Efficacy of non-pharmacological interventions for the restoration of postoperative intestinal motility of patients with colorectal cancer: a systematic review and meta-analysis of randomized controlled trials

Abstract

Purpose

As a major indicator for rehabilitation progress, the restoration of postoperative intestinal motility of patients with colorectal cancer (CRC) is highly publicized. Various non-pharmacological interventions (NPIs) are helpful for bowel recovery. This study aims to evaluate the efficacy of NPIs in enhancing the restoration of postoperative intestinal motility.

Methods

Systematic searches were conducted in the Cochrane Library, PubMed, EBSCO, and Embase databases from inception to March 2025, with language restrictions to English. Eligible studies that quantified the effect of NPIs for the restoration of postoperative intestinal motility of patients with colorectal cancer were selected. The primary outcomes were time to first flatus and time to first defecation. The secondary outcomes were time to tolerate fluid food or solid food and length of hospital stay (LOS). Statistical analysis was performed in Review Manager 5.4, Stata v.15.1, and R 4.4.3 software.

Results

A total of 23 randomized controlled trials (RCTs) involving 3264 patients and 11 NPIs were selected. Gum chewing significantly shortened the time to first flatus (MD =  − 5.88, 95% CI − 10.33 to − 1.43; I² = 83.1%) and defecation (MD =  − 15.18, 95% CI − 21.67 to − 8.69; I² = 83.1%). Transcutaneous tibial nerve stimulation and electroacupuncture also significantly reduced the time to first flatus (MD =  − 14.95, 95% CI − 18.28 to − 11.61; I² = 39%) and defecation (MD =  − 7.29, 95% CI − 12.43 to − 2.14; I² = 0%), respectively. Network meta-analysis (NMA) identified 100 mg caffeine as the most effective intervention for shortening the time to first flatus, while decaffeinated coffee was the most effective for reducing the time to first defecation and length of hospital stay. In traditional pairwise meta-analysis and NMA, 65.7% (23/35) and 10% (1/10) of comparisons were rated as high confidence of evidence, respectively.

Conclusion

Dietary treatments are characterized by their low cost, diverse taste profiles, and lack of dependence on medical settings, making them more amenable to large-scale clinical practice promotion. Particularly, 100 mg caffeine, decaffeinated coffee, and gum chewing are worth paying attention to and recommending. For physical treatments, the limited number of studies and restricted application countries, coupled with higher costs and more stringent implementation requirements, pose significant challenges; therefore, they should be approached with caution. Given the limited evidence currently available, clinical decision-making regarding NPIs should integrate these findings with a focused consideration of the clinical setting, patient needs and preferences, and professional judgment.

Supplementary Information

The online version contains supplementary material available at 10.1007/s00384-025-04968-w.

Introduction

Colorectal cancer (CRC) is a significant global health concern. It is the third most commonly diagnosed cancer and the second leading cause of cancer-related mortality worldwide [1]. In 2020, there were approximately 1.9 million new cases and over 930,000 deaths attributed to CRC [2]. In the USA, it is estimated that there will be 154,270 new cases and 52,900 deaths from colorectal cancer in 2025 [3]. The incidence and mortality rates of CRC vary by region, with higher prevalence in developed countries and increasing trends in low- and middle-income regions [4].

Surgery is the primary treatment for CRC patients, and the recovery of postoperative intestinal motility is crucial for patient outcomes [4]. Postoperative gastrointestinal disorder (POGD) is a clinical syndrome that occurs after anesthesia and surgery, primarily characterized by a range of gastrointestinal symptoms. Among these, the weakening or disappearance of bowel sounds and gastrointestinal peristalsis, as well as the delayed passage of first postoperative flatus and stools, is particularly notable. Other symptoms may include vomiting, abdominal pain, distension, diarrhea, and constipation [5]. The incidence of POGD in CRC surgery ranges from 17.4 to 34.9%, with specific procedures such as intestinal resection and open surgical methods contributing to higher rates [1], negatively impacting postoperative quality of life [1] and prolonging hospital stays, which increases medical costs.

Currently, both pharmacological and non-pharmacological interventions (NPIs) are used to promote the recovery of intestinal motility after colorectal cancer surgery [4]. Pharmacological interventions, such as Alvimopan, a peripherally acting µ-opioid receptor antagonist, are used to accelerate the recovery of intestinal function. However, its costs may limit its widespread use after colorectal surgery [6, 7]. There were studies indicating that the use of prokinetics and nonsteroidal anti-inflammatory drugs (NSAIDs) had no significant effects on the restoration of intestinal motility [8, 9]. In addition, the beneficial effect of NSAIDs on the return of bowel function should be weighed against the potential increased risk of anastomotic leak of colorectal anastomoses [10]. NPIs, including gum chewing [11], electroacupuncture (EA) [12], acupressure [13], transcutaneous electrical acupoint stimulation (TEAS) [14], and transcutaneous tibial nerve stimulation (TTNS) [15], have gained attention due to their minimal invasiveness, cost-effectiveness, and low risk of side effects [4, 16]. In addition, NPIs have greater potential for promotion and popularization through the lack of prescription barrier, which allows patients, nurses, and rehabilitation therapists to take them autonomously. The American Society for Enhanced Recovery (ASER) and Perioperative Quality Initiative Joint Consensus Statement on Postoperative Gastrointestinal Dysfunction Within an Enhanced Recovery Pathway for Elective Colorectal Surgery [17] recommends consideration of coffee and gum chewing as adjuncts to enhanced recovery protocols in promoting recovery of GI function. However, individual patient differences in appetite and preferences, as well as cultural differences, can affect treatment compliance.

