Introduction

Alternative therapies for the treatment of inflammatory bowel disease (IBD) are under-researched and are therefore omitted from current best practice guidelines. Examples of alternative therapies include exercise interventions, dietary supplements, psychotherapies and body-based interventions including acupuncture. IBD remains an incurable condition that encompasses Crohn’s disease (CD), ulcerative colitis (UC) and indeterminate colitis and is characterised by relapsing and remitting clinical episodes [1,2,3]. Debilitating pathognomonic symptoms include diarrhoea, rectal bleeding, abdominal cramping and weight loss. Other significant symptoms include severe fatigue, arthralgia, low mood and anxiety which are commonly seen in other chronic diseases [4,5,6].

Disease-modifying medical therapies are continuously under investigation for the treatment of IBD, with novel mechanistic developments emerging regularly. However, biologic therapies remain at the forefront of treatment of moderate to severe IBD since the initial introduction of infliximab in 1997 [7,8,9,10]. A relatively high proportion of patients experience primary loss of response to biologic therapy and a further 40% experience secondary loss of response [11,12,13,14,15], most commonly seen in anti-TNF agents. This highlights a potential role for alternative or complementary therapies, including the routine use of personalised exercise programmes. There is an existing knowledge gap in whether there is a role for these strategies and how best to use them within the existing armamentarium. An ECCO topical review on alternative and complementary therapies in IBD, published in 2019, acknowledged that patients increasingly use alternative or complimentary therapies. Reportedly, up to 50% of patients with IBD are estimated to use different forms of complimentary or alternative medicine at some point in their disease course despite a lack of appropriate evidence and guidance [16]. To address this knowledge gap, we believe there is a need for an up-to-date systematic review on the topic of structured exercise programmes and their effect on IBD outcomes, in particular with a focus on the inflammatory response and changes in body composition.

Specific guidelines for exercise in patients with IBD are lacking; however, we note that following a review of published exercise studies, Eckert et al. [17] suggested that patients with IBD should participate in moderate physical activity at least 3 times a week, for at least 30 min a day, incorporating both resistance and aerobic training. This echoes the WHO physical activity recommendations, which advises that patients aged 18–65 years should carry out 150–300 min of moderate intensity or 75–150 min of vigorous-intensity aerobic physical activity a week [18]. The American College of Sports Medicine defines moderate-intensity exercise as ranging from 40 to 60% of maximal exercise capacity (e.g. brisk walking) and high-intensity exercise above 65% of maximal exercise capacity (e.g. jogging) [19]. Low intensity exercises include slow-paced walking, yoga and Qigong. Resistance and aerobic training are two of the most common forms of exercise training. Resistance training has been defined as the use of both static and dynamic muscle actions under tension with an external load (e.g. use of free weights, machine weight training) [20] and aerobic training as any activity that includes the use of large muscle groups that is rhythmic in nature and maintained continuously (e.g. walking, jogging, swimming) [21].

Regular exercise may serve to help regulation of inflammation and promotion of an anti-inflammatory state. Current evidence proposes three possible anti-inflammatory mechanisms of exercise, including (i) the reduction in visceral adiposity, (ii) an increase in production and release of anti-inflammatory cytokines from skeletal muscle and (iii) a reduction in the expression of monocyte and macrophage Toll-like receptors leading to subsequent inhibition of downstream pro-inflammatory cytokine responses [22]. The production of IL-6 (myokine) by contracting skeletal muscle during exercise stimulates the production of the anti-inflammatory cytokines IL-ra and IL-10 which in turn inhibit the production of TNF-alpha, the well-known pro-inflammatory cytokine in IBD (Fig. 1) [23]. Whilst TNF-alpha is a recognised potent trigger of inflammation in IBD, it is also believed to play a role in the metabolic syndrome with data suggesting it impairs insulin signalling, leading to insulin-resistant states [24,25,26]. Therefore, the IL-6 mediated suppression of TNF-alpha, as a product of regular exercise, offers protection against inflammation and may also lead to the reduction in TNF-alpha-associated insulin resistance [27]. Whilst a reduction in systematic inflammation by regular exercise has been observed in other chronic diseases, this has not yet been proven in IBD [28].

