Flow of participants through the study. NJT indicates nasojejunal tube.
Doglietto GB, Papa V, Tortorelli AP, Bossola M, Covino M, Pacelli F, Italian Total Gastrectomy Study Group. Nasojejunal Tube Placement After Total GastrectomyA Multicenter Prospective Randomized Trial. Arch Surg. 2004;139(12):1309-1313. doi:10.1001/archsurg.139.12.1309
Anastomotic disruption of the Roux-en-Y esophagojejunostomy after total gastrectomy is an infrequent complication that may lead to severe morbidity and mortality. Consequently, a nasojejunal tube (NJT) is frequently placed when this operation is performed. However, no studies have compared routine vs no placement of an NJT in patients undergoing total gastrectomy for gastric cancer, to our knowledge.
Randomized controlled trial to assess the need for routine nasojejunal decompression after total gastrectomy with Roux-en-Y esophagojejunostomy in patients with gastric cancer.
Tertiary care centers.
Two hundred thirty-seven patients undergoing total gastrectomy for gastric cancer were randomly assigned to NJT placement (NJT group) or not (no-NJT group). The patients were monitored for postoperative complications, mortality, and postoperative course.
Main Outcome Measures
Incidence of esophagojejunostomy leak.
The rates of anastomotic leak were similar in both groups (6.9% and 5.8% for the NJT group and no-NJT group, respectively; P = .71), as were the rates of major postoperative complications (25.9% and 21.5%, respectively; P = .42) and overall postoperative mortality (0.9% and 0.8%, respectively; P = .50). There were no differences between the 2 groups in the mean±SD time to passage of flatus (4.6 ± 1.3 and 4.5 ± 1.7 days, respectively) or to starting a liquid diet (7.8 ± 2.6 and 7.7 ± 1.6 days, respectively), postoperative length of hospital stay (13.5 ± 7.3 and 13.9 ± 10.9 days, respectively), postoperative pain, or postoperative abdominal distention.
Routine placement of an NJT after Roux-en-Y esophagojejunostomy is unnecessary in elective total gastrectomy for gastric cancer.
Although anastomotic disruption of the Roux-en-Y esophagojejunostomy after total gastrectomy has become in recent years an infrequent complication, it may lead to severe morbidity and mortality1- 3; consequently, a nasojejunal tube (NJT) is frequently placed when this operation is performed. However, to our knowledge, no studies have compared routine vs no placement of an NJT in patients undergoing total gastrectomy for gastric cancer.
The present study reports the results of a cooperative, multicenter randomized clinical trial designed to assess the need for routine nasojejunal decompression after total gastrectomy with Roux-en-Y esophagojejunostomy in patients with gastric cancer. The primary objective of this trial was to determine whether the placement of an NJT reduces the incidence of esophagojejunostomy leak in such patients.
The protocol of the study was approved by the ethical committees of the participating centers. Informed consent was obtained from the patients before entering the study.
All patients between the ages of 18 and 80 years who were admitted to surgical units of the participating centers from June 1, 2001, to December 31, 2002, and who were candidates to undergo total gastrectomy for carcinoma of the stomach were potentially eligible for the study. Forty-four patients who underwent only explorative laparotomy (29 patients), subtotal gastrectomy (4 patients), or bypass procedures (11 patients) were excluded from the study (Figure). The patients who entered the study were assigned at the end of operation, before closing the abdomen, to placement of an NJT (NJT group) or not (no-NJT group) by computer-generated random numbers. The patients in the NJT group underwent placement of a 12F to 14F single-lumen NJT. The continuity of the alimentary tract was restored with an end-to-side mechanical Roux-en-Y esophagojejunostomy (25-mm circular stapling device), while the extension of lymphadenectomy was at the surgeon’s discretion. All patients received short-term perioperative antibiotic prophylaxis with a third-generation cephalosporin and a subcutaneous injection of low-molecular-weight heparin sodium (3000 IU) as deep venous thrombosis prophylaxis.
During the postoperative course, all patients received parenteral fluid administration until oral intake was resumed. No patients in the NJT group received postoperative enteral nutrition. Routine radiographic examination using water-soluble contrast material was performed on the seventh postoperative day, and, in the absence of anastomotic leak, a liquid diet was started. If oral intake was not resumed after 7 days or if a complication occurred portending prolonged restriction of oral intake, enteral nutrition or total parenteral nutrition could be instituted by the surgeons of the participating centers according to the availability of access and the type of complication. The NJT was inserted or reinserted when clinically indicated. Demographic data, pathological variables, operative factors, preoperative risk factors, and associated pathologic conditions were recorded.
