World J. Surg. 13, 84-87, 1989
World Journal of Surgery 9 1989 by the Soci~t~ lnternationale de Chirurgie
Original Scientific Reports Colonic Involvement in Acute Neerotizing Pancreatitis: Results of Surgical Treatment J.L. Bouillot, M . D . , J . H . A l e x a n d r e , M . D . , and N.P. V u o n g , M.D. Department of Surgery, Broussais Hospital, and Department of Anatomopathology, Saint Michel Hospital, Paris, France In this series of 15 personal cases, the authors emphasize the unusual colonic complications which occur during acute necrotizing pancreatitis. These lesions always indicate a particularly severe pancreatitis and depend on 2 factors: extension of pancreatic necrosis into the mesocolon with encasing pericolic tumoral fibrosis, and parietal ischemic necrosis secondary to shock and thrombosis with infection. Laparotomy followed by colectomy is indicated in these severely ill patients. Prognosis is poor (8 deaths of 15 patients), sometimes in spite of extensive pancreatic excision. The existence of colonic complications gives evidence of the particular severity of certain forms of pancreatitis.
Indications for surgical operations were based on clinical status and secondarily on clinical and biological gravity criteria in spite of intensive resuscitation. The average day of operation from the onset of the disease was 9.7. Seven patients had been transferred due to the severity of their disease after a surgical procedure had been performed at another institution. Results
Macroscopic Findings (Table 1) In the course of acute necrotizing pancreatitis, extension of tissue necrosis beyond the pancreatic bed is a common occurrence and makes a major contribution to the high morbidity and mortality of the disease. Among these extrapancreatic lesions, it is estimated that colonic involvement affects only 1% of acute pancreatitis patients [1]. It may take the form of pericolic infiltration, necrosis of the colon wall [2, 3], colonic perforation [4], or fistulization of a pancreatic pseudocyst into the colon [5, 6]. These complications call for an aggressive surgical management to eradicate all necrotic and often infected foci. This article reviews our experience with these complications of acute pancreatitis. Material and Methods
From 1974 to 1984, a total of 100 consecutive patients with acute necrotizing pancreatitis necessitating surgical therapy were managed in the Department of Gastrointestinal Surgery at Broussais Hospital. Fifteen of these patients presented with colonic involvement, which was sufficiently severe to entail a specific surgical procedure. There were 14 men and 1 woman; their ages varied from 23 to 77 years (average, 51 yr). The etiology of the pancreatitis was: gallstones (5 cases), alcohol (5 cases), posttraumatic (2 cases), postsurgical (2 cases: gastrectomy and common bile duct obstruction), metabolic (1 case). Using the criteria of Ranson et al. [7], the average score was 4.2, but all signs were not evaluated in each case.
Reprint requests: J.L. Bouillot, M.D., Hopital Broussais, 96 rue Didot, 75674 Paris Cedex 14, France.
The operative procedure included cholecystectomy, biliary drainage, and wide exposure of the pancreatic gland by debridement. The operative management of pancreatic lesions consisted of open lesser sac drainage (3 cases), necrosectomy (1 case), distal pancreatectomy (l case), and total pancreatectomy (10 cases). The colon lesions were located in the right colon in 7 cases, in the transverse colon in 11 cases, and in the left colon in 2 cases (more than 1 colon segment was involved in 5 patients). At laparotomy, colonic perforation was apparent in 3 patients; in the 12 other patients, the lesions involved the mesocolon and the serosa of the colon (causing a true encasement of the colon in 7 cases and simulating a necrotizing process in 5 cases). These lesions were treated at the same time as the pancreatic lesions except for 1 patient in whom a perforation appeared on the seventh day. Eight patients had colectomies with double colostomies performed, and 5 patients had colectomies with immediate anastomosis. In 1 case, the perforation was exteriorized primarily, but a secondary colectomy was necessary for a recurrent perforation; in the remaining patient, Simple suture of a perforation of the transverse colon was performed. Only 6 patients underwent 1 surgical procedure, all the others needed between 2 and 6 operations for postoperative complications. In 2 cases, the colectomy has had to be extended secondarily because of the extension of the colonic lesions observed during another surgical procedure.