Fig. 1
figure 1

The effects of exercise on pro-inflammatory cytokines. Image created using Biorender

Full size image

Some data have proposed associations between body composition in relation to muscle mass in addition to adiposity and outcomes in IBD. Visceral obesity defined by an abnormally high deposition of fat around the intra-abdominal organs is associated with lower degree of response to biologics in CD, stricturing disease with higher rates of surgery and also post-operative recurrence [29,30,31,32,33,34,35]. Sarcopenia is classically associated with older persons and is characterised by both a loss of skeletal muscle mass, strength and performance [36]. Sarcopenia may be present in over 40% of IBD patients and is associated with disease severity and need for rescue therapy in acute severe colitis in addition to the need for surgery in CD [37, 38]. Patients with IBD often lead more sedentary lifestyles due to their disease burden and debilitating fatigue symptoms which puts them at risk of developing sarcopenia due to the lack of muscle training and increases their risk of developing higher levels of visceral adiposity.

We aim to carry out a systematic review to examine the effects of structured exercise on the inflammatory response in IBD patients. Our secondary aim is to examine the effect of structured exercise programmes on body composition given both an increase in visceral obesity and the presence of sarcopenia have deleterious effects on outcomes in IBD patients. We will also include an overview of the included exercise studies in patients with IBD in our reporting to ensure a comprehensive review and to provide a platform for a well-rounded discussion. The study protocol has been registered with PROSPERO (CRD42022310035).

Methods

Study selection

Randomised controlled trials, pilot studies, observational studies, case–control or cross-sectional studies investigating the effects of structured physical activity programmes were included. All forms of exercise were included once they formed part of a formal intervention. Studies that included once-off exercise events were included if they measured relevant outcomes. Studies from 1980 onwards were included (see Table 1). The Methodological Expectations of Cochrane Intervention Reviews (MECIR) manual and the Cochrane Handbook for Systematic Reviews of Interventions were followed during the review [39].

Table 1 Study inclusion and exclusion criteria
Full size table

Types of participants

Patients with confirmed ulcerative colitis, Crohn’s disease and indeterminate colitis were considered for inclusion (Table 1). All studies included were carried out in adults > 18 years old; paediatric and adolescent studies were excluded.

Search methods for identification of studies

Electronic searches

The databases Embase, PubMed/MEDLINE and Cochrane Central Register of Controlled trials (CENTRAL) for applicable RCTs were searched from August to December 2021 in accordance with the PRISMA guidelines. Two researchers (NMG and KA) carried out the searches for each database. The Boolean search method was used on PubMed using ‘OR’ and ‘AND’. The PubMed strategy included the use of both MeSH terms and Title/Abstract. All the exercise terms were grouped together with Title/Abstract OR MeSH Terms, followed by the IBD terms and then the exercise terms were combined with the IBD terms using “AND”.

This search was as follows: “exercise”[Title/Abstract] OR “exercise”[MeSH Terms] OR “exercise movement techniques”[MeSH Terms] OR “physical activity”[Title/Abstract] OR “physical and rehabilitation medicine”[MeSH Terms] OR “resistance training”[Title/Abstract] OR “resistance training”[MeSH Terms] OR “resistance exercise”[Title/Abstract] OR “aerobic training”[Title/Abstract] OR “aerobic exercise”[Title/Abstract]) AND (“inflammatory bowel disease”[Title/Abstract] OR “inflammatory bowel diseases”[MeSH Terms] OR “colitis”[MeSH Terms] OR “ulcerative colitis”[Title/Abstract] OR “crohn s disease”[Title/Abstract] OR “crohns disease”[Title/Abstract] OR “IBD”[Title/Abstract].

