Long‐term results of a short‐term home‐based pre‐ and postoperative exercise intervention on physical recovery after colorectal cancer surgery (PHYSSURG‐C): a randomized clinical trial

The aim of this work was to assess the effect of a short‐term, home‐based exercise intervention before and after colorectal cancer surgery on 12‐month physical recovery within a previously reported randomized control trial (RCT).


INTRODUC TI ON
Curative treatment of colorectal cancer consists primarily of surgery, leading to a recovery process that is sometimes prolonged by postoperative complications [1].Efforts have been made to reduce the burden of surgery by perioperative measures, including the 'enhanced recovery after surgery (ERAS)' concept [2].Both preoperative physical activity and physical functional capacity [3][4][5][6][7] have been associated with postoperative recovery after colorectal cancer surgery.The levels of physical activity and fitness have also been associated with the risk of developing and mortality after a colorectal cancer diagnosis [8][9][10][11][12][13].This has led to several attempts to improve postoperative recovery by exercise intervention alone or in combination with other interventions [14][15][16][17][18].
A 6-week high-intensity exercise intervention in patients undergoing major abdominal surgery with or without cancer showed dramatic effects on postoperative complications [16].Two intervention trials with sample sizes dimensioned for testing effects on clinically relevant outcomes could not show any effect of moderate-intensity physical activity 2-4 weeks preoperatively in patients with colorectal cancer [14,18].A recent systematic review concluded that there is no evidence of any effect on postoperative complications, length of stay, readmissions or mortality in patients with colorectal cancer [19].Two recent studies gave hope of lower complications after colorectal cancer surgery from exercise intervention, albeit limited by highly selected populations resulting in recruitment rates of 10%-21% and either a low sample size or termination at one-third of the planned sample size with changed primary outcome [15,17].Even if the primary aim with prehabilitation studies is to improve postoperative recovery, the underlying hypothesis is that this could translate into long-term benefits.Hence, long-term follow-up is warranted.
In this 12-month follow-up of PHYSSURG-C, we aimed to assess the effect of a short-term, home-based exercise intervention before and after colorectal cancer surgery on long-term physical recovery, compared with usual care.

Study design
PHYSSURG-C is a pragmatic randomized, controlled, open-label, multicentre trial and the design has been described previously, including the full study protocol [18,20].Participants were recruited at five regional hospitals and one university hospital in Sweden.
Ethical permission was obtained from the Regional Ethics Board in Gothenburg (2014-10-30, DNR:597-14).The study was registered at Clini calTr ials.gov with the trial registration number NCT02299596.
The date of first registry was 17 November 2014.

Participants
The eligibility criteria were individuals aged ≥20 years planned for elective colorectal cancer surgery at any of the recruiting hospitals.
Exclusion criteria were emergency surgery, local surgery, cytoreductive surgery with subsequent hyperthermic intraperitoneal chemotherapy, inability to understand given information due to language or intellectual barriers and inability to perform study-specific procedures (including surgery being planned too soon).Participants gave written consent and were recruited in association with their regular visits at included hospitals.

Randomization and masking
Participants were recruited by a research nurse.After consenting to participation, participants were randomly assigned to one of the two study groups with a 1:1 allocation using a computer system creating the allocation sequence.There was a block size of four, unknown to staff.The randomization system included a screening log.Allocation was stratified according to planned surgical method (laparoscopic or open), tumour site and neoadjuvant treatment (colon, rectum without preoperative radiotherapy or rectum with preoperative radiotherapy).The system assigned participants study numbers sequentially and the study group could not be changed.The recruiting research nurse distributed baseline and follow-up questionnaires which were returned to the study surgery.There is still not enough evidence to support clinical guidelines on preoperative exercise to improve outcome after colorectal cancer surgery.

K E Y W O R D S colorectal cancer, exercise, postoperative complications, prehabilitation
What does this paper add to the literature?
The effect of prehabilitation on clinically relevant outcome measures of long-term physical recovery after colorectal cancer surgery is not known.We could not see any effect from an unsupervised pragmatic intervention on long-term physical recovery.This further highlights the uncertainty of clinical benefit from prehabilitation in colorectal cancer surgery.
secretariat by post.Before inclusion, all participants received general information regarding the aim of the study; this contained no specific information on type or dose of physical activity to reduce the risk of contamination of the control group.Neither participants nor staff were blinded to group allocation, while research staff reviewing postoperative complications and length of hospital stay were masked regarding allocation.

