Authors: Jonathan Sivakumar, Qianyu Chen, Cuong Phu Duong
Categories: Review Article, Delayed gastric conduit emptying, Esophagectomy, Incidence
Source: Esophagus
Authors: Jonathan Sivakumar, Qianyu Chen, Cuong Phu Duong
Delayed gastric conduit emptying (DGCE) is a significant and prevalent complication following esophagectomy, adversely affecting recovery and quality of life. The true burden of DGCE remains uncertain due to considerable variability in reported incidence rates. This study aimed to determine the incidence of DGCE following esophagectomy and how this is impacted with varying diagnostic criteria. A systematic review was conducted across major databases, including PubMed, MEDLINE, Embase, Web of Science, and Cochrane, to identify studies reporting the incidence of DGCE following esophagectomy. The incidence rates were pooled and analyzed using a random-effects model, with subgroup analyses for potential sources of heterogeneity such as pyloric interventions, conduit dimensions, and anastomotic height. Among 5176 screened records, 125 studies met the eligibility criteria. The pooled incidence of early DGCE was 15.9% (95% CI 11–21%), and late DGCE was 9.4% (95% CI 7.1–11.9%). Significant heterogeneity was observed across studies, driven by variations in diagnostic methods. Subgroup analysis indicated that prophylactic pyloric drainage was not associated with a statistically significant effect on early DGCE (OR 0.76; p = 0.38) or late DGCE (OR 0.71; p = 0.44). DGCE represents a significant burden for esophageal cancer survivors, with considerable variability in its reported incidence, underscoring the urgent need for a standardized diagnostic criterion. The adoption of the recently published international consensus definition is crucial for reducing the heterogeneity, as well as improving the identification and management of DGCE.
The online version contains supplementary material available at 10.1007/s10388-025-01133-8.
The main curative treatment modality for esophageal cancer (OC) is tumor resection, with or without perioperative chemotherapy or chemoradiation. Esophagectomy is a major operation which involves OC resection with radical lymphadenectomy, followed by mobilization and tubularization of the remnant stomach to create an esophago-gastric anastomosis in the thoracic cavity [1, 2]. As OC survivorship has increased, there has been a growing emphasis on functional status and health-related quality of life following esophagectomy.
Delayed gastric conduit emptying (DGCE) is major contributor to impaired quality-of-life for post-esophagectomy patients, with an incidence of up to 50% [3–5]. This clinical entity is characterized by impaired motility of the gastric conduit, leading to nausea, vomiting, early satiety, and reflux symptoms [6, 7]. While the pathophysiology of DGCE is not completely understood, it is thought to be primarily related to division of the vagus nerve, the neural pathway necessary for effective gastric peristalsis and pyloric sphincter relaxation [5, 8]. The diagnostic criterion for DGCE has varied between institutions and in published literature. To address this issue, a more structured definition of DGCE was proposed at an international consensus meeting [9].
The true burden of DGCE remains uncertain, with considerable inconsistency in how this outcome is reported. This review aims to determine the incidence of DGCE and examine its variation across the different diagnostic definitions.
A comprehensive review of the literature was undertaken through to October 2024 to identify articles pertaining to the incidence of DGCE. The PubMed, Medline, Embase, Web of Science, and Cochrane Library electronic databases were systematically searched from inception. The following medical subject heading (MeSH) terms were used as a minimum in combination with Boolean operators and free-text terms, applied in multiple different “Delayed gastric conduit emptying”, “Gastric conduit dysfunction”, “Gastroparesis”, “Functional complications”, “Oesophageal cancer surgery”, and “Esophagectomy”. A complete search strategy for a single database defined with all keywords and subject headings is included (Electronic supplementary material, Table S1). The search strategy was supplemented by manually screening the references of relevant published studies. Screening of articles and their selection was performed independently by two authors (J.S., Q.C.). Disagreements at any point of this process were solved by consensus or by consulting a third reviewer.
Abstracts of all the retrieved studies were screened to determine need for full-text review. The detailed examination of remaining full-text articles was undertaken to ascertain their suitability for inclusion in the review based on whether they reported on delayed gastric conduit emptying. The studies were selected if they met the following inclusion (a) reported the incidence of delayed gastric conduit emptying; (b) original paper with independent data; (c) published as a full-text article in a peer-reviewed journal. No language restrictions were applied. The studies were excluded according to the following (a) data based on colonic or jejunal interposition; (b) redundant data duplicated or partially duplicated in selected publications of the same cohort; (b) data specific to pediatric patients; (c) incomplete data; (e) reports on animal models; and (f) case reports, editorial letters, reviews and conference abstracts.
Following a screen of abstracts to filter out studies outside of the review criteria, a detailed examination of full-text articles of the remaining studies was undertaken to ascertain their suitability for inclusion in the review. Data extraction was performed by two authors (J.S., Q.C.). The recorded data included author information, study design, study period, type of procedure, definition of DGCE, allocation into early or delayed DGCE, whether an intra-operative pyloric intervention was performed, type of gastric conduit and level of esophago-gastric anastomosis. The reporting of DGCE incidence was categorized into early DGCE if onset occurred during the index admission or within the first 14 days post-operatively, while all other cases were classified as late DGCE. In the absence of explicit temporal documentation, cases were conservatively designated as late DGCE by default. Incidence was reported for each outcome of interest, including comparisons between potential sources of heterogeneity–prophylactic pyloric intervention, conduit dimensions, and anastomosis height.
A comprehensive critical appraisal of the included studies was conducted to ensure methodological rigor and the reliability of findings. The Cochrane Risk of Bias 2 (RoB 2) tool was applied for randomized controlled trials (RCTs) to examine five domains that address key aspects of trial design, implementation, and outcome reporting [10]. The Risk of Bias in Non-Randomized Studies of Interventions (ROBINS-I) tool was employed for cohort and case–control designs to evaluate seven domains that systematically cover potential confounding factors, methodological limitations in participant selection, and deviations from intended interventional protocols [11]. Any disagreements between reviewers during the appraisal process were resolved through discussion and, where necessary, consultation with a third author to achieve consensus.
