Anxiety and depression as potential risk factors for limited pain management in patients with elderly knee osteoarthritis: a cross-lagged study

Abstract

Background

Pain management for knee osteoarthritis (KOA) patients is challenging. Pain arises from both physiological and psychological interactions, with anxiety and depression potentially contributing as risk factors that hinder effective pain management in KOA patients.

Methods

Before treatment(T1), A total of 206 elderly inpatients with KOA were enrolled based on initial screening criteria. After treatment (T2), patients were selected based on inclusion and exclusion criteria, and completed follow-up through phone or online questionnaires. The interval between T1 and T2 was three months. Outcome measures included the Visual Analogue Scale (VAS) for pain intensity, Beck Anxiety Inventory (BAI) for anxiety, and Geriatric Depression Scale (GDS) for depression. Descriptive and bivariate analyses were used to evaluate the pain, anxiety and depression of the participants. A cross-lagged model was used to examine the temporal and causal associations among pain, anxiety, and depression.

Results

91% of elderly patients with KOA experienced at least mild depression. Furthermore, 31% of patients reported mild or higher levels of anxiety. At the same time, pain, depression, and anxiety were significantly correlated and mutually predictive(all p < 0.01). Across the different time points, Depression and anxiety at T1 positively predicted pain at T2,with correlation coefficients of 0.19 (p < 0.05) and 0.07 (p < 0.05), respectively.

Conclusions

Anxiety and depression may be potential risk factors limiting the effectiveness of pain management in KOA patients. Clinical treatment should regularly evaluate anxiety and depression levels and integration of psychological interventions or appropriate antianxiety and antidepressant medications.

Clinical trial number

Not applicable, for the investigative research nature of the study.

Introduction

Knee Osteoarthritis (KOA) is a degenerative joint disease involving all joint tissues, characterized by the progressive erosion of articular cartilage, subchondral bone remodeling, osteophyte formation, synovial disease and ligament relaxation or contracture [1]. The meniscus, periarticular muscles, sub-patellar fat pads and joint capsule may also have a role in the development of KOA [1, 2]. This condition causes joint pain, stiffness, and reduced function, often limiting daily activities and significantly impairing quality of life. KOA is a highly prevalent chronic pain condition, accounting for 85% of osteoarthritis cases globally [3], with a symptomatic incidence of approximately 8.1% [4], leading to a high rate of disability. With the aging population in China, the prevalence of KOA, an age-related disease [5], is also increasing [6]. Studies indicate that the incidence of KOA symptoms among Chinese individuals over 60 years of age is 19.4% [7]. This condition significantly impacts patients’ quality of life and mental health, posing a substantial economic burden on society.

Pain is the hallmark symptom of KOA [4], characterized by intense, intermittent pain on a background of persistent aching pain. Osteoarthritis pain includes nociceptive, neuropathic, and nociplastic components. Nociceptive pain indicates ongoing joint inflammation and surrounding tissue damage, while neuropathic pain indicates a degree of nerve damage [8]. Despite advancements in biological and pharmacological therapies that have refined KOA treatment strategies [9], many patients still experience limited pain relief, and the mechanisms underlying this persistence remain unclear [10]. Pain management is crucial for alleviating clinical symptoms, improving clinical prognosis, and enhancing quality of life. It remains the primary goal of KOA treatment [11, 12]. Therefore, it is extremely important to explore the potential risk factors for the limited effect of KOA pain management, and to improve the treatment strategy.

It is a common clinical phenomenon that physical diseases are often accompanied by mental symptoms. Pain is characterized as an unpleasant sensory and emotional experience that is linked to actual or potential tissue damage [13]. Shaaron Leverment [14] found that patients’pain and negative emotions affect sleep quality and contribute to a poor disease prognosis. In rheumatoid arthritis and axial spondylarthritis, worsening inflammation is closely related to sleep and emotional disorders. Limited management of these factors has led to a decline in patients’ quality of life [15, 16]. As mentioned above, many studies have explored the relationship between negative emotions and somatic symptoms in chronic pain diseases. And research indicates that osteoarthritis is associated with depression and anxiety [17, 18]. However, there are still few studies focus on the population of elder KOA patients, and the research methods are mostly parallel design, which cannot infer the relationship from time and causality. In other words, no study to clarify whether the depressive symptoms in KOA patients before treatment will lead to limited effect of standardized pain management treatment. We hypothesize that anxiety and depression may be potential risk factors for limited pain management in KOA patients, and they will have cross-causality at different time points. This hypothesis requires further validation.

