- Email: [email protected]
Current experience with pancreatogastrostomy
Current
Experience
With Pancreatogastrostomy
G. Robert Mason, MD, Robert J. Freeark, MD, Muywood Illinois BACKGROUND: The reconstruction techuique for the pancreatic remnant remaiuiug after pancreatoduodenectomy has most frqently been paucreatojejunostomy. Although the mortality rate baa been reduced to rather low levels in many centers, the leakage rate from this auastomosis remaina high, in the range of 10% or greater. An alternative reconstruction, pauereatogastrostomy, has heen known for ahnost 50 years and has been performed on small numbers of patients. The leakage rate for this auastomoais is less than 1% in literature reports in more than 200 patieuts. The purpose of this report was to add to the previously reported experience with this technique and to compare it with standard reconstruction as performed in a major American medical center by experienced surgeons. METHODS: The medical records of all patients operated on at the Loyola University Medical Center and the Edward Hines, Jr., Veteran’s Affairs Hospital from August 1986 to May 1993, with a procedure code relating to the pancreas, were reviewed. RESULTS: A total of 58 Whipple procedures were ident&ed, iucl~ 34 pancreatogastrostomies, 23 pancreatojejunostomies and 1 stapled pancreatic stump. No leaks in any pancreatogastrostomies were observed iu the 38 patients so treated, whereas 4 leaks and 2 deaths related to the auastomosis occurred in the group of 23 patients with panereatojejunostomies. The average length of stay was 14.2 days for the pancreatojejunostomy group and 15.5 days for the pancreatogastrosmmy group, excluding duration data from those who died or experienced leakage. There was no sign&ant difference in the length of stay between paucreatojejmmstomy and pancreactogactrostomy; there was a statistically signiticaut lengthening of stay for those patients whose auastomosis leaked versus those whose did not leak. The 10 patients having a pylorus-sparing operation bad an average postoperative stay of 16 days, including both types of reconstruction. CONCLUSIONS: The gradual adoption of this procedure at a major medical center has led to the abandonment of pancreatojejunostomy as a reconstruction technique for the pancreatic remnant remaining after paucreatoduodenectomy.
ince 1935, when Whipple et al’ first described a sucS cessful technique for pancreatoduodenectomy, numerous variations of this procedure have been described. The original two-stage operation was done in one stage within 5 years by Whipple? Trimble et al,3 and Brunschwige4 Brunschwig and Trimble et al both discussed their concerns, at some length, with the pancreatoenteric anastomosis, particularly for fistula formation as a complication of the procedure. Brunschwig recommended ligation of the pancreatic stump, feeling that “successful anatomic implantation of the pancreatic stump into bowel in man has not yet been demonstrated to afford normal or appreciable secretion of pancreatic juice into the bowel.” Others have felt differently and have made various anastomoses of pancreas remnant to small intestine. The mortality and morbidity rates in the best of these series has been reduced to 0% to 2%, and 36%, respectively,” but the problem of pancreatic fistulae remains prominent at 9% to 18%,“17 and may be implicated in complications such as bleeding, abscess, and sepsis. Although these leaks may not result in the death of the patient, the presence of complications in any series is likely to be related to an overall greater mortality rate and longer hospital stay. In animal experiments, Tripodi and Sherwin8 described transplantation of the pancreas into the stomach in 1934, but the procedure was not done clinically for patients in association with pancreatoduodenectomy until 1946, as reported by Waugh and Clagett’ from the Mayo Clinic. Since that time there have been sporadic reports of pancreatogastrostomy being performed with only one leak clearly related to this anastomosis, reported in 1975,‘* although there is one other report which describes a transient leak of pancreatic juice, which, in theory, could be from any of the three reconstructive anastomoses.’ ’ The purpose of this report is to add to the previously reported experience of pancreatogastrostomy for reconstruction after pancreatoduodenectomy and to compare it with standard reconstruction as performed in a major American medical center by experienced surgeons. PATIENTS
AND
METHODS
The medical records of all patients operated on at the Loyola University Medical Center and Edward Hines, Jr., Veteran’s Affairs Hospital from August 1986 to May 1993, with a procedure code relating to the pancreas, were requested from the medical records departments of the two hospitals. There were no other inclusion or exclusion criteria applied. RESULTS
From the Department of Surgery, Loyola University Medical Center and Surgical Service, Edward Hines, Jr., Veteran’s Affairs Hospital, Maywood, Illinois. Requests for reprints should be addressed to G. Robert Mason, MD, Department of Surgery, Loyola University Medical Center, 2160 South First Avenue, Maywood, Illinois 60153. Manuscript submitted October 7, 1993 and accepted in revi& form March 7, 1994.