Our search protocol was agreed upon by both NMG and KA. The protocol was reviewed and approved by a clinical librarian in the Royal College of Surgeons Ireland (RCSI).

Searching other resources

To identify other trials, we:

  • Examined the reference lists of included studies and review articles for additional citations

  • Searched ongoing trials and research registers including ClinicalTrials.gov (www.clinicaltrials.gov) ‘exercise’, ‘physical activity’ and ‘inflammatory bowel disease’ or ‘ulcerative colitis’ or ‘Crohn’s disease’

  • Contacted trial authors to identify further published and unpublished trials and asked if they were willing to disclose their unpublished data

  • Searched published abstracts from conference proceedings, including the European Crohn’s and Colitis Organisation Congress, Digestive Disease Week and United European Gastroenterology Week

  • Searched relevant journals (Journal of Crohn’s and Colitis; Inflammatory Bowel Diseases; Gastroenterology, Alimentary Pharmacology and Therapeutics)

Data collection and analysis

Selection of studies

Two independent review authors (NMG and KA) screened and examined the eligibility of the titles and abstracts identified by the search based on the predetermined inclusion criteria described above. Full‐text papers were sought for all studies meeting the inclusion criteria and were reviewed by two independent review authors (NMG and KA).

Where data was available for IBD participants, trials with a heterogeneous sample of disorders were included. Authors of included studies were contacted regarding information that was unclear or missing. If a disagreement occurred about the selection of a study, arbitration was sought from a third author with content expertise (KB) and a final inclusion decision made by consensus.

Data extraction and management

For each included study, NMG and KA independently extracted and documented the relevant data using standardised data extraction forms. All trial authors were contacted to provide additional or unpublished material where relevant. Any disagreements regarding inclusion or exclusion were resolved through arbitration with a senior author (KB) as necessary.

Measurement of risk of bias and quality assessment

The GRADE (Grades of Research, Assessment, Development and Evaluation) online assessment tool was used to analyse the certainty of evidence for each outcome (both body composition changes in response to exercise and inflammatory response to exercise). The principal domains reviewed by NM and KA were study design, consistency of effect, imprecision, indirectness, inconsistency and publication bias with the overall certainty graded as high, moderate, low or very low.

The Cochrane ‘Risk of Bias’ (RoB 2.0) tool was used to examine the risk of bias for randomised studies at individual study level for the main outcomes of inflammatory response and body composition changes and the risk of bias was determined as ‘low risk’, ‘some concerns’ and ‘high risk’ [40]. The ‘Risk of Bias In Non-randomised studies of Interventions’ (ROBINS-I) was used for non-randomised studies with each domain classified as ‘low’, ‘moderate’, ‘no information’, ‘serious’ or ‘critical risk’ [41]. The overall judgement was based upon the highest classified risk; therefore, if one domain was high risk, the overall risk of bias was deemed as high.

Results

Exercise in IBD

All studies identified for this review are summarised in Table 2. For the purpose of the review, an overview of all included studies is given, and studies have been further stratified to focus on the effect of exercise on inflammatory pathways and body composition in patients with IBD.

Table 2 An overview of results of included studies
Full size table

Study characteristics

In total, 1516 records were screened for eligibility, and 148 records were reviewed for eligibility (full text), of which 16 were included and a further 7 studies were identified from hand searching references (Fig. 2). Of the 23 studies, 9 studies examined the effects of structured exercise in CD [42,43,44,45,46,47,48,49,50], 3 studies examined the effects in UC [51,52,53], and 11 studies examined the effects in unspecified patients with IBD [54,55,56,57,58,59,60,61,62,63,64]. Patients were in clinical remission or had mildly active disease in 15 studies [42,43,44,45,46,47,48, 50,51,52,53,54, 57, 59, 61]; two studies included patients with moderate disease [55, 64]; in 4 studies, the disease activity was unspecified [49, 56, 58, 63]; and 2 studies excluded patients on biologic therapies [60, 62]. Of the studies included, 11 were randomised controlled trials [42, 43, 45, 51, 52, 54, 55], 2 were pilot RCTs [46, 47], and the remaining 10 studies were observational (prospective cohort or case-control) [44, 48, 49, 56, 57, 60, 62, 63].