To address this issue, a cross-lagged regression model was chosen to analyze the temporal and causal relationships among pain, anxiety, and depression. This model examines bidirectional influences between variables over time, providing insights into the dynamic interplay of psychological and physiological factors in KOA. Furthermore, it controls for prior levels of the dependent variables, reducing confounding risk and enhancing causal inferences [19, 20]. This approach is particularly useful in understanding how psychological states, such as anxiety and depression, may predict future pain experiences and vice versa, adding depth to the investigation beyond parallel analyses.

Participants and methods

Participants

The study recruited elderly (age ≥ 55 years) KOA patients hospitalized in the Department of Orthopedics at the Affiliated Traditional Chinese Medicine Hospital of Southwest Medical University in Luzhou, China, from June 2023 to February 2024. A total of 206 patients who met the initial screening criteria completed pre-treatment questionnaires. Follow-up surveys were conducted three months after discharge for patients meeting the inclusion and exclusion criteria, resulting in 106 patients completing both surveys. All participants provided informed consent online.

Initial screening criteria and inclusion criteria

Initial Screening Criteria: ①Age ≥ 55 years; ②Diagnosed of KOA, Kellgren-Lawrence grade П or III; ③BMI: 18.5–28; ④Cognitively capable of completing or responding to the questionnaire independently; ⑤Willingness to participate in the study and sign the informed consent form. Inclusion Criteria; ①On the basis of Initial Screening Criteria and completed the T1 questionnaire;②Underwent 1–4 weeks of standardized treatments (including health education, nonsteroidal anti-inflammatory drugs, physiotherapy and symptomatic slow-acting drugs for osteoarthritis).

Exclusion criteria

①Presence of severe internal diseases or other significant trauma;②Unwillingness to continue participation or loss to follow-up;③Hospitalization duration less than 7 days or more than 30 days;④Surgery and injection.

Investigation data

Visual analogue scale (VAS)

The Visual Analogue Scale (VAS) was employed in this study to assess pain intensity. Patients were asked to mark their pain level on a horizontal line ranging from 0 (no pain) to 10 (worst pain imaginable). This method is suitable for quickly and accurately evaluating patients’ subjective pain experiences and is widely used in clinical settings because of its high reliability and validity. Additionally, as a sleep-related scale, the VAS can also provide some indication of the patients’ sleep quality alongside their pain levels.

Beck anxiety inventory (BAI)

The Beck Anxiety Inventory (BAI), developed by American psychologist Beck and colleagues in 1988 [21], assesses the severity of various anxiety symptoms. The BAI consists of 21 self-report items, each describing anxiety symptoms. The BAI range from 0 to 63, higher scores indicate more anxiety; The respondents rated how much they have been bothered by each symptom over the past week on a 4-point scale: 1 (not at all), 2 (mildly, but it did not bother me much), 3 (moderately, it wasn’t pleasant at times), and 4 (severely, it bothered me a lot).

The BAI has high internal consistency reliability, with Cronbach’s alpha coefficients ranging from 0.88 to 0.92, and test-retest reliability over one week range from 0.71 to 0.75. The BAI scores have significant positive correlations with various anxiety self-assessment scales, with correlation coefficients between 0.56 and 0.83, demonstrating good convergent validity. The BAI’s correlation coefficients with the Beck Depression Inventory (BDI) range from 0.54 to 0.63, which is significantly lower than the correlation between the State-Trait Anxiety Inventory (STAI) and the BDI, indicating good discriminant validity [22, 23]. The BAI is suitable for adults aged 17 and older. The BAI’s concise item content and straightforward administration and analysis make it a highly practical self-reported measure of anxiety, that is widely used today.