Procedures
Participants in the control group received usual care both pre-and postoperatively.Participants in the intervention group met individually with a physiotherapist, when the participants' previous experiences and barriers to physical activity were explored and types of aerobic physical activity were discussed [21].They also received written and oral information regarding the intervention, consisting of two elements to be performed 14 ± 4 days preoperatively, to align with the time frames stipulated in the standardized care pathways for colorectal cancer in Sweden: 1. Thirty minutes of daily aerobic activity added to the individual's normal physical activity routine.This element was individualized in terms of type of aerobic activity and where the intervention was performed.The intensity of the activity was instructed to be of relative medium-intensity activity according to Borg's rating of perceived exertion scale [22].

Data collection
The research nurse registered baseline demographic information at inclusion.Participants were asked to complete questionnaires at inclusion and 4 weeks and 12 months postoperatively.The baseline level of physical activity was assessed with the Saltin-Grimby physical activity level scale [24].Two investigators, masked regarding allocation, collected information on postoperative complications, length of hospital stay, readmission and reoperations from healthcare records at the hospitals performing the index surgery.

Outcomes
The primary outcome in PHYSSURG-C was self-assessed physical recovery 4 weeks postoperatively, and the results have been published together with other measures of physical recovery within 90 days postoperatively [18].For a full list of the outcomes in the study, see the published protocol [20].
The secondary outcomes reported in this 12-month follow-up were: 1. Self-reported physical recovery 12 months postoperatively.
2. Reoperation 91-365 days postoperatively.Defined as all types of surgery under general anaesthesia performed that could be related to the colorectal cancer surgery or the oncological treatment, between days 91-365 after primary surgery.
3. Readmissions.All re-admissions 91-365 days after surgery that could be related to the index surgery or treatment were registered.

Statistical analysis
As described previously, the sample size was estimated for the primary outcome at 4 weeks postoperatively [18,20].The statistical analyses were performed according to a prespecified statistical analysis plan.Self-reported physical recovery was analysed with a proportional odds logistic regression model, with the result presented as an odds ratio (OR).Reoperations and readmittances were analysed with logistic regression models.Analyses were performed both in the intention-to-treat population as well as in the per protocol population, where activity ≥8 days preoperatively according to the exercise diary was considered as adherence to the intervention.
Missing values were handled by a listwise principle.
Statistical analyses were performed using R, version 4.2.2.The proportional odds logistic model was estimated using the MASS package.
The study stopped at full accrual.At 12 months postoperatively, 616 (81%) participants were available for follow-up of reoperations and readmissions, of whom 518 (84%) contributed with self-reported physical recovery.Mean age was 68 years and 39% of participants were female.Participants in the intervention and control groups were balanced at baseline regarding age, gender, education, ethnicity, occupational status, smoking, risk drinking, baseline level of physical activity, body mass index, comorbidity, tumour stage, type of surgery and American Society of Anesthesiologists (ASA) classification (Table 1).A total of 192 (67%) of the 287 participants in the intervention group registered activity ≥8 days preoperatively and were included in the intervention group in the per protocol analysis.There were no reported adverse events related to the intervention.
At 12 months postoperatively, 47% of participants reported that they were fully physically recovered (Table 2) compared with 14% at 4 weeks postoperatively [18].There were no differences between participants in the intervention and control groups, neither in the intention-to-treat analysis (OR 0.91, p = 0.60; Table 3) nor in the per protocol analysis (OR 1.12, p = 0.56; Table 3).
During the 91-365 days postoperative time frame, 97 (16%) of the participants were admitted to hospital for causes that could be related to the index surgery or treatment (Table 2).This did not differ between the intervention and control groups (OR 0.88, p = 0.58 for the intention-to-treat analysis; OR 0.80, p = 0.41 for the per protocol analysis; Table 3).During the same time frame, 31 (9.4%) of participants in the control group and 25 (8.7%) of participants in the intervention group underwent at least one reoperation, with a majority undergoing one reoperation (Table 2).This did not differ between the groups in our multivariable analyses (OR 0.97, p = 0.91 for intention-to treat analysis; OR 1.07, p = 0.84 for per protocol analysis; Table 3).