Fig. 2
figure 2

Flow diagram of included studies

Full size image

Exercise programmes and overview of outcomes measured

The types of structured exercise programmes ranged from combined aerobic and resistance training (n = 7) to walking (n = 6), running (n = 1), yoga (n = 3), high and moderate intensity interval training (n = 6), Qigong (n = 1) and swimming (n = 1). The outcomes of the studies varied from assessments of quality of life, disease activity, inflammatory response, body composition, physical fitness, mental health outcomes, feasibility measurements and fatigue. Programmes that included combined aerobic and resistance training demonstrated an increase in physical fitness and high compliance rates and were deemed safe with no significant deterioration in disease activity [45, 54, 57]. Both high and moderate-intensity interval training were deemed to have acceptable safety profiles [46, 52, 55] in addition to walking interventions which also led to an improvement in patient quality of life [42, 44, 64]. Yoga and Qigong [49, 51, 58, 61] were largely beneficial for improving psychological symptoms and quality of life without demonstrating any specific benefit in terms of disease control. Higher drop-out rates were seen in the study by Seeger et al. in the endurance exercise group due to “loss of motivation’ and “bad weather” [47].

Duration of the exercise programmes

The duration of exercise programmes varied from a once-off treadmill test to a year-long structured programme [43, 50]. The majority of programmes were from 8 to 12 weeks in duration [42, 44, 46, 47, 51,52,53,54,55,56,57, 59, 61, 64]. The majority of studies of 12-week duration or more found an improvement in quality of life and fatigue scores were measured [42, 51, 57,58,59, 64]. A number of studies reported once-off exercise events including walking marches, marathons, triathlons and a treadmill test [48, 50, 60, 62, 63] with outcomes focusing on the inflammatory response to the exercise which will be discussed later (Table 4). The longest study of 52 weeks reported a difference in compliance of 7% less at 1 year for the exercise arm versus controls; however, the compliance rate remained high at 88% for the exercise group [43].

Body composition analysis in exercise studies in IBD patients

Of the studies included in this systematic review, 4 studies examined the effect of structured exercise on body composition (Table 3). The use of dual energy X-ray absorptiometry (DEXA) was the most commonly used imaging modality for the determination of body composition measurements [43, 45, 54]. The 3 studies included which used DEXA were all randomised controlled trials in design. The focus of body composition measurements differed between studies with Cronin and colleagues focusing on total fat mass, truncal fat mass, body fat percentage and total lean tissue mass, whereas publications by Robinson et al. and Jones et al. focused on bone mineral density measured by DEXA [43, 45, 54]. Cronin and colleagues carried out an 8-week aerobic and resistance exercise programme of moderate-intensity in patients with IBD with quiescent or mild disease. The primary outcome of the study was to determine a change in body composition of patients participating in the exercise programme. Results of the study found that favourable changes in body composition were observed in the exercise group with a median decrease in total body fat percentage of 2.1% versus a mean gain of 0.1% body fat in the control group (p = 0.022). In addition to the changes in body fat percentage, an increase in total lean tissue mass of 1.59 kg was observed in the exercise group compared to 1.38 kg among controls (p = 0.003) [54].