The patients were monitored for postoperative complications, mortality, and postoperative course by physicians not associated with the surgical teams. Complications were classified by objective criteria as major or minor and as infectious or noninfectious (Table 1) according to a previously described classification.4
The postoperative days when the first passage of flatus was observed and when a liquid diet was permitted were recorded, as well as the postoperative length of hospital stay. Also recorded were postoperative pain, subjectively expressed by patients on the third and sixth postoperative days according to a visual analog scale from 0 (no pain) to 10 (worst imaginable pain), and postoperative abdominal distention (measured at the umbilical level on the third postoperative day) compared with the preoperative value. Finally, the postoperative courses of patients with esophagojejunostomy leak were analyzed to assess the evolution of the complication.
The primary objective for comparison was the incidence of esophagojejunostomy leak. Based on previous investigations performed on our surgical population,5 we anticipated a 10% incidence of esophagojejunostomy leak in the no-NJT group. A reduction of this rate by half (to 5%) in the NJT group would be considered clinically important.
Detecting a difference of this magnitude or greater, with a level of statistical significance of P<.05 and a power of 80% with a 2-tailed test of proportions, would require 112 patients in each group. Therefore, the goal for the accrual of patients was 230 patients for the final outcome analysis.
Continuous variables were compared by analysis of variance and categorical variables by the Fisher exact test. Analysis of categorical covariates was performed using the Mantel-Haenszel technique. All statistical analyses were 2-tailed and based on the intention-to-treat concept (ie, patients randomized for placement of an NJT in whom a dislocation of the tube occurred postoperatively were analyzed in the NJT group).
During the 19-month patient accrual, 237 patients recruited at 17 university and hospital surgical departments who underwent total gastrectomy for gastric cancer were randomly assigned to placement of an NJT (n=116) or not (n=121). The 2 groups were similar regarding demographic data, tumor location, and American Joint Committee on Cancer/Union Internationale Contre le Cancer TNM stage (sixth edition), as presented in Table 2. All patients in the 2 groups underwent total gastrectomy for gastric cancer, and the main operative factors are listed in Table 3.
One (0.9%) of 116 patients assigned to the NJT group and 1 (0.8%) of 121 patients assigned to the no-NJT group died during the postoperative period, both experiencing an esophagojejunal anastomosis leak. The difference was not statistically significant (P = .50).
The rates of anastomotic leak were similar in both groups (NJT group, 6.9% and no-NJT group, 5.8%; P = .71). The rates of major postoperative complications were also similar: 30 (25.9%) of 116 patients had nasojejunal decompression and 26 (21.5%) of 121 patients in the no-NJT group had such complications (P = .43). The rates of individual complications and the relative risks and confidence intervals are presented in Table 4.
The rates of major infectious complications were similar in the 2 groups: 14 (12.1%) of 116 patients in the NJT group and 13 (10.7%) of 121 patients in the no-NJT group had such complications. There was a slightly higher incidence of pneumonia in the NJT group: 14 (12.1%) of 116 patients in the NJT group and 10 (8.3%) of 121 patients in the no-NJT group experienced this complication (P = .33). Similarly, there were no differences between the NJT and no-NJT groups in major noninfectious complications, the rates being 13.8% and 10.7%, respectively.
The rates of minor postoperative infectious complications and noninfectious complications were comparable in the NJT and no-NJT groups. The postoperative length of hospital stay, time to passage of flatus, and time to starting a liquid diet were similar in the 2 groups, as were the postoperative pain and postoperative abdominal distention (Table 5).
The complications related to placement of an NJT were tube dislocation (9 cases), digestive bleeding (2 cases), nasal mucosa necrosis (2 cases), conspicuous (>250 mL/d) intestinal fluid loss (2 cases), and rhinopharyngitis (1 case). One patient in the no-NJT group required insertion of an NJT on the second postoperative day because of recurrent episodes of vomiting, with resolution of the symptom. All patients in the no-NJT group with esophagojejunostomy leak underwent placement of an NJT after the diagnosis of the complication.