THE AMERICAN
A total of 58 Whipple procedures were identified, including 34 pancreatogastrostomies, 23 pancreatojejunostomies, and 1 stapled pancreatic stump (Table). In the pancreatojejunostomy group there were 7 women and 16 men; in the pancreatogastrostomy group there were 10 women and 24 men. In 11 patients the pylorus was spared (8 in the pancreatogastrostomy group and 3 in the pancreatoje-
JOURNAL
OF SURGERY@
VOLUME
169
FEBRUARY
1995
217
CURRENT
EXPERIENCE
WITH
PANCREATOGASTROSTOMSON
AND
FREEARK
gastrostomiesdisrupted,leaked,or contributed to the mortality or morbidity of any patient. The 2 deathsin the pancreatogastrostomygroup were from myocardial infarction and peritonitis, asdescribed. In the pancreatojejunostomygroup, there were also 2 deaths, and these were directly related to breakdown of this anastomosis.Two other leaks led to a prolonged hospital stay. The average length of stay was 14.2 days for the pancreatojejunostomygroup and 15.5days for the pancreatogastrostomygroup, excluding duration data from thosewho died or experiencedleakage.There was no significant difference in the length of stay between pancreatojejunostomy and pancreactogactrostomy;there was a statistically significant lengthening of stay for those patients whose anastomosisleaked versus those whose did not leak. The 10 patientshaving a pylorus-sparingoperation had an averagepostoperativestay of 16 days, including both types of reconstruction.
TABLE Annual Frequency of Pancreatogastrostomy and Pancreatojejunostomy for Reconstruction After Pancreatoduodenectomy’ PancreatoAll Whipple Pancreatojejunostomy Procedures Year gastrostomy 0 3 3 1986 0 4 4 1987 2 3 5 1988 5 1989 2 3 17 1990 8 9 10 1991 8 2 1992 10 0 10 1993 4 0 4 Total 34 24 58 ‘Data from the medical records of all oatients oDerated on at the Lovola University Medical Center and the Hines Veterans Admlnistration Hospital from August 1986 to May 1993, with a procedure code relating to the pancreas. I
100
-
/’ ,’
$ e601
:
a Q) I -ii. ’ g 40 -
+
pancreatqejunostomy
:
3
75 gl 20 0’
:i
/’
: : : .I’
fi 1986
n 1967
: :
:
/ 1988
I 1989
/ 1990
I 1991
l
1992
1993
I Year Trend toward use of pancreatogastrostomy as compared to pancreatolejunostomy tar reconstructlon after pancreatoduodenectomy at one institution. Data are from the medical records of all patients operated on at the Loyola University Medical Center and the Edward Hines, Jr., Veteran’s Affairs Hospital from August 1986 to May 1993, with a procedure code relating to the pancreas. L F igure.
junostomy group); ofthese, 5 were women.One of the pancreatojejunostomypatient in this group experienceda leak, but survived. The Whipple procedureswere performed by eight different surgeons. There were 5 deaths,at 12, 14, 21, 26, and 45 days. The death of a patient with pancreatogastrostomyat 12 days was attributed to adult iespiratory distresssyndromecausedby fluid overload. The death at 14 days was by myocardial infarction associatedwith gastrointestinalbleeding in a patient with pancreatogastrostomy. The deathsat 21 and 45 days were associated with leaksfrom pancreatojejunostomy.The deathat 26 days was shown at postmortemto be related to peritonitis secondary to breakdownof the gastroenterostomyin a patient whosepancreatogastrostomywas intact. In addition to the patientswho died from leakage,therewere 2 other patients with pancreatojejunostomies who had evidence of leakage of this anastomosis, but survived, with hospital staysof 28 and 45 days. There was no evidence that the pa&reato218
THE AMERICAN
JOURNAL
OF SURGEKY”
VOLUME
169
COMMENTS The patients describedhere were not part of a prospective scientific study and were not subjectedto a randomization protocol. The data are presentedto show the gradual adoption, by a group of experienced surgeons,of pancreatogastrostomyasan alternative reconstructionprocedure to pancreatojejunostomybecauseof problems encountered with the latter (Figure). In the pancreatojejunostomy group, leakage of this anastomosisaccounted for the 2 deaths,anddoubledthe hospitalstay for the 2 patients with nonlethal leaks. The leakagerate in this group was 4 of 23 or 17%, as opposedto 0% for the pancreatogastrostomygroup. The use of an alternative reconstruction of this sort obviously doesnot eliminate the possibility of morbidity or mortality from other causes.In current reports the mortality rate for pancreatoduodenectomyin majormedicalcentersis in the areaof 7% to 9%,‘* although serieshave beenreported with no mortality.