Hand exercise for women with rheumatoid arthritis and decreased hand function: an exploratory randomized controlled trial

Participants were females with RA reporting ADL task performance problems involving the hands, recruited from May 2014 to January 2016 through rheumatologists at outpatient clinics in Copenhagen and announcements in daily press. Potential participants were given further information and pre-screened for eligibility via telephone. If inclined to participate, they were examined by a rheumatologist to determine if they fulfilled inclusion criteria.

Inclusion criteria are as follows: females aged > 18 years; diagnosed with RA (ACR/EULAR 2010 Criteria) [20]; involvement of minimum of one tender wrist, MCP, or PIP joint; stable medication 3 months prior to participation, and self-reported decreased ability to perform ADL tasks involving the hands. Exclusion criteria are as follows: significant osteoarthritis of the hand (assessed by the rheumatologist); hand surgery within 6 months; other pain condition involving muscles and/or joints; prednisolone therapy; alternative treatments during study period; change in medical treatment during study period; inability to understand Danish; and finally any other contradictions for participating assessed by the rheumatologist. Participants were asked not to participate in occupational or physical therapy interventions elsewhere during the study period, aquatic physical therapy was allowed.

Participants gave written informed consent. The study was approved by the ethics committee of the Capital Region of Denmark (H-3-2014-025) and registered at ClinicalTrials.​gov (NCT02140866). The study was conducted in accordance with standards of the responsible committee on human experimentation and the Declaration of Helsinki. If participants experienced increased disease activity during the study period, US and rheumatologist examinations were made to determine if modifications or withdrawal were needed.

The investigator-initiated study was designed as an RCT with parallel groups. Participants entered an 8-week program and were randomly assigned to either intervention group (hand-exercise therapy and CIP, CIP_EXERCISE) or control group (CIP alone, CIP_CONTROL) with a 1:1 equal allocation ratio utilizing a concealment process. Randomization was made using sealed envelopes and carried out by the project secretary. The outcome assessors and data analysts were kept blinded to the allocation, and participants were instructed not to convey their group allocation. The persons performing US examination and evaluation were blinded to results of clinical examination and group allocation.

The CIP consisted of an introduction to compensatory strategies including joint protection (JP), assistive devices, and alternative methods of performing ADL [21, 22]. The program was client-centered focusing on improving the ability to perform those ADL tasks that the single participant identified as purposeful to her life. It consisted of four 1-h sessions during an 8-week period. One occupational therapist (OT) (ISH) with > 10 years of experience performed all sessions. The first session aimed to identify the participant’s individual resources and problems in ADL task performance, goal setting focused on ADL task performance, and introduction to assistive devices. As homework, assistive devices were applied at home and the utility evaluated. In the second session, the OT presented JP principles and discussed these in relation to the participant’s ADL task problems. She supervised the participant to integrate JP principles and alternative methods of doing in ADL tasks. In the third session, the OT followed up on the use of JP principles and assistive devices at home and together they identified any additional ADL task, representing a challenge, and performed this under supervision. Optionally, additional assistive devices were handed out. As homework, the participant practiced JP principles and assistive devices in the chosen task. Fourth session, the OT followed up on homework, evaluated goals, and introduced how to apply for assistive devices and housing accessibility solutions in the home municipality. This fourth session, if relevant, was done as a telephone meeting.

The hand-exercise program lasted for 8 weeks as a strength increase is possible within this time frame and was designed based on recent research [6, 8, 23‐25]. First, the exercise intervention was performed four times per week with one session supervised by a physical therapist (PT), and the other three were home based. After 14 days. One PT (CB) with 3 years of experience performed the individual exercise sessions. Home-based exercise sessions were recorded in a diary. During the period, the exercise intervention was increased to once daily and the load was increased. The program consisted of three parts: (1) warm-up/mobility (10 min), (2) muscle strength training (20 min), and (3) cool-down (5 min). The warm-up was performed to prepare the joints for the muscle strength training and to improve flexibility; the muscle strength training was designed to ensure that relevant muscle was targeted within a period of 20 min. Resistance was supplied by exercise bands and Thera-putty. The amount of resistance was based on the weakest hand. The resistance intensity was set according to the participant’s self-reported experience of load using the Borg Scale [26]. This load was chosen to minimize the risk of flare-up symptoms due to overload and to allow for progression. For cool-down, some of the warm-up exercises were receded. All participants received a detailed illustrated description of the exercise program (Additional file 2). In case of any flare-up in symptoms, the participant was set to only conduct the warm-up and cool-down part. Post exercise soreness and temporary fatigue was tolerated. Detailed description of the exercise program is presented in Additional file 3. If a participant failed to meet at an appointment, a phone call was made to maintain fidelity.

Primary outcome was observed, ADL motor ability measured by Assessment of Motor and Process Skills (AMPS).

Secondary outcomes were observed, ADL process ability (AMPS), self-reported ADL ability (ADL-Questionnaire, ADL-Q), self-reported disability (Stanford Health Assessment Questionnaire Disability Index, HAQ-DI), overall disease activity (DAS28), grip strength, and pain. Exploratory outcome was disease activity assessed by US.

