Efficacy of an acute bout of isometric wall squat exercise on pain sensitivity and clinical pain intensity in adults with knee osteoarthritis: a randomised controlled trial in outpatient physiotherapy clinics in Saudi Arabia

Abstract

Objective

To investigate the efficacy of two volumes of acute isometric wall squat exercise on pain sensitivity (primary outcome: pressure pain threshold (PPT)) and intensity (secondary outcome) in adults with knee osteoarthritis.

Design

A parallel-group randomised controlled trial.

Setting

Outpatient physiotherapy clinics at hospitals in Saudi Arabia.

Participants

Participants (n = 90; mean (SD) age 49.0 (13.0) years) with chronic knee osteoarthritis, diagnosed by a specialist orthopaedic physician using radiographic evidence and clinical judgement, were recruited via two tertiary hospitals in Saudi Arabia.

Intervention

Participants were stratified by sex and randomly allocated to one of three groups: isometric single-repetition wall-squat (ISO-SR; one repetition of a 3-min wall squat or to volitional fatigue at a 100° knee joint angle), isometric multiple-repetition wall-squat (ISO-MR; three repetitions of the same intervention) and control (quiet sitting for 7 min).

Primary and secondary outcome measures

PPT was measured at the calf (primary site of interest), lumbar spine and forearm using a pressure algometer. Pain intensity was assessed using a visual analogue scale. Primary analyses compared changes between intervention and control groups. All participants were included in the ITT analysis.

Results

The ISO-MR group had higher mean PPT ratings immediately postintervention in the calf (mean difference=9.4 (95% CI 5.6 to 13.2), p<0.001), forearm (4.0 (95% CI 1.2 to 6.9), p=0.006) and lumbar spine (5.4 (95% CI 2.4 to 8.4), p<0.001), compared with control. The ISO-SR group reported a greater increase in PPT when compared with control at the calf (5.8 (95% CI 2.0 to 9.6), p=0.003) and lumbar spine (3.7 (95% CI 0.7 to 6.8), p=0.015). There were no differences in clinical pain intensity between the groups.

Conclusions

Both the ISO-SR and ISO-MR exercises to volitional fatigue reduced pain sensitivity in patients with knee osteoarthritis. Neither exercise volume changed clinical pain intensity.

Trial registration number

ClinicalTrials.gov registry (ID: NCT05605444).

STRENGTHS AND LIMITATIONS OF THIS STUDY.

• The randomised controlled trial design and allocation concealment enhanced internal validity.

• Validated outcome measures (pressure pain threshold and visual analogue scale) were assessed immediately before and after the intervention.

• A simple, clinically feasible isometric wall squat protocol was tested in outpatient physiotherapy clinics.

• Prespecified isometric hold durations were not consistently achieved; despite instructions to hold as long as possible, observed time-under-tension was lower than intended and maximum hold capacity was not verified.

• The absence of assessor blinding may have introduced measurement bias.

Introduction

Knee osteoarthritis is a prevalent chronic musculoskeletal condition and a leading cause of disability globally.¹ Pain is the primary symptom reported by individuals with knee osteoarthritis and is often accompanied by heightened pain sensitivity and deficits in endogenous pain modulation mechanisms.2,4 This persistent pain substantially interferes with daily physical function, reduces participation in daily activities and leads to progressive physical deconditioning, which negatively impacts quality of life.^{5 6} Given the substantial personal and healthcare burdens associated with knee osteoarthritis, identifying effective, non-pharmacological pain relief strategy is a critical priority in contemporary clinical management.

Exercise-induced hypoalgesia (EIH) is a phenomenon in which a single bout of exercise leads to subsequent reduction in pain sensitivity.^{7 8} EIH is proposed to occur through activation of central inhibitory mechanisms and modulation of pain processing within the central nervous system.^{7 8} It is generally considered a short-term effect, lasting from several minutes to a few hours after exercise, with limited evidence supporting sustained benefits across the entire day.^{7 8} Pain sensitivity is assessed by determining the pain threshold, which is the minimum intensity of a noxious stimulus perceived as painful.⁹ Methods to quantify pain sensitivity include thermal, chemical and mechanical stimuli. However, mechanical pressure pain threshold (PPT) method is preferred for evaluating pain sensitivity in the context of EIH.¹⁰

The phenomenon of EIH, first investigated in pain-free athletes about 40 years ago,¹¹ has recently gained more attention as a possible non-pharmacological treatment of musculoskeletal pain.^{9 12 13} The first two meta-analyses showed positive effects of aerobic and isometric exercises on pain perception in pain-free individuals, though the results for isometric exercise were inconsistent.^{10 12} Naugle et al¹⁰ (25 studies, 622 participants) reported a large pooled effect size for isometric exercise on PPT; however, this finding was based on non-randomised trials, limiting their reliability. In contrast, Wewege and Jones¹² (13 studies, 537 participants) concluded that the effect of isometric exercise was negligible and highlighted that their findings were derived from very low-quality randomised controlled trials. Both studies emphasised the need for more rigorous research to examine the causal effects of isometric exercise, particularly in populations with chronic musculoskeletal pain. Furthermore, a recent meta-analysis supports the use of aerobic exercise in the management of pain, especially at higher intensities, but the best forms of exercise, especially isometric exercise, are still unclear.¹⁴