‘” The leakagerate for pancreatojejunostomyhas beenreported in the 9% to 12% range for several decadesand doesnot seemto have changedappreciably in recent years. The actual incidence of such leaks in institutions not reporting their statisticsmay well be higher and may be responsiblefor the continuing publication of articles relating to treatment and technical alternatives, such as irradiation, duct ligation, and treatment with somatostatin.‘4-‘6Seriesin the current literature still indicate relatively infrequent performance of pancreatogastrostomy, but in the reported cases,the incidence of breakdown or leakageis minimal. There are 228 casesof pancreatogastrostomy found by our searchthat have had a total of four leaks reported. As noted above, only one of theseis documented as being from the pancreatogastrostomy.Two other caseshave beenmadeknown to usby personalcommunication (G. Black, MD, 1993, and P. Donahue, MD, 1993). Five of the total of 6 caseswere resolved without operation or seriousmorbidity. In addition to the casesreported here and in the literature, we also add to the original seriesof 9 patients reported by Telford and Mason.” another 15 pancreatogastrostomies performed by the same authors, and their associates,without evidence of leak of this anastomosis. FEBRUARY
1995
CURRENT
EXPERIENCE
These procedures were performed not only in university medical centers but also in community hospitals without the usual support systems that are thought to be necessary in tertiary care and comparable referral centers. With this report, there are now 277 patients known to us as having had this procedure, with a possible leak in 6, an incidence of 2%.1*-35Given the statistics in the literature for leakage from a pancreatoenteric anastomosis, the number of leaks in a similar number of pancreatojejunostomies should be between 25 and 50. There are, however, at most, 6 cases. The anastomoses reported here include the intussusception typz of anastomosis, described by Mackie and others, as well as the duct to gastric mucosa anastomosis.‘“*36~37 Long-term patency seems more assured from mucosa-tomucosa anastomosis, based on experimental work. Clearly, both yield satisfactory results in the immediate postoperative period. In the long term, we have reoperated on 3 patients whose intussuscepted anastomosis steno& over time and produced the signs and symptoms of chronic pancrea&is. The reoperative procedure is similar to that used for the Puestow type of pancreatoenteric anastomosis and has been remarkably technically easy, reflecting the absence of adhesions, probably an indication of lack of leakage at the original operation site. The lack of reported reoperations for this problem is likely to be related to the lifespan of the patient population rather than an actual incidence.
WITH
I. Whipple AO, Parsons WB, Mullins CR. Treatment of carcinoma of the ampulla of Vater. Ann Surg. 1935;102:763-779, 2. Whipple AO. Present day surgery of the pancreas. NEJM 1942; 226515-526. 3. Trimble IR, Parsons JW, Sherman CP. A one stage operation for the cure of carcinoma of the ampulla of Vater and of the head of the pancreas. Surg Gynecol & Obstet. 1941;73:71 l-722. 4. Brunschwig A. One Stage Pancreatoduodenectomy. Surg Gynecol & Obster. 1943;77:581-586. 5. Crist DW, Sitzman JV, Cameron JL. Improved hospital morbidity, mortality, and survival after the Whipple procedure. Ann Surg. review of pancreatoduodenectomy.
7. Cameron JL, Pitt HA, Ye0 CJ. One hundred and forty-tive consecutive pancreatectomies without mortality. Ann Surg. 1993;217:430-438. 8. Tripodi AM, Sherwin CF. Experimental transplantation of the pancreas into the stomach. Arch Surg. 1934;28:345-356, 9. Waugh JM, Clagett OT. Resection of the duodenum and the head of the pancreas for carcinoma. Surgery. 1946;20:224-232. 10. Mackie JA, Rboads JE, Park CD. Pancreatogastrostomy: a further evaluation. Ann Surg. 1975;181:541-545. 1I. Bradbeer JW, Johnson CD. Pancreaticogastrostomy after pancreaticoduodenectomy. Ann R Co11 Surf: Engl. 1990; 108:64M43. 12. Trede M, Schwall G. The complications of pancreatectomy. Ann Surg.
AND -__--- FREEAHK
1956;44:299-302.
