- Email: [email protected]
Total gastrectomy in the treatment of advanced gastric cancer
Total Gastrectomy John A. Butler,
MD,
in the Treatment of Advanced Gastric Cancer
Terry J. Dubrow, MD, Thomas Trezona, MD, Michael Klassen, Robert J. Nejdl, MD, Torrance, California
To assess the role of total gastrectomy in the treatment of advanced gastric cancer, we reviewed the records of 27 patients who underwent the procedure from 1979 to 1988. Operative mortality was Ppercent (1 of 27), and postoperative morbidity occurred in 48 percent of the patients. Twenty-five of 26 patients were tolerating solid food at the time of discharge; 21 were able to maintain oral alimentation until just prior to their death. Median survival following the operation was 15 months (range: 2 to 110 months), with a 62 percent absolute l-year survival rate and a 38 percent 2-year survival rate. On the basis of these results, we conclude that in patients with advanced gastric carcinoma, total gastrectomy with Roux-Y esophagojejunostomy can he performed with an acceptable morbidity and mortality, provides significant palliation by restoring the patient’s ability to eat, and should be performed when technically feasible, even in the presence of gross residual disease.
lthough the incidence of gastric cancer has declined dramatically over the past 50 years, it is the 7th leading cause of death b cancer in men and the 9th leading cause in women [13r. Despite the successful results achieved in the treatment of superficial or “early” gastric cancer, most patients present with advanced disease [2]. Multiple reports have documented a lack of efficacy of total gastrectomy as a curative procedure when compared with subtotal resections [3]. The increased morbidity and mortality associated with the procedure, together with the nutritional difficulties encountered in the subsequent management of these patients with advanced disease, has led many physicians to conclude that total gastrectomy is contraindicated for locally advanced gastric cancer. We performed this study to review our recent experience with total gastrectomy and determine its role in the management of these patients.
A
From the Departments of Surgery, Harbor-UCLA Medical Center, Torrance, California, and Kaiser Medical Center, Harbor City, California. Requests for reprints should be sent to John A. Butler, MD, Harbor-UCLA Medical Center, Box 25, 1000 West Carson Street, Torrance, California 90502. Presented at the 41st Annual Meeting of the Southwestern Surgical Congress, Monterey, California, April 23-26,1989.
602
THE AMERICAN JOURNAL OF SURGERY
PATIENTS
MD,
AND METHODS
The tumor registries of Los Angeles County HarborUCLA Medical Center and Kaiser-Harbor City Hospital were reveiwed over a IO-year period from 1979 to 1988. Records of all patients who underwent a total gastrectomy for cancer were reviewed. Patients who had either a near-total gastrectomy or proximal esophagogastrectomy were excluded from the analysis. Patients with a histologic diagnosis of squamous cell carcinoma or malignant lymphoma were also excluded. Twenty-seven patients with gastric adenocarcinoma who underwent total gastrectomy were identified. Their hospital records were reviewed for demographic data, past history, symptoms, tumor size and location, and pathologic characteristics. The operative procedure, with particular reference to the method of restoring intestinal continuity, postoperative morbidity and mortality, additional therapy, and overall survival, was also examined. A patient who died without leaving the hospital after surgery was classified as an operative mortality. Complete follow-up until death or the time of this report was documented for 26 (96 percent) patients. One patient was lost to follow-up 39 months postgastrectomy, at which time he was believed to be free of disease. Crude, uncorrected, absolute survival times are used in view of the fact that only one patient died of causes unrelated to gastric cancer. RESULTS
Twenty-seven patients underwent total gastrectomy during the lo-year period reviewed. Eighteen patients were men and 9 were women. The median age of the patients was 63, with a range from 21 to 93 years. Seven of the 27 (26 percent) were over 70 years of age. Three patients had a gastric resection for peptic ulcer disease 12, 25, and 37 years prior to the development of gastric cancer; one other patient had a prior vagotomy and pyloroplasty. Seventeen of the 27 patients (63 percent) presented with obstructive symptoms, nine of whom could tolerate only liquids. Ten patients (37 percent) presented with evidence of upper gastrointestinal bleeding. Recent weight loss ranging from 10 to 30 pounds was documented in 17 patients. Twenty-two of the patients had an Hl American Joint Committee on Cancer (AJCC) performance status; the remaining five with a normal (HO) status underwent a diagnostic work-up due to bleeding. On the basis of the TNM staging system, 14 patients (52 percent) had stage IV disease, 9 (33 percent) stage III, 3 (11 percent) stage II, and one stage I disease. The indications for total gastrectomy are listed in Table I. Nine patients (33 percent) demonstrated linitis plastica on histologic examination.
