Isometric thumb extension exercise as part of a multimodal intervention for de Quervain’s syndrome: A randomised feasibility trial

Abstract

Introduction

de Quervain’s syndrome is a painful condition commonly presented to hand therapists. Exercise is utilised as an intervention, but isometric exercise has not been investigated. We aimed to assess the feasibility and safety of isometric thumb extension exercise for de Quervain’s syndrome and to explore differences between high-load and low-load isometric exercise.

Methods

This parallel-group randomised clinical feasibility trial included individuals with de Quervain’s syndrome. All participants underwent a 2 week washout period where they received an orthosis, education, and range of motion exercises. Eligible participants were then randomised to receive high or low-load isometric thumb extension exercises, performed daily for 4 weeks. Feasibility and safety were assessed by recruitment and drop-out rates, adherence, adverse events, and participant feedback via semi-structured interviews. Secondary outcomes included patient-reported outcomes for pain and function, and blinded assessment of range of motion and strength.

Results

Twenty-eight participants were randomised. There were no drop-outs after randomisation, and no serious adverse events. Adherence to exercise was 86.7%, with 84% of participants stating they would choose to participate again. There were clinically and statistically significant improvements in pain and function over time (p < 0.001) but not in range of motion or strength. There were no statistically significant between-group differences.

Conclusions

Isometric thumb extension exercise within a multimodal approach appears a safe and feasible intervention for people with de Quervain’s syndrome. A large multi-centre trial would be required to compare high- and low-load isometric exercises. Further research investigating exercise and multimodal interventions in this population is warranted.

Introduction

de Quervain’s syndrome is a condition of the first extensor compartment of the wrist, characterised by radial-sided pain and impaired function. Prevalence is reported as 1.3% in women and 0.5% in men ¹ and most commonly develops between 40 and 59 years of age. ² It is also known to develop in women peri- and postpartum, although prevalence in this population is unknown.^3,4

Operative management is commonly recommended after a period of unsuccessful non-operative management.^5–9 It has been found to be effective in reducing pain and disability in retrospective studies,^6,7,10 prospective case series ¹¹ and randomised clinical trials (RCT) comparing different surgical techniques.^12,13 However, there is a risk of complications, including injury to the superficial branch of the radial nerve, scar complications, and tendon subluxation.^6,13,14 Reported complication rates range from 0% ⁷ to 25%. ¹⁵ Whether operative management is superior to non-operative interventions is unknown as there are no RCTs comparing these treatments.

Non-operative management encompasses intra-compartmental corticosteroid injections, non-steroidal anti-inflammatory drugs, orthoses, exercise, manual therapy, and electrotherapy. Success rates of corticosteroid injections for de Quervain’s syndrome in RCTs range between 50% ¹⁶ and 100%. ¹⁷ Multiple injections are sometimes required for lasting symptom resolution. ¹⁸ Other therapeutic management strategies lack evidence, ⁵ despite being commonly implemented. ¹⁹

A retrospective review of occupational therapy records reported that 83.8% of patients received an exercise intervention for de Quervain’s syndrome. ²⁰ Despite this, evidence investigating exercise for this condition is limited. Randomised trials,^21,22 case series^23–25 and case studies ²⁶ with eccentric exercises as part of a multi-modal approach have also demonstrated improvements in pain and function.

Isometric exercise may be a more appropriate starting point for resistance exercise than eccentric exercise in people with de Quervain’s syndrome. Isometric load does not involve proximodistal glide of the tendons through the first extensor compartment, a consideration given this gliding and associated friction are hypothesised contributors to the painful presentation of this condition. ²⁷ The static nature of isometric load also allows controlled positioning of the wrist and thumb to minimise postures that increase compression (e.g. ulnar deviation).

Isometric exercise has been tested in tendinopathies and found to decrease pain.^28–30 However, the presence of exercise-induced hypoalgesia, being the reduced response to a painful stimulus, ³¹ has not been consistently demonstrated in localised musculoskeletal conditions. ³² Intensity is an important consideration in the prescription of loading programs and exercise-induced hypoalgesia has been observed following both high- and low-load intensity isometric exercises. ³¹ It is hypothesised that high-load programs may be more effective than low-load programs but heterogeneity between published studies means limited direct comparisons are available.^32,33 Further, how this hypothesis translates to conditions involving an overlying retinaculum has not been investigated.

