The implementation of Enhanced Recovery After Surgery (ERAS) protocols in the management of esophagectomy has led to improved postoperative outcomes and earlier discharge [1]. Though consensus has been reached on many aspects of ERAS protocol, the exclusion of nasogastric tube (NGT) in the postoperative period is still controversial and not widely practiced. A recent Cochrane review has recommended against routine NGT use in common abdominal surgeries as it leads to delayed bowel recovery, increased pulmonary complications, a trend towards increased hospital stay, similar anastomotic leak rates, and increased patient discomfort due to throat pain, nasal mucosal damage, sinusitis, gastritis, and epistaxis [2].

Esophagectomy is considered different from other gastrointestinal surgeries because a vagotomized gastric conduit is used to restore gastrointestinal continuity. Due to concern for fluid accumulation and distension of the gastric conduit, which may lead to pulmonary aspiration and anastomotic leak, surgeons are reluctant to avoid an NGT completely after esophagectomy. The standard ERAS protocol suggests either early removal of NGT on a postoperative day (POD) 2 or 3 or not inserting it perioperatively at all [1]. Some studies have reported good outcomes with an ERAS protocol and NGT exclusion following esophagectomy [3–5], but most were retrospective and in the setting of open transhiatal or transthoracic esophagectomy. With minimally invasive esophagectomy (MIE) giving superior perioperative outcomes and becoming the standard, the role of NGT exclusion in this setting is less clear.

This study’s premise was that excluding NGT during ERAS after MIE did not increase the risk of cervical esophagogastric anastomotic leak or pulmonary complications, compared to ERAS protocol with early removal of NGT.

This study was a retrospective analysis conducted at a single institution, specifically the Department of Surgical Gastroenterology at the All India Institute of Medical Sciences, Jodhpur. The data were collected retrospectively from patients who underwent minimally invasive Mckeown esophagectomy (robotic or thoracoscopic-assisted) for biopsy-proven carcinoma of the esophagus or gastroesophageal junction Siewert type I and II between July 2018 and June 2022. Patients who underwent esophagectomy for benign disease or other tumors, conversion to open thoracotomy, Ivor-Lewis esophagectomy, or combined esophagogastrectomy for Siewert type III tumors were excluded.

Patient evaluation and treatment

Patients were staged at baseline with a contrast-enhanced computed tomography (CECT) of the neck, thorax, and abdomen or a whole-body F-18 fluorodeoxyglucose positron-emission tomography (PET)-CT scan. Those with clinical T2–T4 or N+ or >3 cm length tumors without distant metastases and any other contraindication underwent neoadjuvant treatment. Neoadjuvant treatment was either chemoradiation therapy or chemotherapy and was decided in a multidisciplinary meeting based on the tumor histology, differentiation, patient performance status, age, and comorbidities. After 4 weeks of completion of neoadjuvant therapy, all patients were evaluated with a PET-CT scan for response assessment.

The patient’s general condition, cardiopulmonary, and nutritional status were optimized. Incentive spirometry was done under supervision, a high protein oral or enteral diet was given, and smoking cessation was followed for at least 1 week. Patients then underwent minimally invasive McKeown esophagectomy with thoracic esophageal mobilization, either thoracoscopically (Rubina, Karl Stoz) or robotic-assisted (Da Vinci Xi system, Intuitive Surgical). All patients underwent standard two-field lymph node dissection, and lymph nodes along the recurrent laryngeal nerves were dissected selectively only if they were evident on preoperative imaging (PET-CT or CECT).

The abdominal phase was performed either laparoscopically or by laparotomy. The stomach was mobilized with the preservation of right gastric and gastroepiploic vessels and the division of left gastric and gastroepiploic vessels. In all cases, a gastric conduit of about 3 to 4 cm wide was created using a linear cutting stapler (NTLC, Ethicon). Then, the conduit was pulled up via the posterior mediastinum and the cervical esophagogastric anastomosis (CEGA) was performed by a semi-mechanical or handsewn technique. All patients underwent stretching of the pylorus before the formation of the gastric conduit. The surgical team and surgical steps remained similar throughout the study period.