VOLUME 158 DECEMBER 1989
TABLE II
TABLE I Indlcatlow for
Total Gastrectomy
Po8topratlvo CompMailonB n
Locally advanced disease (stage IV) Proximal location Linltis plasticaon histologic examination Priorgastricreeectlon
cardbv~cular
14 6
’
5 2
Includes rensl fallue, thr~lebltls. and small-bowel obstructbn. l
Eight patients underwent a splenectomy in conjunction with gastrectomy and omentectomy. Five patients had an extended total gastrectomy (gastrectomy, splenectomy, and distal pancreatectomy). Two of the five also underwent resection of the transverse colon due to extension of the primary into the mesocolon. The operation was performed through an abdominal incision in 26 of the 27 cases. Gastrointestinal continuity was restored exclusively by means of a functional end-to-end Roux-Y esophagojejunostomy. There were 20 stapled and 7 hand-sutured anastomoses; the stapler has been used in every case since 1983. At the conclusion of the procedure, residual intraabdominal disease was present in 12 of 27 cases (44 percent). Six patients had gross residual tumor, four had microscopically positive margins, and two patients had both gross residual tumor and positive margins. One patient died on postoperative day 20, for an operative mortality rate of 4 percent. A second patient was home for only a brief period before returning to the hospital and dying two months after the resection. Postoperative morbidity occurred in 48 percent of the patients (13 of 27). The complications are listed in Table II. Three patients required reoperation for hemorrhage from a mesenteric vessel, necrotizing pancreatitis, and smallbowel obstruction. Barium swallow demonstrated an anastomotic leak in only 1 of 26 cases; this closed with conservative management. Oral intake was begun a median of 7 days postoperatively, and 25 of 26 patients left the hospital tolerating solid food. Median postoperative hospitalization was 17 days, with a range from 10 to 36 days. Postoperative therapy was given to three patients with advanced disease; two received chemotherapy and one a combination of chemotherapy and radiation therapy. Median survival for the entire group was 15 months, with a range from 2 to 110 months. The uncorrected absolute l- and 2-year survival rates are 62 and 38 percent, respectively. Two patients are alive without evidence of disease at 36 and 74 months, one patient was disease-free when lost to follow-up at 39 months, and one patient is alive with liver metastases at 26 months. Twenty of the remaining 21 patients have died of gastric cancer; one patient died without evidence of disease 110 months after gastrectomy. After deleting the data of the four patients with stage I or II disease who underwent total gastrectomy on the basis of the proximal location of the tumor, the median survival is 12 months (range: 2 to 43 months), with l- and 2-year survival rates of 55 and 27 percent. Five patients developed bowel obstructions an average
5 3 3 3 5
hlmonery s@s Wound infectlo(r OthBr’ m,
pamreatltls,
of nine months after resection. Four had diffuse peritoneal metastases at laparotomy. Three of the four underwent bypass and lived for three, three, and eight months, respectively; the other patient had a feeding jejunostomy and died one month later. One patient was successfully managed nonoperatively, was disease-free for an additional two years, and was then lost to follow-up. The remaining 21 patients were able to maintain oral alimentation until the period immediately prior to their death.
COMMENTS An estimated 14,000 people will die of gastric cancer in the United States in 1989 [I]. Despite the improved results reported for patients with superficial gastric cancer, most patients present with advanced disease. Surgery remains the only effective means of palliation because of the lack of any appreciable response to either chemo therapy or radiation therapy in the treatment of advanced disease. Although subtotal gastric resection is routinely recommended as a palliative procedure, the role of total gastrectomy in patients with advanced disease is a source of considerable controversy [4]. Multiple studies have documented an operative mortalit rate of 15 to 25 percent following total gastrectomy [5&J I n an analysis of 15 series of gastric cancer treatments, Longmire [7] reported an average operative mortality rate of 22 percent. In addition, the nutritional difficult& encountered after total gastrectomy have been an equally important consideration, particularly in patients with advanced disease whose resection is strictly palliative. The recognition of bide reflu esophagitis and its prevention by a Roux-Y jejunal reconstruction, however, have obviated most of the nutritional problems associated with the procedure. The two factors that determine the success of palliative resection are operative mortality and duration of survival. Recent series demonstrate an improvement in operative mortality. Diehl et al [S] reported a 7 percent operative mortality rate in patients undergoing total gastrectomy. Saario and colleagues [9] reported an 8 percent mortality rate in 100 consecutive patients. In a retrospec tive review covering a 20- ear period from 1955 to 1975, Shiu and co-workers [10 r noted a decrease in mortality for total gastrectomy from 23 percent for the first 15 years to 8 percent over the last 5 years of the study. Gur mortality rate of 4 percent is consistent with these results. The most important factor in the improvement of opera-
