Prehabilitation for patients undergoing endoscopic submucosal dissection: a randomized controlled trial

Abstract

Background

Endoscopic submucosal dissection (ESD) is a minimally invasive therapy for early gastrointestinal neoplasms, but patients may experience reduced functional capacity and delayed recovery post-procedure. Prehabilitation, has shown benefits in major surgery by improving physical fitness and recovery. We hypothesized that a structured prehabilitation regimen before ESD would enhance postoperative functional recovery and patient outcomes compared to standard care.

Methods

Single-center, assessor-blinded randomized controlled trial of 100 adults undergoing elective ESD, randomized to 4-week prehabilitation (supervised aerobic/resistance training 3×/week, individualized nutrition with protein supplementation, and psychological support) or usual care (standard instructions without structured exercise or nutrition). The primary outcome was change in six-minute walk distance (6MWD) from baseline to 1 month. Secondary outcomes included handgrip strength, quality of life (SF-36), psychological health (HADS), postoperative complications, length of stay, and 90-day readmissions.

Results

Baseline characteristics were similar between groups. At 1-month post-ESD, six-minute walk distance had decreased in both groups with no significant between-group difference (mean change from baseline − 40.9 m in prehabilitation vs. − 53.3 m in controls, p = 0.439). By 3 months, the prehabilitation group surpassed their baseline walking distance (18.9 m), whereas controls remained below baseline (− 14.1 m), indicating superior functional recovery with prehabilitation (p = 0.022). Prehabilitation also led to greater improvements in early postoperative strength and patient-reported outcomes. At 1 month, handgrip strength increased more in the prehabilitation group (mean 5.6 kg vs. 4.1 kg, p = 0.001). Anxiety symptoms improved significantly with prehabilitation (HADS anxiety score change − 1.17 vs. − 0.22, p = 0.003), with a parallel significant reduction in depression scores (p = 0.002). Prehabilitation patients reported higher gains in mental quality-of-life by 3 months (SF-36 mental component + 15.4 vs. + 11.4 points, p < 0.001). There were no significant differences in perioperative safety outcomes, average hospital stay, and 90-day readmission rates.

Conclusions

A 4-week multimodal prehabilitation program for ESD patients significantly improved postoperative functional capacity and psychological well-being without increasing complication rates or hospital stay. Prehabilitation is a safe and effective strategy to enhance recovery and patient-centered outcomes in ESD.

Clinical trial registration

This trial was retrospectively registered in the Chinese Clinical Trial Registry (ChiCTR) (registration number: ChiCTR2500103062) on May 23, 2025.

Level of evidence

Level I (Randomized Controlled Trial).

Supplementary Information

The online version contains supplementary material available at 10.1186/s12876-025-04418-5.

Background

Prehabilitation is a multimodal preoperative conditioning program (e.g., structured exercise, nutritional support, psychological preparation) designed to enhance patients’ fitness and resilience before surgery, thereby improving postoperative outcomes. A growing body of evidence, including systematic reviews and meta-analyses, indicates that prehabilitation can significantly reduce postoperative morbidity and accelerate recovery in major elective surgeries [1–4].

In addition to lowering complication rates and shortening hospital stays, prehabilitation has been associated with improved functional capacity and quality of life around the time of surgery. A network meta-analysis found that combined prehabilitation yielded superior postoperative six-minute walk distance and improved health-related quality of life compared to usual care [5]. Furthermore, a multicenter trial in colorectal surgery reported that a 4-week supervised prehabilitation program significantly reduced serious postoperative complications and hastened return to functional independence relative to standard care [6].

Prehabilitation may also confer economic benefits by offsetting the costs of complications and prolonged hospital stays. A systematic review of 45 economic evaluations found prehabilitation to be cost-effective in most analyses, with about 60% of studies showing improved outcomes and equal or lower overall costs compared to usual care [7].

Despite robust evidence in open surgery, the role of prehabilitation in endoscopic procedures such as Endoscopic Submucosal Dissection (ESD) remains undefined. ESD is a minimally invasive technique for removing early gastrointestinal neoplasms that avoids open resection, resulting in less surgical trauma and faster recovery. However, many ESD candidates are older or frail and may still experience sedation-related side effects, bleeding risk, or postoperative debility. No randomized trial has examined whether prehabilitation’s documented benefits in major surgery translate to ESD. This represents an important knowledge gap and an opportunity to extend prehabilitation principles to a new patient population.

The purpose of the study is to determine whether such prehabilitation can improve postoperative recovery outcomes in ESD patients, specifically by enhancing physical functional status, reducing postoperative complications, shortening the length of hospital stay, improving health-related quality of life, and potentially contributing to cost-effective care.

Methods

Study design and setting

This study was a single-center, assessor-blinded randomized controlled trial conducted at a tertiary teaching hospital (Shanghai Sixth People’s Hospital, Gastroenterology Department) from May 2022 to December 2024. It aimed to compare a 4-week preoperative prehabilitation program with standard care in patients undergoing endoscopic submucosal dissection (ESD). Patients were randomized in a 1:1 ratio to either the prehabilitation group or the control group. Due to the nature of the intervention, only outcome assessors were blinded (patients and clinicians delivering the intervention were aware of group allocation). The trial design and reporting adhere to the CONSORT 2025 guidelines for randomized trials.

Ethics

The study was approved by the Ethics Committee of Shanghai Sixth People’s Hospital (Approval No. 2022-KY-046 (K)), and was retrospectively registered in the Chinese Clinical Trial Registry (ChiCTR) (registration number: ChiCTR2500103062) on May 23, 2025. All procedures were conducted in accordance with the Declaration of Helsinki and relevant institutional guidelines. Written informed consent was obtained from all individual participants before enrollment in the study.