Despite a growing number of meta-analyses,^{10 12} the optimal exercise prescription to reliably induce an acute EIH response from a single exercise bout in patients with chronic musculoskeletal pain remains unknown. Furthermore, previous meta-analyses on EIH have included mainly healthy or other chronic pain populations (eg, Naugle et al, no knee osteoarthritis studies; Wewege and Jones, only one knee osteoarthritis study), underscoring the limited evidence in knee osteoarthritis and the need for targeted investigation. Moreover, investigating this in people with knee osteoarthritis is important for several other reasons including the high prevalence of knee osteoarthritis, which is estimated to be 365 million cases globally and is among the highest contributors to disability.¹⁵ Although international clinical guidelines16,19 recommend aerobic and resistance exercise training for managing knee osteoarthritis, many individuals do not adhere to exercise.^{20 21} This is partly due to current knee pain or fear of exacerbating the condition.²² Osteoarthritis, which is characterised by chronic joint pain and stiffness, poses significant challenges to effective pain management.²³

In people with knee osteoarthritis, heightened pain sensitivity is commonly observed, leading to persistent pain and disability.²⁴ Inducing EIH may help temporarily decrease pain sensitivity and intensity by modulating these mechanisms. Further, acute reductions in pain through EIH could attenuate pain flares, which are frequently experienced by individuals with knee osteoarthritis and are associated with functional limitations and decreased quality of life.²⁴ If exercise can reliably and immediately reduce pain, it might immediately address these barriers via repeated exposure and improve adherence by alleviating pain-related fears and enhancing expectations of long-term clinical efficacy. Therefore, it is essential to identify exercise prescriptions that can lead to an immediate and meaningful reduction in pain (sensitivity and intensity).

This study aimed to examine the efficacy of two volumes of isometric wall squat exercise (isometric multiple-repetition wall-squat (ISO-MR) and isometric single-repetition wall-squat (ISO-SR)) in reducing pain sensitivity in this population. The secondary aims included examining its efficacy in decreasing clinical pain intensity, as well as conducting exploratory analysis to evaluate the differences and interaction effects of total time under tension (TUT) and expectations for change in PPT and pain intensity.

Methods

Study design

This study was a parallel-group randomised controlled trial. The Consolidated Standards of Reporting Trials (CONSORT) flow diagram illustrates the flow of participants through each stage of the study (figure 1). The trial protocol was approved by the General Administration for Research and Studies of the Directorate of Health Affairs, Makkah, Saudi Arabia (H-02-K-076-0722-770). The trial protocol was registered in the ClinicalTrials.gov registry (ID: NCT05605444) and was conducted in accordance with the Declaration of Helsinki.²⁵ Written informed consent was obtained from all participants prior to the study. This clinical trial was reported following the guidelines of the CONSORT guidelines (online supplemental checklist 1).²⁶

Figure 1. Flow diagram of participants through study.

Participants

This study included participants who were aged ≥18 years, diagnosed by orthopaedic surgeons using radiological imaging and clinical judgement with unilateral or bilateral symptomatic knee osteoarthritis, had chronic knee pain for at least 3 months (with an average pain intensity of at least 3/10 on the visual analogue scale (VAS) over the last week), and were able to perform exercise based on the subjective reporting by participants (by answering ‘yes’ to the question ‘can you perform exercises?’. This study excluded pregnant women and individuals who had severe musculoskeletal disorders (knee varus/valgus), neuromuscular disorders (peripheral neuropathy and balance problems), uncontrolled diabetes, sensory problems, lower limb surgery, orthopaedic impairments (eg, ligament injuries, severe deformities or recent fractures in the last 6 months) and/or any contraindication to exercise that prevented them from performing exercises as per the ‘American College of Sports Medicine’ guidelines.²⁷ Participants with varus/valgus alignment were excluded to decrease biomechanical variability and ensure a focused investigation of the intervention’s efficacy on neuromuscular function and pain modulation.

Participants were recruited using consecutive sampling techniques from outpatient clinics at King Abdul Aziz Hospital & King Faisal Hospital, Makkah, Saudi Arabia. Participants who met the eligibility criteria and provided written informed consent were immediately randomised and allocated to one of three groups. The allocated intervention was applied during the same visit, and outcome measures were collected immediately before and immediately after the intervention. Qualified physiotherapists assessed all participants for eligibility. Data were collected between November 2022 and April 2023.