THE AMERICAN JOURNAL OF SURGERY
VOLUME 158 DECEMBER 1989
603
tive mortality is the avoidance of complications due to breakdown of the esophageal anastamosis, which is the major cause of both the mortality and morbidity of this procedure [ZZ,12].Whereas some authors have suggested that the decrease in leaks is due to the use of stapling devices [13], others have achieved equivalent results with sutured anastamoses [14]. The one advantage the stapler does confer is the ability to perform an anastamosis higher up on the esophagus in cases where the tumor extends proximally without having to enter the thoracic cavity, thus avoiding the increased morbidity associated with a thoracoabdominal operation. The fact that we were able to construct the anastamosis through an abdominal ap preach in 26 of 27 cases contributed to the low mortality. The second measure of successful palliation, improved survival, is more difficult to evaluate. The majority of our patients presented with obstructive symptomatology and would have required continued hospitalization for intravenous nutritional support. Although gastroenterostomy may have been an alternative approach in some of these patients, most studies show that this procedure provides less palliation and for a significantly shorter period of time when compared with resection, without a reduction in postoperative complications or mortality [15-271. Twenty-five of the 26 patients in our series were discharged from the hospital tolerating solid food. Twentytwo of those patients were able to maintain oral alimentation until just before their death. A median survival of 15 months with uncorrected land 2-year survival rates of 62 percent and 38 percent compares favorably with the results of other series. After deleting the data of the 5 patients with stage I or II disease, survival is 55 percent at 1 year and 27 percent at 2 years. Although one could reasonably argue that survival is not necessarily prolonged in these patients with advanced disease, removing the tumor and restoring an intestinal tract free of gross disease provides maximal palliation and allows patients to return home with an improved quality of life. On the basis of a 4 percent operative mortality rate, we conclude that in patients with advanced gastric carcinoma, total gastrectomy with Roux-Y esophagojejunostomy can be performed with an acceptable morbidity and mortality, provides significant palliation by restoring the patient’s ability to eat, and should be performed when technically feasible, even in the presence of macrosmpic residual disease. REFERENCES 1. Silverberg E, Lubera J. Cancer statistics 1989. CA 1989; 1: 320. 2. Bringaxe W, Chappuis C, Cohn I, et al. Early gastric cancer. Arm Surg 1986; 204: 103-13. 3. Dupont JB Jr, Cohn I Jr. Gastric adenccarcinoma. Cmr Probl Cancer 1980; 4: S-46. 4. ReMine WH. Palliative operations for incurable gastric cancer. World J Surg 1979; 3: 721-9. 5. Adashek K, Sanger J, Longmire WP Jr: Cancer of the stomach: review of consecutive ten-year intervals. Ann Surg 1979; 189: 6-10. 6. Cady B, Ramsden DA, Sten A, Hagit RC. Gastric cancer: contemporary aspects. Am J Surg 1977; 133: 423-9. 7. Longmire WP. Gastric carcinoma: is radical gastrectomy worth-
604
THE AMERICAN JOURNAL OF SURGERY
while? Ann R Coil Surg Engl 1980; 62: 25-30. 8. Diehl JT, Hermann RE, Cooperman AM, Hoerr SO. Gastric carcinoma: a 10 year review. Ann Surg 1983; 198: 9-12. 9. Saario I, Schroder T, Tolpannen E, Lempinen M: Total gastrectomy with esophagojejunostomy. Am J Surg 1986; 151: 244-6. 10. Shiu M, Papacristou D, Kosloff C, et al. Selection of operative procedure for adenocarcinoma of the mid-stomach. Ann Surg 1980; 192: 730-7. Il. Koga S, Kishimoto H, Tanaka K, et al. Results of total gastrectomy for gastric cancer. Am J Surg 1980; 140: 636-8. 12. Calen L, Portier G, Ozoux J, et al. Carcinoma of the cardia and proximal third of the stomach. Am J Surg 1988; 155: 481-5. 13. Campion J, Nomikos J, Launois B: Duodenal closure and esophagojejunostomy experience with mechanical stapling devices in total gastrectomy for cancer. Arch Surg 1988; 123: 979-83. 14. Cady B, Ross R, Silverman M, et al. Gastric adenocarcinoma. Arch Surg 1989; 124: 303-8. 15. Ekbom G, Gleysteen J. Gastric malignancy: resection for paliation. Surgery 1980; 88; 476-81. 16. Stem JL, Denman S, Elias EG, Didolkar M, Holyoke ED. Evaluation of palliative resection in advanced carcinoma of the stomach. Surgery 1975; 77: 291-8. 17. Inberg MV, Byori J, Uiikari SJ. Carcinoma of the stomach, a follow-up of 1963 patients. Acta Chir Stand 1972; 138: 195-201.
