- Email: [email protected]
Pancreatic pseudocysts following acute pancreatitis
Pancreatic
Pseudocysts
Acute Stephen
W. Behrman,
Pancreatitis
MD, W. Scott Melvin, MD, E. Christopher
BACKGROUND: Pancreatic pseudocysts (PP) following acute pancreatitis have traditionally been approached by observation to allow cyst maturation. However, recent evidence suggests a selective approach toward management is indicated. METHODS: We retrospectively reviewed the presentation, operative management, and outcome of patients developing PP following acute pancreatitis since 1988. PP related to chronic pancreatitis were excluded. RESULTS: Twenty-seven patients were identified, 17 with giant PP (>lO cm) and 10 with PP less than 10 cm. Groups were not different with respect to age and etiology of pancreatitis, predominantly biliary. Patients with giant PP had a significantly greater number of Ranson criteria at presentation and underwent drainage procedures earlier following their initial presentation. With respect to giant PP, 7 patients underwent internal drainage all via cystogastrostomy with 5 complications. Nine of 10 patients underwent urgent operation via external drainage. There were 3 pancreatic fistulas in this group. The morbidity and mortality rates for giant PP were 65% and 18% respectively. With respect to smaller PP, 8 underwent internal drainage with 1 death (mortality rate 10%). External drainage was performed in 2 patients with 1 pancreatic fistula (morbidity 10%). CONCLUSIONS: Patients with PP and a high Ranson score following acute pancreatitis are at significant risk for giant PP formation. Expectant management of giant PP is associated with higher morbidity and mortality than small PP suggesting that earlier external drainage, before clinical deterioration, may be beneficial. To be accurate, comparisons of outcomes for various treatment modalities must take into consideration PP size. Am J Surg. 1996; 172:228-231.
T
he management of pancreatic pseudocysts developing following acute pancreatitis has traditionally involved a period of expectant, nonoperative management. This allows time for cyst wall maturation in order that internal drainage may he accomplished, which, in genFrom the Department of Surgery, Grant Medical Center and The Ohio State University, Columbus, Ohio. Requests for reprints should be addressed to Stephen W. Behrman, MD, Assistant Clinical Professor of Surgery, UCSFNalley Medical Center, 445 S. Cedar Avenue, Fresno, California 93702. Manuscript submitted March 8, 1995 and accepted in revised form October 3, 1995.
228
0 1996 by Excerpta All rights reserved.
Medica,
Inc.
following Ellison, MD, Columbus, Ohio
eral, is associated with lower morbidity than external drainage.’ In addition, a certain proportion of pseudocysts may be expected to resolve spontaneously during this interval of time. In contrast, it is felt that cysts associated with chronic pancreatitis have a more indolent natural history and may be internally drained immediately if indicated.‘s3 In the past it was recommended that all pseudocysts failing resolution by 4-6 weeks undergo operative drainage in order to prevent complications of infection, bleeding, and rupture.4 More recently, this concept has been challenged and a nonoperative approach has been recommended in select patients with evidence of resolution occuring with longer periods of follow-up.5,” Unfortunately, interpretation of the literature has been clouded by series that include patients with both acute and chronic pancreatitis with uniform management being recommended without regard to degree of pancreatic disease. In addition, the natural history of pseudocysts has not been well correlated with either the severity of pancreatitis at presentation or size of cyst. Our recent experience suggests that both the severity of pancreatitis at presentation, and the size of subsequent cyst development, have a significant impact on the natural history and subsequent management of pancreatic pseudocysts.
PATIENTS
AND
METHODS
The hospital records of all patients admitted to Grant Medical Center and the Ohio State University Hospitals with a CPT coding of pancreatic pseudocyst, or patients having surgical procedures for such, from January 1988 through September of 1993 were examined. Via intraoperative records and, when performed, dynamic pancreatography, we were careful to exclude patients with underlying pancreatic necrosis. Fifty-seven patients were identified. Patients with chronic pancreatitis were excluded. All pseudocysts were followed by computed tomography. Records were examined with respect to etiology of pancreatitis, Ranson criteria, pseudocyst size, pre-, post-, and total hospital days, preoperative laboratory values, reason for operation, type of operative procedure, and postoperative complications. Definitions Chronic pancreatitis included those patients with evidence of exocrine and endocrine insufficiency together with any of the following: pancreatic calcifications, ERCP findings of irregular caliber of the duct of Wirsung, and a history of chronic pain. In addition, pseudocysts occuring in patients with no identifiable antecedent episode of acute pancreatitis were classified as chronic pseudocysts. Giant pancreatic pseudocyst was defined as those equal to or greater than 10 cm in diameter by CAT scan analysis as described in prior work by Johnson et aL7 0002-961 O/96/$1 PII 50002-961 O(96)OOi
5.00 57-2
J
1PANCREATIC
PSEUDOCYSTWBEHRMAN
ET AL
TABLE Etiology
of Pancreatitis Pseudocyst
Alcohol Biliary Idiopathic Pancreas divisum Hypertriglyceridemia Drug related
size
Figure 1. Correlation pseudocyst size.
urgent operation
n=17
2. Operative
management
r10
cm
2 5 3 -
3 7 3 2 1
(n = 10)
:n = 17)
of pseudocyst
between
cm
Size
Figure cysts.