Histological Findings Pathological examination of the colectomy specimens showed 2 types of lesion: (a) Peripheral lesions, or true pericolitis with
J.L. Bouillot et al.: Acute Necrotizing Pancreatitis
85
Table 1. Patients with pancreatitis:
Age (yr), sex
Ranson's criteria
Etiology
Colonic lesions
Pancreatic Localization surgery
1 2 3 4 5
47, 23, 49, 45, 64,
M M M M M
8 4 4 2 2
Gallstones P0sttraumatic Postsurgical Gallstones Alcohol
Colonic surgery
Outcome
Pericolitis Pericolitis Pericolitis Pericolitis Pericolitis
T + L T T T R
TP TP TP TP DP
Colectomy Colectomy Colectomy Colectomy Colectomy
Died Alive Died Died Alive
Metabolic Gallstones Alcohol
Total necrosis Total necrosis Total necrosis Total necrosis Superficial necrosis Total necrosis Total necrosis Total necrosis
6 7 8
61, M 34, M 34, M
6 7 3
Pericolitis Pericolitis Pericolitis
R+ T T + L T
TP TP TP
Died Died Alive
4
AlcohOl
Total necrosis
Pericolitis
R+ T
TP
61, M
3
Alcohol
Pericolitis
R
DP
11 12
76, M 42, F
4 4
Gallstones Posttraumatic
Superficial necrosis Total necrosis Tail necrosis
Pericolitis Pericolitis
R + T R
TP N
13
42, M
2
Alcohol
T
DP
14
77, M
5
Gallstones
R
DPS
15
75, M
7
Postsurgical
Performati0n by ischemic necrosis Perforation by ischemic necrosis Perforation by ischemic necrosis
T
TP
Colectomy Colectomy Colectomy with anastomosis Colectomy with anastomosis Colectomy with anastomosis Coiectomy Colectomy with anastomosis Colostomy, then colectomy Colectomy with anastomosis Suture
9
34, M
10
Case
Pancreatic lesions
Superficial necrosis Body and tail necrosis Total necrosis
Alive Alive Died Alive Alive Died Died
T = transverse colon, L = left colon, R = right colon, TP = total pancreatectomy, DP = drainage procedure, N = necrosectomy, DPS = distal pancreatectomy with splenectomy. involvement of external layers: mainly serosa and subserosa with sparing of the inner layers (12 cases). These lesions were characterized by a necrosis of fatty tissue, with a large necrotic area infiltrated by fat cells. Around this necrosis were many mononuclear and giant cells. (b) Necrotic and hemorrhagic lesions with perforation (3 cases): The microscopic examination showed an ischemic necrosis involving the 3 layers--mucosa, stibmucosa, muscularis. Some vessels were obliterated by recent thrombi. The perforation was delimited by inflammatory tissue with many altered polynuclear cells.
Morbidity and Mortality Eight of the 15 patients died between the first and the eighteenth postoperative day (operative mortality, 53.3%). The average score, according to Ranson, was 5.4. None of these deaths seemed to be related to the colonic procedure performed: 2 patients died because of persistent infection with recurrent intraabdominal necrosis; other deaths were secondary to anuria (1 case), severe hypoxemia (1 case), massive blood transfusions (1 case), multivisceral failure (1 case), unknown (1 case). The patient who underwent a suture of colonic perforation died from leakage of portal vein suture, which was performed at initial laparotomy because of rupture of the portal vein secondary to a massive pancreatic necrosis. Of the 5 patients who underwent immediate anastomosis, 1 died and postmortem examination showed no evidence of anastomotic leakage; in another patient, medical therapy allowed cure of the colonic fistula. Postoperative complications occurred in 6 of the 7 patients alive (intestinal leakage: 5 cases, pancreatic fistula: 1 case, gastrointestinal bleeding: 2 cases, intraperitoneal bleeding: 1 case, pulmonary complications: 3 cases, subphrenic abscess: 1 case). The length of stay in the hospital varied from 1 to 6 months.
