Applicability and results of the consensus definition of delayed gastric conduit emptying after esophagectomy

Abstract

Purpose

Delayed gastric conduit emptying (DGCE) affects 15%–60% of patients after esophagectomy, depending on the definition used. This study evaluated DGCE incidence using the new consensus criteria and explored its clinical and scientific relevance. It also analyzed nasogastric tube (NGT) output to estimate a safe removal threshold.

Methods

This retrospective single-center cohort study included patients who underwent elective Ivor Lewis esophagectomy (2020–2024).

Results

A total of 160 patients were included. The majority of patients underwent robot-assisted esophagectomy (n = 65 [40.6%]), whereas open esophagectomy was performed in 38 patients (23.7%). Early DGCE was identified in 40.6% of patients, predominantly via NGT output (95.4%), with few cases diagnosed radiologically (4.6%). DGCE was associated with longer hospital stays (24 vs. 18 days, p = 0.034) and higher NGT reinsertion rates (47.0% vs. 28.7%, p = 0.02), but not with differences in surgical approach or major complications (Clavien–Dindo Classification > IIIa: 18.2% vs. 13.8%, p = 0.72). Late DGCE was diagnosed in only 8 patients (5.0%), mainly because of the low use of routine radiological imaging. In the cohort of patients who did not require NGT reinsertion, the median of the average nasogastric tube output for the 2 days before removal was 200 mL per day.

Conclusions

The consensus definition of DGCE is applicable and identifies a clinically relevant incidence of over 40%. An average daily NGT output of 200 mL appears to reliably predict safe tube removal.

Introduction

Delayed gastric conduit emptying (DGCE) is among the most common complications following esophagectomy. The reported incidence rates vary widely, ranging from 15% to 60% [1, 2]. This complication significantly impacts the quality of life of patients, both in the early postoperative period by prolonging hospital stay and in the long term, as symptoms may persist for months after discharge [3]. In some cases, severe reflux symptoms are the primary barrier to hospital discharge, leading to notable economic and logistical burdens for healthcare institutions.

The early diagnosis of DGCE has historically been challenging due to the lack of standardized criteria. However, an expert consensus was published in 2020, formally defining DGCE and classifying it into early and late forms [4].

The role of nasogastric tube (NGT) management in the postoperative period has been widely investigated, with conflicting results. Evidence suggests that early removal of the NGT does not increase the risk of anastomotic leakage or postoperative pneumonia, as reflected in ERAS (Enhanced Recovery After Surgery) recommendations [5]. However, a recent Scandinavian randomized controlled trial demonstrated that the incidence of anastomotic leakage was higher in the no-NGT group than in the NGT group [6]. Nevertheless, prolonged NGT use remains common in clinical practice and may play an important role in the early detection of DGCE.

This study has two primary objectives. The first step is to evaluate the applicability of the new DGCE criteria and determine the incidence of DGCE following esophagectomy at a single tertiary center. Second, we analyzed the associated NGT management strategies, with a particular focus on determining the safe and evidence-based criteria for tube removal based on the NGT volume.

Material and methods

Study design

This single-center retrospective study was conducted at the Department of General, Visceral and Thoracic Surgery, University Medical Center Hamburg-Eppendorf, Hamburg, Germany. This study was conducted in accordance with the principles of the Declaration of Helsinki and was approved by the local ethics committee (approval number: #2024−101322-BO-ff). Written informed consent was obtained from all participants. This study included all patients who underwent elective esophagectomy using the Ivor Lewis procedure for oncological or benign esophageal diseases from 2020 to 2024.

Definition of DGCE

We applied the recently published consensus definition of DGCE [4]. Early DGCE was defined as > 500 mL NGT output per day measured on the morning of postoperative day (POD) 5 or later (but within 14 days of surgery) or increased gastric tube width of > 100% on frontal chest X-ray projection (in comparison to baseline chest X-ray taken on the day of operation) plus the presence of an air-fluid level.

Late DGCE was defined as the occurrence of at least two of the following symptoms: early satiety or fullness, vomiting, nausea, regurgitation, inability to meet caloric needs by oral intake, and delayed contrast passage on upper gastrointestinal imaging. Symptoms were graded from 0 to 3 (“not at all” to “very much”). A modified version of the consensus diagnostic criteria [4] was applied, with omission of standardized symptom questionnaires due to their lack of routine implementation during the study period.

