Tendinopathy Education on the Achilles

This feasibility randomized controlled trial was designed to address the following aims:

Specific Aim 1 examines if education on central pain mechanisms (i.e. biopsychosocial approach) compared to education on peripheral pain mechanisms (i.e. standard of care utilizing a biomedical approach) is more effective at reducing pain and disability in a pilot RCT exercise program for Achilles tendinopathy. We hypothesize that a biopsychosocial approach to education will decrease pain (NPRS) and disability (PROMIS physical function) more than the standard of care for patients with AT.

Specific Aim 2 determines which central pain mechanisms (nociplastic pain, pain psychology, motor dysfunction) are improved by an exercise program, regardless of education group. We hypothesize that exercise will improve all three central pain mechanisms (increased PPT, decreased kinesiophobia, increased number of heel raises).

A total of 66 adults with chronic Achilles tendinopathy (AT) will be enrolled (consented and randomized) in this single-site pilot randomized controlled trial (RCT). All data will be collected at a single site within an academic medical center. A pilot RCT comparing effectiveness of exercise with pain education to standard of care, i.e. exercise with biomedical education. Participants will be randomized to treatment group using a permuted block randomization scheme stratified by sex and AT type (insertional and midportion). Participants will complete up to 7 treatment sessions over an 8-week period. With an anticipated 10% dropout rate, we plan to analyze data on 30 adults with chronic AT per group.

Participants will complete 2 evaluation sessions at baseline and 8-week follow-up. Evaluation sessions will involve the following 4 types of testing.

1. Clinical exam (including ultrasound imaging)
2. Questionnaires (demographics, symptom description, and psychological) Private, identifiable information will be collected via questionnaires administered online and in-person. All data will be entered into a University of Iowa REDCap database, which is maintained by University of Iowa ITS staff.
3. Movement analysis and verbal pain rating Participants' motion and force will be captured as they walk, perform heel raises, hop, and do calf stretches within a 3-dimensional motion analysis with force plates flush with the floor. To monitor pain level throughout participation, we will have participants verbally rate their pain with each task.
4. Sensory testing (Pain pressure threshold at the hamstring and Achilles bilaterally; Conditioned pain modulation at the Achilles)

In addition, participants will complete questionnaires at 12-weeks. Participants will attend 7 treatment sessions over approximately 8 weeks. Over this time period participants will complete the following procedures

1. Questionnaires (exercise adherence, educational quizzes, symptom description) This private, identifiable information will be collected via questionnaires administered online and in-person. All data will be entered into a University of Iowa REDCap database, which is maintained by University of Iowa ITS staff.
2. Exercise participation Participants will be given exercises to do at home in between treatment sessions. Participants will receive instruction from a physical therapist prior to doing the exercise at home.
3. Education Participants will be given homework and online quizzes to do at home in between treatment sessions. The physical therapist will review the educational material with the participant at each visit.

In addition, a research team member will access the medical records of participants with Achilles tendinopathy to gain information about the participant's medical history and imaging related to Achilles tendinopathy.

The sample size was based on a power analysis for the primary outcome measures of Specific Aims 1 and 2.

Specific Aim 1: This analysis will compare the between group (biopsychosocial vs biomedical approach to education) changes in pain and disability scores from baseline to 8 weeks and to 12-week follow-up for participants with AT participating in an exercise program. Based on findings from Moseley et al. for an RCT comparing a 4-week exercise program with education on central pain mechanisms to standard of care in 49 patients with chronic low back pain, we anticipate between group differences with Cohen's d greater than or equal to 0.36 for pain (between group difference across 2 time points=0.75, SD of 1.05 on the NPRS, effect size of f=0.36) and disability (between group difference across 2 time points=1.95, SD of 2.33 on low back pain-specific measure, effect size of f=0.42). Under these assumptions, a sample size of 30 patients per group would be needed to reach 80% power for the time averaged difference between two group means in a repeated measures design with alpha equal to 0.025 (Bonferroni correction of 0.05/2 for 2 outcomes in Aim 1) to detect a between group effect size of 0.36.

Specific Aim 2: This analysis will compare the within group changes (baseline to 8-weeks and to 12 weeks) in central pain mechanisms (nociplastic pain, psychosocial factors, motor dysfunction) with an exercise intervention for participants with AT, regardless of educational group. The power analysis for this aim is based on data from 3 studies that have had a moderate to large (Cohen's d greater than or equal to 0.51) treatment effects on central pain mechanisms. A study led by Dr. Sluka previously found that massage improved PPTs 2 days after an intense exercise protocol with a large effect size (27.6%, SD=14.7%, effect size= 1.88). Cai et al. found that a 4-week cognitive behavioral therapy program reduced fear of movement by 8.1 points with a SD of 5.44 (effect size= 1.5). Based on our preliminary data from the K99 phase, there was a mean improvement of 3.4 repetitions (SD of change= 6.7) in heel rise performance after an anesthetic injection for the AT group (effect size=0.51). We used the smallest and population-specific anticipated effect size of 0.51 to calculate the sample size for Aim 2. Type I error rate of alpha equal to 0.017 (Bonferroni correction of 0.05/3 for 3 outcomes in Aim 2) will be used for this aim to adjust for the multiple comparisons. The sample size of 60 patients calculated for Aim 1 would allow us to detect an effect size of d greater than or equal to 0.43 with 80% power under these assumptions.

The normality of the continuous data will be tested by the Shapiro-Wilk test and by examining the quantile-quantile plot. Normally distributed Continuous variables will be presented as mean plus or minus SD for normally distributed data and median with interquartile ranges for non-normally distributed data. When the normality assumption is not met, transformation, such as log, will be used to complete planned parametric analyses. Type I error rate will be maintained at 0.05 by using Bonferroni adjustment for multiple comparisons. First, potential differences between group demographics at baseline will be examined univariately using two independent samples t-tests and chi-square tests, as appropriate. If differences are observed, these variables will be used as covariates in the multivariable models.

A modified intention-to-treat principle will be followed, which will include outcome data on all participants based on the group they were randomized and complete evaluation at 8 weeks. We will also compare patient characteristics of those who remained in the study to those who dropped out to determine if data at subsequent time points is consistent with missing at random.

For Aim 1 we will assess the effect of the interventions on the primary outcomes for pain (NPRS) and disability (PROMIS physical function) from baseline to 8-weeks and to 12-week follow-up using linear mixed model for repeated measures. Similarly, for Aim 2 a linear mixed model for repeated measures will assess the effect of the interventions by group on the primary outcomes for central pain mechanisms (nociplastic pain: PPT, psychosocial factors: TSK, motor control: heel raises) from baseline to 8 weeks. For Aims 1 and 2, the factors in the linear mixed model will include group and time effects. In addition, the significance of the group*time interaction term, where the group*time interaction tests if the change over time differs between groups, will also be tested. We will also assess for normality, linearity, homoscedasticity, and independence of residuals for the multiple regression models (dependent variables: pain and disability). Statistical significance will be defined by alpha greater than or equal to 0.05.

Analyses for secondary outcome measures of central sensitization, psychological factors, motor control, pain and disability for Aims 1 and 2 will be performed using the same statistical tests described above. Secondary analyses on sex will examine potential sex-based differences in both Aims 1 and 2 to inform sample size estimates for future clinical trials. The analyses described above for each aim will be expanded to explore the effect of sex as a fixed factor and examine interactions with group and time. Using the same data analysis strategy as described for sex, a secondary analysis on AT type (insertional vs. midportion) will also be assessed for Aims 1 and 2 to inform recruitment strategies for future clinical trials.