(cm)
number
of Ranson
criteria
and
elective operation
J Figure cysts.
of giant
pancreatic
pseudo-
Data presented are the mean plus or minus the standard deviation of the mean. Comparison between groups was undertaken utilizing an unpaired Student’s t-test and, where appropriate, analysis of variance. Significance was assessed at the 95% confidence interval.
RESULTS Fifty-seven patients developing pseudocysts were identified. Pseudocysts secondary to chronic pancreatitis developed in 30 patients and were eliminated from this analysis. Twenty-seven patients developed pseudocysts following acute pancreatitis, 17 with giant PP and 10 with smaller PP (average size, 14.9 * 3.0 vs 7.7 2 1.9 cm, respectively). Etiology of pancreatitis was not different between groups with approximately 45% being due to gallstone disease (Table). There was no correlation between the etiology of pancreatitis and the subsequent development of complications during expectant management. There was no difference in THE AMERICAN
3. Operative
management
of smaller
pancreatic
pseudo-
age between those with giant PI’ and those with smaller PI’ (54.1 i 17.9 vs 51.7 -C 23.2 years, respectively). All patients underwent operative management. Those developing giant PI? had a significantly greater number of Ranson criteria (5.3 i 2.4 vs 2.0 z 2.1; P = .002). The number of Ranson criteria directly correlated with increasing pseudocyst size (r’ = .1, P = .OlO; Figure 1) . Those with giant pseudocysts underwent definitive operative intervention earlier than those with smaller cysts (49.2 2 39.0 vs 74.3 -t- 55.4 days, respectively), a trend that approached but did not reach statistical significance. More than one-half of the giant PI’ underwent urgent drainage due to cyst complications, whereas those with smaller cysts, in general, safely underwent expectant management to allow cyst maturation and elective internal drainage. All patients but 1 undergoing urgent operation were managed by external drainage whereas those having elective therapy had internal drainage. There was no difference in preoperative laboratory values, or postoperative length of stay between groups. We specifically analyzed results of operative therapy in patients with giant PP (Figure 2). Nine of 10 patients underwent urgent operation via external drainage: 6 for infected pseudocysts, 2 for rapid enlargement, and 1 for bleeding. We could not identify any risk factors for the development of pseudocyst complications in those with giant cysts. External drainage was performed in all patients undergoing urgent operation and electively in 1 patient for lack of adequate dependent drainage by cystoenterostomy. Drainage was via triple lumen Davol sump drains in all but two cases. There were 3 pancreatic fistulas in this group, all closing with conservative measures, which included the use of sandostatin with or without total parenteral nutrition. Two patients developed peripancreatic abscess. In both, closed suction rather than active sump drainage was employed. Three patients died; one due to pulmonary embolism. The course of the other two patients was complicated by ischemic colitis and progressive sepsis following external drainage of an infected and 1 bleeding pseudocyst. Five of 7 patients with giant PI? having internal drainage suffered complications. All patients in this group had undergone cystogastrostomy. Three developed retroperitoneal sepsis that required emergency re-operation and subsequent external drainage, all with good results. There were no complications following re-exploration. Two patients suffered upper gastrointestinal hemorrhage requiring transfusion. Both had negative arteriograms and were managed conservatively. Overall morbidity for giant PP was 65% and mortality 18%. Excluding complications resulting from cystogastrostomy and closed suction drainage, morbidity was 24%. JOURNAL
OF SURGERY@
VOLUME
172
SEPTEMBER
1996
229
PANCREATIC
PSEUDOCYSTWBEHRMAN
ET AL /
With respect to smaller pseudocysts, 8 of 10 patients were successfully managed expectently and underwent elective internal drainage, 6 via cystogastrostomy (Figure 3 ) . There was 1 death following cystogastrostomy. This patient developed retroperitoneal sepsis following drainage of an 8cm pseudocyst and expired despite reexploration and external drainage. Two patients failed expectent management and safely underwent external drainage for infected pseudocysts. There was 1 pancreatic fistula in this group, which closed with conservative management. Overall morbidity for smaller PP was 10% and mortality 10%.