This study aimed to assess the feasibility of conducting a RCT examining an isometric thumb extension exercise intervention for de Quervain’s syndrome, and the safety of this intervention within a multimodal approach. The secondary aims of this study were to explore if there is a difference between high-load and low-load isometric exercise in the outcomes of patient perception of change, pain and function, strength, and range of motion. High- and low-load exercises were chosen as comparisons based on optimising participant recruitment by offering two active interventions (as opposed to a sham or placebo intervention) and to reflect the different intensities of this intervention in musculoskeletal research.

Methods

This parallel-group randomised controlled feasibility trial is reported in accordance with the Consolidated Standards of Reporting Trials 2010 statement ³⁴ and the Template for Intervention Description and Replication checklist. ³⁵ The trial protocol was prospectively registered with the Australian New Zealand Clinical Trials Registry (ANZCTR number: 12619000600145) and approved by the La Trobe University Ethics Committee (HEC18429).

Participants

Potential participants were considered for inclusion if they were aged 18 years or over and met the diagnostic criteria of “probable” or “almost certain” de Quervain’s syndrome on the de Quervain’s Screening Tool ³⁶ (DQST) (score ≥3). The DQST includes seven signs and symptoms of de Quervain’s syndrome, with one point allocated for each positive finding. Individuals were excluded if they had systemic inflammatory conditions, complex regional pain syndrome, concomitant conditions of the affected upper limb that impacted study participation, previous surgery impacting the first extensor compartment of the wrist, corticosteroid injection in the radial wrist region within 3 months of enrolment into the study, injury management funded by a compensable body, or had undergone rigid wrist and thumb orthosis intervention or isometric thumb exercises lasting greater than 2 weeks for the management of their current episode of de Quervain’s syndrome. People who were pregnant were excluded due to the potential change in hormones, physical demands, and potential capacity to adhere to intervention requirements during the study period. In the case of bilateral presentations, an a priori decision was made to enrol the side with the higher DQST score. If a participant had equal DQST scores bilaterally, the side with greater pain severity was enrolled.

Individuals were recruited through physiotherapy, hand therapy, and hand surgery clinics around Melbourne, Australia. Potential participants were screened via telephone prior to a physical screening for the DQST assessment (BM). For those meeting eligibility criteria, written informed consent was gained.

Interventions

The interventions were delivered in private hand therapy practices in Melbourne, Australia by four physiotherapists and two occupational therapists. The clinicians had a minimum of 5 years and a median 15 years of experience in managing wrist conditions. All participants were provided with education and an orthosis. Those still meeting DQST diagnostic criteria after 2 weeks were randomised into either high- or low-load exercise groups (Figure 1).

Figure 1.

Assessment and intervention protocol timeline; AROM = Active range of motion; MVIC = Maximum voluntary isometric contraction.

Education

At the initial appointment, all participants were provided an education booklet outlining the condition and activity modification advice (Supplementary appendix A). Treating clinicians additionally provided individualised advice throughout the study period, informed by the participant’s functional demands and aggravating activities (work, sport, and hobbies).

Orthosis

All participants were provided a custom-made, thermoplastic long opponens orthosis (Figure 2). Participants were advised to wear the orthosis as much as possible for 2 weeks, including overnight but excluding personal hygiene tasks. After 2 weeks, the treating clinician provided individualised advice about weaning from the orthosis, formulated on the bases of the participant’s pain, capacity to modify activities and clinician experience. Participants were encouraged to continue using their affected hand for function within the limits of the orthosis and comfort, with the aim of minimising potential negative effects of immobilisation.

Figure 2.

Thermoplastic long opponens orthosis. The orthosis was fabricated with the wrist in approximately 20° degrees extension, and the thumb able to touch the middle fingertip. To ensure no additional compression of the region, negative space was created between the thermoplastic and the radial styloid. The orthosis was worn as much as possible for 2 weeks then weaned at the treating clinician’s discretion, based on each individual participant’s pain and functional demands.

Exercise intervention

All participants were prescribed pain-free active range of motion exercises at the initial appointment. These were prescribed and progressed over the 6 week intervention at the discretion of the treating clinician.