Interventions

The intervention was described following the Template for Intervention Description and Replication checklist to ensure standardised and comprehensive reporting.²⁸ The participants were randomised equally into an ISO-MR, ISO-SR or control group. The two wall squat exercise groups consisted of either multiple- or single-repetition isometric wall squat holds. Isometric exercise was defined as muscular contraction with no change in muscle length and no observable movement at the joint. Participants were instructed to place their back against the wall with their feet positioned parallel and spaced shoulder-width apart and then lower themselves down until they reached a knee joint angle of 100°. The knee joint angle was measured by placing a clinical goniometer on the lateral epicondyle of the knee in line with the femur while the anchor arm was in line with the lateral malleolus. The exercise protocols were matched for intensity (body weight held for 3 min or until volitional fatigue) but differed with regard to the total exercise TUT. The ISO-MR group completed three repetitions of isometric wall squats maintained for 3 min or until volitional fatigue at a knee joint angle of 100°. The participants rested for 30 s between repetitions. Neutral language and verbal suggestions regarding the expected exercise effects were avoided. The ISO-SR group performed one repetition of isometric wall squats maintained for 3 min or until volitional fatigue at a knee joint angle of 100°. The same neutral language and avoidance of verbal suggestions of the expected effects of exercise on pain were implemented consistently in the ISO-MR group. The control group did not receive active intervention and sat quietly without conversation for 7 min. All participants were provided with general information about study procedures and safety but were not informed of the specific hypotheses or expected group differences to minimise expectation bias. Duration was calculated as the average attention time of the two active intervention groups. The intervention was administered in the physiotherapy outpatient clinics by physiotherapists with 10–12 years of clinical experience. Participants attended a single intervention session (range, 3–9 min) based on the allocated intervention group. Because the achieved exercise dose in efficacy trials often differs from the intended prescription, we recorded TUT during the session as a measure of dose fidelity. We also used TUT to confirm that the ISO-SR and ISO-MR protocols produced distinct doses of isometric wall squat.

To ensure intervention fidelity, participants’ adherence to the prescribed protocols was monitored. In both intervention groups, ISO-SR and ISO-MR groups, all participants successfully attempted all the repetitions of isometric exercise, and the duration of each repetition was recorded. The average recorded time for each repetition in both groups is provided in online supplemental table 1 and figure 1. These measures were implemented to determine adherence to the allocated intervention and to provide insights into fidelity.

Outcomes

The primary (pain sensitivity) and secondary (pain intensity) outcomes were assessed at baseline and immediately postintervention (1–3 min). Pain sensitivity was assessed as the difference in the change from baseline in PPT between each intervention group and control group at the calf (primary interest site), lumbar and forearm. Pain sensitivity was measured by the PPT using an established protocol.²⁹ The PPT was measured using a pressure algometer (Wagner Digital Force Ten FDX 25 with a 1 cm² rubber tip (Wagner Instruments, Greenwich, CT, USA)). This device has been widely used in different studies internationally30,32 and has high intra- and inter-rater reliability in measuring the PPT among patients with knee OA.^{27 33} If postintervention PPT was higher than the pre-intervention value, it indicated that pain sensitivity had decreased, suggesting that EIH had occurred. If the postintervention PPT was lower than the pre-intervention PPT, it indicated that pain sensitivity had increased, suggesting that EIH had not occurred.

The assessment sites for PPT were as follows: (1) right and left (local site) mid-calf muscles between the two muscle heads of the gastrocnemius 1/3 distally; (2) right and left lumbar (semilocal site) at the midline between the iliac crest and 4 cm away from the midline and (3) right and left forearms (remote site) at the extensor carpi radialis muscle, approximately 3 cm distal and posterior to the lateral epicondyle of the humerus. The local site served as the primary interest site, with the semi-local and remote sites serving as secondary interest sites.

Familiarisation with pressure was applied to the participants at each assessment site at a rate of 10 N/s. Pressure was gradually applied until the pressure turned into pain. The participants were instructed to say (stop) when they first felt pain. The PPTs were assessed while the participants were lying in a prone position on an examination table. A marker was used to mark the location of each test site to ensure consistency. The algometer device was positioned perpendicular to the measurement site for all trials. The pressure was increased at a rate of 10 N/s for all recorded tests. Two trials were conducted at each site. There were approximately 20 s of rest before the individual sites were tested. The raw values of each test were recorded, and all values were retained for analysis.

The current clinical pain intensity in the knee was assessed using the VAS (0–10).³⁴ A rating of zero means that participants do not have knee pain, and a rating of 10 means that pain is the worst possible. This is a reliable and valid tool for measuring pain in patients with chronic musculoskeletal disorders.³⁴ A change in pain intensity at one point was considered a minimal clinically important difference.³⁵

Self-reported expectations of changes in the PPT and pain intensity were collected using two single-item questions with a neutral language script.³⁶ The first question on PPT asked ‘If you think back to the time just before you did the wall squat exercise. What impact did you expect that the squat exercise would have on how much pressure would be needed before you experienced the first sensation of pain? For example, did you expect you’d need less or more pressure before you felt pain, or that there would be no change’. The scale ranged from −10 (expectation of much less pressure needed to reach pain) to 0 (expectation that there would be no difference in pressure to reach pain) to +10 (expectation of a lot more pressure needed to reach pain). The second question asked participants to respond to a similar question about pain intensity at the knee: ‘If you think back to the time just before you did the squat exercise. What impact did you expect that the squat exercise would have on your knee pain? For example, did you expect you’d experience less or more pain, or that there would be no change in your pain’. The scale ranged from −10 (expectation that exercise would significantly decrease knee pain) to 0 (expectation of no change in knee pain) and +10 (expectation that exercise would significantly increase knee pain). Self-reported expectations of pain relief were assessed immediately after the intervention to avoid influencing participants’ performance during the exercise itself.

