Routine intraoperative jejunostomy placement and minimally invasive oesophagectomy: an unnecessary step?^†

Abstract

OBJECTIVES:

Adequate nutrition is challenging after oesophagectomy. A jejunostomy is commonly placed during oesophagectomy for nutritional support. However, some patients develop jejunostomy-related complications and the benefit over oral nutrition alone is unclear. This study aims to assess jejunostomy-related complications and the impact of intraoperative jejunostomy placement on weight loss and perioperative outcomes in patients with oesophageal cancer treated with minimally invasive Ivor Lewis oesophagectomy (MIE).

METHODS:

From a prospectively maintained database, patients were identified who underwent MIE with gastric reconstruction. Between 2007 and 2016, a jejunostomy was routinely placed during MIE. After 2016, a jejunostomy was not utilized. Postoperative feeding was performed according to a standardized protocol and similar for both groups. The primary outcomes were jejunostomy-related complications, relative weight loss at 3 and 6 months postoperative and perioperative outcomes, including anastomotic leak, pneumonia and length of stay, respectively.

RESULTS:

A total of 188 patients were included, of whom 135 patients (72%) received a jejunostomy. Ten patients (7.4%) developed jejunostomy-related complications, of whom 30% developed more than 1 complication. There was no significant difference in weight loss between groups at 3 months (P = 0.73) and 6 months postoperatively (P = 0.68) and in perioperative outcomes (P-value >0.999, P = 0.591 and P = 0.513, respectively).

CONCLUSIONS:

The use of a routine intraoperative jejunostomy appears to be an unnecessary step in patients undergoing MIE. Intraoperative jejunostomy placement is associated with complications without improving weight loss or perioperative outcomes. Its use should be tailored to individual patient characteristics. Early oral nutrition allows patients to maintain an adequate nutritional status.

Minimally invasive oesophagectomy, Oesophagectomy, Jejunostomy, Oesophageal cancer, Nutrition

INTRODUCTION

Oesophageal cancer is the sixth most common cause of death from cancer worldwide and carries a poor prognosis with a 5-year overall survival rate of 34–47 %[1, 2]. The addition of neoadjuvant chemoradiotherapy to surgery improves survival for patients with locally advanced disease and is recommended in current clinical guidelines [3–5]. Over 60% of oesophageal cancer patients develop preoperative malnutrition due to a combination of dysphagia, odynophagia and a tumour-induced catabolic state [6]. Neoadjuvant chemoradiotherapy can improve dysphagia through tumour reduction [1], but can also further aggravate preoperative malnutrition as a result of chemotherapy-related nausea and emesis [7] and radiation-induced oesophagitis [8]. Malnutrition is a significant risk factor for postoperative complications, including wound infection, sepsis and pneumonia and is associated with an increased length of hospital stay [9, 10].

Traditionally, oral nutrition is limited in the immediate postoperative period following (open) oesophagectomy, and a jejunostomy is routinely placed for postoperative nutritional support. However, there is no evidence of increased pulmonary complications or anastomotic complications with early oral nutrition following oesophagectomy [11]. Moreover, several reports demonstrate a reduction in postoperative complications with early oral nutrition as compared with enteral [12] or parenteral nutrition [13]. Therefore, early oral nutrition is a key component in enhanced recovery after surgery protocols [14]. Finally, jejunostomy placement is associated with minor complications in 10–40% of patients, including jejunostomy site infection and intermittent tube obstruction. Major complications necessitating a laparotomy or laparoscopy, typically a result of the jejunostomy being dislodged or torsion, occur in 1–3% of patients [15].

Minimally invasive Ivor Lewis oesophagectomy contributes to an expedited recovery, a reduction in postoperative complications and possibly an earlier resumption of oral intake in the immediate postoperative period as compared with open oesophagectomy [16]. With the potential complications associated with jejunostomy placement and possible reduction of postoperative morbidity with early oral nutrition, we question the need for routine intraoperative jejunostomy placement for patients undergoing MIE.

