Effects of a Structured Exercise Program on Physical Activity and Fitness in Colon Cancer Survivors: One Year Feasibility Results from the CHALLENGE Trial

A growing number of epidemiologic studies have reported an association between postdiagnosis physical activity (PA) and colorectal cancer outcomes (1–3). In a recent systematic review, Des Guetz and colleagues (1) analyzed six observational studies involving over 7,500 colorectal cancer survivors and reported that higher postdiagnosis PA was associated with a lower risk of colorectal cancer–specific mortality [HR = 0.61; 95% confidence interval (CI), 0.44–0.86] and a lower risk of all-cause mortality (HR = 0.62; 95% CI, 0.54–0.71). Although these findings are promising, the studies are limited by the observational designs with high risk of confounding, and the use of self-report measures of PA. Moreover, the critical question from a clinical perspective is whether any feasible exercise intervention can actually improve cancer outcomes. To address this issue, the American Society for Clinical Oncology has recently called for large-scale phase III trials to evaluate the effects of lifestyle interventions on cancer outcomes and to determine the best methods of behavior change (4).

Like all medical interventions, the effectiveness of an exercise intervention to improve cancer outcomes will depend on the willingness and ability of patients to complete it. Compared with biomedical interventions, however, exercise interventions are more challenging for patients because they require substantial personal motivation (i.e., the intervention must be performed by the patient). Moreover, exercise interventions to improve cancer outcomes will likely require substantial and sustained behavior change. Unfortunately, exercise interventions delivered through practical avenues such as telephone counseling, print materials, and web-based programs often induce only small behavioral changes that are unlikely to improve cancer outcomes (5, 6). Conversely, exercise interventions delivered through optimal avenues such as fully supervised programs by qualified professionals often induce larger behavioral changes as well as health-related fitness improvements (7, 8); however, their potential for large-scale dissemination is questionable. Consequently, there is legitimate concern in the oncology community about whether any pragmatic exercise intervention can actually motivate patients to increase their exercise behavior and fitness levels sufficiently to improve cancer outcomes.

To address this issue, the Canadian Cancer Trials Group launched the Colon.21 (CO.21) Trial, also known as the Colon Health and Life-Long Exercise Change (CHALLENGE) Trial (9, 10). The CHALLENGE Trial is an ongoing randomized trial examining the effects of a 3-year structured exercise program (SEP), compared with health education materials (HEM), on disease-free survival among 962 patients with high-risk stage II or stage III colon cancer who have completed adjuvant chemotherapy within the past 2–6 months. The exercise intervention in the CHALLENGE Trial was designed to balance the need for substantial behavior change with the desire for broad dissemination. The behavior change goal in the CHALLENGE Trial is to increase recreational PA from baseline by at least 10 metabolic equivalent task (MET) hours/week in the first 6 months and sustain this change for 3 years. Given the legitimate concerns about inducing substantial and sustained exercise behavior change in colon cancer survivors, the CHALLENGE Trial steering committee incorporated a feasibility analysis into the study protocol. For the study to continue, the SEP group was required to perform an average of at least 5 MET-hours/week more recreational PA than the HEM group after a minimum of 250 participants reached the 1-year follow-up. The purpose of the current article was to report the results of the feasibility analysis including the corollary changes in health-related fitness.

The design and methods of the CHALLENGE Trial (Clinicaltrials.gov identifier: NCT00819208) have been reported elsewhere (9, 10). The CHALLENGE Trial originally opened in Canada and Australia in 2009 and is currently accruing participants in 20 centers in Canada and 22 in Australia. The trial has since been expanded to a small number of centers in the United States, Israel, and Korea with further expansion planned to the United Kingdom and France. All centers are required to secure local ethics approval and all participants are required to provide informed consent.

Colon cancer survivors are eligible for the trial if they have been diagnosed with high-risk stage II or stage III colon cancer, received adjuvant chemotherapy within the past 2–6 months, have an Eastern Cooperative Oncology Group (ECOG) performance status of 0 or 1, are not currently meeting PA guidelines (i.e., the equivalent of 150 minutes/week of moderate intensity exercise), and are able to complete at least two stages of a submaximal treadmill test. Participants are stratified by center, disease stage, body mass index (≤27.5 kg/m² vs. >27.5 kg/m²), and ECOG performance status before being randomly assigned to the SEP or HEM intervention.

