Study design
This double-blind (assessor and participant), parallel-arm, pilot feasibility RCT was conducted in accordance with the National Health and Medical Research Council ethical guidelines(51), and reporting adheres to the Consolidated Standards of Reporting Trials (CONSORT) statement for pilot and feasibility studies(23) (Additional File 1). Ethical approval was gained from the La Trobe University Human Ethics Committee (HEC 16-077). The trial was prospectively registered through the Australia and New Zealand Clinical Trials Registry (ACTRN12616000564459).
Setting
All assessments and treatments were conducted at two private physiotherapy clinics in Australia, located in Hobart or Melbourne.
Participant recruitment and eligibility
Individuals who had undergone a hamstring-tendon autograft ACLR 12 to 15 months prior were recruited from five surgical lists, advertisements at La Trobe University, or via social media (December 2016 to August 2017). Individuals aged 18 to 50 years who were 12 to 15 months post-ACLR were considered eligible if they scored <87.5/100 on the Knee injury and Osteoarthritis Outcome Score (KOOS) QoL subscale (threshold below which has been defined a symptomatic knee(24)), and met one of the following criteria; a) <22 repetitions on the one-leg rise test; b) single-hop <90% limb symmetry index (LSI); or c) <87/100 on the Anterior Knee Pain Scale (AKPS)(44). These functional performance thresholds can be associated with worse symptoms and poorer knee-related QoL in the proceeding years(15, 25, 61), and the AKPS threshold can be associated with worse functional performance at 1-year post-ACLR(16). Exclusion criteria were: i) >5 years between injury and ACLR; ii) subsequent injury (for which medical treatment was sought) or follow-up surgery to the ACLR knee; iii) another musculoskeletal, neurological, or cardiorespiratory condition influencing daily function; iv) unable to speak or read English; and v) unable to attend eight supervised sessions.
Deviations from initial trial protocol
Participants were initially deemed ineligible if they had sustained a previous ACL or knee injury to either limb prior to their recent ACLR. After commencing recruitment, it was evident that a previous knee injury was common in those with persistent symptoms, and these individuals have an increased risk of symptomatic post-traumatic OA(78). The inclusion criteria were adjusted at the start of recruitment to include those with a previous ACL or knee injury. The single-hop performance LSI cut-off for eligibility was changed from 88% to 90%, as recent evidence suggests 90% is the most common criterion used for return-to-sport clearance(9). Hypothesis testing in a regression model was not performed as initially planned, due to the limitations of significance testing in clinical research(36), and this was not considered appropriate for a feasibility trial. Instead, the between-group differences and 95% confidence intervals (CI) were used to verify that a worthwhile effect was contained within the confidence interval(5). We defined a potential worthwhile effect as greater than the minimal detectable change (MDC) score for the respective outcome measures where available. While the primary purpose of feasibility was implied throughout the trial registration and included as such in the trial title, we did not list feasibility as a separate outcome in the trial registration. We have maintained our focus on feasibility by including it as the primary aim of this pilot study. Several other exploratory PROs were outlined in the trial registration but were beyond the scope of this evaluation due to the primary aim of feasibility.
Procedures
Eligible participants underwent a baseline assessment with a blinded assessor (BP) and were randomised into one of two intervention groups. The same blinded assessor completed all follow-up assessments unaware of group allocation. Participant age, sex, body mass index (BMI), injury history, ACLR rehabilitation (i.e. self-reported duration) and surgical details (i.e. self-reported graft type, meniscal procedures), and previous activity level were obtained at baseline. All patient-reported outcomes (PROs) were completed via an online portal (Promptus, DS PRIMA, Melbourne, Australia).
Randomisation and blinding
Non-stratified, permuted block randomisation (random blocks of 3 or 6) occurred at a 2:1 (lower-limb focussed: trunk-focussed) ratio. The randomisation sequence was computer-generated using Excel. The administrative staff at the participating physiotherapy clinic revealed the allocation using sequentially numbered, sealed opaque envelopes. The administrative staff were blinded to block size, and entered the group allocation to the participant’s clinical record for the physiotherapist. Participants were blinded to group allocation, to ensure allocation did not influence adherence, other treatment use, or increase the risk of drop-out. The physiotherapists were unable to be blinded to the allocation but were encouraged to deliver both interventions with equal enthusiasm and assertion of exercise value.
