Article Text

Extended report
Effect of home-based hand exercises in women with hand osteoarthritis: a randomised controlled trial
  1. Toril Hennig1,
  2. Liv Hæhre1,
  3. Vivian Tryving Hornburg1,
  4. Petter Mowinckel2,
  5. Ellen Sauar Norli1,
  6. Ingvild Kjeken2
  1. 1Martina Hansens Hospital, Sandvika, Norway
  2. 2National Advisory Unit on Rehabilitation in Rheumatology, Department of Rheumatology, Diakonhjemmet Hospital, Oslo, Norway
  1. Correspondence to Ingvild Kjeken, National Advisory Unit on Rehabilitation in Rheumatology, Department of Rheumatology, Diakonhjemmet Hospital, PO Box 23, Vinderen, Oslo NO-0319, Norway; ingvild.kjeken{at}


Background Hand exercises are recommended for patients with hand osteoarthritis (HOA), though evidence for their effect is conflicting.

Objective To evaluate, in a randomised controlled trial, the effect of HOA information plus home-based hand exercises (exercise group) compared with information only (control group) in women with HOA.

Methods Interventions were delivered by two occupational therapists. Exercise group participants received eight follow-up calls over the 3-month study and recorded adherence, pain after exercises and adverse events in a diary. Primary outcome was activity performance measured after 3 months by the Patient-Specific Functional Scale (PSFS), with a range of 0–10. Secondary outcomes were measurements of hand function, disease activity, symptoms and number of responders to treatment according to the OMERACT-OARSI criteria.

Results Of 80 women randomised (40 : 40) (mean age (SD) 60.8 years (7.0)), follow-up was 89% (n=71). An intention-to-treat analysis was performed. The adjusted mean difference for the exercise versus control group was 1.4 points (95% CI 0.6 to 2.2, effect size 1.0) for the PSFS score. Thirteen patients in the exercise group versus three participants in the control group reached a positive minimal clinical important difference of 2.2 points in the PSFS total score, while none versus two, respectively, had a negative change (p=0.007). For secondary outcomes, significant mean differences were found in grip strength and thumb web space, in fatigue, joint pain and the Functional Index for HOA activity performance scores. Sixteen exercise-group participants fulfilled the OMERACT-OARSI response criteria versus two control-group participants (p<0.001).

Conclusions Hand exercises were well tolerated and significantly improved activity performance, grip strength, pain and fatigue in women with HOA.

Trial registration number: ISRTCN79019063.

  • Hand Osteoarthritis
  • Occupational Therapy
  • Rehabilitation

Statistics from

Request Permissions

If you wish to reuse any or all of this article please use the link below which will take you to the Copyright Clearance Center’s RightsLink service. You will be able to get a quick price and instant permission to reuse the content in many different ways.


Hand osteoarthritis (HOA) has a high and increasing prevalence and is a great burden both to the individual and society. Its prevalence is difficult to determine owing to differences in study populations and diagnostic methods. In the Framingham Osteoarthritis Study, 16% of women and 8% of men aged 28–92 years had symptomatic HOA (both radiographic changes and patient-reported symptoms).1

Clinical manifestations of HOA are soft-tissue swelling and bony enlargements, most frequently in the distal and proximal interphalangeal finger joints and in the carpometacarpal (CMC1) joint of the thumb.2 CMC1-OA often presents with a combination of reduced cartilage thickness, increased ligament laxity with resultant instability and subluxation of the base of the metacarpal on the trapezium,3 ,4 which in turn results in decreased thumb web space.5 The prevalence of CMC1-OA has been estimated at 13% in people aged 41–50 years, increasing to 68% in people between 71 and 80 years.6 Functional consequences of HOA are pain, reduced mobility and grip strength, activity limitations and participation restrictions, including loss of income and increased dependency.7–12 However, it is debateable whether there are differences in pain and disability between people with and without CMC1-OA.13–15

There is no cure for hand OA. Pharmacological treatment is still confined to symptomatic treatment, while surgical treatment is primarily limited to severe CMC1-OA.

The European League Against Rheumatism (EULAR) recommends that all patients with HOA should receive an exercise regimen including both range of motion (ROM) and strengthening exercises, but emphasises that robust evidence for the effect of exercises is lacking.16

Both the EULAR recommendations and six systematic reviews have summed up the results of non-pharmacological interventions for HOA.16–22 They conclude that evidence for the effect of hand exercises is controversial and that robust randomised controlled trials (RCTs) are needed.

The aim of this study, therefore, was to evaluate the effect of hand exercises in women with HOA.