DISCUSSION James H. Thomas
(Kansas City, KS): Dr. Butler, you mentioned the presence of positive margins, but you did not indicate whether these were, in fact, esophageal or duodenal. Was there a difference between palliation in these individuals and, additionally, what was the reason for not obtaining clear margins? Were you unable to extend, for instance, more distally into the duodenum, and-if so, why? Were you unable to extend more proximally into the esophagus, and if so, why did you hesitate to extend the incision into the chest to obtain clear margins? I am particularly interested in the four individuals with so-called stump cancers. This is an interesting sub group of patients who are known to be at increased risk for development of gastric cancer, at least according to some of the reports from Europe., and I would like to know what the results were in these particular patients. Would you also comment on any of the nutritional prob lems that were encountered in this series of patients? Finally, it is of interest to me that stapling was used rather routinely after 1983. David V. Feliciano (Houston, TX): One of the implications of the paper is that resection should be performed in patients with stage III or stage IV carcinoma, since you did not indicate when you would not do a palliative resee tion in one of these groups. If a solitary hepatic metastasis or a fixed celiac set of nodes was present, would you perform total gastrectomy? I think that this is the group that we see much more commonly than the group whose tumor can be cleanly resected. James A. Edney (Omaha, NE): How would you treat a patient in whom you can’t resect the stomach and who is not obstructed? Does prophylactic gastrojejunostomy have a role? I have had experience with patients who were able to empty preoperatively. Once the gastrojejunostomy is done, however, they can’t empty afterwards, de
VOLUME 158 DECEMBER 1989
spite the fact that patency has been established. How do you treat these patients? Lawrence H. Wilkinson (Albuquerque, NM): For several decades, it has been well known that the Billroth I is more physiologic than the Billroth II procedure. More than 35 years ago, Dr. Arthur Hunnicutt developed the concept of using an ileocolic pouch to restore intestinal continuity after total gastrectomy. My reason for commenting is to suggest a reevaluation of the simplicity and physiologic advantages of this pouch, considering the new surgical devices that are available today. The suggestion is this: Once the pouch is rotated up, make a 5-cm linear incision in the tenia on the anterior wall of the pouch. Insert the small end-to-end anastomosis stapling instrument through this opening and pass it up through the ileum into the esophagus where it is fired and removed. Select a larger end-to-end anastomosis instrument; insert it through the same opening in the colon and pass it into the proximal duodenum where it is fued and removed. Then close the opening in the colon pouch and perform the ileocolic anastomosis.
Carey P. Page (San Antonio, TX): In the United States, the ratio of colon to gastric cancer is about six to one. In Mexico, colon cancer is a rare finding and gastric cancer is extremely prevalent, especially in young men less than 40 years of age. This phenomenon may alter some of the classic demography of gastric carcinoma in the southwestern United States with the large increases in recent immigrants from Mexico. Have you observed this phenomenon? Raymond C. Read (Little Rock, AR): This is a retrospective analysis apparently of total gastrectomy, and the investigators’ results are excellent. I would lie to have more information as to whether in certain cases they elect not to do a total gastrectomy. If in more than 40 percent of these cases, the cancer was not removed, what is the advantage of doing a pancreatectomy and a total gastrectomy when a lesser procedure that includes gastroenterostomy could have perhaps been performed in some of these cases?
John A. Butler (closing): Dr. Thomas, six of the patients had positive margins: four at the esophageal end, one at the duodenal end, and one at the esophageal and duodenal ends. These patients did not subsequently develop obstruction at the site of the anastomosis. The four patients with stump carcinoma did somewhat better than the overall group, surviving for 10 to 46 months after resection. In terms of preventing nutritional problems, I think that the recognition of bile reflu gastritis and its prevention with a Roux-Y esophagojejunostomy essentially eliminates most of those problems. We agree that the stapled anastomosis does not function better than the hand-sewn anastomosis. Dr. Feliciano, the focus of our research was palliation for patients with advanced disease, so we would perform resection in all patients who we thought had a reasonable survival potential. Therefore, the lindmg of a positive celiac node or even a single hepatic metastasis would not prevent us from doing a palliative resection. Dr. Edney, we have also found a prophylactic gastrojejunostomy to be of no value in the treatment of these patients. For the patient with a nonobstructing but unre sectable tumor, there is no palliative surgical procedure available. I would lie to emphasize, however, that most tumors can be resected. Dr. Wilkinson, concerning the ileocolic pouch, we have found that patients do very well from a nutritional standpoint with a Roux-Y esophagojejunostomy, and it is our opinion that the various pouches that have been developed are not required. I agree with Dr. Page that at our hospital we are seeing an increasing number of Spanish-American and Asian patients with gastric cancer. More important, these are the patients who are presenting at an early age with advanced disease due to a delay in diagnosis. Dr. Read, I agree that, if you could remove the entire tumor with a subtotal gastrectomy, that would be the procedure of choice. Most of the patients in our group, however, had proximal lesions or linitis plastica histology that involved the entire stomach. A gastrojejunostomy in these patients does not provide significant palliation in terms of relieving obstruction and has the equivalent morbidity and mortality of a gastric resection.