This study aims to assess jejunostomy-related complications and the impact of an intraoperative jejunostomy on postoperative outcomes and weight loss at 3 and 6 months postoperatively following minimally invasive Ivor Lewis oesophagectomy for oesophageal cancer.

METHODS

The Partners Human Research Committee approved this retrospective study (IRB #2001P001701) and waived the need for informed consent.

Patient selection

All patients undergoing minimally invasive Ivor Lewis oesophagectomy (MIE) for oesophageal cancer at the Massachusetts General Hospital between 2007 and 2018 were identified from a prospectively maintained institutional database. Patients undergoing MIE for a benign indication or salvage, emergency or palliative resections or resections with non-gastric interposition were excluded. Also, any patients who died prior to discharge during the initial admission for MIE were excluded.

Surgical procedure and perioperative management

All cases were performed by a single surgeon and the technique has been described in detail elsewhere [17]. In brief, a laparoscopic mobilization of the stomach and creation of a gastric conduit were followed by a thoracoscopic mobilization of the oesophagus and oesophageal reconstruction. No pyloric drainage procedure (e.g. pyloric dilatation, pyloric Botox injections or a pyloromyotomy) was performed during MIE.

Jejunostomy tube placement

From November 2007 to January 2016, an intraoperative jejunostomy was routinely placed during MIE. In regard to technique, at ∼30 cm distal to the ligament of Treitz, the jejunum was tacked to the abdominal wall and a 14-Fr jejunostomy tube was placed via a Seldinger technique in the small bowel. The jejunum was stammed to the abdominal wall around the tube, and a stitch was placed distantly to avoid torsion of the bowel. After January 2016, no routine intraoperative jejunostomies were placed during MIE.

When a jejunostomy was placed prior to MIE (e.g. during neoadjuvant treatment) or replaced during oesophagectomy, patients were included in the intraoperative jejunostomy group. Patients with a surgically placed jejunostomy after MIE were included in the postoperative jejunostomy group.

Postoperative feeding

Postoperative feeding was by standard protocol. All patients were kept nil-by-mouth during the first 5 postoperative days. In patients with a jejunostomy, enteral nutrition was initiated on the second postoperative day at a rate of 20 ml/h. If tolerated, this dosage was increased by 20 ml/h every 6–8 h until the nutritional goal of 25–30 kcal/kg and 1.0–1.3 g/kg protein was reached. If a barium swallow on the fifth postoperative day did not demonstrate an oesophageal leak or delayed gastric emptying, oral intake was initiated with clear liquids. Tube feeding was nocturnally cycled when nutritional goals were reached. Diet was advanced to full liquids during the second postoperative week, soft solids in the third postoperative week and a regular diet in the fourth postoperative week. In patients without an intraoperative jejunostomy, oral intake was initiated on the fifth postoperative day as described above. Oral intake was advanced in the same manner as compared with patients with an intraoperative jejunostomy.

Jejunostomy removal

The jejunostomy was typically removed during the first outpatient clinic visit, ∼3 weeks after discharge. Prolonged jejunostomy tube feeding was classified as need for jejunostomy tube feeding >30 days postoesophagectomy in patients with an intraoperative jejunostomy.

Jejunostomy-related complications

Complications were assessed up to 180 days postoesophagectomy. This included both patients with an intraoperative jejunostomy and patients in whom a jejunostomy was placed after MIE. Jejunostomy-related complications such as unplanned replacement, an emergency department visit or office visit, or a jejunostomy-related admission or surgery were included. Instances where the jejunostomy tube was inadvertently discontinued but enteral feeding was no longer indicated (and therefore did not result in jejunostomy replacement) were not categorized as jejunostomy-related complications.

