Is there a benefit to preoperative weight loss in gastric bypass patients? A prospective randomized trial

BackgroundRoux-en-Y gastric bypass surgery is the leading surgical treatment of morbid obesity in the United States. The role of preoperative weight loss in gastric bypass surgery remains controversial. We performed a prospective randomized trial to determine whether preoperative weight loss results in better outcomes after laparoscopic gastric bypass.MethodsA total of 100 patients undergoing laparoscopic gastric bypass surgery from May 2004 to October 2005 were randomized preoperatively to either a weight loss group with a 10% weight loss requirement or a group that had no weight loss requirements. The patients were followed prospectively. The variables analyzed included perioperative complications, operative time, postoperative weight loss, and resolution of co-morbidities.ResultsData were available for 26 patients in the weight loss group and 35 in the nonweight loss group. The 2 groups had similar preoperative characteristics, conversion and complication rates, and resolution of co-morbidities. The initial body mass index was 48.7 kg/m² and 49.3 kg/m² for the weight loss group and nonweight loss group, respectively (P = NS). The preoperative body mass index was 44.5 kg/m² and 50.7 kg/m² for the weight loss group and nonweight loss group, respectively (P = 0.0027). The operative time was 220.2 and 257.6 minutes for the 2 groups (P = 0.0084). The percentage of excess weight loss at 3 and 6 months for the weight loss group and nonweight loss group was 44.1% and 33.1% (P = 0.0267) and 53.9% and 50.9% (P = NS), respectively. The interval to surgery from the initial consultation was 5.4 months and 5.2 months for the 2 groups (P = NS).ConclusionsPreoperative weight loss before laparoscopic Roux-en-Y gastric bypass was associated with a decrease in the operating room time and an improved percentage of excess weight loss in the short term. Preoperative weight loss, however, did not affect the major complication or conversion rates, and its long-term effects were not apparent through this study. Also, preoperative weight loss did not have any bearing on the resolution of co-morbidities.     

Outcomes and predictive factors of “not self-completion” in gastric endoscopic submucosal dissection for novice operators

Background

Endoscopic submucosal dissection (ESD) has become the standard endoscopic treatment for gastric neoplasms because of its safety and high rate of curability; however, it is not easy for novice operators to learn the technique of ESD. In this study, predictive factors of gastric neoplasms in which novices could not finish ESD by self-completion were evaluated.

Methods

Eighty consecutive ESD procedures performed by four novice operators were retrospectively analyzed. Standard ESD procedures were performed using an insulation-tipped (IT) knife under supervision. Self-completion rates, procedure time, and complete resection rates were evaluated, and predictive factors for “not self-completion” were assessed.

Results

The overall self-completion rate and en bloc plus R0 resection rate were 87.5% (70/80) and 95.7 % (67/70), respectively. In “not self-completion” cases (n = 10), the procedure time was longer and resected specimens were larger than those in self-completion cases (83.7 ± 47.3 min vs. 189.5 ± 106.8 min, p < 0.05; 641.2 ± 487.8 vs. 1,116 ± 480.4 mm², p < 0.01). Predictive factors of “not self-completion” were tumor location in the middle or upper third of the stomach or in the greater curvature and size of resected specimens larger than 900 mm². The self-completion rate of ESD was decreased in cases with more than two predictive factors.

Conclusions

For novice operators, tumor location and resected areas are predictive factors for failure to finish gastric ESD by self-completion. Selection of cancer lesions could be a key factor for effectiveness of ESD training.     

Is lymphadenectomy needed for all submucosal gastric cancers?

OBJECTIVE: To find out if it is feasible to extend the indication for local resection of submucosal gastric cancer without increasing the risk of lymph node metastases. DESIGN: Retrospective study. SETTING: University hospital, Japan. SUBJECTS: 104 patients with gastric cancer confined to the submucosal layer who underwent conventional gastrectomy with lymphadenectomy. INTERVENTIONS: The risk of nodal metastases was analysed retrospectively depending on the depth of submucosal invasion, size of the tumour, and other clinicopathological findings. MAIN OUTCOME MEASURES: The degree of submucosal invasion, size of the tumour, and incidence of lymph node metastasis. RESULTS: 15/104 patients (14%) had lymph node metastases. No patient in whom submucosal invasion was less than 500 microm or tumour was less than 15 mm in diameter developed lymph node metastases. Fewer patients had lymphatic permeation (37/89) and venous involvement (21/89) in the group without lymph node metastases. CONCLUSION: These data seem to support the hypothesis that early, minimally invasive, gastric cancer measuring < 15 mm in diameter could be treated by endoscopic mucosal or local resection, and gastrectomy with lymphadenectomy might be unnecessary.     

