Randomised clinical trial: combined impact and resistance training in adults with stable Crohn’s disease

47 assigned exercise (n = 23) or control (n = 24) groups up for 6 months. The exercise group usual care plus a 6-month combined impact and resistance training programme, involving three, 60-minute sessions per week and a gradual tapering of supervision to self-management. The control group received usual care alone. The primary outcomes were BMD (via dual energy X-ray absorptiometry) and muscle function (measures of upper and lower limb strength and endurance) at 6 months. Results: At 6 months, BMD values were superior in the exercise group with statistical significance at lumbar spine (adjusted mean difference 0.036 g/cm 2 , 95% CI 0.024-0.048; P < 0.001), but not at femoral neck (0.018 g/cm 2 , 0.001-0.035; P = 0.059) or greater trochanter (0.013 g/cm 2 , −0.019 to 0.045; P = 0.415) after correcting for multiple outcomes. The exercise group also had superior values for all muscle function outcomes ( P < 0.001; unadjusted mean differences ranging 22.6‒48.2%), and lower fatigue severity ( P = 0.005). Three exercise-related adverse events were recorded: two instances of light-headedness and one of nausea. The intervention improved BMD and muscle function in adults with CD appears as a suitable model of exercise for reducing future risk of osteoporotic fractures and disability.


| INTRODUCTION
People with Crohn's disease (CD) have an increased risk of osteoporosis and osteoporotic fractures compared with the general population. 1 Factors contributing to bone loss in CD include chronic inflammation, reduced vitamin and mineral absorption, extensive small-bowel disease or resection, corticosteroid use, older age, physical inactivity, smoking and nutritional deficiencies. 2 Bone mineral density (BMD) is one of the most important determinants of fracture risk, accounting for 60%-70% of the variance in bone strength. 3 A 2016 Guideline/Consensus paper from the European Crohn's and Colitis Organisation (ECCO) stated that weight-bearing exercise, stopping smoking and maintaining adequate dietary calcium (1 g/d) may help to prevent bone loss in inflammatory bowel disease; 4 however, few prospective trials of preventative interventions have been conducted in this increased-risk group.
Physical activity and exercise are important determinants of bone health. 5 Gravity-derived impact loads and muscle forces during exercise produce strain within the axial skeleton, which is osteogenic, stimulating bone formation. 6 Observational studies and randomised controlled trials (RCTs) indicate that physical activity and exercise can have beneficial effects on bone mass and strength across the age spectrum, thus reducing the risk of fractures. 5 Load intensity is one of the most important training variables, with guidelines and evidence syntheses indicating that adults should undertake a combination of impact activities (eg jumping) and high-intensity resistance training to optimise bone health. [7][8][9] Resistance training can also improve muscle mass and function, 8 which may be of importance in CD where low muscle mass and strength is common and predictive of osteopenia/osteoporosis. 10,11 Empirical evidence on the effects of exercise in CD is sparse, with only a handful of intervention studies, all of which have focused on modes of exercise that are sub-optimal for improving bone health (eg walking, cycling, yoga). 12 Only one RCT has investigated the effect of exercise on BMD, but the intervention was a low-impact resistance training, ie not a specific bone-loading programme. 13 To address this evidence gap, we conducted an RCT called PROTECT (PROgressive resistance Training Exercise and Crohn's disease Trial), which aimed to evaluate the effects of a 6-month combined impact and resistance training programme on BMD and muscle function in adults with CD.
The programme was designed to be practical to deliver by including a gradual tapering of supervision to self-management. Secondary aims were to explore the safety of the exercise programme and its effects on other health markers (eg fatigue, health-related quality of life), and to evaluate differences in BMD and muscle function between patients and matched healthy controls.