Perioperative weight loss

Relative weight loss was defined as follows:

$$\frac{\mathrm{Weight}\mathrm{ }\mathrm{at}\mathrm{ }\mathrm{specific}\mathrm{ }\mathrm{time}\mathrm{ }\mathrm{interval}\mathrm{ }‐\mathrm{ }\mathrm{weight}\mathrm{ }\mathrm{at}\mathrm{ }\mathrm{the}\mathrm{ }\mathrm{of}\mathrm{ }\mathrm{surgery}\mathrm{ }}{\mathrm{Weight}\mathrm{ }\mathrm{at}\mathrm{ }\mathrm{the}\mathrm{ }\mathrm{time}\mathrm{ }\mathrm{of}\mathrm{ }\mathrm{surgery}}.$$

Variables

Complications and comorbidities were categorized according to the Society of Thoracic Surgeons (STS) guidelines [18]. Anastomotic leak included patients requiring medical (i.e. drainage by interventional radiology, nil-per-os or delay in oral intake) or surgical intervention. Pneumonia was defined using the STS criteria as the presence of 3 of the following criteria: fever, chest X-ray with infiltrate, positive sputum culture, treatment with antibiotics or leucocytosis [18]. Weight at the time of diagnosis, at the end of neoadjuvant treatment and at 3 and 6 months after oesophagectomy was collected retrospectively. Also, jejunostomy-related complications within 180 days postsurgery, and readmissions related to failure to thrive and pyloric drainage procedures within 60 days were examined. Postoperative major morbidity included the following complications based on the STS guidelines: unexpected return to the operating room (OR), anastomotic leakage, reintubation, initial ventilatory support >48 hours, renal failure according to RIFLE criteria [19], pneumonia and recurrent laryngeal nerve injury [20].

Univariable and multivariable analyses of factors associated with relative weight loss

A univariable and subsequent multivariable logistic regression was performed to assess independent contributions to >10% weight loss at 3 and 6 months postoperatively. Risk factors examined included no intraoperative jejunostomy placement, male gender, preoperative weight loss >15%, prior abdominal surgery, coronary artery, peripheral vascular, or cerebrovascular disease, diabetes, current smoking, BMI >30 (i.e. obesity), neoadjuvant treatment, adjuvant treatment, tumour histology, pathological stage III or IV, postoperative major morbidity, readmissions and pyloric drainage procedures within 60 days postoesophagectomy.

Statistical analyses

All statistical analyses were performed using Stata version 15 [21]. Categorical variables were reported using frequencies with percentages and were compared using the Fisher’s exact test. Normally distributed continuous data were reported using the mean with standard deviation and were compared using the Student’s t-test. Non-normally distributed data were reported using the median with the first and third quartile of the continuous distribution and were compared using the Wilcoxon rank-sum test. Factors were included in the multivariable logistic regression if P-value was <0.2 in the univariable analyses. Results of the logistic regression were reported using Odds ratios (ORs) with corresponding 95% confidence intervals (CIs). A P-value <0.05 was considered statistically significant.

RESULTS

Between November 2007 and February 2018, a total of 190 consecutive patients underwent MIE, of whom 136 patients underwent MIE with an intraoperative jejunostomy. Figure 1 demonstrates a flow diagram of patients during the study period. During the study interval when routine intraoperative jejunostomies were placed, 2 patients did not receive an intraoperative jejunostomy due to extensive adhesions related to prior abdominal surgery. During the study interval in which no routine intraoperative jejunostomy was performed, 6 patients received an intraoperative jejunostomy due to preoperative failure to thrive (5 patients had received a preoperative) and were included in the intraoperative jejunostomy group. Two patients were excluded, as they died prior to discharge from non-jejunostomy-related causes.

Figure 1:

Flow diagram of jejunostomy placement.

Patients

Table 1 lists the baseline demographics. Patients were predominantly male (81%), overweight (mean BMI 26.8), and 80% received neoadjuvant and/or adjuvant treatment. At baseline, there were no statistically significant differences between patients with an intraoperative jejunostomy as compared with patients without an intraoperative jejunostomy.