The incidence data were extracted as proportions, with the total number of patients developing DGCE as the numerator and the total number of patients undergoing esophagectomy as the denominator. Pooled incidence from included studies was calculated Freeman–Tukey transformation adjusting for single proportion. Pooled odds ratio (OR) was calculated using random-effects restricted maximum likelihood method. To compare rates of early and late DGCE depending on different anastomotic heights, conduit dimensions or whether prophylactic pyloric drainage was performed, two-sample Z test for proportions was used to compare two incidences, and Marascuilo’s procedure was used to compare multiple proportions. Inter-study heterogeneity was evaluated using both the I^2^ and H^2^ statistics. Tau-squared (τ^2^) values were also calculated to estimate between-study variance in a random effects model, with values closer to zero indicating less variability. Funnel plots and Egger’s tests were used to assess for publication bias. A 95% confidence interval (CI) was calculated to present the overall pooled estimate. Statistical significance was accepted at a p value of 0.05. Stata18.5 was used for statistical analysis.
This systematic review and meta-analysis was conducted in accordance with the Preferred Reporting Items for Systematic Reviews and Meta-Analyses (PRISMA) statement checklist and Assessing the Methodological Quality of Systematic Reviews (AMSTAR) guidelines [12, 13]. The review protocol was registered in the International Prospective Register of Systematic Reviews (PROSPERO) database (CRD42024577878) on 19 th August 2024 [14].
The systematic search yielded 5176 records from all databases, of which 529 duplicate entries and 28 ineligible trial registry records were removed prior to screening. Following title and abstract screening, 4010 records were excluded for irrelevance, leaving 609 full-text articles assessed for eligibility. Five additional studies were identified through reference searches. Of the 614 total articles evaluated, 125 studies met the inclusion criteria and were included in this review, with 51 studies focusing on early delayed gastric conduit emptying (DGCE) and 85 on late DGCE (Fig. 1).Fig. 1Flow diagram for systematic review.Source: Page MJ, et al. BMJ 2021;372:n71. https://doi.org/10.1136/bmj.n71. This work is licensed under CC BY 4.0. To view a copy of this license, visit
The included studies were published between 1990 and 2024, with the characteristics summarized in Table 1. The studies were predominantly retrospective cohort studies (n = 76), and the sample sizes ranged from 8 to 1294 patients, with variable follow-up periods from 30 days to 7 years among the 68 studies that reported this.Table 1Characteristics and outcomes of studies reviewedAuthorCountryStudy designStudy typeSurgical techniqueSample sizeFollow-up periodClassificationDefinition of early DGCEIncidence of early DGCEDefinition of late DGCEIncidence of late DGCEAbdelrahma 2024UKCohort studyRetrospectiveEsophagectomy; Ivor-Lewis100100% follow-up to 5 yearsEarly and lateModified Delphi process consensus20/100 (20%)Need for post-operative pyloric dilatation beyond 14 days (median 3 months; range 0–13 months)12/100 (12%)Antonoff 2014USACohort studyRetrospectiveEsophagectomy; transthoracic (56.0%), transhiatal (44.0%)29347.8% follow up to beyond 12 monthsLate––Need for post-operative pyloric dilatation long-term38/293 (13%)Babic 2022GermanyCohort studyProspectiveEsophagectomy; Ivor-Lewis816–EarlyModified Delphi process consensus226/816 (27.7%)––Bagheri 2014IranRCTProspectiveEsophagectomy; approach not specified60–EarlyRoutine upper gastrointestinal contrast swallow study at day 7 post-operatively15/60 (25.0%)––Benedix 2017GermanyCohort studyRetrospectiveEsophagectomy; transthoracic182100% Follow up to beyond 9 monthsEarlyAny of the (1) > 7 day requirement for NGT, (2) > 3 day use of erythromycin, (3) gastric conduit obstruction/dilatation on post-operative barium swallow or endoscopy71/182 (39.0%)––Bolger 2023CanadaCohort studyProspectiveEsophagectomy; approach not specified171–Early and lateBoth of the (1) subjective clinical evaluation of DGCE; (2) need for post-operative pyloric dilatation during index admission16/171 (9.4%)Both of the (1) subjective clinical evaluation of DGCE; (2) need for post-operative pyloric dilatation long-term11/171 (6.4%)Bolton 2013USACohort studyRetrospectiveEsophagectomy; approach not specified120–EarlySubjective clinical evaluation of DGCE—criteria not detailed20/108 (18.5%)––Boshier 2018UKCohort studyRetrospectiveEsophagectomy; two-stage (76.0%), three-stage (24.0%)100–EarlyRoutine upper gastrointestinal contrast swallow study at day 4–5 post-operatively35/100 (35.0%)––Brunner 2023GermanyCohort studyProspectiveEsophagectomy; Ivor-Lewis70100% Follow-up to 2 yearsLate––Modified Delphi process consensus19/70 (27.1%)Cerfolio 2009USACohort studyProspectiveEsophagectomy; Ivor-Lewis221Mean follow-up of 40 monthsEarlyRoutine upper gastrointestinal contrast swallow study at day 4 post-operatively186/221 (84.1%)––Chang 2017ChinaCohort studyRetrospectiveEsophagectomy; approach not specified1294–EarlySubjective clinical evaluation of DGCE—typical symptoms at day 7–14 post-operatively19/1294 (1.5%)––Chen 2013ChinaCohort studyRetrospectiveEsophagectomy; McKeown142Median follow-up of 26 months (Range 6–57 months)Late––Subjective clinical evaluation of DGCE—Criteria not detailed6/142 (4.2%)Chen 