Effectiveness of prehabilitation during neoadjuvant therapy for patients with esophageal or gastroesophageal junction cancer: a systematic review

A literature search was performed on March 10, 2023, to identify relevant studies assessing prehabilitation regimens including exercise in patients with esophageal or gastroesophageal junction cancer during NAT, in the MEDLINE (PubMed), Web of Science, CENTRAL, CINAHL, and Physiotherapy Evidence (PEDro) databases. Unpublished literature was searched using the ClinicalTrials.gov and OpenGrey databases. The search included the following keywords: “esophagus,” “gastric,” “neoadjuvant,” “physical therapy,” “prehabilitation,” and “respiratory training.” The references of the included articles were screened, and a manual search was performed to identify missing articles. The complete electronic search strategy is available in the electronic supplementary material (Online Resource 1). Two reviewers (TI and ST) independently assessed titles and abstracts to include relevant references. Subsequently, the full-text articles selected by title and abstract screening were assessed for eligibility. Data extraction from the included studies was performed independently by two reviewers (TI and ST). In cases of disagreement regarding the inclusion or data extraction, a third author (TK) was consulted. The authors of the included studies were contacted to obtain unpublished data. The Preferred Reporting Items for Systematic Reviews and Meta-Analyses (PRISMA) guidelines [18] were followed (Online Resource 2). The searched references were uploaded to Rayyan [19] (Qatar Computing Research Institute, Ar Rayyan, Qatar), and duplicate references were removed. The protocol for this systematic review was registered with PROSPERO (https://​www.​crd.​york.​ac.​uk/​prospero/​display_​record.​php?​ID=​CRD42023395998).

We included randomized controlled trials (RCTs) and quasi-RCTs evaluating the feasibility or efficacy of prehabilitation during NAT. Retrospective cohort and single-arm studies were included to investigate the feasibility of this approach. There were no restrictions on the country or language. All published and unpublished papers, including conference abstracts, were eligible. No animal studies were included. No exclusions were made on the basis of the observation period. Systematic reviews, meta-analyses, narrative reviews, cluster RCTs, cross-sectional studies, crossover trials, case series, case reports, and letters were excluded. Studies that met the following criteria were included: those involving patients with esophageal or gastroesophageal junction cancer receiving NAT; those incorporating prehabilitation including aerobic exercise, resistance training, and respiratory training during NAT; and those assessing preoperative physical fitness. Studies that met the following criteria were excluded: participants with unstable comorbidities that precluded exercise therapy or exercise function testing, such as severe cardiac or respiratory dysfunction or orthopedic disease; cognitive disorders (Mini-Mental State Examination score less than 24); and nonradical surgery.

The primary outcomes were exercise capacity (cardiopulmonary exercise test and 6-min walk test) and postoperative complications (postoperative pulmonary complications and surgical site infections, etc.). The secondary outcomes included an assessment of physical fitness other than exercise capacity, tolerance to NAT (completion rate of scheduled treatment), the adherence of prehabilitation (attendance/session*100), and adverse events of prehabilitation.

To understand the characteristics of the included studies, we collected information on participants’ characteristics (age, sex, body mass index [BMI], diagnosis, and NAT regimen) and prehabilitation protocols (type, intensity, time, and frequency).

We assessed the risk of bias in the RCTs using the revised Cochrane risk-of-bias tool for randomized trials (RoB 2) [20]. The RoB2 tool categorizes the risk of bias into “low,” “some concerns,” or “high” groups. The risk of bias in non-randomized comparative studies was assessed using the Newcastle–Ottawa Scale (NOS) [21]. For NOS scores ranging from 0 to 9, we considered scores of 0–3, 4–6, and 7–9 as low, moderate, and high-quality studies, respectively. Single-arm studies were assessed using the National Heart, Lung, and Blood Institute Quality Assessment Tool for before-after (pre- and post-intervention) studies [22]. Based on the 12-item rating, “Good,” “Fair,” and “Poor” scores were determined. All risk-of-bias assessments were conducted in duplicate by two well-trained reviewers (TI and ST). Disagreements between the reviewers were resolved through discussion or by inviting an additional reviewer (TK). The quality of evidence was determined by one researcher (TI) and confirmed and finalized by another researcher (ST). The GRADE (Grading of Recommendations, Assessment, Development, and Evaluation) method was used to check the quality of the evidence [23]. Summary of the findings, prepared according to the Cochrane Handbook is presented in Table 1. Table 1 shows the relative effects per outcome, number of participants, quality of evidence, and additional comments. The standard mean difference (SMD) and risk ratio (RR) were used as relative effects.

