Preoperative anxiety and depression predict two-year outcomes after total hip or knee arthroplasty

Abstract

BACKGROUND

Although total joint arthroplasty effectively treats end-stage joint diseases, 10%-30% of patients experience poor outcomes. Preoperative anxiety and depression are common, but their long-term prognostic impact remains unclear. This study aimed to investigate the association between preoperative anxiety/depression and two-year outcomes after total hip or knee arthroplasty.

AIM

To evaluate the impact of preoperative anxiety and depression on joint function, pain, quality of life, and complication occurrence at 2 years postoperatively in patients undergoing total hip arthroplasty (THA) and total knee arthroplasty (TKA).

METHODS

A case-control study design was employed, retrospectively enrolling 320 patients who underwent primary unilateral THA or TKA between January 2021 and December 2022, with 246 patients completing a 2-year follow-up (follow-up rate 76.9%). The Hospital Anxiety and Depression Scale (HADS) was used to assess preoperative psychological status, dividing patients into an exposed group (HADS-Anxiety Subscale ≥ 8 points and/or HADS-Depression Subscale ≥ 8 points) and a non-exposed group (HADS-Anxiety Subscale < 8 points and HADS-Depression Subscale < 8 points). The 1:1 propensity score matching was applied to balance baseline characteristics, with 92 cases included in each group after matching. Harris score and Hospital for Special Surgery score were used to evaluate joint function, the Visual Analogue Scale to assess pain, the Short Form-36 Health Survey to evaluate quality of life, and postoperative complications were recorded. Repeated measures analysis of variance was used to compare outcome indicators at different postoperative time points between groups, and multivariate logistic regression was performed to analyze the independent effect of preoperative anxiety and depression.

RESULTS

After matching, baseline characteristics were balanced between the two groups (P > 0.05). At 1 and 2 years postoperatively, joint function scores in the exposed group were lower than in the non-exposed group: Harris scores at 2 years postoperatively for THA patients were 82.6 ± 8.7 vs 89.3 ± 6.9 (P < 0.001), and Hospital for Special Surgery scores for TKA patients were 80.8 ± 8.3 vs 87.1 ± 7.5 (P = 0.001). At 1 and 2 years postoperatively, Visual Analogue Scale scores at rest and during activity in the exposed group were higher than in the non-exposed group (P < 0.001). At 2 years postoperatively, Short Form-36 Health Survey scores and improvement magnitudes in all dimensions were lower in the exposed group than in the non-exposed group (P < 0.05). The total complication rate at 2 years postoperatively in the exposed group was 29.3% (27/92), higher than 14.1% (13/92) in the non-exposed group (P = 0.010). Multivariate logistic regression analysis showed that after adjusting for confounding factors, preoperative anxiety and depression were independent risk factors for poor joint function at 2 years postoperatively (odds ratio = 3.84, 95% confidence interval: 1.92-7.68, P < 0.001) and complication occurrence (odds ratio = 2.67, 95% confidence interval: 1.28-5.57, P = 0.009).

CONCLUSION

Preoperative anxiety and depression are independent risk factors for poor joint function and complication occurrence at 2 years postoperatively in THA and TKA patients, producing persistent negative impacts on pain relief and quality of life improvement. Clinical practice should emphasize perioperative psychological assessment and intervention to improve overall patient prognosis.

Key Words: Anxiety; Depression; Arthroplasty; Replacement; Hip; Treatment outcome; Complications; Propensity score

Core Tip: Anxiety and depression, as measured by the Hospital Anxiety and Depression Scale, before surgery have a long-term adverse effect on total hip arthroplasty and total knee arthroplasty outcomes. This study, which incorporates propensity score matching and two-year follow-up, shows anxiety and depression is an independent predictor of joint function, pain relief, quality of life, and complications. This research highlights the importance of psychological screening and interventions during the perioperative period in arthroplasty.

INTRODUCTION

Total hip arthroplasty (THA) and total knee arthroplasty (TKA) are effective interventions for treating end-stage joint diseases such as osteoarthritis and femoral head necrosis, significantly improving patients’ joint function, relieving pain, and enhancing quality of life[1]. The number of joint arthroplasty surgeries performed each year is growing with the world’s ageing population and improvements in surgical technologies. Yet, despite significant improvements in surgical approaches and implant materials, 10%-30% of patients continue to have unsatisfactory outcomes following joint arthroplasty, including persistent pain, lack of recovery in joint function, or little improvement in quality of life[2]. Biomedical factors, such as surgical methods and prostheses, are not the only factors that influence postoperative outcomes of joint arthroplasty, recent studies have discovered that patients' psychological factors, especially preoperative anxiety and depression, are also important[3].