After 2 weeks, the participants commenced a 4 week loading program consisting of isometric thumb extension exercises. Thumb extension was selected over abduction as it has been theorised that de Quervain’s syndrome predominantly affects extensor pollicis brevis.^37–40 Participants were randomised to perform the exercise at either 25% of their maximum voluntary isometric contraction (MVIC) (low-load group) or 70% MVIC (high-load group). The low- and high-load dosages were based on literature of other painful conditions.^28,29,41,42

Thumb extension MVIC was assessed by the treating clinician using a digital force meter (Model: EHS-631, Berkley®, IA, USA). Participants were positioned in neutral forearm rotation, approximately 20° degrees of wrist extension, and thumb in line with the radius. Participants were asked to extend their thumb with maximal effort for 5 seconds. This was repeated three times with a 90 second rest between each attempt. The highest reading recorded (the MVIC) was used to calculate the prescribed home exercise resistance. Participants attended weekly appointments for 4 weeks, and this procedure was repeated at each appointment (Figure 1). The prescribed exercise resistance was adjusted to reflect the most recent MVIC assessment, ensuring intensity consistent with their group allocation. Groups were matched for exercise-related time under load (20 second holds, five repetitions, once daily) and rest between repetitions (1 minute). An exercise device (Figure 3) including a spring gauge was designed for the purpose of this study to allow consistent application of load. The exercise doses and exercise device were pilot tested prior to commencement of the study. If the participant could not tolerate the prescribed exercise, the dosage was modified by reducing the hold time and increasing the repetitions. If this was not tolerated, participants were advised to rest from the exercise for 1 week before MVIC reassessment and recommencing isometric exercise. If the exercises were not tolerated again, the same procedure was repeated once more before a recommendation to cease the exercise intervention.

Figure 3.

Exercise device provided to participants for home exercise program. Participants were instructed to rest their forearm on a table with their hand over the edge and wrist in a neutral position. The spring gauge was lifted to the prescribed load using the unaffected hand. Participants placed the proximal phalanx of the affected thumb under the padded gauge, maintaining the position of the spring gauge and the thumb neutral. The unaffected hand moved to stabilise the base of the device. Participants held this position for 20 seconds, before lifting the gauge off the affected thumb using their unaffected hand.

Outcomes

The primary outcomes of this study were feasibility and safety. Feasibility was measured using recruitment rate, number of drop-outs, intervention adherence, and participant feedback.

Qualitative data (participant feedback) collected via semi-structured interviews assessed participant experience and satisfaction with the treatment and study procedures with the aim of informing feasibility assessment and future study design (Supplementary appendix B). The interview outline was determined a priori. Closed ended questions were included and coded into pre-specified categories, for example addressing willingness to participate, to allow simple descriptive analysis. The interviews were recorded and transcribed with the participant’s written informed consent.

Adherence to the interventions was recorded via a participant-completed diary. Safety was assessed through adverse events, such as persistent, increased discomfort related to the intervention (e.g. pain preventing exercise performance the following day). A serious adverse event was defined as a participant being unable to continue in the study due to pain with exercise, or a wound requiring dressing.

The secondary outcomes were participant perception of improvement, function, pain, grip strength, pinch strength, and range of motion. Participant perception of improvement was assessed using a global rating of change scale. This is an 11-point scale in which a score of five represents “complete recovery,” zero represents “no change,” and minus five represents “very much worse.” Pain and function were measured using the Quick Disabilities of the Arm, Shoulder and Hand questionnaire ⁴³ (QuickDASH) and Patient Rated Wrist and Hand Evaluation ⁴⁴ (PRWHE).

Functional strength was measured using a hand-held dynamometer (JAMAR® PLUS+, Paterson Medical®, IL, USA) in handle position two ⁴⁵ and a pinch strength gauge (JAMAR® PLUS + Digital Pinch Gauge, Performance Health Supply Inc., WI, USA). Participants were seated with shoulder adducted to their side, elbow at 90° degrees flexion, forearm in mid-prone and feet flat on the floor. Three attempts were made for each grip and three-point pinch, with the maximum strength measured used in analysis.

Active range of motion was assessed for three directions of movement. Ulnar deviation and thumb extension were measured using goniometry with the participant’s forearm pronated and resting on the table. The Kapandji Index ⁴⁶ was used as a measure of composite thumb flexion. Whilst typically used as a measure of opposition, a score of six or above is reliant on composite thumb flexion. ⁴⁶

Outcomes were measured at baseline (excluding global rating of change), at randomisation (2 weeks), at the end of the study period (6 weeks) and at a follow-up assessment (14 weeks).

