Post‐diagnosis physical activity and sedentary behaviour and colorectal cancer prognosis: A Global Cancer Update Programme (CUP Global) systematic literature review and meta‐analysis

Low physical activity and high sedentary behaviour have been clearly linked with colorectal cancer development, yet data on their potential role in colorectal cancer survival is limited. Better characterisation of these relationships is needed for the development of post‐diagnosis physical activity and sedentary behaviour guidance for colorectal cancer survivors. We searched PubMed and Embase through 28 February 2022 for studies assessing post‐diagnosis physical activity, and/or sedentary behaviour in relation to all‐cause and cause‐specific mortality and recurrence after colorectal cancer diagnosis. Total and recreational physical activity were assessed overall and by frequency, duration, intensity, and volume using categorical, linear, and non‐linear dose–response random‐effects meta‐analyses. The Global Cancer Update Programme (CUP Global) independent Expert Committee on Cancer Survivorship and Expert Panel interpreted and graded the likelihood of causality. We identified 16 observational studies on 82,220 non‐overlapping patients from six countries. Physical activity was consistently inversely associated with colorectal cancer morbidity and mortality outcomes, with 13%–60% estimated reductions in risk. Sedentary behaviour was positively associated with all‐cause mortality. The evidence had methodological limitations including potential confounding, selection bias and reverse causation, coupled with a limited number of studies for most associations. The CUP Global Expert panel concluded limited‐suggestive evidence for recreational physical activity with all‐cause mortality and cancer recurrence. Total physical activity and its specific domains and dimensions, and sedentary behaviour were all graded as limited‐no conclusion for all outcomes. Future research should focus on randomised trials, while observational studies should obtain objective and repeated physical activity measures and better adjustment for confounders.

with colorectal cancer morbidity and mortality outcomes, with 13%-60% estimated reductions in risk.Sedentary behaviour was positively associated with allcause mortality.The evidence had methodological limitations including potential confounding, selection bias and reverse causation, coupled with a limited number of studies for most associations.The CUP Global Expert panel concluded limitedsuggestive evidence for recreational physical activity with all-cause mortality and cancer recurrence.Total physical activity and its specific domains and dimensions, and sedentary behaviour were all graded as limited-no conclusion for all outcomes.
Future research should focus on randomised trials, while observational studies should obtain objective and repeated physical activity measures and better adjustment for confounders.
colorectal cancer survival, evidence grading, meta-analysis, physical activity, systematic review What's new?
The roles of physical activity and sedentary behaviour in colorectal cancer prognosis are unclear.
Here, as part of CUP Global, the authors performed a systematic review and meta-analysis on the associations of post-diagnosis physical activity and sedentary behaviour with all-cause and cause-specific mortality and recurrence among colorectal cancer survivors.Whilst the evidence was graded as limited, it suggested that a physically active lifestyle and avoidance of sedentary behaviour may be associated with longer overall survival.The study highlights the importance of more consistent and objective exposure assessment in future studies and the need for randomised trials to provide stronger evidence.