DISCUSSION Pancreatic pseudocysts have traditionally been approached by a uniform method of management. A period of expectant observation to allow for cyst wall maturation is followed by elective internal drainage in 4-6 weeks, patients often requiring hospitalization during this entire time with or without total parenteral nutrition.’ Delay of operative intervention beyond this period has been associated with a poor chance of cyst resolution and a significant risk for the development of pseudocyst complications including infection, bleeding, and rupture.’ More recently, these concepts have been challenged and studies have suggested that many of these patients may he safely managed nonoperatively with a significant proportion either resolving or decreasing in size, and with virtually no morhidity.5,” In addition, evidence suggests that pseudocysts associated with chronic pancreatitis behave, and should be managed differently, than those associated with acute pancreatitis.‘,’ Furthermore, pseudocyst size appears to he an important variable with respect to operative management.’ It is becoming clear that the natural history and operative management requires a selective approach based on underlying pancreatic disease, the severity of pancreatitis, and the size of pseudocyst formation. Unfortunately, interpretation of the literature with respect to management of pancreatic pseudocysts is complicated by lack of characterization of the above. Our study suggests that giant pancreatic pseudocysts (> 10 cm) developing following acute pancreatitis have a different natural history and require new approaches toward management. Pseudocysts developing in those with chronic pancreatitis appear to do so in a more indolent nature.“.’ In general, cell wall maturation has occurred, and immediate internal drainage may be performed. The rate of spontaneous resolution in this group IS low secondary to ductal stricture, and preoperative analysis of pancreatic and biliary ductal anatomy via ERCP is indicated. K~‘c) On the basis of these data, PP arising in those with chronic pancreatitis have a natural history, and require modifications in therapeutic interventions different from those with acute pancreatitis. With regard to PP complicating acute pancreatitis, cyst wall maturation is necessary if internal drainage is to he successful. Further, concern for late complications still exists unless operative intervention supervenes.’ Early operation and external drainage of PP is considered less appealing due to complications of catheter drainage consisting mainly of pancreatic fistula and abscess, both significantly prolonging postoperative management. Recently, longterm nonoperative management of PP has been advocated. Prospective studies by Sarr et al and Yeo et al have dem230
THE
AMERICAN
JOURNAL
OF SURGERYa
VOLUME
172
onstrated impressive rates of spontaneous resolution at periods longer than 6 weeks from presentation.5’6 Morbidity has been minimal, although in one series there were several life-threatening complications necessitating emergent operation.” Nonetheless, this approach is appropriate based on selective management. Both studies fail to group PI? based on whether they resulted from acute or chronic pancreatitis, and the severity of pancreatitis at presentation with respect to management is not addressed. Both studies suggest that resolution of pseudocysts greater than 5-6 cm is poor. Therefore, guidelines for management of PP following acute pancreatitis need to be addressed since clearly this form of therapy is not indicated for all. Our study has identified a group of patients that are at high risk for farlure of nonoperative therapy. In this series, patients developmg giant PP following acute pancreatitis had a more signiticant degree of pancreatitis based on Ranson criteria with a mean of 5. Prior studies have associated this number of Ranson criteria with a mortality rate of 40-50%.” In a&t’ I Ion, a significant proportion of these patients failed a 4-6-week period of expectant management. Furthur, approximately 65% of these giant PP developed complications, the maiority of which were related to attempts at internal drAinage via cystogastrostomy. In contrast to prior studies, we found no correlation between etiology of pancreatitis and the development of pseudocyst complications.“.” In summary, those with giant pseudocysts present with severe pancreatitis followed by the early development of large pseudocysts, which have a high propensity for the development of complications and, therefore, failure of expectant management. Analysis of our operative therapies and outcomes are noteworthy for several points. First, early operation (ie, failure of expectant management) was necessary in over 50% of these patients, suggesting that perhaps earlier intervention before complications occur might yield lower morbidity and mortality particularly since many of these patients are in an older age group. Indeed, 2 of the 3 deaths in this group occured in elderly patients having emergency operations for pseudocyst complications. Second, external drainage of giant PP is a safe and effective therapy with acceptable morbidity predominantly m the form of pancreatic fistula. Thirteen of 17 patients ultimately underwent external drainage, 3 of these following failure of internal drainage, with the development of only 3 pancreatic &t&s. All fistulas ultimately closed with adjuvant somatostatin therapy. Third, our results suggest that the form of external drainage is important. Two patients developed pancreatic abscess. In both, closed suction rather than sump drainage was employed. Fourth, as suggested by Johnson et al, adequate dependent drainage of these large cysts is crucial.’ In common with their findings, 5 of 7 patients having cystogastrostomy developed complications in the form of retroperitoneal ahscess and upper gastrointestinal hemorrhage presumably from pooling of gastric contents in an incompletely drained retroperitoneum. Those with retroperitoneal abscess responded quickly to conversion to external drainage. Therefore, better dependent Internal drainage 1s indicated in all large pseudocysts in the form of Roux-en-Y cystojejunostomy or, failing that, external drainage should be utilized. We had no experience with either percutaneous or endoscopic drainage of giant PP. Given the high failure rate SEPTEMBER