Preoperative anxiety and depression are the most common psychological problems in joint arthroplasty patients, ranging from 20% to 40% of patients in the literature[4]. Preoperative anxiety and depression mainly result from chronic pain, joint dysfunction and surgery-related concerns. Earlier studies have demonstrated that preoperative anxiety and depression can influence postoperative recovery in several ways, such as by reducing pain threshold, affecting the function of the neuroendocrine system, reducing compliance with rehabilitation training and increasing inflammatory response[5]. But previous studies have mainly focused on short-term postoperative outcomes, and there is still limited evidence-based medical evidence on the effect of anxiety and depression on medium-to-long-term (such as 2-year) postoperative joint function, pain, quality of life and complications, particularly their independent effect after adjusting for potential confounding factors[6]. Moreover, the sample sizes of most studies are small, and they do not use propensity score matching to adjust patient characteristics differences between groups to avoid potential selection bias. Thus, this study is designed to comprehensively assess the effect of preoperative anxiety and depression on joint function, pain, quality of life, and incidence of complications 2 years after surgery in THA and TKA patients, and to reveal its independent effect via multivariable analysis, in order to provide evidence for clinical joint arthroplasty perioperative psychological screening and intervention, and to improve the medium- to long-term prognosis of joint arthroplasty patients[7].

MATERIALS AND METHODS

Study design and subjects

A total of 320 patients who underwent primary unilateral THA or TKA in our hospital’s orthopedic department between January 2021 and December 2022 were retrospectively enrolled. All patients were tracked until 2 years postoperatively or until the occurrence of study endpoint events. Ultimately, 246 patients completed the full 2-year postoperative follow-up, with a follow-up completion rate of 76.9%. Among the 74 lost-to-follow-up patients, 23 were lost due to changed contact information, 18 refused to continue participating in follow-up, 16 relocated, 10 were hospitalized for other diseases, and 7 died (all deaths were unrelated to surgery). Lost-to-follow-up patients showed no statistical differences from patients who completed follow-up in baseline characteristics such as age, gender, body mass index (BMI), and surgery type (P > 0.05).

Based on preoperative anxiety and depression assessment results, patients were divided into two groups: (1) Exposed group (case group): Patients with preoperative anxiety and/or depression symptoms [Hospital Anxiety and Depression Scale-Anxiety Subscale (HADS-A) ≥ 8 points and/or Hospital Anxiety and Depression Scale-Depression Subscale (HADS-D) ≥ 8 points]; and (2) Non-exposed group (control group): Patients without preoperative anxiety or depression symptoms (HADS-A < 8 points and HADS-D < 8 points).

Inclusion criteria: (1) Age 18-80 years; (2) Primary unilateral THA or TKA performed for benign diseases such as osteoarthritis or femoral head necrosis; (3) Completed preoperative anxiety and depression assessment; (4) Completed full 2-year postoperative follow-up; and (5) Complete clinical data.

Exclusion criteria: (1) Revision surgery or bilateral simultaneous surgery; (2) Joint arthroplasty performed for malignant tumors, severe infections, etc.; (3) Complicated by severe dysfunction of vital organs such as heart, lung, liver, or kidney; (4) History of severe psychiatric disorders such as schizophrenia, bipolar disorder, dementia, or long-term use of antipsychotic medications; (5) Severe bone defects requiring special treatment discovered preoperatively or intraoperatively; and (6) Patients who received systematic psychological treatment or anti-anxiety/depression medication postoperatively (to control confounding factors).

Data collection

Baseline data included: (1) General demographic characteristics: Gender, age, BMI, education level, marital status; (2) Surgery-related data: Surgery type (THA/TKA), surgical side, primary disease diagnosis; (3) Comorbidities: Hypertension, diabetes, cardiovascular disease, etc.; and (4) Preoperative laboratory indicators: Hemoglobin, albumin, blood glucose, etc.

Preoperative psychological assessment: The HADS was used to assess patients' anxiety and depression status within 1 week before surgery. HADS includes two dimensions: The anxiety subscale (HADS-A) and the depression subscale (HADS-D), each containing 7 items. Each item uses a 4-point scale (0-3 points), with a total score range of 0-21 points. HADS-A determines the state of anxiety (tension, worry, panic, restlessness); HADS-D determines the state of depression (loss of interest, loss of pleasure, psychomotor retardation, pessimism about the future). An increase in scores reflects more serious anxiety or depression symptoms. The cutoff was determined as 8 points with 8 or more points showing the presence of depression or anxiety symptoms.

Outcome indicators

Postoperative joint function was the primary outcome indicator, while pain, quality of life, and complications were the secondary outcome indicators. Specific indicators and methods of assessment were as follows: (1) Joint function: For THA patients, the Harris Hip Score was used, including four dimensions: Pain (44 points), function (47 points), deformity (4 points), and range of motion (5 points), with a total score of 100 points. The higher the score, the better the hip joint function: < 70 points = poor, 70-79 points = fair, 80-89 points = good, 90-100 points = excellent. For TKA patients, the Hospital for Special Surgery (HSS) Knee Score was used, including six dimensions: Pain (30 points), function (22 points), range of motion (18 points), muscle strength (10 points), flexion deformity (10 points), and stability (10 points), with a total score of 100 points. The higher the score, the better the knee joint function. Time points for assessment: Pre-operation and 1 and 2 years after operation; and (2) Pain evaluation: Visual Analogue Scale (VAS) was adopted, where the line is 10 cm and points are 0-10 points, with 0 points being no pain and 10 points being the worst pain. Higher scores indicate more severe pain. Pain intensity at rest and during activity was assessed at each time point. Assessment time points were preoperatively, at 1 year postoperatively, and at 2 years postoperatively; (3) Quality of life assessment: The Short Form-36 Health Survey (SF-36) was used for evaluation, including eight dimensions: Physical functioning (10 items), role physical (4 items), bodily pain (2 items), general health (5 items), vitality (4 items), social functioning (2 items), role emotional (3 items), and mental health (5 items), totaling 36 items. Each dimension score ranges from 0 point to 100 points, using a standardized scoring method, with higher scores indicating better quality of life. Assessment time points were preoperatively and at 2 years postoperatively; and (4) Recording of relevant complications.