A meta-analysis of the studies comparing the two groups was performed if at least two published RCTs reported the same outcome within the same timeframe. Published non-RCTs and unpublished studies were included in the meta-analysis only when one or fewer published RCTs examined for each outcome. We pooled the mean differences and 95% confidence intervals (CIs) for the continuous variables (reporting mean and standard deviation or standard error of the mean) for each trial. For cases in which the outcome units were different, we attempted to integrate the data by calculating SMD. RRs with 95% CIs were selected as summary statistics for dichotomized outcomes. The meta-analysis was performed using the Review Manager software (RevMan 5.4, Cochrane) [24]. Missing data were referred to the principal investigator, and in cases where they were unavailable, the substitution method was not used. When a meta-analysis was not possible, results were described narratively. In cases with I² > 50%, we assessed substantial heterogeneity, and subgroup analysis was performed by age when we have gathered sufficient data. The Cochran chi-squared test (Q test) was used for the I² statistic, and a p-value of < 0.10 was considered statistically significant. Considering that the intervention effects varied owing to setting differences across studies, we applied a random-effects model. If the mean differences and 95% CIs were not reported, the study was excluded from the meta-analysis. To confirm the robustness of the results, a meta-analysis was performed using non-RCTs for sensitivity analysis. We assessed the possibility of reporting bias using funnel plots.

A systematic search of the databases yielded 7342 results. Duplicates were removed, leaving 3,902 records. In the primary screening, 3440 articles were reviewed by title and abstract, and in the secondary screening, 27 articles were evaluated for eligibility using the full text. Ultimately, 12 articles [13‐15, 25‐33] were eligible for inclusion. A systematic search of these databases is shown in Fig. 1.

The study characteristics are summarized in Table 2. Three studies were included in each category: RCTs [13, 14, 25], quasi-RCTs [15, 26, 27], retrospective cohort studies [28‐30], and prospective single-arm intervention studies [31‐33]. The literature consisted of nine original articles [13‐15, 26, 28‐30, 32, 33] and three unpublished studies [25, 27, 31]. The total number of participants in the 12 included studies was 664. The mean age of the participants ranged from 59.2 to 67.0 years, 74.7 to 100% were men, and the mean BMI ranged from 21.1 to 28.9 kg/m2. Three studies [13, 26, 33] used neoadjuvant chemoradiotherapy, five [15, 27, 29, 30, 32] used neoadjuvant chemotherapy, and three [14, 28, 31] used neoadjuvant chemoradiotherapy or chemotherapy.

The characteristics of prehabilitation during NAT are presented in Table 3. Prehabilitation programs mainly consisted of aerobic exercise and resistance training, and the duration ranged 4–16 weeks, with one to seven sessions per week. Aerobic exercise duration ranged 15–60 min, performed at low to moderate intensity [13, 14, 29, 32, 33] moderate/vigorous-intensity activity [28, 30], or high-intensity interval training [15, 26, 31]. Aerobic exercise included programs with expert instruction [13‐15, 26, 29, 31], self-training [28‐33], and remote communication using phone calls or applications [28, 30‐33]. Resistance training was performed at moderate to high intensity with expert instruction [15, 26, 29] or self-training [14, 28‐32]. Additionally, some studies incorporated nutritional interventions [13, 14, 28‐31, 33] and medical counseling [14, 29‐31] within their prehabilitation programs.

The risk of bias was assessed only for original articles. We presented the risk of bias assessment for exercise capacity as the primary outcome of the published two RCTs (Fig. 2). The bias risks for the two RCTs [13, 14] were categorized as “some concerns” and “high,” respectively. The bias in the outcome measurement was judged to be “high” for one RCT [13]. Three comparative studies, excluding RCTs, were judged to be of “high” [15, 28] or “moderate” quality [26] using the NOS (Online Resource 3). In comparison, one non-RCT [26] had a risk of bias. Two non-RCTs [15, 28] had a risk of bias in the outcome assessment. Four single-arm studies were judged to be of good [29] or fair [30, 32, 33] quality (Online Resource 4). Three studies [30, 32, 33] did not report the blinding of outcome assessments, suggesting the possibility of lower quality.