1996
1 PANCREATIC
of operative internal cystogastrostomy, endoscopic drainage by cystogastrostomy cannot be recommended despite success in the few cases reported to date.14 Percutaneous CTguided drainage has been utilized with impressive degrees of the data is of success.‘5.‘h However, again, interpretation not possible due to the heterogeneity of the patient population. Adequate drainage of giant PI? through these small, nonsuction catheters would appear to be tenuous and could promote catheter related sepsis if left in place for prolonged periods.‘7m’” Nonetheless, their role in the early management of giant PP to allow improvement in the patient’s medical status prior to definitive operative therapy needs to be defined. SUMMARY Giant PP in the face of severe acute pancreatitis have a different natural history from other pseudocysts with a high propensity for the development of complications. Expectant management of these cysts is associated with a high degree of failure due to the development of pseudocyst complications and the need for emergency operation. Earlier operative drainage of these cysts by active external sump drainage may be indicated. Internal drainage of these giant PP by cystogastrostomy should be abandoned in favor of better dependent drainage in the form of Roux-en-Y cystojejunostomy or external darainage. Finally, future comparisons of pseudocyst management must take into consideration underlying pancreatic disease, the severity of pancreatitis at presentation, and pancreatic pseudocyst size. The general attitude of conservatism toward pancreatic pseudocysts in the patient convalescing from a bout of pancreatitis is justified. Behrman and Ellison point out (I believe correctly) that the giant variant of pseudocyst pursues a different pathway entirely and should be approached in an altogether different manner. REFERENCES 1. Warren WD. Marsh WH. Muller WH. Exnerimental nroduction of pseudocysts of the pancreas with preliminary observations on mternal drainage. Surg Gynecol Ubstet. 1957;1@5:385-392. 2. Mullins RJ, Malangoni MD, Bergamini TM, et al. Controversies in the management of pancreatic pseudocysts. Am J Surg. 1988;155:165-172. 3. Warshaw AL, Rattner DW. Timing af surgical drainage for pancreatic pseudocyst. Ann Surf. 1985;202:720-724.
THE
AMERICAN
PSEUDOCYSTWBEHRMAN
ET AL 1
4. Bradley III E, Clements J. Gonzale: A. The natural history of pancreatic pseudocysts: a unified concept of management. Am J surg. 1979;137:135-141. 5. Yeo CJ, Bastidas JA, Lynch-Nyhan A, er al. The natural history of pancreatic pseudocysts documented hy clmputed tomography. slug Gynecol Obstet. 1990;170:41 l-417. 6. Vitas GJ, Sarr MC;. Selected management of pancreatic pseudocysts: Operative versus expecrant management. Surgery. 1992;111:124-130. 7. Johnson LB, Rattner DW, Warshaw AL. The effect of size of giant pancreatic pseudocysts on the outcome of internal drainage procedures. SurgGynecolObstet. 1991;173:171-174. 8. Williams KJ, F&an TC. Pancreatic pseudocyst: recommendations for operative and nonoperative management. Am Surg. 1992;58:199-205. 9. O’Connor M, Dolars J, Ansel H, et al. Preoperative endoscopic retrograde cholan~iopancreatography m the surgical management of pancreatic pseudocysts. AmJ Surg. 1986;151:18-24. 10. Nealon WH, Townsend CM, Thompson JC. Preoperative endoscoplc retrograde cholangiopancreatograph\i (ERCP) in patients with pancreatic pseudocyst associated ti;ith resolving acute and chronic pancreatitis. Ann Surg. 1989;209:532-540. 11. Ranson JHC, Rifkind KM, Turner JW. Ohlective early identification of severe acute pancreatltis. Am J Gustroenterol. 1974;61:443-448. 12. Nguyen BT, Thompson JS, Edney JA, et al. Influence of the etiology of pancreatitis on the natural hlstory of pancreatic pseudocysts. Am J Surg. 1991;162:527-531. 13. Imrie CW, Buist LJ, Shearer MG. Importance of cause in the outcome of pancreatic pseudocysts. Am J Surg. 1988;156:159162. 14. Howell DA, Holhrook RF, B osco JJ. et al. Endoscopic needle localization of pancreatic pseudocysts before transmural drainage. Gastroint Endoscoer. 1993;5:693-698. 15. VanSonnenherg E, Wittich GR, Casola G, et al. Percuraneous drainage of infected and noninfected pancreatic pseudocysts: experience in 101 cases. Radioloa. 1989;170: IOO7~1009. 16. Gerzof SC, Johnson WC, Robhins AH, et al. Percutaneous drainage of infected pancreatic pseudocysts. Arch Surg. 1984;119:888-893. 17. Adams DB, Anderson MC. Percutaneous catheter drainage compared with internal dramage in the mmagement of pancreatic pseudocyst. Ann Surg. 1992;?15:571-578. 18. Criado E, LX+efano AA, Weiner TM, et al. Long term results of percutaneous catheter drainage of pancreatic pseudocysts. Surg Gynecol Obstet. 1992;175:293-298. 19. Rao R, Fedorak 1, Prmz RA. Effect of failed computed tomography-guided and endoscopic drainage on pancreatic pseudocyst management. Surgery. 1993;114:843-849.