Is intraperitoneal levobupivacaine with epinephrine useful for analgesia following laparoscopic cholecystectomy? A randomized controlled trial

BACKGROUND AND OBJECTIVE: A randomized controlled trial of patients undergoing laparoscopic cholecystectomy. To evaluate whether the instillation of levobupivacaine, with epinephrine, intraperitoneally, reduces morphine consumption; further, to consider the degree of abdominal and shoulder pain in the presence of local anaesthetic at the site of surgical incision. METHODS: Patients received a standardized procedure and were allocated randomly to receive either (a) 2.5 mg mL(-1) levobupivacaine 30 mL, with epinephrine 5 microg mL(-1), or (b) normal saline 30 mL, with epinephrine 5 microg mL(-1), to the bed of the gall bladder and above the liver, immediately before wound closure. RESULTS: Median (interquartile range) total abdominal pain during inspiration in the levobupivacaine group was significantly (P = 0.041) lower (71 (21-129) mm) than that in the placebo group (123 (71-179) mm). However, there was no significant difference between the two groups with respect to total abdominal pain at rest, right shoulder pain or the consumption of rescue morphine and rescue dihydrocodeine. There were no differences between the two groups with respect to the administration of cyclizine, total nausea and total sedation scores. In addition, the number of episodes of vomiting was not significantly different. CONCLUSIONS: Intraperitoneal administration of levobupivacaine with epinephrine is associated with modest analgesia following laparoscopic cholecystectomy.     

Is there a role for antimuscarinics in renal colic? A randomized controlled trial

PURPOSE: We determined whether the addition of hyoscine butylbromide (Buscopan) reduces the amount of opioid analgesia required and the need for ongoing opioid analgesia in acute renal colic. MATERIALS AND METHODS: Patients with a clinical diagnosis of acute renal colic were prospectively randomized to receive Buscopan or placebo in addition to morphine and intravenous fluids, with or without indomethacin. Observations and pain scores were recorded 15 every minutes during the first hour then hourly thereafter. Morphine was administered in 2.5 mg increments until pain relief was achieved. Recordings were ceased at 4 hours or on discharge from the emergency department, whichever occurred first. The diagnosis of renal colic was confirmed by computerized tomography. The dose of morphine (mg/kg) and the proportion of patients who required further morphine were compared between the 2 groups. RESULTS: Of the 192 patients randomized, data were available for 178 on an intent to treat basis. A total of 85 patients received Buscopan while 93 received placebo in addition to standard therapy. Patients in the Buscopan group required a median of 0.12 mg/kg of morphine while those in the placebo group received a median of 0.11 mg/kg, and this difference was not significant (p =0.4). There was also no significant difference in the proportion of patients who required additional morphine, 33% in the Buscopan group and 38% in the placebo group (p =0.5). Subgroup analysis of the 138 patients with a confirmed stone also showed no difference between the groups. CONCLUSIONS: There is no evidence that Buscopan reduces opioid requirements or the need for ongoing opioid analgesia in acute renal colic.     

Colorectal endoscopic submucosal dissection for elderly patients at least 80?years of age

Background

Endoscopic submucosal dissection (ESD) has been used recently for successful en bloc resection of even large lesions, although no consensus appears in medical literature concerning its application to elderly patients. This prospective cohort study aimed to evaluate the efficacy and safety of colorectal ESD for patients 80?years of age or older.

Methods

Colorectal ESD procedure findings were compared with clinical outcomes, including associated complications and mortalities, for two age groups totaling 196 consecutive patients with 202 colorectal lesions. Of the 196 patients, 31 patients (16%) were 80?years of age or older (group E), and 165 patients (84%) were younger than 80?years (group Y).