Statistical analysis

SPSS 26.0 software was used for statistical analysis. The propensity score matching method was employed to balance baseline characteristics between the two groups, with preoperative anxiety and depression as the treatment variable. Covariates, including age, gender, BMI, surgery type, and comorbidities, were incorporated into a logistic regression model to generate propensity scores. The 1:1 nearest neighbor matching was used with a caliper value set at 0.02. When assessing matching quality, all covariates were required to have standardized differences < 0.1. Continuous variables were expressed as mean ± SD, and an independent samples t-test was used for between-group comparisons. Categorical variables were expressed as n (%), and the χ² test was used for between-group comparisons; Fisher’s exact test was used when the theoretical frequency was < 5. Longitudinal data were analyzed using repeated measures analysis of variance (ANOVA) to test for time effects, between-group effects, and interaction effects. Post-hoc pairwise comparisons were corrected using the Bonferroni method, with correction factors determined based on the number of comparisons. Multivariate logistic regression was used to analyze the impact of preoperative anxiety and depression on postoperative outcomes, adjusting for relevant confounding factors. Continuous variables were dichotomized based on clinically relevant thresholds, and the Hosmer-Lemeshow test was used to assess model goodness of fit. All tests were two-sided, with P < 0.05 considered statistically significant.

Ethical considerations

This study was approved by our hospital’s ethics committee. As a retrospective study, informed consent was waived, but patient privacy protection principles were strictly observed, and all data were anonymized. The study followed the ethical principles of the Declaration of Helsinki. Any important health-related information about patients discovered during the study was promptly communicated to patients and their attending physicians.

RESULTS

Comparison of baseline characteristics

Before matching, there were 108 cases in the exposed group and 138 cases in the non-exposed group. After 1:1 propensity score matching, 92 patients were included in each group. After matching, the two groups showed good balance in baseline characteristics, including age, gender, BMI, surgery type, comorbidities, and preoperative joint function scores, with no statistically significant differences (P > 0.05), and standardized differences of all covariates were < 0.1 (Table 1).

Table 1 Comparison of baseline characteristics between the two groups before and after propensity score matching, n (%).

Variable	Before matching			After matching
	Exposed group (n = 108)	Non-exposed group (n = 138)	P value	Exposed group (n = 92)	Non-exposed group (n = 92)	P value
General demographic characteristics
Age (years), mean ± SD	62.4 ± 8.7	60.8 ± 9.3	0.172	62.1 ± 8.5	61.8 ± 8.9	0.811
Gender			0.043			0.856
Male	38 (35.2)	68 (49.3)		36 (39.1)	37 (40.2)
Female	70 (64.8)	70 (50.7)		56 (60.9)	55 (59.8)
BMI (kg/m2), mean ± SD	26.3 ± 3.4	25.6 ± 3.2	0.098	26.1 ± 3.3	25.9 ± 3.4	0.682
Education level			0.156			0.791
Junior high school or below	45 (41.7)	48 (34.8)		38 (41.3)	36 (39.1)
High school/technical secondary school	38 (35.2)	52 (37.7)		33 (35.9)	35 (38.0)
College or above	25 (23.1)	38 (27.5)		21 (22.8)	21 (22.9)
Marital status			0.287			0.924
Married	89 (82.4)	121 (87.7)		77 (83.7)	78 (84.8)
Unmarried/divorced/widowed	19 (17.6)	17 (12.3)		15 (16.3)	14 (15.2)
Surgery-related data
Surgery type			0.029			0.887
THA	58 (53.7)	92 (66.7)		52 (56.5)	53 (57.6)
TKA	50 (46.3)	46 (33.3)		40 (43.5)	39 (42.4)
Surgical side			0.521			0.832
Left	51 (47.2)	71 (51.4)		45 (48.9)	46 (50.0)
Right	57 (52.8)	67 (48.6)		47 (51.1)	46 (50.0)
Primary disease			0.412			0.765
Osteoarthritis	76 (70.4)	104 (75.4)		67 (72.8)	69 (75.0)
Femoral head necrosis	24 (22.2)	26 (18.8)		19 (20.7)	18 (19.6)
Other	8 (7.4)	8 (5.8)		6 (6.5)	5 (5.4)
Comorbidities
Hypertension	46 (42.6)	51 (37.0)	0.379	38 (41.3)	37 (40.2)	0.879
Diabetes	28 (25.9)	30 (21.7)	0.447	23 (25.0)	22 (23.9)	0.867
Cardiovascular disease	15 (13.9)	14 (10.1)	0.369	12 (13.0)	11 (12.0)	0.826
Preoperative laboratory indicators
Hemoglobin (g/L), mean ± SD	128.6 ± 15.3	131.4 ± 14.8	0.143	129.2 ± 15.1	130.1 ± 14.6	0.687
Albumin (g/L), mean ± SD	39.8 ± 4.2	40.6 ± 3.9	0.112	40.1 ± 4.1	40.3 ± 4.0	0.738
Fasting blood glucose (mmol/L), mean ± SD	5.8 ± 1.3	5.6 ± 1.2	0.238	5.7 ± 1.3	5.7 ± 1.2	0.914

THA: Total hip arthroplasty; TKA: Total knee arthroplasty; BMI: Body mass index.

