Benefits of jejunostomy feeding in patients who underwent gastrectomy for cancer treatment

Abstract

BACKGROUND

Gastric cancer is associated with significant undernutrition responsible for an increase in morbidity and mortality after gastrectomy.

AIM

To evaluate the impact of enteral nutrition by jejunostomy feeding in patients undergoing gastrectomy for cancer.

METHODS

Between 2003 and 2017, all patients undergoing gastrectomy for cancer treatment were included retrospectively. A group with jejunostomy (J + group) and a group without jejunostomy (J - group) were compared.

RESULTS

Of the 172 patients included, 60 received jejunostomy. Preoperatively, the two groups were comparable with respect to the nutritional parameters studied (body mass index, albumin, etc.). In the postoperative period, the J + group lost less weight and albumin: 5.74 ± 8.4 vs 9.86 ± 7.5 kg (P = 0.07) and 7.2 ± 5.6 vs 14.7 ± 12.7 g/L (P = 0.16), respectively. Overall morbidity was 25% in the J + group and 36.6% in the J - group (P = 0.12). The J + group had fewer respiratory, infectious, and grade 3 complications: 0% vs 5.4% (P = 0.09), 1.2% vs 9.3% (P = 0.03), and 0% vs 4.7% (P = 0.05), respectively. The 30-day mortality was 6.7% in the J + group and 6.3% in the J - group (P = 0.91).

CONCLUSION

Jejunostomy feeding after gastrectomy improves nutritional characteristics and decreases postoperative morbidity. A prospective study could confirm our results.

Key Words: Feeding jejunostomy; Gastric cancer; Undernutrition; Morbimortality; Survival

Core Tip: Jejunostomy feeding in patients undergoing gastrectomy for cancer significantly improves postoperative nutritional status and reduces complications. A study of 172 patients showed that those with jejunostomy had less weight loss, better albumin levels, and fewer respiratory, infectious, and grade 3 complications compared to those without jejunostomy. Overall morbidity was lower, though 30-day mortality rates were similar between the groups.

INTRODUCTION

Gastric cancer is a serious pathology responsible for significant mortality. Worldwide, it is the fifth most common cancer but the third leading cause of cancer deaths, with a fatality rate of 75%[1]. According to the National Cancer Institute, gastric cancer is more common in the Caribbean arc, especially in Martinique, affecting 1.58 times more men and 2.31 times more women than elsewhere in the world. Gastrectomy, the only potentially curative treatment, is a source of significant morbidity, which, according to the literature, varies from 11% to 46%[2]. The poor nutritional status of patients with gastric cancer, secondary to various factors such as impaired gastric mechanics or decreased dietary intake secondary to dysphagia, is a source of increased morbidity and mortality. Undernutrition is an independent risk factor for morbidity and mortality[3]. The management of this undernutrition would therefore reduce morbidity. The recommendations of the French Society of Anaesthesia and Resuscitation go in this direction by suggesting that any patient malnourished undergoing high-risk nutritional surgery should benefit from enteral nutritional support[4].

In Martinique, it seems that patients are particularly fragile and malnourished, with a consultation time that is often delayed. Gastric cancers appear to be frequently diagnosed at an advanced or even symptomatic stage (high occlusive syndrome, haemorrhagic tumours). Recently, the fragility of these patients has motivated the surgeons of our department to introduce jejunostomy feeding. Our objective was therefore to evaluate the impact of enteral feeding by jejunostomy in patients undergoing gastrectomy for cancer treatment.

MATERIALS AND METHODS

We conducted a monocentric retrospective study in the University Hospital of Fort de France in Martinique between 1 January 2003 and 30 April 2017, seeking to compare different postoperative results in all patients who had undergone gastrectomy for cancer treatment. The inclusion criterion was any adult patient (≥ 18 years) who had been diagnosed with gastric cancer and who had undergone gastrectomy. There were no exclusion criteria. Two groups were compared: A jejunostomy group (J + group) and a group without jejunostomy (J - group). The evaluation of morbimortality was the primary endpoint. Secondary criteria were length of stay and survival time.

Jejunostomy tube placement according to Fontan[5]: The technique according to Fontan was chosen with direct insertion of the tube into the jejunal lumen without an intraparietal path. The insertion orifice of the tube was fixed directly to the parietal peritoneum without prior burying.

Data statement

We had the authorization of the National Commission for Informatics and Liberties for the recovery of the names of patients corresponding to our search criteria (No. 2065145 v 0). The operative procedures were found thanks to the coding of the Common Classification of Medical Acts (CCAM), and the diagnoses were found using the 10^th version of the International Classification of Diseases (ICD10) motivated by the Program of Medicalization of Information Systems. The Department of Medical Information of the University Hospital of Fort de France was contacted to find patients corresponding to the CCAM and ICD10 codes.