Participants

Eligible participants were adult patients (aged ≥ 18 years) scheduled for elective ESD to resect early gastrointestinal lesions (upper or lower GI tract, such as early gastric neoplasia, colon polyps, early colorectal tumors, or superficial esophageal lesions). The inclusion and exclusion criteria were shown in Table 1.

Table 1.

Eligibility criteria

Inclusion Criteria

Exclusion Criteria

1) Adult patients (≥ 18 years) scheduled for an elective ESD (Endoscopic Submucosal Dissection) to resect an early gastrointestinal lesion. 2) At least a 4-week interval before the planned ESD 3) Sufficient baseline functional capacity to engage in light-to-moderate exercise training. 4) Willingness to participate and comply with the study

1) Patients requiring urgent or emergent ESD who cannot delay the procedure for 4 weeks are excluded 2) Patients with serious comorbidities that would prevent safe participation in an exercise program are excluded: • Unstable cardiovascular disease (e.g. recent myocardial infarction, uncontrolled angina, or severe aortic stenosis) or uncontrolled heart failure. • Advanced pulmonary disease precluding exercise (e.g. severe COPD with inability to tolerate activity). • Uncontrolled hypertension (blood pressure too high to exercise safely despite medication). • Severe musculoskeletal or neurological disorders that prevent basic mobility or exercise (e.g. patients who are wheelchair-bound and unable to perform the required physical activities). 3) Patients with conditions such as advanced dementia or uncontrolled severe psychiatric disorders that impair daily functioning are excluded 4) Patients already participating in a structured prehabilitation or high-intensity exercise program, or those who have recently completed such a program, are excluded 5) Any uncontrolled condition that could compromise the intervention or study outcomes will exclude the patient. Examples include: • Active infection or high fever. • Advanced malignancy or other advanced disease with an expected survival of < 6 months. • Patients currently receiving hospice/palliative care. • Need for another major surgical procedure during the study period (which could confound the results of the prehabilitation and ESD outcomes). 6) Patients who are pregnant are excluded

Participants

Eligible participants were adult patients (aged ≥ 18 years) scheduled for elective ESD to resect early gastrointestinal lesions (upper or lower GI tract, such as early gastric neoplasia, colon polyps, early colorectal tumors, or superficial esophageal lesions). The inclusion and exclusion criteria were shown in Table 1.

Interventions

Prehabilitation group (prehab)

Patients randomized to the prehabilitation arm underwent a comprehensive 4-week preoperative rehabilitation program in the weeks leading up to their ESD. This multimodal prehabilitation program was designed to optimize patients’ physical and psychological status before surgery, and it comprised several components (Supplementary Materials).

The prehabilitation program was delivered by a multidisciplinary team, including rehabilitation physicians or physiotherapists (overseeing exercise training), dietitians, and psychologists, under the supervision of the surgical team. Each patient’s program was individualized based on an initial assessment of physical ability, nutritional status, and psychological needs, ensuring that the intervention was safe and appropriately challenging. Throughout the 4 weeks, the research team maintained weekly contact (in-person or by phone) and reviewed patient exercise logs to monitor adherence and provide encouragement or adjustments as needed. The prehabilitation intervention continued up until the day before surgery, totaling approximately 4 weeks of duration for each patient. Any adverse events or difficulties encountered during the prehab period were documented.

Control group (usual care)

Patients randomized to the control arm received the usual preoperative care and standard educational instructions provided to ESD patients at our institution. This usual care program was shown in supplementary materials.

The standard preoperative education was typically delivered in a single session (during the pre-procedure clinic visit) and supplemented with written pamphlets. Importantly, patients in the control group did not receive any structured exercise program, supervised training, or specialized nutritional supplementation during the preoperative period beyond general healthy lifestyle advice. They were instructed to continue with their usual daily activities as they normally would while awaiting surgery.

All participants in both groups were expected to follow the regimen assigned to them. The study team tracked compliance in both arms; for the prehab group this included attendance at training sessions and adherence to nutrition plans, while for the control group any engagement in additional exercise or interventions was noted. Any significant deviations from the assigned protocol or use of extra interventions were recorded for analysis.

Outcomes and data collection

All outcomes were assessed at baseline (preoperatively) and at 4, 12, and 24 weeks after the ESD procedure. The trial’s primary outcome was functional exercise capacity, measured by the distance walked in the 6-minute walk test (6MWT) [8]. This primary endpoint was prespecified at 4 weeks postoperatively, reflecting early post-surgical recovery. The 6MWT was chosen as it is a validated objective test of aerobic endurance and functional recovery in postoperative patients. Follow-up 6MWT assessments at 12 and 24 weeks were also recorded to evaluate longer-term functional outcomes.

Secondary outcomes included a range of clinical, physical, and patient-reported measures, each evaluated at the same postoperative time points (4, 12, and 24 weeks):

1. Body mass index (BMI), calculated as weight (kg) divided by height squared (m²), was tracked to assess nutritional status [9].

2. Postoperative complications (Clavien–Dindo grade ≥ II), defined as any adverse events requiring at least pharmacological, endoscopic, or surgical intervention, were recorded to compare safety outcomes between groups. This includes major ESD-related complications such as significant gastrointestinal bleeding, perforation, or stricture requiring intervention [10]. The incidence of such complications was assessed cumulatively up to each follow-up point.

3. Length of hospital stay (LOS) for the index hospitalization (in days) was documented as an indicator of immediate postoperative recovery and resource use. A shorter postoperative LOS was hypothesized for the prehabilitation group due to improved recovery [11].

4. Unplanned readmissions, defined as any unscheduled hospital readmission after the initial discharge, were tracked at each follow-up interval. This outcome captures post-discharge complications or issues necessitating return to hospital care. Results were expressed as the proportion of patients with ≥ 1 unplanned readmission by 4, 12, and 24 weeks [12].

