.49–7.62), preoperative sleep disturbance (OR, 4.53; 95% CI, 2.54–8.05), and surgery type (general vs. orthopedics, OR, 0.39; 95% CI, 0.23–0.66) were associated with POA (Table S2). Multivariate logistic regression analysis in Figure 3 indicated that preoperative anxiety (OR, 3.60; 95% CI, 1.76–7.40), preoperative sleep disturbance (OR, 3.34; 95% CI, 1.82–6.13) were significantly associated with the development of POA. Therefore, age and preoperative sleep disturbance were found to be independent risk factors for POA in this particular cohort.

Figure 3.

Forest plot of independent predictors of postoperative anxiety based on the multivariate logistic regression model. BMI, body mass index; CI, confidential interval; OR, odds ratio; PNB, peripheral nerve block.

3.5. Subgroup Analyses

The results of subgroup analyses are presented in Table S3. There was no significant difference in POA incidence by gender or frailty status. POA occurred in 52% of both male and female patients (OR = 1.01, 95% CI: 0.81–1.25, p = 0.95). Similarly, POA incidence among frail patients (52%) was comparable to that in non‐frail patients (48%) (OR = 0.98, 95% CI: 0.56–1.72, p = 0.95). However, the surgical type was significantly associated with POA incidence. Patients undergoing orthopedic surgery had the highest incidence (65%), with a significantly increased odds of POA (OR = 2.58, 95% CI: 1.53–4.36, p = 0.0003) compared to general surgery. Urologic surgery was also associated with a higher POA rate (60%) (OR = 2.12, 95% CI: 1.14–3.97, p = 0.0173).

4. Discussion

Our study revealed a high incidence of postoperative anxiety (POA) in elderly patients, with orthopedic (64.6%) and urologic (60%) surgeries being most associated.

This aligns with previous research showing significant rates of POA, although studies focusing on this topic remain limited. For example, Caumo et al. reported that 76.9% of hysterectomy patients experiencing moderate to severe pain exhibited high state anxiety 24 h postoperatively [32]. In our study, we observed a higher incidence of POA within the first 7 days after surgery, which is understandable given our focus on elderly individuals. Due to the decline in physical function and the potential presence of chronic diseases in elderly patients, their susceptibility and vulnerability lead to a higher probability of perioperative complications [33, 34]. The type of surgery appears to significantly influence POA development. Orthopedic and urologic procedures, which often involve functional impairments and prolonged rehabilitation, may exacerbate concerns about postoperative recovery and independence, thereby elevating anxiety levels.

Contrary to some earlier studies, our analysis did not find education level to be significantly associated with POA. Matsumura et al. indicate that a lower education level was an independent risk factor for postpartum depression [35]. In addition, among people in France without surgery, individuals with a lower level of education had a higher risk of anxiety‐depressive state [36]. However, the focus of this study was to assess POA in the elderly as well as its importance in the elderly undergoing surgery. New research conducted by Samudio‐Cruz et al. revealed that a higher education level reduces the risk of depression and anxiety, but their effect is less consistent in older adults after stroke; the education level with cognitive reserve may explain the results [37]. Cognitive reserve refers to the brain's resilience against neuropathological damage and is often associated with education, occupational complexity, and lifelong cognitive engagement. It may allow elderly individuals to better adapt to psychological stressors, attenuating the influence of educational attainment on anxiety [38]. This could potentially lead to the protective effect of a high level of education on mental disorders in older patients diminishing, or even becoming a harmful factor. However, since cognitive reserve was not directly measured in our study, this hypothesis remains speculative and warrants further research using validated tools.

Our study has found that patients who are more anxious in the preoperative showed an estimated 3.6 times higher risk of reporting POA. This result is in agreement with previous research [3, 6]. However, due to the lack of systematic research evaluating POA, our results suggest that intervention in preoperative anxiety could potentially reduce the incidence of POA, thereby promoting patient recovery.

Preoperative sleep disturbance was another independent risk factor for POA (OR: 3.34, 95% CI: 1.82–6.13). This is consistent with existing literature, including a population‐based study by Kim, which found that self‐reported insomnia was closely associated with elevated anxiety and depressive symptoms, particularly among patients with chronic conditions such as migraines [39]. Sleep and anxiety are believed to share a bidirectional relationship, wherein impaired sleep quality increases vulnerability to anxiety, and elevated anxiety levels, in turn, further disrupt sleep architecture. Mechanistically, disruptions in circadian rhythms, heightened hypothalamic–pituitary–adrenal (HPA) axis activity, and altered neuroimmune pathways have all been implicated in this interplay [40]. Data from epidemiological studies propose that individuals suffering from insomnia tend to exhibit a higher prevalence of anxiety symptoms [41]. Similarly, those who are anxious are more prone to experience sleep problems [42]. POA significantly worsened postoperative sleep quality among elderly patients, suggesting a potential bidirectional relationship [40]. Our data also show that POA significantly worsens postoperative sleep quality, reinforcing the clinical relevance of this interplay. These findings suggest that perioperative sleep screening and potential interventions to improve sleep quality—such as non‐pharmacologic behavioral therapies or appropriate pharmacologic support—may serve as a modifiable target to reduce POA risk, especially in high‐risk elderly populations. Despite confirming the association between preoperative sleep disturbance and increased risk of POA, our study did not evaluate the role of any specific interventions targeting sleep improvement. This represents a key area for future research. Melatonin, for instance, has been shown to reduce both preoperative and postoperative anxiety, and its use may be particularly relevant in elderly populations due to its favorable safety profile and dual impact on circadian rhythm and emotional regulation. Similarly, non‐pharmacologic strategies such as sleep hygiene education, cognitive behavioral therapy for insomnia (CBT‐I), and relaxation techniques have demonstrated efficacy in improving sleep quality and reducing anxiety in both surgical and non‐surgical populations [43].