All demographic, histological, nutritional, anaesthetic, operative, postoperative, and length-of-stay data were collected. The complications were graded according to the Dindo-Clavien classification[6] (grade 1: Any deviation of normal postoperative consequences without the need for intervention, except the administration of antiemetics, antipyretics, analgesics, diuretics, or electrolytes; grade 2: A complication requiring pharmacological treatment, different from that used for a grade 1 complication; grade 3: A complication requiring surgical, endoscopic, or radiological treatment; grade 4: A life-threatening complication requiring hospitalization in an intensive care unit; grade 5: Death). Vital status data, dates of the latest news, and death dates were mostly present. The vital missing data could be retrieved by contacting the Martinican Association for Epidemiological Research in Cancer. Nutritional data - weight, body mass index (BMI), albuminemia, prealbuminemia, prognostic nutritional index (PNI)[7], PNI status < 46[8], and nutritional status defined by ANAES (undernutrition and severe undernutrition)[9] - were recorded preoperatively at least one month before the operation and up to two months postoperatively. When two values of the same variable were available preoperatively, the value closest to the date of intervention was used. In the postoperative period, the value closest to two months after the intervention date was used to have a clearer picture of the patient’s nutritional status, whether the patient was enterally fed or not.

Statistical analysis

All data acquisition was carried out in an Excel spreadsheet (Microsoft, Redmond, WA, United States). Categorical variables were expressed as number of cases and percentages, while continuous variables were expressed as mean and standard deviation. Categorical variables were compared by Pearson’s χ² test for all multinomial and binomial variables except when one of the expected contingency table counts was less than 5, in which case Fisher's exact test was used. Continuous variables were compared using Student's t-test after systematic verification of the homogeneity of the variances by a Levene test. Overall and disease-free survival times were analysed by the Kaplan-Meier method, and survival curves of the two groups were compared by the Mantel-Cox test (or log-rank test). Survival times were expressed by their median value with a 95%CI, and their dispersion was expressed by the interquartile range. All tests were bilateral, and P < 0.05 was considered significant. Analyses were conducted on the entire cohort, with subgroup analysis for partial and total gastrectomies. A multivariate analysis adjusted for age, sex, dementia, prior Helicobacter pylori infection, history of stroke, heart failure, atrial fibrillation and severe preoperative undernutrition was performed. Statistical analysis was performed using SPSS, version 20.0, for Mac (Armonk, NY, United States).

RESULTS

Patient distribution

A total of 172 patients underwent gastrectomy for cancer treatment between 1 January 2003 and 30 April 2017. Of these, 74 (43%) had a total gastrectomy, 89 (52%) had a partial gastrectomy, and 9 (5%) had an atypical gastrectomy. Forty-eight (65%) of the patients with a total gastrectomy had jejunostomy, while 12 (13%) of the patients with a partial gastrectomy had jejunostomy. No patients with an atypical gastrectomy received jejunostomy (Figure 1). The baseline characteristics of the two groups were comparable except for the male sex, which was more frequently represented in the J + group than in the J - group: 43 (71.7%) vs 57 (50.9%) (P = 0.01; Table 1). The preoperative nutritional status of both groups was similar, although the BMI values of the J + group patients tended to be lower: 22.7 ± 4.6 vs 24.2 ± 4.8 kg/m² (P = 0.09; Table 2). Tumours in the J + group were more advanced than those in the J - group, especially T4 tumours: 19 (31.7%) vs 15 (13.4%) (P = 0.036; Table 3). Over time, the number of patients who received jejunostomy increased, especially from 2010 to 2011 (Figure 2).

Figure 1 Patient distribution flowchart.

Figure 2 Jejunostomy placement over time.

Table 1 Baseline demographic characteristics of the population, cohort N = 172, n (%).

Variables	J + (n = 60)	J - (n = 112)	P value
Age (years, mean ± SD)	69.4 ± 12.6	70.2 ± 11.2	0.65
Male	43 (71.7)	57 (50.9)	0.01a
H. pylori infection	11 (18.3)	13 (11.6)	0.25
Smoking	6 (10.2)	12 (10.8)	1
Medical background	40 (66.7)	77 (69.4)	0.73
Dementia	5 (8.3)	4 (3.6)	0.28
Stroke	6 (10)	7 (6.2)	0.38
AF	7 (11.7)	8 (7.2)	0.4
CAD	4 (6.7)	7 (6.2)	1
Heart failure	6 (10)	5 (4.5)	0.2
LEAD	4 (6.7)	7 (6.2)	1
Diabetes	15 (25)	28 (25)	1
Other cancer	12 (20)	18 (16.1)	0.53
Abdominal surgery	18 (30)	46 (41.1)	0.19
AC/APA	21 (35.6)	26 (23.9)	0.11
Corticoids/IT	1 (1.7)	2 (1.8)	1
WHO score
0	18 (38.3)	37 (38.9)	0.58
1	20 (42.6)	43 (45.3)
2	5 (10.5)	12 (12.6)
3	4 (8.5)	3 (3.2)

^aP < 0.05.

