The moderating role of negative affect on pain in patients with rheumatoid arthritis who have undergone shoulder arthroplasty

Abstract

Objective

Patients with rheumatoid arthritis undergoing shoulder arthroplasty often experience significant pain and functional limitations. Improvements in physical outcomes are well-documented, but psychological factors like negative affect remains underexplored. This study aimed to examine whether changes in negative affect moderate the relationship between functional and pain in rheumatoid arthritis undergoing shoulder arthroplasty.

Methods

We analyzed 92 patients with rheumatoid arthritis undergoing shoulder arthroplasty. Baseline and 6-month postoperative outcomes were assessed, including pain, shoulder function (SPADI), and negative affect. Moderation analysis was conducted to examine whether changes in negative affect influenced the relationship between functional improvements and pain reduction.

Results

The majority of patients were female (80%) with a mean age of 65 ± 11.1 years. Postoperative assessments showed significant improvements in pain and function. Pain scores decreased from a mean of 5.6 (SD = 2.6) to 1.2 (SD = 1.7) (p <.001), while SPADI scores improved from 63.7 (SD = 18.7) to 24.7 (SD: 15.9), (p <.001). Negative affect showed a reduction from 1.4 (SD = 0.7) to 1.2 (SD: 0.4), (p =.067). Persistent pain was present in 16% of patients postoperatively. Moderation analysis revealed a significant interaction between improvements in function and changes in negative affect on pain (β = 0.16, p =.017), suggesting that patients with concurrent emotional improvement experienced enhanced pain relief.

Conclusion

Shoulder arthroplasty in patients with rheumatoid arthritis reduces pain and improves function. Notably, improvements in negative affect may amplify the benefit of functional recovery on pain outcomes. These findings highlight the importance of integrating psychological well-being into postoperative care for this population.

Introduction

While many rheumatoid arthritis (RA) patients experience substantial improvements in pain, function, and quality of life following surgery, a subset continues to report persistent postoperative pain [1]. RA often involves the glenohumeral joint, though the prevalence varies widely across studies [2]. Advances in treatment particularly the broader use of disease-modifying [3] antirheumatic drugs and biologics have contributed to a decline in the number of Shoulder Arthroplasty (SA) performed in this population. RA related factors contribute uniquely to these outcomes. Unlike primary osteoarthritis, RA is a systemic inflammatory disease that often affects soft tissues, bone quality, and healing capacity. Chronic synovitis, tendon involvement, and bone erosion can all influence the surgical outcome. Additionally, RA patients are more likely to have rotator cuff pathology, complicating implant selection and functional recovery [4].

Persistent pain after SA in RA patients may arise from residual inflammation or synovitis, poor bone integration or implant loosening particularly in patients on long-term corticosteroids suboptimal soft tissue balance, rotator cuff insufficiency, or central sensitization and pre-existing chronic pain syndromes [5, 6].

Management strategies should include careful preoperative planning, optimization of disease control, and appropriate implant selection, such as reverse total shoulder arthroplasty in cases with cuff deficiency. In fact, reverse shoulder arthroplasty is now commonly used to manage RA because of the elevated risk associated with traditional approaches [7]. Addressing patient expectations, as well as early identification of ongoing pain and prompt rehabilitation, is critical [8]. In some cases, persistent pain may also reflect non-mechanical sources, such as depression, anxiety, or systemic inflammation [9].

A more personalized approach to perioperative care may help reduce the risk of persistent pain and improve long-term surgical outcomes. Pain and physical function are closely interrelated outcomes following SA. Improvements in shoulder function are often accompanied by reductions in pain, as patients regain mobility and strength. However, the magnitude and trajectory of change in pain and function can vary widely. Improvements in shoulder function are often accompanied by reductions in pain, as patients regain mobility and strength. However, the magnitude and trajectory of change in pain and function can vary widely. Interestingly, past research suggests that psychological factors, particularly negative affect, may influence these outcomes, with higher levels of distress linked to poorer pain relief and functional recovery [10–14]. Postoperative recovery typically shows the most rapid improvements within the first few months, with sustained gains over time. This study aimed to examine whether changes in negative affect moderate the relationship between functional improvement and pain reduction in patients with RA undergoing SA. We hypothesized that reductions in negative affect would strengthen the association between functional improvement and pain reduction in patients with RA undergoing shoulder arthroplasty.

