Graded motor imagery for women at risk for developing type I CRPS following closed treatment of distal radius fractures: a randomized comparative effectiveness trial protocol

Distal radius fractures (DRF) are the most common fractures of the upper extremity in the United States [1‐3]. These fractures account for up to 18% of all fractures in persons over the age of 65 years and are nearly 5 times more likely to occur in women than in men [4]. With advancing age, the incidence for women increases rapidly after age 55, almost doubling every 10 years until 90 years of age [5] with the average age of occurrence being 56 years [6]. Women in an urban setting are 30% more likely to sustain a DRF than those in a rural setting [7]. The risk of sustaining a DRF is 20% greater in winter months (RR = 1.2), and 45% more likely to occur on days that begin with snow and ice on the ground (incidence rate ratio, 1.45) [8‐10]. Given the globe’s aging population, osteoporotic fractures of the wrist, humerus, spine or hip are expected to increase further in the coming years, and the aftercare will be an increasing burden on healthcare resources [11]. Typical medical care for non-operative DRF involves closed reduction in the emergency department or orthopedic office, followed by immobilization in a forearm based cast for four to six weeks, and early active range of motion to non-immobilized joints [12, 13].

Women sustaining a wrist fracture are likely to have a clinically important functional decline in self care, productivity, and leisure performance [14]. Disability levels based on the Patient Rated Wrist and Hand Evaluation (PRWHE) are reported to be high (75%) at the first week after DRF, and modertate (43%) at 8 weeks post injury [15]. Wrist and forearm active range of motion (AROM) and hand grip strength are the most commonly assessed physical impairments in all persons who are status-post DRF and are, on average, reduced by 40 to 50% at eight weeks post injury [16]. A case-control designed study revealed that distal radius fractures in both men and women have large effects (ƞ ²  = .49) on grip strength relative to matched controls [17] and, in turn, grip strength is a significant predictor of wrist function as measured by the PRWHE [18, 19]. Recent research also suggests that persons with DRF experience significant impairments in sensorimotor functions such as wrist joint proprioception and moving light touch awareness [17]. There are also indications that sensorimotor functions have significant associations (fine motor: r = −.52, joint position sense error: r = .63) with pain levels [17].

Resting pain and pain with activity are common aftereffect following DRF and have been documented to persist in 32.6 and 41.9% of cases respectively at two years post injury [20]. At the time of initial follow up after first definitive medical/surgical treatment of DRF, 81% of persons have been reported to describe having “severe” to “very severe” pain [15].

Type I complex regional pain syndrome (CRPS) is a condition that presents following an injury or illness where nerves are not damaged and is characterized by regional pain, increased sensitivity to touch, swelling, and impairment of motor function that are out out of proportion relative to the extent and location of the original injury/illness [21‐24], and has been reported to occur in as many as 37% of persons who sustain DRF [25]. Jellad et al. [25] report that women are 5.8 times more likely than men to develop CRPS after DRF. Persons who sustain low to medium energy impact DRF are 7.7× more likely to develop CRPS than those who sustain high impact fractures, [25]. Additionally, Moseley et al. [26] report that persons with a 0–10 numerical pain scale (NPS) rating of greater than 4 within one week after closed reduction and casting for DRF were 15.1 times more likely to develop CRPS, than those with less than a 5/10 NPS rating.

Many theories exist on the mechanisms for developing Type I CRPS [27]. Recent evidence indicates that CRPS likely has a component mediated by the cerebral cortex [28, 29]. Neuroimaging studies have revealed CRPS-associated shrinkage of the areas of the primary somatosensory cortex (S1 and S2) that represent the painful limb [30, 31]. These cortical changes may be linked with pain sensitization [28, 30]. Other researchers have illustrated disinhibition of the motor cortex and a disrupted body schema in persons with CRPS [32, 33]. Peripheral sensorimotor changes such as those observed by Karagiannopoulos et al. [17] may be indicative of cortical changes occurring following DRF and cast immobilization.

As many as 37–58% of persons undergoing closed treatment and cast immobilization following DRF go on to develop Type I CRPS [34, 35]. These numbers justify attention, given that most DRF are managed non-surgically via closed reduction and cast immobilization [36, 37]. Therefore, it is important to tailor therapeutic interventions for DRF that address or prevent development of CRPS. Recent evidence suggests that CRPS may develop due to central nervous system changes [27]. Thus, modulation cortical activity might be a noteworthy therapeutic avenue for individuals at risk of developing a CRPS.