Comparison of preoperative and postoperative joint function scores between the two groups

After matching, there was no statistically significant difference in preoperative joint function scores between the two groups (P > 0.05). At 1 year and 2 years postoperatively, joint function scores in the exposed group were lower than in the non-exposed group, with statistically significant differences (P < 0.05). Repeated measures ANOVA showed that time effects, between-group effects, and interaction effects were all statistically significant (P < 0.05). Further post-hoc pairwise comparisons after Bonferroni correction (correction factor of 2, adjusted α = 0.025) showed that differences between the two groups at 1 year and 2 years postoperatively remained statistically significant (P < 0.025) (Figure 1 and Table 2).

Figure 1 Postoperative functional recovery: Comparison of Harris Hip and Hospital for Special Surgery Knee Scores between exposure groups over 2 years. A: Comparison of Harris Hip Scores between groupsin total hip arthroplasty patients. The graph shows the trends in Hardship scores for the non-exposure group (n = 53) and exposure group (n = 42) at preoperative, 1-year postoperative, and 2-year postoperative time points. Both groups had similar baseline scores (approximately 45) preoperatively, with improvement over time. At 2 years postoperatively, the non-exposure group reached approximately 90 points, while the exposure group reached approximately 83 points. Error bars indicate standard deviation; B: Comparison of Hospital for Special Surgery Knee Scores between groups in total knee arthroplasty patients. The graph shows the trends in Hospital for Special Surgery knee scores for the non-exposure group (n = 39) and exposure group (n = 40) at preoperative, 1-year postoperative, and 2-year postoperative time points. Both groups had similar baseline scores (approximately 50) preoperatively, with significant improvement over time. At 2 years postoperatively, the non-exposure group reached approximately 87 points, while the exposure group reached approximately 81 points. Error bars indicate standard deviation. HSS: Hospital for Special Surgery.

Table 2 Comparison of preoperative and postoperative joint function scores between the two groups (mean ± SD, points)¹.

Surgery type	Group	n	Preoperative	1 year postoperative	2 years postoperative	F time	P time	F between-groups	P between-groups	F interaction	P interaction
THA (Harris Score)	Exposed group	52	46.3 ± 8.5	78.4 ± 9.2	82.6 ± 8.7	486.32	< 0.001	18.74	< 0.001	6.85
	Non-exposed group	53	47.1 ± 8.2	85.7 ± 7.4	89.3 ± 6.9
	t-value		0.485	4.526	4.389
	P value		0.629	< 0.001	< 0.001
TKA (HSS Score)	Exposed group	40	48.6 ± 9.3	76.2 ± 8.6	80.8 ± 8.3	412.67	< 0.001	15.28	< 0.001	5.92	0.004
	Non-exposed group	39	49.2 ± 8.9	83.4 ± 7.2	87.1 ± 7.5
	t-value		0.295	4.058	3.546
	P value		0.769	< 0.001	0.001

¹P-values for between-group comparisons at postoperative time points have been Bonferroni-corrected (since two time points were compared - 1 year and 2 years postoperatively - the correction factor is 2, with adjusted α level of 0.025).

THA: Total hip arthroplasty; TKA: Total knee arthroplasty; HSS: Hospital for Special Surgery.

Comparison of preoperative and postoperative pain scores between the two groups

There were no statistically significant differences in preoperative VAS scores at rest and during activity between the two groups (P > 0.05). At 1 year and 2 years postoperatively, VAS scores at rest and during activity in the exposed group were higher than in the non-exposed group, with statistically significant differences (P < 0.05). Repeated measures ANOVA showed that for pain at rest, time effect (F = 652.38, P < 0.001), between-group effect (F = 21.45, P < 0.001), and interaction effect (F = 8.73, P < 0.001) were all statistically significant; for pain during activity, time effect (F = 589.26, P < 0.001), between-group effect (F = 25.82, P < 0.001), and interaction effect (F = 9.54, P < 0.001) were all statistically significant. Simple effect analysis showed that differences in pain scores between the exposed group and non-exposed group at 1 year and 2 years postoperatively were all statistically significant (P < 0.001), and the magnitude of pain improvement in the exposed group was smaller than in the non-exposed group, suggesting that preoperative anxiety and depression had a persistent negative impact on postoperative pain relief (Table 3).

Table 3 Comparison of preoperative and postoperative pain Visual Analogue Scale scores between the two groups (mean ± SD, points)¹.