DISCUSS ION
This home-based unsupervised moderate intensity exercise intervention 2 weeks before and 4 weeks after curative surgery for colorectal cancer did not improve long-term physical recovery or reduce the risk of reoperation or readmission.This applied both to the intention-to-treat analysis and to the per protocol analysis restricted to those who reported adherence to the intervention.

TA B L E 1 (Continued)
with low fitness and randomized them to usual care or a 3-week supervised moderate-to high-intensity exercise intervention [17].
They concluded that the exercise intervention reduced postoperative complications in high-risk patients.However, participants were unbalanced at baseline, with more smokers, a higher comorbidity load and more open surgery in participants in the control group.Furthermore, they reported differences in the number of patients with any postoperative complication but no difference in the comprehensive complications index, length of stay or reoperations.A recent international study in 251 participants with colorectal cancer reported a ≈50% lower risk of a comprehensive complications index ≥20 (effectively a complication requiring medication or more intensive interventions) following 4 weeks of supervised high-intensity exercise [15].However, the population was selected with a recruitment rate of 22%, and the study was terminated at 35% of planned accrual due to the COVID pandemic and the primary outcome was changed.Hence, it remains to be shown whether these results could be repeated in an RCT reaching full power to show an effect on clinically relevant outcomes.
Furthermore, those results only apply to surgical high-risk patients with colorectal cancer (11% of the population) and cannot be extrapolated to the full colorectal cancer population.Still, these results are partly in line with the impressive reduction of postoperative complications seen in a Spanish study where patients scheduled to undergo major abdominal surgery, including but not limited to cancer, were randomized to a personalized, supervised, 6-week high-intensity exercise intervention [16].[9,[11][12][13].
This study has several strengths.One is the large sample size, allowing for sufficient power to detect clinically meaningful effects.
Another is the study design, with wide inclusion criteria and an intervention that is feasible to implement in clinical practice.The fact that the secondary outcome measures analysed in this manuscript were preselected at the outset of the study reduces the risk of data fishing.The exercise intervention in PHYSSURG-C was designed to align with general recommendations on physical activity, allowing for other health benefits than strictly postoperative ones [10].
Furthermore, with a cancer diagnosis being a life-defining event, the addition of a postoperative intervention aimed at long-term behavioural changes during survivorship is a strength.However, the combined short-and long-term results from PHYSSURG-C indicate that the short-term unsupervised moderate-intensity exercise intervention was not sufficient to improve measures of postoperative physical recovery.This is a limitation, and the current combined evidence favours high-intensity exercise for 4-6 weeks for the next large-scale RCT.However, we consider it a strength that PHYSSURG-C has evaluated the effect of general recommendations on the effects of prehabilitation in an unselected colorectal cancer population, and that we can state with relative certainty that they do not have any effect of relevance.Another limitation is the relatively low recruitment rate (41%) which reduces external validity.However, the other three RCTs in colorectal cancer populations have reported recruitment rates of 29% [14], 22% [15] or 11% [17], reflecting highly selected populations and even more limited external validity.include a weak recommendation on prehabilitation based on lowquality evidence of benefits for postoperative recovery [2].While there are a couple of promising studies [16,28], the only other highquality study in colorectal patients reaching full accrual also showed no effect [14].We consider it reasonable to believe that the current and previous results from PHYSSURG-C describe the effect to be expected from general recommendations on unsupervised preoperative physical activity [18].Unsupervised exercise isn't costly, has several other known health benefits and should be recommended during long-term disease-free cancer survivorship [29].However,