Sample size and power calculation

A power analysis was conducted to determine the appropriate sample size for this study with three groups. Pilot data from our lab with participants (n=21) who performed an isometric wall squat exercise for 3 min or until volitional fatigue showed a within-group effect size (Cohen’s d) of 0.5–0.8. Previous research on adults with knee osteoarthritis reported a between-group effect size (Cohen’s d) of 0.6, which was measured in the knee following isometric knee extension.³⁷ Based on these findings, we conservatively estimated a between-group effect size of 0.35 for comparisons between each intervention group and the control group for PPT.

Using G*Power 3.1, an analysis of variance (ANOVA) with repeated measures and within–between-factor interaction effects indicated that a total sample size of 81 participants was required to detect a statistically significant difference between the groups with 80% power at a 2.5% significance level (Bonferroni correction to adjust for multiple comparisons based on an intended significance level of 5%). We considered a moderate correlation between repeated measures to be r=0.6. We recruited 30 participants per group to account for missing data and maintain a balanced design.

Randomisation

The randomisation was computer-generated by a researcher not otherwise involved in the study. Participants were stratified based on sex (male and female) and then allocated to each group in a 1:1 ratio (15 female/15 male for each group (ISO-SR, ISO-MR and control)). To maintain allocation concealment, a researcher not otherwise involved in the study prepared the sequentially numbered, opaque, sealed envelopes containing the randomisation assignments. These envelopes were prepared in advance and opened only after enrolling participants and finishing the baseline assessment. The participants were not informed about the specific goals or hypotheses of the study or the differences between the intervention and control groups; however, they were aware that they were participating in a study on the effects of exercise. The participants attended a 2-hour testing session at the outpatient clinic, where their demographic details were collected. This was followed by the collection of baseline outcome measures, delivery of the intervention and the subsequent assessment of outcome measures.

Statistical analysis

Data analysis was performed using R software (V.4.2.0; R Foundation for Statistical Computing, Vienna, Austria). Baseline characteristics were summarised using descriptive statistics; continuous variables are expressed as the mean and SD, while categorical variables are reported as frequencies and percentages. Separate linear mixed-effects models (LMMs) were constructed to examine the effects of interventions over time, using the ‘lme4’ package for model fitting and the ‘lmerTest’ package to facilitate hypothesis testing, including calculations of p-values and 95% CIs. The ‘emmeans’ package was employed for post hoc analyses, specifically for estimating marginal means and performing comparisons between group levels.

The primary analyses focused on examining differences in the change from baseline between groups, specifically comparing each intervention group to the control group, focusing on group-by-time interaction effects. All the other analyses conducted were considered secondary.

Separate LMMs were constructed for each outcome measure (PPT at calf, lumbar and forearm; pain intensity). Each model included fixed effects for Group, Time and their interaction (Group × Time), with participant ID as a random effect to account for the correlation of repeated measures within individuals. This approach controlled for individual variability and differential responses to exercise interventions across groups. This approach also allowed site-specific inference and avoided confounding due to anatomical variability (ie, regions close to or remote from the exercising body area). In these between-group models, the Group×Time coefficients represent the difference in pre-to-post change for each treatment group relative to the reference group (Control), expressed in N/cm² for PPT and VAS units (0–10) for knee pain intensity. Positive values indicate a larger increase in PPT (reduced pain sensitivity) versus Control. Negative values indicate a larger decrease in knee pain intensity versus Control.

For the within-group analyses, separate LMMs were developed for each intervention group (ISO-SR, ISO-MR and control). These models included random (participants) and fixed effects for time (pre- and postintervention) to capture within-group changes over time for each outcome (PPT and pain intensity). In the within-group models, the Time coefficient estimates the mean pre-to-post change within that group expressed in N/cm² for PPT and VAS units (0–10) for knee pain intensity.

All models were estimated using restricted maximum likelihood to reduce biased variance component estimates, particularly when comparing models with different fixed effects structures. The analysis adopted an intention-to-treat (ITT) approach to analyse participants according to their original group assignments. An unadjusted significance level (α) of 0.05 was used for all inferential statistics. This was a single-session, pre–post design with outcomes evaluated immediately after the intervention. All participants completed the intervention and postassessments, resulting in no missing data. Accordingly, the ITT and per-protocol analyses were identical, and all randomised participants were included in the final analysis as allocated.

An exploratory analysis was conducted to investigate the differences between the two intervention groups for squat hold TUT (recorded to assess dose fidelity and confirm dose separation between intervention groups), and across the three groups for expectations of changes in PPT and knee pain intensity. Owing to the violations of normality assumptions for both the squat-hold TUT and the expected data, non-parametric analyses using the Mann–Whitney U test and Kruskal–Wallis test were employed to compare squat-hold TUT and self-reported expectations, respectively, between groups.