Pain state	Group	n	Preoperative	1 year postoperative	2 years postoperative
At rest	Exposed group	92	4.8 ± 1.6	1.8 ± 0.9	1.5 ± 0.8
	Non-exposed group	92	4.7 ± 1.5	1.2 ± 0.7	0.9 ± 0.6
	t-value		0.432	4.986	5.437
	P value		0.666	< 0.001	< 0.001
During activity	Exposed group	92	6.9 ± 1.8	2.6 ± 1.1	2.2 ± 1.0
	Non-exposed group	92	6.8 ± 1.7	1.7 ± 0.9	1.3 ± 0.8
	t-value		0.379	5.974	6.526
	P value		0.705	< 0.001	< 0.001

¹Simple effect analysis showed that between-group differences were mainly reflected at the 1-year and 2-year postoperative time points. Repeated measures ANOVA showed that the time effect, between-group effect, and interaction effect were all statistically significant.

VAS: Visual Analogue Scale.

Comparison of preoperative and 2-year postoperative quality of life scores between the two groups

There were no statistically significant differences in preoperative SF-36 dimension scores between the two groups (P > 0.05). At 2 years postoperatively, the exposed group had lower scores than the non-exposed group in all eight dimensions: Physical functioning, role physical, bodily pain, general health, vitality, social functioning, role emotional, and mental health, with statistically significant differences (P < 0.05). Further calculation of the change from preoperative to 2 years postoperative showed that the magnitude of improvement in all dimensions was smaller in the exposed group than in the non-exposed group, with statistically significant differences (P < 0.05) (Table 4).

Table 4 Comparison of preoperative and 2-year postoperative Short Form-36 Health Survey quality of life scores between the two groups (mean ± SD, points)¹.

Dimension	Exposed group (n = 92)		Non-exposed group (n = 92)		Change (2 years postoperative - preoperative)
	Preoperative	2 years postoperative	Preoperative	2 years postoperative	Exposed group	Non-exposed group
Physical functioning	42.6 ± 12.3	71.4 ± 13.8	43.2 ± 11.9	81.7 ± 11.2	28.8 ± 10.5	38.5 ± 9.8
Role physical	38.5 ± 15.6	68.3 ± 16.4	39.1 ± 14.8	79.2 ± 13.7	29.8 ± 12.7	40.1 ± 11.3
Bodily pain	35.7 ± 13.2	72.8 ± 14.6	36.3 ± 12.8	83.5 ± 12.3	37.1 ± 11.9	47.2 ± 10.6
General health	48.3 ± 14.5	64.7 ± 15.2	49.1 ± 13.9	74.9 ± 13.6	16.4 ± 9.8	25.8 ± 9.2
Vitality	46.2 ± 13.7	62.5 ± 14.3	47.1 ± 13.2	72.8 ± 12.8	16.3 ± 10.4	25.7 ± 9.6
Social functioning	51.4 ± 15.3	70.2 ± 14.9	52.3 ± 14.8	81.6 ± 13.2	18.8 ± 11.2	29.3 ± 10.5
Role emotional	44.8 ± 16.2	66.9 ± 15.7	45.6 ± 15.7	78.3 ± 14.5	22.1 ± 12.3	32.7 ± 11.8
Mental health	49.5 ± 14.8	65.3 ± 15.6	50.3 ± 14.2	76.8 ± 13.9	15.8 ± 10.7	26.5 ± 9.9

¹Change represents the 2-year postoperative score minus the preoperative score, indicating the magnitude of improvement; change comparison used an independent samples t-test.

SF-36: Short Form-36 Health Survey.

Comparison of complication occurrence at 2 years postoperatively between the two groups

During the 2-year follow-up period, the total complication rate in the exposed group was 29.3% (27/92), higher than 14.1% (13/92) in the non-exposed group, with a statistically significant difference (χ² = 6.582, P = 0.010) (Table 5).

Table 5 Comparison of complication occurrence at 2 years postoperatively between the two groups, n (%).

Complication type	Exposed group (n = 92)	Non-exposed group (n = 92)	χ2 value	P value
Periprosthetic infection	6 (6.5)	1 (1.1)
Venous thromboembolism	8 (8.7)	2 (2.2)
Prosthetic loosening	5 (5.4)	1 (1.1)
Prosthetic dislocation	2 (2.2)	1 (1.1)
Heterotopic ossification	3 (3.3)	2 (2.2)
Chronic pain	7 (7.6)	2 (2.2)
Poor wound healing	2 (2.2)	1 (1.1)
Delayed wound healing	2 (2.2)	3 (3.3)
Joint stiffness	2 (2.2)	1 (1.1)
Total complication rate	27 (29.3)	13 (14.1)	6.582	0.01

Multivariate analysis of the impact of preoperative anxiety and depression on postoperative outcomes

This multivariate analysis was based on the total sample size after propensity score matching (184 cases, 92 per group). With poor joint function at 2 years postoperatively (Harris score < 80 points for THA patients or HSS score < 80 points for TKA patients) as the dependent variable, and preoperative anxiety and depression, age, gender, BMI, surgery type, hypertension, and diabetes as independent variables, multivariate logistic regression analysis was performed. A total of 48 patients had poor joint function at 2 years postoperatively (32 in the exposed group, 16 in the non-exposed group), with an incidence of 26.1% (48/184). Continuous variables age and BMI were dichotomized based on clinically relevant thresholds: Age was divided at 65 years (consistent with the demographic definition of elderly), and BMI at 28 kg/m² (Chinese population obesity diagnostic standard). Results showed that after adjusting for confounding factors, the presence of preoperative anxiety and depression were independent risk factor for poor joint function at 2 years postoperatively [odds ratio (OR) = 3.84, 95% confidence interval (CI): 1.92-7.68, P < 0.001]. The Hosmer-Lemeshow goodness-of-fit test for the model showed χ² = 6.342, P = 0.609, indicating good model fit.

