Effect of prehabilitation in patients undergoing gastric or oesophageal cancer resections: A systematic review and meta-analysis

Abstract

BACKGROUND

Gastric and oesophageal cancers are among the leading causes of cancer-related mortality globally. Treatment in the form of surgical resection and neoadjuvant therapy is considered the gold standard; however, these procedures are associated with significant postoperative morbidity and prolonged recovery times. Prehabilitation aims to combat this by physically and psychologically optimising patients preoperatively to handle the stress of surgery and improve postoperative outcomes.

AIM

To evaluate the effectiveness of prehabilitation in patients undergoing gastric or oesophageal cancer resections.

METHODS

Standard medical databases such as MEDLINE, EMBASE, PubMed, and Cochrane Library were searched to find randomised, controlled trials comparing prehabilitation vs no-prehabilitation in patients undergoing gastric or oesophageal cancer resections. All data were analysed using Review Manager Software 5.4, and the meta-analysis was performed with a random-effect model analysis.

RESULTS

A total of 16 studies were included (n = 1479), recruiting adult patients undergoing gastric or oesophageal cancer resections. In the random effect model analysis, prehabilitation was associated with fewer postoperative complications compared to no prehabilitation (Odds ratio: 0.43, 95%CI: 0.31-0.58, Z = 25.28, P < 0.00001). Additionally, prehabilitation was associated with a lower postoperative readmission rate (risk ratio: 0.58, 95%CI: 0.35-0.96, Z = 2.10, P = 0.04) and a shorter length of stay [standardised mean difference (SMD): -0.19, 95%CI: -0.37 to -0.02, Z = 2.21, P = 0.03] compared to no-prehabilitation. Prehabilitation was also found to improve preoperative 6-minute walking distance (6-MWD) compared to no-prehabilitation (SMD: 0.72, 95%CI: 0.21-1.23, Z = 2.75, P = 0.006). No statistical difference was found in postoperative 6-MWD, mortality, and severe complications.

CONCLUSION

Prehabilitation was found to be effective in reducing postoperative morbidity and improving preoperative physical health in patients undergoing gastric or oesophageal cancer resections.

Key Words: Prehabilitation; Oesophageal cancer; Gastric cancer; Cancer resection; Postoperative complications

Core Tip: Gastric and oesophageal cancer resections are major surgeries often linked with significant complications and lengthy recovery. This meta-analysis of 16 randomised controlled trials demonstrates that prehabilitation–which involves structured physical, nutritional, and psychological preparation before surgery–can significantly reduce postoperative complications, hospital readmissions, and length of stay. It also improves patients’ preoperative physical fitness, as reflected in better performance on the six-minute walk test. These findings support the incorporation of prehabilitation into routine care pathways for upper gastrointestinal cancer surgery.

INTRODUCTION

Cancer of the stomach and oesophagus is a major contributor to the overall global burden of cancer, featuring high incidence, mortality, and poor prognosis[1]. In 2020 alone, there were 600000 new cases of oesophageal cancer and 1.1 million cases of gastric cancer globally, which were accompanied by 540000 and 770000 deaths, respectively[1,2]. Neoadjuvant therapy followed by surgical resection in the form of gastrectomy and oesophagectomy remains the primary and most commonly performed treatment of choice for these malignancies, particularly in cases where the tumor becomes increasingly infiltrative or advanced[3-5]. Given the nature of these cancers and the complexity of these procedures, it is not uncommon for a multitude of complications, such as anastomotic leaks, infections, nutritional issues, and more, to arise following surgical intervention. Previous studies have revealed a correlation between neoadjuvant therapy and a decrease in cardiopulmonary fitness, which has been linked to poorer postoperative outcomes[6]. The high morbidity associated with such surgical interventions underscores the need for improved patient care strategies to enhance patient outcomes and improve disease prognosis.