Materials and methods

Study design

Since January 2006, patients with RA undergoing anatomic or reverse SA at a Swiss orthopedic clinic have been prospectively documented in a local registry. This retrospective cohort study included those with completed 6-month follow-up who were recorded in the Shoulder Arthroplasty Registry (SAR). A detailed summary of data collection procedures has been published previously [15]. Functional evaluations and patient-reported outcome measures are systematically collected using validated instruments and questionnaires. Data collection and management are facilitated through the utilization of the Research Electronic Data Capture system [16].

Participants

This retrospective cohort study included RA patients who underwent either anatomic or reverse SA with completed 6-month follow-up examination and who have been documented in the local SAR. Cases involving revision were excluded, as were patients with duplicate registry entries (either for the same shoulder or the contralateral shoulder).

Self- report assessments

Persistent postsurgical pain

Persistent postsurgical pain was defined as persistent pain with a score of 3 or higher on the pain Numeric Rating Scale (NRS) (0 = no pain, 10 = worst pain) [17]. Pain was assessed at baseline and at the 6 month follow up using a self- report patient questionnaire.

Patient-related and sociodemographic information included patient age at the time of surgery, sex, body mass index, insurance type, baseline smoking status (yes, no), baseline alcohol consumption status (no, occasional, daily) and the American Society of Anesthesiologists (ASA) physical status classification system. In addition, patients’ baseline mental health status was included using the Anxiety and Depression dimension of the of the European Quality of Life 5 Dimensions 5 Level scale [18].

Disease-related covariates included dominance of the affected shoulder, bilateral shoulder arthroplasty (yes or no), surgery admission type (illness or accident), procedure type (anatomic vs. reverse total shoulder arthroplasty (TSA)), surgery duration in minutes and additional procedures performed during surgery such as Biceps Tenotomy (yes or no). We also assessed Shoulder Pain and Disability Index (SPADI) [19] which is a disability assessment in shoulder arthroplasty.

Clinical assessments

Additionally, we included four indexes of functional scores: active flexion and abduction, active external rotation with arms comfortably at the side and internal rotation (thumb reaching thigh, buttock, lumbosacral, L3, Th12 and Th7 vertebrae).

Data analysis

Descriptive statistics were used to summarize the study population. Continuous variables were reported as means and standard deviations, while categorical variables were presented as frequencies and percentages. All continuous outcome variables were assessed for normality using Shapiro-Wilk tests and Q-Q plots, confirming approximate normal distribution and supporting the use of parametric analyses. To assess changes over time, paired t-tests were conducted to compare baseline and 6-month follow-up scores for pain, function, and negative affect. All analyses were performed using R (version 2023.06.2), with a significance level set at p <.05.

To examine whether changes in negative affect moderated the relationship between improvements in shoulder function and reductions in pain, a moderation analysis was performed using change scores. The outcome variable was the change in pain between baseline and 6 months postoperatively. The predictor was change in shoulder function, and the moderator was change in negative affect. All continuous predictors were mean-centered before computing the interaction term. Analyses were conducted in RStudio using the interactions package.

Results

The majority of the 92 patients who were analyzed were female (80%) with an average age at surgery of 65 years ± 11.1 and an average body mass index of 25.1 ± 5.1. Patients received a shoulder arthroplasty on their dominant side in 56% of cases. Insurance type was equally distributed between private (52%) and general (48%) insurance. Most patients were not active smokers (92%) and 31% did not consume any alcohol, while 58% reported occasional alcohol consumption, 11% were daily drinkers. The rotator cuff was intact in 19%, while 31% suffered from massive tears. In total, 16% of RA patients developed persistent pain after SA at 6 months (see Table 1).

Table 1.