| Study design
PROTECT was a two-arm, randomised, parallel-group and assessorblind trial. Participants were randomised 1:1 to receive usual care plus a 6-month impact and resistance training programme (intervention) or usual care only (control). Study assessments were conducted at baseline and at 3 and 6 months after randomisation. Recruitment was from a large

| Participants
We included male and female patients aged 16 years or older with a clinical diagnosis of CD. Patients had to have a stool calprotectin of <250 μg/g, stable medication (>4 weeks), and quiescent or mildly-active disease, as indicated by a Crohn's Disease Activity Index (CDAI) of <150 or 150-219, respectively. Exclusion criteria were contraindication to exercise testing or training, 14 pregnant, planned pregnancy or major surgery within the first 6 months after randomisation, and current participation in 2 or more sessions per week of resistance exercise (self-reported) or another clinical trial where concurrent participation was deemed inappropriate.
Recruitment methods included liaisons with members of the direct care team during routine hospital appointments, postal invitations to patients who had previously consented to be contacted about future studies, and advertisements in hospital clinics and on social media. All participants provided written informed consent before enrolment.

| Randomisation and masking
Participants were randomly assigned 1:1 to one of the two trial arms using a computer-generated randomisation schedule stratified by gender and baseline disease status (quiescent [CDAI < 150] or mildly-active [CDAI 150-219]). The randomisation process was managed by an investigator at Northumbria University (GT) who was not involved in recruitment, intervention delivery or data collection. A researcher (KJ) emailed this investigator for notification of group allocation once the participant had consented and completed baseline assessments. Following notification, the researcher contacted the participant to inform them of their allocation. Due to the nature of the intervention, participants and intervention facilitators were not masked to group allocation. Outcome assessors were masked to group allocation, and participants were asked to not disclose their allocation. This instruction was adhered to by all participants.

| Procedures
Both groups received usual care, which comprised evidence-based medical treatment optimisation. Participants allocated to the control group did not receive any supervised exercise or exercise advice during the trial; however, following their final assessment they were offered a telephone-based consultation with a researcher who discussed their individual facilitators/barriers to exercise, and provided general advice on incorporating exercise into their lifestyle.
Participants allocated to the exercise group were encouraged to complete three, 60-minute exercise sessions on nonconsecutive days each week for 26 consecutive weeks, commencing the week following their group allocation. Twelve sessions were supervised and 66 were unsupervised, with a gradual tapering of supervision to self-management as follows: two supervised sessions per week in weeks 1 and 2; one supervised session per week in weeks 3 and 4; one supervised session every fortnight in weeks 5/6 and 7/8; one supervised session every month from week 9 onwards. The supervised sessions aimed to motivate participants to exercise regularly and teach them how to exercise correctly.
Each session involved a warm-up, a main conditioning phase and a cool-down. The warm-up lasted approximately 5 minutes and included several pulse-raising exercises (eg marching on the spot, squat and punch) and dynamic stretches (eg big arm circles, forward and backward leg swings). The main conditioning phase lasted approximately 50 minutes and involved a combination of impact and high-effort resistance exercises to maximise improvements in BMD and muscle function. [7][8][9] The impact exercises included rope skipping (up to 5 minutes) and several multi-directional jumps (eg squat jump, broad jump, scissor jump; 2-3 sets of 10-15 repetitions for five different jumps). The resistance exercises targeted the major muscle groups of the upper-body, lower-body and mid-section (eg squat, lunge, press-up, reverse fly, lateral raise, bicep curl, triceps extension, bridge; 2-3 sets of 10-15 repetitions for 8-10 exercises) with resistance provided using the participant's own body weight and TheraBand® elastic bands. The intended intensity for each set was moderate-to-hard, which was self-rated using the Resistance Intensity Scale for Exercise. 15 The cool-down lasted approximately 5 minutes and included static stretches for the major muscle groups.
Each exercise participant was given four TheraBand elastic bands and a skipping rope to keep. They also received an exercise booklet (see Supporting Information) that included information about how to perform and progress the exercises, tables for self-monitoring of adherence, information about TheraBand care and travelling to the University, and contact details of the research team. Participants were free to contact the intervention facilitator (KJ) by telephone or email at any time with questions.
We had planned to use a telehealth app (Florence; www.getfl orence.co.uk/) to send the exercise participants regular motivational messages to support exercise adherence, but this was unavailable during the period of intervention delivery. The intervention facilitator did however contact these participants every 4 weeks to provide motivation and support.
Face-to-face assessment visits were conducted before (baseline), and 3 and 6 months after group allocation. Demographic and clinical characteristics were assessed at baseline only. The following outcomes were assessed at baseline and both follow-up time-points: muscle function, body mass, stature, resting blood pressure and heart rate, health-related quality of life, fatigue, physical activity, and medications. BMD was assessed at baseline and 6 months only.
Adverse events and exercise enjoyment were assessed at the two follow-up time-points. The acceptability of the intervention was also assessed using participant feedback via telephone interviews conducted after the 6-month assessments.
With local institutional approval, 33 healthy adults were also recruited from the local community to allow a case-control comparison of BMD, muscle function and quality of life outcomes using the trial participants' baseline data. Participants were matched for age (±5 years), gender, physical activity status, body mass index category and ethnicity. Healthy control participants completed the assessments twice, one week apart, to allow test-retest reliability statistics (two-way random effects intra-class correlation coefficient, ICC) to be calculated. ICC values of between 0.75 and 0.90 and greater than 0.90 were considered indicative of good and excellent reliability, respectively.