2021TaiwanCohort studyRetrospectiveEsophagectomy; Ivor Lewis (39.4%); McKeown (60.6%)526100% follow-up to 3 yearsLate––Subjective clinical evaluation of DGCE post-index admission—Criteria not detailed61/526 (11.6%)Datta 2014USACohort studyRetrospectiveEsophagectomy; approach not specified189100% follow-up to 6 monthsLate––Any of the (1) Post-operative clinical documentation, (2) Requiring pharmacologics for nausea/vomiting, (3) Abnormal gastric emptying scintigraphy with nausea/vomiting43/189 (22.8%)Deana 2021ItalyCohort studyRetrospectiveEsophagectomy; approach not specified110100% follow-up to 12 monthsLate––Esophagectomy Complications Consensus Group—Timeframe not detailed1/110 (0.9%)Decker 2020USACohort studyRetrospectiveEsophagectomy; approach not specified63–EarlyRoutine upper gastrointestinal contrast swallow study as inpatient prior to commencing oral intake16/63 (25.4%)––Deldycke 2015BelgiumCohort studyRetrospectiveEsophagectomy; Ivor-Lewis322100% follow-up to death date or study period completionEarlyRoutine upper gastrointestinal contrast swallow study at day 5 post-operatively119/322 (37.0%)––Deng 2010ChinaCohort studyProspectiveEsophagectomy; approach not specified78Median follow-up of 12 months (Range 8–60 months)EarlyRoutine scintigraphic meal study at day 14 post-operatively (DGCE = > 10% Retention of meal in stomach at 4 h post-ingestion)4/78 (5.1%)––Desprez 2020FranceCohort studyProspectiveEsophagectomy; approach not specified9100% Follow up to beyond 6 monthsLate––Pyloric distensibility < 10 mm2/mmHg based on EndoFLIP measurement beyond 6 months post-esophagectomy6/9 (66.7%)Ding 2021ChinaCohort studyRetrospectiveEsophagectomy; approach not specified25100% Follow up to 5 yearsLate––Criteria not detailed1/25 (4%)Djerf 2015SwedenCohort studyProspectiveEsophagectomy; Ivor-Lewis11100% Follow up to beyond 2 yearsLate––Radionuclide-labeled gastric scintigraphy at 2 years post-operatively0/11 (0.0%)Doran 2014UKCohort studyRetrospectiveEsophagectomy; Ivor-Lewis (100.0%)207–Late––Need for post-operative pyloric dilatation long-term13/207 (6.2%)Eldaif 2014USACohort studyProspectiveEsophagectomy; Ivor Lewis; McKeown; transhiatal322100% Follow up to beyond 6 monthsEarlyRoutine upper gastrointestinal contrast swallow study at day 5–7 post-operatively30/322 (12.4%)––Elliot 2017IrelandCohort studyProspectiveEsophagectomy; two-stage (69.2%); three-stage (23.1%); transhiatal (7.7%)1493% Follow up to beyond 3 monthsLate––Criteria not detailed0/14 (0%)Ericson 2013SwedenCohort studyRetrospectiveEsophagectomy; approach not specified195–Late––Need for post-operative pyloric dilatation long-term13/195 (6.7%)Feenstra 2023Nether-landsCohort studyProspectiveEsophagectomy; approach not specified65100% Follow up to 30 daysEarlyRoutine upper gastrointestinal contrast swallow study at day 3–4 post-operatively8/65 (12.3%)––Findlay 2015UKCohort studyRetrospectiveEsophagectomy; left thoracoabdominal (74.2%), Ivor Lewis (15.9%), three stage (2.3%), transhiatal (4.6%)132–Late––Need for post-operative pyloric dilatation13/132 (9.8%)Finley 1995CanadaCohort studyRetrospectiveEsophagectomy; Transhiatal (74.2%), Transthoracic (18.3%), thoracoabdominal (7.5%)29596% Follow up at 3 monthsEarlyRoutine upper gastrointestinal contrast swallow study at day 7 post-operatively16/295 (5.4%)––Fok 1991Hong KongRCTProspectiveEsophagectomy; Lewis Tanner (100%)200Mean follow-up of 17 monthsEarly and lateRoutine upper gastrointestinal contrast swallow study at day 10 post-operatively7/97 (7.2%)Subjective clinical evaluation of DGCE—Criteria not detailed13/200 (6.5%)Forshaw 2006UKCohort studyProspectiveEsophagectomy; left thoracoabdominal (100.0%)38–EarlyRoutine upper gastrointestinal contrast swallow study at day 6 post-operatively2/38 (5.3%)––Fransen 2023Nether-landsPropensity-matched analysisProspectiveEsophagectomy; transhiatal (78.7%), transthoracic (62.9%)1225100% Follow up at 3 monthsLate––Esophagectomy Complications Consensus Group—Timeframe not detailed19/1225 (1.6%)Frederick 2020USACohort studyRetrospectiveEsophagectomy; Two-stage (70.5%); Three-stage (21.5%); other (8.0%)149Mean follow-up of 26.9 months (Range 0–126.4 months)EarlyAny symptomatic, radiographic or endoscopic evidence of delayed gastric emptying by day 5 post-operatively114/149 (76.5%)––Fritz 2018GermanyCohort studyProspectiveEsophagectomy; Ivor Lewis (100%)170–EarlySubjective clinical evaluation of DGCE as inpatient28/170 (16.5%)––Fuchs 2016USACohort studyRetrospectiveEsophagectomy; Transhiatal (100%)41Mean follow-up of 12 months (Range 6–45 months)Late––Need for post-operative pyloric intervention8/41 (8.8%)Fujimoto 2022JapanCohort studyRetrospectiveEsophagectomy; McKeown (100%)33100% Follow up to beyond 18 monthsLate––Modified Delphi process consensus10/33 (30.3%)Fujita 2012JapanRCTProspectiveEsophagectomy; approach not specified154–EarlyRoutine upper gastrointestinal contrast swallow study at day 6 post-operatively13/154 (8.4%)––Gao 2020ChinaCohort studyProspectiveEsophagectomy; Ivor Lewis (100%)—MIE34EarlySelective upper gastrointestinal contrast swallow study during index admission1/34 (2.9%)––Giugliano 2017USACohort studyProspectiveEsophagectomy; Three-stage (77.4%), Ivor Lewis (22.6%)146100% Follow up to beyond 6 monthsLate––Need for post-operative pyloric intervention within 6 months38/146 (26.0%)Glatz 2017GermanyPropensity-matched analysisRetrospectiveEsophagectomy; approach not specified120–Late––Esophagectomy Complications Consensus Group—Timeframe not