Study design

In this RCT, information plus home-based hand exercises (exercise group) were compared with information only (control group). The study was conducted at the outpatient clinic at Martina Hansens Hospital, Bærum, Norway. Participants were assessed before randomisation and start of treatment and at follow-up after 3 months (main end point).


Between February 2011 and December 2012, patients were consecutively screened for eligibility by two occupational therapists (OTs). The inclusion criteria were female gender, HOA diagnosed by rheumatologists or orthopaedic surgeons according to the American College of Rheumatology criteria,23 age between 18 and 80 years, stable medication over the past 3 months, a minimum of three self-reported HOA-related activity limitations identified by patients in the Patient-Specific Functional Scale (PSFS)24 and ability to communicate in Norwegian. Exclusion criteria were hand surgery within the past 6 months, steroid injections within the past 2 weeks, impaired hand function due to trauma or diseases other than HOA and cognitive or mental impairment. Participants who underwent hand surgery or received steroid injections during the trial period were excluded at the 3-month control.


Two OTs with long experience in HOA treatment delivered the interventions (box 1). A review of exercise programmes in HOA showed that there is no consensus about the design of such programmes.19 The exercise programme was therefore developed by the Norwegian Network for OTs in rheumatology, following an evidence-based methodology to reach consensus for the selection of exercises in an HOA programme.25 However, the American College of Sports Medicine's recommendations for developing muscular strength and flexibility in frail elderly people were used to determine exercise intensity, session frequency and length of the exercise period,26 as no agreement was reached about these elements in the OT network.

Box 1

Interventions delivered to participants

Leaflet and advice

All participants received a leaflet containing the following information about hand osteoarthritis (HOA) and ergonomic principles:

Typical symptoms of HOA are pain, stiffness and reduced grip strength. As a consequence, many also experience problems in performing daily activities.

Based on research, people with HOA are recommended to continue with their daily activities within the limits of their level of pain and energy. To use and move your joints will enhance circulation of the synovial fluid, which in turn will nurture the cartilage in the joints. To use your hands in a wide variety of daily activities is therefore important to maintain joint mobility and hand strength.

Rather than stopping certain activities, adjust the way they are performed. If you like walking or hiking, try choosing a less challenging terrain, use walking sticks, take more breaks and/or walk a shorter distance.

Housework and gardening can be divided into several short periods and you could consider lowering your standards a little.

Many people adjust their working methods. Some suggestions for alternative working methods are:

  • Use both hands: The force is then distributed between both hands over more joints. Carrying a frying pan or milk carton with both hands is an example of an ergonomic working method.

  • Transfer weight from smaller to larger joints: Use a rucksack or a trolley bag when carrying is an example of a strategy that decreases the load on the smaller finger joints.

  • Use lighter equipment: A light frying pan may be easier to handle and carry than a heavy iron pan.

  • Use alternative hand grips: Choose tools with ergonomic or larger grips. If one joint hurts, try to find a grip in which the strain is moved to other joints. There is also a lot of variation in how tightly people grip a pen when writing. Try to find a grip where you can write without using more force than necessary to hold the pen.

  • Balance activity and rest: Muscles that are tired are less capable of supporting and protecting osteoarthritic joints. Taking regular breaks is important to ensure variation in strain and movements. Frequent breaks may also prevent pain.

Hand exercises

Participants in the exercise group additionally received a home-based hand exercise programme aimed at maximising the stable and pain-free functional range of motion of the finger joints, increasing grip strength, maintaining joint stability and preventing or delaying development of fixed deformities. A rubber ball made of polyethylene with a diameter of 7 cm was used to provide resistance in the grip strengthening exercise, while rubber bands were used to provide resistance to the thumb abduction/extension exercise.

Participants were instructed to perform three exercise sessions a week, with each exercise to be performed with 10 repetitions during the first 2 weeks, increasing to 12 repetitions over the next 2 weeks and if possible, to 15 repetitions for the rest of the 3-month exercise period. See also online supplementary appendix I for the exercise programme with illustrations.


Appointments were made for weekly occupational therapist (OT) follow-up calls in the first month and every second week for the rest of the 3-month study period. During the calls, the OT answered questions and discussed adherence and adjustment to the programme.

It is suggested that strengthening the thumb extensors and abductors is important to maintain the thumb web space.4 ,5 ,27–29 Furthermore, excessive exercises to reach full flexion and opposition of the thumb, as well as improving pinch or key grip strength, may result in increased subluxation and pain in unstable joints.4 ,5 ,28 ,29

The aims of the programme were therefore to maximise a stable and pain-free functional ROM of the finger joints, increase grip strength, maintain joint stability and prevent or delay the development of fixed deformities.27 ,29 ,30 Adverse events, adherence to, and pain after, exercise were recorded by participants in a diary (table 1).