5. Hand grip strength, measured in kilograms with a dynamometer, was used as an index of muscle strength and physical function. Grip strength was measured on the dominant hand and recorded at baseline and follow-up visits, as it correlates with overall frailty and recovery [13].

6. Fried Frailty Index, a composite frailty score based on the five criteria of the Fried frailty phenotype (unintentional weight loss, exhaustion, low physical activity, slowness, and weakness), was assessed to evaluate each patient’s frailty status. This index ranges from 0 to 5, with higher scores indicating greater frailty. Changes in the frailty index from baseline were monitored over the 24-week postoperative period [14].

7. Health-related quality of life was measured using the 36-Item Short Form Health Survey (SF-36). In particular, the Physical Component Summary (PCS) and Mental Component Summary (MCS) scores were analyzed as continuous outcomes. These summary scores (derived from the SF-36 subscales) reflect overall physical and mental well-being, respectively, with higher scores denoting better health status. SF-36 questionnaires were administered at baseline and at 4, 12, and 24 weeks post-ESD [15].

8. Psychological status was evaluated with the Hospital Anxiety and Depression Scale (HADS), specifically its anxiety and depression subscale scores. HADS is a validated 14-item questionnaire (7 items each for anxiety and depression; score range 0–21 per subscale) used to detect mood disorders in medical patients. Both HADS subscale scores were recorded at each follow-up to assess the impact of prehabilitation on postoperative anxiety and depressive symptoms [16].

All outcome assessors (blinded to participant group allocation) received standardized training in the administration of each outcome instrument prior to participant enrollment. Outcome data were collected at multiple time points: at baseline (pre-intervention) and at 1, 3, and 6 months after ESD, corresponding to schedule in-person follow-up visits. At each follow-up visit, trained assessors conducted the functional tests (6MWT and grip strength) and supervised the completion of patient-reported outcome questionnaires, ensuring standardized administration of all measures. All data were recorded on paper-based case report forms (CRFs) at the time of collection and stored securely for subsequent analysis.

For participants in the prehabilitation group, adherence to the 4-week pre-surgical intervention was actively monitored through weekly follow-up phone calls and review of patient-completed exercise logs. Adherence to the nutritional component was monitored by the dietitian through weekly follow-up phone calls, during which patients reported their intake of the prescribed protein supplements and compliance with dietary recommendations. Additionally, at the end of the 4-week program, participants completed a questionnaire rating how closely they followed the nutrition plan. Participants in both groups were also asked to maintain symptom diaries throughout the study to document any health issues or adverse events between visits. These diaries, alongside routine follow-up evaluations, facilitated comprehensive adverse event monitoring and were used to cross-verify self-reported symptoms with clinical assessments. All recorded outcomes and events were ultimately verified and documented in the CRFs by the blinded outcome assessors, ensuring consistency and completeness of the trial data.

Sample size calculation

The sample size for this trial was determined based on the primary outcome (6MWT distance at 1 month). We anticipated a between-group difference of 30 m in the 6MWT, which represents the minimal clinically important difference according to previous studies [8]. Assuming a standard deviation of approximately 8 m for the 6MWT measure, we calculated that a total of 56 patients (28 per group) would provide 80% power to detect a 30 m difference between the two groups at a two-sided significance level of α = 0.05. To account for potential dropouts or loss to follow-up (estimated at about 30%), we planned to recruit 80 patients in total (40 patients per group). This sample size was deemed sufficient to test the primary hypothesis while maintaining statistical power in the event of some attrition.

Randomization, allocation and blinding

Randomization was performed using a computer-generated random sequence to assign participants to either the prehabilitation group or the control group in a 1:1 allocation ratio. An independent researcher who was not involved in patient recruitment or outcome assessment prepared the randomization sequence. To ensure allocation concealment, the assignments were sealed in sequentially numbered opaque envelopes. When a patient consented and was ready for enrollment, the next envelope in sequence was opened to reveal the group assignment. This process prevented the enrolling clinicians from predicting or influencing the upcoming allocations.

Blinding: Given the nature of the intervention, it was not feasible to blind the participants or the healthcare providers delivering the prehabilitation (patients in the prehab group actively participated in exercise and received coaching, which by necessity made them aware of their group). However, the outcome evaluators were kept blinded to group assignments. The staff conducting follow-up assessments, for example, the personnel administering the 6MWT at 1 month and collecting questionnaire data, did not know whether a patient had undergone prehabilitation or standard care. Participants were instructed not to reveal their group during assessments. This assessor blinding was implemented to reduce bias in outcome measurement. In summary, the trial employed a single-blind design (blinded outcome assessment) to maximize objectivity in the collected data.

Statistical analysis

All analyses were conducted following the intention-to-treat (ITT) principle, including all randomized patients in the groups to which they were assigned, regardless of protocol adherence or dropout. For the primary outcome and other key endpoints, any missing data were addressed using multiple imputation methods in order to minimize bias and preserve statistical power. We used multiple imputation by chained equations to handle missing outcome data under the assumption that data were missing at random, generating several imputed datasets and pooling the results for inference.

For continuous outcomes measured at multiple time points (such as the 6MWT distance and quality-of-life scores over the follow-up), we employed a linear mixed-effects model for repeated measures to compare differences between the two groups over time. This modeling approach accounts for the within-patient correlation of repeated observations. In our model, we included fixed effects for treatment group (prehabilitation vs. control), time (as a categorical variable for each postoperative follow-up point), and the interaction term group × time. The inclusion of the interaction term allowed us to assess whether the trajectory of recovery over time differed between the two groups. A patient-specific random intercept was used to model within-subject correlations across the repeated measurements at the three postoperative time points (1, 3, and 6 months after surgery), recognizing that measurements from the same individual are not independent. From this mixed model, we obtained estimates of the between-group difference in 6MWT distance at 1 month (primary endpoint) as well as at later time points, adjusted for baseline, along with 95% confidence intervals and p-values. A similar analytic approach was used for other repeated continuous outcomes (EQ-5D index values and other functional tests over time). For binary outcomes (complication occurrence or readmission), comparisons between groups were made using chi-square tests or Fisher’s exact test, and risk differences or odds ratios with 95% confidence intervals were reported. Time-to-event outcomes (if any were considered, such as time to discharge) would be analyzed with appropriate survival analysis techniques, but in this study most outcomes were cross-sectional or measured at set intervals.