Prehabilitation is a multifaceted preoperative approach aimed at reducing physical deterioration, complications, and improving patient quality of life postoperatively[7]. This consists of exercise, nutritional, and psychological interventions all aimed at optimising patients to handle the stress of surgery[7]. Programs such as enhanced recovery after surgery (ERAS) have been proven to have positive effects on patient outcomes, leading to shorter hospital stays without increased morbidity or mortality, however, these programs primarily focus on the postoperative period and overlook potential preoperative interventions that could further improve outcomes[8]. Furthermore, it has been shown that patients are more accepting of lifestyle interventions preoperatively, which can be attributed to the fatigue, low mood, and lack of eating associated with the postoperative period[9]. The effectiveness of patient optimisation prehabilitation programs has been demonstrated across multiple studies, however, this meta-analysis is the first of its kind comparing exclusively randomised controlled trials (RCTs) looking at the impact of prehabilitation on postoperative outcomes of patients awaiting resections for oesophageal or gastric cancer.

MATERIALS AND METHODS

This meta-analysis was conducted in accordance with the preferred reporting items for a systematic review and meta-analysis, through the Preferred Reporting Items for Systematic Reviews and Meta-analyses guidelines[10].

Data sources and search strategy

For this systematic review and meta-analysis, standard medical databases MEDLINE, EMBASE, PubMed, and The Cochrane Library were searched for relevant RCTs. Articles published until January 2025 were searched for, using the headings “prehabilitation for oesophageal cancer” and “prehabilitation for gastric cancer”. Boolean operators (AND, OR, NOT) were utilised to refine and optimise search results. The titles from the search results were independently reviewed by two authors and were chosen either be included or excluded from the study. The referenced list of the included studies was searched to find any additional articles. The accuracy of the extracted data was additionally confirmed by a third author.

Study selection

Only RCTs involving human participants directly comparing the effects of prehabilitation to no prehabilitation in gastric or oesophageal cancer patients awaiting surgery were chosen. Studies showing rehabilitation were not included. Quasi-experimental studies, cohort studies, and case-control studies were excluded. There was no restriction on the country of origin for the studies, language, or hospital of origin. The main outcomes we were searching for included the preoperative and postoperative 6-minute walking distance (6-MWD), length of stay (LOS), total postoperative complications, severe postoperative complications (Clavien Dindo ≥ 3), postoperative mortality, and readmission rates. The process and results of the literature retrieved are shown in Figure 1.

Figure 1  Preferred Reporting Items for Systematic Reviews and Meta-analyses flow chart explaining identification of studies via databases.

Data extraction

Two reviewers independently extracted data from the included studies using a predefined meta-analysis data form, with a third reviewer verifying the accuracy. There were no disagreements among the reviewers regarding study selection or data extraction.

Statistical analysis

The software Review Manager 5.4[11], provided by the Cochrane Collaboration, was used for statistical analysis. The odds ratio (OR) with a 95%CI was calculated for binary data, and the standardised mean difference (SMD) with a 95%CI was used for continuous data variables. The random-effects model[12,13] was used to measure the combined outcomes for dichotomous variables. Heterogeneity was explored through the χ² test, with significance set at P < 0.05, and was quantified[14] using I², with a maximum value of 30% identifying low heterogeneity. If the standard deviation was not available, it was calculated through the guidelines set out by the Cochrane Collaboration[15]. The Mantel-Haenszel method was used to calculate OR under the random-effect model[16] analysis. Only RCTs that were clinically comparable and had tested the same variables, such as 6-MWD or LOS, were pooled together.

In a sensitivity analysis, 0.5 was added to each cell frequency for trials in which no event occurred in either the treatment or control group, according to the method recommended by Deeks et al[16]. This process assumed that both groups had the same variance, which may not have been confirmed, and variance was either estimated from the range or the P value. The estimate of the difference between both techniques was pooled, depending on the effect weights in results determined by each trial’s estimated variance. Forest plots were used for the graphical display. The square around the estimate represented the accuracy of the estimation (sample size), and the horizontal line represented the 95%CI. The methodological quality of the included trials was initially assessed by the Cochrane trial quality assessment tool[17].

RESULTS

The primary database search led to five hundred and twenty-three studies, of which five hundred and six were excluded after initial screening. The final review included sixteen studies (Figure 1).

Characteristics and demographics of included studies

Sixteen RCTs on 1479 patients were included in our updated systematic review comparing patient outcomes between prehabilitation and no prehabilitation in patients undergoing surgical intervention in the form of resection for gastric or oesophageal cancer. Oesophageal and gastric cancer resections were done through open, laparoscopic, robotic, and thorascopic surgery. Principles advised by the Cochrane Collaboration were used in this analysis. The trials included were studied in Japan, the United Kingdom, Spain, Lithuania, China, Ireland, the Netherlands, Canada, and India. Primary demographic characteristics of the studies included are specified in Table 1, and the protocol used in each study is given in Table 2[7,17-31].