The postoperative course of patients with esophagojejunostomy leak was similar in both groups, with 1 death in each group and complete resolution of the complication in the other cases. Concerning the management of the esophagojejunostomy leak, 2 patients (1 in each group) needed reoperation, while the other patients were treated conservatively. The mean±SD time to resolution of the leak was similar in the 2 groups, 20.7 ± 14.3 and 21.1 ± 13.5 days for the NJT and no-NJT groups, respectively (P = .82).
Previous studies6- 12 have shown that routine nasogastric decompression after gastrointestinal surgery does not decrease postoperative morbidity and mortality. In contrast, some studies10,13 have shown that the use of a nasogastric tube may result in an increased incidence of pulmonary complications, including atelectasis and pneumonia.
Prophylactic nasojejunal decompression after total gastrectomy is considered differently from decompression elsewhere in the abdomen because anastomotic disruption of the Roux-en-Y esophagojejunostomy is a complication that may lead to severe morbidity and mortality; therefore, nasojejunal decompression has become a surgical dogma. Despite the lack of studies concerning this topic, many surgeons continue to use an NJT after total gastrectomy, believing that its use permits the decompression of the digestive tract after the surgical procedure, especially when a D2 complete lymphadenectomy is performed; in fact, the sectioning of sympathetic and parasympathetic nerve fibers during skeletonization of the celiac axis and truncal vagotomy may severely impair intestinal motility.14,15 Moreover, in case of esophagojejunostomy disruption, the presence of an NJT may be useful in the management of a leak. Finally, an NJT may be used in malnourished patients for administering early postoperative enteral feeding.
Regarding the nutritional use of the NJT, although some authors claim that early postoperative enteral nutrition may be advantageous in malnourished patients,16 it has been shown in a recently published study17 that enteral feeding following major abdominal surgery does not reduce postoperative complications and mortality compared with parenteral nutrition. Therefore, the use of an NJT for nutritional purposes in malnourished patients remains a choice based on the surgeon’s preference rather than a mandatory procedure.
Concerning the role of the NJT in the incidence of esophagojejunostomy leak and in the postoperative course of patients with this complication, no studies have compared routine vs no placement of an NJT in patients undergoing total gastrectomy for gastric cancer, to our knowledge. For that reason, a prospective multicenter randomized trial was conducted to assess the need for routine nasojejunal decompression after total gastrectomy with Roux-en-Y esophagojejunostomy. In our study, the incidence of postoperative complications, including esophagojejunostomy leak and mortality, was similar in the 2 groups, as were the postoperative courses. In contrast, patients in the NJT group experienced a slightly higher incidence of pulmonary complications. Moreover, the incidence of complications related to NJT placement was not negligible. Therefore, the results of the present study suggest that routine placement of an NJT after Roux-en-Y esophagojejunostomy is unnecessary in elective total gastrectomy for gastric cancer.
Correspondence: Fabio Pacelli, MD, Digestive Surgery Unit, Department of Surgical Sciences, Catholic University School of Medicine, Via della Mendola 47, 00135 Rome, Italy (firstname.lastname@example.org).
Accepted for Publication: June 30, 2004.
Group Members: The Italian Total Gastrectomy Study Group includes the authors and the following investigators (listed according to the number of recruited cases): Roberto Coppola, Pierfilippo Crucitti, and Domenico Borzomati, Università Campus Bio-Medico, Roma; Aurelio Pic ciocchi and Domenico D’Ugo, Università Cattolica, Roma; Roberto Tersigni and Livio Paganelli, Ospedale S Camillo-Forlanini, Roma; Ignazio M. Civello and Giuseppe Brisinda, Università Cattolica; Andrea Valeri and Pietro Tonelli, Ospedale Careggi, Firenze; Giovanni Natalini and Francesco Guiggi, Ospedale di Marsciano e Todi; Alberto Del Genio, Vincenzo Maffettone, and Gianluca Russo, II Università degli Studi di Napoli; Eduardo Landi and Walter Siquini, Università di Ancona; Vincenzo Memeo and Michele De Fazio, Università di Bari; Giuseppe S. Bondanza and Enrico Ciferri, Ospedale di Genova-Pontedecimo; Piero Chirletti, Università La Sapienza, Roma; Eugenio Santoro and Massimo Carlini, Istituto Regina Elena, Roma; Giuseppe Pappalardo, Università La Sapienza; Giovan B. Grassi and Marco Catarci, Ospedale S Filippo Neri, Roma; Davide F. D’Amico and Riccaro Ranzato, Università di Padova; and Filippo Ferrarese and Giovanni D’Eredità, Università di Bari.