25. Millboume E. Pancreaticogastrostomy in pancreaticoduodenal resection for carcinoma of the head of the pancreas or the papilla of Vater. Acta Chir Scand. 1958; 116: 12-27. 26. Pataky ZS, Popik E. Operative modifizierung bei pankreas opfresection. C&w-g. 1959;30:464465. 27. Bracey DW. Complete rupture of the pancreas. Br J Surg. 196 1; 48:575-576. 28. Hurwitz
REFERENCES
1987:206:358-365. 6. Peters J, Carey L. Historical Am JSurg. 1991;161:219-225.
--____
13. Trede M, Schwall G, Saeger H-D. Survival after pancreatodue denectomy: I 18 consecutive resections without an operative mortality. Ann Surg. 1990;211:447-458. 14. Ishikawa 0, Ohigishi H, Imaoka S, et al. Concomitant benefit of preoperative irradiation in preventing pancreas fistula after pancre atoduodenectomy. Arch Surg. 1991;126:885-889. 15. Papachristock DN, D’ Agostino H, Former JG. Ligation of the pancreatic duct in pancreatectomy. Br J Surg. 1980;67:260-262. 16. Buchler M, Friess H, Hermanek P, et al. The somatostatin analogue sMS 201-995 (Sandostatin) prevents postoperative complications after pancreatic resection: results from the German Multicentric Randomized Controlled and Double Blind Trial. Society for Surgery of the Alimentary Tract Annual Meeting, May 20-22,199 1. Abstract #820. 17. Telford CL, Mason GR. Pancreaticogastrostomy: clinical experience with a direct pancreatic-duct-to-gastric-mucosa anastomosis. Am J Surg. 1984;147:832-837. 18. Wells CA, Shepherd JA, Gibbon N. Pancreatogastrostomy. L.ancer. 1952;1:588. 19. Sames CP. Pancreaticogastrostomy. hncet. 1952;1:718. 20. Ingebritsen R, Langfeldt E. Pancreaticogastrostomy. Inrnc.~r. 1952;2:270-211. 21. Dill-Russell AS. Pancreaticogastrostomy. L.uncet. 1952;2:589-590. 22. Nanson EM. An unusual case of carcinoma of the head of the pancreas. Er J Surg. 1954;41:439-441. 23. Walker F. Anastomose zwischen Ductus pancreaticus und Magen (Wirsungogastrostomie). Krebsnrzt. 1955;10:83. 24. Silverstone M. Pancreaticoduodenectomy and panctreaticogastrostomy: a five-year survival, with notes on the metabolism. Er J Surg.
Pancreatogastrostomyis an invention that many qf usattribute to Mackie and Rhoads. This paper adds a dimension and depth to its acceptability that suggestsall of us shouldfind this procedure more helpful than is rqflected by its breadth of use.
PAh’CREATOGASTROSTOMYIMASON _____-____..-.-
A. In discussion: Park CD. Mackie JA. Rhoads JE. Pancreaticogastrostomy. Am J Surg. 1967;113:85-90. 29. Mackie JA, Rhoads JE. The restoration of pancreatic flow by pancreaticogastrostomy following Whipple resection. Chir Gastroenterol. 1977; 11~347-352. 30. Reding R. Die Pankreatogastrostomie als modification des Wippleschen operation. Zentrolbl Chir. 1978;103:943-946. 3 1. Flautner F, Tihanyi T, Szecseny A. Pancreatogastrostomy: an ideal complement to pancreatic head resection with preservation of the pylorus in the treatment of chronic pancreatitis. Am J Surg. 1985;150:608-611 32. Strauch GO. The use of pancreatogastrostomy after blunt traumatic pancreatic transection. Ann Surg. 1972; 176: I& 18. 33. Kapur BML. Pancreaticogastrostomy in pancreaticoduodenal resection for ampullaq carcinoma: experience in thirty-one cases. Surgery. 1986:100:489493. 34. lcard P, Dubois F. Pancreaticogastrostomy following pancreatoduodenectomy. Ann Surg. 1988;207:253-256. 35. Delcore R, Thomas JH, Pierce GE, Hermreck AS. Pancreatogastrostomy:a safe drainage procedure after pancreatoduodenectomy. Surgerv. 1990;108:641-643. 36. Ferguson DL, Wangensteen OH. Experimental anastomoses of the pancreatic duct. Ann Surg. 1950: 132: 1066-i-1074. 37. Telford GL, Mason GR. An improved technique for pancreaticogastrostomy after pancreaticoduodenectomy, Am J Surg. I98 1; 142:386-387.
1988; 207:39-47.