JOURNAL
OF SURGERY”
VOLUME
172
SEPTEMBER
1996
231
Pseudocysts
Acute Stephen
W. Behrman,
Pancreatitis
MD, W. Scott Melvin, MD, E. Christopher
BACKGROUND: Pancreatic pseudocysts (PP) following acute pancreatitis have traditionally been approached by observation to allow cyst maturation. However, recent evidence suggests a selective approach toward management is indicated. METHODS: We retrospectively reviewed the presentation, operative management, and outcome of patients developing PP following acute pancreatitis since 1988. PP related to chronic pancreatitis were excluded. RESULTS: Twenty-seven patients were identified, 17 with giant PP (>lO cm) and 10 with PP less than 10 cm. Groups were not different with respect to age and etiology of pancreatitis, predominantly biliary. Patients with giant PP had a significantly greater number of Ranson criteria at presentation and underwent drainage procedures earlier following their initial presentation. With respect to giant PP, 7 patients underwent internal drainage all via cystogastrostomy with 5 complications. Nine of 10 patients underwent urgent operation via external drainage. There were 3 pancreatic fistulas in this group. The morbidity and mortality rates for giant PP were 65% and 18% respectively. With respect to smaller PP, 8 underwent internal drainage with 1 death (mortality rate 10%). External drainage was performed in 2 patients with 1 pancreatic fistula (morbidity 10%). CONCLUSIONS: Patients with PP and a high Ranson score following acute pancreatitis are at significant risk for giant PP formation. Expectant management of giant PP is associated with higher morbidity and mortality than small PP suggesting that earlier external drainage, before clinical deterioration, may be beneficial. To be accurate, comparisons of outcomes for various treatment modalities must take into consideration PP size. Am J Surg. 1996; 172:228-231.
T
he management of pancreatic pseudocysts developing following acute pancreatitis has traditionally involved a period of expectant, nonoperative management. This allows time for cyst wall maturation in order that internal drainage may he accomplished, which, in genFrom the Department of Surgery, Grant Medical Center and The Ohio State University, Columbus, Ohio. Requests for reprints should be addressed to Stephen W. Behrman, MD, Assistant Clinical Professor of Surgery, UCSFNalley Medical Center, 445 S. Cedar Avenue, Fresno, California 93702. Manuscript submitted March 8, 1995 and accepted in revised form October 3, 1995.
228
0 1996 by Excerpta All rights reserved.
Medica,
Inc.
following Ellison, MD, Columbus, Ohio
eral, is associated with lower morbidity than external drainage.’ In addition, a certain proportion of pseudocysts may be expected to resolve spontaneously during this interval of time. In contrast, it is felt that cysts associated with chronic pancreatitis have a more indolent natural history and may be internally drained immediately if indicated.‘s3 In the past it was recommended that all pseudocysts failing resolution by 4-6 weeks undergo operative drainage in order to prevent complications of infection, bleeding, and rupture.4 More recently, this concept has been challenged and a nonoperative approach has been recommended in select patients with evidence of resolution occuring with longer periods of follow-up.5,” Unfortunately, interpretation of the literature has been clouded by series that include patients with both acute and chronic pancreatitis with uniform management being recommended without regard to degree of pancreatic disease. In addition, the natural history of pseudocysts has not been well correlated with either the severity of pancreatitis at presentation or size of cyst. Our recent experience suggests that both the severity of pancreatitis at presentation, and the size of subsequent cyst development, have a significant impact on the natural history and subsequent management of pancreatic pseudocysts.
PATIENTS
AND
METHODS
The hospital records of all patients admitted to Grant Medical Center and the Ohio State University Hospitals with a CPT coding of pancreatic pseudocyst, or patients having surgical procedures for such, from January 1988 through September of 1993 were examined. Via intraoperative records and, when performed, dynamic pancreatography, we were careful to exclude patients with underlying pancreatic necrosis. Fifty-seven patients were identified. Patients with chronic pancreatitis were excluded. All pseudocysts were followed by computed tomography. Records were examined with respect to etiology of pancreatitis, Ranson criteria, pseudocyst size, pre-, post-, and total hospital days, preoperative laboratory values, reason for operation, type of operative procedure, and postoperative complications. Definitions Chronic pancreatitis included those patients with evidence of exocrine and endocrine insufficiency together with any of the following: pancreatic calcifications, ERCP findings of irregular caliber of the duct of Wirsung, and a history of chronic pain. In addition, pseudocysts occuring in patients with no identifiable antecedent episode of acute pancreatitis were classified as chronic pseudocysts. Giant pancreatic pseudocyst was defined as those equal to or greater than 10 cm in diameter by CAT scan analysis as described in prior work by Johnson et aL7 0002-961 O/96/$1 PII 50002-961 O(96)OOi
5.00 57-2
J
1PANCREATIC
PSEUDOCYSTWBEHRMAN
ET AL
TABLE Etiology
of Pancreatitis Pseudocyst
Alcohol Biliary Idiopathic Pancreas divisum Hypertriglyceridemia Drug related
size
Figure 1. Correlation pseudocyst size.
urgent operation
n=17
2. Operative
management
r10
cm
2 5 3 -
3 7 3 2 1
(n = 10)
:n = 17)
of pseudocyst
between
cm
Size
Figure cysts.