Despite the existence of a systematic review [18] on single NPIs such as coffee and gum chewing, the emergence of new evidence [19] may potentially alter the outcomes of previous systematic reviews. In addition, some prior reviews have become less applicable due to the introduction of new interventions.

In addition, there is a lack of head-to-head comparisons of different NPIs, making it difficult to determine the optimal treatment. Network meta-analysis (NMA) can provide valuable insights by conducting direct and indirect comparisons among various interventions and ranking their effectiveness [20]. There is no consensus-based definition of NPIs in health care [21]; physical intervention and dietary treatment are adopted to categorize the NPIs in this study.

Therefore, this study aims to assess the efficacy of different forms of NPIs to enhance the recovery of intestinal motility in patients undergoing colorectal surgery and comprehensively compare and rank the effectiveness of NPIs. The results may provide a basis for clinical practice and enhance supportive care for cancer patients undergoing colorectal surgery.

Method

We followed the Preferred Reporting Items for Systematic Reviews and Meta-Analysis (PRISMA) guidelines, and this study was registered with PROSPERO (CRD42024603606).

Information sources and retrieval strategy

Systematically searched the PubMed, Cochrane Library, Embase, and EBSCO databases from database inception to March 2025. The retrieval strategy was performed using search terms related to the combination of “intestinal motility” and “colorectal cancer” including all synonymous variations tailored to the respective databases. The search strategy was described in Appendix 1.

Eligibility criteria

Inclusion criteria

1. Type of study population (P): Patients with colorectal cancer who undergo surgery.

2. Type of intervention (I): NPIs such as gum-chewing, coffee, tea, caffeine, acupressure, electroacupuncture (EA), transcutaneous electrical tibial nerve stimulation (TTNS), transcutaneous electrical acupoint stimulation (TEAS), etc.

3. Type of comparison (C): Participants in the control group received “treat as usual” (TAU, including standard care, usual care) or NPIs different from those of the experimental group.

4. Type of outcome indicators (O): Bowel function recovery was evaluated according to (1) primary outcomes: time to first flatus, time to first defecation; (2) secondary outcomes: time to tolerate liquid diet, time to tolerate solid diet and the length of hospital stay(LOS).

5. Type of Study Designs (S): Randomized controlled trial published in English.

Exclusion criteria

1. Duplicate publications or unavailability of the full text;

2. Reference to “randomization” in the full text without a specific description of the randomization method. For example, randomized grouping is done using random number tables, randomization software, the touchball method, etc.

3. Combined interventions or mixed NPIs, where it was not possible to judge the effect of a single intervention;

4. Incomplete data.

Data extraction

Two researchers (Lanlan Zheng and Xuan Zhang) independently screened and extracted information and data from the literature, and when disagreements arose, a third party (Hui Yang) joined the discussion to resolve them. After a comprehensive search of the database literature, the references were traced, initial de-duplication was carried out using EndNote X9 (Thomson Reuters LLC, Philadelphia), titles and abstracts were read for initial screening, and full text was read to select conforming literature and evaluate the quality of the literature for final inclusion.

Excel was used for data extraction, and the main contents included (1) basic information: first author, year of publication, country of study, etc.; (2) patient characteristics: age, sample size (experimental group, control group), participant sex, etc.; (3) experimental group and control group interventions, frequency, program of intervention time, outcome indicators, etc. The means and standard deviations of preintervention and postintervention time were extracted from the studies. Several studies reported continuous variables as the means with ranges; we used the method proposed by Hozo et al. [22] to convert the variables into means and standard deviations. The screening process was performed in accordance with the Preferred Reporting Items of Systematic Reviews and Meta-Analyses (PRISMA) guidelines.

Primary and secondary endpoints

The primary endpoints include time to first flatus and time to first defecation. The secondary endpoints were: length of hospital stay(LOS), time to tolerate liquid diet, and time to tolerate solid diet.

Quality appraisal

The Cochrane risk-of-bias tool for randomized trials (RoB2) [23] was utilized to assess the risk of bias in the RCTs and to determine the methodological quality by two independent reviewers (Lanlan Zheng and Xuan Zhang). The overall risk-of-bias judgment of a specific trial was assessed on the basis of the following five domains, which included the randomization process, deviations from intended interventions, missing outcome data, measurement of the outcome, and selection of reported results. The risk-of-bias judgments were categorized into three levels of bias risk: “high risk,” “low risk,” and “some concerns” based on stipulated evaluation criteria. A third researcher (Hui Yang) adjudicated the inclusion decision in the event of any disagreement.

The certainty of the evidence for each outcome was assessed using the Grading of Recommendations Assessment, Development and Evaluation (GRADE) profiler Guideline Development Tool [24]. The assessment criteria in this tool are risk of bias, inconsistency, indirectness, imprecision, and publication bias. The quality of evidence for each outcome is classified as very low, low, moderate, or high.

Confidence in Network Meta-Analysis (CINAeM) software, a developed web application based on the Grading of Recommendations, Assessment, Development and Evaluation (GRADE) framework, was used to determine the strength of evidence [25].

Statistical analysis

Meta-analysis

Pairwise meta-analyses for direct comparisons were conducted by using Review Manager 5.4 (The Cochrane Collaboration, 2020) and Stata 15.1 (StataCorp, College Station, TX). In the pairwise meta-analysis, the pooled effect sizes were expressed as mean difference (MD) and standardized mean difference (SMD) for continuous outcomes with corresponding 95% CIs. Cochrane’s Q test and I² statistics were used to evaluate heterogeneity across the studies. A p-value of equal to or less than 0.10 in Cochrane’s Q test and I² of greater than 75% was considered to indicate high heterogeneity. A subgroup analysis of setting outcomes was performed to evaluate the effectiveness of NPIs in patients who underwent either open or laparoscopic surgery. Sensitivity analysis was conducted, excluding studies predominantly responsible for heterogeneity. We did not assess publication bias because the number of included studies for each outcome was fewer than 10.