Discussion
Necrotizing pancreatitis with coionic involvement is rare since, by 1979, only 75 cases had been reported [8]. The largest series show only a few cases [8-11]. The large number of our study reflects the particular severity of our patients, often transferred from other hospitals because of the extension of the pancreatic necrosis for which the only surgical alternative is total pancreatectomy [12, 13]. Nevertheless, in recent years, we did not meet any more of such lesions, possibly because of an improvement in the medical management of Pancreatitis, and reduction of indications for early surgical operation. From the pathological point of view, colonic involvement can manifest itself in 3 ways: (a) A mesocolic mass secondary to necrosis of fatty tissue causing an external inflammatory compression. None of our 15 patients had a mesocolic mass. These lesions do not, in themselves, justify laparotomy since appropriate medical treatment would, in most cases, permit normal anatomical repair in a few weeks, perhaps at the cost of delayed stenosis [14]. (b) Pericolitis with involvement of only the parietal layers: these lesions are in direct contiguity with pancreatic necrosis. Because of the close anatomical relation between the pancreas and the transverse mesocolon, necrosis can develop and extend between the peritoneal layers [3, 4]. It can result in a true encasing of the colon without involvement of the musculariayers and changes in the submucosa and mucosa limited to edema [2, 15]. (c) Ischemic necrosis of the colon wall was observed in 3 of our patients: It is usually complicated by perforation which necessitates emergency surgery. Its pathogenesis is explained by a low blood flow or diffuse thrombosis with necrotizing enterocolitis resulting in ischemic necrosis of the entire colon wall leading to perforation. Several authors advance the hypothesis that these lesions are related to septic
86
phenomena, but we did not succeed in assessing this explanation [2, 10]. The preoperative diagnosis o f colonic involvement is difficult except in cases of perforation. Practically, it is diagnosed during laparotomy undertaken when clinical (fever, abdominal tenderness, jaundice, ileus) or biological signs of gravity appear during the course of the disease that may evocate a complication of the pancreatitis, it occurs toward the end of the first week. At this stage, the necrotic lesions are most often obvious and permit detection of pancreatic and peripancreatic lesions, allowing appropriate surgical procedures. The place of computed tomographic scan is not yet established in detecting such lesions [16]. The decision for resection of necrotic tissue is made after complete evaluation of the lesions. We resect ali areas of necrosis, performing total pancreatectomy only when necrosis involves more than 75% of the pancreatic tissue [17]. If less tissue is involved, we used to perform necrosectomy and open lesser sac drainage [18]. Concerning the colon, the decision for resection is easy in the case of perforation. In the case of colonic encasement; it is difficult, in the absence of colonoscopy, to assert the integrity of the mucosa. In theory, in this situation, one might be able to abstain from resection because of the possible regression of these lesions documented by repetitive colonoscopies [14]. However, the decision in the operating room is difficult because true pericolitis may simulate a real necrosis (5 cases in our series), and only histological examination can assert t h e integrity of the colon wall. In addition, the lesions are evolutive, and a complementary colectomy was secondarily necessary in 2 patients. A colectomy allows an excision of all necrotic and infected areas in order to be sure that only intact tissues remain. In the absence of resection, there is a risk of secondary abscess or fistula, which has led other authors to undertake such an approach [5, 6]. In doubt, it seems advisable to perform the colostomy, allowing a direct vision of the colon and limiting the resection for judicious cases. Although we did not note special mortality with immediate anastomosis, such an approach is open to criticism and cannot be recommended [19]. Conclusion