The perioperative ERAS protocol followed is described in Table 1. From July 2018 to June 2020, NGT was placed in all patients in our unit and removed by POD 3. From July 2020 to June 2022, we excluded NGT while implementing all other components of the ERAS protocol unchanged. Thus, two groups were formulated as follows: group I, NGT placed across CEGA into the gastric conduit and removed on POD 3; group II, NGT not placed after CEGA.

Table 1 . Perioperative ERAS protocol for minimally invasive esophagectomy followed in this study.

Preoperative period

Preoperative counselling by a clinical nurse for at least 20 minutes

Incentive spirometry 15 times every 4 hours starting 1 week before surgery

Preoperative isotonic carbohydrate drink 800 mL at 12 hours and 400 mL at 2–3 hours (to be drunk in 20 minutes) before surgery in all patients except uncontrolled diabetes mellitus (HbA1c >7)

Thromboprophylaxis (combined mechanical and pharmacological): LMWH 12 hours before surgery and thromboembolic deterrent stockings

Intraoperative period

General anesthesia with preemptive epidural analgesia

Maintaining normothermia

Goal-directed fluid therapy 5–6 mL/kg/hour intraoperatively and adding vasopressor when mean blood pressure drops by 20%

No neck drain and abdominal drain

Immediate extubation after surgery or extubation on POD 1

Shifting the patient to a ward with continuous monitoring of vitals

Postoperative period

POD 1a)

Negative fluid balance

Chest X-ray

FJ trial feed 1 L at 40 mL/hour

Out of bed 1 hour morning and evening with assisted walking

Start LMWH

Saline nebulization and steam inhalation every 6 hours

Incentive spirometry 10 times each hour while awake

POD 2

Remove urinary and epidural catheter

FJ feed 2 L, half strength at 60–80 mL/hour

Stoppage of intravenous fluids by evening

Out of bed for 4 hours with assisted 2 walks of 100 m

Continue nebulization, steam inhalation, and spirometry as the previous day

POD 3

Remove central line and ICD if output <200 mL/24 hours

FJ feed 2 L, full strength at 80 mL/hour

Out of bed for 4 hours with assisted/non-assisted 3–4 walks of 150 m

Continue nebulization, steam inhalation, and spirometry as previous day

POD 4

Oral contrast study with nonionic contrast (Iohexol) (POD 5 if POD 4 is Sunday)

Allow oral clear liquids if the contrast study shows no leak

Stop LMWH

Non-assisted 3–4 walks more than 150 m

Continue FJ feeds 2 L, nebulization, and spirometry as the previous day

POD 5

Oral liquids/semisolid diet

FJ feeds reduced to 1 L at 80 mL/hour

Continue non-assisted walks, nebulization, and spirometry as previous day

Discharge the patient

^a)A 1 or 2 days margin will be allowed for each component..

HbA1c, glycosylated hemoglobin; LMWH, low molecular weight heparin; POD, postoperative day; FJ, feeding jejunostomy; ICD, intercostal drainage tube..

Although strict adherence to all ERAS components within the specified timelines was encouraged, a grace period of 1 to 2 days was permitted; exceeding this was considered noncompliance with ERAS protocols. Discharge criteria included the ability to tolerate an oral semisolid diet without difficulty, passage of stool, ambulation, and independence in daily activities.

Data collection

Demography, clinicopathological characteristics, imaging findings, details of neoadjuvant treatment received, operative details, postoperative complications, and compliance for all ERAS protocol components were recorded. The day of commencement of oral feeds, NGT reinsertion, length of hospital stay, 30-day readmission, 30-day reoperation, and 30-day mortality were also recorded. The complications of esophagectomy were recorded according to the Esophagectomy Complications Consensus Group (ECCG) and graded according to the Clavien-Dindo (CD) system [6].

Statistical analysis

Continuous data were expressed as median with interquartile range, while discrete data were expressed as percentages. Comparisons between groups were made using the Mann-Whitney U test and chi-square tests for continuous and discrete data. A two-tailed p-value of less than 0.05 was considered significant. Data analysis was done using IBM SPSS version 23.0 (IBM Corp.).