Participants randomized to the HEM intervention receive general health education materials promoting PA and healthy nutrition as well as standard surveillance follow-up. Participants randomized to the SEP intervention receive an exercise guidebook developed specifically for colon cancer survivors (11) and work with a PA consultant (PAC) who delivers the intervention over a 3-year period. PACs are certified exercise specialists (e.g., exercise physiologists, personal trainers, physiotherapists) who receive either 1.5 days of training in Edmonton or from another PAC (i.e., “train the trainer” model). The training includes overviews of the trial, the role of the PACs, the theory of planned behavior, the behavioral support sessions, the tool box, handling drop outs, completion of forms, communication with the cancer center, and a practical demonstration of the health-related fitness testing. PACs are also required to submit a report to the Physical Activity Working Group and participate in a 1-hour conference call every 2 months to discuss any issues related to fitness testing, intervention delivery, and adherence.

The SEP intervention is based on the highly successful Diabetes Prevention Program (12) and Look AHEAD (13) studies with modifications and enhancements for colon cancer survivors (14–16). The key components of the SEP intervention include: (i) clear and challenging exercise goals, (ii) some supervised exercise sessions, (iii) free or low-cost access to a fitness facility, (iv) frequent and ongoing contacts including some face-to-face sessions with qualified staff, (v) individual tailoring of the intervention, (vi) written materials, and (vii) the application of behavior modification techniques based on a validated theoretical model. The goal of the SEP intervention is to increase recreational aerobic PA by at least 10 MET-hours/week from baseline in the first 6 months and sustain this change for 3 years. Participants are able to choose the type, frequency, intensity, and duration of aerobic exercise to meet the intervention goal.

The SEP intervention is comprised of three phases (Table 1). The first 6 months is an adoption phase that consists of bi-weekly face-to-face counseling sessions combined with supervised exercise, with the option of an additional supervised exercise session during alternate weeks. The second 6 months is a consolidation phase that consists of bi-weekly behavioral support sessions, which may be face-to-face with supervised exercise, or by telephone. The final 2 years are a maintenance phase that consists of monthly behavioral support sessions, which may be face-to-face with supervised exercise, or by telephone. The SEP intervention is guided by the theory of planned behavior (TPB; ref. 17) which has been validated for predicting and understanding exercise behavior in colorectal cancer survivors (15, 18–20). The counseling sessions cover 17 evidence-based behavior change techniques (Table 2; refs. 21, 22).

The primary outcome for the feasibility analysis was self-reported recreational PA as assessed by the validated and reliable Total Physical Activity Questionnaire (TPAQ; ref. 23). The TPAQ measures all types (occupational, household, recreational, and transportation) and parameters of PA (frequency, duration, and intensity) performed over the past month. The recorded activity is converted to MET-hours/week using the Compendium of Physical Activities (24). In the CHALLENGE Trial, the TPAQ is administered to all participants at baseline and every 6 months for 3 years. The feasibility analysis was restricted to recreational PA at baseline and 1 year.

Secondary feasibility outcomes included cardiorespiratory fitness, body weight, and circumferences, and objective physical functioning assessed at baseline, 6 months, 1 year, 2 years, and 3 years. Cardiorespiratory fitness is assessed using a submaximal, multistage, modified Balke treadmill protocol (25) with a validated prediction formula to estimate maximum volume of oxygen consumption (VO_2max; ref. 26). Standing height and weight are measured by a balance beam scale and stadiometer. Waist and hip circumferences are determined using an anthropometric measuring tape (27, 28). Objective physical functioning is assessed by the Seniors' Fitness Test (29).