Results

The median ages were 82?years in group E and 68?years in group Y. The frequency of chronic concomitant diseases was significantly higher in group E (65%) than in group Y (27%) (p?=?0.003). No significant pressure decrease or need for oxygenation was observed in either group. In addition, groups E and Y did not differ significantly in terms of mean lesion sizes (40.9 vs. 39.7?mm) en bloc resection rates (84% vs. 93%), curative rates (78% vs. 84%), median procedure times (65 vs. 70?min), or associated complications (no perforation or delayed bleeding cases [0%] vs. 5 perforations [3%]) The median postprocedure hospitalization period was 3?days in both groups. Except for 10 cases requiring subsequent lymph node dissection surgery, follow-up colonoscopy examinations showed no recurrences or ESD-related mortalities in either group.

Conclusion

Colorectal ESD is a safe and effective treatment for elderly patients (age????80?years) despite a significantly higher frequency of chronic concomitant diseases than among younger patients.     

Are drains useful for lumbar disc surgery? A prospective, randomized clinical study

OBJECTIVE: In this prospective, observer-masked clinical study, we evaluated if insertion of a drain had a significant role in decreasing the existence and the size of postoperative epidural hematoma, which is believed to be a factor causing epidural fibrosis in patients undergoing lumbar discectomy. METHOD: Fifty patients undergoing lumbar disc surgery were randomly assigned to two groups: with or without insertion of a drain in the epidural space. A drain was inserted in 22 patients, whereas 28 were left without a drain. All patients were evaluated, by means of magnetic resonance imaging (MRI) on the first postoperative day, specifically looking for the existence and the size of epidural hematoma. The size of epidural hematoma was graded as none, minimal, moderate, or prominent. The patients were clinically examined preoperatively and at the follow-up of 6 months by Oswestry Disability Index and recurrence of pain. A follow-up MRI was repeated at 6 months, and the subsequent development of epidural fibrosis was evaluated. RESULTS: Epidural hematoma was detected in 36% of patients with a drain and in 89% of patients without a drain (P=0.000). There were significant less number of minimum, moderate, and prominent sized hematomas in the group with a drain (P=0.000). On the 6-month follow-up, epidural fibrosis was found in 58.3% of patients without a drain and in 31.6% of patients with a drain (P=0.08). Late clinical outcome (improvement in Oswestry Index and no recurrent pain) was better in the group with drain, but not statistically significant (P=0.4). CONCLUSIONS: Occurrence of hematoma in the epidural space is common after lumbar disc surgery even if meticulous hemostasis has been achieved. Insertion of a drain decreases both the incidence and the size of hematoma on the first postoperative day as detected by MRI. This may have practical implications for the prevention of significant postoperative fibrosis and obtaining better surgical outcome.     

Is extended lymph node dissection necessary for gastric cancer in elderly patients?

OBJECTIVE: To compare the outcome after limited and extended gastric resections to find out whether extended lymph node dissection is indicated for gastric cancer in elderly patients. DESIGN: Retrospective study. SETTING: University hospital, Japan. SUBJECTS: 182 patients over 75 years of age with gastric cancer who had gastric resections from 1980 to 1995. INTERVENTIONS: 161 patients had limited lymph node dissection (limited group) and 21 had extended lymph node dissection (extended group). MAIN OUTCOME MEASURES: Histopathological features, morbidity, mortality, and long-term survival. RESULTS: Postoperative morbidity was 27% (n = 44) in the limited group and 57% (n = 12) in the extended group, and postoperative mortality was 1% (n = 2) in the limited group and 10% (n = 2) in the extended group; these differences are significant (p = 0.005 and p = 0.002). The 5-year survival did not differ significantly between the two groups. Only the T classification and presence of lymph node metastases had a significant influence on the outcome of gastric cancer in elderly patients. CONCLUSIONS: The presence of lymph node metastases is a critical factor in the prognosis of gastric cancer, and extended lymph node dissection has therefore been recommended. However, extended lymph node dissection in elderly patients did not influence the 5-year survival; in addition, the mortality and morbidity in the extended group were higher than in the limited group. Extended lymph node dissection is therefore usually not indicated for gastric cancer in elderly patients.     

Is a thyroid follicular neoplasm a good indication for endoscopic surgery?

Endoscopic thyroidectomy is a safe and feasible alternative as compared with conventional open thyroidectomy in patients with a small thyroid cancer or a benign thyroid tumor. However, despite the many advantages of endoscopic surgery, it can result in unexpected complications. Recently, the authors experienced a case of follicular thyroid cancer recurrence that developed around the operative bed and along the port insertion site after endoscopic thyroidectomy for a large follicular neoplasm. The authors suggest that a smaller follicular neoplasm is a good indication for endoscopic thyroidectomy, but that a large follicular neoplasm should not be viewed as indicators for endoscopic surgery, because of the possibility of malignancy and rupture during manipulation.     