No subgroup analyses were planned for this trial, as the study sample size was modest and the aim was to evaluate the overall effect in the general study population. To assess the robustness of the primary analysis, we performed a sensitivity analysis on the per-protocol population, defined as those participants who completed the intervention as allocated (for the prehab group, those who attended the majority of prehabilitation sessions with no major protocol violations, and for the control group, those who did not engage in any unplanned prehabilitation activities). The primary outcome (6MWT at 1 month) was re-analyzed in this per-protocol subset to check if the results were consistent with the ITT analysis.

All statistical tests were two-sided, and a p-value < 0.05 was considered statistically significant for hypothesis testing. Data were analyzed using IBM SPSS Statistics software (Version 27; IBM Corp., Armonk, NY). The trial results are reported with effect size estimates and confidence intervals where applicable. No interim analysis was performed, and the final analysis was carried out after completion of data collection for all participants.

Results

Participants

The flow chart of the study was shown in Fig. 1. The intention-to-treat population consisted of 100 patients (50 per group). Baseline demographic and clinical characteristics were similar between the prehabilitation and usual care arms (Table 2). Both groups had 29 male participants (58.00%), and mean age was 41 (9.10) vs. 40 (7.93) years in the prehabilitation versus usual care groups. Mean body mass index was identical in the two groups at 21 (2.08) vs. 21 (2.56) kg/m², and baseline functional capacity (6-minute walk distance 551 (50.06) vs. 555 (55.07) m) was likewise comparable (Table 2). Baseline patient-reported outcome measures were nearly identical in both groups. A sensitivity analysis of the per-protocol population confirmed that baseline characteristics remained balanced between groups (Supplementary Table 1).

Fig. 1.

CONSORT 2025 Flow Diagram of the study. Flow diagram of the progress through the phases of a randomised trial of two groups (that is, enrolment, intervention allocation, follow-up, and data analysis)

Table 2.

Baseline characteristics of the prehab group and usual care group (intention-to-treat population)

ITT	Sample characteristic	Prehab (N = 50)	Usual care (N = 50)	P value
Baseline characteristics	Male (no.[%])	29 (58.00)	29 (58.00)	1.000
	Age (yrs)	41 (9.10)	40 (7.93)	0.476
	BMI (kg/m²)	21 (2.08)	21 (2.56)	0.864
	Education Level (no.[%])
	Below high school	25 (50.00)	25 (50.00)	1.000
	≥high school	25 (50.00)	25 (50.00)
	Insurance Type (no.[%])
	Government	31 (62.00)	34 (68.00)	0.790
	Commercial	15 (30.00)	12 (24.00)
	Self-financed	4 (8.00)	4 (8.00)
	Current Smoker (no.[%])	2 (4.00)	2 (4.00)	1.000
	Current Alcohol use (no.[%])	2 (4.00)	2 (4.00)	1.000
	CRP level(mg/L)	14 (3.93)	15 (3.91)	0.741
	Albumin level(g/dL)	33 (3.74)	33 (3.70)	0.708
	Hemoglobin level(g/dL)	107 (7.96)	108 (10.88)	0.722
Underlying diagnosis prompting ESD	Early Gastric Lesions (no.[%])	19 (38.00)	14 (28.00)	0.288
	Esophageal Lesions (no.[%])	12 (24.00)	26 (52.00)	0.004
	Colorectal Lesions (no.[%])	5 (10.00)	4 (8.00)	0.727
	Duodenal Lesions (no.[%])	14 (28.00)	6 (12.00)	0.046
	Selected Submucosal Tumors (no.[%])	50 (100.00)	50 (100.00)	1.000
Medical History	Diabetes type 2 (no.[%])	2 (4.00)	2 (4.00)	1.000
	Hypertension (no.[%])	7 (14.00)	4 (8.00)	0.338
	Cardiovascular diseases (no.[%])	7 (14.00)	6 (12.00)	0.766
	Atrial fibrillation (no.[%])	2 (4.00)	2 (4.00)	1.000
	Obstructive Sleep Apnea (OSA) (no.[%])	2 (4.00)	2 (4.00)	1.000
	Helicobacter pylori, HP (no.[%])	13 (26.00)	13 (26.00)	1.000
	Hepatitis (no.[%])	16 (32.00)	15 (30.00)	0.829
	Gastroesophageal Reflux Disease, GERD (no.[%])	17 (34.00)	19 (38.00)	0.677
	Baseline 6MWT distance(m)	551 (50.06)	555 (55.07)	0.740
	Grip strength (kg force)	30 (4.15)	30 (4.64)	0.939
	Fried Frailty index	1 (0.91)	1 (0.95)	0.915
Patient-Reported Outcome Measures	Total Physical SF-36 subscale	59 (6.14)	59 (5.81)	0.640
	Total Mental SF-36 subscale	55 (4.53)	55 (4.69)	0.746
	HADS Anxiety score	13 (3.03)	13 (2.92)	0.789
	HADS Depression score	13 (2.52)	13 (2.00)	0.222

ITT intention-to-treat, BMI body mass index, CRP C-reactive protein, 6MWT 6-minutes walking test, SF-36 36-Item Short Form Survey, HADS Hospital Anxiety and Depression Scale