Table 1 Characteristics of included studies.

Ref.	Country	Operation	Number of patients	Age	Female:male	Duration of follow up	Trial running time
Aiko et al[18], 2012	Japan	Open oesophagectomy					2007- 2012
Prehabilitation			14	65 ± 2	1:13	1 week
No prehabilitation			12	61 ± 3	2:10
Allen et al[17], 2022	United Kingdom	Open oesophagectomy and gastrectomy					2016-2018
Prehabilitation			26	65 ± 6	4:22	3 years
No prehabilitation			28	62 ± 9	4:24
Barberan-Garciaet al[19], 2018	Spain	Oesophagectomy and gastrectomy (type not reported)				Until discharge, mean LOS: 8 ± 8, 13 ± 20	2013–2016
Prehabilitation			73	71 ± 11	19:43
No prehabilitation			71	71 ± 10	12:51
Bausys et al[20], 2023	Lithuania	Laparoscopic and open gastrectomy					2020-2022
Prehabilitation			61	61 ± 11	26:35	90 days
No prehabilitation			61	64 ± 10	19:42
Chen et al[21], 2024	China	Laparoscopic and open gastrectomy					2019-2023
Prehabilitation			57	73 (70.5-77)	24:33	30 days
No prehabilitation			58	74 (68-78)	30:28
Guinan et al[22], 2018	Ireland, Netherlands	Open oesophagectomy					2014-2016
Prehabilitation			28	63.07 ± 8.8	8:20	9 days
No prehabilitation			32	65.06 ± 67.78	10:22
He et al[23],2022	China	Gastrectomy (type not reported)					2020–2021
Prehabilitation			31	63.2 ± 12	7:24	1 month
No prehabilitation			35	60.5 ± 9.4	12:23
Liu et al[24],2020	China	Open and thorascopic oesophagectomy					January–June 2018
Prehabilitation			26	62.04 ± 5.12	5:21	1 month
No prehabilitation			24	64.58 ± 5.87	10:14
Loughney et al[25], 2024	Ireland	Laparoscopic, thorascopic, open, and robotic oesophagectomy and gastrectomy					2019-2020
Prehabilitation			36	62.8 ± 9.2	9:27	6 weeks
No prehabilitation			35	61.5 ± 8.8	10:25
Minnella et al[7], 2018	Canada	Open and minimally invasive oesophagectomy and gastrectomy					2013-2017
Prehabilitation			26	67.3 ± 7.4	8:18	8 weeks
No prehabilitation			25	68 ± 11.6	5:20
Swaminathan et al[26], 2020	India	Open gastrectomy					2017-2019
Prehabilitation			29	56.03 ± 14.95	11:18	2 weeks
No prehabilitation			29	56.82 ± 11.27	9:20
Valkenet et al[27], 2018	Netherlands	Open, laparoscopic, and robotic oesophagectomy					2013-2016
Prehabilitation			120	63. ± 7.5	31:89	15 days
No prehabilitation			121	62.7 ± 8.9	24:97
Wang et al[28], 2015	China	Gastrectomy (type not reported)					2010-2012
Prehabilitation			100	38-72	70:130	10 days
No prehabilitation			100
Yamana et al[29], 2015	Japan	Open and thorascopic oesophagectomy				Until discharge	2011-2014
Prehabilitation			30	68.33 ± 7.64	6:24
No prehabilitation			30	65.9 ± 9.5	7:23
Zhao et al[30], 2018	China	Oesophagogastric cancer resections				Until discharge, mean LOS: 7 ± 1, 8 ± 2	2012-2016
Prehabilitation			33	62 (26-74)	9:57
No prehabilitation			33
Zheng et al[31], 2024	China	Open and laparoscopic gastrectomy				Until discharge, mean LOS: 8.69 ± 1.92, 9.57 ± 2.16	Not reported
Prehabilitation			61	66.07 ± 8.97	51:9
No prehabilitation			60	63.72 ± 8.84	45:16

LOS: Length of stay.