Treating physiotherapists and treatment fidelity
Treating physiotherapists were experienced (≥5 years treating musculoskeletal patients) in ACLR rehabilitation, and completed a 4-hour training session (led by BP) related to delivering both interventions. A manual, outlining the exercise prescription and progressions, manual treatment algorithm, education material, and trial procedures (attendance sheet, clinical notes, adherence monitoring) was provided to each physiotherapist (Additional File 2). Prescribed exercises were entered via Physitrack© smartphone application for participants to access via the participant-facing application PhysiApp© (Physitrack Ltd, London, UK).
Interventions
Participants were randomised to a lower-limb focussed or trunk-focussed exercise-therapy intervention, which were both delivered in eight face-to-face 30-minute physiotherapy sessions over 16-weeks. Both interventions are reported below according to the Template for Intervention Description and Replication (TIDieR) guidelines(38) and the Consensus on Exercise Reporting Template (CERT)(64) (Table 1).
Table 1. Summary of intervention delivery and components for both groups
What
|
Lower-limb focussed intervention
|
Trunk-focussed intervention
|
Who provides
|
Physiotherapists who have all undergone study-specific training
|
How
|
1-to-1 face-to-face sessions to assess and progress unsupervised exercise-therapy program
|
Where
|
Physiotherapy sessions: Private clinics in Hobart and Melbourne
Unsupervised exercise-therapy program: Clinic/public gym, or home
|
When & how much
|
Physiotherapy 1-to-1 sessions: 30 minutes duration, weekly for 4 weeks then every 2 to 3 weeks for 12 weeks
Unsupervised exercise-therapy program: instructions provided via PhysiApp©, 30 to 45 minutes duration, minimum 3 sessions per week, unsupervised
|
Tailoring
|
- Standardised lower-limb exercises (i.e. strength, power, balance), functional retraining (e.g. plyometric, agility) and cardiovascular program
- Choice of priority exercises* (from the standard set) was individualised
- Exercise progression was individualised
- Individualised education (e.g. exercise rationale, goal setting)
- Passive therapy treatment algorithm if appropriate (e.g. taping)
|
- Standardised, non-specific trunk strengthening exercises
- Progression of exercises was individualised
- Optional stretching
- Standardised education (e.g. rationale for trunk exercises)
|
Both groups: exercises progressed based on assessment of pre-defined criteria at each session (i.e., pain, swelling, technique) and resistance training principles
|
How well
|
Attendance at physiotherapy recorded by physiotherapists and clinic
Unsupervised exercise program adherence recorded by participants in PhysiApp© smartphone app or paper diaries, and monitored by physiotherapists via Physitrack©
|
* Physiotherapists could choose 3 to 4 priority exercises (out of a possible 8), based on the participant’s needs and goals. If necessary, all 8 exercise types were included, but it was not compulsory for all eight to be incorporated.
Lower-limb focussed exercise-therapy intervention
The lower-limb focussed intervention included standardised (with individualised progression) lower-limb, functional and cardiovascular exercises, and individualised, ACL-specific education (Additional File 2). The protocol was informed by current evidence-based recommendations(72), and developed by the research team, two of whom regularly (weekly) treated patients after ACLR (CB and RC). The lower-limb focussed exercise-therapy program targeted typical strength and functional impairments(79), and altered movement patterns(79) during sport-specific tasks related to ACL injury mechanisms (i.e., landing, and cutting). The eight areas in the exercise-program were: 1) movement retraining (e.g. landing); 2) lower-limb strength (e.g. squats); 3) balance (e.g. perturbation exercises); 4) hip-abductor strength; 5) calf strength; 6) trunk strength; 7) hip-extensor and knee-flexor strength; and 8) cardiovascular exercise (e.g. cycling, running, graded sport-specific activities). Each of the eight exercises had three or more phases of difficulty for individualised progression (Additional File 2). Physiotherapists were provided with a summary of the participant’s injury history, goals, 3 to 4 priority exercises, and suggested starting phases based on baseline assessment. Physiotherapists could add target exercises based on participant need, but it was not compulsory for all eight exercises to be incorporated. Exercise progression was based on: i) good technique; ii) minimal irritability (i.e. <2/10 pain during/after and no swelling); iii) resistance training principles related to muscular strength and power(1); and iv) participant-specific goals and feedback. Strength exercises were prescribed in 3 sets of 12 repetitions (each repetition performed as 2 seconds concentric, 1 second isometric, 2 seconds eccentric), and could be progressed to a power dosage prescribed in 3 to 5 sets of 5 to 10 repetitions (<1 second concentric, 0 isometric, 2 seconds eccentric)(1). Treating physiotherapists were encouraged to use the face-to-face sessions to check exercise technique, and adjust loads so that participants were reaching fatigue (i.e. they could not physically perform >2 more repetitions) after their prescribed dosage(1). Thirty-minutes was considered an appropriate appointment duration to supervise at least 1 set of prescribed exercises (the other 2 sets could be completed unsupervised in the clinic gym), and provide education.
