Feasibility of a 12-Month Exercise Intervention in Postsurgical Colorectal Cancer Patients

Background Extensive physical activity (PA; ≥18 MET∗h/week, MET metabolic equivalent of tasks hours) postcancer diagnosis has shown favorable effects on colorectal cancer disease-free survival. However, the feasibility of introducing this high volume of PA in this patient group is unclear. Therefore, the aim of the F-PROTECT study was to evaluate the feasibility of extensive and prolonged PA (≥18 MET∗h/week over 12 months) in colorectal cancer patients with the primary objectives to (1) recruit 50 patients within 12 months and (2) reach an attendance rate of ≥70%. Methods Single-armed, bicentric, prospective intervention study in colorectal cancer patients (≤80 years; UICC II/III Union for International Cancer Control) after histopathological confirmed R0-resection who were consecutively recruited from visceral surgery units of 10 clinics in Germany. Recruitment rates were calculated using screening logs. Intervention was a 12-month endurance-focused exercise program with supervised and home-based training. Attendance rates defined as ≥70% participation in training sessions were calculated by training diaries. Results Out of 521 patients who were screened for eligibility, 50 (23 female; 59 ± 10 years, UICC 44% II, 56% III; adjuvant chemotherapy 60%) were recruited within 15 months. Mean duration between surgery and first training was 103 ± 57 days. Training attendance rate was 64% (including 9 dropouts). Six (12%) participants reached ≥18 MET∗h/week in ≥70% of training sessions between 4–12 months. 28 adverse events (n = 9 serious) occurred, however, were not assessed as training related. Conclusions The present intervention involving a combination of supervised and home-based exercise training in postsurgical colorectal cancer patients was not feasible. Strategies specifically designed for this patient group must be developed and investigated to motivate long-term PA. Registration. The study was prospectively registered at clinicaltrials.gov (NCT01991847).


Introduction
Colorectal cancer (CRC) is the third most common cancer worldwide and is responsible for about 10% of all cancer deaths [1].As a result of optimized therapy methods, the relative 5-year survival rate of CRC has improved over the last decades and is above 60% in most European countries [2].One cornerstone of the posttreatment management in CRC survivors is a healthy lifestyle.Physical activity (PA) intervention trials have been shown to improve the quality of life, PA level, and exercise capacity in CRC survivors [3].Furthermore, there is growing evidence from cohort studies that PA even has the potential to improve overall and CRCspecifc mortality in CRC survivors.A current meta-analysis over 13 studies investigated the association between PA and prognosis in 19.135 nonmetastatic CRC patients after a curative resection.Compared to a PA level of <3 metabolic equivalent task hours per week [4] (MET * h/week), PA levels of ≥18 MET * h/week were associated with a 36% reduction in overall mortality and a 34% increase in cancer-specifc survival [5].Accordingly, lifestyle intervention programs are required for CRC patients after primary treatment to increase their PA level.However, difculties in recruiting CRC patients for lifestyle interventions must be considered [6].It has been shown that adherence to PA recommendations is low in cancer survivors, particularly in CRC [7].In addition, it is unclear whether an extensive exercise training volume of ≥18 MET * h/week can be maintained over a prolonged period of time in this population, as most previous exercise intervention studies have examined only short intervention periods lasting maximum 6 months [3,8,9].
Terefore, the feasibility of the potential role of tertiary prevention by exercise in CRC therapy study (F-PROTECT) was initiated to investigate the feasibility of an extensive exercise intervention over 12 months in postsurgical CRC patients.Te main objectives were to (1) recruit 50 patients within 12 months and (2) reach an attendance rate (defned as ≥70% participation in training sessions) ≥70%.

Study Design.
Te F-PROTECT study is a prospective, single-armed, bicentric feasibility study, which was conducted in Munich and Ingolstadt (German cities) between January 2014 and April 2016.Te study has been carried out in accordance with the Code of Ethics of the World Medical Association (Declaration of Helsinki).Te study was approved by the ethics committee of the Faculty of Medicine of the Technical University of Munich, Germany (5888/13).Te study was prospectively registered at clinicaltrials.gov(NCT01991847).

Recruitment.
Te aim was to recruit 50 patients within 12 months.Patients were consecutively enrolled from visceral surgery units of ten clinics (nine in the metropolitan area of Munich and one in the smaller city of Ingolstadt).Tey were asked to participate in the study during their inpatient stay before or after surgery by clinical staf (physicians, study nurse) or study investigators or contacted by letter after their inpatient stay.Te time between recruitment and surgery had to be less than six months.Recruitment was continuously documented in screening logs for each clinic.Written informed consent was obtained from each patient before study enrolment.Te initially defned eligibility criteria were age <75 years, histologically confrmed, primary colon cancer, stadium UICC II or III, and histopathological confrmed R0-resection.After 7 months of recruitment, an addendum to the study protocol was accepted by the ethics committee for adjusting inclusion criteria (extension of the upper age barrier from <75 years to ≤80 years and inclusion of rectal carcinoma patients).