A group with jejunostomy (J + group) and a group without jejunostomy (J - group) were compared. AC: Anticoagulant; AF: Atrial fibrillation; APA: Antiplatelet agent; CAD: Coronary artery disease; H. pylori: Helicobacter pylori; IT: Immunosuppressive therapy; LEAD: Lower-extremity artery disease; WHO: World Health Organization.

Table 2 Preoperative nutritional characteristics, cohort N = 172, n (%).

Variables	J + (n = 60)	J - (n = 112)	P value
Weight (kg, mean ± SD)	64 ± 12.9	66.7 ± 14.3	0.23
BMI (kg/m2, mean ± SD)	22.7 ± 4.6	24.2 ± 4.8	0.09
Weight loss (kg, mean ± SD)	12.2 ± 6.4	10.1 ± 4.3	0.15
Proportion weight loss (%, mean ± SD)	15.9 ± 8.5	13.9 ± 5.5	0.21
Duration of weight loss (months, mean ± SD)	3.9 ± 5.2	4.7 ± 4.3	0.52
Preoperatory albuminemia (g/L, mean ± SD)	32.3 ± 6.9	32.6 ± 8.6	0.87
Preoperatory prealbuminemia (g/L, mean ± SD)	0.19 ± 0.06	0.17 ± 0.05	0.3
PNI (mean ± SD)	41.2 ± 8.8	39.8 ± 12	0.74
PNI < 46	18 (54.5)	13 (81.2)	0.22
Undernutrition	21 (36.2)	34 (35.7)	1
Severe undernutrition	16 (28.1)	17 (17.9)	0.55

A group with jejunostomy (J + group) and a group without jejunostomy (J - group) were compared. Continuous variables are expressed as the mean and standard deviation. Categorical variables are expressed as numbers and percentages. Weight loss is expressed as an absolute and a relative average value. PNI: Prognostic nutritional index; BMI: Body mass index.

Table 3 Tumour characteristics, cohort N = 172, n (%).

Variables	J + (n = 60)	J - (n = 112)	P value
Adenocarcinoma	47 (79.7)	87 (78.4)	1
Grade of differentiation
Good	5 (10.9)	7 (9)	0.76
Moderate	18 (39.1)	38 (48.7)	0.35
Poor	20 (43.5)	31 (39.7)	0.71
Contingent of independent cells	23 (38.3)	33 (30.3)	0.31
GIST	9 (15.5)	22 (20.2)	0.53
Low risk	1 (11.1)	8 (40)	0.2
Medium risk	1 (11.1)	3 (15)	1
High risk	7 (77.8)	9 (45)	0.13
Tumour ≤ 3 cm	10 (16.7)	20 (18.9)	0.84
T
0	0 (0)	3 (2.7)	0.036a
1	5 (8.3)	12 (10.7)
2	6 (10)	21 (18.8)
3	23 (38.8)	39 (34.8)
4	19 (31.7)	15 (13.4)
N0	18 (32.1)	31 (34.4)	0.86
M0	50 (90.9)	79 (84.9)	0.45
Clear margins	45 (75)	92 (84.4)	0.15
Vascular invasion	25 (43.1)	48 (44)	1
Perineural invasion	27 (46.6)	39 (36.1)	0.24

^aP < 0.05.

A group with jejunostomy (J + group) and a group without jejunostomy (J - group) were compared. GIST: Gastrointestinal stromal tumor.

Morbimortality

Concerning morbidity and mortality, there was no difference in 30-day mortality between the groups: 6.7% vs 6.3% for the J + and J - groups, respectively (P = 1). In total, 56 patients presented at least one complication, with a total of 84 complications recorded. The overall complication rate was 25% in the J + group, compared to 36.6% in the J - group, with no statistically significant difference (P = 0.12; Table 4). Regarding the specific complications, the rate of anastomotic fistula was comparable: 2 (3.3%) vs 6 (5.4%) for the J + and J - groups, respectively (P = 0.71). Patients in the J + group had fewer infectious complications than those in the J - group: 2 (1.2%) vs 16 (9.3%) (P = 0.03). They also tended to present fewer respiratory complications: 0 (0%) vs 6 (5.4%) (P = 0.09; Table 4). There were no differences in complications according to the Dindo-Clavien classification for grades 1, 2, 4, and 5 between the groups. However, patients in the J + group had fewer grade 3 complications: 0 (0%) vs 8 (4.7%) (P = 0.05). Regarding high-grade complications (≥ 4), we observed two grade 4 complications all in J - group patients: One case of respiratory distress due to hypoxemic pneumonia caused by Klebsiella pneumoniae infection and one case of massive pulmonary embolism. Additionally, there were 11 grade 5 complications, including 4 in patients in group J +, including four cases of septic shock due to hypoxemic pneumonia, one case of aspiration pneumonia in a J + patient, one case of bilateral massive pulmonary embolism in a J + patient, one case of duodenal fistula leading to septic shock, one case of cardiorespiratory arrest, one case of hemorrhage on postoperative day 2, one case of extensive small bowel necrosis in a J + patient on postoperative day 11, and one death of unknown cause (J + patient).

