Carli 2020.

Study characteristics
Methods	Design: parallel‐arm single‐blinded, 2‐site randomised controlled trial.  Setting: two tertiary centres, Montreal, Quebec, Canada. Recruitment: participants were enrolled from the 7th of September 2015 to the 19th of June 2019. Consecutive patients eligible for participation were screened for frailty. Eligible patients were randomised on a 1:1 ratio to receive either a 4‐week prehabilitation programme (Prehab group) or a similar 4‐week postoperative rehabilitation programme (Rehab group). Both groups were assessed after randomisation by a kinesiologist, nutritionist and psychology‐trained nurse. Follow‐up: follow‐up was up to four weeks postoperatively. Surgical care followed ERAS guidelines. Blinding: outcome assessors, surgeons and statisticians were blinded to group assignment.
Participants	Screened: 418 patients randomised: 120 patients analysed: 110 patients Inclusion criteria: frail patients older than 65 years of age, scheduled for surgical treatment of non‐metastatic colorectal cancer. Patients were considered frail when scores of the Fried Frailty Index were ≥2. Exclusion criteria: patients were excluded in case of a Fried Frailty Index of 1, did not speak French or English, had metastatic disease and/or had a contraindication for exercise. Baseline characteristics: age in the Rehab group was higher (≥ 75 years of age: Prehab 32 (58.2%), Rehab 42 (76.4%) and patients had higher American Society of Anesthesiologists scores (ASA score of 3: Prehab 33 (60.0%), Rehab 43 (78.2%). The majority had a colonic resection with a minimal invasive surgical approach.
Interventions	The content of the multimodal programme was identical in both groups. However, the timing of the start of the programme differed between groups. Prehab (n = 55): the personalised, home‐based programme was prescribed by a kinesiologist, nutritionist and psychology‐trained nurse after the baseline visit. The programme continued for four weeks until surgery; no postoperative programme.  Rehab (n = 55): baseline assessment was similar. The patients were prescribed an identical, personalised home‐based programme. However, the programme started postoperatively after discharge from the hospital and continued for four weeks. Patients were informed about the programme only a few days before surgery. Exercise Home‐based exercise (both groups): a personalised home‐based programme was prescribed containing aerobic activities (moderate‐intensity, 30‐minute daily walk) and resistance training (three times per week elastic band routine). Guidelines of the American College of Sports Medicine were followed. In‐hospital supervised sessions (both groups): once a week patients performed an in‐hospital training session supervised by a trained kinesiologist. The training sessions consisted of a 30‐minute moderate‐intensity exercise on a recumbent stepper (including a 5‐minute warming up), a 25‐minute resistance exercise using an elastic band, and five minutes of stretching. Details of the programme were similar to Bousquet‐Dion 2018.  Nutritional intervention Nutritional status was determined by a registered dietitian using a 3‐day food diary, the SGA, and assessment of macronutrient intake and food choices. Dietary advices were provided together with counselling on caloric balance, bowel movement regularity, and glycaemic control. In case the patient did not meet a daily protein intake of 1.5g/kg of body weight (ESPEN guidelines), whey protein supplementation was prescribed and patients were instructed to ingest the supplementation within one hour of the exercise. Mental intervention The psychological intervention focused on perioperative fatigue, anxiety, and depression. A psychology‐trained nurse provided personalised coping strategies together with a compact disc containing instructions for guidance with the home‐based exercises. counselling regarding smoking and alcohol cessation was included in the consult and when indicated, nicotine replacement therapy was offered. Booklet All patients received an instructional booklet, including a diary to record daily activities. Follow‐up Patients were contacted on a weekly basis, by telephone, to report adherence to the home‐based programme.
Outcomes	Primary Postoperative complications within 30 days postoperatively was the primary outcome. Complication rate was expressed as mean and median CCI. Furthermore, the number and percentages of patients having a complication scored with Clavien‐Dindo grade was displayed, together with the number and percentages of severe complications (definition of severe not specified). Secondary Reported postoperative outcomes 30 days after surgery included length of hospital stay, re‐admissions and emergency department visits. Primary length of stay in the hospital and total hospitalisation were presented as median and IQR for both groups. The number of emergency department visits and re‐admissions are presented for both groups as numbers (%).  The following outcomes were assessed at baseline, before surgery, and four weeks after surgery: ‐ 6MWT. Results were given in metres (mean, SD) and number of patients (n, %) who improved their scores preoperatively compared to baseline, and number of patients who recovered to their baseline score four weeks after surgery. A change of at least 20 metres was considered to be clinically meaningful. ‐ Energy expenditure was determined using the Community Healthy Activity Model Programme for Seniors (CHAMPS) questionnaire. The results were given in kcal/kg/week (median, IQR) per time point in both groups. The results were dichotomised to light and moderate‐vigorous energy expenditure. ‐ SF‐36‐scores were presented as the total physical and total mental subscales (mean, SD) for all three time points. ‐ Anxiety and depression symptoms were assessed using HADS and were expressed as median with IQR for all three time points.  ‐ Mean compliance with the programme is presented as % for both groups. Both compliance to the in‐hospital training sessions and the self‐reported compliance to the home‐based exercises.
Notes	Trial registration number: NCT02502760 Funding source: this trial was funded by a research grant from the Peri Operative Program charitable foundation and a peer‐reviewed grant from the Rossy Cancer Network. Conflict of interest: Dr. Carli reported the grant from the Rossy Cancer Network. Dr. Liberman reported nonfinancial support from Servier Laboratories and personal fees from Ipsen, Merck & Co, and Pfizer, Inc. The sponsors had no role in the trial.
Risk of bias
Bias	Authors' judgement	Support for judgement
Random sequence generation (selection bias)	Low risk	Quote: "...achieved via computer‐generated random numbers..." Comment: probably done
Allocation concealment (selection bias)	Low risk	Quote: "...placed in sealed, opaque, consecutively numbered envelopes" Comment: probably done
Blinding of participants and personnel (performance bias) All outcomes	High risk	Quote: "...nature of the intervention, it was not possible to blind patients or intervention staff" Comment: No blinding. The authors did try to minimise performance bias; they did not present one of the programmes as potentially superior to the other.
Blinding of outcome assessment (detection bias) All outcomes	Low risk	Quote: "Outcome assessors, surgeons and statisticians were blinded to group assignment." Comment: probably done
Incomplete outcome data (attrition bias) All outcomes	High risk	Comment: protocol published as supplementary material. Sample size per group differs between protocol and article. Not all randomised patients were included in the intention‐to‐treat analysis (reasons for dropout were mentioned).
Selective reporting (reporting bias)	Low risk	Comment: protocol available in supplementary material. All outcomes discussed.
Other bias	Unclear risk	Comment: there is insufficient information to assess an other potential bias.