Is bowel preparation before excretory urography necessary? A prospective, randomized, controlled trial

PURPOSE: A prospective, randomized, single blind, controlled trial was performed to determine whether routine bowel preparations are necessary for satisfactory visualization of the urinary system during IVP and compare the side effects of several bowel preparation methods. MATERIALS AND METHODS: A total of 200 patients between 17 and 70 years old at the outpatient department who were prescribed IVP were enrolled in this prospective study. Patients were divided into 4 groups and randomly received liquid supper, senna, magnesium sulfate solution and PEG-ELS as bowel preparation before IVP. Control and contrast images were assessed by a radiologist and a urologist blinded to the bowel preparation. IVP was given a visualization score, a fecal residue score, a bowel gas score and an overall quality score. Images were also divided into anatomical sections to assess visualization, and the effect of fecal residue and bowel gas. Questionnaires were prepared about the side effects of bowel preparation. RESULTS: For contrast images there was no significant difference in terms of visualization,, fecal residue, bowel gas and overall quality scores among the 4 groups. The same result was seen on the control image except for fecal residue scores, when the senna and PEG-ELS groups had a lower fecal residue score than the liquid supper group (p = 0.001). Interestingly when considering sectional images, the right renal region of the control image showed that the senna and PEG-ELS groups had better visualization scores with less bowel gas and fecal residue than the liquid supper only group. Patients in all 3 laxative bowel preparation groups except the liquid supper group reported side effects. There was no statistically significant difference among these 3 groups in the attitude toward ingesting laxatives again. CONCLUSIONS: Our results suggest that routine bowel preparation prior to IVP does not improve the visibility or overall quality of control or contrast images, while patients experience certain side effects due to the bowel preparation. The laxative bowel preparation may be helpful only when visualizing the right renal region on the control image.     

Is ureteral stenting necessary after uncomplicated ureteroscopic lithotripsy? A prospective, randomized controlled trial

PURPOSE: We conducted a prospective, randomized controlled study to investigate the advantages and disadvantages of ureteral stenting after ureteroscopic lithotripsy. MATERIALS AND METHODS: A total of 60 patients who underwent ureteroscopic lithotripsy were equally randomized into a stented or a nonstented group. The inclusion criteria were stone 6 to 10 mm., absence of polyp or stricture in the ureter and no mucosal injury during ureteroscopy. The operation was performed with a 6Fr rigid ureteroscope without ureteral dilation and stones were fragmented with a 1.9Fr electrohydraulic lithotriptor without extraction. A 7Fr double pigtail stent was placed in the stented group for 3 days after ureteroscopy. Urinalysis, plain x-ray and renosonography were performed before and after lithotripsy in each patient. Subjective symptoms and pain score were recorded on admission to the hospital and 3 days postoperatively. RESULTS: The stone-free rate was 100% in each group and preoperative hydronephrosis equally resolved in both groups. Mean pain score plus or minus standard deviation improved significantly in the nonstented (6.33 +/- 1.81 preoperatively to 2.30 +/- 1.93 postoperatively, paired Student's t test p <0.0001) and stented (7.10 +/- 1.03 to 2.30 +/- 2.22, p < 0.0001) group. There was no statistical difference in pain reduction between the 2 groups (p = 0.18). The amount of extra parenteral analgesic used was similar in both groups. One patient in the nonstented group visited the emergency room for postoperative renal colic, 25 (83.3%) patients in the stented group complained of at least 1 irritative bladder symptom and only 4 (13.3%) in the nonstented group experienced bladder discomfort. CONCLUSIONS: After uncomplicated ureteroscopic electrohydraulic lithotripsy patients without ureteral stenting tend to have similar renal function recovery and satisfactory pain reduction with less irritative symptoms compared to those treated with a ureteral stent. We suggest that it is not necessary to place a ureteral stent routinely after uncomplicated ureteroscopic electrohydraulic lithotripsy for stones smaller than 1 cm.     

Is early oral feeding safe after elective colorectal surgery? A prospective randomized trial.