Intervention.
Every patient participated in an individualized endurance-based exercise intervention over 12 months.Te aim of the intervention was to increase the patient's PA level up to ≥18 MET * h/week [4] within the frst 12 weeks and to maintain this PA level during the intervention period.One MET is equivalent to resting energy expenditure.Multiple METs express exercise intensities that are performed over a certain time, e.g., hours per week.Based on the results summarized in [10], we decided to defne 18 MET * h/week as the target PA level.
Te exercise intervention consisted of both supervised and home-based training sessions (STS and HTS).Frequency of STS (on average 45 sessions) decreased over the intervention period while frequency of HTS (at least 120) increased.STS was consecutively reduced: during the frst 3 months, STS took place twice a week (months 4-6: once a week; months 7-9: every other week), and during months 10-12, STS took place monthly.HTS started with twice a week for the frst three months, to 2-3 times in months 4-9, to 3-4 times in months 10-12.For both STS and HTS, during the frst six months of training, the intensity was in the range of 60-70% of the peak oxygen consumption (peak VO 2 ).In months 7-9, the intensity varied through the use of the interval method between 60 and 80% of peak VO 2 in order to achieve an increase in physical exercise capacity.In the last three months, the intensity range was increased to 70-80% of peak VO 2 .Each training session lasted 60 minutes and included a 10-minute warm-up (50-60% of peak VO 2 ) and a 5-minute cool-down at low intensity.
Due to the changing performance status of CRC patients, especially during chemotherapy, intermittent training was performed at the beginning of the training or as needed (for example, 5-6 minutes of exercise followed by 1-2 minutes of rest).Te load phases were then gradually prolonged, and the rest periods were shortened accordingly.As soon as possible, the transition was made to the continuous method.In the further course, the application of the interval method was possible to increase the intensity, e.g., by interval runs with an alternation between intensive walking and light jogging.STS consisted of treadmill running, bicycleergometry, or elliptical.HTS consisted of multiple diferent activities.During the frst 3 months, sports such as hiking, walking, Nordic walking, and cycling were practiced.In the further course, other sports were added depending on the physical capacity and preferences of the study participants (e.g., modifed ball games, exercise ball games, movement and holding exercises from the areas of Tai-Chi, Qigong, or Yoga), in order to diversify the training.Alternatively, dynamic strength training was performed.However, the focus was on aerobic endurance training.
To optimize patient compliance, STS were free of charge and conducted at licensed rehabilitation centers close to participants' homes.[11] including the colorectal module, the EORTC QLQ Fatigue-Module (FA-13) [12], the Hospital Anxiety and Depression Scale (HADS) [13,14].A patient satisfaction questionnaire was handed out at the fnal study visit including the questions "How high was your motivation for training (a) at the rehabilitation center and (b) at home?" Te response options were classifed into fve levels from very low (0) to very high (4).Cardiopulmonary ftness was determined by peak VO 2 and assessed by cardiopulmonary exercise testing (CPX) at V1 and V3.CPX was performed on a bicycle, with a loading scheme of 10/10/1 for very poorly performing patients or 25/25/3 for more powerful patients.Tat is, the load started at 10 or 25 watts, and there was an increase of 10 or 25 watts every minute or every 3 minutes, respectively, until the maximum workload was reached with the target of exhausting them with a respiratory exchange ratio (RER) > 1.05.Te heart rates were documented at the respective intensity increases.On the basis of the maximum oxygen uptake during testing (peak VO 2 ) and the heart rates determined during the testing, the patient could be given the optimal training pulse range.For aerobic endurance training, this corresponds to the heart rates determined at 60-80% of the peak VO 2 .