THE AMERICAN JOURNAL OF SURGERYm VOLUME 169 FEBRUARY 1995
219
Experience
With Pancreatogastrostomy
G. Robert Mason, MD, Robert J. Freeark, MD, Muywood Illinois BACKGROUND: The reconstruction techuique for the pancreatic remnant remaiuiug after pancreatoduodenectomy has most frqently been paucreatojejunostomy. Although the mortality rate baa been reduced to rather low levels in many centers, the leakage rate from this auastomosis remaina high, in the range of 10% or greater. An alternative reconstruction, pauereatogastrostomy, has heen known for ahnost 50 years and has been performed on small numbers of patients. The leakage rate for this auastomoais is less than 1% in literature reports in more than 200 patieuts. The purpose of this report was to add to the previously reported experience with this technique and to compare it with standard reconstruction as performed in a major American medical center by experienced surgeons. METHODS: The medical records of all patients operated on at the Loyola University Medical Center and the Edward Hines, Jr., Veteran’s Affairs Hospital from August 1986 to May 1993, with a procedure code relating to the pancreas, were reviewed. RESULTS: A total of 58 Whipple procedures were ident&ed, iucl~ 34 pancreatogastrostomies, 23 pancreatojejunostomies and 1 stapled pancreatic stump. No leaks in any pancreatogastrostomies were observed iu the 38 patients so treated, whereas 4 leaks and 2 deaths related to the auastomosis occurred in the group of 23 patients with panereatojejunostomies. The average length of stay was 14.2 days for the pancreatojejunostomy group and 15.5 days for the pancreatogastrosmmy group, excluding duration data from those who died or experienced leakage. There was no sign&ant difference in the length of stay between paucreatojejmmstomy and pancreactogactrostomy; there was a statistically signiticaut lengthening of stay for those patients whose auastomosis leaked versus those whose did not leak. The 10 patients having a pylorus-sparing operation bad an average postoperative stay of 16 days, including both types of reconstruction. CONCLUSIONS: The gradual adoption of this procedure at a major medical center has led to the abandonment of pancreatojejunostomy as a reconstruction technique for the pancreatic remnant remaining after paucreatoduodenectomy.
ince 1935, when Whipple et al’ first described a sucS cessful technique for pancreatoduodenectomy, numerous variations of this procedure have been described. The original two-stage operation was done in one stage within 5 years by Whipple? Trimble et al,3 and Brunschwige4 Brunschwig and Trimble et al both discussed their concerns, at some length, with the pancreatoenteric anastomosis, particularly for fistula formation as a complication of the procedure. Brunschwig recommended ligation of the pancreatic stump, feeling that “successful anatomic implantation of the pancreatic stump into bowel in man has not yet been demonstrated to afford normal or appreciable secretion of pancreatic juice into the bowel.” Others have felt differently and have made various anastomoses of pancreas remnant to small intestine. The mortality and morbidity rates in the best of these series has been reduced to 0% to 2%, and 36%, respectively,” but the problem of pancreatic fistulae remains prominent at 9% to 18%,“17 and may be implicated in complications such as bleeding, abscess, and sepsis. Although these leaks may not result in the death of the patient, the presence of complications in any series is likely to be related to an overall greater mortality rate and longer hospital stay. In animal experiments, Tripodi and Sherwin8 described transplantation of the pancreas into the stomach in 1934, but the procedure was not done clinically for patients in association with pancreatoduodenectomy until 1946, as reported by Waugh and Clagett’ from the Mayo Clinic. Since that time there have been sporadic reports of pancreatogastrostomy being performed with only one leak clearly related to this anastomosis, reported in 1975,‘* although there is one other report which describes a transient leak of pancreatic juice, which, in theory, could be from any of the three reconstructive anastomoses.’ ’ The purpose of this report is to add to the previously reported experience of pancreatogastrostomy for reconstruction after pancreatoduodenectomy and to compare it with standard reconstruction as performed in a major American medical center by experienced surgeons. PATIENTS
AND
METHODS
The medical records of all patients operated on at the Loyola University Medical Center and Edward Hines, Jr., Veteran’s Affairs Hospital from August 1986 to May 1993, with a procedure code relating to the pancreas, were requested from the medical records departments of the two hospitals. There were no other inclusion or exclusion criteria applied. RESULTS
From the Department of Surgery, Loyola University Medical Center and Surgical Service, Edward Hines, Jr., Veteran’s Affairs Hospital, Maywood, Illinois. Requests for reprints should be addressed to G. Robert Mason, MD, Department of Surgery, Loyola University Medical Center, 2160 South First Avenue, Maywood, Illinois 60153. Manuscript submitted October 7, 1993 and accepted in revi& form March 7, 1994.