Primary and secondary outcomes

In the intention-to-treat population, the primary outcome of 6MWT distance improved significantly more in the prehabilitation group (Table 3; Fig. 2). At 12 weeks, prehabilitated patients increased their 6MWT distance by 18.94 m from baseline on average, whereas the usual care group declined by 14.09 m (between-group difference: 11.74 m, 95% CI 2.26 to 21.22; p = 0.015). There was no significant difference at 4 weeks (both groups showed a decline, p = 0.439), and by 24 weeks the between-group difference had attenuated (difference: 8.33 m, 95% CI − 0.96 to 17.61; p = 0.079). Among secondary outcomes, prehabilitation led to greater early improvements in physical function and mental health (Table 4). Grip strength at 4 weeks increased by 5.58 kg in the prehab group vs. 4.10 kg in controls (difference: 0.67 kg, 95% CI − 0.29 to 1.64; p = 0.001), although this difference was no longer significant by 12 or 24 weeks (p = 0.135 and 0.298). The SF-36 mental component score was significantly higher in the prehab group at both 4 and 12 weeks (4-week difference: 0.62, 95% CI − 0.996 to 2.231; p = 0.010; 12-week difference: 1.78, 95% CI 0.24 to 3.32; p = 0.024), with the advantage diminishing by week 24 (p = 0.052). In contrast, the SF-36 physical component improved similarly in both groups (12-week difference: − 0.05, 95% CI − 1.89 to 1.80; p = 0.959). Prehabilitation also yielded greater short-term reductions in anxiety and depression: at 4 weeks, HADS anxiety and depression scores decreased more in the prehab arm (Anxiety difference: − 0.64, 95% CI − 1.61 to 0.33; p = 0.003; Depression difference: − 0.07, 95% CI − 0.88 to 0.74; p = 0.002), although between-group differences in these psychological scores were no longer significant at 12 or 24 weeks (e.g., HADS-A p = 0.098 at 12 weeks). There were no significant differences between groups in BMI change, incidence of ESD-related complications, length of hospital stay, unplanned readmission rates, or frailty index at any time point (all p > 0.05). Sensitivity analyses using the per-protocol population (Supplementary Fig. 1, Supplementary Tables 2, and Supplementary Table 3) yielded consistent results, with similar effect estimates and significance levels as the main analysis.

Fig. 2.

Changes of primary and secondary outcomes in each follow-up time point (error bars represent 95 CIs) based on intention-to-treat population

Table 3.

Changes in outcomes of the prehab group and usual care group at week 4, week 12 and week 24 (intention-to-treat population)

ITT	4 week			12 week			24 week
	Prehab (N = 50)	Usual care (N = 50)	P value	Prehab (N = 50)	Usual care (N = 50)	P value	Prehab (N = 50)	Usual care (N = 50)	P value
6MWT distance(m)	−40.94 (62.39)	−53.28 (89.62)	0.439	18.94 (68.06)	−14.09 (68.62)	0.022	4.88 (62.47)	−0.11 (78.77)	0.734
BMI (kg/m²)	0.53 (0.21)	0.56 (0.26)	0.494	0.73 (0.27)	0.72 (0.33)	0.979	0.93 (0.86)	1.13 (0.42)	0.159
Number of complicationsa	0.06 (0.24)	0.07 (0.25)	0.958	0.13 (0.34)	0.20 (0.40)	0.378	0.00 (0.00)	0.02 (0.15)	0.310
LOS (days)	0.08 (0.35)	0.11 (0.43)	0.754	0.09 (0.35)	0.11 (0.38)	0.756	0.00 (0.00)	0.04 (0.29)	0.310
Unplanned readmissionb (no.[%])	6.25%	6.52%	0.957	6.38%	8.70%	0.673	0.00%	2.17%	0.304
Grip strength (kg)	5.58 (2.06)	4.10 (2.12)	0.001	5.59 (3.07)	4.76 (2.52)	0.159	4.94 (3.40)	5.12 (2.15)	0.756
Fried Frailty index	−0.10 (1.12)	0.15 (1.19)	0.284	−0.15 (1.22)	0.17 (1.23)	0.207	−0.50 (1.09)	−0.50 (1.09)	1.000
Total Physical SF-36 subscale	6.08 (0.58)	6.02 (0.61)	0.617	12.98 (1.11)	11.11 (1.14)	0.000	15.00 (1.19)	14.61 (1.24)	0.121
Total Mental SF-36 subscale	6.35 (3.30)	5.07 (0.57)	0.010	15.43 (3.81)	11.41 (1.15)	0.000	18.60 (3.90)	18.43 (1.15)	0.778
HADS Anxiety score	−1.17 (1.58)	−0.22 (1.43)	0.003	−4.21 (1.67)	−3.30 (1.58)	0.008	−5.67 (2.07)	−5.59 (1.96)	0.848
HADS Depression score	−1.25 (1.41)	−0.20 (1.72)	0.002	−4.23 (1.59)	−3.33 (1.84)	0.012	−6.02 (1.84)	−5.67 (1.62)	0.335

ITT intention-to-treat, BMI body mass index, LOS length of hospital stays, 6MWT 6-minutes walking test, SF-36 36-Item Short Form Survey, HADS Hospital Anxiety and Depression Scale

^aComplications include Gastrointestinal bleeding, perforation, esophageal stenosis

^bFor the “Unplanned readmission” variate, we calculate the difference between the proportion of times of unplanned readmission at weeks 4, 12, and 24 with the proportion in baseline

Table 4.