THE AMERICAN JOURNAL OF SURGERY
VOLUME 158 DECEMBER 1989
605
MD,
in the Treatment of Advanced Gastric Cancer
Terry J. Dubrow, MD, Thomas Trezona, MD, Michael Klassen, Robert J. Nejdl, MD, Torrance, California
To assess the role of total gastrectomy in the treatment of advanced gastric cancer, we reviewed the records of 27 patients who underwent the procedure from 1979 to 1988. Operative mortality was Ppercent (1 of 27), and postoperative morbidity occurred in 48 percent of the patients. Twenty-five of 26 patients were tolerating solid food at the time of discharge; 21 were able to maintain oral alimentation until just prior to their death. Median survival following the operation was 15 months (range: 2 to 110 months), with a 62 percent absolute l-year survival rate and a 38 percent 2-year survival rate. On the basis of these results, we conclude that in patients with advanced gastric carcinoma, total gastrectomy with Roux-Y esophagojejunostomy can he performed with an acceptable morbidity and mortality, provides significant palliation by restoring the patient’s ability to eat, and should be performed when technically feasible, even in the presence of gross residual disease.
lthough the incidence of gastric cancer has declined dramatically over the past 50 years, it is the 7th leading cause of death b cancer in men and the 9th leading cause in women [13r. Despite the successful results achieved in the treatment of superficial or “early” gastric cancer, most patients present with advanced disease [2]. Multiple reports have documented a lack of efficacy of total gastrectomy as a curative procedure when compared with subtotal resections [3]. The increased morbidity and mortality associated with the procedure, together with the nutritional difficulties encountered in the subsequent management of these patients with advanced disease, has led many physicians to conclude that total gastrectomy is contraindicated for locally advanced gastric cancer. We performed this study to review our recent experience with total gastrectomy and determine its role in the management of these patients.
A
From the Departments of Surgery, Harbor-UCLA Medical Center, Torrance, California, and Kaiser Medical Center, Harbor City, California. Requests for reprints should be sent to John A. Butler, MD, Harbor-UCLA Medical Center, Box 25, 1000 West Carson Street, Torrance, California 90502. Presented at the 41st Annual Meeting of the Southwestern Surgical Congress, Monterey, California, April 23-26,1989.
602
THE AMERICAN JOURNAL OF SURGERY
PATIENTS
MD,
AND METHODS
The tumor registries of Los Angeles County HarborUCLA Medical Center and Kaiser-Harbor City Hospital were reveiwed over a IO-year period from 1979 to 1988. Records of all patients who underwent a total gastrectomy for cancer were reviewed. Patients who had either a near-total gastrectomy or proximal esophagogastrectomy were excluded from the analysis. Patients with a histologic diagnosis of squamous cell carcinoma or malignant lymphoma were also excluded. Twenty-seven patients with gastric adenocarcinoma who underwent total gastrectomy were identified. Their hospital records were reviewed for demographic data, past history, symptoms, tumor size and location, and pathologic characteristics. The operative procedure, with particular reference to the method of restoring intestinal continuity, postoperative morbidity and mortality, additional therapy, and overall survival, was also examined. A patient who died without leaving the hospital after surgery was classified as an operative mortality. Complete follow-up until death or the time of this report was documented for 26 (96 percent) patients. One patient was lost to follow-up 39 months postgastrectomy, at which time he was believed to be free of disease. Crude, uncorrected, absolute survival times are used in view of the fact that only one patient died of causes unrelated to gastric cancer. RESULTS
Twenty-seven patients underwent total gastrectomy during the lo-year period reviewed. Eighteen patients were men and 9 were women. The median age of the patients was 63, with a range from 21 to 93 years. Seven of the 27 (26 percent) were over 70 years of age. Three patients had a gastric resection for peptic ulcer disease 12, 25, and 37 years prior to the development of gastric cancer; one other patient had a prior vagotomy and pyloroplasty. Seventeen of the 27 patients (63 percent) presented with obstructive symptoms, nine of whom could tolerate only liquids. Ten patients (37 percent) presented with evidence of upper gastrointestinal bleeding. Recent weight loss ranging from 10 to 30 pounds was documented in 17 patients. Twenty-two of the patients had an Hl American Joint Committee on Cancer (AJCC) performance status; the remaining five with a normal (HO) status underwent a diagnostic work-up due to bleeding. On the basis of the TNM staging system, 14 patients (52 percent) had stage IV disease, 9 (33 percent) stage III, 3 (11 percent) stage II, and one stage I disease. The indications for total gastrectomy are listed in Table I. Nine patients (33 percent) demonstrated linitis plastica on histologic examination.