Table 4 Postoperative complications, cohort N = 172, n (%).

	J + (n = 60)	J - (n = 112)	P value
Global complications	15 (25)	41 (36.6)	0.12
Thromboembolic	2 (1.2)	4 (2.3)	1
Respiratory (other than aspiration pneumonia and thromboembolic)	0 (0)	6 (5.4)	0.09
Aspiration pneumonia	2 (3.3)	0 (0)	0.12
Haemorrhagic	1 (1.7)	6 (5.4)	0.42
Infectious	2 (1.2)	16 (9.3)	0.03a
Deep collection	2 (1.2)	3 (1.7)	1
Ileus	4 (2.3)	10 (5.8)	0.77
Occlusion	0 (0)	3 (1.7)	0.55
Fistula	2 (3.3)	6 (5.4)	0.71
Others	3 (1.7)	9 (5.2)	0.54
Renutrition syndrome	1 (0.6)	0 (0)	0.34
Probe obstruction	1 (0.6)	0 (0)	0.34
Abscess around the probe	1 (0.6)	0 (0)	0.34

^aP < 0.05.

A group with jejunostomy (J + group) and a group without jejunostomy (J - group) were compared.

Among the secondary outcomes, the lengths of stay varied between groups. The length of stay was longer for the J + group: On average 20.49 ± 9.32 vs 16.45 ± 9.47 days (P = 0.01). These longer durations were at the expense of conventional hospitalization: 17.75 ± 8.50 vs 14.06 ± 7.71 days (P = 0.01), with no significant difference between the groups regarding the postoperative length of stay (i.e., between the day of the intervention and the day of discharge) (Table 5). Concerning the postoperative nutritional characteristics, these worsened in both groups, but this worsening was less among patients in the J + group: A tendency towards less weight loss, 5.74 ± 8.4 vs 9.86 ± 7.5 kg (P = 0.07), and a smaller decrease in albumin, 7.2 ± 5.6 vs 14.7 ± 12.7 g/L (P = 0.15; Table 6).

Table 5 Length of stay, cohort N = 172, n (%).

Variables	J + (n = 60)	J - (n = 112)	P value
Total hospitalization (days, mean ± SD)	20.49 ± 9.32	16.45 ± 9.47	0.01a
Conventional	17.75 ± 8.50	14.06 ± 7.71	0.01a
Postoperative	18.13 ± 8.32	17.96 ± 12.52	0.71
PICU	1.65 ± 1.55	1.87 ± 3.61	0.68
ICU	0.84 ± 3.45	0.1 ± 0.71	0.11
CCU	0.07 ± 0.37	0.02 ± 0.19	0.33

^aP < 0.05.

A group with jejunostomy (J + group) and a group without jejunostomy (J - group) were compared. PICU: Postoperative intensive care unit; ICU: Intensive care unit; CCU: Critical care unit.

Table 6 Postoperative nutritional characteristics.

Variables	J +	J -	P value
Weight (kg, mean ± SD)	58.86 ± 11.8	56 ± 9.8	0.34
Weight loss (kg, mean ± SD)	5.74 ± 8.4	9.86 ± 7.5	0.07
Proportion weight loss (%, mean ± SD)	10.25 ± 8	14.19 ± 9.5	0.12
Albuminemia (g/L, mean ± SD)	31.4 ± 5.5	32.3 ± 10.9	0.8
Loss of albuminemia (g/L, mean ± SD)	7.2 ± 5.6	14.7 ± 12.7	0.15
Proportion albuminemia loss (%, mean ± SD)	20 ± 15.6	40 ± 33.1	0.16

A group with jejunostomy (J + group) and a group without jejunostomy (J - group) were compared.

In subgroup analysis, whether in the partial (Table 7) or total (Table 8) gastrectomy cohort, there was no significant difference on morbidity between the groups with or without jejunostomy. In multivariate analysis (Table 9), in the entire cohort, a history of dementia was a significant predictor of overall complications: Odds ratio (OR) = 6.42, 95%CI: 1.19-52.4, P = 0.04, and infectious complications: OR = 9.75, 95%CI: 1.03-88.8, P = 0.04. Additionally, among partial gastrectomies, a prior Helicobacter pylori infection increased the risk of other complications: OR = 6.41, 95%CI: 0.85-43.91, P = 0.04.

Table 7 Subset analysis in partial gastrectomies, cohort N = 89, n (%).