With complication occurrence at 2 years postoperatively as the dependent variable, multivariate logistic regression analysis was performed. A total of 40 patients experienced complications at 2 years postoperatively (27 in the exposed group, 13 in the non-exposed group), with an incidence of 21.7% (40/184). Results showed that the presence of preoperative anxiety and depression were independent risk factor for complication occurrence at 2 years postoperatively (OR = 2.67, 95%CI: 1.28-5.57, P = 0.009). The impact of diabetes on complications approached statistical significance (OR = 2.04, 95%CI: 0.99-4.19, P = 0.053), suggesting that diabetes may also be a potential risk factor for complications, but requires larger sample sizes for further verification. The Hosmer-Lemeshow goodness-of-fit test for the model showed χ² = 7.865, P = 0.447, indicating good model fit (Table 6).

Table 6 Multivariate logistic regression analysis of the impact of preoperative anxiety and depression on 2-year postoperative outcomes¹.

Dependent variable	Independent variable	β	SE	Wald χ2	P value	OR	95%CI
Poor joint function at 2 years postoperatively	Preoperative anxiety and depression (yes vs no)	1.346	0.353	14.532	< 0.001	3.84	1.92-7.68
(n = 48/184, 26.1%)	Age (≥ 65 years vs < 65 years)2	0.682	0.328	4.325	0.038	1.98	1.04-3.77
	Gender (female vs male)	0.245	0.315	0.605	0.437	1.28	0.69-2.37
	BMI (≥ 28 kg/m2 vs < 28 kg/m2)3	0.534	0.336	2.526	0.112	1.71	0.88-3.31
	Surgery type (TKA vs THA)	0.412	0.308	1.786	0.181	1.51	0.83-2.76
	Hypertension (yes vs no)	0.328	0.321	1.043	0.307	1.39	0.74-2.61
	Diabetes (yes vs no)	0.587	0.342	2.947	0.086	1.8	0.92-3.52
Complication occurrence at 2 years postoperatively	Preoperative anxiety and depression (yes vs no)	0.982	0.374	6.895	0.009	2.67	1.28-5.57
(n = 40/184, 21.7%)	Age (≥ 65 years vs < 65 years)2	0.738	0.352	4.397	0.036	2.09	1.05-4.16
	Gender (female vs male)	0.186	0.342	0.296	0.586	1.2	0.61-2.37
	BMI (≥ 28 kg/m2 vs < 28 kg/m2)3	0.624	0.361	2.989	0.084	1.87	0.92-3.79
	Surgery type (TKA vs THA)	0.298	0.335	0.791	0.374	1.35	0.70-2.60
	Hypertension (yes vs no)	0.453	0.347	1.704	0.192	1.57	0.80-3.10
	Diabetes (yes vs no)	0.712	0.368	3.743	0.053	2.04	0.99-4.19

¹Numbers in parentheses represent the number of outcome events/total number of cases (incidence %).

²Age classification threshold of 65 years based on the demographic definition of the elderly.

³Body mass index classification threshold of 28 kg/m² based on Chinese adult obesity diagnostic standard (WS/T 428-2013); Hosmer-Lemeshow goodness-of-fit test P-values for both models are > 0.05, indicating good model fit.

OR: Odds ratio; CI: Confidence interval; THA: Total hip arthroplasty; TKA: Total knee arthroplasty; BMI: Body mass index.

DISCUSSION

This study found that patients with preoperative anxiety and depression had significantly lower joint function scores at 1 and 2 years postoperatively compared to patients with normal psychological status, and preoperative anxiety and depression were an independent risk factor for poor joint function at 2 years postoperatively (OR = 3.84). This result is generally consistent with previous research findings[8,9]. Some studies have reported that preoperative depression symptoms can increase the risk of poor functional recovery by 2-4 times[10], which is consistent with the results of this study. However, most previous studies have only focused on short-term outcomes at 3-6 months postoperatively[11,12], whereas this study extended the follow-up period to 2 years, revealing the persistent impact of preoperative psychological status on medium-to-long-term prognosis, a finding of significant clinical value.

The mechanisms by which preoperative anxiety and depression affect postoperative joint function recovery may be multifaceted. First, patients with anxiety and depression often exhibit pain-catastrophizing thinking and lower pain thresholds[13], which may lead to a stronger subjective perception of postoperative pain, thereby affecting enthusiasm for early functional exercise. Second, anxiety and depression can affect tissue healing and neuromuscular function recovery through hypothalamic-pituitary-adrenal axis dysfunction and elevated inflammatory factor levels[14,15]. Additionally, patients with anxiety and depression generally have poor compliance with rehabilitation training[16], which directly affects the progression of postoperative joint function recovery. The repeated measures ANOVA in this study showed significant interaction effects, suggesting differences in the functional recovery trajectories of the two groups of patients, with the exposed group’s rate of improvement significantly slower than the non-exposed group, further supporting the above mechanistic hypotheses.