Participants recorded medication taken for their wrist pain but were asked to abstain from receiving any other interventions during the study period. Interventions received during the follow-up period of the study were also recorded.

Randomisation

Participants were randomised following the 2 week orthosis period. Equal allocation ratio was used. Block randomisation was performed and stratified based on whether participants were within 12 months postpartum, as the postpartum population may present a different patient group based on potential hormone association. ⁴⁷ Randomisation was performed by drawing of unmarked, opaque envelopes from a larger envelope, completed by an author with no direct participant contact (SD). Treating clinicians were notified of allocation via a telephone call.

Blinding

Due to the nature of the intervention, the treating clinician could not be blinded. Participants were aware that they would be randomised into one of two similar exercise programs and were informed of their prescribed resistance to complete the home exercises. However, they were not informed of their group allocation. The outcomes of range of motion, grip and pinch strength were measured by a blinded assessor.

COVID-19 modifications

To minimise impact of missed appointments during COVID-19 restrictions, the trial protocol was modified. When a participant could not attend for treatment, the clinician consulted via telephone or telehealth. Due to an inability to measure thumb extension strength remotely, participants continued with the same resistance prescribed at the previous appointment. All other treatment was provided remotely. If participants could not attend assessment appointments, patient-reported outcome measures were mailed.

Further modifications included use of personal protective equipment and screening for COVID-19 symptoms or exposure. These modifications were determined by authors, based on La Trobe University policy and Victorian Department of Health and Human Services advice, and approved by the La Trobe University Ethics Committee (HEC18429).

Statistical analysis

Descriptive statistics were utilised to report demographic characteristics, expectation of improvement, and interventions received during the follow-up period. Primary outcomes of recruitment rate, adverse events, adherence, and drop-outs were also reported using descriptive statistics. Data were analysed on intention-to-treat principles. Continuous outcomes were analysed using a linear mixed model multivariate repeated measures analysis. This deviated from the registered protocol, a decision made to better reflect multiple assessments within the same subject, to cater for missing data ⁴⁸ and to reduce the risk of type I error. ⁴⁹ Each continuous outcome was modelled with dependent variables of time, sex, group allocation, and postpartum status. As no adjustment was made for multiple time comparisons for each variable, the interpretation of the results was conservative with a declared significance p <0.01. Participant perception of improvement scores were compared using an odds ratio. A sample size calculation was not required as the study was for feasibility purposes.

Results

Twenty-nine participants were recruited between June 2019 and April 2021 and follow-up completed by August 2021. Recruitment was suspended for 9 months due to COVID-19 restrictions. With consideration of this formal suspension, the recruitment rate was 2.1 participants/month. Recruitment was closed due to impacts of COVID-19 restrictions on participant recruitment and attendance. The researchers were blinded to group allocation and results when this decision to discontinue was made.

Twenty-eight participants were randomised, 11 were women within 12 months postpartum (Figure 4). There were no significant differences in baseline characteristics, except duration of symptoms (p = 0.006) (Table 1). No participants withdrew from treatment or the study. A total of three treatment sessions (two participants) were conducted remotely and four data collection points were missed (four participants), all related to COVID-19 restrictions or participant illness (Figure 4).

Figure 4.

Flowchart of participants through the study.

Table 1.