(cm)
number
of Ranson
criteria
and
elective operation
J Figure cysts.
of giant
pancreatic
pseudo-
Data presented are the mean plus or minus the standard deviation of the mean. Comparison between groups was undertaken utilizing an unpaired Student’s t-test and, where appropriate, analysis of variance. Significance was assessed at the 95% confidence interval.
RESULTS Fifty-seven patients developing pseudocysts were identified. Pseudocysts secondary to chronic pancreatitis developed in 30 patients and were eliminated from this analysis. Twenty-seven patients developed pseudocysts following acute pancreatitis, 17 with giant PP and 10 with smaller PP (average size, 14.9 * 3.0 vs 7.7 2 1.9 cm, respectively). Etiology of pancreatitis was not different between groups with approximately 45% being due to gallstone disease (Table). There was no correlation between the etiology of pancreatitis and the subsequent development of complications during expectant management. There was no difference in THE AMERICAN
3. Operative
management
of smaller
pancreatic
pseudo-
age between those with giant PI’ and those with smaller PI’ (54.1 i 17.9 vs 51.7 -C 23.2 years, respectively). All patients underwent operative management. Those developing giant PI? had a significantly greater number of Ranson criteria (5.3 i 2.4 vs 2.0 z 2.1; P = .002). The number of Ranson criteria directly correlated with increasing pseudocyst size (r’ = .1, P = .OlO; Figure 1) . Those with giant pseudocysts underwent definitive operative intervention earlier than those with smaller cysts (49.2 2 39.0 vs 74.3 -t- 55.4 days, respectively), a trend that approached but did not reach statistical significance. More than one-half of the giant PI’ underwent urgent drainage due to cyst complications, whereas those with smaller cysts, in general, safely underwent expectant management to allow cyst maturation and elective internal drainage. All patients but 1 undergoing urgent operation were managed by external drainage whereas those having elective therapy had internal drainage. There was no difference in preoperative laboratory values, or postoperative length of stay between groups. We specifically analyzed results of operative therapy in patients with giant PP (Figure 2). Nine of 10 patients underwent urgent operation via external drainage: 6 for infected pseudocysts, 2 for rapid enlargement, and 1 for bleeding. We could not identify any risk factors for the development of pseudocyst complications in those with giant cysts. External drainage was performed in all patients undergoing urgent operation and electively in 1 patient for lack of adequate dependent drainage by cystoenterostomy. Drainage was via triple lumen Davol sump drains in all but two cases. There were 3 pancreatic fistulas in this group, all closing with conservative measures, which included the use of sandostatin with or without total parenteral nutrition. Two patients developed peripancreatic abscess. In both, closed suction rather than active sump drainage was employed. Three patients died; one due to pulmonary embolism. The course of the other two patients was complicated by ischemic colitis and progressive sepsis following external drainage of an infected and 1 bleeding pseudocyst. Five of 7 patients with giant PI? having internal drainage suffered complications. All patients in this group had undergone cystogastrostomy. Three developed retroperitoneal sepsis that required emergency re-operation and subsequent external drainage, all with good results. There were no complications following re-exploration. Two patients suffered upper gastrointestinal hemorrhage requiring transfusion. Both had negative arteriograms and were managed conservatively. Overall morbidity for giant PP was 65% and mortality 18%. Excluding complications resulting from cystogastrostomy and closed suction drainage, morbidity was 24%. JOURNAL
OF SURGERY@
VOLUME
172
SEPTEMBER
1996
229
PANCREATIC
PSEUDOCYSTWBEHRMAN
ET AL /
With respect to smaller pseudocysts, 8 of 10 patients were successfully managed expectently and underwent elective internal drainage, 6 via cystogastrostomy (Figure 3 ) . There was 1 death following cystogastrostomy. This patient developed retroperitoneal sepsis following drainage of an 8cm pseudocyst and expired despite reexploration and external drainage. Two patients failed expectent management and safely underwent external drainage for infected pseudocysts. There was 1 pancreatic fistula in this group, which closed with conservative management. Overall morbidity for smaller PP was 10% and mortality 10%.