To offer a holistic perspective, the effect of preoperative psychological distress can be explained by a “physiological and behavioral load” model. The impact of preoperative anxiety and depression can be described in terms of a dual “load” that combines with the surgical stress response. The physiological load stems from neuroendocrine-immune dysregulation: Prolonged psychological distress causes dysfunction of the hypothalamic-pituitary-adrenal axis, which leads to elevated cortisol and pro-inflammatory cytokines (interleukin-6, tumor necrosis factor-α) that establish an inflammatory environment that disrupts tissue repair and immune surveillance. This physiological load accounts for the increased infection rates in the exposed group (6.5% vs 1.1%), because lowered immune efficiency decreases the body’s ability to fight infection. At the same time, the behavioral load operates through dysfunctional pain coping (catastrophising), fear-avoidant behaviours (avoidance) and non-compliance with physiotherapy. These behaviours result in reduced activity, subtherapeutic adherence to physiotherapy, and a reluctance to load the affected joint, all of which affect recovery. The interaction of these loads with the trauma of surgery has a synergistic effect: Physiological dysregulation contributes to pain amplification (central sensitization) and behavioral changes restrict functional performance. This integrated perspective explains why the exposed group exhibited not only delayed functional recovery (as measured by Harris and HSS scores) but also ongoing pain and reduced improvement in quality of life (all SF-36 sub-scales) as physiological and behavioral factors interact, favouring pain and limited improvements in quality of life.

Postoperative pain is an important aspect of patient satisfaction following joint arthroplasty. Our study demonstrated that patients with preoperative anxiety and depression experienced significantly greater pain at rest and during activity at 1 and 2 years after surgery than the control group; and their pain relief was significantly less. This is in line with other studies that show preoperative psychological distress is a key predictor of postoperative chronic pain[17,18]. In our study, we found 7.6% of patients in the exposed group had chronic pain at 2 years after surgery, compared to 2.2% in the non-exposed group, and this highlights the importance of preoperative psychological status evaluation to help identify patients at-risk of chronic pain.

Central sensitization may be involved in the link between preoperative anxiety and depression and chronic pain. Long-term anxiety and depression has been found to alter the way the central nervous system processes pain, thereby decreasing the ability to modulate pain and increasing the risk of developing chronic pain[19]. Moreover, anxiety and depression patients often suffer from sleep disturbance[20], which increases pain sensitivity and slows down healing. In this study, the researchers compared the VAS scores of pain at rest and during activity in the exposed and non-exposed groups, and found that the pain scores of the exposed groups were significantly higher than those in the non-exposed groups, indicating that the influence of preoperative anxiety and depression on pain is comprehensive and sustained.

Quality of life is a crucial measure of the overall benefit of joint arthroplasty. In this study, the SF-36 scale was used to compare the patients’ quality of life 2 years after surgery, and it was found that the exposed group had significantly lower scores in each of the eight dimensions than the non-exposed group, and the improvement in each dimension was also significantly less. This conclusion is in line with other studies[21,22]. It is particularly interesting to note that the magnitude of improvement in the mental health and role emotional dimensions of the exposed group was only around 60% of the non-exposed group, which indicates that psychological issues prior to surgery not only have an impact on physical function recovery but also a lasting effect on mental health.

The quality of life is a multidimensional concept, such as physiological, psychological and social aspects. Because of low joint function and pain, patients with preoperative anxiety and depression have reduced daily living activities, which in turn, limit social participation and social communication[23]. In turn, unsatisfactory outcomes of surgery may also aggravate patient psychological distress, creating a vicious circle. A few studies have shown that psychological intervention can effectively improve the quality of life of patients after joint arthroplasty[24], which has certain implications for clinical management. The findings of this study indicate that for these patients with preoperative anxiety and depression, it is challenging to get satisfactory results with surgical intervention alone, and that the perioperative period requires a more holistic bio-psycho-social medical model treatment.

In this study, the total complications at 2 years after surgery in the study group (anxiety and depression before surgery) were significantly higher than in the control group (29.3% vs 14.1%, P = 0.006), and multivariate analysis demonstrated that anxiety and depression before surgery were independent risk factors for complications (OR = 2.67). This result is in line with some reports[25], but is not consistent with others[26]. This may be due to the different study designs, sample sizes, follow-up periods, and definition of complications. This study had a longer follow-up period (2 years) and a larger sample size (184 cases after matching) and adjusted for confounding factors using propensity score matching, so the results were more credible.

The mechanisms by which preoperative anxiety and depression increase postoperative complication risk may include multiple aspects. First, anxiety and depression can affect immune function through neuroendocrine pathways, reducing the body’s resistance to infection[27]. In this study, the periprosthetic infection rate in the exposed group (6.5%) was significantly higher than in the non-exposed group (1.1%), supporting this hypothesis. Second, poor compliance with rehabilitation training in patients with anxiety and depression may increase the risk of complications such as venous thromboembolism and prosthetic loosening[28]. Additionally, anxiety and depression may affect the wound healing process, and in this study, the incidence of wound healing-related complications in the exposed group also showed an increasing trend. Notably, this study found that the impact of diabetes on complications approached statistical significance (P = 0.053), suggesting that diabetic patients may be another high-risk group for postoperative complications, which is consistent with literature reports[29].