Table 1:

Baseline patient characteristics

Characteristics	No intraoperative jejunostomy (n = 53)	Intraoperative jejunostomy (n = 135)	P-value
Gender, n (%)			>0.999
Female	10 (19)	25 (19)
Male	43 (81)	110 (81)
Age, mean ± SD	64.0 ± 12.2	63.2 ± 9.4	0.667
BMIsurg, mean ± SD	27.9 ± 3.8	26.5 ± 4.8	0.067
Tumour histology, n (%)			0.563
Adenocarcinoma	50 (94)	122 (90)
Squamous cell carcinoma	3 (6)	13 (10)
ASA-score, n (%)			0.276
I	1 (2)	1 (1)
II	26 (49)	53 (39)
III	26 (49)	81 (60)
Hypertension, n (%)	26 (49)	64 (47)	0.872
Peripheral vascular disease, n (%)	4 (8)	7 (5)	0.507
Coronary artery disease, n (%)	10 (19)	20 (15)	0.511
Diabetes, n (%)	7 (13)	19 (14)	>0.999
Current smoking, n (%)	7 (13)	16 (12)	0.807
Number of packyears (first–third quartile), median	25 (15–45)	30 (14.5–45)	0.877
FEV/predicted, mean ± SD	98.3 ± 21.0	93.1 ± 18.1	0.126
Prior abdominal surgery, n (%)	11 (21)	25 (19)	0.837
Prior cardiothoracic surgery, n (%)	5 (9)	10 (7)	0.765
Additional CT and/or RT, n (%)			0.793
Neoadjuvant only	39 (74)	89 (66)
Adjuvant only	2 (4)	5 (4)
Perioperative	3 (6)	12 (9)
Pathological tumour stage,an (%)			0.172
T0	15 (28)	24 (18)
Tis	1 (2)	3 (2)
T1	12 (23)	47 (35)
T2	10 (19)	14 (10)
T3	15 (28)	43 (32)
Pathological nodal stage,an (%)			0.892
N0	37 (70)	88 (65)
N1	12 (23)	31 (23)
N2	4 (8)	13 (10)
N3	0 (0)	3 (2)

a

AJCC/UICC 7th edition.

ASA: American Society of Anesthesiologists; BMI_surg: body mass index at time of surgery; CT: chemotherapy; FEV: forced expiratory volume; RT: radiation therapy; SD: standard deviation.

Jejunostomy-related complications

Figure 2 demonstrates a flow diagram of jejunostomy-related complications per patient. Among the 135 patients with an intraoperative jejunostomy, 10 patients (7.4%) developed a jejunostomy-related complication. Among the 53 patients without an intraoperative jejunostomy, 4 patients (7.5%) received a jejunostomy after MIE (i.e. postoperative jejunostomy), of whom 2 patients developed a jejunostomy-related complication. Table 2 demonstrates a list of all jejunostomy-related complications within 180 days postsurgery. The most common jejunostomy-related complication in patients with an intraoperative jejunostomy was leakage around tube necessitating intervention (33%) followed by obstruction (22%). Jejunostomy-related complications resulted in 9 emergency department visits, 7 admissions, 8 interventional radiology visits for jejunostomy replacement and 3 operative procedures. In 1 patient, the jejunostomy tube migrated into the peritoneal cavity requiring jejunostomy removal. Subsequently, this patient developed failure to thrive and required a second jejunostomy placement. Ultimately, this jejunostomy was found to have withdrawn into the abdominal wall necessitating incision and drainage followed by admission for intravenous antibiotic therapy. Finally, 2 patients required surgical replacement of the jejunostomy because the jejunostomy had fallen out or because it was clogged.

Figure 2:

Flow diagram of jejunostomy-related complications per patient.