Geriatric depression scale (GDS)

The Geriatric Depression Scale (GDS), developed by Brink and colleagues in 1982 [24], is specifically designed for the elderly individuals to screen for depression, and addresses the unique somatic symptoms often misdiagnosed as depression in older adults. The GDS comprises 30 items representing core symptoms of depression in the elderly individuals, including low mood, reduced activity, irritability, withdrawal, distressing thoughts, and negative evaluations of the past, present, and future. Each item is a question with a “yes” or “no” response, indicating the participant’s feelings over the past week. Ten items are reverse-scored (where “no” indicates depression), and twenty items are scored positively (where “yes” indicates depression). The GDS range from 0 to 30, higher scores reflect more severe depression.

Studies by Brink et al.(1982) [24], and Yesavage et al. (1983) [25] have shown that the GDS has good test-retest reliability and internal consistency. Its correlation coefficients with other commonly used depression scales, such as the SDS, HRSD, and BDI, range from 0.73 to 0.84, indicating good validity. Clinical assessments have revealed that the GDS has a higher agreement rate for diagnosing depression in the elderly individuals than do the BDI and SDS. The GDS is appropriate for individuals aged 56 and older and is specifically standardized for the elderly individuals, making it superior for this demographic. Sheikh et al. (1986) [26] reviewed six different depression scales (GDS, HRSD, teh Geriatric Psychophysiological Complaint Questionnaire, the SDS, the BDI, and the CES-D) and reported that the GDS assessed six out of thirteen critical symptoms unique to depression in elderly individuals, which was greater than any other scale.

Data analysis

Descriptive statistics and Pearson correlation analyses were conducted using IBM SPSS Statistics version 26.0 (IBM Corp., Armonk, NY, USA) to investigate the relationships between pain, anxiety, and depression. Mplus version 8.3 (Base Program and Combination Add-On, 64-bit; Muthén & Muthén, Los Angeles, CA, USA) was employed to construct a cross-lagged regression model, assessing the mutual influence and quasi-causal relationships among the variables.

Model fit indices included: χ²/df, Tucker-Lewis Index (TLI), Comparative Fit Index (CFI), and Root Mean Square Error of Approximation (RMSEA).

The cross-lagged model was used to assess the stability and mutual influence of the variables. If the path coefficient from a variable at T1 to another variable at T2 was significant, it indicated a causal relationship between the variables.

Results

Demographic information

Between June 2023 and February 2024, a total of 206 participants meeting the initial screening criteria completed the T1 questionnaire. In accordance with the inclusion and exclusion criteria, 120 participants were selected for the T2 follow-up questionnaire. Of these, 108 participants completed both questionnaires, resulting in a 10% loss to follow-up. Data from 2 participants were excluded due to logical inconsistencies, leaving a final sample of 106 participants for analysis. The basic demographic characteristics are shown in Table 1, and the general conditions of pain, depression, and anxiety among the participants are presented in Table 2.

Table 1.

Demographic characteristics of participants

Variables	Groups	Number	Percentage(rounded)
Age	55–65	66	62%
	65–75	28	26%
	>75	12	11%
Gender	Female	78	74%
	Male	28	26%
Pain	Mild(0–3)	14	13%
	Moderate(4–6)	72	68%
	Severe(7–9)	20	19%
Depression	None(0–9)	10	9%
	Mild (10–19)	80	76%
	Moderate or Severe(20–30)	16	15%
Anxiety	None(0–7)	72	68%
	Mild (8–15)	24	23%
	Moderate(16–25)	8	8%
	Severe(26–63)	2	2%

Table 2.

General conditions of pain, depression, and anxiety among participants

Symptom	Before / After treatment	Mean	SD	P value (diffrient groups)
Pain
	Before	4.21	1.57	P<0.001
	After	2.64	1.62
Depression
	Before	15.32	4.67	P<0.001
	After	10.13	4.46
Anxiety
	Before	7.74	6.90	P<0.001
	After	4.81	4.91

The results indicated that 91% of elderly patients with KOA experienced at least mild depression, whereas 15% experienced moderate to severe depression. Additionally, 31% of patients reported mild or high levels of anxiety, with 10% experiencing moderate to severe anxiety(Table 2). These findings are in line with those of previous studies, highlighting the prevalence of comorbid pain, depression and anxiety among elderly KOA patients [27].