Participant baseline characteristics

	Total (n = 28)	Low-load group (n = 12)	High-load group (n = 16)
Age mean (SD)	40.1 (13.0)	44.6 (14.3)	36.6 (11.3)
Sex
Male n (%)	6 (21.4%)	3 (25.0%)	3 (18.8%)
Female n (%)	22 (78.6%)	9 (75.0%)	13 (81.3%)
Height (cm) mean (SD)	165.2 (7.7)	164.7 (9.2)	165.7 (6.6)
Weight (kg) mean (SD)	68.1 (11.6)	68.3 (13.6)	68.0 (10.2)
BMI mean (SD)	24.9 (3.7)	25.0 (3.4)	24.8 (4.0)
Bilateral presentation n (%)	7 (25%)	3 (25%)	4 (25%)
Side included
Dominant n (%)	10 (35.7%)	5 (41.7%)	5 (31.3%)
Non-dominant n (%)	18 (64.3%)	7 (58.3%)	11 (68.8%)
Women within 12 months postpartum n (%)	11 (42.9%)	5(41.7%)	6 (37.5%)
Number of episodes median [range]	1 [1–2]	1 [1–2]	1 [1–1]
Duration of symptoms (months) median [range]	2.5 [0.33–17]	3 [1.5–7]	2 [0.33–17]
Work type
Office based	16 (57.1%)	5 (41.7%)	11 (68.8%)
Home duties	6 (21.4%)	3 (25.0%)	3 (18.8%)
Manual	2 (7.1%)	2 (16.7%)	0 (0.0%)
Mixed	4 (14.3%)	2 (16.7%)	2 (12.5%)
Work hours/week	25.1 (17.2)	21.3 (18.2)	27.8 (16.6)
Work repetitiveness (VAS,/10)	7.1 (2.1)	7.9 (2.2)	6.6 (1.9)
Work heaviness (VAS,/10)	3.5 (2.5)	3.7 (2.5)	3.4 (2.6)
Upper limb related recreational hours/week	2.7 (3.1)	2.8 (3.2)	2.7 (3.2)
Previous treatments
Orthosis	18 (66.7%)	7 (63.6%)	11 (68.8%)
Corticosteroid injection	1 (3.7%)	0 (0.0%)	1 (6.3%)
Resistance exercise	4 (14.3%)	1 (8.3%)	3 (18.8%)
Oral NSAIDs	6 (22.2%)	3 (27.3%)	3 (18.8%)
Topical NSAIDs	3 (11.1%)	1 (9.1%)	2 (12.5%)
Expectation of improvement
Completely recover	8 (28.6%)	3 (25.0%)	5 (31.3%)
Much improve	19 (67.9%)	9 (75.0%)	10 (62.5%)
Slightly improve	1 (3.6%)	0 (0.0%)	1 (6.3%)
No change or worse	0 (0.0%)	0 (0.0%)	0 (0.0%)
DQST score at baseline median [range]
Baseline	6 [3–7]	6 [3–7]	6 [5–7]
At randomisation	6 [3–7]	6 [4–7]	6 [3–7]

BMI = body mass index; DQST = de Quervain’s screening tool; NSAID = nonsteroidal anti-inflammatory drug; VAS = visual analogue scale

Adverse events

One participant in the low-load group rested from the isometric thumb extension exercises for 2 days due to exercise-related pain. One participant reported pain that increased immediately following the exercise but despite this, they maintained prescribed dosage. There were four minor adverse events relating to orthoses (skin irritation, rubbing, tightness around the thumb), all before randomisation and all in participants subsequently allocated to the high-load group. There were no serious adverse events.

Adherence

Participants performed a mean of 86.7% (SD 16.11) of the prescribed isometric exercise sessions and this did not differ between groups (Supplementary appendix C). Reasons for missing exercise sessions included “forgetting” (n = 12), “too busy” (n = 5), forgetting the exercise device if away from home (n = 3), illness (n = 2), pain associated with isometric exercise (n = 1), and pain not associated with the exercise (n = 2). Adherence to range of motion exercises and time wearing the orthosis was similar between the groups (Supplementary appendix C).

Participant interviews

Of 25 participants who consented to be interviewed, 21 participants said they would participate again and two were unsure. Two participants, both in the low-load group, said they would not participate again and cited wanting a faster recovery. During the follow-up period, one of these participants went on to surgery and the other to corticosteroid injection. All but two participants reported they would recommend the study interventions to others with de Quervain’s syndrome. Two of the participants who underwent a corticosteroid injection within the follow-up period advocated for a trial of therapy prior to proceeding to more invasive interventions.

Fifteen participants correctly identified their group allocation, eight were unsure, and two guessed incorrectly (both in the high-load group). Group preference varied; 11 participants would have chosen to be in the high-load group, with many citing their perception that working harder would lead to improved results. Six participants believed that being in the low-load group was better, citing a preference for a gentle introduction to exercise, and two would have preferred starting with low-load and building to a higher load. Six participants had no preference. Barriers to participation included pain with exercise, attending appointments (finding time, travel, parking), wearing the thermoplastic orthosis, being consistent with exercise, and the duration of time to recover.

When comparing the study interventions to other available interventions, such as surgery and corticosteroid injections, some participants expressed a preference to try conservative management initially. Some reasons cited were a desire to avoid surgery (n = 5) and wanting to avoid or being frightened of corticosteroid injections (n = 5). There was a preference expressed for the concept escalation of care, starting with most conservative interventions and only progressing to more invasive interventions if adequate improvement was not achieved (n = 8).