Table 3 Results of body composition analysis
Full size table

Both Robinson et al. and Jones et al. aimed to examine the effect of structured exercise on bone mineral density (BMD) using DEXA. Robinson and colleagues aimed to determine the effect of low impact exercise on BMD over 12 months by dynamically loading the axial skeleton and the hip and lumbar spine with at-home, low impact exercises (predominantly core-based exercises). BMD was measured at the hip and lumbar spine at baseline and at 12 months. Although a non-significant increase was observed in BMD at the hip and spine in the exercise group versus controls (p = 0.05), the gain in BMD at the greater trochanter was significantly superior to that of controls (difference in means, 4.67; 95% confidence interval: 0.86–8.48; p = 0.02). The greater number of exercises carried out, the greater the gain in BMD (femoral neck: r2 = 0.28; 95% confidence interval: 0.10–0.45; p = 0.04) [43]. In comparison, Jones et al. used combined impact and resistance training to illicit changes in BMD in patients with stable CD. Exercises included were of a higher impact compared to that utilised by Robinson et al. A 6-month period was assigned for follow-up, and at this timepoint, BMD values were observed to be superior in the exercise group at the lumbar spine (adjusted mean difference 0.036 g/cm2, 95% CI 0.024–0.048; p < 0.001), but not at femoral neck (0.018 g/cm2, 0.001–0.035; p = 0.059) or greater trochanter (0.013 g/cm2, − 0.019 to 0.045; p = 0.415) after correcting for multiple outcomes.

DeSouza et al. adapted a different approach to the inclusion of body composition measurements by measuring thigh circumference manually in response to progressive quadriceps resistance training in female patients with IBD. 19 patients were included (10 = CD, 9 = UC) who carried out an 8-week machine-based quadricep training programme. An increase in thigh strength greater than 40% was observed; however, thigh circumference did not increase from baseline to the time of programme completion at 8 weeks. The authors suggest that the increase in thigh strength was likely relative to neural adaptation rather than muscle hypertrophy itself [56].

Inflammatory response to exercise in IBD

In total, we include 14 studies which objectively measure the inflammatory response to structured exercise programmes (Table 4). Seven studies examined the response of C-reactive protein (CRP) to exercise [46, 47, 51, 54, 55, 58, 59], and only one study, which also included patients with lymphocytic colitis, demonstrated a statistically significant improvement in CRP after Qigong from baseline, with the CRP remaining elevated post the exercise intervention (730.0 versus 836.0 ng/mL, respectively, p = 0.01) [58]. We note that faecal calprotectin (FCP) response was not observed in the same study [58]. Similarly, in the vast majority of other studies examining the effects of exercise on FCP, no significant improvement was observed in response to the studied intervention [46, 47, 51, 55, 63]. On the contrary, in the study by Tew et al., the FCP was elevated in two patients; however, one of these patients already had an elevated FCP in the few months prior to the exercise intervention [46].

Table 4 Results of inflammatory response to exercise in IBD patients
Full size table

Four studies examined the effects of exercise (prolonged walking, running and combined aerobic-resistance training) on pro-inflammatory cytokine levels, of which three studies carried out a formal analysis including of IL-6, IL-8, IL-10, IL-1β and TNF-α pre and post exercise [50, 54, 60]. Lamers et al. found that there was no difference in IL-6, IL-8, IL-10, IL1b or TNF-a in a mixed linear model but that a transient increase in both IL-6 and IL-10 was observed between IBD-walkers and non-IBD non-walkers at baseline and post day 1 exercise during a 4-day walking march. The authors conclude that exercise of moderate intensity caused a rise in the pro-inflammatory cytokine expression rather than IBD disease activity itself [60].

More novel markers of inflammatory activity measured in other studies included T-lymphocyte subsets, serum eosinophilic cationic proteins, faecal polymorphonuclear neutrophil elastase, faecal lactoferrin and markers of immunometabolism (plasma endotoxin, oxidative stress) [46, 50,51,52, 61, 62]. Studies by Elsenbruch et al. and Tew et al. included the effects of their exercise programmes on circulating T lymphocytes [46, 52]. Elsenbruch et al. examined an array of lymphocyte subsets in UC patients in remission in response to a moderate 10-week exercise programme and found no substantial effect of the exercise therapy on helper T cells (CD3 + CD4 + lymphocytes), cytotoxic/suppressor T cells (CD3 + CD8 + lymphocytes), natural killer (NK) cells (CD3– CD16 + CD56 + lymphocytes), B-cells (CD3–CD20 + lymphocytes) or monocytes (CD14 + leukocytes) [52]. To date, formal analysis on circulating T lymphocytes has not been completed.