The prognosi s of colonic complications during the evolution of acute necrotizing pancreatitis remains severe. There were 8 deaths of 15 patients in our report. It seems that the severity of these cases is more a result of important pancreatic lesions (with hemorrhagic and infectious components) than a result of colonic lesions. These cases only give evidence of the severity of the necrotizing process. No patient in our experience died because of colonic complications, including the cases in whom an immediate anastomosis was performed. This must be an encouragement to eradicate necrotic lesions, wherever they are, even if some patients appear to be beyond surgical help. R~sum6
A partir d'une s6rie de 15 cas, les complications coliques peu habituelles observ6es pendant la pancr6atite aigu6 sont analyS6es. Ces 16sions indiquent toujours une pancr6atite s6v6re et sont en rapport avec 2 facteurs: extension de la nfcrose pancr6atique dans le m6soc61on entourant le c61on transverse
World J. Surg. Vol. 13, No. 1, Jan./Feb. 1989
avec de la fibrose pseudotumorale, et n6crose pari6tale isch6mique secondaire au choc et ~ des thromboses infectieuses. Une colectomie est indiqu6e chez ces patients. Le pronostic est mauvais (8 morts), parfois malgr6 des r6sections pancr6atiques 6tendues. L'existence de complications coliques au cours de certaines formes de pancr6atites est un indice de s~v6rit6. Resumen
En esta serie de 15 casos personales, los autores presentan una rara complicaci6n col6nica de la pancreatitis aguda necrotizante. Tales lesiones, que siempre indican una pancreatitis particularmente severa, dependen de 2 factores: la extensi6n de la necrosis pancrefitica al mesocolon con encapsulamiento y fibrosis tumoral peric61ica, y la necrosis isqu6mica parietal secundaria al shock y trombosis con infecci6n. La laparotomfa seguida de colectomfa estfi indicada en estos pacientes criticamente enfermos. E1 pron6stico es pobre (8 muertes en 15 pacientes), a pesar de la resecci6n pancre~ttica extensa. L a existencia de complicaciones col6nicas es evidencia de la particular agresividad de ciertas formas de pancreatitis. References
I. Lukasch, W.M.: Complications of acute pancreatitis. Arch. Surg. 94:848, 1967 2. Fiament, J.B., Pluot, M., Convers, G., Delattre, J.F., Rives, J.: N6crose colique au cours de la pancr6atite aigue n6crotico-hfmorragique. Med. Chir. Dig. 10:31, 1981 3. Katz, P., Dorman, M.J., Aufses, A.H.: Colonic necrosis complicating post operative pancreatitis. Ann. Surg. 179:403, 1974 4. Fazio, V., Cady, B.: Spontaneous perforation of the transverse colon complicating acute pancreatitis. Report of a case. J. Dis. Colon Rectum 17:106, 1974 5. Berne, T. V., Edmonson, A.: Colonic fistulization due to pancreatitis. Am. J. Surg. 111:359, 1966 6. Chaikhouni, A., Regueyra, F . I , Stevens, J.R., Tidrick, R.T= Colonic fistulization in pancreatitis. Case report and literature review. Dis. Colon Rectum 23:27I, 1980 7. Ranson, J.H.C., Rifkind, K.M., Roses, D.F., Fink, S.D., Spencer, F.C.: Prognostic signs and the role of operative management in acute pancreatitis. Surg. Gynecol. Obstet. 139:69, 1974 8. Abcarian, H., Eftaika, M., Kraft, A.R., Nyhus, L.M.: Colonic complications of acute pancreatitis. Arch. Surg. 114:995, 1979 9. Alexandre, J.H., Germain, M.: La n6crose colique au cours des pancr~atites aigueso Ann. Chir. 26:857, t972 10. Dubois, A., Charras, A., Lapeyrie, H., Janbon, C.: N6crose colique et pancr6atite aigue nfcrotico-h6morragique. Gastroenterol. Clin. Biol. 6:510, 1982 11. Leborgne, J., Pannier, M., Leneel, J.C., Potiron, L., Visset, J.: L6sions coliques au cours des pancr6atites n6crosantes. Apropos de 4 observations. Ann. Chir. 30:377, 1976 12. Alexandre, J.H., Camilleri, J.P., Assan, R., Guerrieri, M.T., Bonan, A.: Indications et r6sultats de la pancr6atectomie totate dans le traitement des pancr6atites aigues n~crosantes. Chirurgie 103:858, 1977 13. Alexandre, J.-H., Guerrieri, M.T.: Role of pancreatectomy in the treatment of necrotizing pancreatitis. World J. Surg. 5:369, 1981 14. Mann, N.S.: Colonic involvement in pancreatitis. Am. J. Gastroenterol. 73:357, 1980 15. Champetier, J., Bouchet, Y., Brabant, A., Guignier, M., Durand, A., Charignon, G., Corallo, J.: L'extension au colon de la n6crose pancr6atique. Lyon Chir. 70:196, 1974 16. Ranson, J.H.C., Balthazar, E., Lavallace, R., Cooper, M.: Computed tomography and the pi'ediction of pancreatic abscess in acute pancreatitis. Ann. Surg. 201:656, 1985 17. Vuong, N.P., Guerrieri, M.T., Alexandre, J.H., Camilleri, J.P.: Early histological changes in acute necrotizing hemorrhagic pancreatitis. Pathol. Res. Pract. 178:273, 1984