Correlation analysis

For this analysis, Pearson correlation analysis was used to examine the relationships between pain (Pain1, P1 and Pain2, P2), depression (Depression1, D1 and Depression2, D2), and anxiety (Anxiety1, A1 and Anxiety2, A2) at two different time points, T1 and T2. Age and gender were included in the assessment to account for their potential impact on self-reported scores and to ensure a comprehensive analysis. Age was significantly correlated with the three variables at T1 but not at T2, And Gender, in most cases, did not significantly affect the correlations.Both of them contradicted the assumptions of the cross-lagged design, leading to its exclusion from subsequent analyses.

The correlation coefficient between P1 and P2 was 0.022 (p < 0.05), that between D1 and D2 was 0.333 (p < 0.01), and that between A1 and A2 was 0.612 (p < 0.01). These statistically significant coefficients indicate a certain level of stability for pain, depression, and anxiety between the two time points. At T1, the correlation coefficient was 0.496 (p < 0.01) between P1 and D1, 0.582 (p < 0.01) between P1 and A1, and 0.403 (p < 0.01) between D1 and A1. At T2, the correlation coefficient was 0.351 (p < 0.01) between P2 and D2, 0.493 (p < 0.01) between P2 and A2, and 0.395 (p < 0.01) between D2 and A2. All the correlations were significantly positive, supporting the assumptions of correlation and synchronicity, and thus aligning with the cross-lagged design hypothesis. (Table 3; Fig. 1).

Table 3.

Correlation analysis of pain, depression, and anxiety before and after treatment

	sex	age	P1	D1	A1	P2	D2
age	-0.270**
P1	-0.162	0.699**
D1	-0.041	0.507**	0.496**
A1	-0.146	0.428**	0.582**	0.403**
P2	0.213*	-0.044	0.022*	0.201*	0.071*
D2	0.175	0.17	0.165	0.333**	0.067	0.351**
A2	-0.05	0.362**	0.480**	0.343*	0.612**	0.493**	0.395**

*Note * p < 0.05; ** p < 0.01; *** p < 0.001

Fig. 1.

Correlation heatmap

Cross-lagged analysis

On the basis of the results of the correlation analysis, we constructed a cross-lagged model to further explore the predictive relationships between pain, depression, and anxiety. We used Mplus 8.3 to fit the model, incorporating baseline variables such as gender, age, marital status, education level, and type of health insurance as control variables. The model fit indices indicated a good fit: χ²/df = 3.82, RMSEA = 0.038, CFI = 0.931, TLI = 0.963, SRMR = 0.033.

The cross-lagged analysis revealed that pain, depression, and anxiety positively predict each other at the same time point. At different time points, the symptoms of pain, anxiety, and depression significantly improved after systematic treatment. The correlation coefficient from D1 to P2 was 0.19 (p < 0.05), and that from A1 to P2 was 0.07 (p < 0.05), indicating that D1 and A1 had a positive predictive effect on P2. However, the effects from P1 to D2, P1 to A2, D1 to A2, and A1 to D2 were not significant (Figure 2).

Fig. 2.

Cross-lagged model. *Note * p < 0.05; ** p < 0.01; *** p < 0.001. The solid lines indicate statistically significant paths, whileas the dashed lines indicate statistically non-significant paths

Typical cases

As shown in Fig. 3(a) and Fig. 3(b), the patients’ baseline information, including age, gender, ethnicity, BMI, and KOA progression, was generally comparable, with no other relevant medical history. At T1, both participants had a pain VAS score of 4. After two weeks of identical conservative KOA treatment, participant A’s VAS score decreased to 2, while participant B’s pain score remained at 4, indicating limited pain improvement for participant B. Additionally, participant B had higher BAI and GDS scores at both T1 and T2, suggesting higher levels of anxiety and depression.

Fig. 3.

(a). Participant A’s knee X-ray. (b).Participant B’s knee X-ray. Note: (a). Participant A, with left KOA, experiencing pain and limited mobility for 10 years; (b). Participant B, with left KOA, experiencing pain and limited mobility for 9 years

Discussion

Our findings indicate that standardized pain management generally improved both pain levels and symptoms of depression and anxiety in patients. This reinforces previous research showing that effective pain management is essential for alleviating psychological symptoms in individuals with depression and anxiety [28, 29]. Unlike previous studies, our cross-lagged analysis confirmed the positive predictive effects of D1 and A1 on P2 among elderly KOA patients. We observed that patients who did not experience significant pain improvement often had higher scores for depression and anxiety before treatment.