2 .
Inspiratory muscle training.The participant's maximal inspiratory pressure was determined at residual volume with a MicroRPM respiratory pressure meter (CareFusion, Höchberg, Germany) [23].The participants were instructed to perform inspiratory muscle training with 30 × 2 breaths, twice daily, starting with a resistance of 30% of maximal inspiratory pressure (Philips Respironics, Eindhoven, The Netherlands).The participants were instructed on how to increase the resistance.The postoperative intervention started when participants were discharged from the hospital.Participants in the intervention group were instructed to resume the same dose and intensity of aerobic activity as during the preoperative intervention and to continue for 4 weeks.No inspiratory muscle training was to be performed postoperatively.Participants in the intervention group were contacted by telephone by the research nurse 1 week into the preoperative intervention and 3 weeks into the postoperative intervention.This included follow-up of the intervention and allowed for modifications as needed.Adverse events related to the intervention were also registered by the research nurse in the follow-up phone calls.During hospitalization, participants in both groups received the same information on the importance of early mobilization postoperatively and were instructed to use deep breathing exercises hourly with positive expiratory pressure according to local routine.Local routines for pre-and postoperative care, including varying degrees of adherence to the ERAS protocol, were followed.
the preoperative time frame, where the patient has received a cancer diagnosis and is adapting emotionally, is challenging, and exercise recommendations should be evidence-based and cost-effective.If implemented in clinical practice, we think current evidence only supports supervised high-intensity exercise for a minimum of 4 weeks[15,16].To conclude, a pre-and postoperative unsupervised moderateintensity exercise intervention had no effect on short-or long-term physical recovery after elective colorectal cancer surgery.There is still not enough evidence to support clinical guidelines on preoperative exercise to improve outcome after colorectal cancer surgery.AUTH O R CO NTR I B UTIO N S Aron Onerup: Conceptualization; investigation; funding acquisition; writing -original draft; methodology; writing -review and editing; data curation.Ying Li: Writing -review and editing; methodology; investigation; formal analysis.Kevin Afshari: Methodology; writingreview and editing; investigation.Eva Angenete: Conceptualization; investigation; methodology; writing -review and editing; project administration; supervision.Hanna de la Croix: Conceptualization; funding acquisition; investigation; methodology; writing -review and editing.Carolina Ehrencrona: Methodology; writing -review and editing; data curation.Anette Wedin: Investigation; writing -review and editing; methodology; project administration.Eva Haglind: Conceptualization; investigation; funding acquisition; methodology; writing -review and editing; project administration; supervision; resources.FU N D I N G I N FO R M ATI O NThis study was funded by grants from AFA insurance (150072), the Swedish state under the agreement between the Swedish government and the county councils, the ALF-agreement (ALFGBG-718221, ALFGBG-4307771, ALFGBG-493341 and ALFGBG-784821), the Swedish Cancer Society (CAN 2016/519; 2019/190303), Assar Gabrielsson's foundation (FB19-07 and FB 16-95), Anna-Lisa and Bror Björnsson's Foundation, Dr Felix Neubergh's Foundation (2017-347), Gothenburg Medical Society (GLS-688001, GLS-778731 and GLS-883991), Lions Cancer Fund West (2019:6), Mary von Sydow's Foundation (1216), the Swedish Society of Medicine (SLS-499811) and the healthcare committee Region Västra Götaland (Hälso-och sjukvårdsstyrelsen) (VGFOUREG-309261, VGRFOUREG-659011, Flowchart of study cohort.

244 for self- reported physical recovery, n = 329 for other outcomes) Intervention (n = 262 for self- reported physical recovery, n = 287 for other outcomes)
[26]models were adjusted for tumour site (colon or rectum), neoadjuvant therapy (none, radiotherapy or chemo/radiotherapy) and type of surgery (open or laparoscopic).The missing values were handled on a listwise principle.pathwayguidelinesmaybeeffective.The Dutch study prolonged time to surgery for participants in the intervention group to give time for the intervention[17], and we have recently reported that this should be safe in colon cancer patients, based on analyses in a population-based sample[26].To our knowledge, none of the other large RCTs have reported long-term follow-up of their interventions.However, Trépanier They reported a 50% reduction in postoperative complications.Combined, these results indicate that high-intensity training for a longer duration than what is currently possible within most national cancer TA B L E 2 Descriptive information about outcomes for the intention to treat population.Outcome Control (n = TA B L E 3 Adjusted analyses for intention to treat (n = 616) and per protocol (n = 521) analyses.