Network meta-analysis

The network meta-analysis was performed only when a minimum of two studies demonstrated homogeneity in participants, interventions, outcomes, and other relevant factors [26]. R 4.4.3 software and the GeMTC package were used to conduct a network meta-analysis based on a Bayesian framework for both direct and indirect comparisons. STATA 15.1 (StataCorp, College Station, TX) generated the network plot. Two Markov chains were chosen, with an initial run of 20,000 iterations and an additional 50,000 iterations. To assess the consistency model, 20,000 annealing steps and 50,000 iterations are performed. The iterative results—comparing the Deviance Information Criterion (DIC) values from this model to those from an inconsistency model—were analyzed. This study featured a Bayesian network meta-analysis of five NPIs related to coffee and caffeine intake, as well as tea, assessing their impact on time to first flatus, time to first defecation, and LOS.

Heterogeneity was mainly determined according to the I² values. The loop-specific approach and the node-splitting method were used to assess the global and local inconsistencies [27]. When the network plot contains closed loops, local inconsistency was assessed through the node-splitting method. A p-value greater than 0.05 suggests that the data were consistent. A p-value < 0.05 was considered to indicate statistical significance. A league table was used to analyze the results of the comparisons among coffee or caffeine interventions based on a network meta-analysis. The effect ranking of NPIs was measured by calculating the surface under the cumulative ranking area (SUCRA). For the outcome indicators of this study, a lower SUCRA value indicates a greater likelihood that the treatment was superior [28].

Results

Research screening and description

A total of 3353 studies were retrieved using the search strategy in four databases. After removing 839 duplicate articles, 2514 articles were screened, and then 40 full-text articles were assessed for eligibility. Finally, 23 studies and 3264 patients were included in this study [11–16, 29–45]. The PRISMA flowchart (Fig. 1) outlined the literature screening process and the reasons for the exclusion.

Fig. 1.

Flow chart of the study identification, screening, eligibility assessment, and inclusion processes

Basic characterization

This study, encompassing 23 studies from ten countries published between 2002 and 2024, evaluated 11 NPIs for restoring intestinal motility in CRC patients. The NPIs included seven dietary treatments (gum chewing, 100 mg caffeine, 200 mg caffeine, caffeinated coffee, decaffeinated coffee, tea, and soluble dietary fiber enteral nutrition) and four physical treatments (EA, acupressure, TTNS, and TEAS).

Among the 23 studies, the three most commonly tested NPIs were gum chewing, coffee and caffeine intake, as well as EA. Chewing gum was also the most widely used NPIs in research applications across countries, followed by caffeine and coffee intake. However, the application of EA in research was limited to China, and the research application countries for other physical treatments were also relatively limited.

The basic characteristics of the included literature were shown in Table 1. Besides coffee and caffeine intake, non-significant outcomes for other NPIs were shown in Table 2, respectively. The operational definitions of the included NPIs were shown in Supplementary Table S1.

Table 1.