Table 1

Exercise diary

Both groups received medical treatment as usual. After the 3-month assessments, participants in the control group received exercises and all participants received information about assistive devices and thumb base splints if needed.

Primary outcome

The primary outcome was change in activity performance assessed by the PSFS at 3 months (see online supplementary appendix II).24 Patients identified up to five activity limitations caused by their HOA and rated their performance of each activity on an 11-point Numeric Rating Scale (NRS) (0–10, 0=unable to perform activity). A total score was computed as the mean of the scores for the listed activities.

Secondary outcomes

Secondary outcomes comprised symptoms, disease severity, function and number of participants fulfilling the OARSI-OMERACT criteria.

Fatigue, pain, stiffness and patient global assessment of disease activity were recorded on a NRS (0–10, 0=no activity or symptoms). Additionally, exercise group participants recorded pain on a NRS after each exercise session.

Maximum grip strength was measured in Newtons by the Grippit electronic instrument, according to published testing procedures.31 Recording of right and left hand pain was standardised by asking the participant to mark the level of pain after measurements of grip strength on a NRS.

Joint mobility of digits 2–5 was recorded as flexion deficit in millimetres for each finger and summarised and computed as one variable for each hand, while thumb joint mobility was recorded as opposition deficit.32

The grip size instrument (12 transparent Plexiglas cylinders with a diameter from 1 cm to 12 cm) was used to assess thumb web space. Participants were asked to grip around one cylinder at a time and the largest size for which the therapist could see full contact between the cylinder and the total arch of the participant's thumb and second digit was recorded.

The Functional Index for HOA (FIHOA) was used as a patient-reported outcome of activity performance.33 ,34 FIHOA consists of 10 items with a four-point Likert scale: (0=possible without difficulty). A summary score ranging from 0 to 30 was calculated (0=good performance).

The OARSI-OMERACT responder criteria is a composite index that presents the results of changes after treatment in the three domains of pain, function and patient's global assessment as a single variable (responder yes/no).35

Evaluation at baseline also included demographic data, symptoms and disease duration, comorbidity, hand dominance and use of medication. All finger joints in both hands were examined by a trained OT for the presence of pain on pressure, while interphalangeal joints were examined for bony enlargement (1: present, 0: absent). Joint counts were computed as a summary of manifestations for each hand (range 0–15 for pain and 0–9 for bony enlargements). Nodal HOA was defined as one or more joints with nodules.

Conventional radiography of both hands was graded by an experienced rheumatologist, who estimated the degree of CMC1-OA according to the Kellgren–Lawrence method, with grades from 0 to 4 (grade 4 indicating large osteophytes, severe sclerosis and narrowing joint space).36 Furthermore, the presence of erosive OA (one or more joints with erosion=erosive OA) was examined.37

Sample size

We used the baseline scores of PSFS in a study of patients with knee dysfunction for sample calculation.38 Based on this study, the mean (SD) PSFS-total score was set as 3.1 (1.8). The minimal clinical important difference in PSFS has been estimated as 2.2 points.39 We expected a 20% loss to follow-up after 3 months and found that a sample size of 25 patients for each group was required to detect a difference of 2.2 points between groups with a significance level of 0.05 and a power of 80%. Owing to uncertainty related to the estimates of PSFS mean and SD, which might be different in HOA, we decided to include 40 participants in each group.


A statistician not involved in the study made a computer-generated randomisation list with a block size of 10. Concealed, opaque envelopes were used to allocate patients to the exercise or control group. The envelopes were locked away and opened by the patient after baseline assessments and information about HOA were completed. In this trial, patients and therapists delivering the intervention were aware of the treatment assigned. To achieve observer blinding, patients were asked not to inform the assessor about their group allocation in the 3-month assessments. Moreover, the statistician who performed the main statistical analyses was blinded to group allocation.

Statistical analysis

Analysis was performed on an intention-to-treat basis. Treatment effects (mean differences in outcomes between the two groups after 3 months) were estimated by an analysis of covariance, using the scores at 3 months as the dependent variable and baseline values and group as covariates.40 Furthermore, treatment effect sizes (Cohen's d) were computed as the adjusted between-group difference in scores divided by the pooled SD of the baseline scores for this outcome.41

Additionally, differences in number of participants in each group, with a change of ±2.2 or more in PSFS-change score or classified as OMERACT-OARSI responders after 3 months, were examined by χ2 test. Within-group differences were examined using paired sample t tests. The analyses were performed using IBM SPSS (International Business Machines SPSS V.21) and the Statistical Analysis System (SAS, V.9.3, SAS Institute Inc, Chapel Hill, North Carolina, USA).