Effectiveness estimates from linear mixed effects models of the prehab group and usual care group at week 4, week 12 and week 24 (intention-to-treat population)

ITT	4 week			12 week			24 week
	Coefficient	95% CI	P value	Coefficient	95% CI	P value	Coefficient	95% CI	P value
6MWT distance(m)	3.169	(−8.473, 14.811)	0.439	11.736	(2.255, 21.218)	0.015	8.325	(−0.964, 17.614)	0.079
BMI (kg/m²)	0.558	(−0.114, 1.231)	0.494	0.643	(−0.050, 1.336)	0.069	0.631	(−0.066, 1.329)	0.076
Number of complicationsa	0.002	(−0.047, 0.051)	0.958	−0.017	(−0.074, 0.039)	0.542	−0.019	(−0.066, 0.027)	0.409
LOS (days)	−0.016	(−0.084, 0.053)	0.754	−0.026	(−0.092, 0.041)	0.448	−0.031	(−0.085, 0.024)	0.268
Unplanned readmissionb (no.[%])	−0.021	(−1.647, 1.604)	0.957	−0.186	(−1.316, 0.943)	0.746	/	/	/
Grip strength (kg)	0.673	(−0.290, 1.636)	0.001	0.727	(−0.227, 1.682)	0.135	0.500	(−0.442, 1.442)	0.298
Fried Frailty index	−0.023	(−0.260, 0.215)	0.284	−0.116	(−0.309, 0.077)	0.239	−0.072	(−0.251, 0.106)	0.427
Total Physical SF-36 subscale	−0.478	(−2.385, 1.428)	0.617	−0.048	(−1.893, 1.796)	0.959	−0.139	(−2.018, 1.740)	0.885
Total Mental SF-36 subscale	0.618	(−0.996, 2.231)	0.010	1.778	(0.235, 3.321)	0.024	1.499	(−0.011, 3.009)	0.052
HADS Anxiety score	−0.638	(−1.606, 0.331)	0.003	−0.804	(−1.757, 0.149)	0.098	−0.695	(−1.583, 0.192)	0.125
HADS Depression score	−0.073	(−0.882, 0.735)	0.002	−0.287	(−1.060, 0.486)	0.467	−0.249	(−0.984, 0.486)	0.507

ITT intention-to-treat, CI confidential interval, BMI body mass index, LOS length of hospital stays, 6MWT 6-minutes walking test, SF-36 36-Item Short Form Survey, HADS Hospital Anxiety and Depression Scale

^aComplications include Gastrointestinal bleeding, perforation, esophageal stenosis

^bFor the “Unplanned readmission” variate, we calculate the difference between the proportion of times of unplanned readmission at weeks 4, 12, and 24 with the proportion in baseline

Cost-effectiveness analysis

At 24 weeks, the mean total cost per patient was CNY 2707 (SD 59.90) in the prehabilitation group compared to CNY 2008 (SD 211.73) in the usual care group (p = 0.000; Table 5). This difference was driven by intervention-related costs in the prehabilitation group (CNY 1980 per patient vs. CNY 0 in usual care), partially offset by lower patient transportation costs (CNY 394 vs. CNY 2008, respectively; Table 5). Participants in the prehabilitation group were encouraged to walk or cycle to the hospital as part of their exercise regimen, and many reported opting for walking or biking instead of using paid transportation (e.g., taxis, buses) for their hospital visits, which likely contributed to the markedly lower transportation expenses in the prehabilitation group; notably, no formal transportation assistance was provided to either group, and all travel costs were collected via patient self-report. The resulting incremental cost at 24 weeks was CNY 699 per patient (intention-to-treat analysis) and CNY 708 (per-protocol; Table 6). Coupled with improved outcomes in the prehabilitation group, this yielded an ICER of CNY 84 per additional meter walked in the 6-minute walk test (8.325 m gain; ITT) and CNY 466 per one-point increase in the SF-36 mental health score (1.499-point gain; ITT; Table 6). Prehabilitation also reduced complications (−0.019 per patient; ITT) and improved frailty (−0.073) and anxiety (−0.695) scores relative to usual care, resulting in negative ICER values (e.g., −35924 for complications; ITT; Table 6). Results were similar in the per-protocol analysis (Tables 5 and 6).

Table 5.

Costs of the prehab group and usual care group at week 24

ITT	Prehab (N = 50)	Usual care (N = 50)	P value
Total cost	2707 (59.90)	2008 (211.73)	0.000
Intervention-related costs (CNY)	1980 (0.00)	0 (0.00)	/
Personnel cost (PT time, coaching hours) (CNY)	0 (0.00)	0 (0.00)	/
App development & maintenance (per patient) (CNY)	120 (0.00)	0 (0.00)	/
Facility and equipment usage (CNY)	141 (29.71)	0 (0.00)	/
Resistance bands/Exercise equipment provided (CNY)	71 (26.79)	0 (0.00)	/
Patient transportation costs (CNY)	394 (51.16)	2008 (211.73)	0.000

PP	Prehab ( N  = 44)	Usual care ( N  = 43)	P value
Total cost	2708 (60.25)	2000 (215.50)	0.000
Intervention-related costs (CNY)	1980 (0.00)	0 (0.00)	/
Personnel cost (PT time, coaching hours) (CNY)	0 (0.00)	0 (0.00)	/
App development & maintenance (per patient) (CNY)	120 (0.00)	0 (0.00)	/
Facility and equipment usage (CNY)	141 (27.91)	0 (0.00)	/
Resistance bands/Exercise equipment provided (CNY)	71 (27.09)	0 (0.00)	/
Patient transportation costs (CNY)	397 (49.57)	2000 (215.50)	0.000

ITT intention-to-treat, PP per-protocol, CNY Chinese Yuan, PT physiotherapist, APP application (on Mobile)

Table 6.