Cramer and colleagues did not observe a change in these faecal markers during a 24 week yoga programme in UC patients who were in clinical remission [51].

Sharma et al. carried out an observational study examining the effects of an 8-week yoga programme on IBD patients which included the observation of the change in both serum eosinophilic cationic proteins and soluble interleukin-2 receptors (induce T-lymphocyte cytotoxicity and stimulates natural killer cell activity [65]) in response to the programme. Neither markers changed significantly as a result of the programme which included 100 IBD patients in clinical remission.

Finally, Benazatto et al. observed the effect of a once-off ‘moderate intensity’ treadmill exercise on antioxidant status in patients in remission with Crohn’s disease. Plasmatic endotoxin, total plasmatic antioxidant status and total oxidative status were assessed by enzyme-linked immunosorbent assay (ELISA), chromogenic limulus amoebocyte lysate (LAL) and spectrophotometry, respectively, at rest and after exercise. Whilst a higher intestinal permeability to lactulose/mannitol was observed in patients with CD versus healthy controls, no significant difference was recorded in plasma endotoxin levels between the two groups [50].

The type and duration of exercise differs significantly between the studies included. The type of exercise varies from walking to yoga, running, Qigong and interval training. The inclusion of IBD disease-types also differs between studies, with some including all IBD patients and others limiting the studies to either UC or CD. Furthermore, the inflammatory markers measured differ significantly between studies. For all studies excluding 1, patients were in remission or had mildly active disease which may have impacted on results of inflammatory markers in the absence of clinically active disease states [55]. Hence, significant heterogeneity exists between studies, and therefore, it was not possible to carry out a meta-analysis.

Risk of bias and quality assessment

Risk of bias

The risk of bias was assessed using the ROB 2.0 tool for 11 RCTs which were all deemed to have some concerns for bias for both inflammatory and body composition outcomes [43, 45,46,47, 51, 52, 54, 55, 58, 59, 61] (Figs. 3 and 4). The absence of pre-published study protocols for a number of studies meant there may be bias due to deviations in the intended interventions. Not at all studies followed the intention-to-treat principle. The lack of blinding of investigators in some studies may also have led to an increase in bias.

Fig. 3
figure 3

Risk of bias graph for inflammatory response to exercise in RCTs

Full size image
Fig. 4
figure 4

Risk of bias graph for body composition in RCTs

Full size image

The ROBINS-I tool was used to evaluate the risk of bias in the included observational studies examining body composition and inflammatory response to measured exercise interventions. Six observational studies were included none of which were deemed to be of low risk of bias [48, 50, 56, 60, 62, 63]. Common areas for concern for bias were due to confounding and selection of participants with two studies being deemed as serious risk (Supplementary Figs. 1 and 2).

Quality of evidence

The GRADE (Grades of Research, Assessment, Development and Evaluation) online assessment tool was used to analyse the certainty of evidence for both outcomes (both body composition changes in response to exercise and inflammatory response to exercise). Results of body composition assessments in response to exercise using DEXA were reported in 3 separate randomised controlled trials with a risk of bias identified due to the lack of blinding given the nature of the intervention, single operator status and a significant difference in exercise type between studies. However, the overall quality of evidence was deemed to be moderate. Only one study by Cronin et al. used DEXA imaging to examine the effects of exercise on body and lean muscle composition [54]. The quality of evidence was also reviewed for the inflammatory response to the exercise interventions for 9 RCTs and 5 observational studies. Significant differences in patient selection and type of exercise were identified between the RCTs. Allocation concealment was often unclear, and there were significant differences in the type of inflammatory markers measured between studies; therefore, the level of certainty of evidence was deemed low (Table 5).