| INTRODUCTION
Colorectal cancer is the third leading cause of cancer incidence and the second leading cause of cancer-related mortality worldwide, with an estimated >1.9 million incident cases and >0.9 million deaths, respectively, in 2020. 1 The effectiveness of early detection via cancer screening programmes in reducing mortality [2][3][4] coupled with better cancer treatment options have led to increased survival rates among individuals diagnosed with colorectal cancer. 5The estimated 5-year survival for colorectal cancer is generally over 60% in western populations and high income countries, [5][6][7] and on average 20% lower in transitioning countries. 8Among patients diagnosed with earlier stages of the disease, survival rates are even higher, and the 5-year prevalence of the disease is estimated to be over 5.2 million people worldwide, making colorectal cancer the second most prevalent cancer, after breast cancer. 1 Extensive research has been conducted to identify risk factors affecting the incidence of colorectal cancer.In addition to genetic predisposition and increasing age, several environmental and lifestyle factors have been associated with the development of the disease.Among these, the protective role of physical activity on colorectal cancer risk, specifically on colon cancer, is well-established. 9,10The Third Expert Report from the World Cancer Research Fund/American Institute for Cancer Research (WCRF/AICR) concluded that there is convincing evidence that physical activity reduces colon cancer incidence, 9 while the 2018 Physical Activity Guidelines Advisory Committee (PAGAC) Scientific Report 11 graded this association as being supported by strong evidence.
Despite the abundance of evidence on colorectal cancer incidence, evidence regarding the association between physical activity and sedentary behaviour and colorectal cancer prognosis is relatively limited.A small number of meta-analyses investigating the role of post-diagnosis physical activity on outcomes after colorectal cancer [12][13][14][15][16][17][18] suggested an inverse association with all-cause mortality.
These meta-analyses have primarily focused on the association of categories of physical activity levels with colorectal cancer prognosis, but none explicitly assessed the different dimensions and domains of activity.Moreover, apart from one ongoing trial 19 there is a lack of randomised controlled trials (RCTs) assessing the effect of physical activity on colorectal cancer survival.The role of sedentary behaviour as a distinct risk factor from physical activity, defined as the energy expenditure of less than 1.5 metabolic equivalents of task (METs) while one assumes a sitting or reclining posture, 20 is even less studied. 21e to the lack of strong evidence, physical activity and sedentary behaviour recommendations targeting cancer outcomes tailored to cancer survivors are limited.Formal evaluation of the existing observational evidence base performed by PAGAC concluded moderate evidence strength for all-cause and cancer-specific mortality in colorectal cancer, based on individual studies published until February 2018. 11The 2018 American College of Sports Medicine (ACSM) Roundtable also concluded that physical activity is beneficial for survival 22 and other cancer-related health outcomes. 23However, the precise quantity of physical activity required to effectively reduce cancer-specific or all-cause mortality remains undetermined. 22The evidence on sitting time as a proxy of total sedentary time and cancer-specific mortality was limited. 22[26] The aim of WCRF International's new Global Cancer Update Programme (CUP Global) formerly known as the WCRF/AICR Continuous Update Project 27 is the development of disease-and stage-specific guidance and recommendations.Therefore, as part of CUP Global, we present the findings of a systematic literature review on the association of post-diagnosis physical activity and sedentary behaviour and colorectal cancer prognosis.[30]

| METHODS
This systematic review is part of the ongoing CUP Global programme and is reported following the Preferred Reporting Items for Systematic Reviews and Meta-Analyses (PRISMA) guidelines 31 (Supplementary Table S1).The protocol can be found elsewhere. 32

| Search strategy, selection criteria and data extraction
Comprehensive details on the methodology are available in the Supplementary Texts 1 and 2. Briefly, we searched PubMed and Embase through 28th February 2022.We included observational studies, RCTs, or pooled analyses of such studies evaluating postdiagnosis (at or shortly after diagnosis, during treatment, after the end of treatment) physical activity, sedentary behaviour, or their pre-to-post diagnosis or post-diagnosis changes.We included studies with at least 100 participants with primary colorectal cancer or its subtypes diagnosed during adulthood.The term 'colorectal' referred to any colorectal cancer survivor, regardless of cancer subtype, colon or rectum.Outcomes of interest included all-cause, colorectal cancer site and subsite-specific, and cardiovascularspecific mortality, second primary cancer, and cancer recurrence/ disease-free survival.Study selection was performed at title and abstract level, followed by full-text review, and supplemented with manual reference screening.If more than one publication from the same population was identified, we included the one with the higher number of events per outcome assessed.Data were extracted into the CUP Global database and included, among others, publication details, study characteristics (number of participants, events, and inclusion and exclusion criteria), exposure and outcome details, relative risks (RR) 95% confidence intervals (CI), standard errors, and/or p-values, and details on adjustment variables.Study selection and data extraction were performed by one reviewer and at least 10% checked by another; disagreements were resolved by consensus.