Trunk-focussed exercise-therapy (control) intervention
An active control intervention was chosen to ensure that both treatment groups received equal exposure to physiotherapy(37). The trunk-focussed intervention was considered the active control, and included standardised (with individualised progression) trunk strengthening exercises, stretching and education. Physiotherapists could choose a minimum of three trunk strengthening exercises (from a maximum of five options), and each exercise had three or more phases of difficulty (Additional File 2). Exercises were prescribed according to resistance training principles; typically prescribed in 3 sets of 60 seconds (isometric), and progressed to achieve adequate fatigue (i.e., could not physically perform >5 more seconds)(1). Lower-limb and trunk stretching appropriate to the participant, could be prescribed (Additional File 2). The trunk exercises were predominantly isometric, non-weight-bearing, and had minimal lower-limb involvement and thus, were not expected to impact knee-related QoL, symptoms, or function. This was chosen as the control intervention as trunk muscle strength deficits has not been reported following ACLR, nor has addressing trunk strength been investigated to impact knee-related outcomes following ACLR. Trunk exercises were considered to provide a credible intervention to enhance control participant’s blinding to group allocation and minimise drop-outs.
Unsupervised exercise-therapy program (both groups)
Participants in both groups were prescribed an unsupervised exercise-therapy program relevant to their allocation, to be completed 3 times per week, at home or in a gym, to optimise likelihood of muscular strength and power improvements(29). Physiotherapists entered participant’s exercises via the Physitrack© app, for the participant to use PhysiApp© to guide exercises and record adherence on their own smartphone, tablet or computer. Paper diaries of the exercise-therapy programs were provided as required. PhysiApp© included video examples (created specifically for the trial) of correct (and incorrect) technique for each exercise (Additional File 3), and exercise dosage (e.g. number of sets/repetitions, time under tension, external load, rest time) according to resistance training and muscle adaptation guidelines(1, 69). Co-interventions were discouraged but if participants chose to receive other treatment, they recorded them on an “other treatments calendar”. The trunk-focussed unsupervised program could be completed at home with minimal equipment. When gym equipment was required for the lower-limb focussed unsupervised program, gym access was provided free of charge.
Education component (both groups)
Both groups received education, including face-to-face discussion and/or provision of handouts (Additional File 3). Handouts for the lower-limb focussed group covered the following topics: i) surgical information and post-operative expectations; ii) goal setting and return-to-sport criteria; iii) injury prevention; iv) psychosocial influences on recovery; and v) post-traumatic OA risk. The purpose of the education for the lower-limb focussed group was to provide informational support regarding ACL-specific topics and address common knowledge gaps regarding evidence-based rehabilitation(6), and psychosocial support for kinesiophobia, fear of re-injury, confidence, or negative lifestyle modifications (e.g. weight gain) known to influence outcomes(11, 56, 70). For the trunk-focussed group, physiotherapists could deliver standardised education on the rationale for trunk strengthening (e.g. theoretical influence of lumbo-pelvic stability on lower-limb biomechanics), or provide handouts/ face-to-face discussion on the topics “surgical information and post-operative expectations”, “psychosocial influences on recovery”, and “post-traumatic OA risk” (Additional File 3).
Primary outcome: feasibility
Feasibility was assessed according to previously published recommendations(47). Proceeding to a full-scale RCT was deemed feasible if all criteria were met, or reasonable amendments could be made to achieve these criteria in future trials(3).
Recruitment, adherence and retention was evaluated by: i) recruitment rate (criterion: 4 participants per month); ii) proportion of eligible participants who were willing to enrol (criterion: >80%); iii) physiotherapy attendance rate (criterion: >80%); iv) adherence to unsupervised exercise-therapy program (criterion: >80%); and v) proportion of drop-outs (criterion: <20%).
Acceptability of the study protocol was assessed via the appropriateness of the inclusion criteria (criterion: at least 1 in 3 eligible), and acceptability of the intervention content, delivery, adherence monitoring, and barriers or facilitators to adherence. Acceptability was determined via informal interviews conducted with participants and physiotherapists (Additional File 4).