detailed20/120 (16.7%)Godazandeh 2008IranRCTProspectiveTranshiatal (80.0%), right thoracotomy (20.0%)30–Late––Radionuclide-labeled gastric scintigraphy at 2 months post-operatively20/30 (66.7%)Hadzijusufovic 2019GermanyCohort studyProspectiveEsophagectomy; Ivor Lewis (100%)115Median follow-up of 16 months (range 4–32 months)Early and lateNeed for post-operative pyloric dilatation before day 14 post-operatively11/115 (9.6%)Subjective clinical evaluation of DGCE—Criteria not detailed21/115 (18.3%)Hagens 2023Nether-landsCohort studyProspectiveEsophagectomy; approach not specified159-Late––Esophagectomy Complications Consensus Group—Timeframe within 30 days post-operatively4/159 (2.5%)Holscher 2007GermanyCohort study–Esophagectomy; approach not specified83Median follow-up of 7.2 monthsEarlyNeed for post-operative pyloric dilatation during index admission3/83 (3.6%)––Huang 2019ChinaCohort study–Esophagectomy; Ivor Lewis (100%)—MIE156–EarlySubjective clinical evaluation of DGCE within day 7 post-operatively49/156 (31.4%)––Johnson 2019AustraliaCohort studyProspectiveEsophagectomy; Two-stage (100%)62–Late––Criteria not detailed4/62 (6.5%)Kandagatla 2022USACohort studyRetrospectiveEsophagectomy; Ivor Lewis (100%)112–Late––Subjective clinical evaluation of DGCE at 12 months post-operatively3/112 (2.7%)Kao 1994TaiwanRCTProspectiveEsophagectomy; approach not specified38–Late––Radionuclide-labeled gastric scintigraphy—Timeframe not detailed18/38 (47.4%)Kent 2007USACohort studyRetrospectiveEsophagectomy; approach not specified12Median follow-up of 5.3 monthsEarly and lateRoutine upper gastrointestinal contrast swallow study within first week post-operatively0/12 (0%)Need for post-operative pyloric intervention4/12 (8.3%)Kim 2008South KoreaCohort studyProspectiveEsophagectomy; Transthoracic (67.0%); transhiatal (10.0%), transcervical three-field (24.0%)257Median follow-up of 26 monthsLate––Radionuclide-labeled gastric scintigraphy—Timeframe not detailed21/257 (8.2%)Klevebro 2023SwedenCohort studyProspectiveEsophagectomy; Ivor Lewis (57.1%), McKeown (31.3%), transhiatal (9.8%), left thoracoabdominal (1.8%)—open = 27, MIS = 85112–EarlyRoutine upper gastrointestinal contrast swallow study at day 2–3 post-operatively8/112 (7.1%)––Kuvendjiska 2023GermanyCohort studyRetrospectiveEsophagectomy; Ivor Lewis (100.0%)—open = 154; hybrid MIE = 152306Median follow-up of 21 monthsLate––Criteria not detailed30/306 (9.8%)Kuvendjiska 2019GermanyCohort studyRetrospectiveEsophagectomy; approach not specified157–EarlyRoutine upper gastrointestinal contrast swallow study at day 5 post-operatively27/157 (17.2%)––Lanuti 2011USACohort studyRetrospectiveEsophagectomy; Ivor Lewis (46.0%), Left thoracoabdominal (34.0%), Transhiatal 11.0%), Modified McKeown (3.0%), Minimally invasive (6.0%)436–EarlyRoutine upper gastrointestinal contrast swallow study at day 4–7 post-operatively76/436 (17.4%)––Lee 2005South KoreaCohort studyRetrospectiveEsophagectomy; approach not specified56Mean follow-up of 4.9 months (Range 1–30 months)Late––Radionuclide-labeled gastric scintigraphy—timeframe not detailed21/56 (37.5%)Lee 2000Hong KongCohort studyRetrospectiveEsophagectomy; Lewis Tanner (55.4%), three-stage (21.5%), esophago-gastrectomy (16.9%), transhiatal (6.2%)65Mean follow-up of 84.8 months (Range 10–178 months)EarlySubjective clinical evaluation of gastroparesis during index admission3/65 (4.6%)––Li 2020ChinaRCTProspectiveEsophagectomy; Three-field (50.0%), Two-field (50.0%)400100% Follow up to beyond 3 monthsLate––Criteria not detailed2/400 (0.5%)Li 2014ChinaCohort studyProspectiveEsophagectomy; left thoracotomy (52.5%), Ivor Lewis (32.0%), McKeown (11.2%), minimally invasive (4.2%)356Mean follow-up of 32.3 months (Range 3–66 months)Early and lateRoutine upper gastrointestinal contrast swallow study at day 5–7 post-operatively26/356 (7.3%)Clinical and endoscopic evaluation of DGCE 30 days post-esophagectomy—Criteria not detailed30/356 (8.4%)Li 2015ChinaPropensity-matched analysisRetrospectiveEsophagectomy; cervical-right thoracoabdominal (29.0%), right thoracic–abdominal esophagectomy (16.2%), left thoracic esophagectomy (54.8%)—open = 318, MIE = 89407Median follow-up of 27 months (Range 1–99 months)Late––Criteria not detailed7/407 (1.7%)Lin 2013ChinaCohort studyProspectiveEsophagectomy; approach not specified150100% Follow up to beyond 6 monthsLate––Criteria not detailed4/150 (2.7%)Liu 2022ChinaCohort studyRetrospectiveEsophagectomy; approach not specified251-Late––Criteria not detailed2/251 (0.8%)Liu 2019ChinaCohort studyRetrospectiveEsophagectomy; approach not specified68100% Follow up to 5 yearsLate––Criteria not detailed1/68 (1.5%)Luketich 1998USACohort study-Esophagectomy; Transhiatal (50%), Two-stage (50%)8-Late––Criteria not detailed1/8 (12.5%)Ma 2014ChinaCohort studyRetrospectiveEsophagectomy; Sweet (81.7%), Ivor Lewis (18.3%)915Median follow-up of 33.6 months (Range 24–72 months)EarlySelective upper gastrointestinal contrast swallow study or > 400 mL daily nasogastric tube output during index admission21/915 (23.0%)Mannell 1990South AfricaRCTProspectiveEsophagectomy; approach not specified40Median follow-up of 9 months (Range 2–26 months)Late––Subjective clinical evaluation of DGCE/gastric stasis—criteria not detailed10/40 (25.0%)Mantoan 2018ItalyCohort studyRetrospectiveEsophagectomy; Ivor Lewis, McKeown65100% Follow up to 3 yearsLate––Need for post-operative pyloric dilatation9/65 (13.8%)Marchese 2018UKCohort studyRetrospectiveEsophagectomy; Ivor