| Evidence synthesis
We reviewed total physical activity (defined as any combined activity domain, including occupational, recreational, household, transportation) and the specific domains, irrespective of which physical activity dimension was assessed (frequency, duration, intensity).Specifically for recreational physical activity, the most investigated domain, we further reviewed frequency, duration, and intensity (activity of any intensity and specifically moderate-to-vigorous intensity).We also investigated recreational physical activity volume as a separate dimension, measured in MET-h/week.Furthermore, we performed metaanalyses on sedentary behaviour when its definition was sufficiently consistent across studies (capturing similar types of sedentary behaviours).
We analysed cancer recurrence/disease-free survival as defined by the studies (Supplementary Table S2).When more than one of these outcomes were reported, the one with the highest number of events was included in the meta-analyses.
We performed DerSimonian-Laird random-effects meta-analyses 33 for categorical (highest vs. lowest levels) and linear doseresponse comparisons when at least two and three studies provided relevant information, respectively.When at least five studies provided sufficient information, we also performed one-stage nonlinear dose-response meta-analyses 34 using restricted cubic splines with knots placed at the 10th, 50th, and 90th percentiles of the exposure distribution.The nonlinear component was tested using a likelihood ratio test. 35 assessed the presence of small-study effects (such as publication bias) by visual inspection of the funnel-plots and used the Egger's regression asymmetry test, 36 when at least 10 studies were included in a meta-analysis.If small study effects bias was suggested, we estimated an adjusted summary RR using the trim-and-fill method. 37Heterogeneity was described by the between-study variation (tau 2 ) and by the range of study estimates in each meta-analysis.The I 2 metric of inconsistency 38 with cut-offs at I 2 = 30%, 50%, and 75% indicated low, moderate, high, and substantial proportion of total variability in effect estimates due to between-study heterogeneity.The 95% prediction intervals depicted the range of estimates likely to contain the value of a new study. 39 performed pre-defined stratified meta-analyses by sex, cancer anatomical subsite and stage, and molecular subtype (based on sequencing or immunohistochemistry approaches).We performed sensitivity meta-analyses by excluding, when possible, survivors with locally advanced and metastatic tumours, to explore potential impact of bias by reverse causation.We also performed leave-one-out sensitivity meta-analyses to assess the magnitude of the effect of exclusion of each study on the summary estimate. 40

| Risk of bias assessment
A modified version of the Risk of Bias for Nutrition Observational Studies (RoB-NObs) tool 41 was utilised to assess the risk of bias in the studies included in the meta-analyses.The tool was originally developed by the U.S. Department of Agriculture (USDA) Nutrition Evidence Systematic Review after modifications to the Cochrane's collaboration Risk Of Bias In Non-randomized Studies of Interventions (ROBINS-I). 42The tool underwent further refinement and testing by the Imperial College London (ICL) review team to ensure its suitability for investigating exposure-outcome associations in cancer survivorship studies.This involved adapting the tool's prompting questions and providing additional guidance to encompass adiposity, physical activity, and dietary/nutritional exposures (the working document version dated 11/07/2023 can be found in Supplementary Table S3).
The tool consists of seven domains, including confounding, participant selection, exposure classification, departures from intended exposures, missing data, outcome measurement, and selective reporting.In the context of cancer survivorship, we have specifically designated age, adiposity, and cancer stage and treatment as pivotal confounding variables a-priori of the studied associations, due to their substantial impact on both the levels of physical activity and sedentary behaviour engagement along with their recognised role as robust prognostic factors for survival in colorectal cancer patients.

| Evidence grading criteria
The CUP Global independent Expert Committee on Cancer Survivorship and Expert Panel convened by WCRF International interpreted the findings, independently of ICL.The Expert Committee made the preliminary conclusions, and the Expert Panel made the final conclusions.The quality of the evidence was graded for all exposures as strong (subgrades evaluating likelihood of causality: convincing, probable, substantial effect on risk unlikely) or limited (subgrades evaluating likelihood of causality: limited-suggestive, limited-no conclusion) according to a list of pre-defined criteria assessing the quantity, consistency, magnitude, and precision of the summary estimates, evidence of biological gradient, the study design, risk of bias, generalisability, and biological plausibility of the results (Supplementary Table S4).The grades of the quality of the evidence reflect the independent Expert Committee's and Expert Panel's confidence that the association estimates are correct.

| RESULTS
A flowchart of the selection process is presented in Figure 1.We excluded 453 publications on exposures other than physical activity, 64 on pre-diagnosis-only exposures,  seven on a mixture of preand post-diagnosis exposures, [107][108][109][110][111][112][113] and one where the outcome was colorectal polyps. 114 Ovrall, 28 publications met the eligibility criteria.
When excluding survivors with locally advanced and metastatic tumours, the summary RR for all-cause mortality was similar but with lower heterogeneity than when including all colorectal cancer patients (RR = 0.70; 95% CI = 0.56-0.87;I 2 = 21%; tau 2 = 0.09; RRs range = 0.53-0.80;four studies, three publications) (Supplementary Figure S2).The leave-one-out sensitivity analyses did not identify any influential studies (Supplementary Figure S3).