THE AMERICAN
A total of 58 Whipple procedures were identified, including 34 pancreatogastrostomies, 23 pancreatojejunostomies, and 1 stapled pancreatic stump (Table). In the pancreatojejunostomy group there were 7 women and 16 men; in the pancreatogastrostomy group there were 10 women and 24 men. In 11 patients the pylorus was spared (8 in the pancreatogastrostomy group and 3 in the pancreatoje-
JOURNAL
OF SURGERY@
VOLUME
169
FEBRUARY
1995
217
CURRENT
EXPERIENCE
WITH
PANCREATOGASTROSTOMSON
AND
FREEARK
gastrostomiesdisrupted,leaked,or contributed to the mortality or morbidity of any patient. The 2 deathsin the pancreatogastrostomygroup were from myocardial infarction and peritonitis, asdescribed. In the pancreatojejunostomygroup, there were also 2 deaths, and these were directly related to breakdown of this anastomosis.Two other leaks led to a prolonged hospital stay. The average length of stay was 14.2 days for the pancreatojejunostomygroup and 15.5days for the pancreatogastrostomygroup, excluding duration data from thosewho died or experiencedleakage.There was no significant difference in the length of stay between pancreatojejunostomy and pancreactogactrostomy;there was a statistically significant lengthening of stay for those patients whose anastomosisleaked versus those whose did not leak. The 10 patientshaving a pylorus-sparingoperation had an averagepostoperativestay of 16 days, including both types of reconstruction.
TABLE Annual Frequency of Pancreatogastrostomy and Pancreatojejunostomy for Reconstruction After Pancreatoduodenectomy’ PancreatoAll Whipple Pancreatojejunostomy Procedures Year gastrostomy 0 3 3 1986 0 4 4 1987 2 3 5 1988 5 1989 2 3 17 1990 8 9 10 1991 8 2 1992 10 0 10 1993 4 0 4 Total 34 24 58 ‘Data from the medical records of all oatients oDerated on at the Lovola University Medical Center and the Hines Veterans Admlnistration Hospital from August 1986 to May 1993, with a procedure code relating to the pancreas. I
100
-
/’ ,’
$ e601
:
a Q) I -ii. ’ g 40 -
+
pancreatqejunostomy
:
3
75 gl 20 0’
:i
/’
: : : .I’
fi 1986
n 1967
: :
:
/ 1988
I 1989
/ 1990
I 1991
l
1992
1993
I Year Trend toward use of pancreatogastrostomy as compared to pancreatolejunostomy tar reconstructlon after pancreatoduodenectomy at one institution. Data are from the medical records of all patients operated on at the Loyola University Medical Center and the Edward Hines, Jr., Veteran’s Affairs Hospital from August 1986 to May 1993, with a procedure code relating to the pancreas. L F igure.
junostomy group); ofthese, 5 were women.One of the pancreatojejunostomypatient in this group experienceda leak, but survived. The Whipple procedureswere performed by eight different surgeons. There were 5 deaths,at 12, 14, 21, 26, and 45 days. The death of a patient with pancreatogastrostomyat 12 days was attributed to adult iespiratory distresssyndromecausedby fluid overload. The death at 14 days was by myocardial infarction associatedwith gastrointestinalbleeding in a patient with pancreatogastrostomy. The deathsat 21 and 45 days were associated with leaksfrom pancreatojejunostomy.The deathat 26 days was shown at postmortemto be related to peritonitis secondary to breakdownof the gastroenterostomyin a patient whosepancreatogastrostomywas intact. In addition to the patientswho died from leakage,therewere 2 other patients with pancreatojejunostomies who had evidence of leakage of this anastomosis, but survived, with hospital staysof 28 and 45 days. There was no evidence that the pa&reato218
THE AMERICAN
JOURNAL
OF SURGEKY”
VOLUME
169
COMMENTS The patients describedhere were not part of a prospective scientific study and were not subjectedto a randomization protocol. The data are presentedto show the gradual adoption, by a group of experienced surgeons,of pancreatogastrostomyasan alternative reconstructionprocedure to pancreatojejunostomybecauseof problems encountered with the latter (Figure). In the pancreatojejunostomy group, leakage of this anastomosisaccounted for the 2 deaths,anddoubledthe hospitalstay for the 2 patients with nonlethal leaks. The leakagerate in this group was 4 of 23 or 17%, as opposedto 0% for the pancreatogastrostomygroup. The use of an alternative reconstruction of this sort obviously doesnot eliminate the possibility of morbidity or mortality from other causes.In current reports the mortality rate for pancreatoduodenectomyin majormedicalcentersis in the areaof 7% to 9%,‘* although serieshave beenreported with no mortality.