DISCUSSION Pancreatic pseudocysts have traditionally been approached by a uniform method of management. A period of expectant observation to allow for cyst wall maturation is followed by elective internal drainage in 4-6 weeks, patients often requiring hospitalization during this entire time with or without total parenteral nutrition.’ Delay of operative intervention beyond this period has been associated with a poor chance of cyst resolution and a significant risk for the development of pseudocyst complications including infection, bleeding, and rupture.’ More recently, these concepts have been challenged and studies have suggested that many of these patients may he safely managed nonoperatively with a significant proportion either resolving or decreasing in size, and with virtually no morhidity.5,” In addition, evidence suggests that pseudocysts associated with chronic pancreatitis behave, and should be managed differently, than those associated with acute pancreatitis.‘,’ Furthermore, pseudocyst size appears to he an important variable with respect to operative management.’ It is becoming clear that the natural history and operative management requires a selective approach based on underlying pancreatic disease, the severity of pancreatitis, and the size of pseudocyst formation. Unfortunately, interpretation of the literature with respect to management of pancreatic pseudocysts is complicated by lack of characterization of the above. Our study suggests that giant pancreatic pseudocysts (> 10 cm) developing following acute pancreatitis have a different natural history and require new approaches toward management. Pseudocysts developing in those with chronic pancreatitis appear to do so in a more indolent nature.“.’ In general, cell wall maturation has occurred, and immediate internal drainage may be performed. The rate of spontaneous resolution in this group IS low secondary to ductal stricture, and preoperative analysis of pancreatic and biliary ductal anatomy via ERCP is indicated. K~‘c) On the basis of these data, PP arising in those with chronic pancreatitis have a natural history, and require modifications in therapeutic interventions different from those with acute pancreatitis. With regard to PP complicating acute pancreatitis, cyst wall maturation is necessary if internal drainage is to he successful. Further, concern for late complications still exists unless operative intervention supervenes.’ Early operation and external drainage of PP is considered less appealing due to complications of catheter drainage consisting mainly of pancreatic fistula and abscess, both significantly prolonging postoperative management. Recently, longterm nonoperative management of PP has been advocated. Prospective studies by Sarr et al and Yeo et al have dem230
THE
AMERICAN
JOURNAL
OF SURGERYa
VOLUME
172
onstrated impressive rates of spontaneous resolution at periods longer than 6 weeks from presentation.5’6 Morbidity has been minimal, although in one series there were several life-threatening complications necessitating emergent operation.” Nonetheless, this approach is appropriate based on selective management. Both studies fail to group PI? based on whether they resulted from acute or chronic pancreatitis, and the severity of pancreatitis at presentation with respect to management is not addressed. Both studies suggest that resolution of pseudocysts greater than 5-6 cm is poor. Therefore, guidelines for management of PP following acute pancreatitis need to be addressed since clearly this form of therapy is not indicated for all. Our study has identified a group of patients that are at high risk for farlure of nonoperative therapy. In this series, patients developmg giant PP following acute pancreatitis had a more signiticant degree of pancreatitis based on Ranson criteria with a mean of 5. Prior studies have associated this number of Ranson criteria with a mortality rate of 40-50%.” In a&t’ I Ion, a significant proportion of these patients failed a 4-6-week period of expectant management. Furthur, approximately 65% of these giant PP developed complications, the maiority of which were related to attempts at internal drAinage via cystogastrostomy. In contrast to prior studies, we found no correlation between etiology of pancreatitis and the development of pseudocyst complications.“.” In summary, those with giant pseudocysts present with severe pancreatitis followed by the early development of large pseudocysts, which have a high propensity for the development of complications and, therefore, failure of expectant management. Analysis of our operative therapies and outcomes are noteworthy for several points. First, early operation (ie, failure of expectant management) was necessary in over 50% of these patients, suggesting that perhaps earlier intervention before complications occur might yield lower morbidity and mortality particularly since many of these patients are in an older age group. Indeed, 2 of the 3 deaths in this group occured in elderly patients having emergency operations for pseudocyst complications. Second, external drainage of giant PP is a safe and effective therapy with acceptable morbidity predominantly m the form of pancreatic fistula. Thirteen of 17 patients ultimately underwent external drainage, 3 of these following failure of internal drainage, with the development of only 3 pancreatic &t&s. All fistulas ultimately closed with adjuvant somatostatin therapy. Third, our results suggest that the form of external drainage is important. Two patients developed pancreatic abscess. In both, closed suction rather than sump drainage was employed. Fourth, as suggested by Johnson et al, adequate dependent drainage of these large cysts is crucial.’ In common with their findings, 5 of 7 patients having cystogastrostomy developed complications in the form of retroperitoneal ahscess and upper gastrointestinal hemorrhage presumably from pooling of gastric contents in an incompletely drained retroperitoneum. Those with retroperitoneal abscess responded quickly to conversion to external drainage. Therefore, better dependent Internal drainage 1s indicated in all large pseudocysts in the form of Roux-en-Y cystojejunostomy or, failing that, external drainage should be utilized. We had no experience with either percutaneous or endoscopic drainage of giant PP. Given the high failure rate SEPTEMBER