Table 2 Treatment protocol adopted in included studies.

Ref.	Intervention	Control
Aiko et al[18], 2012	Immune enhanced diets 5 days prior to surgery. Enteral feeding and parenteral mixture (UNICALIQ l) following surgery	Standard diets before surgery. Enteral feeding and parenteral mixture (UNICALIQ l) following surgery
Allen et al[17]	Supervised and home-based exercise program. Six sessions of medical coaching	Standardoesophagogastric care pathways. Encouraged to remain active, given dietary recommendations and psychological support
Barberan-Garciaet al[19], 2018	Personalised daily activity and exercise programs. Motivational interviewing	Standard care. Encouraged to make healthy lifestyle choices
Bausys et al[20], 2023	Personalised home-based prehabilitation program consisting of exercise, nutritional, and psychological support	Did not receive advice on prehabilitation related interventions, some were recommended supplementation prior to surgery
Chen et al[21], 2024	3-week individualised prehabilitation program consisting of exercise, respiratory training, nutritional, and psychological support	Did not receive any prehabilitation advice. ERAS recommendation
Guinan et al[22], 2018	Inspiratory muscle training program	Standard pathway
He et al[23],2022	Oral nutritional supplement for seven days	Standard dietary advice
Liu et al[24],2020	Oral nutrition supplement for all patients	ERAS, oral nutrition supplement if needed
Loughney et al[25], 2024	Standardised nutritional pathway. Exercise prehabilitation	Standardised nutritional pathway
Minnella et al[7], 2018	Exercise and nutrition programs	ERAS. Standard care
Swaminathan et al[26], 2020	Preoperative volume-oriented spirometry. Oral maltodextrin drink the night before and day of surgery. Oral midazolam on day of surgery. ERAS	Standard care
Valkenet et al[27], 2018	Inspiratory muscle training	Standard care
Wang et al[28], 2015	Daily nutritional protein-based supplement 1 week prior to surgery	Standard preoperative care
Yamana et al[29], 2015	Preoperative respiratory exercise training program	No interventions
Zhao et al[30], 2018	Oral nutrition supplement 7 days before surgery	Routine preoperative diet
Zheng et al[31], 2024	Preoperative pulmonary related training program	No interventions

ERAS: Enhanced recovery after surgery.

Methodological quality of included studies

The methodological quality of the included RCTs is shown in Table 3[7,17-31]. Randomisation was achieved through either computer generation, envelopes, block randomisation, random number tables, or was not recorded. Concealment was reported in nine studies.

Table 3 Quality assessment of included studies.

Ref.	Randomisation	Power calculation	Blinding	Concealment	ITT	Inclusion criteria	Exclusion criteria
Aiko et al[18], 2012	Envelope method	Not reported	Not reported	Not reported	Not reported	Reported	Reported
Allen et al[17]	Computer generated variable block size	Reported	Not reported	Reported	Reported	Reported	Reported
Barberan-Garciaet al[19], 2018	Computer generated random numbers	Not reported	Double	Not reported	Reported	Reported	Reported
Bausys et al[20], 2023	Computer generated	Reported	Not reported	Reported	Reported	Reported	Reported
Chen et al[21], 2024	Computer generated	Reported	Single	Reported	Reported	Reported	Reported
Guinan et al[22], 2018	Web based	Not reported	Single	Not reported	Not reported	Reported	Reported
He et al[23],2022	Random number tables	Reported	Single	Not reported	Not reported	Reported	Not reported
Liu et al[24],2020	Computer random number generation	Reported	Single	Not reported	No ITT	Reported	Reported
Loughney et al[25], 2024	Computer generated	Reported	Noblinding	Not reported	Reported	Reported	Not reported
Minnella et al[7], 2018	Computer generated	Reported	Single	Reported	Not reported	Reported	Reported
Swaminathan et al[26], 2020	Block randomisation	Reported	Not reported	Reported	Modified ITT	Reported	Reported
Valkenet et al[27], 2018	Web based	Reported	Single	Reported	Modified ITT	Reported	Reported
Wang et al[28], 2015	Computer generated	Not reported	Not reported	Reported	Not reported	Not reported	Not reported
Yamana et al[29], 2015	Not reported	Reported	Not reported	Reported	Not reported	Not reported	Not reported
Zhao et al[30], 2018	Table of random numbers	Not reported	Not reported	Not reported	Not reported	Reported	Reported
Zheng et al[31], 2024	Computer based random number table	Reported	Double	Reported	Not reported	Reported	Reported

ITT: Intention to treat.