Fig. 1 shows the study profile diagram with exclusion criteria. After applying the selection criteria, 50 patients were included: 30 in group I (NGT) and 20 in group II (No NGT). Table 2 shows the baseline characteristics and perioperative outcomes of the patients. There was no significant difference in age, sex, body mass index, performance status, or comorbidities. Patients in both groups had tumors in the middle or lower thoracic esophagus, with all having squamous histology. All patients received either neoadjuvant chemoradiotherapy or chemotherapy, which was similar between the groups. The percentage of thoracoscopic or robotic-assisted esophageal mobilization in each group was also comparable. In group I, CEGA was performed by handsewn technique in 17 patients (56.7%) and semi-mechanical technique in 13 patients (43.3%); whereas, in group II, all 20 patients underwent CEGA by semi-mechanical technique.

Table 2 . Baseline characteristics and treatment details of patients in both groups.

Characteristic	Group I (NGT)	Group II (No NGT)	p-value
No. of patients	30	20
Age (yr)	51 (45–62)	49 (40–61)	0.53
Sex, male/female	13 (43.3)/17 (56.7)	11 (55.0)/9 (45)	0.41
Body mass index (kg/m2)	21.7 (19.4–23.3)	21.5 (18.4–24.6.0)	0.78
ECOG, 0/1/2	7 (23.3)/22 (73.3)/1 (3.3)	10 (50.0)/9 (45.0)/1 (5.0)	0.12
Comorbidity
Diabetes mellitus	1 (3.3)	2 (10.0)	0.55
Hypertension	1 (3.3)	4 (20.0)	0.14
Tumor location			0.64
Middle thoracic	13 (43.3)	10 (50.0)
Lower thoracic	17 (56.7)	10 (50.0)
Esophagogastric junction	0 (0)	0 (0)
Clinical stage			0.06
cT2, N0/N1/N2	2/5/1	0/1/0
cT3, N0/N1/N2	5/14/3	6/11/2
Neoadjuvant therapy			0.91
Chemotherapy	13 (43.3)	9 (45.0)
Chemoradiotherapy	17 (56.7)	11 (55.0)
Preoperative hemoglobin (g/dL)	9.5 (8.6–10.7)	10 (9.0–10.8)	0.76
Preoperative albumin (g/dL)	2.5 (2.4–3.0)	2.6 (2.3–2.9)	0.54
Operation type			0.27
Thoracoscopic	18 (60.0)	15 (75.0)
Robotic-assisted	12 (40.0)	5 (25.0)
Anastomotic technique			NA
Handsewn	17 (56.7)	0 (0)
Semi-mechanical	13 (43.3)	20 (100)
Intraoperative blood loss (mL)	200 (100–300)	200 (175–300)	0.42
Operation time (min)	440 (360–480)	420 (360–480)	0.76

Values are presented as number only, median (interquartile range), or number (%)..

NGT, nasogastric tube; ECOG, Eastern Cooperative Oncology Group; NA, not applicable..

Figure 1. Study profile diagram with exclusion criteria and two groups for comparison. ERAS, Enhanced Recovery After Surgery.

Postoperative outcomes and adherence to ERAS protocol are presented in Table 3. More patients in group II were extubated immediately after surgery (18 of 20 [90.0%] vs. 17 of 30 [56.7%], p = 0.01), while other characteristics like intensive care unit stay, day of ambulation, Foley’s catheter removal, intercostal drain removal, and oral intake were comparable between the two groups. More patients in group II had early epidural removal. After an oral contrast study on POD 4 was negative for the leak, oral intake were started, which was statistically not significant between the two groups (63.3% vs. 65.0%, p = 0.91).

Table 3 . Compliance and postoperative outcomes of minimally invasive McKeown esophagectomy managed with ERAS protocol parameters.

Variable	Group I (NGT) (n = 30)	Group II (no NGT) (n = 20)	p-value
Compliance
Immediate extubation	17 (56.7)	18 (90.0)	0.01
ICU stay ≤1 day	28 (93.3)	19 (95.0)	>0.99
Ambulation ≤1 day	29 (96.7)	18 (90.0)	0.55
Foley catheter removal ≤2 days	24 (80.0)	17 (85.0)	0.72
Epidural removal ≤2 days	19 (63.3)	1 (5.0)	0.01
ICD removal ≤4 days	21 (70.0)	17 (85.0)	0.31
Oral intake initiation day (POD 4)	19 (63.3)	13 (65.0)	0.91
Postoperative outcome
Anastomotic leaka) (ECCG grade I/II/III)	5/0/0	2/0/0	0.68
Pulmonary complications	5 (16.7)	5 (25.0)	0.41
RLN Palsy	2 (6.7)	2 (10.0)	0.67
Chylothorax	3 (10.0)	0 (0)	NA
NGT reinsertion	1 (3.3)	1 (5.0)	>0.99
Major complicationsb)	4 (13.3)	1 (5.0)	0.63
Readmission	2 (6.7)	1 (5.0)	>0.99
Length of stay (day)	7 (6–7)	6 (5–7)	0.03

Values are presented as number (%), number only, or median (interquartile range)..