	J + (n = 12)	J - (n = 77)	P value
Global complications	4 (33.3)	30 (38.9)	0.96
Thromboembolic	0 (0)	2 (2.6)	1
Respiratory (other than aspiration pneumonia and thromboembolic)	0 (0)	6 (7.8)	0.7
Aspiration pneumonia	1 (8.3)	0 (0)	0.28
Haemorrhagic	0 (0)	3 (3.9)	1
Infectious	0 (0)	10 (12.9)	0.4
Deep collection	0 (0)	2 (2.6)	1
Ileus	1 (8.3)	8 (10.4)	1
Occlusion	0 (0)	2 (2.6)	1
Fistula	1 (6.5)	5 (8.3)	1
Others	1 (8.3)	7 (9.1)	1
Renutrition syndrome	0 (0)	0 (0)	1
Probe obstruction	0 (0)	0 (0)	1
Abscess around the probe	1 (8.3)	0 (0)	0.28

A group with jejunostomy (J + group) and a group without jejunostomy (J - group) were compared.

Table 8 Subset analysis in total gastrectomies, cohort N = 74, n (%).

	J + (n = 48)	J - (n = 26)	P value
Global complications	11 (22.9)	7 (26.9)	0.92
Thromboembolic	2 (4.2)	1 (3.8)	1
Respiratory (other than aspiration pneumonia and thromboembolic)	0 (0)	0 (0)	1
Aspiration pneumonia	1 (2.1)	0 (0)	1
Haemorrhagic	1 (2.1)	1 (3.8)	1
Infectious	2 (4.2)	4 (15.4)	0.21
Deep collection	2 (4.2)	1 (3.8)	1
Ileus	3 (6.3)	1 (3.8)	1
Occlusion	0 (0)	0 (0)	1
Fistula	1 (2.1)	1 (3.8)	1
Others	2 (4.2)	2 (7.7)	0.92
Renutrition syndrome	1 (2.1)	0 (0)	1
Probe obstruction	1 (2.1)	0 (0)	1
Abscess around the probe	0 (0)	0 (0)	1

A group with jejunostomy (J + group) and a group without jejunostomy (J - group) were compared.

Table 9 Multivariable analysis.