| Recreational physical activity
Twelve studies (12 publications) were identified. 122,124,125,128,130,132,135- 139,1419][140] Four studies measured duration of recreational activity 124,128,137,138 and three measured physical activity frequency. 122,125,141The primary analysis on recreational physical activity included all studies, irrespective of the dimension of activity measured, however, due to heterogeneity across definitions and dimensions measured, meta-analyses are also presented by specific dimensions.
Since studies often included patients with various cancer stages, the predefined subgroup meta-analyses by stage could not be performed.When excluding survivors with locally advanced and metastatic tumours, the summary RR and heterogeneity for all-cause mortality were similar to the results including all colorectal cancer patients (RR = 0.65; 95% CI = 0.56-0.75;tau 2 = 0.02; RRs range = 0.41-0.78; 122,124,125,128,130,132,135,136(Supplementary Figure S7).The leave-one-out sensitivity analyses did not identify any influential studies (Supplementary Figure S8).

| Recreational physical activity frequency
Three studies (three publications) were identified. 122,125,141The highest versus lowest levels of physical activity frequency showed an inverse association with all-cause mortality (RR = 0.77; 95%
Dose-response meta-analyses were not possible due to limited data.
Sequential omission of each study did not change the results (Supplementary Figure S19).
Non-linear dose-response analysis was only possible for all-cause mortality.The evidence from six studies was suggestive of a non-linear inverse association (P non-linearity = .07;2408 deaths) 130,132,135,136,138,139 indicating a consistent inverse association up to 20 MET-h/week and remained relatively constant for higher activity levels (Supplementary Figure S20).There was no strong evidence of non-linearity (P non-linear- ity = .27;four studies; 1075 deaths) when excluding survivors with locally advanced and metastatic tumours, but the data were sparce 130,132,135,136 (Supplementary Figure S21).

| Other types of physical activity
A pattern of inverse associations with all-cause mortality was observed across the different types of physical activities analysed (walking, bicycling, gardening, housework, and sport) (RRs ranged from 0.34 to 0.90, with the 95% CI not including the null in two out of the six comparisons) 127,138 (Supplementary Figure S25).No evidence of association was observed in the only study analysing walking and cancer recurrence. 138

| Physical activity change
Two studies (two publications) 133,136 were identified, which investigated changes in physical activity levels around the time of diagnosis.
A meta-analysis was not performed due to large methodological differences, but we descriptively synthesised these results.In one study, 133

Sedentary behaviour
The body of evidence is weak showing higher risk of all-cause mortality and indicating a higher risk for colorectal cancer-specific mortality with higher levels of sedentary behaviour, which may be influenced by reverse causation and other biases.
Sitting recurrence, and to recreational physical activity dimensions (volume, frequency, intensity, duration), total physical activity, change of physical activity, and sedentary behaviour.