The basic characteristics of the included literature

Study Country	Population(P) a. Sample(intervention group vs control group) b. Age(mean±SD，or mean,intervention group vs control group)	Intervention(I) a. Types of Intervention b. Periodic c. Frequency (duration or times/day)	Control(C) Treatment	Outcome(O) ①②③④⑤⑥ Intervention group vs control group	Side effects T: test group C: control group
China [14]	a.303 vs 307 b.61.9±2.8 vs 62.9±2.8	a.TEAS b.3 days c.30 min, 1 time/day	Sham TEAS without electrical impulse	①: (63.4±29)h vs (76.3±37.9)h ②: (106.7±65.6)h vs (121.5±65.6)h	Not reported
China [16]	a.44 vs 41 b.54.3 vs 53.1	a. EA b.6 days c.1 time/day	Standard postoperative care without acupuncture	①: (68.26±23.38)h vs (65.24±17.50)h ②: (119.04±47.97)h vs (119.38±60.21)h	No adverse events
China [13]	a.30 vs 30 b.61.9±13.1 vs 62.8±15.7	a.Acupressure b.Not reported c.3 minutes, 3 times/day	Routine postoperative care with sham acupressure	①: (66.72±22.08)h vs (86.64±27.60)h ②: (114±28.08)h vs (126±47.04)h ③: (82.32±23.28)h vs (102.24±34.08)h ④: (108.96±28.08)h vs (122.64±47.76)h	Not reported
China [43]	a.19 vs 20 b.63±9 vs 60±10	a.EA b.4 days c.30 min, 1 time/day	Sham acupuncture	①: (23±14)h vs (32±18)h ②: (68±45)h vs (72±53)h ⑤: (11±3)d vs (12±3)d	Not reported
China [12]	a.125 vs 123 b.60.2±11.0 vs 60.2±11.8	a.EA b.4 days c.30 min, 1 time/day	Sham EA at nonacupoints without electric current	①: (46.62±15.9)h vs (58.13±14.4)h ②: (80.48±21.9)h vs (87.86±20.03)h ③: (104.29±24.98)h vs (115.09±33.76)h ④: (180.85±46.43)h vs (194.84±47.63)h ⑤: (8.14±2.1)d vs (8.32±2.33)d	T: 5 cases of hematoma; 1 case of sharp pain; 2 cases of residual needling sensation after needle removal C: 4 cases of hematoma; 1 case of sharp pain; 2 cases of residual needling sensation after needle removal
China [14]	a.303 vs 307 b.61.9±2.8 vs 62.9±2.8	a.TEAS b.3 days c.30 min, 1 time/day	Sham TEAS without electrical impulse	①: (63.4±29)h vs (76.3±37.9)h ②: (106.7±65.6)h vs (121.5±65.6)h	Not reported
China [16]	a.44 vs 41 b.54.3 vs 53.1	a. EA b.6 days c.1 time/day	Standard postoperative care without acupuncture	①: (68.26±23.38)h vs (65.24±17.50)h ②: (119.04±47.97)h vs (119.38±60.21)h	No adverse events
China [13]	a.30 vs 30 b.61.9±13.1 vs 62.8±15.7	a.Acupressure b.Not reported c.3 minutes, 3 times/day	Routine postoperative care with sham acupressure	①: (66.72±22.08)h vs (86.64±27.60)h ②: (114±28.08)h vs (126±47.04)h ③: (82.32±23.28)h vs (102.24±34.08)h ④: (108.96±28.08)h vs (122.64±47.76)h	Not reported
China [43]	a.19 vs 20 b.63±9 vs 60±10	a.EA b.4 days c.30 min, 1 time/day	Sham acupuncture	①: (23±14)h vs (32±18)h ②: (68±45)h vs (72±53)h ⑤: (11±3)d vs (12±3)d	Not reported
China [12]	a.125 vs 123 b.60.2±11.0 vs 60.2±11.8	a.EA b.4 days c.30 min, 1 time/day	Sham EA at nonacupoints without electric current	①: (46.62±15.9)h vs (58.13±14.4)h ②: (80.48±21.9)h vs (87.86±20.03)h ③: (104.29±24.98)h vs (115.09±33.76)h ④: (180.85±46.43)h vs (194.84±47.63)h ⑤: (8.14±2.1)d vs (8.32±2.33)d	T: 5 cases of hematoma; 1 case of sharp pain; 2 cases of residual needling sensation after needle removal C: 4 cases of hematoma; 1 case of sharp pain; 2 cases of residual needling sensation after needle removal
Italy [45]	a.170 vs 170 b.70.6±8.6 vs 66.8±10.3	a.TTNS b.3 days c.continuous stimulation for 72h	Sham TTNS without electrical impulse	①: (30.5±22.06)h vs (42.5±30.11)h ②: (81±24.98)h vs (88±23.65)h ⑤: (4.97±2.22)d vs (5.44±2.05)d	Not reported
France [15]	a.15 vs 19 b.67.2±7.9 vs 65.1±5.4	a.TTNS b. until GI motility was recovered C.20 min, 3 times/day	Sham TTNS without electrical impulse	①: (35.28±4.56)h vs (51.84±7.68)h ⑤: (7.47±0.33)d vs (7.71±0.44)d	Not reported
Netherlands [29]	a.52 vs 60 b.66±9 vs 67±11	a.Gum chewing b.Not reported c.Not reported	Dermal patch	⑤: (9.5±4.9)d vs (14±14.5)d	Not reported
Japan [33]	a.21 vs 22 b.66.4±13.9 vs 68.0±9.6	a.Gum chewing b.from PoD 1 until the first day of regular oral dietary intake c.5 min, 3 times/day	Not to chew gum	①: (53±2)h vs (49±26)h ②: (94±44)h vs (109±35)h ③: (74.4±24)h vs (72±40.8)h ④: (108±36)h vs (105.6±40.8)h ⑤: (19.8±12.6)d vs (16.4±16.9)d	Not reported
Britain [11]	a.199 vs 203 b.65.5±14.1 vs 66.9±11.6	a.Gum chewing b.5 days c.10 min, 4 times/day	Not to chew gum	①: (56.4±18)h vs (48±35.76)h ④: (48±35.76)h vs (48±35.76)h ⑤: (7.7±4.48)d vs (7.7±4.48)d	Not reported
China [30]	a.41 vs 41 b.68±10 vs 69±12	a.Gum chewing b. from PoD 1 until discharge c.30 min, 3 times/day	Satndard postoperative care	①: (23.84±19.62)h vs (42.22±31.62)h ⑤: (6.58±4.62)d vs (7.28±5.31)d	Not reported
China [34]	a.190 vs 189 b.53.3±14.9 vs 54.1±16.2	a.Gum chewing b.5 days c.10 min, 3 times/day	Not to undertake any postoperative intervention.	①: (61.11±9.03)h vs (62.93±8.56)h ②: (119.46±33.09)h vs (125.48±32.12)h ⑤: (10.75±3.26)d vs (10.78±2.54)d	Not reported
America [31]	a.17 vs 17 b.60.0±6.1 vs 63.0±8.5	a.Gum chewing b. from PoD 1 until discharge c.1 hour, 3 times/day	Not to chew gum	①: (65.4±14.8)h vs (80.2±19.1)h ②: (63.2±5.4)h vs (89.4±24)h ⑤: (4.3±0.43)d vs (6.8±1.38)d	Not reported
Japan [32]	a.10 vs 9 b.58.6±9.1 vs 60.6±6.0	a.Gum chewing b.until the day begins, oral intake c. 3 times/day	Not to chew gum	①: (50.4±12)h vs (76.8±21.6)h ②: (74.4±26.4)h vs (139.2±52.8)h ⑤: (13.5±3)d vs (14.5±6.1)d	No adverse events
China [35]	a.43 vs 46 b.62.7±7.84 vs 63.37±7.23	a.Gum chewing b.Not reported c.30 min, 3 times/day	Standard postoperative care	①: (42.33±3.46)h vs (49.2±1.42)h ②: (66.07±2.36)h vs (78.37±1.62)h	Not reported
Switzerland [39]	a.49 vs 47 b.63.3±3.8 vs 68.5±3.6	a.Caffeinated coffee b.Not reported c.3 times/day	Tea	①: (39.91±5.14)h vs (30.91±4.51)h ②: (66.8±50.53)h vs (77.3±47.37)h ⑤: (6.09±0.67)d vs (7±0.45)d	Not reported
Israel [37]	a. 30 vs 28 b. 56.9±12.77 vs 55.4±15.48	a.100mg caffeine b.7 days c.3 times/day	Water	①: (51.98±21.25)h vs (57.26±18.83)h ②: (59.86±35.5)h vs (82.12±41.64)h ④: (93.39±39.08)h vs (108.35±41.01)h ⑤: (6.36±2.34)d vs (7±1.56)d	Not reported
Lithuania [40]	a. 30 vs 30; 30 vs 30 b. 67.3±6.8 vs 62.4±10.8 vs 66.3±9.1	a. Coffee with or without caffeine b. Not reported c. 3 times/day	Water	②: (90±36.72)h vs (72±36)h vs (99.36±27.6)h	Not reported
America [38]	a. 37 vs. 31 vs. 31 b. 59.5±15.0 vs 63.7±14.7 vs 61.6	a. Caffeinated or decaffeinated coffee b. 7 days c. 3 times/day	Water	①: (56.64±18.48)h vs (48±37.2)h vs (43.68±39.84)h ②: (72±36.96)h vs (56.64±18.72)h vs (72±37.2)h ⑤: (4.36±2.31)d vs (3.36±0.78)d vs (4±1.55)d	Not reported
Germany [42]	a. 35 vs 36 b. Not reported	a. Caffeinated coffee b. Not reported c. 3 times/day	Water	②: (62.1±21.5)h vs (73.7±22)h ⑤: (10±3.2)d vs (11.6±4.6)d	Not reported
China [44]	a.30 vs 30 b.52.2±14.2 vs 53.2±13.1	a. SDFEN b. 7 days c. Not reported	Enteral nutrition without dietary fiber	①:(58.93±6.5)h vs (63.03±4.8)h ⑤: (7.4±0.72)d vs (7.63±0.61)	Not reported
Japan [41]	a.7 vs 10 b.74.6±6.1 vs 68.7±10.2	a. Tea b. from PoD1 until the first bowel movement c. 1 time/day	Water	②: (45±22.7)h vs (49.4±42.2)h ⑤: (10.5±3)d vs (8.9±1.1)d	T: 1 case of pseudogout C: No adverse events