Incremental outcome and incremental cost-effectiveness ratio

Incremental Outcome	ITT	PP
Incremental 6MWT distance(m)	8.325 (−0.964, 17.614)	9.572 (−0.842, 19.986)
Incremental BMI (kg/m²)	0.631 (−0.066, 1.329)	0.541 (−0.111, 1.192)
Incremental Number of complicationsa	−0.019 (−0.066, 0.027)	−0.010 (−0.051, 0.032)
Incremental LOS (days)	−0.031 (−0.085, 0.024)	−0.022 (−0.071, 0.026)
Incremental Unplanned readmissionb (no.[%])	/	/
Incremental Grip strength (kg force)	0.500 (−0.442, 1.442)	0.719 (−0.254, 1.691)
Incremental Fried Frailty index	−0.073 (−0.251, 0.106)	−0.104 (−0.289, 0.081)
Incremental Total Physical SF-36 subscale	−0.139 (−2.018, 1.740)	−0.947 (−2.802, 0.909)
Incremental Total Mental SF-36 subscale	1.499 (−0.011, 3.009)	1.163 (−0.427, 2.753)
Incremental HADS Anxiety score	−0.695 (−1.583, 0.192)	−0.666 (−1.550, 0.217)
Incremental HADS Depression score	−0.249 (−0.984, 0.486)	−0.302 (−1.070, 0.466)
Overall Incremental Cost	699	708

ICER	ITT	PP
6MWT distance(m)	84	74
BMI (kg/m²)	1107	1310
Number of complicationsa	−35,924	−73,206
LOS (days)	−22,891	−31,491
Unplanned readmissionb (no.[%])	/	/
Grip strength (kg force)	1397	985
Fried Frailty index	−9678	−6819
Total Physical SF-36 subscale	−5038	−748
Total Mental SF-36 subscale	466	609
HADS Anxiety score	−1006	−1062
HADS Depression score	−2806	−2345

ITT intention-to-treat, PP per-protocol, ICER Incremental cost-effectiveness ratio, BMI body mass index, LOS length of hospital stays, 6MWT 6-minutes walking test, SF-36 36-Item Short Form Survey, HADS Hospital Anxiety and Depression Scale

^aComplications include Gastrointestinal bleeding, perforation, esophageal stenosis

^bFor the “Unplanned readmission” variate, we calculate the difference between the proportion of times of unplanned readmission at weeks 4, 12, and 24 with the proportion in baseline

Adherence to intervention

Adherence to the prehabilitation program was high among participants in the intervention group (Table 7). On average, patients completed 26.3 (6.5) of 30 exercise log entries and 3.1 (1.2) of 4 weekly self-assessments. They received 3.3 (0.8) follow-up phone calls over 4 weeks. Patient-reported adherence and satisfaction ratings were uniformly high across all domains, with mean scores ranging from 4.6 to 4.8 out of 5 for satisfaction and for following exercise, nutrition, psychological, and educational components of the program.

Table 7.

Adherence to prehabilitation program (intention-to-treat population)

ITTa	Prehab (N = 50)
Completion of daily exercise logbooks (totally 30 times)	26.3 (6.5)
Completion of weekly self-examination on logbooks (totally 4 times)	3.1 (1.2)
Times of receiving follow-up phone calls (4 times)	3.3 (0.8)
b To what extent were you satisfied with the exercise protocol?	4.7 (0.5)
b To what extent did you do the exercise program as recommended?	4.8 (0.4)
b To what extent were you satisfied with the Nutritional support?	4.6 (0.6)
b To what extent did you do the nutritional support as recommended?	4.8 (0.6)
b To what extent were you satisfied with the psychological support?	4.6 (0.5)
b To what extent did you do the psychological support as recommended?	4.8 (0.5)
b To what extent were you satisfied with the preoperative education?	4.7 (0.5)
b To what extent did you do the preoperative education as recommended?	4.8 (0.5)

ITT intention-to-treat

^aAdherence data are reported as mean values with standard deviations in parentheses

^*The survey of ‘Satisfied with prehabilitation program ‘was conducted at outpatient clinic at the 4-week follow-up using a questionnaire (each question on the questionnaire uses a 5-point Likert scale, 1 = Strongly Satisfied, 5 = Strongly Dissatisfied)

Discussion

Our study evaluated the effects of a 4-week multimodal prehabilitation program in patients scheduled for elective endoscopic submucosal dissection. The primary outcome, change in 6-minute walk test distance at 4 weeks postoperatively, showed no significant between-group difference. However, greater functional gains in the prehabilitation group were observed at 12 weeks, which did not persist to 24 weeks. Improvements in grip strength, mental health, and anxiety and depression scores were also more pronounced in the prehabilitation group during early follow-up, though most differences diminished over time. No significant differences were observed in postoperative complication rates, hospital length of stay, unplanned readmissions, or frailty scores. The total cost associated with the prehabilitation intervention was higher than usual care, driven by program-related expenses. Incremental cost-effectiveness ratios suggested favorable cost per unit benefit for selected outcomes, including functional recovery and mental health.

Our findings are consistent with other literature on surgical prehabilitation, reinforcing that a multimodal prehabilitation program can enhance perioperative outcomes without added risk. In major abdominal and oncologic surgery, numerous trials and meta-analyses have demonstrated that prehabilitation improves patients’ functional capacity and may reduce postoperative complications. For example, a recent meta-analysis of 27 RCTs in surgical oncology found significant gains in preoperative 6-minute walk distance and an 40% reduction in overall complication risk with prehabilitation vs. standard care [17]. Likewise, a large multicenter trial in colorectal cancer (PREHAB study) reported that 4 weeks of multimodal prehabilitation led to fewer severe complications (17% vs. 30%) and medical complications compared to usual perioperative care [18]. These broad improvements, seen across diverse major abdominal surgeries, align with the positive trends observed in our cohort and suggest that prehabilitation effectively bolsters patients’ physiological reserves before surgery.