| Outcomes
The primary outcomes were BMD and muscle function at 6 months. Resting blood pressure and heart rate were assessed using an automated monitor (V100 Dinamap Vital Signs Monitor, GE Carescape, UK). Disease-specific health-related quality of life was assessed using the total score on the long version of the Inflammatory Bowel Disease Quality of Life Questionnaire (IBDQ). 17 Generic health-related quality of life was assessed using the EQ-5D-5L utility index. 18 Fatigue was assessed using the Fatigue Assessment Scale from the IBD Fatigue (IBD-F) Scale. 19 Physical activity was assessed as minutes per week of physical activity using the Scottish Physical Activity Questionnaire. 20 Exercise enjoyment was assessed using the total and sub-scale scores on the Physical Activity Enjoyment Scale. 21 We documented all serious adverse events, and all nonserious adverse events that were either deemed to be related to participation in the research or resulted in withdrawal from the exercise programme or study. An event was classed as serious if it met any of the following criteria: fatal; life threatening (ie event in which patient is at risk of death at the time of the event occurring); requiring unplanned or prolonged hospitalisation; resulting in persistent or significant disability or incapacity; or resulting in a congenital abnormality or birth defect. Nonserious adverse events were defined as any untoward medical occurrence that did not fulfil any of the serious adverse event criteria.

| Statistical considerations
We used the distribution-based approach for the sample size calculation because a minimum clinically important difference has not been established for BMD in adults with CD. A target difference of 0.4 SD (ie a small-to-moderate effect) was selected because this effect size was reported for femoral neck BMD in a meta-analysis of combined impact and resistance training programmes in post-menopausal women. 9 Using the sample size calculation methods of Borm et al, 22 and assuming 80% power, a two-sided α level of 0.05, and a pre-post correlation between of r = 0.9, we required 38 participants in total.
Allowing for 20% loss to follow-up at 6 months, we calculated that 50 participants were needed to be recruited and randomised.
We used descriptive statistics to show the baseline characteristics of the participants who were enrolled, by treatment group. The effect of the intervention was evaluated using separate analysis of covariance models for the 3-and 6-month outcomes. The outcome at 3 or 6 months was the dependent variable and group (intervention, control) was the independent variable. The baseline value of the outcome, gender and disease activity status were included as covariates.
The primary analysis used a modified intention-to-treat population that included all randomised participants for whom treatment was allocated as per the randomised list regardless of circumstances after randomisation and who had both baseline and follow-up outcome data (ie complete-case analysis). Best-case and worst-case sensitivity analyses were also done to explore the impact of missing data, using the group mean plus 1 SD of the group mean as a 'beneficial outcome' and the group mean minus 1 SD of the group mean as a 'harmful outcome'. 23 We report adjusted and unadjusted means for each group and the adjusted mean difference (with 95% CI and p value) between treatment groups at 3 and 6 months. All statistical tests were two-sided at the 5% significance level. In the primary analysis, p-values were corrected for multiple testing using the Benjamini-Hochberg procedure. 24 All analyses were performed using IBM SPSS Statistics (version 24, IBM Corporation, UK).