Additionally, separate post hoc LMMs were used to test whether TUT or expectations moderated the PPT response. In each model, fixed effects included Group (single-rep reference), Time (pre vs post; prereference), the moderator (TUT, continuous, for intervention groups only; or expectations, continuous, for all groups), and all interactions, with participant ID as a random intercept. PPT measurements from calf, lumbar and forearm were pooled to estimate an overall moderation effect. This approach allowed us to account for within-subject variability by including random intercepts for participants. The Moderator (TUT or expectations) × Group × Time interaction term quantifies whether the association between each moderator and pre-to-post change in PPT differs between intervention groups.

Patient and public involvement

This study did not involve patients or members of the public in the design, development of the research questions or outcome measures, recruitment, conduct or dissemination plans.

Results

We screened 122 adults for eligibility; 90 met the inclusion criteria and provided informed consent to participate in the study prior to being randomised (figure 1). The mean (SD) age of participants was 49.0±13.0 years and mean (SD) of body mass index was 28.7±5.2 kg/m². Most participants were married (68.9%), had a diploma degree or below (61.1%), were employed full-time (64.4%) and were non-smokers (83.3%). Participants reported a mean (SD) pain frequency of 5.0 (1.7) days experienced over the past week and a current pain intensity of 3.6 (1.6). Table 1 presents the characteristics of the participants.

Table 1. Participant characteristics presented as mean (SD).

	Total number of participants (n=90)	Experimental (ISO-MR) (n=30)	Experimental (ISO-SR) (n=30)	Control (n=30)
Characteristics
Age (years)	49.0 (13.0)	45.8 (10.8)	47.0 (14.8)	54.1 (11.9)
Sex, n (%)
Male	45 (50.0)	15 (50.0)	15 (50.0)	15 (50.0)
Female	45 (50.0)	15 (50.0)	15 (50.0)	15 (50.0)
Body mass index (kg/cm2)	28.7 (5.2)	28.4 (5.5)	27.6 (4.3)	30.1 (5.5)
Pain frequency (days/week)	5.0 (1.7)	4.7 (1.8)	4.8 (1.9)	5.7 (1.3)
Prepain current (visual analogue scale (VAS), 0–10)	3.6 (1.6)	3.2 (1.7)	3.3 (1.3)	4.2 (1.5)
Western Ontario and McMaster Universities Arthritis Index (WOMAC), 0–96	37.5 (21.2)	46.7 (22.6)	32.9 (18.1)	32.7 (20.2)
Smoking status, n (%)
Current smoker	15 (16.7)	3 (10.0)	4 (13.4)	8 (26.7)
Not smoker	75 (83.3)	27 (90.0)	26 (86.7)	22 (73.3)
Marital status, n (%)
Married	62 (68.9)	22 (73.3)	17 (56.7)	23 (76.7)
Single	23 (25.6)	7 (23.3)	11 (36.7)	5 (16.7)
Widowed	3 (3.3)	1 (3.3)	0 (0.0)	2 (6.7)
Divorced	2 (2.2)	1 (3.3)	1 (3.3)	0 (0.0)
Education, n (%)
Bachelor or above	35 (38.9)	15 (50.0)	13 (43.3)	7 (23.3)
Diploma or below	55 (61.1)	15 (50.0)	17 (56.7)	23 (76.7)
Employment, n (%)
Full time	58 (64.4)	20 (66.7)	18 (60.0)	20 (66.7)
Part time	6 (6.7)	3 (10.0)	2 (6.7)	1 (3.3)
Self-employed	1 (1.1)	0 (0.0)	0 (0.0)	1 (3.3)
Unemployed	5 (5.6)	1 (3.3)	3 (10.0)	1 (3.3)
Retired	20 (22.2)	6 (20.0)	7 (23.3)	7 (23.3)

Groups: isometric multiple-repetition wall-squat (ISO-MR), isometric single-repetition wall-squat (ISO-SR) and Control.

Primary outcomes

The efficacy of an acute bout of exercise on the PPT in adults with knee osteoarthritis is shown in table 2. There were group-by-time interaction effects for each body region, where participants receiving ISO-MR exercises reported a higher mean change in PPT immediately postintervention in the calf (mean difference=9.4 N/cm² (95% CI 5.6 to 13.2), p<0.001), forearm (4.0 (95% CI 1.2 to 6.9), p=0.006) and lumbar (5.4 (95% CI 2.4 to 8.4), p<0.001) when compared with the control group. Moreover, participants receiving ISO-MR exercise reported a higher mean change in PPT rating immediately postintervention in forearm (4.8 (95% CI 1.7 to 7.9), p=0.003) but not at the calf (3.6 (95% CI −0.4 to 7.5), p=0.077) or at the lumbar region (1.7 (95% CI −1.7 to 5.0), p=0.327) when compared with the ISO-SR group. Furthermore, the single-repetition exercise group reported a higher mean PPT rating in the calf (5.8 (95% CI 2.0 to 9.6), p=0.003) and lumbar spine (3.7 (95% CI 0.7 to 6.8), p=0.015), but not in the forearm (−0.7 (95% CI −3.6 to 2.2), p=0.617) when compared with the control group.