Table 5 GRADE assessment
Full size table

Discussion

Current peer-reviewed studies of the impact of exercise among IBD patients range from low intensity exercises such as yoga, Qigong and walking [42, 51, 58] to moderate intensity aerobic and resistance training [52, 54, 55] to high-intensity interval and impact resistance training [45, 46, 57, 63]. This systematic review was carried out to focus primarily on the inflammatory response to structured exercise in patients with IBD and body composition changes observed in response to structured exercise in this cohort of patients. Previous demonstrated benefits of exercise in patients with IBD include an improvement in quality of life, fatigue, cardiorespiratory fitness and confirmation of the safety of various exercises in mild to moderate disease [46, 47, 54, 59, 64]. A systematic review of the exercise-induced inflammatory response to IBD, published post registration of this current review, focused on 5 studies only and did not use the Cochrane tools for risk of bias [28]. A recently published systematic review and meta-analysis by Jones et al. reported an overview of physiological and psychological outcomes in adult IBD patients in response to exercise programmes of at least 4-week duration; hence, it did not include studies of shorter duration, which excluded the review of a number of studies that examine novel biomarkers of inflammation which is a strength of this current review.

Few studies included in the review examined the potential anti-inflammatory of exercise in addition to changes in body composition.

In a short follow-up period of 8 weeks, Cronin and colleagues demonstrated favourable changes in body composition in the exercise group compared with healthy controls, whereas no demonstrable change in pro-inflammatory cytokines was observed with exercise. Patients included were known to have quiescent or mild disease which may account for the lack of inflammatory response to the prescribed exercise in addition to the small sample size of patients enrolled. A reduction in body fat itself could account for an anti-inflammatory response and reduction in pro-inflammatory cytokines due to the pro-inflammatory nature of adipose tissue [30, 33, 66, 67]. We observed that the baseline BMI for active patients in this study was in the ‘overweight’ but not obese category with a mean BMI of 28 kg/m2. Previous studies have shown a relationship between higher BMI and raised CRP in CD [68] and the presence of visceral fat, fat wrapping and a raised visceral: subcutaneous fat ratio are associated with a more complicated disease course, higher postoperative complications and post-operative recurrence of CD [31, 69, 70]. Given the clinical importance of visceral and mesenteric fat in these patients, future research looking at the body composition changes associated with structured exercise should include these as standard body composition outcomes given that available studies did not address these parameters.

Both Robinson et al. and Jones et al. examined the effect of exercise on bone mineral density (BMD) in patients with CD. Robinson et al. observed that significant gains in BMD at the greater trochanter were statistically significant in the exercise group who carried out core exercises (p = 0.02), whereas the same was not observed at the hip and spine, potentially due to poor compliance with the exercise reported by the authors [43]. In comparison, Jones et al. demonstrated that combined impact and resistance training improved BMD at the lumbar spine (p < 0.001), but the same effect was not seen at the femoral neck or greater trochanter; however, muscle function outcomes were also superior in the exercise group versus controls (p < 0.001). Our results are similar to those reported in the review by Jones et al. [71].

A wide variety of inflammatory biomarkers were assessed in the current exercise studies in IBD. Given the majority of studies included patients with mild or quiescent disease, it is unsurprising that statistically significant changes in biomarkers in response to exercise were rarely observed. Only one study demonstrated a change in CRP, and it is unlikely the observed reduction in CRP was as a result of the exercise intervention [58]. The inclusion of studies of short duration or once-off events allowed for the review of a number of novel inflammatory biomarkers. Faecal polymorphonuclear neutrophil elastase and faecal lactoferrin have been shown in prior studies to correlate with the clinical activity index (CAI) score and to differentiate mucosal healing versus endoscopic inflammation in UC [72]. Eosinophil infiltration in the lamina propia has been included in the Geboes histological score for UC as eosinophils have been identified as significant contributor cells to immune infiltration in IBD [73]. Additionally, one of the hallmarks in the early diagnosis of IBD is the presence of eosinophilia-associated basal plasmacytosis [74]. Reports suggest that circulating endotoxin as a result of increased intestinal permeability is found in IBD and associated with increasing disease severity [75]. Novel biomarkers of inflammation included in studies in this review were pro-inflammatory cytokines, circulating T-lymphocyte subsets, serum eosinophilic cationic proteins, faecal polymorphonuclear neutrophil elastase, faecal lactoferrin and markers of immunometabolism [51, 52, 54]. Whilst the studies included did not demonstrate significant changes in the biomarkers investigated as a response to exercise, they are useful to include to inform biomarker panel development for future studies in the appropriate patient population. We suggest that the exclusion of patients with active disease, including those on biologic therapy, and the duration and nature of the exercises (yoga and walking) are the most likely factors leading to a lack of significant response of inflammatory biomarkers to the exercise programmes.