ERAS, Enhanced Recovery After Surgery; ICU, intensive care unit; ICD, intercostal drainage tube; POD, postoperative day; ECCG, Esophagectomy Complications Consensus Group; RLN, recurrent laryngeal nerve; NA, not applicable; NGT, nasogastric tube..

^a)Esophagectomy Major Complications Consensus Group grade I/II/III. ^b)Clavien-Dindo grade ≥IIIa..

Among the postoperative complications, the anastomotic leak was not statistically significant between the two groups (16.7% vs. 10.0%, p = 0.68), and all had grade I leaks, according to ECCG. Although pulmonary complications were more prevalent in group II (25.0%) vs. group I (16.7%), it was insignificant. One patient in each group had NGT reinsertion for conduit dilatation. Three patients had chylothorax in group I and none had it in group II. Two patients had readmission in group I, while one patient was readmitted in group II. The median length of hospital stay was 7 days in group I, while it was 6 days in group II (p = 0.03).

On comparing the group with (n = 7) or without (n = 43) the anastomotic leak (Table 4), none of the variables was found significant between the groups except for readmission, which was higher in patients with an anastomotic leak (p = 0.01). NGT was not a significant factor in affecting anastomotic leaks.

Table 4 . Comparison between the patients who developed anastomotic leak and who did not.

Variable	Anastomotic leak (n = 7)	No anastomotic leak (n = 43)	p-value
Sex, male/female	3/4	21/22	0.76
Smoking, yes /no	1/6	9/34	0.68
Tobacco chewing, yes/no	4/3	15/28	0.26
Diabetes mellitus, yes/no	1/6	2/41	0.32
Coronary artery disease, yes/no	0/7	2/41	0.56
Tumor location, middle/lower	4/3	19/24	0.52
Neoadjuvant treatment, CT/CRT	4/3	18/25	0.45
Anastomotic technique, stapled/handsewn	3/4	30/13	0.16
NGT, yes/no	5/2	25/18	0.50
Oral initiation day, POD 4/>4 days	3/4	29/14	0.20
Clavien-Dindo grade, <IIIa/≥IIIa	5/2	40/3	0.07
Readmission, yes/no	3/4	1/42	0.01

CT, chemotherapy; CRT, chemoradiation therapy; NGT, nasogastric tube; POD, postoperative day..

The NGT was introduced for gastric decompression during the postoperative period after all major gastrointestinal surgeries. It became a standard of care with the premise that it may decrease anastomotic leaks and pulmonary complications due to aspiration. The perceived benefit of routine NGT use was not scrutinized or validated in prospective trials before its widespread adoption. Literature now provides evidence for the contrary, that it lacks the benefit of routine NGT use and added risks/discomfort to patients during the postoperative period. A Cochrane review of 37 randomized trials found that routine NGT use in abdominal surgeries, compared to selective or no NGT use, led to similar anastomotic leak rates, delayed recovery, and oral intake with increased pulmonary complications and patient discomfort [2].

The ERAS protocol is a multimodal perioperative strategy to improve the patient’s preoperative, intraoperative, and postoperative physiologic status, leading to reduced perioperative stress response, fluid overload, and catabolism. This results in faster recovery and shorter hospital stay duration without increasing postoperative morbidity and mortality. With ERAS protocols advocating earlier initiation of diet to decrease starvation, NGT lost its relevance and began to be removed early or omitted altogether in the postoperative period of most abdominal surgeries. However, NGT continues to be used after esophagectomy by most surgeons for fear of fluid accumulation in the denervated gastric conduit, which may lead to distension, anastomotic leak, and aspiration pneumonitis.