Assessment of Physical Activity.
Leisure time PA (LTPA) was measured via the International Physical Activity Questionnaire (IPAQ) (V1-V3).Te IPAQ guidelines were followed including truncation rules to avoid the misclassifcation of extreme values [15].If weekly hours for walking, moderate, or vigorous intensity LTPA exceeded 180 minutes for each category, it was truncated to a maximum of 180 minutes per week for each category.Besides this subjective measurement, the AiperMotion440 accelerometers were used to objectively assess PA level.Te device was handed out to the patients at the study visits V1-V3.Participants were instructed to wear it for a period of ten consecutive days on an elastic belt placed at the height of the iliac crest of the nonparetic side, except when sleeping at night, showering, or doing any water-based activity.Seven days were included in the evaluation (minus the frst two days and the last day).Patients with less than four valid days were excluded from the analysis.Date of birth, body weight and height, date and time were entered into the device to determine the evaluation period as well as the metabolic rate.Furthermore, MET * h/week was calculated on the basis of training diaries based on the Compendium of Physical Activities by Ainsworth et al. [4] considering months 4-12 (months 1-3: initiation phase for reaching 18 MET * h/ week).Lost training diaries were defned as missing values.
Incomplete documentation was calculated as no PA (0 MET * h/week).
2.8.Safety.All (serious) adverse events ((S)AEs) that occurred during the study were documented in the case report forms by study investigators, regardless of their causal relationship to the intervention.Tey were recorded continuously and had to be reported back to the study center by using safety report forms.Furthermore, the patients were asked by physicians about their medical history as part of a standardized anamnesis during all study visits with regard to the occurrence of diseases.AEs were defned as diseases, signs of disease, or symptoms that occur or worsen after the inclusion of the patient in the study.SAEs included patient death, life-threatening events, events that cause permanent or signifcant damage to health or that require hospitalization or prolong hospital stay or that result in permanent disability (inability to work/earn), and secondary malignancies.Excluding the 9 dropouts, attendance was 78% (Figure 2).Patient's motivation for STS was indicated as very high by 44.4% and as high by 36.1%, while this was 13.2% and 50.0%for HTS.
Changes in the patient characteristics over the course of the intervention are shown in Figure 4 and Table 2.

Discussion
Te F-PROTECT study has shown that the present intervention involving a combination of supervised and homebased exercise training in postsurgical CRC patients was not feasible according to the defned criteria.Te recruitment period of the targeted 50 patients exceeded the goal of 12 months by 3 months, which is also refected in the low recruitment rate of 28% (50/181).Te attendance to STS was 6% below the 70% to be achieved.Furthermore, only 12% of participants reached the intended ≥18 MET * h/week of exercise in ≥70% of training weeks up to week 13.

Recruitment.
A recent systematic review and metaanalysis found an overall recruitment rate of 38% based on 16 exercise intervention trials in CRC patients.Importantly, recruitment rates difered according to the duration of the intervention (62% for <12 weeks; 33% for ≥12 weeks).In addition, they found notable diferences in the timing of treatment initiation.Te recruitment rate in studies that started the intervention after treatment was 28%, while it was 33% during treatment and 88% before treatment [8].In our study, the majority of patients were recruited during their inpatient stay with ongoing cancer therapy.Te median time from surgery to study inclusion was 9.5 weeks.Te short time span from cancer diagnosis and surgery to study inclusion, as well as the stress from chemotherapy, may have discouraged patients from participating in an exercise intervention study.However, Brown et al. have shown that recruitment was most successful during a more restrictive phase of recruitment (completed surgical resection and adjuvant chemotherapy ≤24 months before entering the study compared to ≤36 months).Tere is probably a window 4 Translational Sports Medicine of time relatively close to therapy when patients are particularly motivated to make lifestyle changes [17].
On average, we enrolled 3.3 patients at 10 departments over 15 months.A further study in CRC patients investigating a home-based interval-walking intervention delivered by a smart-phone application that has also applied in-clinic recruitment achieved a number of 2.4 patients (39 over a period of 16 months) per month at 5 oncology departments [18].A way to increase the recruitment numbers could be the use of population-based cancer registries along the lines of Brown et al., with which they were able to contact more than 1.500 potential candidates via letters.Te screening was conducted via telephone interview.Anyhow, they enrolled 39 patients over 7 months (5.4 patients per month) at a single institution [17].Furthermore, it might be useful for future trials to not exclude patients based on their metrics but to also account for the study staf's ability to motivate patients to engage in tertiary preventive measures.Accompanying, the fnal selection of patients could be done to a higher extent by the main trial centers with an expertise in sports medicine.Te possible lack of sports medicine knowledge of the clinicians on-site at the surgical clinics could have led to overcaution and reluctance to propose patients for the study.Sports medicine departments have experience in evaluating the ftness of patients independently of comorbidities or old age factors that are not necessarily accompanied by reduced ability to exercise.A far distance to the trial center or to the training locations was identifed as the main reason for the refusal of study participation in previous exercise intervention studies for cancer patients [6,[19][20][21].We tried to prevent this problem and ofered diferent training locations located near the patients' homes.Te distance to travel to the trial center was still considered as a key reason for refusal.Terefore, the number of visits to the trial center should be limited in the future in order to increase recruitment and compliance.