VOLUME 158 DECEMBER 1989
TABLE II
TABLE I Indlcatlow for
Total Gastrectomy
Po8topratlvo CompMailonB n
Locally advanced disease (stage IV) Proximal location Linltis plasticaon histologic examination Priorgastricreeectlon
cardbv~cular
14 6
’
5 2
Includes rensl fallue, thr~lebltls. and small-bowel obstructbn. l
Eight patients underwent a splenectomy in conjunction with gastrectomy and omentectomy. Five patients had an extended total gastrectomy (gastrectomy, splenectomy, and distal pancreatectomy). Two of the five also underwent resection of the transverse colon due to extension of the primary into the mesocolon. The operation was performed through an abdominal incision in 26 of the 27 cases. Gastrointestinal continuity was restored exclusively by means of a functional end-to-end Roux-Y esophagojejunostomy. There were 20 stapled and 7 hand-sutured anastomoses; the stapler has been used in every case since 1983. At the conclusion of the procedure, residual intraabdominal disease was present in 12 of 27 cases (44 percent). Six patients had gross residual tumor, four had microscopically positive margins, and two patients had both gross residual tumor and positive margins. One patient died on postoperative day 20, for an operative mortality rate of 4 percent. A second patient was home for only a brief period before returning to the hospital and dying two months after the resection. Postoperative morbidity occurred in 48 percent of the patients (13 of 27). The complications are listed in Table II. Three patients required reoperation for hemorrhage from a mesenteric vessel, necrotizing pancreatitis, and smallbowel obstruction. Barium swallow demonstrated an anastomotic leak in only 1 of 26 cases; this closed with conservative management. Oral intake was begun a median of 7 days postoperatively, and 25 of 26 patients left the hospital tolerating solid food. Median postoperative hospitalization was 17 days, with a range from 10 to 36 days. Postoperative therapy was given to three patients with advanced disease; two received chemotherapy and one a combination of chemotherapy and radiation therapy. Median survival for the entire group was 15 months, with a range from 2 to 110 months. The uncorrected absolute l- and 2-year survival rates are 62 and 38 percent, respectively. Two patients are alive without evidence of disease at 36 and 74 months, one patient was disease-free when lost to follow-up at 39 months, and one patient is alive with liver metastases at 26 months. Twenty of the remaining 21 patients have died of gastric cancer; one patient died without evidence of disease 110 months after gastrectomy. After deleting the data of the four patients with stage I or II disease who underwent total gastrectomy on the basis of the proximal location of the tumor, the median survival is 12 months (range: 2 to 43 months), with l- and 2-year survival rates of 55 and 27 percent. Five patients developed bowel obstructions an average
5 3 3 3 5
hlmonery s@s Wound infectlo(r OthBr’ m,
pamreatltls,
of nine months after resection. Four had diffuse peritoneal metastases at laparotomy. Three of the four underwent bypass and lived for three, three, and eight months, respectively; the other patient had a feeding jejunostomy and died one month later. One patient was successfully managed nonoperatively, was disease-free for an additional two years, and was then lost to follow-up. The remaining 21 patients were able to maintain oral alimentation until the period immediately prior to their death.
COMMENTS An estimated 14,000 people will die of gastric cancer in the United States in 1989 [I]. Despite the improved results reported for patients with superficial gastric cancer, most patients present with advanced disease. Surgery remains the only effective means of palliation because of the lack of any appreciable response to either chemo therapy or radiation therapy in the treatment of advanced disease. Although subtotal gastric resection is routinely recommended as a palliative procedure, the role of total gastrectomy in patients with advanced disease is a source of considerable controversy [4]. Multiple studies have documented an operative mortalit rate of 15 to 25 percent following total gastrectomy [5&J I n an analysis of 15 series of gastric cancer treatments, Longmire [7] reported an average operative mortality rate of 22 percent. In addition, the nutritional difficult& encountered after total gastrectomy have been an equally important consideration, particularly in patients with advanced disease whose resection is strictly palliative. The recognition of bide reflu esophagitis and its prevention by a Roux-Y jejunal reconstruction, however, have obviated most of the nutritional problems associated with the procedure. The two factors that determine the success of palliative resection are operative mortality and duration of survival. Recent series demonstrate an improvement in operative mortality. Diehl et al [S] reported a 7 percent operative mortality rate in patients undergoing total gastrectomy. Saario and colleagues [9] reported an 8 percent mortality rate in 100 consecutive patients. In a retrospec tive review covering a 20- ear period from 1955 to 1975, Shiu and co-workers [10 r noted a decrease in mortality for total gastrectomy from 23 percent for the first 15 years to 8 percent over the last 5 years of the study. Gur mortality rate of 4 percent is consistent with these results. The most important factor in the improvement of opera-