①time to first flatus; ②time to first defecation; ③time to tolerance of liquid diet; ④time to tolerance of solid diet; ⑤length of hospital stay; ⑥time to first sensation of hunger; SDFEN Soluble Dietary Fiber Enteral Nutrition; TTNS Transcutaneous Tibial Nerve Stimulation; TEAS Transcutaneous Electrical Acupoint Stimulation

Table 2.

Primary and secondary non-significant outcomes

Intervention	Outcome	No. of included studies	No. of included patients		MD [95% CI]/ SMD [95% CI]	p-value	Heterogeneity level
			Treatment	Control			I2 (%)	p-value
Gum chewing	Secondary
	Length of hospital stay (days)	7 [11, 29–34]	530	541	− 0.98 [− 2.12, 0.16]	0.09	89	< 0.01
Electroacupunture	Primary
	Time to first flatus(hours)	3 [12, 16, 43]	188	184	− 6.29 [− 15.32, 2.74]	0.17	0	0.01
	Secondary
	Length of hospital stay (days)	2 [12, 43]	144	143	− 0.24 [− 0.77, 0.29]	0.37	0	0.41
Acupressure	Primary
	Time to first defecation	1 [13]	30	30	− 12.00 [− 31.60, 7.60]	0.23	NA	NA
	Secondary
	Time to tolerate solid diet	1 [13]	30	30	− 13.68 [− 33.51, 6.15]	0.18	NA	NA
SDFEN	Primary
	Length of hospital stay	1 [44]	30	30	− 0.23 [− 0.57, 0.11]	0.18	NA	NA

Risk of bias assessment and quality of evidence

The summary of the risk of bias was reported in Supplementary Fig. S1. Overall, seven trials were rated as having some concerns regarding risk of bias, and one RCT related to acupressure was judged as having a high risk of bias, which was not included in quantitative synthesis. In addition, 15 studies were rated as low risk of bias. The certainty of direct and network evidence was shown in Supplementary Tables S2 and S3.

Traditional pairwise meta-analysis

The results of pairwise meta-analysis were presented in Supplementary Table S2 and Appendix 2.

Dietary treatments

Gum chewing

For patients with colorectal cancer undergoing colorectal surgery, gum chewing (MD=−5.88, 95%CI (−10.33,−1.43); I²=83.1%), (MD=−15.18, 95%CI (−21.67, −8.69); I²=83.1%) was statistically significantly more effective than TAU control in reducing time to first flatus and first defecation, respectively. The subgroup analyses were based on a surgical approach. In laparoscopic surgery, gum chewing showed significant difference in reducing time to first flatus (MD=−21.09, 95%CI (−30.36,−11.82); I²=0%). In open surgery  (MD=−24.03, 95%CI (−34.53,−13.53); I²=0%), laparoscopic surgery (MD=−64.80, 95%CI (−102.98,−26.62); I²=NA), mixed open and laparoscopic surgeries (MD=−14.26, 95%CI (−23.03,−5.49); I²=50%), gum chewing all showed significant improvement in shortening the time to first defecation.