Importantly, our results also resonate with evidence from similar high-risk populations such as the elderly and frail. Meta-analyses focusing on older adults undergoing major abdominal surgery have concluded that multimodal prehabilitation significantly increases preoperative functional capacity and can shorten hospital length of stay and reduce complication rates in frail or high-risk patients [19]. Pang et al. reported a one-third reduction in postoperative complications among patients ≥ 65 years old who underwent prehabilitation, along with meaningful improvements in walking capacity after surgery [20]. Such findings underscore that even in vulnerable groups with limited physiologic reserve, the benefits of prehabilitation are attainable. In our study’s relatively early-stage GI tumor population (many of whom were older and/or frail), we similarly noted improvements in functional outcomes without an increase in complications, supporting the applicability of prehabilitation in this subgroup. That said, the literature also highlights the importance of balancing enthusiasm with realistic expectations. Not all trials have shown dramatic outcome differences; for instance, Carli et al. found no significant reduction in 30-day complications with prehabilitation in a frail colorectal cancer cohort managed within an enhanced recovery pathway [19]. The contrasting results might be explained by differences in patient characteristics, surgical approaches (minimally invasive techniques and optimized ERAS protocols can lower baseline complication rates), and study design (prehabilitation versus postoperative rehabilitation timing). Such nuances remind us that while prehabilitation tends to be beneficial on average, its impact can vary, and identifying which patients benefit most is an important consideration for future research.

On balance, the accumulated evidence, including our current study, suggests that the benefits of prehabilitation outweigh potential drawbacks for suitable surgical candidates. We observed no intervention-related adverse events or delays to surgery in our program, echoing reports from other trials that prehabilitation is safe and well-tolerated. Potential harms, such as exercise-related injuries or the risk of postponing definitive cancer treatment, were minimal in our experience. In fact, for early-stage cancer patients, a short preoperative conditioning period can be incorporated into the usual wait for surgery without compromising oncologic timelines [18]. Patients in the prehabilitation group tended to have better preserved functional status and quality of life through the perioperative period in our study, which may translate to easier recoveries and possibly long-term health benefits. These positive effects must be weighed against the resources and effort required: multimodal prehabilitation demands patient engagement and multidisciplinary support (exercise coaching, nutritional supplements, etc.), which can be challenging for some. Nonetheless, considering the overall improvements in outcomes and no increase in complications or readmissions observed in this and other studies [17, 18, 20], prehabilitation offers a net benefit. In summary, our results reinforce the growing body of evidence that properly supervised prehabilitation is a valuable adjunct to standard perioperative care, improving patient fitness and surgical recovery while posing little risk.

Clinically, our findings tentatively support multimodal prehabilitation as a beneficial adjunct for patients undergoing ESD by optimizing pre-procedure fitness [21]. However, the heterogeneous nature of ESD patients (varying age, comorbidities, and lesion complexity) means the observed benefits may not generalize to all individuals. Any routine implementation must weigh real-world practicalities: added program costs, resource demands, feasibility across healthcare settings, and potential patient burden (e.g., adherence to exercise and nutrition regimens). While some evidence suggests prehabilitation can be cost-effective, current uptake in practice is limited, reflecting the still-evolving evidence base [22, 23]. Thus, further research is needed to confirm these preliminary benefits and assess cost-effectiveness and feasibility across diverse populations before recommending prehabilitation as a standard pre-ESD protocol.

Potential sources of bias and confounding should also be considered when interpreting our findings. Although randomization aimed to balance baseline characteristics, residual confounding by unmeasured variables, such as participants’ intrinsic motivation, socioeconomic background, or prior exercise habits, may have influenced adherence and recovery outcomes. As both exercise and dietary compliance were self-reported, recall and social desirability biases could have led to overestimation of adherence and intervention effects. In addition, outcome assessors were blinded, but participants and care providers were aware of group allocation, which may have introduced performance bias, especially in patient-reported outcomes. Finally, differences in perioperative management practices or external lifestyle factors during follow-up could have contributed to variability in functional recovery beyond the prehabilitation effect itself.

This study has several strengths, including its randomized controlled design, adherence to CONSORT guidelines, use of blinded outcome assessment, and comprehensive follow-up to 24 weeks. The intervention was standardized yet individualized, and adherence was closely monitored with both objective and self-reported measures. However, certain limitations must be acknowledged. The single-center setting may limit generalizability, and the relatively small sample size may have reduced power to detect modest differences in some secondary outcomes. Additionally, the study population was limited to elective ESD patients with a minimum 4-week lead time, which may exclude higher-risk or urgent cases. Cost data were collected at the patient level but may not reflect broader healthcare system costs. Another limitation is that exercise adherence in the prehabilitation group was recorded through patient self-reported logbooks and questionnaires. This reliance on self-report may have led to overestimation of actual exercise participation due to recall and social desirability bias. Finally, it is important to note that our study population consisted of relatively young, physically fit, and non-frail individuals. As such, our findings may not be fully generalizable to older or more frail populations, who are often the primary target of prehabilitation programs.

Conclusion

Prehabilitation before ESD significantly improved muscle strength and reduced anxiety at 4 weeks postoperatively, and enhanced functional exercise capacity and quality-of-life by 3 months, compared to standard care. Complication rates, hospital stay, and readmissions did not differ significantly between groups.

Authors’ contributions

Conceptualization: Y.S. and B.J.; Methodology: Y.S. and S.L.; Formal Analysis: Y.Y., R.W and M.X.; Investigation: B.J. and L.H.; Writing – Original Draft Preparation: B.J. and S.L.; Writing – Review & Editing: Y.S.

Funding

No funding or financial support was received for the conduct of this study or the preparation of this manuscript.

Data availability

The datasets used and/or analysed during the current study are available from the corresponding author on reasonable request.

Declarations

Ethics approval and consent to participate

The study was approved by the Ethics Committee of Shanghai Sixth People’s Hospital (Approval No. 2022-KY-046 (K)). All procedures were conducted in accordance with the Declaration of Helsinki and relevant institutional guidelines. Written informed consent was obtained from all individual participants before enrollment in the study.

Consent for publication

Not applicable.

Competing interests

The authors declare no competing interests.