Lewis (100%)90–Early and lateNeed for post-operative pyloric dilatation during index admission4/90 (4.4%)Need for post-operative pyloric intervention after discharge6/90 (6.7%)Margolis 2003USACohort studyRetrospectiveEsophagectomy; approach not specified80–Late––Criteria not detailed8/80 (10.0%)Martin 2009USACohort studyProspectiveEsophagectomy; approach not specified45–Early and lateRoutine upper gastrointestinal contrast swallow study at day 6 post-operatively4/43 (9.3%)Subjective clinical evaluation of DGCE at 3 months post-operatively3/45 (6.7%)Maus, 2016GermanyCohort studyRetrospectiveEsophagectomy; approach not specified403Study follow-up period of 12 months (Compliance N/S)Late––Need for post-operative pyloric dilatation60/403 (14.9%)Mehran, 2011USACohort studyProspectiveEsophagectomy; approach not specified—open = 44, MIE = 4488Mean follow-up of 15.2 months (Range 0.8–60.7 months)Late––Need for post-operative pyloric dilatation6/88 (6.8%)Mertens, 2021Nether-landsCohort studyProspectiveEsophagectomy; approach not specified239Median follow-up of 10 months (Range 6–27 months)EarlyAny of the following at 8–10 days post-operatively: (1) daily NGT output > 500 mL; (2) retention of > 300 mL after spigotting NGT for 4 h; (3) Conduit dilatation > 50% compared with index X-ray; (4) Replacement of NGT due to symptoms15/239 (6.3%)––Mohajeri, 2016IranRCTProspectiveEsophagectomy; approach not specified30–EarlyRoutine upper gastrointestinal contrast swallow study at day 7 post-operatively18/30 (60%)––Moons, 2021BelgiumPropensity-matched analysisProspectiveEsophagectomy; approach not specified—open = 284; MIE = 168; hybrid = 44496–Late––Esophagectomy Complications Consensus Group—timeframe not detailed104/496 (21.0%)Nafteux, 2011BelgiumCohort studyProspectiveEsophagectomy; Approach not specified; MIE = 16616686.5% Follow-up to 12 monthsLate––Gastric conduit dilatation on x-ray and need for either prokinetics or pyloric dilatation19/166 (11.4%)Nevins, 2020UKCohort studyProspectiveEsophagectomy; Ivor Lewis (75.3%), left thoracoabdominal (20.6%), McKeown (4.1%)97–Early and latePersistent daily NGT aspirates > 400 mLs from day 3 post-operatively29/97 (29.9%)Subjective clinical evaluation of DGCE and gastric conduit dilatation on x-ray at 1 month post-operatively13/97 (13.4%)Nevo, 2022CanadaCohort studyProspectiveEsophagectomy; Ivor Lewis, McKeown94–EarlyNeed for NGT re-insertion or inability to remove NGT due to symptoms of DGCE3/94 (31.9%)––Nguyen, 2010USACohort studyRetrospectiveEsophagectomy; approach not specified; MIE = 140140100% Follow-up to 3 monthsEarlyAny of the (1) > 7 Day requirement for NGT, (2) Gastric conduit dilatation on x-ray, (3) Clinical symptoms of DGCE7/140 (5.0%)––Nguyen, 2000USACohort study-Esophagectomy; two-stage (91.7%), transhiatal (8.3%)—MIE12Mean follow-up of 12.6 monthsLate––Need for post-operative pyloric intervention3/12 (25.0%)Nobel, 2019USACohort studyRetrospectiveEsophagectomy; approach not specified283–Early and latePersistent daily NGT aspirates > 300 mLs from day 3 post-operatively201/283 (71.0%)Need for post-operative pyloric dilatation at 3 months13/283 (4.6%)Noshiro, 2007JapanCohort studyRetrospectiveEsophagectomy; Approach not specified—MIE70Mean follow-up of 32 monthsLate––Criteria not detailed4/70 (5.7%)Oezcelik, 2011USACohort studyRetrospectiveEsophagectomy; En bloc transthoracic (66.9%), Transhiatal (30.0%), Minimally invasive (7.1%)277Median follow-up of 17 monthsEarlyRoutine upper gastrointestinal contrast swallow study at day 7 post-operatively33/266 (12.4%)Subjective clinical evaluation of DGCE—Criteria not detailed22/277 (7.9%)Palmes, 2007GermanyCohort studyRetrospectiveEsophagectomy; abdomino-thoracic (91.4%), Transhiatal (6.6%), cervico-abdomino-thoracic (2.0%)175Study follow-up period of 12 months (Compliance N/S)EarlyRoutine upper gastrointestinal contrast swallow study at day 4 post-operatively57/175 (32.6%)––Park, 2019South KoreaCohort studyRetrospectiveEsophagectomy; approach not specified291–Late––DGCE resulting in readmission to hospital—Criteria not detailed3/291 (1.0%)Perry, 2009USACohort studyProspectiveEsophagectomy; thoracoscopic-laparoscopic (42.9%), Transhiatal (57.1%)70–Late––Criteria not detailed4/70 (5.7%)Pines, 2011IsraelCohort studyProspectiveEsophagectomy; transhiatal (100.0%)100Median follow-up of 19.5 monthsLate––Criteria not detailed13/100 (13.0%)Predescu, 2018RomaniaCohort studyRetrospectiveEsophagectomy; approach not specified8293.9% Follow-up to 12 monthsLate––Conduit neuro-motor dysfunction—Criteria not detailed6/82 (7.3%)Prokakis, 2021GreeceCohort studyRetrospectiveEsophagectomy; Approach not specified49–Late––Criteria not detailed4/49 (8.2%)Puccetti, 2022USACohort studyProspectiveEsophagectomy; Ivor Lewis (48.8%), left thoracoabdominal (46.0%), McKeown (9.3%)43–EarlyRoutine upper gastrointestinal contrast swallow study at day 2–3 post-operatively7/43 (16.3%)––Rasmussen, 2021DenmarkCohort studyRetrospectiveEsophagectomy; Ivor Lewis (100.0%)120Median follow-up of 27 monthsLate––Need for post-operative pyloric dilatation63/120 (52.5%)Reinstaller, 2022GermanyCohort studyRetrospectiveEsophagectomy; approach not specified137–Late––Criteria not detailed69/137 (50.4%)Reyhani, 2020UKCohort studyProspectiveEsophagectomy; left thoracoabdominal (100%)—MIE74–Late––Need for post-operative pyloric dilatation at 6 months23/74 (31.1%)Rong, 2022ChinaPropensity-matched analysisRetrospectiveEsophagectomy; approach not specified—open = 138; MIE = 132270–Late––Post-esophagectomy gastroparesis—Criteria not detailed3/168 (1.8%)Saeed, 2024USAPropensity-matched analysisRetrospectiveEsophagectomy; Ivor Lewis (100%)—robotic, hybrid475–Late––Both of the (1) Presence of one or more DGE-related symptoms; (2) Delayed contrast passage on a barium swallow study58/475 (12.2%)Sarkaria, 2019USACohort studyProspectiveEsophagectomy; Ivor Lewis, Thoracoabdominal, McKeown—open = 106, robotic = 65170–Late––Post-esophagectomy gastroparesis—Criteria not detailed1/170 (0.6%)Schuchert, 2004USACohort studyRetrospectiveEsophagectomy; approach not specified222–Late––Criteria not detailed4/222 (1.8%)Senkowski, 2006USACohort studyProspectiveEsophagectomy; approach not specified—MIE20–EarlyRoutine upper gastrointestinal contrast swallow study at day 4–5 post-operatively2/20 (10.0%)––Shen, 2022ChinaRCTProspectiveEsophagectomy; approach not specified—MIE118–Late––Esophagectomy Complications Consensus Group—Timeframe not detailed1/118 (0.8%)Shi, 2021ChinaCohort studyRetrospectiveEsophagectomy; Ivor Lewis (50.0%), McKeown (50.0%)136Median follow-up of 23.5 months (Range 6–36 months)Late––Criteria not detailed2/136 (1.5%)Skancke, 2017USACohort studyRetrospectiveEsophagectomy; Ivor Lewis (55.6%), transhiatal (44.4%)—open = 18, MIE = 927Study follow-up period of 12 months (Compliance N/S)Late––Criteria not detailed1/27 (3.7%)Sokouti, 2015IranRCTProspectiveEsophagectomy; Transhiatal (100.0%)51100% Follow up to 12 monthsLate––Radionuclide-labeled gastric scintigraphy at 2 months19/51 (37.3%)Stewart, 2017USACohort studyProspectiveEsophagectomy; Ivor Lewis (96.2%), transhiatal (1.4%), McKeown (1.4%)—MIE71100% Follow up to 3 monthsEarlyBoth of the (1) Symptoms of DGCE; (2) Features of DGCE on esophagram or upper endoscopy5/71 (7.0%)––Sun, 2014ChinaCohort studyProspectiveEsophagectomy; approach not specified—MIE68–EarlyBoth of the (1) Inability to tolerate soft diet by day 10 post-operative; (2) Prolonged gastric emptying time > 25% compared with preoperative radionuclide-labeled gastric scintigraphy0/68 (0%)––Sun, 2019ChinaRCTProspectiveEsophagectomy; McKeown (100.0%)—MIE86–EarlyPresence of DGCE during index admission—Criteria not detailed1/86 (1.2%)––Sutcliffe, 2008UKCohort studyRetrospectiveEsophagectomy; transhiatal (58.8%), mckeown (8.5%), left thoraco-abdominal (22.0%), Ivor Lewis (10.7%)177Median follow-up of 37 months (Range 1–77 months)Late––Both of the (1) Symptoms of DGCE; (2) Endoscopic or radiological evidence of gastric distension and pyloric stenosis21/177 (11.9%)SwansonUSACohort studyRetrospectiveEsophagectomy; transhiatal (80.0%) thoracoabdominal (12.0%), VATS 3-Stage (8.0%)25Median follow-up of 22 months (Range 1–84 months)Early and lateRoutine upper gastrointestinal contrast swallow study at week 1–2 post-operatively3/25 (12.0%)Need for post-operative pyloric dilatation1/20 (5.0%)Takahashi, 2022USACohort studyRetrospectiveEsophagectomy; Ivor-Lewis (57.1%); McKeown (34.6%); transhiatal (8.3%)—MIE254–EarlyBoth of the (1) Symptoms of DGCE within 2 weeks post-operatively; (2) Persistently high nasogastric output or increased gastric conduit dilatation with air-fluid level on x-ray61/254 (24.0%)––Tang, 2022ChinaPropensity-matched analysisRetrospectiveEsophagectomy; McKeown (37.0%); sweet (56.8%); Ivor Lewis (6.2%)—open = 311; MIE = 175486–Late––Subjective clinical evaluation of gastroparesis—Criteria not detailed12/486 (2.5%)Tapias, 2013USACohort studyRetrospectiveEsophagectomy; Ivor Lewis (45.6%), minimally invasive (8.0%), thoracoabdominal (32.1%), transhiatal (10.3%), modified McKeown (3.8%)474Mean follow-up of 43 monthsLate––Criteria not detailed35/474 (7.4%)Tham, 2019UKCohort studyProspectiveEsophagectomy; Ivor Lewis (100.0%)228–EarlyPersistent daily NGT aspirates from day 5 post-operatively (Inability to remove NGT)40/228 (17.5%)––Tham, 2022UKCohort study-Esophagectomy; Ivor Lewis (100.0%)65100% Follow-up to 6 weeksEarlyPersistent daily NGT aspirates from day 5 post-operatively (Inability to remove NGT)24/65 (36.9%)––Uzun, 2024GermanyCohort studyRetrospectiveEsophagectomy; Ivor Lewis (100.0%)—MIE439–Late––Criteria not detailed52/439 (11.8%)Van der Sluis, 2022GermanyCohort studyProspectiveEsophagectomy; Ivor Lewis (86.5%), McKeown (13.5%)—open = 107, hybrid = 101, MIE = 91, RAMIE = 123422–Late––Esophagectomy Complications Consensus Group—Timeframe not detailed43/422 (10.2%)Velanovich, 2003USACohort study-Esophagectomy; approach not specified58Mean follow-up of 15 months (Range 1–60 months)Late––Subjective clinical evaluation of DGCE—Criteria not detailed11/58 (19.0%)Wang, 2021ChinaCohort studyRetrospectiveEsophagectomy; McKeown (100.0%)—MIE108–Late––Criteria not detailed1/108 (0.9%)Wang, 2021ChinaCohort studyProspectiveEsophagectomy; McKeown (100.0%)—MIE192100% Follow-up to beyond 3 monthsLate––Criteria not detailed2/192 (1.0%)Wu, 2022ChinaCohort studyRetrospectiveEsophagectomy; modified McKeown (100.0%)45–Late––Esophagectomy Complications Consensus Group—Timeframe not detailed0/45 (0.0%)Xu, 2023ChinaCohort studyRetrospectiveEsophagectomy; McKeown (100.0%)—Robotic211–Late––Criteria not detailed6/211 (2.8%)Yajima, 2009JapanCohort studyRetrospectiveEsophagectomy; transthoracic (66.0%), transhiatal (34.0%)141Median follow-up of 60 months (Range 18–204 months)Late––Endoscopic examination for presence of chyme in gastric conduit—Timeframe not detailed26/141 (18.4%)Yetasook, 2013USACohort studyRetrospectiveEsophagectomy; Ivor Lewis (100.0%)—Hybrid24Median