Table 2:

List of jejunostomy-related complications within 180 days postoesophagectomy

	Complications	No.	Total (%)	Intervention	No.	Row (%)	Presentation	No.	Row (%)
Jejunostomy placed during MIE	Jejunostomy leakage (not peritoneal)	6	33	IR jejunostomy replacement	3	50	IR visit	3	50
				Percutaneous jejunostomy replacement	2	33	During admission	2	33
				Closure of leakage	1	17	ED visit	1	17
	Jejunostomy obstructed/clogged	4	22	Percutaneous jejunostomy replacement	2	50	During admission	2	50
				Flushing of jejunostomy	1	25	ED visit	1	50
				Surgery: jejunostomy replacement	1	25	Office visit	1	50
	Jejunostomy dislodged	2	11	IR jejunostomy replacement	1	50	IR visit	1	50
				Percutaneous jejunostomy replacement	1	50	ED visit	1	50
	Jejunostomy site infection	2	11	IV antibiotics	1	50	Admission	1	50
				Jejunostomy removal	1	50	ED visit	1	50
	Jejunostomy leakage (peritoneal)	1	6	Jejunostomy removal	1	100	Admission	1	100
	Jejunostomy malfunction	1	6	Pump replacement	1	100	Office visit	1	100
	Jejunostomy migration to abdominal wall	1	6	Surgery: incision and drainage	1	100	Admission	1	100
	Jejunostomy related complicated wound infection	1	6	IV antibiotics	1	100	ED visit	1	100
	Total	18			18			18
Jejunostomy placed after MIE	Jejunostomy obstructed/clogged	6	55	Flushing of jejunostomy	4	67	ED visit	4	67
				IR jejunostomy replacement	2	33	IR visit	2	33
	Jejunostomy dislodged	5	45	IR jejunostomy replacement	4	80	Admission	3	60
				Surgery: jejunostomy replacement	1	20	IR visit	2	40
	Total	11			11			11

ED: emergency department; IR: interventional radiology; IV: intravenous; No.: number of complications.

Perioperative relative weight loss

Figure 3 and Table 3 demonstrate the impact of an intraoperative jejunostomy on relative weight loss. At 3 or 6 months postoperative, no statistically significant difference in median relative weight loss was observed between patients without an intraoperative jejunostomy as compared with patients with an intraoperative jejunostomy (−8.4% vs −8.6%, P = 0.73; −9.8% vs −10.9%, P = 0.68, respectively).

Figure 3:

Impact of an intraoperative jejunostomy on relative weight loss.

Table 3:

Impact of an intraoperative jejunostomy on relative weight loss

Events	No intraoperative jejunostomy		Intraoperative jejunostomy		P-value
	Median relative weight change (%)	First–third quartile	Median relative weight change (%)	First–third quartile
Time of diagnosis	2.2	0.0–6.3	2.4	0.0–9.4	0.77
End of neoadjuvant treatment	−0.4	−2.5 to 1.7	−1.4	−3.6 to 0.1	0.27
Time of surgery	NA	NA	NA	NA	NA
3 Months postoperative	−8.4	−11.7 to −5.6	−8.6	−12.3 to −5.1	0.73
6 Months postoperative	−9.8	−14.3 to −4.5	−10.9	−15.5 to −5.6	0.68

NA: not applicable.

Factors associated with weight loss

Table 4 demonstrates the results of a multivariable logistic regression looking at factors associated with weight loss. Factors independently associated with >10% weight loss at 3 months postoperative were BMI at the time of surgery above 30.0 (OR 2.55, 95% CI 1.05–6.40; P = 0.041) and adjuvant treatment (OR 3.50, 95% CI 1.20–11.21; P = 0.026). Only BMI >30 at the time of surgery was independently associated with >10% weight loss at 6 months postoperative (OR 3.67, 95% CI 1.31–7.10; P = 0.007). The omission of an intraoperative jejunostomy was not independently associated with weight loss at 3 or 6 months postoperatively.