Data acquisition

Data on the daily volume of fluid drained via NGT were recorded for up to 30 postoperative days. The postoperative day on which the tube was removed was additionally documented for patients whose NGT was removed later. The lack of data reporting NGT output beyond the fifth POD was a reason for excluding a patient from the DGCE incidence analysis. Furthermore, patients with an NGT output not documented for 2 days prior to NGT removal were also excluded from the NGT output analysis. Two metrics were created for a more accurate determination of the NGT volumes at the time of the final removal. The first metric shows the volume of NGT on the day of removal or the day before. The second metric shows the average value over the last 2 days. This allowed for balancing the values in cases of large variations in the amount of output over 2 days.

One of the goals of this study was to analyze NGT management. Therefore, the patients were divided into two subgroups: those whose NGT was removed without reinsertion and those who required NGT reinsertion. In these two subgroups, the reasons for NGT removal and the reasons for reinsertion were analyzed separately.

A cohort of patients was selected to determine the NGT output volume considered safe for removal based on the following criteria: (1) the nasogastric tube (NGT) was removed without subsequent need for reinsertion, and (2) documented NGT output was available for the two days preceding removal. The average output was calculated over these two days.

The follow-up period was primarily limited to the postoperative hospital stay, during which patients were monitored for clinical symptoms and NGT-related parameters. Late DGCE after discharge was identified only in patients readmitted with relevant symptoms.

Statistical analysis

Data were analyzed using R with a user-friendly statistical interface (EZR) [7]. Descriptive statistics are presented as a mean ± standard deviation for normally distributed data and the median (interquartile range [IQR]) for non-normally distributed data. The Fisher Exact test, Mann–Whitney U test, and χ² test were used. A difference was considered significant if p < 0.05.

Results

A total of 160 patients were included in the study (Table 1). Most patients were men (n = 123; 76.9%). The median age was 63 years (IQR 57–71). 97% of the patients (n = 156) presented with oncological diagnoses. A total of 100 patients (62.5%) received neoadjuvant treatment: 62.5% chemotherapy (n = 100) and 16.9% chemoradiation (n = 27).

Table 1.

Patient characteristics

Subcategory	Value
Total cohort [n (%)]	160 (100%)
Male [n (%)]	123 (76.9)
Female [n (%)]	37 (23.1)
Age [median (IQR)]	63 (57–71)
BMI [median (IQR)]	25.2 (22.5–27.8)
ASA Score [n (%)]
ASA 1	1 (0.6)
ASA 2	63 (39.4)
ASA 3	93 (58.1)
ASA 4	3 (1.9)
Pathology [n (%)]
Adenocarcinoma	120 (75.0)
Squamous cell carcinoma	36 (22.5)
Other	4 (2.5)
Neoadjuvant treatment [n (%)]
Chemotherapy	100 (62.5)
Chemoradiation	27 (16.9)

Abbreviations: IQR interquartile range, BMI body mass index, ASA American Society of Anesthesiologists

Of the 160 patients, 122 underwent minimally invasive esophagectomy: robot-assisted (RAMIE) in 65 cases (40.6%), solely laparoscopic/thoracoscopic esophagectomy in 30 cases (18.7%), and hybrid esophagectomy in 27 cases (16.9%). An open surgical approach was performed in 38 cases (23.7%). Significant postoperative morbidity (Clavien–Dindo classification [CDC] grade ≥ IIIa) was present in 18.7% of the cases (Table 2). The overall 90-day mortality rate was 5.6%. The mean duration of hospital stay was 20 days (IQR 15–37), and 56 patients (35.0%) required readmission.

Table 2.