Type of cohort	Variables	Age			Sex (M)			Dementia			H. pylori infection			Sroke			Heart failure			Atrial fibrillation			Severe undernutrition
	Estimates	OR	95%CI	P value	OR	95%CI	P value	OR	95%CI	P value	OR	95%CI	P value	OR	95%CI	P value	OR	95%CI	P value	OR	95%CI	P value	OR	95%CI	P value
Entire cohort	Global complications	1	0.97-1.03	0.51	0.79	0.41-1.55	0.49	6.42	1.19-52.4	0.04a	1.74	0.67-4.45	0.25	0.26	0.03-1.26	0.14	0.66	0.12-2.88	0.61	0.6	0.13-2.16	0.47	1.28	0.61-2.64	0.52
	Thromboembolic	0.95	0.88-1.02	0.18	1.37	0.24-10.54	0.73	-	-	-	0.78	0.03-6.63	0.85	-	-	-	8.36	0.33-110.9	0.12	-	-	-	0.51	0.03-3.54	0.56
	Respiratory	1.05	0.96-1.16	0.33	0.36	0.04-2.1	0.27	7	0.21-113	0.19	4.15	0.42-32.3	0.17	-	-	-	1.47	0.05-16.94	0.78	1.94	0.09-16.88	0.59	0.2	0-2.05	0.27
	Aspiration pneumonia	1.17	0.93-1.77	0.28	2.13	0.05-603	0.7	12.79	0.12-863	0.21	2.34	0.01-102	0.66	-	-	-	12.1	0.11-760	0.21	-	-	-	0.31	0-17.5	0.61
	Haemorrhagic	0.99	0.93-1.06	0.8	0.51	0.1-2.41	0.39	-	-	-	1.34	0.07-9	0.8	-	-	-	-	-	-	-	-	-	1.11	0.15-5.5	0.91
	Infectious	0.99	0.95-1.04	0.69	0.53	0.19-1.48	0.23	9.75	1.03-88.8	0.04a	0.66	0.09-2.9	0.63	-	-	-	0.66	0.03-4.97	0.73	1.72	0.23-8.36	0.54	0.64	0.16-2.03	0.48
	Deep collection	1	0.92-1.09	0.92	0.13	0-1.03	0.09	-	-	-	-	-	-	-	-	-	13.84	0.4-488	0.11	-	-	-	3.21	0.44-26.99	0.24
	Ileus	1.03	0.98-1.09	0.22	1.05	0.33-3.47	0.94	-	-	-	1.36	0.19-5.94	0.71	-	-	-	0.87	0.04-6.01	0.91	1.68	0.22-8.16	0.56	2.06	0.62-6.48	0.22
	Occlusion	0.98	0.87-1.1	0.22	0.4	0.02-4.84	0.94	-	-	-	2.77	0.1-40.5	0.71	9.62	0.33-215.5	0.99	-	-	-	-	-	-	-	-	-
	Fistula	1.01	0.95-1.08	0.81	0.21	0.03-1	0.07	0.92	0.02-14.57	0.96	1.53	0.14-9.89	0.68	2.1	0.07-23.7	0.59	2.38	0.09-27.2	0.52	-	-	-	2.52	0.49-12.9	0.25
	Others	1.01	0.95-1.07	0.79	1.36	0.39-5.42	0.64	2.53	0.22-20.3	0.41	2.11	0.39-8.61	0.33	1.6	0.16-10.13	0.65	-	-	-	-	-	-	1.03	0.23-3.82	0.97
Partial gastrectomies	Global complications	1.02	0.98-1.07	0.34	0.8	0.29-2.17	0.66	-	-	-	4.41	0.93-25.3	0.07	0.53	0.03-4.15	0.59	1.11	0.13-7.9	0.92	0.61	0.08-3.3	0.59	0.86	0.29-2.4	0.77
	Thromboembolic	0.98	0.82-1.22	0.83	0.57	0.01-20.3	0.74	-	-	-	-	-	-	-	-	-	39.9	0.88-540	0.06	-	-	-	0.98	0.02-32.7	0.99
	Respiratory	1.03	0.93-1.15	0.63	0.38	0.04-2.8	0.37	4.27	0.12-72.3	0.33	7.56	0.61-96.3	0.09	-	-	-	2.39	0.07-46.7	0.57	3.92	0.16-49.8	0.31	0.09	0-1.18	0.14
	Aspiration pneumonia	-	-	-	-	-	-	-	-	-	-	-	-	-	-	-	-	-	-	-	-	-	-	-	-
	Haemorrhagic	1.01	0.91-1.14	0.86	2.03	0.17-47.1	0.58	-	-	-	-	-	-	-	-	-	-	-	-	-	-	-	2.03	0.17-47.1	0.99
	Infectious	1.1	0.98-1.17	0.17	0.8	0.17-3.57	0.77	11.3	0.91-271.8	0.07	0.39	0.01-4.33	0.53	-	-	-	0.98	0.03-11.3	0.99	1.43	0.06-12.1	0.77	0.56	0.08-2.88	0.515
	Deep collection	-	-	-	-	-	-	-	-	-	-	-	-	-	-	-	-	-	-	-	-	-	-	-	-
	Ileus	1.03	0.97-1.12	0.34	1.36	0.3-6.53	0.69	-	-	-	1.1	0.05-8.21	0.94	-	-	-	1.33	0.06-11.7	0.82	1.11	0.05-9.6	0.93	2.27	0.51-10.4	0.27
	Occlusion	0.98	0.82-1.22	0.85	0.49	0.01-16.38	0.67	-	-	-	-	-	-	17.8	0.52-929	0.09	-	-	-	-	-	-	-	-	-
	Fistula	1.02	0.94-1.16	0.63	0.07	0-0.74	0.07	0.29	0-12.11	0.56	6.96	0.39-137.9	0.17	16.35	0.42-702	0.1	2.07	0.04-45.9	0.66	-	-	-	4.51	0.53-54.47	0.18
	Others	1.01	0.93-1.1	0.8	1.72	0.29-13.95	0.57	2.81	0.16-31.48	0.43	6.41	0.85-43.91	0.05	3.53	0.28-31.38	0.27	-	-	-	-	-	-	0.67	0.07-4	0.68
Total gastrectomies	Global complications	0.98	0.93-1.03	0.38	0.73	0.22-2.52	0.61	-	-	-	0.73	0.1-3.62	0.72	-	-	-	-	-	-	1.01	0.05-8.95	0.99	2.3	0.62-8.4	0.2
	Thromboembolic	0.93	0.82-1.02	0.15	2.5	0.16-85.16	0.54	-	-	-	2.98	0.11-49.1	0.44	-	-	-	-	-	-	-	-	-	-	-	-
	Respiratory	-	-	-	-	-	-	-	-	-	-	-	-	-	-	-	-	-	-	-	-	-	-	-	-
	Aspiration pneumonia	1.11	0.91-1.63	0.46	-	-	-	-	-	-	-	-	-	-	-	-	-	-	-	-	-	-	-	-	-
	Haemorrhagic	1	0.89-1.16	0.95	0.42	0.01-11.99	0.57	-	-	-	-	-	-	-	-	-	-	-	-	-	-	-	3.17	0.12-86.87	0.43
	Infectious	0.93	0.84-1.00	0.08	0.28	0.03-1.77	0.19	-	-	-	0.84	0.03-8.17	0.89	-	-	-	-	-	-	10.75	0.33-309	0.14	0.85	0.04-8.44	0.9
	Deep collection	1	0.9-1.12	0.96	0.18	0.01-2.4	0.21	-	-	-	-	-	-	-	-	-	-	-	-	-	-	-	7.53	0.61-184.88	0.13
	Ileus	1.04	0.94-1.19	0.47	1.85	0.19-46.88	0.63	-	-	-	-	-	-	-	-	-	-	-	-	4.92	0.17-87.66	0.27	1.68	0.07-20.5	0.69
	Occlusion	-	-	-	-	-	-	-	-	-	-	-	-	-	-	-	-	-	-	-	-	-	-	-	-
	Fistula	0.93	0.82-1.04	0.24	0.64	0.02-20.36	0.78	-	-	-	-	-	-	-	-	-	-	-	-	-	-	-	-	-	-
	Others	1.02	0.94-1.13	0.62	0.38	0.04-3.54	0.37	-	-	-	-	-	-	-	-	-	-	-	-	-	-	-	1.04	0.05-9.63	0.97