Interventions designed to address centrally mediated pain and motor dysfunction are referred to as Movement Representation Techniques (MRT) [38]. These therapies involve observing or imagining normal and pain free movements, which may be performed simultaneously with sensory stimulation and active motion. The aim of MRT is to facilitate pain free movements of a painful limb [38]. MRT include mirror therapy (MT) and graded motor imagery (GMI). GMI is a multidimensional MRT and consists of three intervention phases: 1) limb laterality recognition, 2) Explicit motor imagery, and 3) the aforementioned, MT. Phase 1 involves viewing images of upper limbs and identifying if they are right or left limbs. In the absence of worsening limb symptomology and after proficiency in identifying limb segments has been established, participants transition into phase 2. Phase 2 involves viewing of images of upper limbs and imagining that the affected limb segment is assuming the postures depicted absent pain. Providing that symptoms are controlled, after 2 weeks participants transition into phase 3. In phase 3 (MT), the mirror image of the unaffected limb is observed engaging in various postures. This phase is completed in 2 weeks. All phases progress participants through imagining, and viewing movements that are least to most likely to proke painful experiences.

The upper limb rehabilitation literature is rich in restorative approaches (i.e., resolving impairments when already present). However, little is known about strategies to prevent upper limb, impairment, disability and pain. Recent longitudinal studies indicate that as many as 1/3 to nearly 1/2 of all persons who undergo closed treatment of DRF develop type 1 CRPS [24, 25] and 80% of these patients are women [25]. Current evidence indicates that cortical changes in upper-limb sensory andsensorimotor representation occur within days or weeks during immobilization [46‐48], and these changes appear to mimic cortical changes in persons with type I CRPS. A multidimentional MRT, Graded motor imagery (i.e., laterality training, motor imagery and mirror therapy), has emerging support in the literature as a restorative treatment strategy for patients with CRPS [38, 39, 41, 42, 44], however, studies have not examined GMI as a treatment strategy to prevent the onset of CRPS. Implementation of the present protocol will allow investigation into the effects of a modified GMI (mGMI) intervention during the cast immobilization period for women with closed DRF.

This is a protocol for a randomized comparative effectiveness trial, where the therapy outcomes of two six-week long intervention programs, aSOC only programand an mGMI+SOC program, are compared. The experimental design will be as is described in Table 1.

To ensure that criteria for participation are met, patients will be recruited from outpatient clinics staffed by physicians specializing in Orthopaedic, Sports and Family Medicine. The clinics are associated with a large public university in the Midwestern United States. At the time of the first orthopaedic treatment for closed DRF, patients will be offered the opportunity to participate. Prior to enrollment and allocation, oral and written informed consent will be sought. We will seek to enroll 66 participants who have received closed treatment of a distal radius fracture. Through use of a block randomization method, participants will be allocated to a SOC intervention (n = 33) group or a SOC + mGMI (n = 33) group. Following the consent process, the PI, not a blinded evaluator or interventionist, will actuate the randomization through use of a random block assignment generator [49]. To ensure partcipants are and remain blinded to allotment, participants will be informed that they will be randomized to one of two “occupational therapy treatment approaches” and will be asked not to discuss their treatments with others enrolled in the trial. Evaluators will be blinded to allocation. The feasibility of enrolling these numbers is high given that from 2013 to 2015, the aforementioned Orthopaedics clinic alone cared for an average of 218 women a year who met the below described inclusion criteria. These numbers are likely attributable to the characteristics of the area served by this health system (i.e., urban, northern latitude with slippery weather conditions) which are associated with increased incidence of DRF [7, 9, 10]. The proposed flow of participants from enrollment to completion is reflected in Fig. 1.

DRF will be managed as per the American Academy of Orthopedic Surgeons guidelines [13]. Upon enrollment in the study, participants will be randomized to an mGMI +SOC group or a SOC only group. The SOC group includes a combination of clinic and home based programming to address hand, elbow and shoulder motion as well as edema control whereas the mGMI +SOC group receives the same SOC programming in addition to MRT interventions intended to maintain the affected limb’s cortical representation. Participants in both groups will complete four clinic-based intervention sessions of an hour each across a 6 week period where thefocus will be on facilitating home program competency and advancement as indicated. The first session will occur within one week of cast treatment and home programming will supplement these clinical sessions. The entire therapy protocol for each group is standardized and manualized to ensure intervention fidelity. SOC and all phases of mGMI home programming are standardized and are to be provided in the form of written materials. The format will be one on one with the interventionist and will involve instruction in home programming as described below. All interventions will be carried out by certified OTs, including a certified hand therapist. Table 1 illustrates the intervention timeline. To enhance intervention fidelity both interventions are protocolized and manualized and the evaluator all interventionists will undergo formal training by the PI. Prior to the initiation of the trial, the PI will conduct a competency assessement of the evaluator and the interventionists.

All outcome measures will be administered by a single licensed OT who will be blinded to allocation. Blinded assessments occur within 1 week of cast immobilization (baseline), at four three weeks post cast immbolization, cast removal, and at three months post cast removal. A list of measures and a timeline of all data collection points can be found in Table 2.