Coffee and caffeine intake

For patients with colorectal cancer undergoing colorectal surgery, decaffeinated coffee (MD =  − 20.75, 95% CI (− 31.63, − 9.87); I² = 14%), caffeinated coffee (MD =  − 15.84, 95% CI (− 23.22, − 8.46); I² = 23%), and 100 mg caffeine (MD =  − 22.26, 95% CI (− 42.24, − 2.28); I² = NA) were statistically significantly more effective than TAU control in shortening the time to first defecation. Compared with caffeinated coffee, decaffeinated coffee statistically significantly reduced the time to first defecation by 16.29 (95% CI (− 27.24, − 5.35); I² = 0%)h.

Patients in the 100 mg caffeine group (MD =  − 13.67, 95% CI (− 27.20, − 0.14); I² = 0%) tolerated solid diet significantly earlier than the TAU control group.

Comparing caffeinated coffee against tea, LOS was reduced by 0.91 (95% CI (− 1.14, − 0.68)) days. While comparing tea against caffeinated coffee, the first flatus time was reduced by 9 (95% CI (− 10.93, − 7.07))h (Appendix 2).

Physical interventions

Electroacupuncture

For patients with colorectal cancer undergoing colorectal surgery, EA (MD=−7.29, 95%CI (−12.43, −2.14); I²=0%) was statistically significantly more effective than TAU control in shortening time to first defecation.

Transcutaneous electrical tibial nerve stimulation

For patients with colorectal cancer undergoing colorectal surgery, TTNS (MD=−14.95, 95%CI(−18.28, −11.61); I²=39%), (SMD=−0.25, 95%CI (−0.46, −0.05); I²=1%) was statistically significantly more effective than TAU control in reducing time to first flatus and LOS, respectively.

Transcutaneous electrical acupoint stimulation

For patients with colorectal cancer undergoing colorectal surgery, TEAS (MD=−12.90, 95%CI(−18.25, −7.55)), (MD=−14.8, 95%CI (−25.21, −4.39)), (MD=−51.60, 95%CI (−88.15, −15.05)) were statistically significantly more effective than TAU control in shortening time to first flatus and defecation, respectively.

Sensitivity analyses

Leave-one-out sensitivity analysis for gum chewing on time to first flatus, time to first defecation, and LOS, as well as EA on time to first flatus yielded results consistent with the primary analysis, supporting the robustness and reliability of the findings (Supplementary Appendix 3).

Network meta-analysis

Network plot

Each node represents a type of NPI, and the size of the nodes indicates the sample size. Lines connecting the nodes indicate direct comparisons between two types of NPIs. The number of studies directly comparing their effects is indicated by the thickness of the line connecting the two nodes. The network plots of some NPIs are shown in Fig. 2.

Fig. 2.

Network plot of each intervention under different outcome indicators. Note: (1): water; (2): 100 mg caffeine; (3) 200 mg caffeine; (4) caffeinated coffee; (5) decaffeinated coffee; (6) tea. Each dot represents one intervention, and a line between dots indicates the presence of a direct comparison between interventions; a thicker line indicates that more studies have directly compared the two interventions

Time to first flatus

The thickest connecting line between 100 mg caffeine (2) and water (1) implied that 100 mg caffeine had the highest number of studies (two RCTs). Four studies [36–39], with 280 patients, reported the effectiveness of six NPIs in shortening the time to first flatus. Two triangular closed loops were formed between the 100 mg caffeine, 200 mg caffeine, and water nodes, as well as caffeinated coffee, decaffeinated coffee, and water nodes, respectively.

Time to first defecation

The thickest connecting line between caffeinated coffee (4) and water (1) implied that caffeinated coffee had the highest number of studies (three RCTs). With 335 patients, five studies [37, 38, 40–42] reported the effectiveness of five NPIs in shortening the time to first defecation. A triangular closed loop was performed between the caffeinated coffee, decaffeinated coffee, and water nodes, indicating the existence of simultaneous direct and indirect comparisons between interventions.

Length of hospital stay

Five studies [37–39, 41, 42] reported the length of hospital stay by five NPIs in a total of 341 patients, including 100 mg caffeine, caffeinated coffee, decaffeinated coffee, tea, and water. The thickest connecting line between caffeinated coffee (4) and water (1) implied that caffeinated coffee had the highest number of studies (two RCTs). Two triangular closed loops were formed between the caffeinated coffee, decaffeinated coffee, and water nodes, as well as caffeinated coffee, tea, and water nodes, respectively.

Consistency analysis

Consistency modeling was performed on the data with the time to first flatus, time to first defecation, and LOS as the outcome indicators. After 50,000 iterations, the results of DIC were compared with the results of DIC of the inconsistency, and the values of the difference were within five, indicating that all the above data were in line with the consistency. The potential scale reduction factors (PSRF) were all 1, indicating high data convergence. Details are provided in Table 3. The overall heterogeneity parameters were all I² ≤ 50%, p > 0.1, so all fixed-effects models were used.

Table 3.