follow-up of 10 months (Range 5–15 months)EarlyRoutine upper gastrointestinal contrast swallow study at day 3 post-operatively6/24 (25.0%)––Zhang, 2017ChinaCohort studyRetrospectiveEsophagectomy; approach not specified—open = 250, MIE = 35285100.0% Follow-up to beyond 3 monthsLate––Both of the (1) Symptoms of DGCE; (2) Endoscopic or radiological evidence of gastric distension and pyloric stenosis52/285 (18.2%)Zhang, 2019ChinaRCTProspectiveEsophagectomy; approach not specified10499.0% follow-up to 5 yearsLate––Criteria not detailed6/104 (5.8%)Zhang, 2022ChinaPropensity-matched analysisRetrospectiveEsophagectomy; Ivor Lewis, McKeown, thoracic-cervical dual-incision, left thoracic esophagectomy256Median follow-up of 47.5 months (Range 3–139 months)Late––Subjective clinical evaluation of gastroparesis during index admission3/256 (1.2%)Zhang, 2017ChinaCohort studyRetrospectiveEsophagectomy; Ivor Lewis (100.0%)185Mean follow-up of 38.5 months (Range 6–55 months)Late––Criteria not detailed5/185 (2.7%)Zhao, 2017ChinaCohort studyRetrospectiveEsophagectomy; approach not specified—MIE273Median follow-up of 35 monthsLate––Criteria not detailed7/273 (2.6%)Zhou, 2009ChinaCohort studyRetrospectiveEsophagectomy; modified McKeown (100.0%)—hybrid30–Late––Criteria not detailed1/30 (3.3%)
The methodological quality of the included studies varied. RCTs assessed using the RoB 2 tool predominantly demonstrated moderate methodological quality, with consistent limitations including inadequate blinding and incomplete follow-up reporting (Electronic Supplementary Table S2). The ROBINS-I tool identified that 77.9% of cohort and case–control studies were low-risk of bias, with a subset exhibiting heightened risk of bias stemming from confounding factors (Electronic Supplementary Table S3). As illustrated in Electronic Supplementary Fig. 1, no publication bias was detected for early DGCE incidence (Egger’s test, p = 0.46), whereas asymmetry was identified for late DGCE incidence (Egger’s test, p = 0.0012). Although marked heterogeneity may give rise to such asymmetry, publication bias is a possibility. While substantial heterogeneity may account for this asymmetry, it may also indicate potential publication bias.
The pooled incidence of early DGCE was 15.9% (95% CI 11.0–21.0%), with significant heterogeneity across studies (I^2^ = 97.64%, p < 0.001). Definitions of early DGCE varied considerably, with some studies basing their diagnostic criteria solely on radiographic swallow study findings (n = 26), nasogastric output volume (n = 5), or otherwise adopting Konradsson’s international consensus (n = 2). The remaining studies employed either another institution-specific criterion (n = 12), as disclosed in Table 1, or did not specify their definition (n = 6). Marked heterogeneity persisted within each individual definition group, as evidenced by elevated τ^2^ values across all definitional categories. The sole exception was Konradsson’s international consensus (τ^2^ = 0.01), though only two studies were included in this group (Fig. 2).Fig. 2Pooled incidence of early DGCE, categorized by definition
With regards to late DGCE, the pooled incidence was 9.4% (95% CI 7.1–11.9%), also demonstrating high heterogeneity (I^2^ = 95.59%, p < 0.001). Studies again were wide-ranging in their definition of late DGCE, such as the need for post-operative pyloric intervention to treat delayed emptying symptoms (n = 17), assessment via gastric scintigraphy (n = 6), the Esophagectomy Complications Consensus Group (ECCG) definition (n = 8) [15], or the Konradsson’s international consensus (n = 2) [16]. Additional studies employed an independent institution-specific criterion (n = 10), while 43 studies did not specify their definition for DGCE. Despite the high overall heterogeneity observed in late DGCE incidence, the analysis within individual definition subgroups demonstrated predominantly low τ^2^ values (< 0.1), with the exception of diagnostic criteria based on gastric scintigraphy or post-operative pyloric intervention. This indicates relatively strong agreement in reported incidence rates when consistent diagnostic criteria are applied.
An additional subgroup analysis exploring potential sources of heterogeneity demonstrated that prophylactic pyloric drainage (PPD) was not associated with a statistically significant reduction in early DGCE (OR 0.76; p = 0.38) or late DGCE (OR 0.71; p = 0.44) (Electronic Supplementary FigS. 3 and 4). This analysis, however, exhibited persistently high I^2^ values, indicating considerable residual heterogeneity. Thoracic anastomosis and whole-stomach conduits were associated with marginally higher rates of DGCE, though this association did also not reach statistical significance (Electronic Supplementary Figs. 5–8). Interpretation of these findings was further constrained by the limited number of studies reporting DGCE outcomes stratified by anastomotic height and conduit dimensions (Fig. 3).Fig. 3Pooled incidence of late DGCE, categorized by definition
This systematic review and meta-analysis provide a comprehensive evaluation of DGCE following esophagectomy, a complication with significant implications for postoperative recovery and quality of life. The pooled incidences of early and late DGCE were 15.9% and 9.4%, respectively, with high heterogeneity observed across studies. These findings highlight the persistent burden of DGCE despite advancements in surgical techniques and perioperative care.