INTRODUCTION: The routine use of a nasogastric tube after elective colorectal surgery is no longer mandatory. More recently, early feeding after laparoscopic colectomy has been shown to be safe and well tolerated. Therefore, the aim of our study was to prospectively assess the safety and tolerability of early oral feeding after elective "open" abdominal colorectal operations. MATERIALS AND METHODS: All patients who underwent elective laparotomy with either colon or small bowel resection between November 1992 and April 1994 were prospectively randomized to one of the following two groups: group 1: early oral feeding--all patients received a clear liquid diet on the first postoperative day followed by a regular diet as tolerated; group 2: regular feeding--all patients were treated in the "traditional" way, with feeding only after the resolution of their postoperative ileus. The nasogastric tube was removed from all patients in both groups immediately after surgery. The patients were monitored for vomiting, bowel movements, nasogastric tube reinsertion, time of regular diet consumption, complications, and length of hospitalization. The nasogastric tube was reinserted if two or more episodes of vomiting of more than 100 mL occurred in the absence of bowel movement. Ileus was considered resolved after a bowel movement in the absence of abdominal distention or vomiting. RESULTS: One hundred sixty-one consecutive patients were studied, 80 patients in group 1 (34 males and 46 females, mean age 51 years [range 16-82 years]), and 81 patients in group 2 (43 males and 38 females, mean age 56 years [range 20-90 years]). Sixty-three patients (79%) in the early feeding group tolerated the early feeding schedule and were advanced to regular diet within the next 24 to 48 hours. There were no significant differences between the early and regular feeding groups in the rate of vomiting (21% vs. 14%), nasogastric tube reinsertion (11% vs. 10%), length of ileus (3.8 +/- 0.1 days vs. 4.1 +/- 0.1 days), length of hospitalization (6.2 +/- 0.2 days vs. 6.8 +/- 0.2 days), or overall complications (7.5% vs. 6.1%), respectively, (p = NS for all). However, the patients in the early feeding group tolerated a regular diet significantly earlier than did the patients in the regular feeding group (2.6 +/- 0.1 days vs. 5 +/- 0.1 days; p < 0.001). CONCLUSION: Early oral feeding after elective colorectal surgery is safe and can be tolerated by the majority of patients. Thus, it may become a routine feature of postoperative management in these patients.     

Is routine ureteric stenting needed in kidney transplantation? A randomized trial

BACKGROUND: Whether routine ureteric stenting in low-urological-risk patients reduces the risk of urological complications in kidney transplantation is not established. METHODS: Eligible patients were recipients of single-organ renal transplants with normal lower urinary tracts. Patients were randomized intraoperatively to receive either routine stenting or stenting only in the event of technical difficulties with the anastomosis. All patients underwent Lich-Gregoire ureteroneocystostomy. RESULTS: Between June 1994 and December 1997, 331 kidney transplants were performed at a single center, 305 patients were eligible, and 280 patients were enrolled and randomized. Donor and recipient age, sex, donor source, whether first or subsequent grafts, ureteric length, native renal disease, and immunosuppression were similar in each group. In the no-routine-stenting group 6 of 137 patients (4.4%) received stents after randomization for intraoperative events that in the surgeon's opinion required use of a stent. In an intention-to-treat analysis there was no difference between groups in the primary outcome cluster of obstruction or leak [routine stenting 5 of 143 (3.5%) vs. no routine stenting 9 of 137 (6.6%); P=0.23], or in either of these complications analyzed separately. All urological complications were successfully managed without major morbidity. Living donor organs and shorter ureteric length (after trimming) were univariate risk factors for leaks, although increasing donor age was associated with obstruction. CONCLUSIONS: Routine ureteric stenting is unnecessary in kidney transplantation in patients at low risk for urological complications. Careful surgical technique with selective stenting of problematic anastomoses yields similar results.     

Is there a role for drain use in elective laparoscopic cholecystectomy? A controlled randomized trial

Background

Although the issue of drain use in open cholecystectomy has been adequately addressed by prospective randomized trials, there is lack of evidence on the usefulness of drains in elective laparoscopic cholecystectomy, and the surgeons follow their beliefs and bias on this debate. Therefore, a controlled randomized trial was designed to assess the value of drains in elective laparoscopic cholecystectomy.