Outcome of primary and secondary variables

Preoperative 6-MWD: A total of 6 studies on 529 patients undergoing open, laparoscopic, robotic, or thorascopic upper gastrointestinal cancer surgery reported preoperative 6-MWD. The pooled analysis found that patients who underwent prehabilitation had significantly greater preoperative 6-MWD (SMD: 0.72, 95%CI: 0.21-1.23, Z = 2.75, P = 0.006). Significant heterogeneity was observed among included trials [Tau² = 0.35, χ² = 39.43, df = 5 (P < 0.00001), I² = 87%; Figure 2A][7,19-22,25].

Figure 2 Forest plot showing preoperative and postoperative 6-minute walking distance represented in standardised mean difference with 95%CI. A: Preoperative 6-minute walking distance (6-MWD); B: Postoperative 6-MWD.

Postoperative 6-MWD: A total of 4 studies on 297 patients undergoing open, laparoscopic, robotic, or thorascopic upper gastrointestinal cancer surgery reported 6-MWD at postoperative follow-ups. The pooled analysis found no statistically significant difference between both groups (SMD: 0.27, 95%CI: -0.64 to 1.19, Z = 0.59, P = 0.56). Significant heterogeneity was observed among included trials [Tau² = 0.80, χ² = 42.35, df = 3 (P < 0.00001), I² = 93%; Figure 2B][7,21,22,25].

Length of hospital stay

A total of 12 studies on 1134 patients undergoing open, laparoscopic, robotic, or thorascopic upper gastrointestinal cancer surgery reported length of hospital stay. The pooled analysis found that patients who underwent prehabilitation had significantly shorter hospital stays (SMD: -0.19, 95%CI: -0.37 to -0.02, Z = 2.21, P = 0.03). Significant heterogeneity was observed among trials [Tau² = 0.04, χ² = 21.98, df = 11 (P = 0.02), I² = 50%; Figure 3A][7,17,19-22,24-27,30,31].

Figure 3 Forest plot. A: Forest plot showing length of hospital stay represented in standardised mean difference with 95%CI; B: Forest plot showing total postoperative complications represented in odds ratio with 95%CI; C: Forest plot showing severe postoperative complications represented in risk ratio with 95%CI; D: Forest plot showing hospital readmission represented in odds ratio with 95%CI; E: Forest plot showing mortality represented in odds ratio with 95%CI.

Total postoperative complications

A total of 11 studies on 1033 patients undergoing open, laparoscopic, robotic, or thorascopic upper gastrointestinal cancer surgery reported postoperative complications. The pooled analysis found that patients who underwent prehabilitation had significantly fewer postoperative complications (OR: 0.43, 95%CI: 0.31-0.58, Z = 5.28, P < 0.00001). There was no statistically significant heterogeneity among included trials [Tau² = 0.04, χ² = 11.55, df = 10 (P = 0.32), I² = 13%; Figure 3B][7,17,19-21,23-26,30,31].

Severe postoperative complications (CD ≥ 3)

A total of 8 studies on 681 patients undergoing open, laparoscopic, robotic, or thorascopic upper gastrointestinal cancer surgery reported severe postoperative complications. The pooled analysis found no statistically significant difference between both groups (OR: 0.84, 95%CI: 0.59-1.20, Z = 0.93, P = 0.35). There was no statistically significant heterogeneity observed among included trials [Tau² = 0.00, χ² = 5.02, df = 7 (P = 0.66), I² = 0%; Figure 3C][7,17,18,20,24,25,27,29].

Postoperative readmissions

A total of 7 studies on 720 patients undergoing open, laparoscopic, robotic, or thorascopic upper gastrointestinal cancer surgery reported postoperative readmissions. The pooled analysis found that patients who underwent prehabilitation had significantly lower readmission rates (OR: 0.54, 95%CI: 0.31-0.95, Z = 2.13, P = 0.03). There was no statistically significant heterogeneity observed among included trials [Tau² = 0.00, χ² = 1.92, df = 6 (P = 0.93), I² = 0%; Figure 3D][7,17,20,21,23,25,27].