Secondary outcomes

There were statistically and clinically significant improvements in pain and function as measured by the QuickDASH and PRWHE. The change in mean QuickDASH score was statistically significant at each timepoint (week 2 p = 0.001; weeks 6 and 14 p < 0.001) and had surpassed the minimal clinically important difference (MCID) of 16-points ⁵⁰ by week 6 (Table 2). The change in mean PRWHE was also statistically significant at each timepoint (p’s < 0.001) and achieved the MCID of 17-points ⁵⁰ by week 2 (Table 2). There were no statistically significant improvements in grip strength (p’s = 0.154–0.911), pinch strength (p’s = 0.017–0.100) or range of motion (ulnar deviation p’s = 0.11 to 0.805; thumb extension p’s = 0.208–0.805) at any timepoint; and no significant differences between high- and low-load groups (Table 2). Baseline Kapandji Index scores for thumb flexion were high, with 89.2% of participants scoring nine or 10/10 (Supplementary appendix D). Given this high baseline measure, a meaningful change was unlikely and hence data were not statistically analysed.

Table 2.

Secondary outcomes

		Low-load group n = 12 mean (95% CI)	Low-load group within-group difference*mean (95% CI)	High-load group n = 16 mean (95% CI)	High-load group within-group difference*mean (95% CI)	Between-group difference mean (95%CI)
QuickDASH	Baseline	45.3 (34.5 to 56.2)		47.8 (38.0 to 57.6)
Possible score range: 1–100; a higher score indicates greater disability and symptoms	2 weeks	31.3 (20.5 to 42.1)	14.0 (2.0 to 26.0)	39.4 (29.6 to 49.2)	8.4 (−0.9 to 17.7)
	6 weeks	22.0 (11.0 to 33.0)	23.3 (7.4 to 39.2)	16.8 (7.0 to 26.6)	31.0 (17.1 to 44.8)
	14 weeks	14.3 (2.9 to 25.7)	31.0 (12.6 to 49.4)	14.3 (4.4 to 24.1)	33.5 (18.6 to 48.5)
						−1.3 (−12.0 to 9.3)
PRWHE	Baseline	56.0 (45.4 to 66.6)		57.0 (47.4 to 66.6)
Possible score range: 0–100; a higher score indicates greater pain and disability	2 weeks	34.3 (23.8 to 44.9)	21.7 (9.9 to 33.4)	42.7 (33.1 to 52.3)	14.3 (5.2 to 23.4)
	6 weeks	20.3 (9.6 to 31.0)	35.7 (19.7 to 51.6)	21.3 (11.7 to 30.9)	35.7 (22.2 to 49.2)
	14 weeks	17.1 (6.0 to 28.2)	38.9 (21.4 to 56.4)	18.6 (9.0 to 28.3)	38.4 (23.7 to 52.9)
						−3.0 (−13.4 to 7.4)
Grip strength (kg)	Baseline	31.4 (28.1 to 34.6)		31.9 (29.0 to 34.9)
	2 weeks	31.0 (27.8 to 34.2)	0.4 (−2.8–3.5)	32.0 (29.1 to 35.0)	−0.1 (−2.6 to 2.4)
	6 weeks	34.0 (30.7 to 37.3)	−2.6 (−7.5 to 2.2)	33.4 (31.5 to 37.5)	−2.5 (−6.4 to 1.3)
	14 weeks	32.1 (28.7 to 35.5)	−0.7 (−5.4 to 4.0)	35.6 (32.7 to 38.6)	−3.7 (−8.2 to 0.8)
						−1.4 (−4.6 to 1.8)
Pinch strength (kg)	Baseline	6.2 (5.4 to 7.0)		6.3 (5.6 to 7.0)
	2 weeks	6.7 (6.0 to 7.5)	−0.5 (−1.4 to 0.3)	6.8 (6.1 to 7.5)	−0.5 (−1.3 to 0.3)
	6 weeks	6.9 (6.1 to 7.7)	−0.7 (−1.8 to 0.4)	6.9 (6.2 to 7.6)	−0.6 (−1.6 to 0.4)
	14 weeks	7.2 (6.4 to 8.0)	−1.0 (−2.3 to 0.3)	7.2 (6.5 to 7.9)	−0.9 (−2.0 to 0.2)
						−0.03 (−0.8 to 0.7)
Ulnar deviation range (°)	Baseline	30.3 (24.7 to 35.9)		27.5 (22.5 to 32.6)
	2 weeks	31.2 (25.6 to 36.8)	−0.9 (−6.3 to 4.5)	30.7 (25.7 to 35.8)	−3.2 (−8.3 to 2.0)
	6 weeks	32.5 (26.8 to 38.1)	−2.2 (−9.8 to 5.5)	33.7 (28.5 to 38.8)	−6.1 (−13.2 to 1.0)
	14 weeks	35.9 (30.0 to 41.8)	−5.6 (−15.1 to 3.9)	36.7 (31.6 to 41.7)	−9.1 (−17.0 to −1.2)
						0.3 (−5.2 to 5.8)
Thumb abduction range (°)	Baseline	43.6 (39.1 to 48.2)		43.0 (38.9 to 47.2)
	2 weeks	45.6 (41.1 to 50.2)	−2.0 (−7.1 to 3.1)	42.2 (38.1 to 46.3)	0.8 (−2.7–4.3)
	6 weeks	45.2 (40.6 to 49.8)	−1.6 (−7.4 to 4.3)	47.8 (43.6 to 51.9)	−4.7 (−10.5 to 1.1)
	14 weeks	48.3 (43.5 to 53.1)	−4.7 (−12.4 to 3.1)	46.3 (42.2 to 50.4)	−3.3 (−9.1 to 2.6)
						0.9 (−3.6 to 5.4)