| DISCUSSION
This systematic literature review showed that higher levels of recreational physical activity and its different dimensions were consistently inversely associated with major clinical outcomes in colorectal cancer patients.Due to methodological limitations leading to increased likelihood of biases, small number of studies, and often high betweenstudy variability due to heterogeneity, the CUP Global independent Expert Committee and Expert Panel concluded that the evidence on post-colorectal cancer diagnosis physical activity and sedentary behaviour and survival-related outcomes was limited.The association of recreational physical activity with all-cause mortality that was supported by a relatively larger number of studies and without high levels of between-study heterogeneity, along with the evidence on cancer recurrence that was supported by relatively more consistent studies in terms of population characteristics and timing of exposure measurement reached a limited-suggestive conclusion for the likelihood on causality.
The evidence on recreational physical activity volume also suggested an inverse non-linear relationship with all-cause mortality up to 20 MET-h/week (roughly equivalent to 5 h of moderate-intensity physical activity/week 142 ) compared to zero MET-h/week, which plateaued for higher volumes, although no evidence of non-linearity was found in a sensitivity analysis excluding survivors with locally advanced and metastatic tumours.There was less available evidence for total activity, dimensions of recreational activity, change in activity, and sedentary behaviour on survival-related outcomes and no conclusion could be reached.The evidence base was limited and did not allow for comprehensive evaluation of most of the a-priori defined subgroup analyses.
A plethora of mechanisms have been proposed to explain the link between physical activity and colorectal carcinogenesis, 143 such as changes in the gastrointestinal transit time, immune function, bile acid secretion, lipid and hormone profiles, and insulin and insulin-like growth factor (IGF) alterations.Some of these mechanisms including changes to the IGF-axis and immune response have also been hypothesised to affect colorectal cancer survival, albeit supported by less data, [144][145][146] while new proposed pathways of action include exercise-induced shear stress, extracellular vesicles, and systemic milieu alterations. 144Increased blood pressure and heart rate during exercise leads to higher blood flow and shear pressure, 147 which may reduce the number of circulating tumour cells by immediate cell necrosis promotion. 148,1491][152] For example, exercise leads to increased levels of serum adiponectin and decreased levels of insulin and leptin, which may lead to reduced proliferation and enhanced apoptosis of tumour cells, although evidence specifically in colorectal cancer patients is currently limited. 153Extracellular vesicles, small membrane-surrounded structures which transfer molecules across cells, are an emerging target for treatment of various cancers. 154,155Physical activity enhances skeletal muscle mass, which increases extracellular vesicle secretion in blood, leading to increased circulation of beneficial myokines, cytokines and peptides released from muscle cells, such as interleukin-6 (IL-6).Exerciseinduced myokine secretion has been shown to have an antiinflammatory role in various chronic diseases. 156Physical activity also promotes immune cell mobilisation, which, coupled with increased exercise-induced IL-6 circulation, may lead to increased immune cell tumour infiltration and enhance anti-cancer immunity. 1579][160] Despite the inconsistent response of IGFs to physical activity in healthy individuals, limited evidence supports the beneficial role of exercise on colorectal cancer mortality via IGF pathway alterations. 65,66Nevertheless, the diverse mechanisms though which physical activity is linked to outcomes after cancer may be influenced by the interplay between activity and treatment in cancer survivors, while the existence of comorbidities may also affect the observed effects through similar biological pathways.
To the best of our knowledge, our analysis is the most comprehensive evaluation of the existing body of evidence on post-diagnosis physical activity and sedentary behaviour and colorectal cancer outcomes.Our results are in line with the findings of previous reviews and meta-analyses of physical activity in the field [12][13][14][15][16][17][18] in both direction and magnitude.However, in the previous meta-analyses, studies assessing total and recreational physical activities were frequently combined.The most recent analysis pooling studies of both postdiagnosis total and recreational physical activity found a 37% and a 21% reduced risk of all-cause mortality for highest versus lowest levels (10 studies, any dimension) and for 10 MET-h/week increase (seven studies), 18 respectively.Those results were similar in magnitude to ours on recreational physical activity volume (40% and 24%, respectively).Half to two-thirds of the studies within these metaanalyses employed validated instruments for the quantification of physical activity, reducing the probability of exposure misclassification and consequently enhancing the reliability of the findings.Change in physical activity has been previously meta-analysed using the same two identified studies. 15In the present work, due to the high clinical and methodological heterogeneity across studies, including largely different definitions and timeframes of assessment, we decided against a quantitative synthesis.We identified one meta-analysis on postdiagnosis sedentary behaviour and colorectal cancer-specific mortality showing similar results to ours, however sitting while watching TV and overall sitting time were analysed together. 21art from the updated evidence base, our review is the first to assess specific physical activity domains and dimensions in relation to colorectal cancer outcomes to increase homogeneity of exposure assessment, as well as to evaluate non-linearity.Furthermore, the totality of the evidence was evaluated by the CUP Global independent Expert Committee and Expert Panel following a rigorous evaluation using pre-defined standardised grading criteria.Nevertheless, the results presented should be interpreted considering certain limitations.Cancer survivorship studies of observational nature are subject to a range of methodological limitations, survivor bias for instance, which could not be excluded in most studies.In addition, there was an increased RoB due to incomplete confounding adjustment in most studies.Studies often did not adjust for important confounders such as treatment and/or stage, factors which can have a large effect on both the exposures and outcomes of interest, while most studies did not account for the presence of comorbidities, which may have an important impact on the ability to participate in physical activity and survival.Finally, the literature search ended on 28 February 2022.