Pain management in KOA extends beyond reducing pain severity, it also aims to enhance patients’ overall quality of life. Effective pain management encompasses physical function, psychological well-being, and the ability to engage in daily activities without limitations. Although KOA is recognized as a typical condition associated with pain-depression comorbidity [30], routine screening or evaluation of depressive and anxiety symptoms has not been widely incorporated into risk factors research and clinical practice [31].We recommend routine psychiatric symptom scores on admission for KOA patients. When treating patients with severe depression or anxiety, additional psychological therapy or, if necessary, anti-anxiety or antidepressant medications should be considered. We also suggest that clinicians adopt a multidisciplinary approach that integrating both biological and psychological perspectives. Psychological interventions, such as cognitive-behavioral therapy (CBT), should be integrated to assist patients in managing negative emotions, enhancing coping skills, and improving self-efficacy. This may create a synergistic effect between psychological and physical symptoms, ultimately leading to improved patient outcomes [32, 33].

Notably, anxiety and depression significantly influence pain management outcomes in KOA patients. The relevance of these psychological factors may be heightened within the cultural and healthcare context of China. The cultural tendencies of Chinese people have been shown to be associated with particularly high levels of psychiatric disorders stigmatization [34],and few patients with mental disorders receive regular psychiatric treatment [35]. Additionally, patients’ reluctance to receive psychiatric treatment may be related to higher drug costs. These factors have hindered the development of clinical psychiatric treatment for both psychiatric and pain-related conditions, resulting in limited pain management effectiveness.

This study utilized a unique cross-lagged design involving two time-point assessments to provide a more robust analysis of causal relationships over time. This approach helps mitigate random and systematic biases often associated with cross-sectional data [36, 37]. Additionally, it allows the assessment of both bidirectional and unidirectional effects between variables, shedding light on their intricate interactions. Through strict adherence to the inclusion and exclusion criteria and the use of cross-lagged analysis, we were able to isolate within-individual and between-individual effects, thereby enhancing the precision of our results [38, 39].

Several limitations remain in this study. First, the particular design of our study resulted in certain limitations regarding the population. We constrained baseline conditions, including age, disease stage, BMI, treatment duration, treatment methods, and treatment location. While this approach reflects specific circumstances, it may not accurately represent the entire population of patients with KOA. Moreover, while we have demonstrated the positive predictive effects of D1 and A1 on P2, indicating that anxiety and depression levels impact patients’ perception of pain, it is important to note that no interventional (prospective) studies have been conducted to validate these findings. Finally, due to the unique cross-lagged design, we did not analyze the impact of social factors (e.g., marital status, income, education) on pain, anxiety, and depression levels from a cross-sectional perspective. This approach is inconsistent with our original design as it does not reflect temporal causality between variables. We are considering an improved cross-lagged analysis model to address this limitation.

Author contributions

All the authors contributed to the study conception and design. Material preparation were performed by [Wenhao Yang][Dujiang Yang], and data collection and analysis were performed by [Guangyuan Ma], [Jingchi Li], and [Dingchang He]. The first draft of the manuscript was written by [Wenhao Yang] and [Taiyuan Guan] , [Hui Shi] and [Guoyou Wang] give guidance and revision for the process of research and writing. All the authors read and approved the final manuscript.

Funding

The research was supported by two fellowship from the School-level Projcct of Southwest Medical University, {00210274} and {SKLYB13}.

Data availability

The data that support the findings of this study are available from the authors Email: wangguoyou1981@swmu.edu.cn.

Declarations

Ethical approval

Ethical approval was waived by the local Ethics Committee of southwest medical university in view of the investigative research nature of the study, all the procedures being performed did not interfere with conventional treatment and investigation is beneficial to participants.

Consent to participate

Informed consent was obtained online from all individual participants included in the study.

Competing interests

The authors declare no competing interests.