Study Design and Participants
This single-center randomized controlled trial was conducted at the Colorectal Cancer Clinic, a tertiary referral center in Seoul, Korea. We recruited participants between October 2013 and September 2016. Inclusion criteria were 1) stage Ⅰ-Ⅲ colon or rectal cancer; 2) 19–70 years of age; 3) American Society of Anesthesiologist grade ≤3 at surgery; and 4) the ability to read and understand Korean. Primary exclusion criteria included 1) evidence of recurrent or metastatic disease; 2) postsurgical intensive care unit stay; 3) presence of a stoma after colectomy; and 4) open surgery. The Institutional Ethics Review Board of Severance Hospital (Clinical trial number: KCT0003920) approved the trial. All participants provided informed consent before the initiation of any study-related procedures.
Outcomes and Data Collection
The primary outcome of this study was LOS, defined as the number of days between the day of surgery and the day of discharge. Patients were discharged when the following conditions were met: 1) vital sign stability (blood pressure, body temperature, pulse rate, and respiration rate); 2) soft diet tolerance; 3) clearing of the wound; 4) pain control; 5) no difficulty of voiding; and 6) passage of first stool. Target LOS duration for the current ERP was 5 days for colon cancer and 7 days for rectal cancers.
The secondary outcomes included patient-perceived readiness for hospital discharge (Pt-RHDS). The Pt-RHDS [20, 21] was translated into Korean and back-translated by two independent bilingual scholars to examine the construct validity among Korean colorectal cancer patients. After excluding a question with a similar meaning as another question, the Korean version of the Pt-RHDS questionnaire comprised 22 items that measured four domains. Personal status measured how the patient feels on the day of discharge. Knowledge measures the patients’ knowledge about self-care at home after discharge. Coping ability measured how the patient will be able to cope at home after discharge. Finally, expected support measured how much help the patient will have if/when needed at home after discharge). Each question was scored from 0 (not at all) to 10 (absolutely), and higher scores were interpreted as greater readiness for hospital discharge. To facilitate interpretation, the total score and each RHDS domain were converted to a standardized 100-point scale so that the maximum possible score at each level was 100.
We measured other secondary outcomes including anthropometric measures such as height, weight, waist circumference, and thigh circumference. We measured waist circumference using medical body measuring tape at the midpoint between the lower border of the 12th rib cage and the iliac crest. We measured thigh circumference at the midpoint of the thigh and body composition using bioelectrical impedance analysis (BIA) (Inbody 230, Biospace, Seoul, South Korea). Tests were completed on the day before the surgery and the day of hospital discharge (i.e., twice). Physical function measures included handgrip strength using a grip dynamometer (GRIP-D 5401, TKK, Japan), lower body strength (chair-stand test) and balance ability (time participant could stand on one leg).
Sample Size Calculation
We powered our study to detect a clinically meaningful difference in LOS between the exercise and usual care groups. We required a sample size of 52 to detect a medium effect size (Cohen’s d = 0.8) and the final sample size in this study was 64 (32 per group), which identified an effect size of d = 0.8, with 0.80 power, a two-tailed overall type I error rate of 0.05, and an expected dropout rate of 20%. The effect size used in our power calculation was from the previous study with similar exercise inpatient exercise protocol and the primary outcome measure in colorectal cancer patients (LOS = 7.82±1.07 days in the exercise group, 9.86±2.66 days in the usual care group; effect size = 0.77) [19]. The consensus among the clinical medical staff was that an effect size of this magnitude would be clinically meaningful.
Randomization
A total of 124 patients who met inclusion criteria were assessed for eligibility. Among them 60 patients were excluded due to inability to understand Korean, 2 medical issues, and 57 declined participation of the study. Out of 64 patients, we excluded one participant before randomization due to metastasis found during surgery. A total of 63 patients were randomly assigned to the exercise or usual care groups (1:1 ratio). Randomization was performed using a permuted block design, with stratification by age, sex, body mass index, and cancer type (colon vs. rectal). Allocation concealment was implemented by sequentially numbered, sealed, opaque envelopes. We used a minimization method to balance the prognostic factors between the groups [22, 23]. Study staff were not blinded to treatment assignment. Among the patients randomized, 11 (5 in the exercise group and 6 in the usual care group) were randomized but excluded from the trial due to ileostomy and did not receive any intervention.
Study Groups
Exercise Intervention
We implemented the institutional ERP protocol for all trial participants. The details of the ERP protocol are attached as supplementary material (Supplementary Table 1). Participants randomized to the exercise group engaged in a 15-minute supervised exercise intervention twice a day. A qualified exercise specialist supervised the exercise. The exercise program consisted of three phases according to postoperative day and patients' conditions.
The phase 1 exercise program started on postoperative day 1 (patient condition: patients experienced limited mobility due to pain; patients required help because of pain at the surgical site when sitting or lying down). During this phase, the exercises consisted of stretching and low-intensity resistance exercises (i.e., stretching the neck, shoulder, wrist, and ankle, pelvic stretching, and posterior pelvic tilt). We implemented phase 2 on postoperative days 2 and 3 (patient condition: patients able to perform daily activities without help or to be able to walk for >20 minutes at a time, but still experiencing discomfort). In this phase, in addition to phase 1 exercises, participants performed stretching, resistance exercises (i.e., leg raise, leg circle, bridge, and squeezing a ball with the thighs) and core resistance exercises (i.e., including arm circles, triceps extensions, and posterior pelvic tilt in the supine position). The phase 3 exercise program was performed between postoperative day 4 to hospital discharge (patient condition: can perform daily activities and self-care activities without discomfort. Phase 3 program was a continuation of Phase 2, with balancing exercises including one-leg standing, one-leg calf raises, hip adduction, hip abduction, hip extension, and hip flexion. Each movement was performed in sets of 10 repetitions for isotonic exercises or 10 seconds for isometric exercises.
The participants in the usual care group received the same preoperative and postoperative ERP as the exercise group (Supplementary Table 1). The usual care group did not receive the exercise program.
Statistical analysis
We used descriptive statistics describe and summarize the study sample. Primary and secondary analyses used intention-to-treat (ITT) analyses using multiple-imputation with expectation-maximization algorithm methods [24]. No participants were lost to follow up, so multiple-imputation methods were only used for two participants in the exercise group and one in the control group for RHDS measures. Per-protocol analyses were performed for sensitivity analyses (Supplementary Tables 2). Data normality was examined using the Kolmogorov-Smirnov test before initiating any statistical analyses. A chi-square test or independent t-test was used to test for differences between groups at baseline. All study outcomes were compared using analysis of covariance (after adjusting for sex, cancer type and stage), or the Wilcoxon signed-rank test or Mann-Whitney U-test, depending on the distribution of the data. When analysis of covariance was used, data was presented as mean ± standard deviation. When Mann- Whitney U-test was used, data was presented as median with interquartile range (IQR). Statistical significance was set at P < 0.05, and data were analyzed using SPSS version 20 (IBM Corp., Armonk, NY, USA).