Consistency analysis

		Consistent model	Inconsistent model
Time to first flatus	DIC	18.852746	18.900044
	I2	5%	5%
	PSRF	1	1
Time to first defecation	DIC	21.24217	21.31948
	I2	0%	0.07%
	PSRF	1	1
Length of hospital stay	DIC	12.256679	11.303474
	I2	21%	12%
	PSRF	1	1

Effectiveness and ranking of NPIs

Time to first flatus

Under the consistency model, compared with the water control group, NMA indicated that 100 mg caffeine, 200 mg caffeine, caffeinated coffee, decaffeinated coffee, and tea did not significantly shorten the time to the first flatus. A two-by-two comparison of the interventions revealed that tea was better than caffeinated coffee [− 9.00 (− 10.93, − 7.07)] (p < 0.05) (Fig. 3).

Fig. 3.

League table of the outcomes. A League table of time to first flatus; B League table of time to first defecation; C League table of length of postoperative hospital stay. ^aThe difference between two-by-two comparisons was statistically significant

In this study, smaller SUCRA values indicate better efficacy. As ranked by SUCRA value, the relative efficacy of NPTs showed that 100 mg caffeine (23.3%) ranked in the top one. The remaining NPIs were ranked as follows: water (40%), 200 mg caffeine (46.7%), tea (50.1%), decaffeinated coffee (52.8%), and caffeinated coffee (87.2%). (Supplementary Figure S2).

Time to first defecation

Under the consistency model, compared with the water control group, NMA indicated that decaffeinated coffee [− 22.74 (− 32.39, − 13.10)], 100 mg caffeine [− 22.26 (− 42.24, − 2.28)], and caffeinated coffee [− 9.10 (− 16.83, − 1.38)] shorten the time to first defecation. A two-by-two comparison of the interventions revealed that decaffeinated coffee was better than caffeinated coffee [− 13.64 (− 23.39, − 3.89)].

As ranked by SUCRA value, the relative efficacy of NPIs showed that decaffeinated coffee (17.5%) ranked in the top one. The remaining NPIs were ranked as follows: 100 mg caffeine (23.3%), caffeinated coffee (58%), tea (64.8%), and water (95%). (Supplementary Fig. S2).

Length of hospital stay

Under the consistency model, compared with the water control group, NMA indicated that 100 mg caffeine, caffeinated coffee, decaffeinated coffee and tea did not significantly reduce the length of hospital stay. (Fig. 3).

As ranked by SUCRA value, the relative efficacy of NPIs showed that decaffeinated coffee (41.2%) ranked in the top one. The remaining NPIs were ranked as follows: caffeinated coffee (41.5%), water (42.5%), 100 mg caffeine (56.6%), and tea (68.3%) (Supplementary Fig. S2).

Strength of evidence

In the time to the first defecation, one of ten comparisons was rated as high confidence of evidence, nine as low. High confidence of evidence related to decaffeinated coffee (Supplementary Table S3).

Discussion

Implementation of fast-track protocols has been a successful strategy for overcoming POGD [46]. Nevertheless, the search for affordable and effective preventive strategies to further decrease surgical complications and associated expenses persists. To our knowledge, this is the first systematic review and NMA to summarize and assess the comparative efficacy of NPIs for the restoration of intestinal motility in patients with CRC undergoing colorectal surgery.

Among dietary treatments, we found that gum chewing was significantly effective in shortening the time to first flatus and defecation. It stimulates bowel motility recovery after colorectal cancer surgery by mimicking the act of eating, which activates the cephalic-vagal reflex [47]. This reflex enhances gastrointestinal motility and the release of gastrointestinal hormones, thereby accelerating the time to first flatus and defecation. The “sham feeding” behavior effectively stimulates the oral and vagal nerves, promotes the secretion of digestive fluids, and enhances gastrointestinal motility [48]. Sarmiento et al. [49] recently published a systematic review, concluding that gum chewing was a measure with strong evidence to decrease the time to first flatus and defecation, which was in accordance with our findings. Compared with Sarmiento’s findings, our systematic review, which included subgroup analyses by surgical approach, showed that gum chewing effectively shortened the time to first flatus and defecation after laparoscopic surgery for colorectal cancer. Thus, gum chewing can be recommended to promote bowel motility recovery in CRC patients undergoing laparoscopic surgery. Gum chewing is not restricted by dietary culture; it is also a cost-effective NPI with a very low incidence of side effects. Based on the results of this study, chewing gum is worth recommending for patients undergoing laparoscopic surgery for colorectal cancer.

Based on the combined evidence, we found that 100 mg caffeine may be the optimal intervention among coffee and caffeine intake in shortening the time to first flatus, and decaffeinated coffee may be the optimal intervention in shortening the time to first defecation and the length of hospital stay. The results of the current meta-analysis with seven included studies revealed, in accordance with the previously published literature [50], that caffeine intake accelerated the restoration of intestinal motility. The results of the NMA showed that decaffeinated coffee was the most effective NPI on defecation for patients with CRC undergoing colorectal surgery. Decaffeinated coffee recorded a significantly shorter time to first defecation when compared to caffeinated coffee, which aligned with the results of studies by Zizzo et al. [51] and Vaghiri et al. [50]. What is more, decaffeinated coffee may be the optimal intervention among coffee and caffeine intake in shortening LOS. This suggests that factors beyond caffeine may be significantly involved in bowel function, and the effect of decaffeinated coffee could be the formation of chemically active agents during decaffeination [52, 53]. Caffeine promotes postoperative recovery of gastrointestinal function through vasodilation. Nevertheless, pooled analysis of minimally invasive methods pointed out a significant benefit of 100 mg caffeine with respect to the resumption of a solid diet, while the 200 mg caffeine group did not record a significantly shorter time to first solid diet tolerance when compared to the control group. In this study, 100 mg caffeine was also the optimal NPI among coffee and caffeine intake on the first flatus. However, in certain cultures, these interventions may not be immediately accepted. Therefore, once patient acceptance is confirmed, the optimal approach is to recommend 100 mg caffeine or decaffeinated coffee for clinical use.