The AMPS is a standardized observation-based tool used to measure a person’s observed ADL task performance [27, 28]. The person being evaluated chooses and performs at least two standardized ADL tasks of relevance and appropriate challenge. During AMPS evaluation, two domains are evaluated, i.e., ADL motor ability (the amount of effort, fatigue and/or clumsiness) and ADL process ability (the degree of disorganization, inappropriate use of time, space, objects and ability to adapt actions). The 16 ADL motor and 20 ADL process skills are evaluated in terms of ease, efficiency, safety, and independence using a four-point ordinal scale. The available AMPS software [29], based on a many-faceted Rasch measurement model, makes it possible to convert ordinal raw scores into overall linear ADL motor ability measures and overall linear ADL process ability measures adjusted for task challenge, skill item difficulty, and rater severity. Measures are expressed in logits (log-odds probability units) [28]. The overall ADL motor ability measure indicates how much effort or fatigue the person demonstrated, and the overall ADL process ability measure indicates how efficient the person was observed to be during the ADL task performance. Additionally, both ADL ability measures reflect safety and independence in ADL task performance. ADL ability measures above the 2.0 logit cutoff on the ADL motor scale and above the 1.0 logit cutoff on the ADL process scale indicate effortless, efficient, safe, and independent ADL task performance in everyday life. In contrast, ADL motor ability measures below the 2.0 logits cutoff indicate increased effort or fatigue during task performance. Moreover, ADL ability measures below the 1.50 ADL motor cutoff and/or below the 1.00 ADL process cutoff indicate a need for minimal assistance for community living. Finally, according to the AMPS manual a difference of > 0.30 logits on the AMPS ADL motor and/or ADL process scale defines a clinically relevant change [28]. Studies support that the AMPS ability measures are reliable and valid in RA patients [3].

Self-reported ADL ability was assessed using the ADL-Q, a standardized instrument with 47 ADL tasks developed to measure a person’s perceived quality of ADL performance [3, 30]. The person marks the quality of the ADL task performance using seven response categories reflecting efficiency, effort/fatigue, safety, and independence. Rasch methods are applied [30]. Studies support that the ADL-Q can be used to generate valid measures of self-reported quality of ADL task performance among RA patients [3].

The HAQ-DI is developed to assess disability in RA [21]. The questionnaire consists of 20 questions primarily concerning ADL tasks, a Danish version was used [31].

Overall disease activity was assessed using the DAS28 which is based on count of 28 joints for swelling and tenderness, C-reactive protein level in the blood and the patient’s self-reported impact of disease on a visual analog scale (VAS Global health). The DAS28 score range from 0 to 9.4 [32].

The maximal grip strength was measured in kilogram using a digital hand Dynamometer (North Coast Medical Inc.). The grip strength was measured three times in both hands. The maximal force performed in each hand was used in the analysis.

Hand pain during activity and in rest was measured in both hands on a visual analog scale (VAS), where zero was no pain and 100 was maximal pain.

Synovial hypertrophy and increased synovial perfusion are indications of disease activity assessed by US [33]. Gray scale US is used to examine synovial hypertrophy seen as hypo-echoic structure between the tendon/muscle and the bone [1]. Doppler US added to the gray scale image register movement of the blood as an indication of increased synovial perfusion. US has shown to correlate with measures of disease activity in RA [19, 34, 35]. US examination in RA has shown to display sub clinical disease activity leading to disease progression on X-ray [36]. The wrist and MCP 2–5 were examined both dorsal and palmar. The wrist was scanned in four dorsal and one volar position and the MCP joints in three dorsal and one volar position. Both synovial hypertrophy and Doppler were evaluated using a validated scoring system for RA [37]. One score for synovial hypertrophy, one score for synovial perfusion (Doppler), and one sum score were calculated.

Sample size calculation was based on previous data on AMPS ADL motor ability in women with RA [3]. For a two sample pooled t-test of a normal mean difference with pooled variances (equal variances assumed) and a two-sided significance level of 5% (p ≤ 0.05), assuming a common standard deviation (SD) of 0.36 logits, a sample size of n = 32 participants per group was required to obtain a power of at least 90% to detect a group mean difference of 0.3 logits. It was decided to include n = 45 participants in each group.

Data analyses were carried out according to a pre-established statistical analysis plan (SAP); all analyses were done applying SAS (v. 9.4, SAS Institute Inc., Cary, NC, USA). Descriptive statistics and tests are reported in accordance to the “Enhancing the QUAlity and Transparency Of health Research” (EQUATOR) network [38, 39]. To evaluate the empirical distributions of the continuous outcomes, visual inspection was applied to suggest whether the assumption of normality was reasonable. The PROCUNIVARIATE statement was used for summarizing the data (descriptive statistics).

Intention-to-treat (ITT) analyses were made, i.e., analyzing participant outcomes according to the group to which they were allocated, even if participants did not receive allocated intervention. The ITT principle was done by replacing missing data with the value obtained at baseline.

At week 8, the CIP_EXERCISE group was compared with the CIP_CONTROL group using general linear model (analyses of covariance; ANCOVA) for mean changes from baseline and t tests for comparison of least squares means between groups. The model included change as the dependent variable (Δ), with treatment group as a main effect and the baseline score as an additional covariate. Results were expressed as the difference between the group means and 95% CI with the associated p values, based on the general linear model.

For sensitivity purposes, the analyses were repeated with further adjustment for disease duration and hand pain at baseline as there were group imbalances in these variables at baseline (Additional file 1: Table S1).