Attendance and Dropouts.
In the meta-analysis of Singh et al., a median adherence of 85% (42-91%) was reported among 19 exercise intervention trials with intervention durations ranging between 1 week and 6 months [8].To compare other feasibility studies, Sellar et al. achieved attendance rates of more than 90% with 27 of 29 CRC patients  Translational Sports Medicine  [22].Our attendance rate is relatively low compared to other studies but seems to be realistic when taking the long training period of 12 months into account.Furthermore, supervision was reduced because the goal was to enable patients to maintain PA independently.In the comparable CHALLENGE trial, 83% of participants completed the mandatory STS in the frst six months and 68% completed the optional STS in the second six months [23].Based on the high level of patient's satisfaction with STS, consideration should be given to maintain a mix of supervised and home training for future studies.Tis approach is supported by the overall results of a current meta-analysis showing a signifcant efect on the quality of life and functional capacity regarding supervised or mixed intervention (6 studies) while no signifcant efect was observed for home-based intervention (5 studies) [24].Te meta-analysis by Singh et al. reported a median withdrawal rate of 12% (0-22%) compared to 18% in the F-PROTECT study.Te feasibility study of the CHALLENGE trial reported a 10% dropout rate.Importantly, the trial included specifc behavior change techniques that may have improved motivation to continue participation in the study [23].
Te comparison between compliant versus noncompliant patients in our study indicates that patients with advanced stages of cancer, low overall performance status, and low PA status need to achieve special attention in exercise intervention trials.

Changes in Physical
Activity.Te amount of selfreported LTPA increased statistically signifcant from baseline to 12 months, which is in line with other PA intervention studies [21,23,25,26].However, results of objectively measured PA revealed no statistically signifcant changes in PA level nor metabolic rate.Furthermore, the aim of the PA intervention to reach and maintain a threshold of ≥18 MET * h/week was not successfully realized, as only 6 participants achieved ≥18 MET * h/week in ≥70% of training weeks.Sellar et al. were able to achieve a PA level of ≥18 MET * h/week but acknowledged that further studies need to investigate whether this can be maintained over a longer period of time because their intervention periods were only 12 weeks [20,25].Although the lack of documentation was counted as 0 MET * h/week for calculation purposes, which may have resulted in an underestimation of PA volume when training diaries were taken into account, the results of the F-PROTECT study suggest that this extensive PA volume is not realistically achievable for most CRC patients over a long time period.However, the question arises whether a lower PA level might be sufcient, also considering current recommendations of 150 min/week of moderate intensity PA, which corresponds to 10 MET * h/week.Te systematic review and meta-analysis by Schmid and Leitzmann showed that there seems to be a dose-response efect concerning PA and mortality in CRC patients.Each 10 MET * h/week increase in PA after diagnosis was associated with a 28% lower risk of all-cause mortality [10].Similarly, in a subanalysis by Meyerhardt et al., patients who increased their PA and those who were consistently active of at least 9 MET * h/week had an improvement in CRC-specifc mortality compared with relatively sedentary patients who had no change in PA [27].Accordingly, the goal in the CHALLENGE trial is to increase leisure-time PA by ≥ 10 MET * h/week from baseline during the frst 6 months and to maintain this change for 3 years.Tis was successfully achieved in their one-year feasibility study [23].Future RCTs need to investigate lower PA doses associated with survival benefts and transferability to clinical practice.
Interestingly, there was an increase in BMI as well as in waist circumference and body fat which is in line with the results of a current meta-analysis including 7 RCTs and 803 CRC patients, which has shown no signifcant within groups efects of PA interventions for BMI [3].Tis may be interpreted in the context of the timing of diagnosis and cancer treatment.An observational study in 485 stage II/III CRC patients has shown that body weight decreased during surgery and increased during and after chemotherapy [28].However, as we do not have information on dietary intake and due to the lack of a control group, we can only speculate on these fndings.Importantly, our results reveal that an increase in peak VO 2 was statistically signifcantly correlated with a decrease/less increase in waist circumference (r � −0.413; p � 0.011), which highlights the importance of increasing ftness to improve the prognosis in this patient group.4.4.Safety.Twenty-eight(S)AEs occurred during the study participation.Even though they were assessed as not training-related which is in accordance with the assessment of the safety of exercise in previous interventional trials [8], the high incidence of complications in this patient group is certainly a major barrier to exercise intervention studies.