1996
1 PANCREATIC
of operative internal cystogastrostomy, endoscopic drainage by cystogastrostomy cannot be recommended despite success in the few cases reported to date.14 Percutaneous CTguided drainage has been utilized with impressive degrees of the data is of success.‘5.‘h However, again, interpretation not possible due to the heterogeneity of the patient population. Adequate drainage of giant PI? through these small, nonsuction catheters would appear to be tenuous and could promote catheter related sepsis if left in place for prolonged periods.‘7m’” Nonetheless, their role in the early management of giant PP to allow improvement in the patient’s medical status prior to definitive operative therapy needs to be defined. SUMMARY Giant PP in the face of severe acute pancreatitis have a different natural history from other pseudocysts with a high propensity for the development of complications. Expectant management of these cysts is associated with a high degree of failure due to the development of pseudocyst complications and the need for emergency operation. Earlier operative drainage of these cysts by active external sump drainage may be indicated. Internal drainage of these giant PP by cystogastrostomy should be abandoned in favor of better dependent drainage in the form of Roux-en-Y cystojejunostomy or external darainage. Finally, future comparisons of pseudocyst management must take into consideration underlying pancreatic disease, the severity of pancreatitis at presentation, and pancreatic pseudocyst size. The general attitude of conservatism toward pancreatic pseudocysts in the patient convalescing from a bout of pancreatitis is justified. Behrman and Ellison point out (I believe correctly) that the giant variant of pseudocyst pursues a different pathway entirely and should be approached in an altogether different manner. REFERENCES 1. Warren WD. Marsh WH. Muller WH. Exnerimental nroduction of pseudocysts of the pancreas with preliminary observations on mternal drainage. Surg Gynecol Ubstet. 1957;1@5:385-392. 2. Mullins RJ, Malangoni MD, Bergamini TM, et al. Controversies in the management of pancreatic pseudocysts. Am J Surg. 1988;155:165-172. 3. Warshaw AL, Rattner DW. Timing af surgical drainage for pancreatic pseudocyst. Ann Surf. 1985;202:720-724.
THE
AMERICAN
PSEUDOCYSTWBEHRMAN
ET AL 1
4. Bradley III E, Clements J. Gonzale: A. The natural history of pancreatic pseudocysts: a unified concept of management. Am J surg. 1979;137:135-141. 5. Yeo CJ, Bastidas JA, Lynch-Nyhan A, er al. The natural history of pancreatic pseudocysts documented hy clmputed tomography. slug Gynecol Obstet. 1990;170:41 l-417. 6. Vitas GJ, Sarr MC;. Selected management of pancreatic pseudocysts: Operative versus expecrant management. Surgery. 1992;111:124-130. 7. Johnson LB, Rattner DW, Warshaw AL. The effect of size of giant pancreatic pseudocysts on the outcome of internal drainage procedures. SurgGynecolObstet. 1991;173:171-174. 8. Williams KJ, F&an TC. Pancreatic pseudocyst: recommendations for operative and nonoperative management. Am Surg. 1992;58:199-205. 9. O’Connor M, Dolars J, Ansel H, et al. Preoperative endoscopic retrograde cholan~iopancreatography m the surgical management of pancreatic pseudocysts. AmJ Surg. 1986;151:18-24. 10. Nealon WH, Townsend CM, Thompson JC. Preoperative endoscoplc retrograde cholangiopancreatograph\i (ERCP) in patients with pancreatic pseudocyst associated ti;ith resolving acute and chronic pancreatitis. Ann Surg. 1989;209:532-540. 11. Ranson JHC, Rifkind KM, Turner JW. Ohlective early identification of severe acute pancreatltis. Am J Gustroenterol. 1974;61:443-448. 12. Nguyen BT, Thompson JS, Edney JA, et al. Influence of the etiology of pancreatitis on the natural hlstory of pancreatic pseudocysts. Am J Surg. 1991;162:527-531. 13. Imrie CW, Buist LJ, Shearer MG. Importance of cause in the outcome of pancreatic pseudocysts. Am J Surg. 1988;156:159162. 14. Howell DA, Holhrook RF, B osco JJ. et al. Endoscopic needle localization of pancreatic pseudocysts before transmural drainage. Gastroint Endoscoer. 1993;5:693-698. 15. VanSonnenherg E, Wittich GR, Casola G, et al. Percuraneous drainage of infected and noninfected pancreatic pseudocysts: experience in 101 cases. Radioloa. 1989;170: IOO7~1009. 16. Gerzof SC, Johnson WC, Robhins AH, et al. Percutaneous drainage of infected pancreatic pseudocysts. Arch Surg. 1984;119:888-893. 17. Adams DB, Anderson MC. Percutaneous catheter drainage compared with internal dramage in the mmagement of pancreatic pseudocyst. Ann Surg. 1992;?15:571-578. 18. Criado E, LX+efano AA, Weiner TM, et al. Long term results of percutaneous catheter drainage of pancreatic pseudocysts. Surg Gynecol Obstet. 1992;175:293-298. 19. Rao R, Fedorak 1, Prmz RA. Effect of failed computed tomography-guided and endoscopic drainage on pancreatic pseudocyst management. Surgery. 1993;114:843-849.
JOURNAL
OF SURGERY”
VOLUME
172
SEPTEMBER
1996
231