^aP < 0.05.

H. pylori: Helicobacter pylori; OR: Odds ratio; M: Male.

Recurrence-free survival and overall survival

The median recurrence-free survival time of the cohort was 38 months with 95%CI: 20.1-55.8. There was no difference in median survival time between the J + and J - groups: 18 months with 95%CI: 0-59.7 vs 42 months with 95%CI: 20.3-63.6, respectively (P = 0.22; Figure 3A). The median overall survival time of the cohort was 68 months with 95%CI: 41.2-94.7. There was no difference in overall survival between the J + and J - groups: 79 months with 95%CI: 0-168.1 vs 68 months with 95%CI: 50.1-85.8, respectively (P = 0.36; Figure 3B).

Figure 3 Recurrence-free survival and overall survival. A: Recurrence-free survival; B: Overall survival.

DISCUSSION

Malnutrition (pre- and postoperative) is a recognized factor in increasing morbidity in the management of patients undergoing high-risk surgery[4]. We conducted a retrospective study seeking to compare morbidity and mortality in patients undergoing gastrectomy for cancer treatment. Overall, our groups were initially comparable, both in terms of clinical characteristics and in terms of nutritional characteristics. Histologically, the tumours in the J + group were a little more advanced. We demonstrated a significant positive impact of postoperative enteral nutrition via jejunostomy on some complications. In fact, enterally fed patients had fewer infectious complications, 2 (1.2%) vs 16 (9.3%) (P = 0.03); tended to have fewer respiratory complications, 0 (0%) vs 6 (5.4%) (P = 0.09); and had fewer grade 3 complications according to the Dindo-Clavien classification, 0 (0%) vs 8 (4.7%) (P = 0.05). Furthermore, the postoperative nutritional status, although worse for both groups, was better in patients fed via jejunostomy, both regarding weight loss, 5.74 ± 8.4 vs 9.86 ± 7.5 kg (P = 0.07), and regarding albumin decrease, 7.2 ± 5.6 vs 14.7 ± 12.7 g/L (P = 0.15). This lower morbidity in the group refed by jejunostomy led to an increase in the overall hospitalization length of stay, which in reality depended on the conventional hospitalization length of stay (in the digestive surgery department). This is more likely a reflection of the number of patients with a total gastrectomy than of enteral nutrition. Indeed, the majority of patients with jejunostomy underwent more often a total gastrectomy, 48 (65%), than a partial gastrectomy, 12 (13%). Moreover, for some patients, preoperative hospitalization was necessary to refeed them with parenteral nutrition before surgery. Feeding jejunostomy is crucial for the nutritional management of gastric cancer patients due to the high prevalence of malnutrition in this population. It allows direct enteral feeding, thus preventing post-surgery malnutrition complications, which significantly improves patients' quality of life and clinical outcomes. Early implementation of feeding jejunostomy can prevent nutritional deterioration and optimize recovery following gastrectomy[9].

We did not demonstrate any difference in survival (recurrence-free and overall survival) in our study. It is noteworthy that our findings align with those of a previous study in terms of survival outcomes and nutritional status following the implantation of a jejunostomy[10]. In this recent retrospective study including 125 patients who received preoperative chemoradiotherapy followed by surgery for esophageal cancer, the implementation of a feeding jejunostomy was associated with less weight loss during neoadjuvant treatment: 8 Lbs vs 13 Lbs (P = 0.003) between feeding jejunostomy tube (FJT) placement and no FJT patients, but was not associated with reduced toxicity due to neoadjuvant chemoradiotherapy or improved survival.