Study participants

Of 167 consecutive screened patients, 80 met the inclusion criteria, agreed to participate and were randomly assigned to the exercise (n=40) or control group (n=40) (figure 1).

Figure 1

Study flow diagram. HOA, hand osteoarthritis.

Patients in the two groups did not differ in baseline characteristics, with the exception that a significantly larger number of participants in the exercise group had nodal HOA (p=0.03) (table 2). Use of medication at baseline was similar in both groups and none of the participants received any steroid injection during the study period. However, one participant in each group had received a steroid injection 3 weeks before inclusion. Thirty-seven participants in the exercise group and 35 in the control group completed all follow-up assessments. One participant in the control group withdrew from the trial after completion, leaving a total of 37 and 34 for the final analyses, respectively.

Table 2

Baseline characteristics in 80 patients with hand osteoarthritis (HOA) randomised to both information and hand exercises (exercise group) or information only (control group), reported as number and proportions, mean values and SDs, or median and range


The mean time used to deliver the exercise programme was 1.5 h, including assessment (20 min), instruction (30 min) and follow-up calls (8×5 min). The participants in the exercise group reported a median number (min, max) of exercise sessions of 37 (26, 43). The median time used for each session was 23.6 min (10.4, 42.0).

Mean pain (SD) after exercises in the first week of the study period was 5.6 (2.2), which decreased to 4.2 (2.4) in the last week, yielding a mean difference in pain between weeks 1 and 13 of 1.4 (95% CI 0.5 to 2.3), p=0.003 (figure 2).

Figure 2

Mean pain per week in the intervention group over the study period, measured on a scale from 0 to 10 (0=no pain) and presented as error bars with mean values and 95% CI.

However, one patient withdrew from the trial after 9 weeks owing to high and sustained pain.


Primary outcome

After 3 months, there was a significant adjusted mean difference in favour of the exercise group in the PSFS total score of 1.4 points (95% CI 0.6 to 2.2, p<0.001) (effect size 1.0) (table 3). The number of participants who reached a clinically relevant positive change, no change or a clinically relevant negative change of 2.2 in the PSFS total score were 13, 24 and 0 in the exercise group versus 2, 30 and 2 in the control group (p=0.007).

Table 3

Effect of hand exercises in 80 women with hand osteoarthritis (HOA) randomised to receive information plus hand exercises (exercise group) or information only (control group)

Secondary outcomes

Significant differences between the two groups in favour of the exercise group were found in joint pain (−1.1, p=0.02), grip strength in right (53.5 N, p<0.001) and left (44.6 N, p<0.001) hand, in thumb web space of the right (0.6, p=0.018) and left (0.7, p=0.007) hand, fatigue (mean adjusted differences in change between groups −1.1, p=0.05) and in the FIHOA activity performance score (−3.2, p=0.001).

Sixteen participants in the exercise group fulfilled the OMERACT-OARSI response criteria versus two in the control group (p<0.001).

Significant improvements of within-group differences were found in the exercise group for joint pain, stiffness and FIHOA scores, as well as in right and left hand measurements of grip strength, pain and thumb web space, while measurements of ROM remained stable.

In the control group, there was a general trend towards deterioration, with a significant increase in flexion deficit in the right hand and an aggravation of the FIHOA score.


The results in this trial demonstrate that hand exercises significantly improved activity performance, fatigue and joint pain, right and left hand grip strength and thumb web space in women with HOA. Additionally, there was a trend towards deterioration in the control group, with a significant increase in flexion deficit in the right hand. These findings are in line with the EULAR recommendations that patients with HOA should receive an exercise regimen that includes both ROM and strengthening exercises,16 as ROM exercises seem to be important in helping to maintain joint mobility and thumb web space, while strengthening exercises maintain joint stability and increase grip strength.

The difference between groups in the primary outcome activity performance was 1.4 points in the PSFS. After we started our study, a minimal clinical important difference of 1.2 PSFS points was estimated in patients with upper extremity problems.42 Together with the significant effects on pain and activity as measured by the FIHOA, these results show that hand exercises lead to clinically important functional improvements.

Sixteen participants in the exercise group were classified as OMERACT-OARSI responders after 3 months compared with two in the control group, and the intervention was delivered in a short time at low cost. This cheap and effective programme should therefore be considered for inclusion in the standard care for people with HOA.