Surgical and postoperative characteristics

Subcategory	Value
Total cohort [n (%)]	160 (100%)
Robotic [n (%)]	65 (40.6)
Laparoscopic [n (%)]	30 (18.7)
Hybrid[n (%)]	27 (16.9)
Open [n (%)]	38 (23.7)
Anastomotic Leakage [n (%)]	44 (27.5)
Local defect requiring no therapy change or treated medically or with dietary modification	1 (2.27)
Localized defect requiring intervention but not surgical	36 (81.8)
Localized defect requiring surgical intervention	7 (15.9)
Pneumonia [n (%)]	55 (34.4)
Aspiration [n (%)]	6 (3.7)
Clavien–Dindo Classification [n (%)]
0	6 (3.7)
I	34 (21.2)
II	31 (19.4)
III a	30 (18.7)
III b	25 (15.6)
IV a	22 (13.7)
IV b	3 (1.9)
V	9 (5.6)
Readmission [n (%)]	56 (35)
Length of Stay, days [median (IQR)]	20 (15–37)

Abbreviations: IQR interquartile range

Following the recently published consensus definition of DGCE, 65 patients (40.6%) were diagnosed with early DGCE (within the first 14 days), whereas only 8 patients (5.0%) developed late DGCE (on or after POD 15) (Table 3). Both forms of DGCE were observed in 7 patients (4.4%). However, patients with DGCE had a significantly longer hospital stay duration of 24 days (IQR 16–39) versus 18 days (IQR 15–36) compared with patients without DGCE (p = 0.034). The median POD for NGT removal was POD 7 (95% CI: 6–9) in patients without and POD 14 (95% CI 9–19) in patients with DGCE (p < 0.001). Furthermore, the need for NGT reinsertion was significantly higher (p = 0.02) in DGCE patients (31 reinsertions, 47.0%) compared with those without DGCE (27 reinsertions, 28.7%). DGCE occurred across all surgical approaches without a statistically significant difference between these subgroups.No significant difference was found in the number or severity of complications between patients with and without DGCE. Specific complications, including anastomotic leakage and pulmonary complications, occurred at comparable rates in both groups.

Table 3.

Comparison of patient and perioperative data in patients with and without delayed gastric conduit emptying

Subcategory	No DGCE	Early DGCE	p	Late DGCE	p
Total cohort [n (%)]	94 (58.75)	65 (40.62)		8 (5)
Age [median (IQR)]	62 (57–69)	66 (58–73)	0.07	60 (53–64)	0.29
ASA [n (%)]
I	1 (1.1)	0 (0.0)	0.97	0 (0.0)	1.00
II	38 (40.4)	25 (38.5)		3 (37.5)
III	53 (56.4)	39 (60.0)		5 (62.5)
IV	2 (2.1)	1 (1.5)		0 (0.0)
Hospital stay [median (IQR)]	18 (15–36)	24 (16–39)	0.04	56 (27–75)	0.022
Final NGT removal, POD [median (IQR)] POD [median (IQR)]	7 (5–17)	14 (8–26)	< 0.001	28 (20–41)	0.01
NGT reinsertion
No	67 (71.3)	35 (53.8)	0.044	3 (37.5)	0.14
Yes	27 (28.7)	30 (46.2)		5 (62.5)
NGT output 1, mL [median (IQR)]	0 (0–300)	300 (0–500)	0.001	250 (0–625)	0.66
NGT output 2, mL [median (IQR)]	100 (0–229)	300 (188–525)	< 0.001	75 (38–388)	0.75
DaVinci	40 (42.6)	25 (38.5)	0.86	1 (12.5)	0.20
Hybrid	14 (14.9)	13 (20.0)		2 (25.0)
Open	23 (24.5)	15 (23.1)		2 (25.0)
Laparoscopic	17 (18.1)	12 (18.5)		3 (37.5)
Anastomotic leak
No	70 (74.5)	45 (69.2)	0.47	5 (62.5)	0.68
Yes	24 (25.5)	20 (30.8)		3 (37.5)
Pneumonia
No	61 (64.9)	43 (66.2)	1.00	5 (62.5)	1.00
Yes	33 (35.1)	22 (33.8)		3 (37.5)

Abbreviations: NGT nasogastric tube, IQR interquartile range, ASA American Society of Anesthesiologists, POD postoperative day, NGT output 1 the NGT output on the day of removal or the day before, NGT output 2 the average NGT output over the last 2 days prior to removal

The median NGT volume in the 2 days before removal was significantly higher in patients with DGCE. In patients without DGCE, it was 100 mL (IQR 0–229), whereas in those with DGCE, it was 300 mL (IQR 188–525) (p < 0.001).