‘” The leakagerate for pancreatojejunostomyhas beenreported in the 9% to 12% range for several decadesand doesnot seemto have changedappreciably in recent years. The actual incidence of such leaks in institutions not reporting their statisticsmay well be higher and may be responsiblefor the continuing publication of articles relating to treatment and technical alternatives, such as irradiation, duct ligation, and treatment with somatostatin.‘4-‘6Seriesin the current literature still indicate relatively infrequent performance of pancreatogastrostomy, but in the reported cases,the incidence of breakdown or leakageis minimal. There are 228 casesof pancreatogastrostomy found by our searchthat have had a total of four leaks reported. As noted above, only one of theseis documented as being from the pancreatogastrostomy.Two other caseshave beenmadeknown to usby personalcommunication (G. Black, MD, 1993, and P. Donahue, MD, 1993). Five of the total of 6 caseswere resolved without operation or seriousmorbidity. In addition to the casesreported here and in the literature, we also add to the original seriesof 9 patients reported by Telford and Mason.” another 15 pancreatogastrostomies performed by the same authors, and their associates,without evidence of leak of this anastomosis. FEBRUARY
1995
CURRENT
EXPERIENCE
These procedures were performed not only in university medical centers but also in community hospitals without the usual support systems that are thought to be necessary in tertiary care and comparable referral centers. With this report, there are now 277 patients known to us as having had this procedure, with a possible leak in 6, an incidence of 2%.1*-35Given the statistics in the literature for leakage from a pancreatoenteric anastomosis, the number of leaks in a similar number of pancreatojejunostomies should be between 25 and 50. There are, however, at most, 6 cases. The anastomoses reported here include the intussusception typz of anastomosis, described by Mackie and others, as well as the duct to gastric mucosa anastomosis.‘“*36~37 Long-term patency seems more assured from mucosa-tomucosa anastomosis, based on experimental work. Clearly, both yield satisfactory results in the immediate postoperative period. In the long term, we have reoperated on 3 patients whose intussuscepted anastomosis steno& over time and produced the signs and symptoms of chronic pancrea&is. The reoperative procedure is similar to that used for the Puestow type of pancreatoenteric anastomosis and has been remarkably technically easy, reflecting the absence of adhesions, probably an indication of lack of leakage at the original operation site. The lack of reported reoperations for this problem is likely to be related to the lifespan of the patient population rather than an actual incidence.
WITH
I. Whipple AO, Parsons WB, Mullins CR. Treatment of carcinoma of the ampulla of Vater. Ann Surg. 1935;102:763-779, 2. Whipple AO. Present day surgery of the pancreas. NEJM 1942; 226515-526. 3. Trimble IR, Parsons JW, Sherman CP. A one stage operation for the cure of carcinoma of the ampulla of Vater and of the head of the pancreas. Surg Gynecol & Obstet. 1941;73:71 l-722. 4. Brunschwig A. One Stage Pancreatoduodenectomy. Surg Gynecol & Obster. 1943;77:581-586. 5. Crist DW, Sitzman JV, Cameron JL. Improved hospital morbidity, mortality, and survival after the Whipple procedure. Ann Surg. review of pancreatoduodenectomy.
7. Cameron JL, Pitt HA, Ye0 CJ. One hundred and forty-tive consecutive pancreatectomies without mortality. Ann Surg. 1993;217:430-438. 8. Tripodi AM, Sherwin CF. Experimental transplantation of the pancreas into the stomach. Arch Surg. 1934;28:345-356, 9. Waugh JM, Clagett OT. Resection of the duodenum and the head of the pancreas for carcinoma. Surgery. 1946;20:224-232. 10. Mackie JA, Rboads JE, Park CD. Pancreatogastrostomy: a further evaluation. Ann Surg. 1975;181:541-545. 1I. Bradbeer JW, Johnson CD. Pancreaticogastrostomy after pancreaticoduodenectomy. Ann R Co11 Surf: Engl. 1990; 108:64M43. 12. Trede M, Schwall G. The complications of pancreatectomy. Ann Surg.
AND -__--- FREEAHK
1956;44:299-302.