For each primary and secondary outcome, the mean and SD were calculated for all participants with available data at both baseline and 1 year. The mean and 95% CI for the change at 1 year from baseline were calculated for all participants with available data at both time points. As prespecified in the protocol, a Wilcoxon rank-sum test was used to compare the differences in mean change between the two randomized groups. Also as prespecified in the protocol, the trial would be stopped if the observed between group difference in the average MET-hours/week at 1 year was less than 5. Assuming that the SD of MET-hours/week was 20, with 250 participants included in the analysis, the probability to continue the study (i.e., the observed difference in the average MET-hours/week between two intervention arms is 5 or higher) was higher than 88% if the true difference was 8 MET-hours/week or higher. Because of missing data from some participants, it was decided that all participants randomized before March 1, 2014 would be included in the feasibility analysis. A database was locked for this analysis on April 2, 2015.

Flow of participants through the 1-year follow-up is presented in Fig. 1. Between June 2009 and February 2014, 273 participants (192 in Canada and 81 in Australia) were randomized to the SEP intervention (n = 136) or the HEM intervention (n = 137). The final analyses for the change in PA from baseline to 1 year included 211 participants (106 from the SEP arm and 105 from the HEM arm). The 62 participants not analyzed included 5 who were deemed ineligible after randomization, 33 who were not required to complete the 1 year questionnaire because of progressive disease (n = 31) or withdrawal of consent (n = 2), and 24 who did not complete the baseline or 1 year questionnaire for various reasons. Consequently, the primary analysis included over 90% of eligible participants who did not have progressive disease at 1 year.

Table 3 provides the baseline clinical characteristics for the entire sample (N = 273) and the analyzed sample (n = 211) by randomized group assignment. In terms of the medical profile, 89% of participants were diagnosed with stage III colon cancer, 57% were treated with FOLFOX chemotherapy, and 74% have an ECOG performance status of 0. In terms of demographics, 54% are women and the median age is 60 years, with 31% over the age of 65 years. The groups were well-balanced on the baseline clinical characteristics in the overall and analyzed samples.

During the first 6 months, intervention participants completed an average of 82.8% of their 12 mandatory face-to-face counseling sessions, 82.6% of their mandatory supervised exercise sessions, and 22.1% of their optional but strongly recommended supervised exercise sessions. During the second 6 months, intervention participants completed 73.5% of their mandatory counseling sessions and 68.2% of their optional but strongly recommended supervised exercise sessions.

Table 4 provides the baseline and 1 year data for self-reported PA, cardiorespiratory fitness, and body weight and circumferences. In terms of the primary feasibility outcome, the SEP group reported an increase in PA of 15.6 MET-hours/week from baseline to 1 year compared with an increase of 5.1 MET-hours/week in the HEM group (mean between group difference = +10.5; 95% CI, +3.1–+17.9; P = 0.002). For predicted VO_2max, the SEP group increased by 1.6 mL/kg/minute from baseline to 1 year compared with a decrease of 0.6 mL/kg/minute in the HEM group (mean between group difference = +2.2; 95% CI = −4.6–+9.1; P = 0.068). There were no differences in body weight or circumferences.

Table 5 reports the baseline and 1 year data for objective physical functioning. For the 6-minute walk, the SEP group improved from baseline to 1 year by 59 meters compared with 31 meters in the HEM group (mean between group difference = +29; 95% CI = +0.4–+57; P < 0.001). The SEP group also improved relative to the HEM group for the 30-second chair stand (mean between group difference = +1.6 repetitions; 95% CI = +0.6–+2.7]; P < 0.001), the 8-foot up and go (mean between group difference = −0.4 seconds; 95% CI = −0.7–−0.2; P = 0.004), and the sit and reach (mean between group difference = +2.1 centimeters; 95% CI = −0.6–+4.7; P = 0.08). There were no differences in upper body functioning measures.