Among patients whose NGT was removed without reinsertion (n = 102), 14 were excluded from the analysis because of incomplete data. Two complementary threshold analyses were performed (Table 4). To reduce the influence of day-to-day variability, the second analysis included 55 patients with available NGT output data for both of the two days preceding removal. The median two-day average daily NGT output was 200 mL (IQR, 62.5–375 mL).

Table 4.

Descriptive statistics of NGT output in the cohort of patients without NGT reinsertion

Subcategory/Percentile	0%	25%	50%	75%	100%	n
NGT output 1, mL	0.0	0.0	145	400	2,000	88
NGT output 2, mL	0.0	62.5	200	375	1,300	55

Abbreviations: NGT nasogastric tube, NGT output 1 the NGT output on the day of removal or the day before, NGT output 2 the average NGT output per day during the 2 days prior to removal

Most DGCE cases were retrospectively diagnosed based on NGT output. Among patients with early DGCE, 62 (95.4%) were identified using NGT output criteria, whereas only 3 patients were diagnosed according to radiological criteria. Late DGCE was identified by symptoms and radiological findings in only 8 patients.

A total of 44% (n = 29) of patients with DGCE did not receive any specific treatment. The second largest group comprised 27 patients who underwent conservative therapy with prokinetics. Endoscopic pyloric balloon dilation was the second most frequently performed intervention.

Patients who required NGT reinsertion had a significantly higher incidence of complications. The rate of anastomotic leakage was 60.3% in this group, compared with 8.8% in patients who did not require reinsertion (p < 0.001). Furthermore, these patients experienced a higher frequency of pulmonary complications (Table 5).

Table 5.

Patients undergoing NGT reinsertion

Subcategory	Nasogastric tube reinsertion		p
	no, n = 102	yes, n = 58
Anastomotic leakage [n (%)]
No	93 (91.2)	23 (39.7)	< 0.001
Yes	9 (8.8)	35 (60.3)
Aspiration [n (%)]
No	101 (99)	53 (91.4)	0.024
Yes	1 (1)	5 (8.6)
Pneumonia [n (%)]
No	75 (73.5)	30 (51.7)	0.009
Yes	27 (26.5)	28 (48.3)
Final removal of the NGT and POD [median (IQR)]	7 (5–10)	28 (15–44)	< 0.001
Hospital stay (days) [median (IQR)]	17 (14–23)	37 (25–61)	< 0.001

Abbreviations: NGT nasogastric tube, POD postoperative day, IQR interquartile range, Mean ± SD mean ± standard deviation

The most common reason for NGT reinsertion was confirmed or suspected anastomotic leakage (30 of 58 patients with NGT reinsertion). When considering all reinsertions, the second most frequent reason was the presence of nausea and vomiting, which led to NGT reinsertion in 7 patients. The most common reason for NGT removal was self-removal by patients—28 out of 54 cases at the time of first removal.

Discussion

This study is among the first to perform a detailed analysis of the new DGCE criteria following esophagectomy and to evaluate their applicability in clinical practice and research contexts. In this article, we describe the challenges we encountered and the limitations that may restrict their broader application.

DGCE was a frequent complication and was diagnosed in 41.2% of the patients. Previous studies have reported DGCE rates between 15% and 60%, often relying on varying and nonstandardized diagnostic criteria [1, 2]. In some cases, the diagnosis was based on radiological findings or retained food on upper GI endoscopy, with early DGCE defined as occurring within 30 days [8]. Only four studies, including ours, have applied the current DGCE criteria, reporting incidence rates of 15%–45% [9–11]. Interestingly, Damtoft et al. reported a significantly higher incidence of late DGCE, which may be partly explained by differences in patient follow-up over time [11]. The large difference in the incidence rates suggests that a clear definition is only the first step toward understanding the problem and that further detailed investigation is necessary.

DGCE significantly prolongs hospitalization as observed, in our study participants, length of stay increased by 6 days in cases with early DGCE and by 38 days with late DGCE. This aligns with previous studies that reported extensions of 17 and 8 days, although they did not differentiate between early and late DGCE [12, 13].

Interestingly, the DGCE rate was similar across all surgical approaches. This contrasts with studies suggesting the advantages of minimally invasive surgery [9, 14], although those studies had a higher proportion of open procedures (> 60%). In contrast, the largest group in our study consisted of robot-assisted procedures (40.6%), whereas the remaining subgroups were more evenly distributed and balanced. Intraoperative factors, such as blood loss and operative time, were not associated with DGCE, which is consistent with previous findings.