The ERAS guidelines on esophagectomy advise early removal of NGT by POD 3 or its omission altogether. In this study of minimally invasive McKeown’s esophagectomy patients managed by a uniform perioperative ERAS protocol, the NGT omission group had a similar anastomotic leak rate compared to NGT removal on POD 3. Previous studies have uniformly reported no increase in anastomotic leak rates on NGT omission or early removal within 24 hours following esophagectomy, irrespective of approach (open, transhiatal, or minimally invasive), procedure (McKeown or Ivor-Lewis esophagectomy), anastomotic technique (handsewn vs. semi-mechanical), and geography (east vs. west) [4,5,7–11].

A second perceived benefit of NGT is the reduction in aspiration pneumonitis. Pulmonary complications occur in up to 60% of patients after esophagectomy [12]. They are multifactorial, with preexisting lung disease, the pain of thoracotomy, fluid-related lung congestion, aspiration, etc., likely playing a role. Our study showed a nonsignificant trend toward higher pulmonary complications in the NGT omission group (25.0% vs. 16.7%, p = 0.41). This could be attributed to surgery performed during the coronavirus disease 2019 pandemic, which may have led to slightly higher pulmonary complications. NGT was not a significant factor affecting pulmonary complications in multivariate analysis. While some studies reported NGT omission or early removal within 24 hours to not affect pulmonary complication rates [9,10,13], others reported a decrease in pulmonary complications with NGT omission [14]. One randomized controlled trial (RCT) showed similar tracheal acid aspiration between NGT and no NGT groups and higher respiratory complications in the no NGT group, but the numbers were small with skewed data [15]. In the current era, less pain due to MIE, as well as preoperative incentive spirometry, early ambulation, fluid restriction, and epidural analgesia as part of ERAS protocols, have led to a reduction in pulmonary complications, but it still occurs in 20% to 30% of patients. NGT omission is likely to lead to better patient comfort, a better expectorative effort to clear secretions, and motivation to remain ambulant—all of which may contribute to further decreased pulmonary complications.

Delayed gastric emptying (DGE) is a major complication following esophagectomy encountered in 15%–39% of patients [16]. The etiology of postesophagectomy DGE is multifactorial, which includes dysfunctional peristalsis, incomplete relaxation of the pylorus, and redundant gastric conduit. Gravity is the key factor driving gastric emptying in the early postoperative period, as gastric contraction and viscous force are negligible in the conduit. DGE decreases as the diameter of the gastric conduit decreases. Among the interventions, pyloromyotomy and pyloroplasty are the most commonly performed procedures to reduce DGE. In a study by Deng et al. [17], pyloric finger fracture could effectively decrease the risk of DGE without long-term complications like dumping syndrome and bile reflux. A systematic review including all different pyloric interventions has shown a nonsignificant trend toward a lower risk of DGE. However, significant heterogeneity existed among the different pyloric interventions in the study. In RCTs by Hayashi et al. [9] and Daryaei et al. [10], one patient in no NGT group required reinsertion for conduit dilatation. Similarly, in our study, one patient in groups I and II required NGT reinsertion for conduit dilatation. Both patients were managed conservatively by nil per oral, NGT drainage, prokinetics, and feeding jejunostomy feeds. In our study, pyloric finger fracture was performed in all patients, and the conduit diameter was upto 4 cm. The decreased conduit diameter and intraoperative pyloric dilatation by the pyloric finger fracture technique may explain the low DGE in our patients.

Nevertheless, routine placement of NGT increases the patient’s discomfort leading to delay in ambulation and delayed oral intake. Using a simple patient-reported score, one RCT measured the patient discomfort on the day of NGT removal in the early group (POD 2) vs. the late group (POD 7) and found less patient discomfort in the early removal group [11]. Our study was a retrospective analysis, and patient discomfort was not measured postoperatively prospectively. Further, routine NGT insertion postoperatively with improper taping increases the risk of alar necrosis, and omission of NGT in the postoperative period can also avoid the risk of pressure necrosis.

The median postoperative hospital stay was significantly less in the no NGT group (6 days vs. 7 days). Although the difference is only 1 day, it could be attributed to reduced patient discomfort without NGT in the postoperative period and improved compliance with ERAS protocol. It allowed for early ambulation, early removal of chest tubes and urinary catheters, and early oral intake with better pulmonary toileting, thereby reducing the complications leading to reduced length of hospital stay. A meta-analysis by Weijs et al. [13] reported similar results that hospital stay was significantly shorter with preoperative or early removal of NGT. However, a recent meta-analysis of six RCTs by Kaaki et al. [18] found that early removal of NGT has no impact on length of stay (LOS), whereas early oral intake found shorter LOS. Hence, it is apparent that early oral feeding (EOF) can be initiated easily in the no NGT group to receive its benefit.