Compared to traditional acupuncture, a key advantage of EA is its capacity to precisely control the frequency and intensity of stimulation at acupoints, enabling accurate and measurable adjustments [54]. In this study, EA showed a significant reduction in the time to first defecation after colorectal surgery, which was consistent with other systematic reviews [7, 49, 54]. In contrast to earlier findings, our meta-analysis found no significant association between EA and a shorter time to first flatus compared to standard care or sham-EA. The difference in results may be due to the distinct study designs and inclusion criteria. Previous studies used EA with or without other therapies, making it hard to isolate EA’s specific effect. Our meta-analysis excluded combined or mixed interventions where a single effect could not be judged, focusing only on EA’s independent effect. This rigorous approach likely explains why we found no significant association between EA and a shorter time to first flatus. As far as we know, there were multiple systematic reviews and meta-analyses that evaluated the efficacy of TTNS on fecal incontinence [55], overactive bladder [56], and low anterior resection syndrome [57], while this is the first study to conduct a meta-analysis to examine the efficacy of TTNS on the restoration of intestinal motility in patients with CRC undergoing colorectal surgery. The pooled analysis of two included studies [15, 45] demonstrated a shorter length of hospital stay in the TTNS group than in the control group while no significant difference in time to the first flatus, which may be attributed to the limited number of included studies.

Soluble dietary fiber, which is fermented in the gut and serves as a nutrient, can protect the intestinal barrier, modulate immune function, and potentially reduce the risk of inflammatory responses and postoperative complications. As shown by the results of the only included study for each non-pharmacological intervention (NPI), early postoperative soluble dietary fiber enteral nutrition (SDFEN) [44] is feasible and beneficial for immune function, reducing inflammatory reactions and accelerating gastrointestinal recovery. TEAS [14] can also increase anti-inflammatory actions to prevent postoperative ileus, while acupressure [13] can accelerate gastrointestinal recovery.

The Enhanced Recovery Pathway (ERP) [17] recommends considering coffee and gum chewing as adjuncts to promote gastrointestinal function recovery, while it does not cover other non-pharmacological interventions. Moreover, our study provides a detailed analysis of the effects of different caffeine doses and coffee types on the recovery of intestinal motility, thereby offering a valuable supplement to this ERP recommendation.

Given these promising findings, future research should aim to expand the scope of investigation to further explore the synergistic effects of combining NPIs with established ERPs to further enhance the restoration of intestinal motility in CRC patients.

Strengths and limitations

A major strength of this study is the inclusion of only randomized controlled trials (RCTs), which enhances the reliability of our findings and increases the likelihood that we can draw reasonable conclusions. However, several limitations must be acknowledged. First, our search for relevant studies was restricted to English language publications, which may have overlooked pertinent studies published in other languages. Second, the focus of our network meta-analysis was limited to coffee and caffeine-related articles due to the absence of studies directly comparing other non-pharmacological interventions (NPIs). This restricts our ability to perform a comprehensive network meta-analysis and identify the optimal NPI for enhancing the recovery of intestinal motility. Lastly, a significant limitation is the potential confounding effect of the increasing adoption of various Enhanced Recovery After Surgery (ERAS) components over the study period. Given the wide timespan of the included studies, the evolving ERAS components could have contributed to variability in the results. While we have attempted to account for these factors in our analysis, we acknowledge that this limitation may affect the interpretation of our findings.

Implications for future research

Future research should focus on conducting well-designed randomized controlled trials (RCTs) that directly compare a broader range of NPIs and standardize outcomes related to bowel recovery to facilitate more meaningful comparisons across studies. By overcoming the current gaps in direct comparisons, future meta-analyses will be better positioned to identify an optimal strategy for enhancing bowel recovery, ultimately guiding clinical practice and improving patient outcomes in postoperative care.

Conclusion

The NPIs for postoperative bowel motility recovery in colorectal cancer are abundant, and the findings of this study underscore NPIs’ potential benefits in restoring intestinal motility and enhancing postoperative recovery. Among coffee and caffeine intake, 100 mg caffeine and decaffeinated coffee may be the optimal NPI for enhancing the return of bowel function. Due to low cost and diverse flavor, gum chewing could be recommended as an optimal NPI in CRC laparoscopic surgery. While dietary interventions, in particular, offer a convenient option for clinical implementation and can be effectively introduced via educational materials, they allow patients to make personalized choices based on taste and habits. For physical treatments, the limited number of studies and restricted application countries, coupled with higher costs and more stringent implementation requirements, pose significant challenges. Therefore, it should be approached with caution. Overall, more high-quality studies are still needed to fully assess the impact of various NPIs on intestinal motility restoration.

Supplementary Information

Below is the link to the electronic supplementary material.

Author contribution

Conceptualization: Hui Yang; Methodology: Lanlan Zheng, Hui Yang; Formal Analysis: Lanlan Zheng, Xuan Zhang, Baoyu Ma and Yuan Yuan; Writing-Original Draft Preparation: Lanlan Zheng, Xuan Zhang, Baoyu Ma and Yuan Yuan; Supervision: Hui Yang; Writing-Review & Editing: Hui Yang. All authors reviewed the manuscript.

Data availability

No datasets were generated or analysed during the current study.

Declarations

Competing interests

The authors declare no competing interests.