Adverse events (i.e., any injury or illness requiring medical attention as a result of participating in the trial) ) were noted by the physiotherapist on a standardised recording sheet (criterion: <10% of all participants). Pain-level (on a visual analogue scale; 0=no pain, 10=worst possible pain) during the unsupervised exercise-therapy program was entered on PhysiAppã by participants (criterion: each participant mean <2/10 across all sessions).
Randomisation integrity was determined by contamination between groups (reported by participant or physiotherapist) (criterion: 0% contamination), or knowledge of group allocation by the participants or assessor (criterion: <10% unblinded).
Acceptability of the outcome measures was determined by the time needed to collect the data, and completeness of the outcome measures at baseline and follow-up (criterion: >90%).
Secondary Outcomes
Patient-reported outcomes
Knee-related QoL was assessed via the KOOS-QoL and ACL-QoL. The KOOS-QoL is one of the five KOOS subscales, and evaluates knee-related QoL(63). The KOOS-QoL has the highest content validity of all subscales and the greatest responsiveness in young adults following knee injury(10). The ACL-QoL was designed to assess additional domains (e.g. work-related, social and emotional) of knee-related QoL specific to a young, active ACL-injured population(49). The KOOS-QoL and ACL-QoL are converted to a total score out of 100 (0=extreme problems; 100=no problems). The KOOS-QoL and ACL-QoL have established content validity (Cronbach’s alpha >0.76), test-retest reliability (ICC>0.86), and responsiveness (effect sizes >0.5)(10, 46). The MDC is 8-10 points for KOOS-QoL(63), and 12-points for ACL-QoL(46).
The KOOS subscales of pain, symptoms, and sport were assessed, and all combined with the KOOS-QoL, to calculate an overall KOOS4 score. The KOOS individual subscales are valid, reliable and responsive following ACL injury(10). Psychological readiness for return-to-sport (a common goal of ACLR), and fear of re-injury was measured by the ACL Return to Sport Index (ACL-RSI)(74). The ACL-RSI has established test-retest reliability (ICC=0.89) and responsiveness (MDC=19 points)(45), and validity with higher scores associated with better return-to-sport rates, self-reported symptoms and function(45, 75). The global rating of change (GROC) on a 7-point Likert scale (“much worse” to “much better”) measured separately for knee pain and knee function; and the change in proportion of patients answering “yes” to the patient acceptable symptom state (PASS) question(39) were evaluated. The GROC has good face validity (Pearson’s r=0.72 to 0.90), test-retest reliability (ICC>0.90), responsiveness following knee injury (0.5 to 2.7 points on 7-point scale), and construct validity (e.g. correlated with changes in hop tests)(40). The PASS assists in interpretation of improvement in PROs by evaluating the concept of “feeling good” as opposed to “feeling better”(71)¸ and answering yes to “PASS” corresponds with better KOOS scores after ACL injury(39).
Functional performance outcomes
Functional performance outcomes were measured at baseline and follow-up, including the single-hop (maximum distance on one hop forward)(34), side-hop (maximum number of hops over two parallel lines 40cm apart in 30 seconds)(34), and one-leg rise test (maximum number repetitions from a standardised height)(68). We recorded the raw score (e.g. cm hopped) on the ACLR and contralateral limb, and calculated the LSI (score of ACLR knee divided by contralateral knee, multiplied by 100, expressed a percentage). The hop-tests and one-leg rise have high intra-rater reliability (ICC>0.80) and responsiveness after knee injury(7, 34, 60).
Data analysis
The sample size of 27 was not formally determined. It was based on previous pilot RCTs evaluating health-professional guided interventions for musculoskeletal conditions(41, 67), and deemed sufficient to assess the feasibility criteria. Participants who completed baseline and follow-up evaluations were included in the analysis, as recommended in the CONSORT guidelines(48). Feasibility outcomes were reported descriptively. The majority (>50%) of baseline and follow-up scores, and the change scores for the patient-reported and functional performance outcomes were normally distributed (assessed with Shapiro-Wilk’s test). Therefore, within-group, and between-group differences were reported as mean±SD, and mean and 95% confidence interval (CI), respectively. The treatment effect for the respective outcome measures was considered potentially worthwhile if the MDC was contained within the 95% CI of the mean between-group difference(5). Activity level, GROC, and PASS outcomes were reported descriptively. Decision criteria for progression to a full-scale RCT was based on: i) all feasibility criteria being met, or reasonable recommendations to achieve criteria in future trials and ii) presence of a potentially worthwhile treatment effect for knee-related QoL, symptoms, and function.