The SEP intervention in the CHALLENGE Trial increased self-reported recreational PA by 15.6 MET-hours/week from baseline to 1 year compared with 5.1 MET-hours/week in the HEM group. The increase in the SEP arm exceeded the initial 6-month goal of 10 MET-hours/week and achieved the second 6-month goal of an increase of 10–20 MET-hours/week. The increase in the HEM arm of 5.1 MET-hours/week is not unexpected and likely resulted from the motivated sample, the health education materials, and the fitness testing. The 10.5 MET-hours/week between group difference equates to approximately 3.5 hours/week of moderate intensity exercise (at 3 METs) or 1.75 hours/week of vigorous intensity exercise (at 6 METs). The magnitude of this between-group behavior change difference compares favorably to other exercise interventions in colorectal cancer survivors that have reported increases compared with usual care of between 30 and 60 minutes/week of moderate intensity exercise (14, 30–33).

Moreover, the amount of exercise behavior change achieved in the CHALLENGE Trial is consistent with the amount of exercise associated with improved colorectal cancer outcomes in observational studies. In their systematic review and meta-analysis, Schmid and Leitzmann (3) reported that an increase of 10 MET-hours/week of postdiagnosis PA was associated with a 28% lower risk of all-cause mortality and a 25% lower risk of colorectal cancer–specific mortality. Furthermore, the increase in PA from 16 to 32 MET-hours/week in the SEP arm is within the range of benefit for colorectal cancer outcomes. Specifically, two studies (34, 35) have applied spline curves to test for a threshold of association independent of predetermined MET categories and reported risk reductions starting between 6–12 MET-hours/week and continuing up to 30–35 MET-hours/week.

Consistent with the increase in self-reported recreational PA, the SEP intervention also improved several health-related fitness outcomes. Predicted VO_2max trended towards an improvement of 2.2 mL/kg/minute in the SEP group compared with HEM. VO_2max is an established predictor of mortality in many populations (36, 37). In a systematic review of 33 studies with over 100,000 participants (36), an increase in VO_2max of 3.5 mL/kg/minute was associated with a 13% lower risk of all-cause mortality. We also observed a significant improvement in the 6-minute walk in the SEP group of 59 meters which was superior to the HEM group. A change in the 6-minute walk of between 43 and 54 meters is considered clinically meaningful in several patient populations (38–40). Furthermore, the 6-minute walk is prognostic for survival in patients with several chronic diseases (41–43) including lung cancer (44).

Importantly, the SEP intervention improved the health-related fitness parameters that were expected to improve with an aerobic exercise intervention. Specifically, the SEP intervention improved VO_2max, the 6-minute walk, the 30-second chair stand, the 8-foot up and go, and the sit and reach. It did not improve the fitness parameters that were not expected to improve with an aerobic exercise intervention such as upper body functioning or body weight and circumference. Overall, the changes in health-related fitness are consistent with the focus of the SEP intervention on promoting aerobic exercise rather than strength exercise or weight loss.

The behavioral and health-related fitness improvements in the CHALLENGE Trial are especially noteworthy given that the intervention was delivered over a 1-year period by over 100 different PACs from 42 centers in two countries. The vast majority of exercise interventions in cancer survivors have assessed short-term behavioral changes (<6 months) in a limited number of centers (5, 6). Moreover, the small number of trials that have included longer term follow-up and broader implementation have often produced the smallest behavioral and fitness changes (5, 6), highlighting the well-known trade-off between reach/feasibility and effectiveness.

The effectiveness of the SEP intervention in the CHALLENGE Trial is likely based on the challenging exercise behavior change goal, the inclusion of some supervised exercise sessions, the frequent and continuing contacts, the theoretically-based behavior change techniques, and the ongoing training and monitoring of the PACs (12). A recent systematic review and meta-analysis (45) of exercise behavior change interventions in patients with diabetes noted that interventions were more effective if they included specific behavior change techniques such as planning, prompting, social support, goal setting, time management, personalizing benefits, and overcoming barriers; all of which are included in the CHALLENGE Trial. Moreover, the interventions were also more effective if they utilized ≥10 behaviour change techniques (17 are utilized in the CHALLENGE Trial) and were underpinned by a theory or model of behavior change, such as the theory of planned behavior.