Patient demographics (N = 92)

	Missing	N	Mean (SD)
Age	0%	92	65.5 (11.1)
Sex (n,%)	0%
Female		74 (80)
Male		18 (20)
Body Mass Index (BMI)	41%	54	25.1 (5.4)
Arthroplasty on dominant side (n,%)	2%
No		40 (44)
Yes		50 (56)
Procedure Type (n,%)	0%
Reverse TSA		66 (72)
Anatomic TSA		26 (28)
Bilateral Arthroplasty (n,%)	0%
Unilateral		48 (52)
Bilateral		44 (48)
ASA Risk Category (n,%)	1%
II		31 (34)
III		59 (65)
IV		1 (1)
Admission type (n,%)	0%
Illness		91 (99)
Accident		1 (1)
Surgery Duration (min)	0%	92	98.5 (26.4)
Biceps Tenotomy (n,%)	0%
No		64 (70)
Yes		28 (30)
Insurance type (n,%)	0%
General		44 (48)
Private		48 (52)
Smoker (n,%)	7%
No		79 (92)
Yes		7 (8)
Alcohol Consumption (n,%)	2%
No		28 (31)
Occasionally		52 (58)
Daily		9 (10)
Multiple times a day		1 (1)
Tear Severity (Gerber) (n,%)	1%		0.6 (0.2)
No Rotator Cuff Tear		17 (19)	64.2 (19.2)
Partial tear		27 (30)	30.5 (13.6)
Single full tear		5 (5)	56.9 (15.7)
Two or three tendons (only one full)		14 (15)	31.7 (17.4)
Massive tear		28 (31)	5.8 (2.6)
Equation 5D5L utility index DE (value set)	64%	33
SPADI (0 = best, 100 = worst)	4%	88
CS Constant Murley Score (0 = min 100 = max)	7%	86
Quick DASH score (0 = best, 100 = worst)	2%	90
Subjective Shoulder Value	71%	27
Pain (NRS)	2%	90
Flexion	1%	91	76.3 (32.1)
Abduction	1%	91	65.2 (29.0)
External Rotation	1%	91	24.2 (18.6)
Internal Rotation / Apley’s Test (n,%)	2%
Lateral thigh		5 (6)
Gluteal region (buttocks)		29 (32)
Lumbosacral region		24 (27)
Lumbar vertebra 3 (L3)		20 (22)
Thoracic vertebra 12 (Th12)		12 (13)

Functional outcomes assessed by the SPADI improved significantly over time. The mean SPADI score at baseline was 63.7 (SD = 18.7), decreasing to 24.7 (SD = 15.9) at six months. This corresponds to a mean improvement of 39.1 points (SD = 21.9), t(78) = 15.82, p <.001, 95% CI [34.16, 43.00; Fig. 1]. Patients reported a significant reduction in pain intensity six months postoperatively. The mean preoperative pain score was 5.6 (SD = 2.6), which decreased to 1.2 (SD = 1.7) at follow-up. This represents a mean change of 4.4 points (SD = 3.0), t(82) = 13.36, p <.001, 95% CI [3.74, 5.05; Fig. 2]. Negative affect showed a modest decrease over time, with preoperative scores averaging 1.4 (SD = 0.7) and decreasing to 1.2 (SD = 0.4) at six months. The mean change was 0.2 (SD = 0.6), which did not reach statistical significance, t(29) = 1.54, p =.067, 95% CI [–0.05, 0.39; Fig. 3].

Fig. 1.

Comparison of SPADI scores at baseline and 6-month follow-up

Fig. 2.

Pain intensity (NRS 0–10) at baseline versus 6-month follow-up

Fig. 3.

Negative affect scores at baseline and 6-month follow-up

Moderation analysis examined whether the relationship between improvement in shoulder function (predictor) and reduction in pain (outcome) was moderated by changes in negative affect. The analysis revealed that greater improvement in shoulder function was significantly associated with greater reduction in pain (β = 0.06, p =.018). Negative affect itself was also associated with pain levels, such that increases in negative affect were linked to worse pain outcomes (β = − 1.76, p =.037). Importantly, the interaction between improvement in function and change in negative affect was significant (β = 0.16, p =.017), confirming a moderation effect (Table 2).

Table 2.

Summary of regression model output

Outcome	Predictor	Coefficient	p-value
Pain Reduction	Change in Function	0.06	0.018
Pain Reduction	Change Negative Affect	-1.76	0.037
Pain Reduction	Interaction: Change in Function*Change in Negative Affect	0.16	0.017

This interaction indicates that the association between improved function and reduced pain was stronger for individuals who also experienced improvements in negative affect. In other words, the beneficial impact of functional gains on pain reduction was amplified in the context of reduced negative affect. (Fig. 4).

Fig. 4.