Further, several reviews and meta-analyses have shown that EOF is also a safe practice without any significant increase in complications in a patient undergoing colonic and gastric surgery. However, many surgeons are reluctant to practice EOF following esophagectomy due to the risk of anastomotic leak and aspiration pneumonia. A meta-analysis of two RCTs has shown that EOF does not increase the risk of anastomotic leak and aspiration pneumonia [16]. Patients in the EOF group had a significantly shorter time to pass flatus and bowel movements compared to late oral feeding. Our study initiated orals on POD 4 after the oral contrast study confirmed no CEGA leak. As our study is the first in our population to eliminate NGT in the perioperative protocol, we were cautious about initiating EOF. However, after the positive results that NGT exclusion did not affect the anastomotic leak and major complications (CD grade of ≥IIIa), starting EOF from POD 1 might be safe. A summary of the literature review on NGT exclusion in McKeown esophagectomy is compiled in Table 5, along with the results of our study [9–11].

Table 5 . Summary of literature on nasogastric tube (NGT) exclusion in McKeown esophagectomy.

Study	Year	No. of subjects	NGT removal (day)a)	NGT reinserted	Anastomotic location	Anastomotic type	Anastomotic leak (%)	Pulmonary complications (%)	Length of stay (day)a)
Hayashi et al. [9]	2019	75	Group I: 7 Group II: 1	Group I: 1 Group II: 1	Cervical	Handsewn	Group I: 2.7 Group II: 8.8 p = 0.26	Group I: 21 Group II: 22	Group I: 12 (10–17) Group II: 12 (9–17)
Daryaei et al. [10]	2009	40	Group I: 4.2 ± 1.3 Group II: 0	Group I: 0 Group II: 1	Cervical	Handsewn	Group I: 27 Group II: 0 p = 0.01	Group I: 0 Group II: 11	Group I: 10 ± 3.5 Group II: 13 ± 8.2
Mistry et al. [11]	2012	150	Group I: 6–10 Group II: 2	Group I: 7 Group II: 23	Cervical	Handsewn, stapled	Group I: 11 Group II: 8 p = 0.81	Group I: 33 Group II: 24	Group I: 25 ± 18 Group II: 27 ± 12
Current study	2024	50	Group I: 3 Group II: 0	Group I: 1 Group II: 1	Cervical	Handsewn, semi-mechanical	Group I: 16.7 Group II: 10 p = 0.68	Group I: 16.7 Group II: 25 p = 0.40	Group I: 7 (6–7) Group II: 6 (5–7)

Group 1, NGT group; group II, no NGT group..

^a)Values are presented as mean ± standard deviation or median (interquartile range)..

Our study has a few limitations. First, it is a retrospective study with small sample size. However, our data was maintained prospectively with strict adherence to the ERAS protocol during the study period. Our results will pave the path for conducting an RCT. Second, the anastomotic technique was handsewn initially, which evolved to semi-mechanical later in the study. However, multiple RCTs and meta-analyses have shown no difference in the rate of anastomotic leak between the two techniques. Lastly, we could not perform a quality-of-life analysis, as it would help us better understand the patient functional outcomes.

In conclusion, our study supports that avoiding NGT in the postoperative period is safe without an increase in an anastomotic leak and major complications (CD grade of ≥IIIa).

Ethical statements

The Institutional Ethics Committee of All India Institute of Medical Sciences, Jodhpur has approved the study protocol (No. AIIMS/IEC/2022/4249). Informed written consent was obtained from all the patients.

Author’s contributions

Conceptualization: VN, VKV, SB

Data curation: VN, VKV, SB

Formal analysis: VKV, SS, PV, LA

Methodology: VKV, SS, PV, LA

Project administration: VKV, SS, PV, LA

Writing–original draft: VN, VKV, SB

Writing–review & editing: VKV, SB

All authors reviewed the final version of the manuscript.

Conflict of interest

All authors have no conflicts of interest to declare.

Funding/support

None.

Data availability

The data presented in this study are available upon reasonable request to the corresponding author.