Table 2. Change in pressure pain threshold (PPT, N/cm²) following exercise at calf, lumbar and forearm regions: within-group and between-group comparisons from linear mixed-effects models.

Outcomes	Experimental (ISO-MR) (n=30)	Experimental (ISO-SR) (n=30)	Control (n=30)	Group-by-time interaction effect
				ISO-MR vs Control		ISO-SR vs Control		ISO-MR vs ISO-SR
	Mean (SD)	Mean (SD)	Mean (SD)	Mean difference (95% CI)	P value	Mean difference (95% CI)	P value	Mean difference (95% CI)	P value
PPT, calf
Pre	51.5 (26.5)	55.7 (28.4)	56.6 (31.9)
Post	55.7 (29.1)	56.4 (25.9)	51.4 (31.9)
Within-group difference, mean difference (95% CI)	4.3 (1.4 to 7.1)	0.7 (−2.1 to 3.5)	−5.1 (−7.6 to –2.6)	9.4 (5.6 to 13.2)	<0.001	5.8 (2.0 to 9.6)	0.003	3.6 (−0.4 to 7.5)	0.077
Within-group difference, P value	0.003	0.642	<0.001
PPT, lumbar
Pre	55.9 (26.8)	59.1 (28.0)	56.0 (32.8)
Post	58.4 (31.1)	60.0 (29.3)	53.1 (32.5)
Within-group difference, mean difference (95% CI)	2.6 (−0.0 to 5.2)	0.9 (−1.2 to 3.0)	−2.9 (−4.4 to –1.3)	5.4 (2.4 to 8.4)	<0.001	3.7 (0.7 to 6.8)	0.015	1.7 (−1.7 to 5.0)	0.327
Within-group difference, P value	0.055	0.403	<0.001
PPT, forearm
Pre	41.3 (21.3)	45.3 (23.4)	47.9 (31.0)
Post	44.2 (26.2)	43.4 (25.3)	46.7 (35.0)
Within-group difference, meandifference (95% CI)	2.9 (0.6 to 5.2)	−1.9 (−4.0 to 0.2)	−1.2 (−2.9 to 0.6)	4.0 (1.2 to 6.9)	0.006	−0.7 (−3.6 to 2.2)	0.617	4.8 (1.7 to 7.9)	0.003
Within-group difference, P value	0.013	0.077	0.194

Outcomes were analysed with a linear mixed-effects model including fixed effects for Group, Time (pre vs post) and their interaction (Group×Time), plus a random intercept for participant. Within-group difference denotes the model-estimated pre–post difference within each group (PPT, N/cm²). Between-group difference denotes the contrast of these pre–post changes vs Control (PPT, N/cm²). Groups: isometric multiple-repetition wall-squat (ISO-MR), isometric single-repetition wall-squat (ISO-SR) and Control.

Bold values denote statistically significant differences at p<0.05.

Participants in the control group demonstrated a within-group reduction in PPT (hyperalgesia) at the calf (−5.1 (95% CI −7.6 to –2.6), p<0.001) and lumbar (−2.9 (95% CI −4.4 to –1.3), p<0.001) anatomical regions, but no change at the forearm region (−1.2 (95% CI −2.9 to 0.6), p=0.194).

Secondary outcomes

The efficacy of an acute bout of isometric wall squat exercise on clinical pain intensity among adults with knee osteoarthritis is shown in table 3. There were within-group reductions in pain intensity across all groups at follow-up (p<0.05); however, there was no group-by-time interaction effect between any of the three groups on the change in pain intensity at follow-up.

Table 3. Knee pain intensity (visual analogue scale (VAS), 0–10) at baseline and immediately postintervention, and between-group differences in change estimated by linear mixed-effects models (LMMs).

Outcomes	Experimental (ISO-MR) (n=30)	Experimental (ISO-SR) (n=30)	Control (n=30)	Group-by-time interaction effect
				ISO-MR vs Control		ISO-SR vs Control		ISO-MR vs ISO-SR
	Mean (SD)	Mean (SD)	Mean (SD)	Mean difference (95% CI)	P value	Mean difference (95% CI)	P value	Mean difference (95% CI)	P value
Pain intensity
Pre	3.2 (1.7)	3.3 (1.3)	4.2 (1.5)
Post	2.4 (1.9)	2.4 (2.3)	3.4 (1.8)
Within-group difference, mean difference (95% CI)	−0.8 (–1.5 to –0.1)	−0.9 (–1.7 to –0.1)	−0.8 (–1.5 to –0.0)	0.0 (–1.0 to 1.0)	0.949	−0.1 (–1.1 to 0.9)	0.798	0.1 (–0.9 to 1.1)	0.847
Within-group difference, P-value	0.023	0.030	0.049

Outcomes were analysed with a linear mixed-effects model including fixed effects for Group, Time (pre vs post) and their interaction (Group×Time), plus a random intercept for participant. Within-group difference denotes the model-estimated pre–post difference within each group (VAS, 0–10). Between-group difference denotes the contrast of these pre–post changes vs Control (VAS, 0–10). Groups: isometric multiple-repetition wall-squat (ISO-MR), isometric single-repetition wall-squat (ISO-SR) and Control.