THE AMERICAN JOURNAL OF SURGERY
VOLUME 158 DECEMBER 1989
603
tive mortality is the avoidance of complications due to breakdown of the esophageal anastamosis, which is the major cause of both the mortality and morbidity of this procedure [ZZ,12].Whereas some authors have suggested that the decrease in leaks is due to the use of stapling devices [13], others have achieved equivalent results with sutured anastamoses [14]. The one advantage the stapler does confer is the ability to perform an anastamosis higher up on the esophagus in cases where the tumor extends proximally without having to enter the thoracic cavity, thus avoiding the increased morbidity associated with a thoracoabdominal operation. The fact that we were able to construct the anastamosis through an abdominal ap preach in 26 of 27 cases contributed to the low mortality. The second measure of successful palliation, improved survival, is more difficult to evaluate. The majority of our patients presented with obstructive symptomatology and would have required continued hospitalization for intravenous nutritional support. Although gastroenterostomy may have been an alternative approach in some of these patients, most studies show that this procedure provides less palliation and for a significantly shorter period of time when compared with resection, without a reduction in postoperative complications or mortality [15-271. Twenty-five of the 26 patients in our series were discharged from the hospital tolerating solid food. Twentytwo of those patients were able to maintain oral alimentation until just before their death. A median survival of 15 months with uncorrected land 2-year survival rates of 62 percent and 38 percent compares favorably with the results of other series. After deleting the data of the 5 patients with stage I or II disease, survival is 55 percent at 1 year and 27 percent at 2 years. Although one could reasonably argue that survival is not necessarily prolonged in these patients with advanced disease, removing the tumor and restoring an intestinal tract free of gross disease provides maximal palliation and allows patients to return home with an improved quality of life. On the basis of a 4 percent operative mortality rate, we conclude that in patients with advanced gastric carcinoma, total gastrectomy with Roux-Y esophagojejunostomy can be performed with an acceptable morbidity and mortality, provides significant palliation by restoring the patient’s ability to eat, and should be performed when technically feasible, even in the presence of macrosmpic residual disease. REFERENCES 1. Silverberg E, Lubera J. Cancer statistics 1989. CA 1989; 1: 320. 2. Bringaxe W, Chappuis C, Cohn I, et al. Early gastric cancer. Arm Surg 1986; 204: 103-13. 3. Dupont JB Jr, Cohn I Jr. Gastric adenccarcinoma. Cmr Probl Cancer 1980; 4: S-46. 4. ReMine WH. Palliative operations for incurable gastric cancer. World J Surg 1979; 3: 721-9. 5. Adashek K, Sanger J, Longmire WP Jr: Cancer of the stomach: review of consecutive ten-year intervals. Ann Surg 1979; 189: 6-10. 6. Cady B, Ramsden DA, Sten A, Hagit RC. Gastric cancer: contemporary aspects. Am J Surg 1977; 133: 423-9. 7. Longmire WP. Gastric carcinoma: is radical gastrectomy worth-
604
THE AMERICAN JOURNAL OF SURGERY
while? Ann R Coil Surg Engl 1980; 62: 25-30. 8. Diehl JT, Hermann RE, Cooperman AM, Hoerr SO. Gastric carcinoma: a 10 year review. Ann Surg 1983; 198: 9-12. 9. Saario I, Schroder T, Tolpannen E, Lempinen M: Total gastrectomy with esophagojejunostomy. Am J Surg 1986; 151: 244-6. 10. Shiu M, Papacristou D, Kosloff C, et al. Selection of operative procedure for adenocarcinoma of the mid-stomach. Ann Surg 1980; 192: 730-7. Il. Koga S, Kishimoto H, Tanaka K, et al. Results of total gastrectomy for gastric cancer. Am J Surg 1980; 140: 636-8. 12. Calen L, Portier G, Ozoux J, et al. Carcinoma of the cardia and proximal third of the stomach. Am J Surg 1988; 155: 481-5. 13. Campion J, Nomikos J, Launois B: Duodenal closure and esophagojejunostomy experience with mechanical stapling devices in total gastrectomy for cancer. Arch Surg 1988; 123: 979-83. 14. Cady B, Ross R, Silverman M, et al. Gastric adenocarcinoma. Arch Surg 1989; 124: 303-8. 15. Ekbom G, Gleysteen J. Gastric malignancy: resection for paliation. Surgery 1980; 88; 476-81. 16. Stem JL, Denman S, Elias EG, Didolkar M, Holyoke ED. Evaluation of palliative resection in advanced carcinoma of the stomach. Surgery 1975; 77: 291-8. 17. Inberg MV, Byori J, Uiikari SJ. Carcinoma of the stomach, a follow-up of 1963 patients. Acta Chir Stand 1972; 138: 195-201.