Methods

During a 5-year period (January 2002 to December 2006), 284 patients were randomized to have a drain placed (group A), whereas 281 patients were randomized not to have a drain (group B) placed in the subhepatic space. End points of this trial were to detect any differences in morbidity, postoperative pain, and hospital stay between the 2 groups.

Results

There was no mortality in either group and no statistically significant difference in morbidity or hospital stay between the 2 groups. However, postoperative pain was significantly increased in patients who had a drain placed; median visual analog scale (VAS) score was 5 (range 1 to 8) versus 3 (range 1 to 8), in the non-drained group (P < .0001). Interestingly, in 2 of 3 patients in whom a drain was placed against randomization because of bile leak suspicion, a bile leak occurred.

Conclusions

The routine use of a drain in elective laparoscopic cholecystectomy has nothing to offer; in contrast, it is associated with increased pain. It would be reasonable, however, to leave a drain if there is a worry about an unsolved or potential bile leak, bearing in mind that drain placement, although sometimes providing a false sense of security, does not guarantee either prevention or treatment of postoperative bile collections, bleeding, or bile peritonitis.     

Is mechanical bowel preparation mandatory for elective colon surgery? A prospective randomized study

BACKGROUND: Bowel preparation prior to colonic surgery usually includes antibiotic therapy together with mechanical bowel preparation (MBP). Mechanical bowel preparation may cause discomfort to the patient, prolonged hospitalization, and water and electrolyte imbalance. It was assumed that with the improvement in surgical technique together with the use of more effective prophylactic antibiotics, it was possible that MBP would no longer be necessary. HYPOTHESIS: There is no statistical difference in the postoperative results of patients who undergo elective colon resection with MBP as compared with those who have no MBP.Design and PATIENTS: The study includes all patients who had elective large bowel resection at Campus Golda between April 1, 1999, and March 31, 2002. Emergency operations were not included. The patients were randomly assigned to the 2 study groups (with or without MBP) according to identification numbers. All patients were treated with intravenous and oral antibiotics prior to surgery. The patients in the MBP group received Soffodex for bowel preparation. RESULTS: A total of 329 patients participated in the study, 165 without MBP and 164 with MBP. The 2 groups were similar in age, sex, and type of surgical procedure. Two hundred sixty-eight patients (81.5%) underwent surgery owing to colorectal cancer and 61 patients (18.5%) owing to benign disease. The hospitalization period was longer in the bowel-prepared group (mean +/- SD, 8.2 +/- 5.1 days) as compared with the nonprepared group (mean +/- SD, 8.0 +/- 2.7 days). However, this difference was not statistically significant. The time until the first bowel movement was similar between the 2 groups: a mean +/- SD of 4.2 +/- 1.3 days in the nonprepared group as compared with a mean +/- SD of 4.3 +/- 1.1 days in the prepared group (P = NS). Four patients (1.2%) died in the postoperative course owing to acute myocardial infarction and pulmonary embolism. Sixty-two patients (37.6%) of the non-MBP group suffered from postoperative complications as compared with 77 patients (46.9%) of the MBP group. CONCLUSION: Our results suggest that no advantage is gained by preoperative MBP in elective colorectal surgery.     

Is D2 lymph node dissection necessary for early gastric cancer?

Background The objective of this study was to clarify a survival benefit of D2 lymphadenectomy in patients with early gastric carcinoma (GC). Methods A retrospective study was conducted to examine the incidence of metastasis to level 2 lymph nodes, the causes of postoperative death, and the mode of recurrence in 1041 patients who head early GC and underwent D2 lymphadenectomy with curative intent. Results Postoperative mortality occurred in 129 (12,4%) of 1041 patients, 6 patients (.6%) died of surgical complications 108 (10.2%) died of diseases other than cancer, and 16 (1.5%) died of recurrence. Hematogenous metastasis was the major mode of recurrence (56.3% of recurrences). The incidence of metastasis to level 2 nodes was 2.5% (26 of 1041 patients, 18 of whom were alive). Thus, the estimated survival benefit of radical lymphadenectomy for patients with early GC was calculated to be 1.7% (18 of 1041 patients). Conclusions D2 lymphadenectomy in patients with early GC had little survival benefit because (1) metastasis to level 2 nodes was rare, (2) most causes of death were not related to the tumor, and (3) more than half the recurrences were hematogenous. Use of radical lymphadenectomy for early GC should be limited.