*Difference from baseline; CI = confidence interval; PRWHE = Patient Rated Wrist and Hand Evaluation; QuickDASH = Quick Disabilities of the Arm, Shoulder, and Hand questionnaire

Group allocation did not affect global rating of change at 6 weeks (odds ratio (95% CI) = 0.82 (0.20–3.37), p-value = 0.784) with 41.7% and 43.8% of participants rated themselves as very much improved in the low-load and high-load groups respectively. By 14 weeks, 50% of the low-load and 62.6% of the high-load group rated themselves as very much improved or completely recovered (Figure 5). With an MCID of 2-points, ⁵¹ 93% of participants achieved a clinically important improvement at 6 weeks. One participant (low-load group) rated themselves 2-points worse. At the 14 week assessment, 85% of participants rated themselves improved by at least 2-points and one participant rated themselves 2-points worse (this high-load group participant had received a corticosteroid injection in the week prior to the follow-up assessment).

Figure 5.

Global rating of change scores. Participants rated their degree of change from baseline at each assessment timepoint. Set up as colour figure. Can be amended to black and white.

Self-reported medication-use and additional treatments during the follow-up period are outlined in Supplementary appendices E and F.

Discussion

Isometric thumb extension exercises at either high- or low-loads were safe, well adhered to, and considered acceptable by most participants. No significant difference was found between high- and low-load isometric exercises in any of the secondary outcomes. Both groups demonstrated statistically significant and clinically important improvements in pain and function at 6 and 14 weeks as measured by the QuickDASH and PRWHE, and 93% exceeded the MCID for improvements in global rating of change.

The exercise adherence rate (>80%) can be considered satisfactory. ⁵² This adherence, the absence of drop-outs, low number of adverse events, and willingness to participate again indicate that isometric exercise within a multimodal approach is feasible for future research. We cannot compare the exercise adherence rate to other de Quervain’s syndrome studies as it has not previously been reported.^21,22,25 However, the exercise adherence rate in this study is comparable to previous studies of other tendon-related conditions. ⁵³ There was no significant between-group differences for these outcomes, suggesting that isometric exercise is likely to be acceptable, safe, and feasible in a clinical setting regardless of load intensity.

Repeating this comparative study (high-load vs low-load) as a single centre RCT is likely unfeasible. A post hoc sample size calculation based on the QuickDASH found 151 participants would be required in each group (assuming a standard deviation of 20, alpha 0.05 and power of 0.80). To perform a full scale RCT comparing high- and low-load isometric exercises, a large multi-centred study would be needed to accommodate the required sample size and the low recruitment rate.

It has been hypothesised that high-load may have a greater effect on pain than low-load isometric exercises in musculoskeletal conditions. ³² The effect of isometric thumb extension exercise and whether there is an optimal load in people with de Quervain’s syndrome remains unknown due to the small sample size and the multimodal nature of the intervention. However, detecting a difference in patient-reported outcomes was not the primary aim of this trial. There are the potential of an undetected difference between high- and low-load exercises, that neither high- nor low-load isometric exercise added benefit to the multimodal intervention, or that both exercise intensities were equally beneficial in this study. Given both load intensity exercises were tolerated and adhered to, future research could investigate isometric thumb extension exercise at any intensity.