The variability in DGCE definitions across studies underscores a critical challenge in standardizing outcome reporting. Definitions ranged from subjective clinical evaluations to internationally endorsed criteria or standalone diagnostic investigations. Efforts to address this inconsistency were initially spearheaded by the Esophagectomy Complications Consensus Group (ECCG), which brought together 21 high-volume surgeons from 14 countries to develop a set of Delphi consensus definitions for esophagectomy-related complications. DGCE was characterized by the need for intervention or delayed discharge [15]. Konradsson et al. sought to further refine this definition using a modified Delphi process involving a consortium of 33 global experts [16]. This process incorporated both clinical symptoms and radiologic findings to establish diagnostic criteria for early and late DGCE (Table 2), while also introducing a symptom severity grading scale for late DGCE. This framework holds significant promise for improving diagnostic consistency and benchmarking DGCE management. Although only four of the 43 studies published after the publication of Konradsson’s consensus adhered to these criteria [17–20], the limited uptake likely stems from the early stage of adoption of this definition. This gradual dissemination process highlights the inherent challenges of integrating new definitions into established clinical systems and the inconsistent awareness or training among healthcare providers [21]. As survivorship outcomes continue to improve and awareness of this functional issue grows, the adoption of these criteria is anticipated to expand, promoting greater consistency, and standardization in clinical practice.Table 2International consensus statement on DGCE(adapted from Konradsson et al.)Diagnostic criteria for early DGCEOne of the following criteria should be > 500 mL Diurnal nasogastric tube output measured on the morning of postoperative day five or later (but within 14 days of surgery); OR > 100% Increased gastric tube width on front chest X-ray projection (in comparison to baseline chest X-ray taken on the day of surgery) together with the presence of an air-fluid levelDiagnostic criteria for late DGCEBoth of the following criteria should be The patient should have “quite a bit” or “very much” of at least two of the following Early satiety/fullness Vomiting Nausea Regurgitation Inability to meet caloric need by oral intake Delayed contrast passage on upper gastrointestinal water-soluble contrast radiogram or on timed barium swallow (relying on expert radiologist)
Despite the historical prevalence of contrast swallow studies in the post-operative evaluation after esophagectomy, their notable exclusion from Konradsson’s consensus definition reflects an evidence-based shift in clinical practice [22, 23]. While our review found that over half of the analyzed studies relied on contrast swallow studies as their primary diagnostic tool for early DGCE, mounting evidence suggests this approach has significant limitations. The poor sensitivity and specificity of contrast studies for assessing gastric conduit function and transpyloric emptying, combined with their resource intensity and potential for inconsistent interpretation across institutions, has led many specialized upper gastrointestinal units to move away from their routine use [24–26]. This trend aligns with current ERAS guidelines for esophagectomy [27], which do not mandate contrast swallow studies. The consensus definition’s focus on more objective measures, such as nasogastric tube output and clinical parameters, provides a more robust and practical approach to diagnosing early DGCE [16].
The implementation of standardized diagnostic criteria for DGCE has far-reaching implications for both research and clinical practice. As supported in our analysis, the studies employing explicit diagnostic definitions reported higher incidence rates of DGCE compared to those without defined criteria. This disparity suggests that centers not held accountable to a specific definition are more prone to observer bias, leading to underreporting of DGCE cases. Standardization would not only improve diagnostic accuracy but would also enable meaningful comparisons across institutions and international borders, fostering more robust collaborative research efforts. From a clinical perspective, a uniform criterion also enables early and accurate diagnosis of DGCE, allowing for prompt implementation of tailored interventions to reduce symptom burden and enhance postoperative quality of life.
The subgroup analysis revealed insights into potential modulators of DGCE incidence. Intraoperative pyloric intervention did not confer a statistically significant protective effect against DGCE. Some investigators attribute its suboptimal efficacy to the perioperative localized tissue edema at the pylorus associated with its manipulation [28]. In line with this, the evidence supporting prophylactic pyloric drainage remains equivocal. Several meta-analyses have failed to establish a clear benefit in outcomes from draining the pylorus during esophagectomy [29, 30], with the additional risk of procedural complications introducing substantial uncertainty into the intervention's risk–benefit calculus. Notably, the present subgroup analyses examining prophylactic pyloric drainage continued to demonstrate substantial residual heterogeneity, with persistently high I^2^ values, suggesting that the true sources of heterogeneity could not be fully elucidated despite stratification. The limited efficacy of prophylactic pyloric drainage reflects the nuanced pathophysiology of DGCE, where pyloric dysfunction is significant but not the sole factor determining gastric emptying, particularly given the equally compromised motility of the denervated gastric conduit. Despite the prevailing clinical hypotheses that a more proximal anastomosis and narrow conduit lead to less anatomical redundancy and improved gastric conduit emptying [29, 31–33], neither of these potential sources of heterogeneity demonstrated a statistically significant association. The absence of these findings is in part due to the paucity of studies that compared DGCE incidence data by anastomotic height or conduit diameter. The nuanced interplay between conduit anatomy and DGCE is clearly appreciated by esophagogastric surgeons, however, as more than 80% report constructing a conduit diameter less than 5 cm, and 90% indicate positioning the anastomosis above the azygos vein [34].
A key strength of this study is it allows for more meaningful comparisons across the various definitions of DGCE, spanning multiple institutions and several decades, contributing to a broad and representative understanding of the DGCE burden. Several limitations should also be considered in interpreting this review. The high heterogeneity observed across the included research is a notable drawback, reflecting the variations between studies in both study design and diagnostic criteria for DGCE. Another limitation is that relying on definitions dependent on postoperative pyloric drainage as a surrogate for DGCE may exclude patients who were otherwise managed pharmacologically, potentially underestimating the true incidence of this problem. While the subgroup analysis of anastomosis location may have been constrained by limited study data, potential reporting bias should be considered in its classification. Cervical anastomoses were designated when explicitly documented by authors or when unambiguously described in operative technique details. Given relatively few studies stratified DGCE incidence by anastomotic height, this represents a potential methodological limitation with low statistical power.
This review demonstrates the significant burden of DGCE following esophagectomy, with considerable variability in reported incidence rates primarily stemming from inconsistent diagnostic definitions. The implementation of standardized criteria is crucial to reduce heterogeneity and mitigate underreporting. Integrating the recently proposed international consensus definition into clinical practice is a step toward enhancing diagnostic accuracy, ultimately improving the identification and management of DGCE.
Below is the link to the electronic supplementary material.Supplementary file1 (DOCX 2461 KB)