| RESULTS
Between February 20, 2018 and March 14, 2019, we screened 76 individuals, of whom 47 were recruited and randomly assigned to the exercise or control groups ( Figure 1). Forty-three (91%) participants completed the trial, achieving our target sample size. Table 1 shows the participant characteristics at baseline. Thirty-two (68%) participants were female and the mean age was 49.3 years (SD 13.0).
All participants were of white ethnicity, none were current smokers, and most had quiescent disease (66%) and were in paid employ- Two participants thought that thrice weekly was too often to fit in with their lifestyle, one thought that 60-minute sessions were too long, one thought that the sessions were too hard for their level of fitness, and two thought that the programme duration should be greater than 6 months. Table 2 shows summary statistics for primary and secondary outcome measures at each assessment time-point. Forty-three (91.5%) participants (n = 22 exercise group, n = 21 control group) provided complete outcome datasets at 3 and 6 months.
At 3 months, all muscle function outcomes were superior in the exercise group (Table 4). For example, adjusted mean differences were 12.9 Nm for isokinetic knee extension at 60°/s (95% CI 2.5-23.3; P = 0.016) and 5.2 Nm for isokinetic elbow flexion at 60°/s (95% CI 2.8-7.6; P < 0.001). There were no between-group differences at either time-point for systolic blood pressure, diastolic blood pressure or physical activity (Table 4)    The EQ5D utility index scores also indicated that, at baseline, trial participants had lower health-related quality of life (adjusted mean difference 0.098, 95% CI 0.032-0.165; P = 0.004).

| DISCUSSION
In this RCT involving adults with quiescent or mildly-active CD, the offer of a combined impact and resistance exercise programme with gradual tapering of supervision to independent practice improved BMD and muscle function at 6-month follow-up compared with usual care control. The intervention was also associated with improvements in generic health-related quality of life and fatigue at 6 months. The exercise programme required minimal direct supervision and appeared acceptable and safe for participants.
Consensus guidelines on the management of inflammatory bowel disease state that weight-bearing and resistive exercises should be promoted to help prevent bone loss. 4,26 However, specific exercise prescriptions are not provided, and the recommendation is based largely on research in the general population. 27 Findings from the general population are unlikely to be generalisable to people Skeletal muscle weakness is a peripheral manifestation of CD, 10,11 and a risk factor for falls and fall-related fractures, 32 mobility disability 33 and major post-operative complications. 34 Consequently, a reduction in the risk of these adverse outcomes may be achieved through exercise programmes that improve muscle function. We have shown that combined impact and resistance training enhances a broad range of muscle function outcomes, including measures of upper-and lower-body muscular strength and endurance, in adults with CD.
Therefore, the exercise programme seems to be suitable for improving BMD and muscle function simultaneously. It is also notable that the deficits in muscle function that were observed relative to healthy controls at baseline were no longer apparent at the 6-month follow-up.
Fatigue is a common and burdensome symptom in people with CD, even for those in clinical remission. 35 It can have a negative impact on personal and social life, on work and employment, and the ability to think clearly. 36 At present, there is no consensus on how to manage CD-related fatigue, which may be explained by its varied and incompletely-understood aetiology, and a lack of Although not a primary focus of this study, we also observed that the exercise programme had a beneficial effect on generic health-related quality of life assessed using the EQ-5D utility index ( Table 4).
Inspection of the summary data in Table 2 indicated that the between-group differences at follow-up were as much explained by a deterioration in quality of life in the control group as an improvement in the intervention group. This was an unexpected finding because the control group had unrestricted access to usual care and did not receive any specific advice against exercising. It also stands in contrast with both our IBDQ data (  In summary, a 6-month combined impact and resistance training programme improved BMD and muscle function in adults with CD.
The intervention appears a suitable model of exercise for reducing the future risk of osteoporotic fractures and physical disability in this increased-risk population.