Clinically important and statistically significant improvements in pain and function in both groups at 6 and 14 weeks indicate that further research exploring isometric exercise as part of a multimodal hand therapy intervention may be warranted. It cannot be ruled out that isometric exercises may have added benefit to the multimodal intervention through the potential effect of exercise-induced hypoalgesia.^31,32 This effect potentially contributed positively to participants’ attitudes and response to intervention through buy-in to the intervention and aiding participation in function. ²⁸ Alternatively, exercise of any kind may have had a positive effect through a potential reduction in fear avoidance behaviour, ⁵⁴ increased self-efficacy ⁵⁵ and confidence in a gradual return to functional tasks. If this is the case, exercise selection should consider the type of exercise least likely to aggravate pain. Isometric exercises, as opposed to eccentric or isotonic exercises, limit friction associated with tendon glide and limit compression associated with changes in wrist and thumb postures, making it a more appropriate early exercise in people with de Quervain’s syndrome. Future studies comparing multimodal interventions with or without exercise components will provide useful evidence as to whether the addition of exercise results in improved patient outcomes.

Further investigation of multimodal interventions compared to other intervention options may also be warranted. The outcomes from this study are promising given that orthoses alone have not demonstrated clinically important improvements in 6 weeks ⁵⁶ or symptom resolution in a postpartum population. ⁵⁷ Education and individualised advice may be a key component, as has been demonstrated in other musculoskeletal conditions. ⁵⁸ A multimodal intervention is more reflective of clinical practice and may demonstrate greater benefit than intervention components delivered in isolation. ⁵⁹ Comparison with a “wait and see” approach should be considered as de Quervain’s syndrome is believed to be a self-limiting condition that resolves over 6–18 months. ⁶⁰ Multimodal interventions have the potential to improve the quality of care in this condition by improving patient outcomes and avoiding invasive and costly interventions.

Corticosteroid injection is a commonly implemented intervention. While previous studies suggest that corticosteroid injection may be effective in reducing de Quervain’s syndrome symptoms, ⁶¹ recurrence rates as high as 48% have been reported. ⁶² The literature, supported by the findings in this study, suggests that a proportion of people with de Quervain’s syndrome elect to trial conservative interventions before progressing to more invasive interventions of injection or surgery. ⁶⁰ Exercise in combination with a multimodal intervention may be an acceptable and efficacious option for this population. The findings of this study support the notion that further investigation of multimodal interventions with or without exercise is warranted to facilitate evidence-based care and support patient preferences.

There are several limitations to this study. The ability to determine recruitment feasibility was impacted by COVID-19 restrictions and the decision process regarding the feasibility of future studies not being determined a priori. The study was underpowered for secondary outcomes due to the small sample size. The contribution of exercise and the other components of the multimodal intervention to the overall treatment effect could not be assessed. As such, it is not possible to conclude the efficacy or effectiveness of exercise in this population. Treating clinicians could not be blinded due to their role in the exercise prescription. Fifteen participants correctly guessed their group allocation. This increases the likelihood of performance and detection bias, particularly relevant to the subjective outcome measures, which is a common issue in exercise and physical rehabilitation research. ⁶³ However, as both groups received active interventions, the extent of bias may be less than in studies involving a sham intervention arm. Whilst the study participants were reflective of those seen in allied health settings, those who agreed to participate were potentially biased towards the types of interventions provided in the study or against other interventions such as corticosteroid injections or surgery. Thus, their opinions cannot be generalised as applicable to all people with de Quervain’s syndrome.

Conclusions

The outcomes of this feasibility trial suggest that isometric thumb extension exercises as part of a multimodal intervention are safe and feasible for people with de Quervain’s syndrome. To repeat a comparison of high- and low-load as a full scale RCT, a large multi-centre trial would be required. Based on participant satisfaction and the clinically important improvements in pain and function observed in the study, this multimodal intervention shows promise and further research investigating conservative treatment techniques, including isometric exercise and multimodal interventions, for people with de Quervain’s syndrome is warranted.

ORCID iD

Brodwen McBain https://orcid.org/0000-0001-9594-0713