DISCUSSION James H. Thomas
(Kansas City, KS): Dr. Butler, you mentioned the presence of positive margins, but you did not indicate whether these were, in fact, esophageal or duodenal. Was there a difference between palliation in these individuals and, additionally, what was the reason for not obtaining clear margins? Were you unable to extend, for instance, more distally into the duodenum, and-if so, why? Were you unable to extend more proximally into the esophagus, and if so, why did you hesitate to extend the incision into the chest to obtain clear margins? I am particularly interested in the four individuals with so-called stump cancers. This is an interesting sub group of patients who are known to be at increased risk for development of gastric cancer, at least according to some of the reports from Europe., and I would like to know what the results were in these particular patients. Would you also comment on any of the nutritional prob lems that were encountered in this series of patients? Finally, it is of interest to me that stapling was used rather routinely after 1983. David V. Feliciano (Houston, TX): One of the implications of the paper is that resection should be performed in patients with stage III or stage IV carcinoma, since you did not indicate when you would not do a palliative resee tion in one of these groups. If a solitary hepatic metastasis or a fixed celiac set of nodes was present, would you perform total gastrectomy? I think that this is the group that we see much more commonly than the group whose tumor can be cleanly resected. James A. Edney (Omaha, NE): How would you treat a patient in whom you can’t resect the stomach and who is not obstructed? Does prophylactic gastrojejunostomy have a role? I have had experience with patients who were able to empty preoperatively. Once the gastrojejunostomy is done, however, they can’t empty afterwards, de
VOLUME 158 DECEMBER 1989
spite the fact that patency has been established. How do you treat these patients? Lawrence H. Wilkinson (Albuquerque, NM): For several decades, it has been well known that the Billroth I is more physiologic than the Billroth II procedure. More than 35 years ago, Dr. Arthur Hunnicutt developed the concept of using an ileocolic pouch to restore intestinal continuity after total gastrectomy. My reason for commenting is to suggest a reevaluation of the simplicity and physiologic advantages of this pouch, considering the new surgical devices that are available today. The suggestion is this: Once the pouch is rotated up, make a 5-cm linear incision in the tenia on the anterior wall of the pouch. Insert the small end-to-end anastomosis stapling instrument through this opening and pass it up through the ileum into the esophagus where it is fired and removed. Select a larger end-to-end anastomosis instrument; insert it through the same opening in the colon and pass it into the proximal duodenum where it is fued and removed. Then close the opening in the colon pouch and perform the ileocolic anastomosis.
Carey P. Page (San Antonio, TX): In the United States, the ratio of colon to gastric cancer is about six to one. In Mexico, colon cancer is a rare finding and gastric cancer is extremely prevalent, especially in young men less than 40 years of age. This phenomenon may alter some of the classic demography of gastric carcinoma in the southwestern United States with the large increases in recent immigrants from Mexico. Have you observed this phenomenon? Raymond C. Read (Little Rock, AR): This is a retrospective analysis apparently of total gastrectomy, and the investigators’ results are excellent. I would lie to have more information as to whether in certain cases they elect not to do a total gastrectomy. If in more than 40 percent of these cases, the cancer was not removed, what is the advantage of doing a pancreatectomy and a total gastrectomy when a lesser procedure that includes gastroenterostomy could have perhaps been performed in some of these cases?
John A. Butler (closing): Dr. Thomas, six of the patients had positive margins: four at the esophageal end, one at the duodenal end, and one at the esophageal and duodenal ends. These patients did not subsequently develop obstruction at the site of the anastomosis. The four patients with stump carcinoma did somewhat better than the overall group, surviving for 10 to 46 months after resection. In terms of preventing nutritional problems, I think that the recognition of bile reflu gastritis and its prevention with a Roux-Y esophagojejunostomy essentially eliminates most of those problems. We agree that the stapled anastomosis does not function better than the hand-sewn anastomosis. Dr. Feliciano, the focus of our research was palliation for patients with advanced disease, so we would perform resection in all patients who we thought had a reasonable survival potential. Therefore, the lindmg of a positive celiac node or even a single hepatic metastasis would not prevent us from doing a palliative resection. Dr. Edney, we have also found a prophylactic gastrojejunostomy to be of no value in the treatment of these patients. For the patient with a nonobstructing but unre sectable tumor, there is no palliative surgical procedure available. I would lie to emphasize, however, that most tumors can be resected. Dr. Wilkinson, concerning the ileocolic pouch, we have found that patients do very well from a nutritional standpoint with a Roux-Y esophagojejunostomy, and it is our opinion that the various pouches that have been developed are not required. I agree with Dr. Page that at our hospital we are seeing an increasing number of Spanish-American and Asian patients with gastric cancer. More important, these are the patients who are presenting at an early age with advanced disease due to a delay in diagnosis. Dr. Read, I agree that, if you could remove the entire tumor with a subtotal gastrectomy, that would be the procedure of choice. Most of the patients in our group, however, had proximal lesions or linitis plastica histology that involved the entire stomach. A gastrojejunostomy in these patients does not provide significant palliation in terms of relieving obstruction and has the equivalent morbidity and mortality of a gastric resection.
THE AMERICAN JOURNAL OF SURGERY
VOLUME 158 DECEMBER 1989
605