Previous reviews of hand exercises have reported low or modest effects.16–22 In two recent RCTs, the authors found no effectiveness after 3 months of a group-based multidisciplinary treatment programme that included daily home exercises,43 or of hand exercises versus no hand exercises.44 However, none of these interventions included follow-up calls and the rationale for the design of the exercise programmes is unclear. The selection and monitoring of exercises in our programme is in accordance with recent research and recommendations.45–47 The distinct and positive effectiveness of the programme may therefore be attributed to the combination of an evidence-based programme and the close follow-up during the study period.

Limitations in daily activities have been identified as the major determinant of reduced health-related quality of life of patients with HOA.48 Several studies have shown significant correlations between activity limitations and poor grip strength,1 ,9 ,15 ,49 ,50 and reduced grip strength is described as an important functional consequence by people with HOA.7 There was a considerable increase in grip strength in the exercise group in our study, which might be one factor explaining the significant effect on activity. However, since the relationship between impairment and activity limitations is complex, the process of improvement should be investigated in future studies.

It is often stated that exercises for people with HOA should be of low intensity, performed during periods with little pain and inflammation and not cause pain that persists for more than 2 h after exercising.4 ,27 ,29 Nevertheless, it has been argued that for people with OA, a pain score of 5 on a scale from 0 to 10 is acceptable immediately after training and that pain should be back to its initial level after 24 h.51 There is also a considerable amount of research which demonstrates that intensive exercising was well tolerated and had positive effects on pain and function in knee OA,52 ,53 and in rheumatoid arthritis with respect to hand function,54 disease activity and radiological damage to the hands and feet.55 The results from our study show that even if participants reported considerable pain after exercising, this pain decreased significantly during the study period, and general joint pain improved. The high adherence to the programme also indicates that the exercise programme was safe and well tolerated.

In a call for new strategies within HOA research, it is argued that the concept of function should be broadened to include measures of the activities most relevant to patients.56 Hence, we used the individualised PSFS as a primary outcome. Recently, the validity and reliability of the PSFS in CMC1-OA was confirmed in a Swedish study,57 while the Norwegian version displayed good results in patients with musculoskeletal disorders.58 The mean and SD of PSFS scores in our sample was also in line with the estimates used in the power calculation. The study thus had sufficient power to detect clinically relevant differences produced by hand exercising.

The trial has methodological and clinical limitations. First, it was impossible to design a placebo-controlled trial, since there are no convincing placebos or sham exercise programmes, and therapists delivering the intervention could not be blinded.

However, participants were instructed not to disclose the group allocation to the assessor at the 3-month follow-up. Additionally, most outcomes were collected using either standardised instruments or were patient-reported using validated measures, thereby reducing the chance of results being affected by observer bias.

Second, monitoring drugs during the study period would have allowed us to compare the groups for possible differences in use of medication during the trial period.

Third, the instrument we used to measure thumb web space has not yet been tested for validity and reliability. Since then, the Pollexograph has been launched and tested and would be preferred in future studies.59

Lastly, because most patients with HOA in rheumatology practice are women, we decided to include only women in our study.48 Participants recruited were receiving specialist healthcare and 50% had one or more erosive interphalangeal joint. As a result, the sample is not representative of all patients with HOA. The effectiveness of the programme in men and in primary care should therefore be examined. Moreover, an examination of the long-term effectiveness is warranted.

In summary, an evidence-based, low-cost hand-exercise programme was well tolerated and significantly improved activity performance, grip strength, fatigue and pain in women with HOA.


We thank Professor Kåre Birger Hagen for reading and commenting on the manuscript and Martina Hansens Hospital and the Norwegian Associations for Rheumatism, Hand Therapy and Occupational Therapy for financial support.


Supplementary materials

  • Supplementary Data

    This web only file has been produced by the BMJ Publishing Group from an electronic file supplied by the author(s) and has not been edited for content.

    Files in this Data Supplement:


  • Handling editor Francis Berenbaum

  • Contributors The corresponding author confirms that all authors made substantial contributions to the study conception and design or analysis and interpretation of data, were involved in drafting the manuscript and approved the final version.

  • Funding Martina Hansens Hospital, Norway, the Norwegian Association for Rheumatism, the Norwegian Association of Hand Therapists and the Norwegian Association for Occupational Therapy.

  • Competing interests None.

  • Patient consent Obtained.

  • Ethics approval The regional committee for medical research ethics.

  • Provenance and peer review Not commissioned; externally peer reviewed.

  • Data sharing statement The data are stored on a secured research server at Martina Hansens Hospital, to which all authors have access. Applications for access to anonymised data from the dataset will be reviewed by the hospital's research committee and a decision about access to the data made subject to approval from the regional committee for medical research ethics first provided for the study and to new analyses being proposed.