Bold values denote statistically significant differences at p<0.05.

Exploratory analysis

The Mann–Whitney U test revealed a significant difference in total squat hold time between the ‘ISO-MR’ and ‘ISO-SR’ groups (W=663.5, p=0.002). The median total hold time for the ‘ISO-MR’ group was 52 s (IQR: 43.2 s), whereas the median for the ‘ISO-SR’ group was 22.5 s (IQR: 30.8 s). The difference in medians between the groups was 29.5 s, indicating that the ‘ISO-SR’ group had a significantly lower total hold time compared with the ‘ISO-MR’ group as intended. The squat hold time data are presented in online supplemental table 1 and figure 1.

An LMM was employed to evaluate the interaction effect of the total squat hold time, times and group allocation on the PPT. The model included a random intercept to account for within-subject variability. There was a significant three-way interaction effect between the total hold time, time point and group on the PPT (estimate=0.1, p=0.001).

The Kruskal–Wallis test shows no significant differences in expectations for PPT, and pain intensity were observed among the three groups, with χ² (2, n=90) = 1.0, p=0.617, and χ² (2, n=90) = 1.2, p=0.557, respectively. The median (IQR) was approximately zero for all three groups (online supplemental table 2).

A linear mixed model was used to evaluate the interaction effects of expectations, time and group allocation on the PPT. The interaction between expectations, time point and the ISO-MR group showed no significant effect on PPT when compared with the control group (estimate = −0.1, p=0.791). Similarly, the interaction between expectations, time point and the ISO-SR group was not significant compared with the control (estimate=0.93, p=0.057).

Discussion

The primary aim of this study was to evaluate the efficacy of acute isometric wall squat exercises on the PPT in patients with knee osteoarthritis. This study showed that EIH occurred in the ISO-MR group compared with the control group postintervention at all three anatomical sites. Additionally, an ISO-SR also led to hypoalgesia at the two sites close to the exercising muscles (calf and lumbar) after the intervention compared with the control group.

Secondary analysis comparing both exercise interventions revealed a group-by-time interaction effect at the forearm, but not at the calf and lumbar regions, in favour of the ISO-MR group for PPT at follow-up. These findings suggest that a longer contraction time is likely necessary for global hypoalgesia. This was confirmed by an exploratory LMM, which demonstrated a three-way interaction between total squat hold time (TUT), group allocation and time for change in PPT. In exploratory analysis, greater achieved TUT was associated with larger hypoalgesic changes in the ISO-MR group, suggesting that dose delivered may partly account for variability in acute pain sensitivity responses. Notably, the actual TUT achieved was lower than intended, highlighting the importance of monitoring and reporting dose fidelity in exercise trials.

This is the first study to demonstrate a reduction in pain sensitivity following isometric wall squat exercise in adults with knee osteoarthritis. A previous study by Neelapala et al³⁷ in knee osteoarthritis patients (n=70) showed that 10 repetitions of isometric quadriceps terminal knee extension exercise with a 6 s hold was also effective in reducing pain sensitivity (increased PPT) when compared with the non-intervention control group (5 min quiet rest). However, this study and ours contrast with the findings of an earlier study by Vaegter et al³⁸ that investigated the effect of a single repetition of isometric knee extension in 90° of knee flexion at an intensity of 30% of maximum voluntary contraction for 90 s. However, Vaegter et al³⁸ examined preoperative hypoalgesia in knee osteoarthritis patients awaiting knee replacement (n=14), whereas our study focused on acute EIH in an outpatient population, which may explain differences in findings. Further, this particular study³⁸ did not show a group-by-time interaction effect when compared with the quiet rest control. It is unclear why these contrasting findings existed when the TUT prescribed in this study was greater than that in our study or that of Neelapala et al.³⁷ TUT and the intensity of isometric exercise have previously been reported to moderate the effect size of changes in pain sensitivity in pain-free populations.10 In our study, both groups were asked to hold the isometric wall squat position for up to 3 min or until volitional fatigue. This was intended to achieve high-intensity maximal isometric contractions using body weight. The ISO-SR group completed just 22.5 s (median) of the isometric wall squat; however, this led to similar changes in pain sensitivity as the ISO-MR group for body regions close to the exercising muscles. Therefore, we speculate that intensity is an important consideration for exercise prescription in this context and that exercise to the point of local muscle fatigue is required. A longer TUT may need to be considered when global changes in pain sensitivity are desired. These hypotheses are partially supported by previous evidence of a dose–response relationship observed in healthy populations.³⁹ However, it remains unclear how well these findings translate to adults with knee osteoarthritis, in whom the time to volitional fatigue is shorter than that reported in this study. It also cannot be ruled out that the larger muscle groups involved in the isometric wall squat contributed to the results when compared with these previous studies; however, this does not fully explain the discordant results observed between the two earlier trials, in which both used isolated knee extension exercise. Despite the observation of a statistically significant decrease in pain sensitivity, caution should be made when interpreting these findings due to the absence of a known minimal clinically important difference for measures of PPT. Additional research to evaluate the efficacy of intensity and duration on changes in pain sensitivity is essential for optimising isometric exercise prescriptions.