The duration of exercise programmes is an important factor to consider in order to demonstrate positive effects on inflammation, body composition and habit formation. We previously noted that studies on the effect of exercise on the microbiome that were of too short duration did not demonstrate significant changes in gut microbiome composition, diversity and functionality [76]. The study by Cronin et al. which was of 8-week duration, did not show any significant difference in gut microbiome composition/diversity or in pro-inflammatory markers in response to the exercise programme [54], whereas exercise studies in other populations of at least 12-week duration have demonstrated positive microbiome effects [77, 78]. Given the proven safety of a variety of exercises in patients with IBD, we suggest future studies examining the effects of exercise in IBD should be of at least 12-week duration and of moderate intensity to comply with the current World Health Organization guidelines on physical activity [18].

Limitations of the studies reviewed include significant differences in type and duration of exercise, the inclusion of predominantly patients who were in remission or mildly active disease, heterogeneity between inflammatory markers measured and a lack of blinding of participants and investigators. It is clear that participants carrying out the exercise intervention cannot be blinded to the intervention; however, this may lead to bias when participants complete patient-reported outcome measures; therefore, standardised biomarkers would be useful in addition to blinding of investigators who are interpreting the results. Small samples sizes were seen in the majority of studies and in the pilot studies included which lead to concern regarding adequate power of the studies and may explain the lack of significant changes particularly in inflammatory markers in response to the interventions. We acknowledge that the authors of the individual studies published the analysed results, despite often a lack of favourable outcomes, thus reducing the presence of publication bias.

The strengths of the review are the inclusion of novel studies examining inflammatory biomarkers in response to exercise in IBD, the use of robust reporting and search tools including the Cochrane guidelines (including the Cochrane risk of bias reporting tools for both RCTs and observational studies) and the MECIR manual and the PRISMA search guidelines; registration of the protocol a priori and our reporting was strengthened by the participation of a librarian from our institution’s library. Limitations of the study include a search using the English language only, the exclusion of studies in adolescents may have led to the exclusion of data relevant to the adult IBD population, and given the heterogeneity of the type of study, interventions and outcomes, a meta-analysis was not carried out.

Conclusion

The quality of evidence for a significant inflammatory response to the current exercise studies in IBD is low. Adequately powered studies of adequate duration with the inclusion of patients with active disease are required to assess disease-altering changes in the microbiome which in turn may yield anti-inflammatory effects through the modulation of immuno-metabolic pathways. The development of a biomarker panel would help to standardise reporting of the inflammatory response in future studies.

This review has shown a moderate level of evidence supporting the positive body composition changes as a result of exercise in patients with IBD which supports the findings of previous reviews, although very few studies examined these changes and further adequately powered studies are required.

The majority of studies included patients with mild or quiescent IBD; hence, further studies are required to include patients with more active disease, including those on biologic or small molecule therapies, which is more reflective of real-world patients. In conclusion, the results of this review suggest limited evidence to support the anti-inflammatory effect of exercise programmes in patients with IBD based on data published to date but demonstrate the positive effect of exercise on body composition. Exercise programmes in IBD remain a promising area for future research and as an adjunctive therapy in inflammatory bowel disease management.