Thereby, any relevant studies published after this date were not included.Nonetheless, given the pivotal role RCTs in our evidence grading framework, we performed a comprehensive literature review targeting RCTs published subsequently and until August 31, 2023.However, we did not identify any published physical activity trial among colorectal cancer survivors, apart from the currently ongoing CHALLENGE trial.That trial has an estimated completion date in late 2030. 19,161As such, we anticipate that the conclusions would remain unaltered.
There are complexities in determining the feasibility of trials for survival-related outcomes.Deciding whether mortality should be the primary focus is not a straightforward decision.To conduct intentionto-treat analyses, participants should ideally remain actively engaged until reaching the primary endpoint.Additionally, selecting and maintaining the appropriate level of activity for the comparison group is challenging, particularly considering existing recommendations for physical activity for cancer survivors and the potential for contamination between groups.The evidence, however, is supported by studies of different designs including clinical treatment trials with a secondary aim of evaluating the impact of physical activity on cancer outcomes.The populations of these studies were more homogeneous, composed of individuals with advanced and metastatic tumours and the timing of exposure measurement was also more consistent, capturing activity close to treatment.Physical activity was not consistently assessed across observational studies, as they used different tools for measuring activity levels, measured different types of activity, and assessed different aspects.Our review explicitly assessed several different aspects of physical activity, including frequency, intensity, volume, and duration to enhance definition homogeneity, but studies did not mutually adjust for any of these factors.All studies utilised self-report questionnaires, of which only around half have been validated.Although all studies provided comprehensive descriptions of exposure assessment, physical activity is an umbrella term that describes a breadth of complex behaviours, often inhibiting direct comparability across studies.Even within the same domain or dimension, the comprehensiveness of the assessment varied across studies.Furthermore, selfreported physical activity and sedentary behaviour have been shown to have poor agreement with accelerometer-based measurements generally, 162 and specifically in colon cancer patients, 163 leading to an increased likelihood of exposure misclassification.This misclassification is expected to be non-differential which may result in attenuated associations.The wide range of tools used for exposure assessment and the inclusion of cancer patients of any stage may partly contribute to the heterogeneity of the results across studies.Finally, while TV viewing time was used as a surrogate measure of sedentary behaviour, the two are not directly equivalent, since other sedentariness aspects such as occupational sedentary time are not captured in TV viewing time.
The literature is primarily focused on the assessment of recreational activity.Evidence on total activity is sparse while there is some limited evidence on other types of activity, such as household activities.The timing of physical activity measurement differed across studies, ranging from immediately after diagnosis to over 10 years post-diagnosis.Individuals are more likely to engage in lower levels of activity during and short-term after cancer treatment, 164,165 and their attitudes towards exercise engagement during this period are mixed. 166,167In contrast, studies on long-term survivors, that usually exclude patients with relapse or with most aggressive or treatmentresistant disease, indicate good proportions of adherence to physical activity recommendations. 168,169Such inconsistencies across the timing of measurement may contribute to the observed heterogeneity and aggravate direct comparisons across studies but due to the large differences and low number of studies, subgroup analyses were not feasible.Moreover, almost all studies used a single measurement as the exposure.Colorectal cancer patients have inconsistent levels of activity throughout the course of the disease, and it has been suggested that even if patients engage in a short-term physical activity intervention they tend to revert to their less-active lifestyle. 170Due to the limited number of studies reporting on specific subgroups, we could not perform a comprehensive evaluation of possible sources of heterogeneity, as originally planned.
Our systematic literature review also highlighted a lack of evidence in the literature, particularly in examining the associations of both subjectively and objectively measured sedentary behaviour with outcomes following colorectal cancer, as well as for non-recreational physical activity domains.To enhance our understanding of the impact of both physical activity and sedentary behaviour on colorectal cancer survival, future research should prioritise the assessment of these currently understudied areas.
In consideration of the current evidence base and its methodological limitations, the CUP Global Expert Panel conservatively categorised most of the evidence as limited-no conclusion.Despite the indication of associations across the breadth of the performed analyses, the current evidence is not sufficient to support robust conclusions on causality.The inverse association between recreational activity and all-cause mortality was concluded as limited-suggestive despite being subject to the same biases as the remaining associations, due to the high number of studies providing consistent results (10/11 supported an inverse association, 1 indicative), which was further supported by the results of the linear dose-response metaanalysis.In most cases dose-response meta-analyses could not be performed due to low number of studies, lack of information, and use of a mixture of domains and dimensions assessed, only high versus low levels of activity were compared utilising heterogeneous cut-offs across studies.Previous assessments, such as those from the ACSM 22 that considered the effect of activity across different types of cancer survivors, have concluded a protective effect of physical activity on cancer mortality, albeit with unknown dose required for tangible benefits.However, the CUP Global Expert Panel, focusing specifically on the benefit of activity on survival outcomes of colorectal cancer patients, acknowledged the limitations of the current literature on the field and concluded that the evidence lacks a consistent basis for further recommendations for this specific population.An overview of limitations of cancer survival studies and future research recommendations is presented in the summary manuscript, Box 1 30 of the current manuscript series on colorectal cancer survivors.
[173] The CUP Global Expert Panel recognises that the limitations of the evidence represent an opportunity for further research to clarify the nature of the associations between measures of physical activity and sedentary behaviour and cancer-related outcomes.Furthermore, the panel recognises a need to better inform cancer patients about the links between physical activity and sedentariness and cancer survivorship, beyond making firm recommendations based on high quality evidence.