In addition to the identified risk factors, such as preoperative anxiety and sleep disturbance, there are other potential contributors to POA that merit consideration. For instance, although patients with a known history of mental illness were excluded from our study, subclinical psychiatric conditions or previously undiagnosed mood disorders could have influenced POA development. Moreover, perioperative opioid use, while standardized and recorded, has been associated with neuropsychiatric side effects, including anxiety, particularly during the postoperative tapering phase. While our study accounted for fentanyl‐equivalent consumption, future studies should further examine the psychotropic effects of different analgesic protocols. Pain management strategies, including the use of peripheral nerve blocks (PNB) and dexmedetomidine, may also modulate anxiety by reducing pain intensity or through direct central effects on stress response and sedation. Our multivariate model included these variables, yet their psychological influence might be multifaceted and deserves dedicated analysis in future prospective designs.

Our subgroup analyses showed no significant differences in POA incidence by gender or frailty, diverging from previous studies. This could reflect sample size limitations or resilience in our elderly population. In contrast, the surgical type showed a clear association with POA risk. Patients undergoing orthopedic and urologic surgeries exhibited significantly higher POA incidence compared to those undergoing general surgery. This may reflect the anticipated postoperative pain, functional impairment, and longer rehabilitation periods commonly associated with orthopedic and urologic procedures. This finding can inform targeted psychological interventions for high‐risk surgical populations.

POA can significantly exacerbate postoperative pain (median: 3.0 vs. 4.0). Numerous studies have established a close relationship between pain and perioperative anxiety. For instance, pain has been independently linked with anxiety in various types of surgery [14, 15, 44]. Pain directly triggers mood dysregulation, as evidenced by increased activity in the anterior cingulate gyrus (a brain region associated with mood regulation) when peripheral pain intensifies. Concurrently, anxiety and depression can heighten the perception of pain. This is due to the depletion of serotonin and norepinephrine found in depression and anxiety, which downregulates the pain‐modulating serotonergic and noradrenergic neurons in the periaqueductal gray matter. This results in an amplification of minor pain signals from the body [45]. These findings underscore the complex interplay between pain and anxiety in the perioperative period.

Several limitations must be acknowledged. As a retrospective observational study, it becomes challenging to establish causal relationships between risk factors and POA. In addition, the current study suffers from a small sample size and is derived from a single center, potentially limiting the generalizability of the conclusions drawn in the research. The use of a GAD‐7 score > 4 to define POA may identify mild cases that are not clinically significant, although even mild anxiety has been shown to impair recovery in elderly patients. Our reliance on retrospective assessments for postoperative delirium (POD) may underestimate its true incidence due to underreporting. Additionally, being a single‐center study with a moderate sample size limits the generalizability of our findings. These factors should be taken into consideration when interpreting the results of the study.

5. Conclusion

This study shows that POA is common in elderly patients and is associated with worse pain, poor sleep, and ongoing anxiety after discharge. Preoperative anxiety and sleep disturbance were key risk factors. These findings highlight the need for early screening and targeted interventions to improve outcomes in high‐risk elderly surgical patients.

Author Contributions

Hao Guo: conceptualization, data curation, formal analysis, investigation, methodology, project administration, writing – original draft, writing – review and editing. Li‐Heng Li: conceptualization, formal analysis, investigation, methodology, writing – original draft, writing – review and editing. Xiao‐Hong Lv: conceptualization, data curation, investigation, methodology, writing – original draft, writing – review and editing. Fei Xiao: conceptualization, data curation, investigation, methodology, writing – original draft, writing – review and editing. Yu‐Bo Xie: conceptualization, data curation, formal analysis, funding acquisition, methodology, project administration, software, supervision, validation, visualization, writing – original draft, writing – review and editing.

Conflicts of Interest

The authors declare no conflicts of interest.

Author Responsibility Statement

All authors have read and approved the final version of the manuscript. Yu‐Bo Xie, the corresponding author, had full access to all of the data in this study and takes complete responsibility for the integrity of the data and the accuracy of the data analysis. All individuals acknowledged have given their permission to be named. Their contribution does not imply endorsement of the study's findings.

Transparency Statement

The lead author Yu‐Bo Xie affirms that this manuscript is an honest, accurate, and transparent account of the study being reported; that no important aspects of the study have been omitted; and that any discrepancies from the study as planned (and, if relevant, registered) have been explained.

Supporting information

Table S1: Patient demographics and baseline characteristics between the group POA and non‐POA. Table S2: Univariate and multivariate analysis of influencing factors (Logistic regression). Table S3: Subgroup analysis of influencing factors for POA.

Guo H., Li L.‐H., Lv X.‐H., Xiao F., and Xie Y.‐B., “Incidence, Associated Risk Factors, and Outcomes of Postoperative Anxiety in Elderly: A Retrospective Observational Study,” Health Science Reports 9 (2026): 1‐9, 10.1002/hsr2.71682.

Data Availability Statement

All relevant data are provided within the manuscript and supporting files. Additional data may be available upon reasonable request.