Table 4:

Logistic regression of factors associated with >10% weight loss

Factors	>10% weight loss 3 months postoperative						>10% weight loss 6 month postoperative
	Univariable			Multivariable			Univariable			Multivariable
	OR	CI	P-value	OR	CI	P-value	OR	CI	P-value	OR	CI	P-value
Intraoperative J tube	0.87	0.39–19.92	0.735	1.25	0.32–2.48	0.602	0.76	0.33–1.73	0.510	1.31	0.32–1.82	0.508
Smoking
Never	Reference			Reference			Reference			Reference
Past	0.59	0.27–1.31	0.196	0.55	0.22–1.32	0.185	1.04	0.46–2.30	0.930	NA	NA	NA
Current	0.67	9.19–2.26	0.519	0.69	0.17–2.66	0.589	0.50	0.14–1.67	0.266	NA	NA	NA
Mean BMIsurg >30	2.60	1.19–5.67	0.016	2.55	1.05–6.40	0.041	3.33	1.46–7.63	0.004	3.67	1.31–7.10	0.007
Adjuvant treatment	3.58	1.26–10.16	0.016	3.50	1.20–11.21	0.026	3.32	1.01–10.88	0.048	3.28	1.05–12.50	0.054
Squamous cell histology	0.46	0.10–1.62	0.261	NA	NA	NA	0.35	0.07–1.32	0.142	0.36	0.07–1.41	0.166
Pathological stage III–IV	0.60	0.27–1.28	0.197	0.44	0.16–1.07	0.081	0.92	0.43–1.92	0.819	NA	NA	NA

BMI_surg: body mass index at time of surgery; CI: confidence interval; NA: not applicable; OR: odds ratio.

Postoperative outcomes

Table 5 demonstrates the impact of an intraoperative jejunostomy on postoperative outcomes. No statistically significant differences were observed in postoperative anastomotic leakage and pneumonia, length of hospital stay, readmissions or pyloric dilations within 60 days postoesophagectomy between patients with an intraoperative jejunostomy as compared with patients without an intraoperative jejunostomy. In an analysis of operative time, the first 40 MIE cases were excluded as we have previously established a learning curve for MIE [22]. After correcting for this learning curve effect, the median operating time was 35 min longer for patients with an intraoperative jejunostomy (P = 0.001).

Table 5:

Postoperative outcomes

30-day postoperative outcomes	No intraoperative jejunostomy (n = 53)	Intraoperative jejunostomy (n = 135)	P-value
Complications, n (%)
Anastomotic leakagea	1 (1.9)	2 (1.5)	>0.999
Pneumonia	6 (11.3)	12 (8.9)	0.708
Chyle leakagea	2 (3.8)	2 (1.5)	0.314
Recurrent laryngeal nerve injury	0 (0.0)	1 (0.7)	0.718
Ileus	1 (1.9)	2 (1.5)	1.000
Oesophageal stricture requiring dilation	1 (1.9)	1 (0.7)	0.485
Delayed conduit emptying or obstructiona	1 (1.9)	2 (1.5)	>0.999
Sepsis	2 (3.8)	2 (1.5)	0.316
Min operation time (first–third quartile), median	243 (223–265)	278 (249–317)	0.001
Length of hospital stay, median (first–third quartile)	7 (6–8)	7 (6–8)	0.513
Readmission <30 days of discharge, n (%)	6 (11.3)	15 (11.1)	0.574
Readmission for FTT <60 days,bn (%)	6 (11.3)	10 (7.4)	0.395
Surgical pyloric dilation <60 days,bn (%)	2 (3.8)	5 (3.7)	>0.999

a

Requiring surgical or medical intervention.

b

Postsurgery.

FTT: failure to thrive; Min: minutes.