| CONCLUSIONS
Our work indicates that the accumulated evidence on the putative beneficial association of physical activity with outcomes after colorectal cancer is still limited and impacted by methodological limitations frequently present in observational survivorship studies, such as measurement error, residual confounding, reverse causation, and survival bias.As physical activity is a multi-faceted exposure, additional observational research using standardised assessment and ideally repeated objective measures, is of utmost importance to minimise biases and accurately capture its relationship with outcomes after colorectal cancer.Moreover, future high-quality trials focusing on specific physical activity dimensions and domains are essential for formulating robust conclusions, which could form the base for specific lifestyle guidance for colorectal cancer survivors.Nevertheless, even in the absence of strong evidence, there is potential to use this evidence as guidance for colorectal cancer patients and their health professionals to show that a physically active lifestyle and avoidance of sedentary behaviour may be associated with longer overall survival after a colorectal cancer diagnosis.

2 F I G U R E 2
Summary risk ratios (RRs) and 95% confidence intervals (95% CI) for post-diagnosis physical activity and sedentary behaviour and outcomes after colorectal cancer.CI, confidence interval; MET, metabolic equivalent of task; RR, relative risk.CI = 0.72-0.83;I 2 = 0%; tau 2 = 0.00; RRs range = 0.75-0.78; Stratified analyses by molecular subtypes were only performed in subsets of participants (n = 371-605) of the Nurses' Health Study & increases of total physical activity by >0-2 h/week and by ≥2 h/week from pre-diagnosis to 5 months post-diagnosis compared to no change or decreased levels showed no strong evidence of association with all-cause mortality (RR = 1.27; 95% CI = 0.88-1.83and RR = 1.06; 95% CI = 0.65-1.71,respectively).There was some weak evidence that a post-diagnosis increase (from 5 to 12 months) by 0-2 h/week was associated with a lower risk of all-cause mortality (RR = 0.79; 95% CI = 0.59-1.04),while an increase by ≥2 h/week was associated with a lower risk of all-cause mortality (RR = 0.69; 95% CI = 0.50-0.94)compared to no change or decreased activity levels.In the Nurses' Health Study,136 an increase of at least one level of recreational physical activity (defined as categories of MET-h/week) from a median of 6 months before to a median of 22 months post-diagnosis compared to no change in activity levels over this period was associated with lower all-cause mortality risk (RR = 0.53; 95% CI = 0.30-0.85),while a decrease of physical activity by one level compared to no change showed no association (RR = 1.23; 95% CI = 0.79-1.91). 123,127,133

Table 1
Summary of the judgement of the Independent WCRF Expert Panel.
were among the reasons that led to limited conclusions on the likelihood of causality for the existing evidence.The evidence on postdiagnosis recreational physical activity and all-cause mortality and cancer recurrence received the limited-suggestive sub-grade.A limited-no conclusion sub-grade was given to recreational physical activity and other outcomes apart from all-cause mortality and cancerT A B L E 1