In the literature, few studies have compared the presence and absence of enteral feeding via jejunostomy in patients undergoing gastrectomy for cancer treatment. To date, the largest study is a retrospective study by Sun et al[11], which included 2980 patients, with the primary endpoint being the complication rate at 30 days[11]. Their results are similar to ours, with no difference in complication rate between the groups: 38.8% vs 36.1% (P = 0.32). Unlike our study, their study observed differences in nutritional levels, with greater recent weight loss among patients with jejunostomy: 21% vs 14.8% (P < 0.01). We used numerous variables to assess nutritional status: Weight, albuminemia, BMI, prealbuminemia, and PNI, which are all highly predictive markers of nutritional status[7,12,13]. The absence of differences in preoperative nutritional status between the groups allowed comparisons of interpretable postoperative results to be made.

We did not find any difference between the groups regarding the fistula rate, which was 3.3% in the J + group, compared to 5.4% in the J - group (P = 0.71). However, in the meta-analysis by Lewis et al[14], comparing early enteral feeding to fasting in patients undergoing gastrointestinal resection for cancer treatment, patients fed early enterally had fistula rates ranging from 2% to 7%, compared to 1% to 25% in fasting patients. In this meta-analysis, 7 out of 11 trials showed a reduced risk of anastomotic leakage in patients fed early enterally, with a relative risk in favour of enteral feeding (hazard ratio = 0.53; 95%CI: 0.26-1.08; P = 0.08). Of these 7 trials, 2 used jejunostomy feeding as a means of refeeding[15,16].

Our results showed a reduction in infectious complications in patients refed by jejunostomy: 1.2% vs 9.3% (P = 0.03). In one study, early enteral feeding was found to be a protective factor against overall infectious risk (relative risk = 0.72; 95%CI: 0.54-0.98; P = 0.036)[14]. This protective factor was found for parietal abscesses and the occurrence of pneumonia. Other trials have demonstrated that the preoperative nutritional status is correlated with postoperative infectious complications and that nutritional support helps reduce this excess morbidity in patients operated on for cancer of the gastrointestinal tract[17,18]. Another multicentre trial including 837 patients found contradictory results[19] with the meta-analysis by Lewis et al[14]. The authors revealed more infectious complications in the jejunostomy group (36% vs 19%; P < 0.01), whether surgical site infections (14% vs 6%; P < 0.01) or intra-abdominal infections (11% vs 4%; P < 0.01). However, this study presented numerous confounding biases: Patients with jejunostomy had a greater incidence of weight loss, a lower BMI, more extensive resections, and more advanced tumor node metastasis classification.

Interestingly, our study found a trend towards fewer respiratory complications, which in Lewis et al’s meta-analysis[14] translated into less pneumonia. Parallel to this observation, the postoperative nutritional status of patients refed by jejunostomy tended to be better than that of patients without jejunostomy. It is very likely that there is a beneficial nutritional effect on the diaphragmatic muscle, providing better postoperative ventilation. Indeed, stomach cancer is one of the most common causes of sarcopenia11[20]. However, this sarcopenia, a consequence of protein-energy malnutrition, is an independent factor in surgical complications after total gastrectomy[21]. It is therefore not surprising that by improving nutritional status, we reduce sarcopenia, including diaphragmatic sarcopenia, thus reducing ventilatory disorders and the resulting complications. In the literature, this improvement in nutritional status was found in a pilot study whose objective was to evaluate the impact of enteral nutrition by jejunostomy at home 6 weeks after esophagectomy or gastrectomy for cancer treatment[22]. Patients without jejunostomy lost an average of 3.9 kg (95%CI: 1.6-6.2) more than patients with jejunostomy, with the persistence of this difference at 3 and 6 months. In a more recent retrospective study including 715 patients who underwent gastrectomy for cancer, the authors did not find a significant difference in postoperative complications between patients with a FJT and those without. However, nutritionally, at one month postoperatively, patients with a FJT had higher mean serum albumin levels: 39.6 vs 36.1 g/L (P = 0.005), prealbumin levels: 181 vs 146 mg/L (P < 0.001), hemoglobin levels: 117.8 vs 112 g/L (P = 0.14), Onodera's prognostic nutrition index: 47.9 vs 45.6 (P = 0.03), and a lower neutrophil-to-lymphocyte ratio: 1.96 vs 2.5 (P = 0.005)[23].

Our study must be interpreted in light of its strengths and limitations. Indeed, we conducted a retrospective study prone to confounding biases. Misclassification biases may have arisen, particularly related to missing data, which could lead to over- or underestimation of patients' nutritional status. While all patients had at least one variable to define their nutritional status, overestimation may have occurred. Our study is a comprehensive cohort study that included all patients over 15 consecutive years, enriched notably by the higher incidence of stomach cancer in the Caribbean region. Prospective randomized trials are needed to confirm our findings.

CONCLUSION

A jejunostomy feeding tube placed intraoperatively after gastrectomy for cancer treatment helps reduce infectious and respiratory morbidity, as well as grade 3 complications, and significantly improves the postoperative nutritional status of patients.