Furthermore, postoperative complication rates, including anastomotic leakage and pneumonia, were similar in patients with and without DGCE. A large study of 816 patients found that DGCE is not a direct cause of anastomotic leakage [10]. When both conditions occur together, DGCE is most likely the result of leakage-induced mediastinitis [13]. This explains why patients requiring NGT reinsertion—primarily due to anastomotic leakage are more often associated with pulmonal complications. As these patients frequently undergo endoscopic vacuum therapy complemented with endoluminal feeding tubes, continuous gastric drainage via the gastric channel often leads to a diagnosis of DGCE according to the current criteria.

Intraoperative pyloric drainage procedures, such as pyloroplasty or pylorotomy, prevent DGCE after esophagectomy. However, a meta-analysis showed only a minor improvement in gastric emptying without affecting postoperative morbidity [15].

In our cohort, 95% of early DGCE cases were diagnosed based on gastric tube output, and only 5% using radiological criteria. Due to the lack of a required radiological contrast study, only 8 of the 160 patients (5%) fulfilled the criteria for late DGCE. Patients diagnosed based on dysphagia symptoms or endoscopy were excluded. Similar findings in the literature show that DGCE diagnosis requiring endoscopic intervention was based on clinical and endoscopic assessment in 86% of cases, with upper GI water-soluble contrast radiogram used in only 14% [16]. We believe that the radiological findings may serve as a supportive, but not definitive, diagnostic tool. Mandatory radiological evaluation for late DGCE diagnosis limits the universal applicability of the current criteria.

Early and late DGCE appear to be distinct conditions requiring different diagnostic and therapeutic strategies. In our cohort, the majority of patients required no specific treatment, and only 24.6% underwent pyloric dilation. These findings align with previous studies showing a 34% success rate for endoscopic dilation and generally favorable outcomes with conservative management [8]. Early DGCE is generally not associated with significant clinical risk with timely diagnosis, appropriate NGT management, and conservative treatment.

Treatment of late DGCE often requires a more complex approach. In our cohort, interventional methods such as pyloric balloon dilation and botulinum toxin injection were most commonly used. Another study reported that 67.2% of patients receiving endoscopic treatment had late DGCE, with combined therapy showing the best outcomes and enabling earlier discharge [17]. In cases of a long or twisted conduit, reoperation may be necessary—an option rarely considered in early DGCE [18].

We identified the NGT output volume threshold beyond which reinsertion of the nasogastric tube (NGT) was not required. In this patient cohort, the mean daily output over the last 2 days before removal was 200 mL, which increased to 300 mL in patients with DGCE. Comparable studies are lacking. The retrospective analysis was limited by inconsistent documentation, and many cases were excluded due to missing data.

This study has several limitations. First, the consensus DGCE criteria were applied retrospectively, and due to incomplete data on NGT output, two estimation methods were used. While the consensus definition appears suitable for prospective use, several late DGCE cases were excluded owing to the lack of radiological confirmation. Patients with anastomotic leakage were frequently managed with endoluminal feeding tubes, continuous gastric drainage via the gastric channel often fulfills the diagnostic criteria for DGCE, which may influence the reported incidence of DGCE.

Conclusion

This study offers a detailed evaluation of the new DGCE criteria, confirming their applicability in clinical practice. Early and late DGCE are distinct conditions that require different diagnostic and therapeutic approaches. While the criteria are generally practical, radiological findings should be seen as supportive rather than definitive. An NGT output of ≤ 200 mL appears to be a reliable indicator for safe removal, though the decision should also rely on clinical evaluation, supported by radiological evaluation. Prospective, targeted data collection is necessary to more accurately assess NGT management.

Author contributions

Conceptualization: T.W. and P. Z. Methodology: F. N. Formal analysis and investigation: M. K. Writing - original draft preparation: P. Z. Writing - review and editing: T. W. Funding acquisition: A. RJ., S. S. and A. P. Supervision: T. H., M. R., J. I. All authors reviewed the manuscript.

Funding

No funding.

Data availability

The datasets generated and/or analysed during the current study are not publicly available due to patient privacy but are available from the corresponding author on reasonable request.

Declarations

Competing interests

The authors declare no competing interests.