Prolonged jejunostomy tube feeding

The median time interval between jejunostomy removal and hospital discharge was 17 days (first–third quartile 13–23) and between jejunostomy removal and operative date was 24 days (first–third quartile 20–30). Among patients with an intraoperative jejunostomy, 33 patients required continuous jejunostomy tube feeding >30 days postoesophagectomy. Reasons for prolonged jejunostomy tube feeding were failure to thrive due to inadequate oral intake and/or gastric outlet obstruction (n = 11), failure to thrive due to a complicated hospital course (n = 9), adjuvant treatment (n = 2) or unrelated to nutrition (n = 11) (e.g. related to the schedule of patient and/or surgeon). One patient required continuous jejunostomy tube feeding until the time of death. For 9 patients, the duration of jejunostomy tube feeding was unknown.

Postoperative jejunostomy tube placement

Among patients with an intraoperative jejunostomy, 1 patient (1%) required the placement of a postoperative jejunostomy. This patient had initial adequate oral intake, and the jejunostomy was removed during an outpatient clinic visit. However, the patient developed progressive dysphagia and required a jejunostomy which was used until death. Among patients without an intraoperative jejunostomy, 4 patients (7.5%) required a postoperative jejunostomy due to failure to thrive during a readmission for aspiration pneumonia (n = 2), the need for adjuvant treatment (n = 1) and poor gastric emptying (n = 1).

DISCUSSION

This study demonstrates that an intraoperative jejunostomy does not have an impact on postoperative weight loss at 3 or 6 months in patients undergoing minimally invasive Ivor Lewis oesophagectomy for oesophageal cancer. This may be explained by early resumption of oral nutrition following MIE due to a low postoperative complication rate [23] (equivalent between both groups) and an expedited recovery associated with MIE [16]. Early oral feeding potentially allows patients to meet their nutritional goals or it is possible that patients develop a catabolic state regardless of intraoperative jejunostomy placement, which has has no impact in the early postoperative period. As the jejunostomy is routinely removed at 3 weeks postoperatively when patients have achieved adequate oral caloric intake, the jejunostomy does not appear to play a significant role.

In addition, this study demonstrates jejunostomy-related complications in 7.4% patients with an intraoperative jejunostomy, of whom 30% had more than 1 complication. Despite relatively low incidence of complication, the physical discomfort to the patient and the cost associated with jejunostomy feeding formula, emergency department visits and readmissions are difficult to quantify. Among patients without an intraoperative jejunostomy, postoperative jejunostomy placement was required in 7.5% of patients. Indications for postoperative placement included aspiration pneumonia, failure to thrive related to adjuvant treatment or dysphagia.

These results suggest that intraoperative jejunostomy placement is not indicated for over 92% of oesophageal cancer patients treated with MIE. With careful patient selection and the early initiation of oral feeding, a routine intraoperative jejunostomy can be avoided in the majority of oesophageal cancer patients undergoing MIE.

The patients included in this study were primarily those with distal oesophageal or gastroesophageal junction cancer and the results may not be applicable to Eastern countries, where patients generally present with squamous cell carcinoma and more proximal tumours. Patients with squamous cell cancers may present with more severe dysphagia and significant nutritional deficiency.

The strengths of this study include the generalizability of the study participants. Only oesophageal cancer patients were included who underwent Ivor Lewis oesophagectomy with gastric reconstruction by a minimally invasive approach, which is the current standard of care in centres around the world. In addition, over 80% of patients underwent neoadjuvant treatment. Potential weaknesses include the loss to follow-up data. Weight data were missing in 22% at 3 months and 23% at 6 months postoperatively. In addition, we acknowledge that this is a single-institution, single-surgeon study with limitations. However, with this study design, we were able to eliminate potential bias related to differences in surgical techniques, surgical skills and perioperative management.

Conflict of interest: none declared.

Footnotes

†

Presented at the 32nd Annual Meeting of the European Association for Cardio-Thoracic Surgery, Milan, Italy, 18–20 October 2018.