The moderating effects of negative affect between changes in function and pain at 6 months

Discussion

This study aimed to examine whether changes in negative affect moderate the relationship between functional improvement and pain reduction. At six months postoperatively, most patients experienced notable improvements in pain; however, 16% continued to report persistent pain, defined as greater than 3 out of 10, aligning with previous literature [20, 21]. Functional outcomes also showed overall improvement, as did levels of pain. While often overlooked, negative affect plays an important role in recovery following SA particularly in individuals with RA, who may exhibit heightened pain sensitivity and emotional vulnerability. These findings highlight the need to more systematically assess and address psychological factors in postoperative care and rehabilitation planning. The present findings highlight the complex interplay between physical function, negative affect, and pain outcomes following shoulder arthroplasty. Consistent with prior research, improvements in shoulder function were significantly associated with reductions in pain [8]. This aligns with the theoretical and clinical understanding that functional recovery often alleviates mechanical and activity-related discomfort [22].

Interestingly, changes in negative affect were also significantly associated with pain outcomes. Patients who reported worsening emotional states experienced higher levels of pain, reinforcing the biopsychosocial model of pain, where psychological states influence physical symptomatology. The significant interaction between functional improvement and negative affect change indicates a moderation effect: the relationship between function and pain was stronger among individuals whose emotional well-being also improved. In other words, the benefit of functional gains on pain reduction was most pronounced when accompanied by improvements in mood or emotional health. This finding underscores the importance of addressing emotional and psychological health in postoperative care. Several limitations should be considered when interpreting these findings. First, the retrospective design introduces inherent biases, including potential selection and information bias, and precludes any conclusions about causality. The relatively small sample size may further limit the generalizability of the results and reduce the power to detect more nuanced effects. Additionally, all measures such as on pain, function, and negative affect were self-reported, making them susceptible to subjective bias. While negative affect was examined, other relevant psychological factors such as anxiety, depression, coping strategies, or social support were not assessed and may have contributed to pain perception and recovery. The study also only included follow-up data up to six months postoperatively, which may not fully capture longer-term outcomes or emotional fluctuations. While we accounted for key demographic variables and preoperative pain levels in our analysis, we did not include factors such as postoperative physical therapy utilization or preoperative opioid use, which may also influence functional recovery. However, opioid use is uncommon in the Swiss healthcare setting, and preoperative opioid prescriptions were rare in our sample. Future prospective studies with comprehensive psychological assessments and longer follow-up periods are warranted.

Collectively, these findings underscore the need for a holistic rehabilitation approach that goes beyond physical recovery to include emotional support. While functional improvement remains a key target in shoulder arthroplasty rehabilitation, the results highlight those changes in emotional well-being, particularly reductions in negative affect play a vital role in pain outcomes. Mental health remains an often-underestimated component of care in patients with RA, who may be more emotionally vulnerable and sensitive to pain. Integrating mental health screening and interventions into standard postoperative care may enhance overall outcomes and support a more comprehensive, patient-centered recovery process in shoulder arthroplasty and RA.

Abbreviations

RA

Rheumatoid Arthritis

SA

Shoulder Arthroplasty

SAR

Shoulder Arthroplasty Registry

TSA

Total Shoulder Arthroplasty

SPADI

Shoulder Pain and Disability Index

CS

Constant Score

NRS

Numeric Rating Scale

ASA

American Society of Anesthesiologists

Author contributions

AL-writing original draft, conceptualization, formal analysis, methodologyTS-reviewing original draft, data curation, software, visualizationDB-reviewing original draft, data curationRE-reviewing original draft, conceptualization, methodologyMS-supervision, conceptualization, reviewing original draft, investigation.

Funding

No specific financial source of funding.

Data availability

The datasets used and/or analysed during the current study are available from the corresponding author on reasonable request.

Declarations

Ethics approval and consent to participate

The study was performed in accordance with the standards of the Ethics Committee of Zurich (Kantonale Ethikkommission [KEK], Stampfenbachstrasse 121, CH-8090 Zurich, Switzerland; KEK-ZH-Nr. 2014–0483) and with the 1964 Helsinki declaration and its later amendments or comparable ethical standards. All patients provided written informed consent prior to patient enrolment / data collection and use of their data for research purposes.

Consent for publication

Not applicable.

Competing interests

MS is a paid consultant of and receives royalties from Stryker Inc.