Mortality

A total of 7 studies on 681 patients undergoing open, laparoscopic, robotic, or thorascopic upper gastrointestinal cancer surgery reported patient mortality. The pooled analysis found no statistically significant difference between both groups (OR: 1.22; 95%CI: 0.54-2.73, Z = 0.48, P = 0.63). There was no statistically significant heterogeneity observed among included trials [Tau² = 0.00, χ² = 1.76, df = 3 (P = 0.62), I² = 0%; Figure 3E][7,17,20,21,25,27,29].

DISCUSSION

Sixteen RCTs with a combined total of 1479 patients (742 who underwent prehabilitation and 737 who did not) were included in our systematic review and meta-analysis comparing outcomes of patients who received prehabilitation vs those who did not prior to surgical resection of gastric or oesophageal cancer. Our findings demonstrate the viability of prehabilitation interventions in improving patient outcomes by reducing morbidity and optimising cardiopulmonary reserve preoperatively. While no significant differences were observed in mortality and severe complications, patients who underwent prehabilitation spent significantly less time in the hospital and were less likely to require readmission. Such results have considerable implications for further improving postoperative outcomes, as prolonged hospital stays have been linked to an increased risk of hospital-acquired infections and other complications[32]. Moreover, prehabilitation has been linked to reduced healthcare costs, potentially as a manifestation of shorter hospital stays and fewer readmissions[33].

The improvement in cardiopulmonary reserve observed in prehabilitation groups, indicated by increased 6-MWDs, also carries clinical significance. A systematic review published in 2021 established a significant correlation between higher peak oxygen uptake and favourable outcomes following surgery[34]. The absence of statistically significant differences between the two groups in postoperative 6-MWD indicates that preoperative interventions alone may not be enough to sustain improvements in physical health. To maximise the effectiveness of prehabilitation, these interventions should be complemented with personalised postoperative rehabilitation programs to prevent any decline and further optimise recovery.

Two previous meta-analyses on the subject were identified: (1) An et al[35] in 2024, which evaluated the effect of prehabilitation in patients undergoing oesophagectomy; and (2) Shen et al[36] in 2024, which evaluated the effect of nutrition-based prehabilitation in patients with oesophagogastric cancer. These studies reported similar outcomes to the ones we discussed; however, the reliability of their results is limited by the low quality of evidence used. Our study was the first to exclusively include all available RCTs on the subject, providing more robust, high-quality evidence. Our analysis revealed greater improvements in patient outcomes, as previous studies found no significant differences in LOS, readmission rates, and 6-MWD scores.

A review of the existing literature shows that this meta-analysis is the most up-to-date and comprehensive review on the effect of prehabilitation on patients undergoing resection for gastric or oesophageal cancer. However, this study does have several limitations. Firstly, the employed interventions varied significantly between included studies. Prehabilitation was either multimodal, nutritional, or exercise-based, introducing significant heterogeneity between the protocols that patients undertook. Additionally, differences in postoperative management were also noted, as certain studies adopted rehabilitation programs (e.g., ERAS) while others did not, possibly skewing results and compromising their generalisability. Secondly, most included trials had short follow-up timings, impeding our ability to draw long-term conclusions. Thirdly, as shown in Figure 2B, several methodological aspects such as blinding, inclusion/exclusion criteria, randomisation techniques, and concealment were not adequately reported, introducing a risk of bias. Fourthly, one of the included studies included patients undergoing resections for other gastrointestinal cancers in addition to gastric and oesophageal cancers, possibly introducing heterogeneity. Finally, statistically significant heterogeneity existed between included trials reporting preoperative 6-MWD, postoperative 6-MWD, and readmission rates, suggesting that findings should be interpreted with caution. Future trials should aim to address existing gaps in the literature and aim to provide insight into the long-term effects of prehabilitation.

CONCLUSION

Prehabilitation has been shown to improve patient outcomes by reducing postoperative morbidity, LOS, and improving preoperative cardiopulmonary reserve in patients undergoing gastric or esophageal cancer resections. Despite this evidence, further research is required to draw generalised long-term conclusions.