J.L. BouiUot et al.: Acute Necrotizing Pancreatitis
87
18. Alexandre, J.H., Bouillot, J.L., Billebaud, T., Dupin, P., Dhote, J.: Traitement chirurgical conservateur des pancr~atites aigues. Nouv. Presse Med. 13:1439, 1984
19. Russel, M.D.: Colonic complications of acute pancreatitis and pancreatic abscess. Am. J. Surg. 146:558, 1983
Invited Commentary
continued observation and support of patients with massive peripancreatic necrosis for a prolonged period of time, followed by necrosectomy and external drainage when necessary, results in a mortality rate of less than 5% in the critically ill group. Among our last 31 such patients who required delayed laparotomy, 3 patients (10%) developed recognizable colonic complications. One of these patients, 4 months after onset, was demonstrated to have a fistula between the retroperitoneal peripancreatic necrotic cavity and the descending colon. This patient was subsequently found to have developed a stricture of the descending colon. He survived following several operations including a delayed resection of the colon. Two other patients developed a fistula of the descending colon, 2-3 months after onset, along Penrose drains which had been brought out through a retrocolic stab wound at the time of delayed necrosectomy and external drainage of the peripancreatic area. One of the latter 2 required a temporary colostomy. All 3 patients survived and recovered. As the authors point out, the peripancreatic necrosis extends not only retroperitoneally toward the diaphragm and pelvis, but also between the leaves of the mesocolon and the mesentery of the small bowel. Whether colonic resection is justified in the management of "pericotitis" under these conditions remains to be demonstrated. We adhere to a more conservativ e approach to both the pancreatic lesion and the visceral complications of the peripancreatic necrosis and believe that the resultant mortality rate will prove to be lower in the majority of surgical clinics.
John M. H o w a r d , M.D. Department of Surgery, Medical College of Ohio, Toledo, Ohio, U.S.A.
During the decade of 1974-1984, Bouillot and colleagues treated 100 patients in whom operative therapy was deemed necessary. This is an impressive experience and all indications point to the fact that the patients had very severe disease. Fifteen (15%) of their patients were found at operation to have involvement of the colon largely by a peripancreatic inflammatory process. Ischemic perforation of the colon was found in 3 instances. The authors assumed that perforations resulted from ischemia secondary to the low flow state of acute pancreatitis. Since 1984, improved supportive care has been successful in preventing this complication. In the 12 patients without perforation, the resected colon was found histologically to have a "pericolitis" involving mainly the serosa and subserosa, but sparing the inner layers. This process was noted at an average time of 9.7 days after onset. Primary colectomy was performed in the majority of their patients with colonic involvement. Ten of the 15 patients simultaneously underwent totalpancreatectomy. Eight of their 15 patients died within the first 3 weeks after operation. Other Parisian surgeons [1, 2] have demonstrated that histological and gross examination of the pancreas, resected because of acute necrotizing pancreatitis, revealed the necrosis to characteristically involve only the outer layers of the pancreas. Comparably, Bouillot and colleagues have demonstrated that the involvement of the resected colon in the majority of patients was limited to the outer layers of this organ. In the absence of perforation, which we have not recognized at an early stage, we should be hesitant to resect the colon early in the course of acute pancreatic necrosis. In our experience ,
References
1. Leger, L., Chicke, B., Louvel, A.: Necrosis in acute pancreatitis: Surgical and anatomopathologicalcomparisons. Seven cases. N0uv. Presse Med. 6:337, 1977 2. Testas, P.: The treatment of acute pancreatitis. Presented at the Conjoint Meeting of the Royal College of Surgeons in Edinburgh, Association Francaise de Chirurgie, and The Egyptian Surgical Society, Edinburgh, Scotland, May 28, 1987