Forest plots using two RCTs [13, 14] are shown for the effects of prehabilitation during NAT on exercise capacity (Fig. 3), grip strength (Fig. 4), skeletal muscle mass (Fig. 5), and tolerance to NAT (Fig. 6). Only one RCT [14] assessed the postoperative complications; hence, meta-analysis was performed using non-RCTs [15, 26, 28] (Fig. 7). In all four studies [14, 15, 26, 28], complications were assessed with Clavien–Dindo classification of 3 or higher. A summary of the effects of prehabilitation during NAT is shown in Table 4. Prehabilitation affected completion rate of NAT (RR [95% CI], 1.51 [1.10, 2.08]) and the change in grip strength during NAT (SMD [95% CI], 1.22 [0.53, 1.90]). However, prehabilitation have no significant effect on exercise capacity (SMD [95% CI], 0.93 [-0.30, 2.17]) and skeletal muscle mass (SMD [95% CI], 0.27 [ – 0.11, 0.65]). Prehabilitation during NAT reduced the risk of postoperative complications (RR [95% CI], 0.61 [0.39, 0.94]). In the sensitivity analysis, including RCTs, non-RCTs, and unpublished literature [13‐15, 25], prehabilitation during NAT had a significant effect on the exercise capacity (SMD [95% CI], 1.63 [0.11, 3.14]; Online Resource 5) and skeletal muscle mass (SMD [95% CI], 0.36 [0.02, 0.70]; Online Resource 6). Due to the small number of RCTs, a subgroup analysis could not be performed.

Feasibility of the prehabilitation during NAT is presented in Table 4. Five studies [13, 14, 26, 29, 31] reported adherence (attendance/sessions*100) to supervised exercise during NAT, ranging 59.6–76.1%. Adherence to low- to moderate- intensity supervised exercise included in multimodal prehabilitation during NAT was 76.0% or 76.1% [14, 29]. Six studies [14, 28, 30‐33] reported 55.0–69.4% adherence to home-based exercises. Adherence to home-based exercises using the application was relatively high (69.4%) [33]. Two studies [13, 31] reported 80.8–100% adherence to dietary interventions. Two studies [14, 32] reported 82.0–100% adherence to psychological care. Nine studies [13, 14, 26, 28‐33] reported 0–14.3% dropout rates of participants with no disease association in the experimental group. Dropout rates from low- to moderate-intensity supervised exercise included in multimodal prehabilitation during NAT were relatively low (0–7.7%). The characteristics of dropouts from prehabilitation included health-related and non-health concerns (travel, lack of time, and lack of energy) [14, 26, 32, 33]. Six studies [13, 14, 26, 29, 31, 32] reported no serious adverse events (0%).

This systematic review, encompassing data from 12 articles involving 664 participants, aimed to assess the effectiveness, acceptability, and safety of prehabilitation during NAT for patients with esophageal or gastroesophageal junction cancer. The meta-analysis, including two RCTs, demonstrated that prehabilitation was more effective than usual care regarding tolerance to NAT and grip strength. Moreover, an extensive meta-analysis including RCTs, non-RCTs, and unpublished literature showed that prehabilitation during NAT might contribute to a decrease in the risk of postoperative complications and the maintenance of exercise capacity and skeletal muscle mass. Multimodal prehabilitation approaches, including supervised exercise and dietary and psychological care, were well accepted. Serious adverse events related to prehabilitation were not reported.

To our knowledge, this study is the first systematic review of the effectiveness, acceptability, and safety of prehabilitation for patients with esophageal or gastroesophageal junction cancer during NAT. Although prehabilitation is considered a standard approach before major surgeries and several reviews have reported its efficacy, acceptability, and safety information on NAT was lacking [12]. Furthermore, a previous review [34] of prehabilitation during NAT for all types of cancer has not focused on important RCTs in patients with esophageal cancer [13, 14, 25]. During NAT, patients with esophageal or gastroesophageal junction cancer experience medical problems such as malnutrition [35] and mental illness [36] due to treatment toxicity. Therefore, we conducted a systematic review of prehabilitation during NAT to summarize the current evidence to develop prehabilitation program specifically for patients during NAT and presented three main findings. First, a meta-analysis of two RCTs [13, 14] showed an effect of prehabilitation on tolerance to NAT and a change in grip strength during NAT. Previous reviews [37, 38] have shown that frailty, namely reduced physiological reserve, leads to vulnerability to the stress of cancer treatment. Therefore, prehabilitation during NAT might contribute to preventing frailty for NAT and surgery against esophageal or gastroesophageal junction as part of multimodal cancer treatment. Second, one RCT and three non-RCTs [14, 15, 26, 28] showed that prehabilitation may reduce the risk of postoperative complications. This finding is essential because postoperative complications after esophagectomy are recognized as one of the most critical concerns. Third, prehabilitation during NAT may be considered safe, showing no significant adverse events related to prehabilitation, irrespective of its frequency, intensity, time, or type. Nevertheless, investigating strategies to enhance adherence to prehabilitation programs is essential. The sample size and number of RCTs are insufficient to make specific recommendations for prehabilitation of patients with esophageal or gastroesophageal junction cancer during NAT; however, these findings will guide the development of future prehabilitation programs.