Study Strengths and Limitations.
Te F-PROTECT study examined a much longer intervention period than most previous feasibility studies in this population, providing important information about the feasibility of maintaining relatively high PA volumes over long periods of time.Moreover, a relatively large number of patients were included compared with other feasibility studies of PA in CRC.Physical ftness was measured by peak VO 2 , which is considered the gold standard for measuring maximal aerobic capacity.
An important limitation is the lack of a control group.However, the aim of this study was to demonstrate the feasibility and not the superiority of the intervention.Tus, our results cannot determine causality.
Furthermore, selection bias must be considered (patients willing to participate may be healthier/ftter).As can be seen from the results, the study participants were relatively active at baseline.However, the cardiorespiratory ftness was relatively low and comparable to other studies involving CRC patients [9,29].Nevertheless, our study Translational Sports Medicine population may represent a high-selective group, which does not necessarily refect the patient strata presenting for treatment.However, attendance to protocols was difcult, probably due to the high amount of required exercise.Additionally, feasibility and attendance rates are critically dependent on multiple factors, such as study setup, patient's preference of exercise type (strenuous vs less strenuous exercise, focusing on diferent aspects of motoric abilities, etc.) center capabilities, local distances to training centers, or underlying population demographics.Terefore, translating results from this study to other sites has to be done with caution.

Conclusions
Te feasibility of an endurance-based intervention over a 12month period involving a combination of supervised and home-based training in postsurgical CRC patients could not be confrmed.
Researchers should therefore develop exercise programs specifcally tailored to CRC patients that they can maintain over the long term.Furthermore, future RCTs should aim to optimally monitor PA (e.g., telemedical) to allow the calculation of a PA continuum to defne a realistic and sufcient threshold for improving prognosis.

Figure 2 :Figure 3 :
Figure 2: Number of patients considering the percentage of performed training sessions.

Figure 4 :
Figure 4: Leisure-time physical activity at baseline, 6 months and 12 months.
Contact with one participant was lost.Tere were no statistically signifcant diferences between dropouts and participants who completed the study for the variables age, sex, stage of disease, adjuvant therapy, and distance to training centers.
Te causal association of the AE with the intervention (physical training) was given if the event occurred during a reasonable period of time after intervention (within 48 hours), could not be explained by the patient's clinical condition, the environment, toxic factors, or other therapeutic interventions, could be attributed to known adverse efects of the intervention, improved or resolved after discontinuation of the intervention, or recurred after the resumption of the intervention.testsandbetweentimepoints(paireddata)using Wilcoxon signed rank tests, as data were not normally distributed.For the comparison of distributions of categorical variables between groups, Fisher's exact test was used.All tests were performed two-sided, and a signifcance level of 5% was used.*h/weekfromweek 13 onwards.Tere were no statistically signifcant diferences in baseline characteristics between sexes.3.2.Recruitment.Participant recruitment took place from 01/2014 to 03/2015.50 patients were recruited within 15 months out of 521 CRC patients who were screened for eligibility.Leading causes for exclusion were UICC IV stadium (20%), UICC I/0 stadium (10%), and age (19%, before the study addendum).131 of 181 eligible patients declined study participation (recruitment rate of 28%), of whom 62% (n � 80) were male and 38% (n � 50) were female (1 was excluded from the calculation because of missing information).Decliners were statistically signifcantly older compared to the included patients (66.2 ± 9.8 vs. 58.9±10.3years,p< 0.001).Te leading cause for declining was a too long distance from the patients' residence to the study center (see Figure1).3.3.Dropouts.Te dropout rate was 18% (n � 9; n � 6 before an intermediate visit, n � 3 before the fnal visit).Dropouts were due to lack of motivation (n � 3), occupational reasons (n � 2), lack of time (n � 1), chemotherapeutical side efects (i.e.polyneuropathy, foot pain) (n � 1), and cancer recurrence (n � 1).participants including the 9 dropouts (n � 32 of n � 50).

Table 1 :
Baseline characteristics in compliant versus noncompliant patients.

Table 1 :
Continued.Test for skewed quantitative data (LTPA, sitting time per day), independent t-test for other quantitative data; SD � standard deviation; QoL � quality of life; LTPA � leisure time physical activity; UICC � Union for International Cancer Control.† measured by IPAQ questionnaire.‡ measured by accelerometers.Peak VO2 � peak oxygen consumption; MET �metabolic equivalent of tasks.
* Fisher's exact test for binary data, Mann-Whitney U

Table 2 :
Patient characteristics from baseline to 12 months.Paired t test (changes from baseline to 12 months, based on individuals with both baseline and 12 months data).SD �