The results presented here demonstrate that there was no change in clinical knee pain intensity between the groups, which is consistent with an earlier study that compared isometric knee extension exercises with placebo transcutaneous electrical nerve stimulation in adults with knee osteoarthritis.⁴⁰ These findings are important because it remains unclear why there were no observed changes in pain intensity despite changes in pain sensitivity. Pain sensitisation in adults with osteoarthritis may be associated with symptom severity.²⁴ In exercise training studies, pain sensitivity⁴¹ changes over time with repeated exposure.^{42 43} One study also showed that changes in pain sensitivity and pain intensity after 12 weeks of exercise training were correlated.⁴⁴ However, despite the correlation between pain sensitivity and pain intensity, these constructs remain functionally different.⁴⁵ It remains unclear whether it is possible to design an exercise programme that results in an immediate reduction in pain intensity because few studies have reported the outcome after an acute bout of exercise.⁴⁶ There is currently no evidence that acute exercise can immediately reduce pain intensity. However, the evidence to date is far from conclusive given the heterogeneity of the included studies. Owing to this uncertainty, future research is required to determine whether isometric exercise can reduce clinical pain intensity, as observed for pain sensitivity in this study. Future studies on EIH should report both pain sensitivity and clinical pain intensity in the context of musculoskeletal pain conditions.

This study is the first to evaluate the efficacy of two different isometric exercise volumes on pain sensitivity and pain intensity in patients with knee osteoarthritis and is an important step forward in clinical exercise research. While this study demonstrates that higher volumes of isometric exercise performed to volitional fatigue may have advantages over lower volumes in inducing EIH, this was not consistently observed for all sites, including the primary site at calf muscle. Future research is warranted to explore the roles of intensity and duration in optimising exercise prescriptions for pain modulation. Moreover, a more optimal exercise protocol may be necessary to ensure that patients experience predictable and meaningful changes in pain intensity; however, it remains unclear whether the reduction in pain sensitivity demonstrated in this study will translate to functional gain if used prior to the performance of a normally painful functional task, such as stair climbing and walking.

One strength of this trial is that it evaluated the effect of expectations on changes in PPT and clinical pain intensity. The results showed that there were no differences in expectations between the groups, and the mean scores remained close to zero (expected change), suggesting that the findings of this study were not influenced by inherent beliefs about benefit or exacerbation. A strength of this study was the systematic monitoring and reporting of actual exercise exposure in each intervention group, which confirmed dose fidelity and enhanced transparency in evaluating intervention efficacy.

This study had some limitations. First, based on clinical experience, we anticipated that participants would achieve isometric squat hold times of approximately 2 min when coached and encouraged during the task. However, the reported actual TUT in this study demonstrates that the participants achieved hold times well below those of the previously reported squat-hold time norms.^{47 48} The participants in this trial were coached and encouraged to maintain the squat position as long as they were capable; however, it remains unknown whether the participants reached their maximum hold times. Future research should record the intensity of the isometric exercise using the Isometric Exercise Scale⁴⁹ at the end of each minute and the end of the exercise to aid in the interpretation of exercise intensity. Second, the findings of this study may have limited generalisability to an older population. While the exercise approach may be tolerated in adults within the age range reported in this study, it is unclear whether older adults would achieve the same total volume of exercise necessary to achieve such outcomes reported here. Future research should explore adaptations or alternative interventions tailored to older and more diverse populations with knee osteoarthritis such as repeated shorter bouts of isometric exercise. Third, although knee osteoarthritis in this study was confirmed by orthopaedic surgeons using radiological imaging and clinical judgement, we did not apply a standardised clinical diagnostic classification system (eg, American College of Rheumatology (ACR) criteria). Thus, while this reflects real-world clinical practice, it may limit the generalisability of our findings to populations defined according to formal diagnostic criteria. Fourth, the outcome assessor was not blinded to group allocation, which may have introduced potential measurement bias. Fourth, this was an unblinded intervention, and participants’ expectations may have impacted on their responses; however, because pain sensitivity was measured using PPT, which is less prone to expectation bias than self-reported ratings, we consider the potential impact on the results to be minor.^{27 33}

Conclusion

The ISO-SR and ISO-MR exercises to volitional fatigue led to reductions in pain sensitivity when compared with controls in patients with knee osteoarthritis. There were no differences in clinical knee pain intensity between the groups. These findings suggest that isometric wall squat exercise may offer immediate pain relief in knee osteoarthritis, supporting its use in clinical practice. More research is needed to investigate whether repeated or long-term protocols can sustain these effects and improve clinical outcomes.

Data availability statement

Data are available upon reasonable request.