This study has the following innovations and advantages: First, the use of propensity score matching methods effectively balanced baseline characteristics between the two groups of patients, reducing the impact of selection bias and confounding factors, and improving the reliability of study results. After matching, the standardized differences of all covariates were < 0.1, achieving good balance. Second, this study extended the follow-up time to 2 years, addressing the shortcoming of previous studies that mostly focused on short-term outcomes, revealing the persistent impact of preoperative psychological status on medium-to-long-term prognosis. Third, this research used multidimensional and standardized measurement instruments (HADS, Harris score, HSS score, VAS, SF-36) which evaluates in a relatively comprehensive way the effects of preoperative anxiety and depression on postoperative joint functioning, pain, quality of life, and complications, giving an otherwise complete evidence chain. Fourth, the independent impact of preoperative anxiety and depression on postoperative adverse outcomes was elucidated using multivariate logistic regression analysis, which set clearer goals of clinical intervention.

This study has the following limitations: First, as a single-center retrospective study, there may be information bias and selection bias, and the generalizability of results should be approached with caution. Second, although the 76.9% follow-up completion rate is within an acceptable range, lost-to-follow-up patients may affect the accuracy of results, although there were no significant differences in baseline characteristics between lost-to-follow-up patients and patients who completed follow-up. Third, this study did not include patients who received psychological treatment or medication after surgery, which helps control for confounding factors but also limits the reflection of study results on actual clinical situations. Fourth, this study only assessed psychological status once preoperatively and did not dynamically monitor postoperative psychological changes, making it impossible to determine whether improvement in psychological status can improve postoperative outcomes. Fifth, the sample size is relatively limited, and some subgroup analyses (such as risk factors for different types of complications) may have insufficient statistical power. Sixth, as an observational study, this study can only reveal associations rather than causal relationships, and cannot completely exclude the influence of unmeasured confounding factors.

It must be noted that our one preoperative timepoint measure is a snapshot of psychological status and is not a dynamic curve of anxiety and depression. HADS scores preoperative could be inappropriate to adequately capture chronic psychological burden, and could not differentiate between state anxiety (situational, which pertains to surgical stress) and trait anxiety (chronic, dispositional). Patients whose postoperative psychological distress persists may contribute a disproportionate number of adverse outcomes, yet our study design is unable to identify these individuals or whether postoperative change in psychological status is associated with functional recovery. Although our single-timepoint measure is a convenient, clinically viable screening method that can be applied in routine practice, future studies that build on this measurement, including repeated measures of psychological variables at various timepoints (preoperative, immediate postoperative, 3-month, 6-month, 1-year and 2-year), would be informative regarding the patterns of psychological responses and their relative influence on outcomes. These longitudinal designs would be able to differentiate between patients who experience chronically high levels of distress, distress improvement, or a novel postoperative distress, allowing more tailored interventions and possibly defining critical periods of psychological support.

The clinical implications of the findings of this study are significant. The first is to include psychological status assessment as part of the daily routine preoperative evaluation process to joint arthroplasty to detect the high-risk groups of psychological issues at the early stages. The HADS scale is straightforward and user-friendly, and can be promoted in clinics. Second, multidisciplinary collaborative comprehensive intervention strategies are to be applied to patients with preoperative anxiety and depression, in the perioperative period, namely the preoperative psychological counseling, cognitive behavioral therapy, medication where needed, and postoperative for each patient of individual pain management instructions and rehabilitation instructions[30]. Third, clinicians must completely educate patients that preoperative mental condition could influence postoperative results, modify their anticipations of surgical results, and prevent disappointments of overestimating anticipations. Fourth, develop long-term postoperative follow-up systems to offer more intensive care and support to patients with high risks of psychological issues and diagnose and treat postoperative psychological issues and complications in time. The results can be used as a strong evidence base by clinicians and healthcare administrators to support the resources required to create the real multidisciplinary patient-centered care paths. The 2.08 fold more complicated rate in the exposed group would imply significant increment in healthcare expenses and this would imply that investment in integrated psychological services could be cost effective.

Several areas can be investigated in the future: First, use prospective randomized controlled trials to study the improvement effect of perioperative psychological interventions on joint arthroplasty patients’ postoperative outcomes, and provide high-level evidence for clinical practice. These trials should examine the effect of specific interventions (cognitive-behavioral therapy, mindfulness-based stress reduction, pharmacotherapy) that reduce the anxiety/depression scores, on long-term function and complications. Second, use longitudinal study designs to track dynamic changes of psychological status of patients at different time points before and after surgery, and investigate the time course and dose-response relationship between the improvement in psychological status and improvement in postoperative outcomes.

CONCLUSION

In conclusion, this study, using propensity score matching and multivariate analysis techniques, validated that anxiety and depression prior to surgery are independent risk factors for poor joint function and complications at 2 years after surgery in total hip and knee arthroplasty, and continue to have a negative effect on pain reduction and quality of life after surgery. This study highlights the need for psychological evaluation and support in the perioperative period and offers scientific evidence to enhance the prognosis of joint arhtroplasty patients. In the future, attention should be paid to patients’ psychological status and a variety of interventions should be used to improve surgical outcomes and patient satisfaction.