25. Millboume E. Pancreaticogastrostomy in pancreaticoduodenal resection for carcinoma of the head of the pancreas or the papilla of Vater. Acta Chir Scand. 1958; 116: 12-27. 26. Pataky ZS, Popik E. Operative modifizierung bei pankreas opfresection. C&w-g. 1959;30:464465. 27. Bracey DW. Complete rupture of the pancreas. Br J Surg. 196 1; 48:575-576. 28. Hurwitz
REFERENCES
1987:206:358-365. 6. Peters J, Carey L. Historical Am JSurg. 1991;161:219-225.
--____
13. Trede M, Schwall G, Saeger H-D. Survival after pancreatodue denectomy: I 18 consecutive resections without an operative mortality. Ann Surg. 1990;211:447-458. 14. Ishikawa 0, Ohigishi H, Imaoka S, et al. Concomitant benefit of preoperative irradiation in preventing pancreas fistula after pancre atoduodenectomy. Arch Surg. 1991;126:885-889. 15. Papachristock DN, D’ Agostino H, Former JG. Ligation of the pancreatic duct in pancreatectomy. Br J Surg. 1980;67:260-262. 16. Buchler M, Friess H, Hermanek P, et al. The somatostatin analogue sMS 201-995 (Sandostatin) prevents postoperative complications after pancreatic resection: results from the German Multicentric Randomized Controlled and Double Blind Trial. Society for Surgery of the Alimentary Tract Annual Meeting, May 20-22,199 1. Abstract #820. 17. Telford CL, Mason GR. Pancreaticogastrostomy: clinical experience with a direct pancreatic-duct-to-gastric-mucosa anastomosis. Am J Surg. 1984;147:832-837. 18. Wells CA, Shepherd JA, Gibbon N. Pancreatogastrostomy. L.ancer. 1952;1:588. 19. Sames CP. Pancreaticogastrostomy. hncet. 1952;1:718. 20. Ingebritsen R, Langfeldt E. Pancreaticogastrostomy. Inrnc.~r. 1952;2:270-211. 21. Dill-Russell AS. Pancreaticogastrostomy. L.uncet. 1952;2:589-590. 22. Nanson EM. An unusual case of carcinoma of the head of the pancreas. Er J Surg. 1954;41:439-441. 23. Walker F. Anastomose zwischen Ductus pancreaticus und Magen (Wirsungogastrostomie). Krebsnrzt. 1955;10:83. 24. Silverstone M. Pancreaticoduodenectomy and panctreaticogastrostomy: a five-year survival, with notes on the metabolism. Er J Surg.
Pancreatogastrostomyis an invention that many qf usattribute to Mackie and Rhoads. This paper adds a dimension and depth to its acceptability that suggestsall of us shouldfind this procedure more helpful than is rqflected by its breadth of use.
PAh’CREATOGASTROSTOMYIMASON _____-____..-.-
A. In discussion: Park CD. Mackie JA. Rhoads JE. Pancreaticogastrostomy. Am J Surg. 1967;113:85-90. 29. Mackie JA, Rhoads JE. The restoration of pancreatic flow by pancreaticogastrostomy following Whipple resection. Chir Gastroenterol. 1977; 11~347-352. 30. Reding R. Die Pankreatogastrostomie als modification des Wippleschen operation. Zentrolbl Chir. 1978;103:943-946. 3 1. Flautner F, Tihanyi T, Szecseny A. Pancreatogastrostomy: an ideal complement to pancreatic head resection with preservation of the pylorus in the treatment of chronic pancreatitis. Am J Surg. 1985;150:608-611 32. Strauch GO. The use of pancreatogastrostomy after blunt traumatic pancreatic transection. Ann Surg. 1972; 176: I& 18. 33. Kapur BML. Pancreaticogastrostomy in pancreaticoduodenal resection for ampullaq carcinoma: experience in thirty-one cases. Surgery. 1986:100:489493. 34. lcard P, Dubois F. Pancreaticogastrostomy following pancreatoduodenectomy. Ann Surg. 1988;207:253-256. 35. Delcore R, Thomas JH, Pierce GE, Hermreck AS. Pancreatogastrostomy:a safe drainage procedure after pancreatoduodenectomy. Surgerv. 1990;108:641-643. 36. Ferguson DL, Wangensteen OH. Experimental anastomoses of the pancreatic duct. Ann Surg. 1950: 132: 1066-i-1074. 37. Telford GL, Mason GR. An improved technique for pancreaticogastrostomy after pancreaticoduodenectomy, Am J Surg. I98 1; 142:386-387.
1988; 207:39-47.
THE AMERICAN JOURNAL OF SURGERYm VOLUME 169 FEBRUARY 1995
219