Our interim feasibility analysis has some limitations. First, we have only reported exercise behavior change after 1 year of a 3-year intervention. Substantial declines in lifestyle behaviors after 6 months or 1 year are well-known (46); however, these declines often occur after an intervention is ceased (47), whereas the SEP intervention in CO.21 will continue for the full 3 years (9). Second, we have relied on a self-report measure of PA which, although validated, has well-known limitations (48). Nevertheless, there is an acceptable correlation between self-reported and accelerometer-assessed PA in colon cancer survivors (49) and the health-related fitness changes are consistent with the self-reported PA results. Finally, our results may only apply to promoting aerobic exercise; not strength exercise, nutrition, multiple health behaviors, or weight loss (50, 51).

In conclusion, the CHALLENGE Trial remains the only randomized controlled trial designed to examine the effects of a SEP intervention on disease-free survival in colon cancer survivors. The results of this feasibility analysis suggest that the SEP intervention is being successfully implemented in a multicenter international trial and is producing substantial behavior change and improvements in health-related fitness that appear sufficient for determining their causal effects on colon cancer outcomes. Moreover, the trial will provide valuable data on exercise motivation and behavior change that will inform the implementation of this program into widespread practice should it be warranted. The role of lifestyle interventions in improving cancer outcomes remains a compelling issue for cancer survivors and cancer care professionals (4).

No potential conflicts of interest were disclosed.

Conception and design: K.S. Courneya, J.L. Vardy, C.M. Friedenreich, P. O'Brien, H.M. Dhillon, D.J. Jonker, S. Gill, R.M. Meyer, D. Tu, C.M. Booth

Development of methodology: K.S. Courneya, J.L. Vardy, C.M. Friedenreich, J.J. Crawford, H.M. Dhillon, S. Gill, R.M. Meyer, C.M. Booth

Acquisition of data (provided animals, acquired and managed patients, provided facilities, etc.): K.S. Courneya, J.L. Vardy, C.J. O'Callaghan, C.M. Friedenreich, K.L. Campbell, H. Prapavessis, J.J. Crawford, P. O'Brien, H.M. Dhillon, D.J. Jonker, N.S. Chua, S. Lupichuk, M.S. Sanatani, S. Gill, S. Begbie, T. Bonaventura, M.E. Burge, C.M. Booth

Analysis and interpretation of data (e.g., statistical analysis, biostatistics, computational analysis): K.S. Courneya, C.J. O'Callaghan, P. O'Brien, D.J. Jonker, R.M. Meyer, S. Begbie, D. Tu, C.M. Booth

Writing, review, and/or revision of the manuscript: K.S. Courneya, J.L. Vardy, C.J. O'Callaghan, C.M. Friedenreich, K.L. Campbell, H. Prapavessis, J.J. Crawford, P. O'Brien, H.M. Dhillon, D.J. Jonker, N.S. Chua, S. Lupichuk, M.S. Sanatani, S. Gill, R.M. Meyer, S. Begbie, T. Bonaventura, M.E. Burge, D. Tu, C.M. Booth

Administrative, technical, or material support (i.e., reporting or organizing data, constructing databases): K.S. Courneya, C.J. O'Callaghan, J.J. Crawford, P. O'Brien, N.S. Chua, R.M. Meyer, D. Tu, C.M. Booth

Study supervision: K.S. Courneya, J.L. Vardy, C.J. O'Callaghan, C.M. Friedenreich, H.M. Dhillon, N.S. Chua, R.M. Meyer, J. Turner

K.S. Courneya and C.M. Booth are supported by the Canada Research Chairs Program. C.M. Friedenreich is supported by a Health Senior Scholar Award from Alberta Innovates-Health Solutions and through the Alberta Cancer Foundation's Weekend to End Women's Cancers Breast Cancer Chair. K.S. Courneya, S. Gill, R.M. Meyer, D. Tu, C.M. Booth, and P. O'Brien received funding from the Canadian Cancer Society Research Institute. J.L. Vardy, H.M. Dhillon, K.S. Courneya, and R.M. Meyer received funding from the National Health and Medical Research Council (Australia).

The costs of publication of this article were defrayed in part by the payment of page charges. This article must therefore be hereby marked advertisement in accordance with 18 U.S.C. Section 1734 solely to indicate this fact.