A previous study [42] demonstrated that prehabilitation was effective in enhancing exercise capacity for patients with rectal cancer during NAT. This effective program was characterized by an excellent adherence rate of 96% [42]. Consequently, this review focused on identifying a prehabilitation program that would be acceptable for patients with esophageal cancer during NAT as a secondary outcome. Considering most reports conventionally consider ≥ 75% adherence as acceptable [43], in our review, attendance for dietary and psychological care was commendable, exercise attendance was 56–76%, with inadequate and acceptable programs. Notably, participation rates were acceptable (76%) [14, 29] for low- to moderate-intensity supervised exercise included in multimodal prehabilitation, with low dropout rates due to health-related or other reasons (0–7.7%), suggesting better feasibility. As patients with esophageal or gastroesophageal junction cancer experience multiple medical problems during NAT, multimodal interventions [14, 29] may be beneficial to promote adherence to prehabilitation. Previous studies [44‐46] theoretically indicated the importance of combining exercise therapy with nutritional and psychological care, supporting our hypothesis. Additionally, multilevel intensity exercise [47] and intensive monitoring may be needed for patients with disincentive factors such as baseline fitness weakness or poor motivation [48]. As for prehabilitation programs involving unsupervised or high-intensity exercise, adherence tends to be insufficient; thus, innovations for enhancing feasibility are warranted.

This review examined the safety of various evidence-based prehabilitation programs. Prehabilitation during NAT may be safe, showing no significant adverse events related to prehabilitation, irrespective of its frequency, intensity, time, or type. Previous reviews found exercise therapy during chemotherapy, including NAT, is generally safe for patients with cancers other than esophageal or gastroesophageal junction cancer [16, 17, 49]. Our review supports this, suggesting that prehabilitation during NAT for patients with esophageal or gastroesophageal junction cancer may also be safe.

The quality of the studies included in this review ranged from low to high; therefore, the results must be interpreted with caution. Specifically, we found that both RCTs and non-RCTs have a high risk of bias in the measurement of outcomes. This issue needs to be addressed in future clinical trials. Three RCTs were conducted, all of which had small sample sizes, raising concerns regarding imprecision. To determine the effectiveness and feasibility of prehabilitation during the NAT period, a meta-analysis should be conducted using clinical trials with larger sample sizes.

This systematic review had several limitations. First, there was heterogeneity in the NAT regimens. Second, there is insufficient RCTs regarding its effect on clinical outcomes. Future studies may reveal additional benefits for postoperative clinical outcomes, considering the potential effectiveness of prehabilitation during NAT. Third, we could not examine what composition would be best for multimodal prehabilitation during NAT. In recent years, multimodal supportive care for patients with advanced cancer are becoming the standard, and additional studies are needed to determine optimal combinations of multiple components. Fourth, the small number of included studies made it difficult to accurately assess reporting bias using funnel plot. Fifth, the sample sizes of the two RCTs are small, the intervention methods vary, and the results differ from the results of the sensitivity analysis, thus the robustness of the results is weak. Finally, due to the small sample size, we have not conducted a subgroup analysis to examine the impact of age on the effectiveness of prehabilitation. Considering the negative impact of postoperative complications on survival in older patients undergoing intensive chemotherapy [50], there is a critical need for prehabilitation during NAT to benefit older patients with advanced esophageal cancer. Further clinical studies are needed to examine the impact of age on the effect of prehabilitation during NAT.

This review demonstrates that prehabilitation during NAT may safely maintain physical fitness, improve tolerance to NAT, and reduce postoperative complications in patients with esophageal or gastroesophageal junction cancer. However, although the potential usefulness of prehabilitation was demonstrated, only two RCTs were eligible for inclusion, which suggested that the high-quality evidence was currently lacking. In the future, the benefit of prehabilitation during NAT and the mechanisms through which prehabilitation improves clinical outcomes should be examined in RCTs with sufficient sample sizes.