Regarding the continuous primary outcomes (i.e., SF-MPQ, PRWHE) a sample size of 30 per group will give 81% power to identify as significant a mean difference of .75 SD in change in these outcomes between the two groups [72, 73]. A sample size of 30 per group will give 81% power to identify as significant a mean difference of 0.75 SD in change in continuous outcomes between the two groups. A sample size of per group will give 81% power to identify 20% vs. 58% as statistically significant for the Dichotomous outcome, Budapest diagnostic criteria. To adjust to an anticipated dropout rate of 10%, we will overenroll in each group by 3 bringing each group’s enrollment to 33. Following this adjustment, the total sample size will be 66 patients.

Continuous variable outcomes (Primary: PRWHE, SF-MPQ; Secondary: Goniometry, grip dynamometry and JPS) will be analyzed using a mixed effects linear models to evaluate change over time between groups. Models will include a random effect to account for the within-subject correlation for repeated measurements, and fixed effects of group, time, and group by time interaction. Model assumptions will be examined before fitting models. If assumptions are not met, nonparametric test (Wilcoxon test) will be used. For dichotomous outcomes (diagnosis of CRPS), Fisher’s exact test will be performed to compare proportions between groups.

In a manner consistent with that described of Walenkamp et al. [73], an achor-based approach to determining the MCID of the PRWHE and SF-MPQ will be taken through use of the Patient’s Global Impression of Change (PGIC) [74] scale as the anchor. The PGIC is a seven-point Likert scale, ranging from ‘very much improved’ to ‘very much worse’, and will be administered at the time of cast removal. The Minimal detectable change (MDC) of these two measures will be determined through use of the statistical methods described by Walencamp et al. [73].

Participants will be allowed frequent rest breaks to limit any amount of fatigue and soreness participants may experience in their hands and arms following testing. If discomfort increases as a result of testing, participants will be instructed on self- managing symptoms through use of physical agents such as cryotherapy and rest. In the event that this research activity results in an injury, treatment will be available, including first aid, emergency treatment and follow-up care as needed. Participants who believe they have suffered a research related injury, will be asked to inform the researchers immediately.

If participants are physically uncomfortable or uncomfortable with the nature of the questions or occasional touch, they will be instructed that they are free to withdraw from participation at any time without any impact to relationships with the university.

At each contact with the participant, the will seek information on adverse events by specific questioning and, as appropriate, by examination. Information on all adverse events will be recorded immediately in the source document, and also in the appropriate adverse event module of the case report form (CRF). All clearly related signs, symptoms, and abnormal diagnostic procedures results will be recorded in the source document. Reports of all serious adverse events (including follow-up information) will be submitted to the IRB within 10 working days if it falls under the UPIRTSO guidelines. Copies of each report and documentation of IRB notification and receipt will be kept in the Clinical Investigator’s binder. A second copy will be sent to the sponsor.

The PI will be responsible for 1) collecting, reporting, and risk management of adverse events, 2) data collection, entry, transmission and analysis, 3) site coordination and enrollment, 4) regulatory issues such as IRB actions, and conflict of interest disclosures, and 5) reporting the interim analysis to IRB. All collaborators will be immediately notified of any adverse events occur.

REDCap will be used for data capture and management. REDCap is a web-based data entry package that is structured so that access is only through a secure login by certified study personnel. Data entry screens will mimic the format of case report forms and include programmed automatic data field checks for real-time data quality control.

Participant identifiers will be stored separate from raw data in a separate secured and encrytped dataset housed within the RedCAP system. In a separate dataset, participants’ names and date of birth will be linked to a participant number. Demographic data, medical comorbidities data, and outcomes data will be housed in a separate dataset in RedCAP and these data will be linked to these participant numbers.

This protocol is not without some limitations. Participants will not be screened for motor imagery abilities. This was a decision made to avoid use of time-consuming research tools, avoid further restricting elgibility, and because persons with cognitive and right-left discrimination impairments (i.e., Dyslexia) who are subsequently predisposed to challenges with GMI will be screened out. For ethical reasons, the control group in the proposed study will receive standard care and thus, this study does not invole a no-treatment or sham comparison group.

Future study would include additional randomized control trials to investigate 1) the most effective treatment intensity and duration for the protocol, and 2) the combined effectiveness of this and other rehabilitative interventions.

This study is expected to conclude in late 2021.The initial phase of study preparation has focused on building communication with referral sources to allow for recruitment and ease of enrollment. Pre-trial preparation has focused on interventionist training, refining intervention scripts to deliver patient education consistently, and construction of necessary intervention materials (e.g., portable mirror boxes and laterality cards). Table 4 describes project deadlines.