Endoscopic drainage of pancreatic pseudocysts

Endoscopic drainage of pancreatic pseudocysts was attempted in 17 patients over an eight year period. There were nine cysts located in the head of the pancreas, six in the body and two in the tail. Endoscopic retrograde cholangiopancreatography was performed in all cases and the pancreatic duct satisfactorily opacified in 16 of the 17 patients. This study identified a communication with the pancreatic duct in seven cases. There were two cases in which multiple cysts were present; in each, one cyst was drained endoscopically and the others surgically. Endoscopic drainage of the cyst was immediately possible in 16 of 17 cases (94%). Late follow-up (mean 26 months) documented cyst disappearance in 11 cases (69%). None of the five patients with persistent cysts has required secondary surgical intervention, and the cysts are asymptomatic and stable or decreasing in size by serial scanning. There was one case ( 6%) in which a pseudocyst recurred following initial resolution. There were two complications (12%) requiring surgical intervention: gastrointestinal perforation with peritonitis in one patient and hemorrhage at the cyst margin from an arterial bleeder in another. There were no deaths at 30 days, but in one case a recurrent acute necrotizing pancrearitis occurred 36 days following endoscopic drainage and the patient died. This death was felt to be unrelated to the endoscopic procedure. In conclusion, internal drainage of pancreatic pseudocysts by endoscopic means can be proposed as an alternative to surgical drainage when the cyst can be identified as bulging into the stomach or duodenum. Immediate drainage is usually effective with a minimal long term recurrence rate. Can J Gastroenterol 1990;4(9):568-571

T H E TREATMENT OF PANCREATIC pseudocysrs remains somewhat controversial with respect to necessity, riming a nd method of in tervention {1,2). The advent of endoscopic retrograde cholangiopancreatography (ERCP), abdominal ultrasound and computed tomography scanning has al, lowed a more comprehensive look into the natural history of pscudocy:,ts both before and after treatment (3,4). It is clear that given the relatively benign course of this disease, a nonsurgical O p· tion would he very appeali ng, if succe~ful. Endoscopic pancreatic cystoenteros tom y has bee n reported in the t reatment of pseudocysrs compl icating acute or ch ron ic pancrearitis and repre, sents a treatment option in selected anatomic configurations (5)(6)(7)(8). The curre nt report details the method, results and potential complications of this technique in the management of the pancreatic p:,eudocyst. patients) pancreatitis. C linica l presenting symptoms are delineated in Table L. The indications for pseudocyst drainage were clinical symptoms or increasing volume in 13 patients, biliary compression in two, spontaneous fistulization in one and recurrence after surgical treatment in vne. Mean fo llow-up after cyst drainage was 26 months (range three to 80). All patients underwent e ither abdomina I ul trasou nd or computed tomography scanning to docu me nt resolution of the cyst. Repeat ERCP was nor routinely performed.

PATIENTS AND METHODS
The cysts cou ld be identified ultrasonographica ll y in 16, by abdominal computed tomography scan in 13, and in all cases hy endoscopically identified internal indentation of the adjacent digestive tract. The mean size of the cysts was 69±44 mm ( range 30 to 200). In two cases there were multiple cysts, and only one cyst in each patient was drained endoscopically; surgical drainage was accomp lish ed fo r the remaining cysts in t hese patients. Of the cysts drained endoscopicall y, there were nine Located in the head of the pancreas, six in the body and two in the tail. The cepha lic cysts resulted in a compression of the second part of the duodenum; cysts of the body and tail resulted in gastric compression. ERCP was performed in all cases to define the ductal anatomy and to determine the presence or absence of a major ductal communication to the cyst . The pancreatic duct was satisfactorily opacified in 16 of 17 patients.
Cystoenterostomy was performed by making a short incision of approximately 5 mm with the diathermy knife, which was en larged to a size of l Oto l S mm after assessment of cyst fluid and opacification of the cyst cavity. The point of maximum compression on the enteric lumen was used as a point for drainage. Hemorrhage, when present, was controlled using the diathermy probe. Nasopancreatic cyst drains were placed when possible to allow irrigation of the cyst cavity and assure adequate continued drain age fo llowi ng local tissue trauma.

RESULTS
ERCP identified a communication with pancreatic ducts in seven cases. In two cases of chronic pancreatitis, a pancreas divisum was discovered; one had an enlarged ventral and on e an en larged dorsal pancreatic duct. Panc reatic ductal dilation was seen in eigh t cases and stenosis in three. The pancreatic duct was normal in three cases ( Table 2).
Drainage of the cysts was immediately possible in 16    the five patients with persistent cysts by echo or computed tomography, the pseudocyst was clinically asymptomatic and appeared stable or reducing in size on serial scans. None of t hese has required secondary surgical intervention to date. There was one ca&e (6%) in which a pseudocyst recurred fo llowing initial resolution. This recurrent cyst was in proximity co the original cyst and it was not possible to tell whether it represented treatment failu re o r progression of disease with subsequent cyst formation. There were two significant complications (12%) requiring su rgical intervention. An early complication of G I perforation with peritonitis occurred in one patient and required emergency laparotomy the day fo llowing endoscopic drainage. Hemorrhage requiring transfusion was seen in one patient, and this patient required emergency surgery 1 S days after the endoscopic procedure. At surgery, an arterial bleeder was identified at the cyst margin and cyscojejunoscomy was performed after I igation of the bleeder. There were no deaths at 30 days. In one case a recurrent acute necrotizing pancreatitis occurred 36 J ays after endoscopic dra inage and the patient died. In reviewing the cou rse of the patient, it was felt t hat the death was not related to endoscopic drainage, as the patient had done well following the procedure. Outcomes are shown in Table 3.

DISCUSSION
The pancreatic pseudocyst accompanying acute pancreatitis may be transient; spontaneous resolution has heen documented 111 up to 90% of cases (2,3,9). Their drainage is more difficult hoth surgically and nonsurgically because of the nature of the cyst wall and the heterogeneous cyst contents. The early cyst wall is thick but phlegmonous in nature and frequently does not adhere well to adjacent gastrointestinal structures. Thus, surgical drainage may not be possible at this stage, since a suture will not hold to the cyst wall, and may be mo1e difficult with higher risk of anastomotic leakage since the plane between the pseudocyst and the stomach or duodenum may not be completely obliterated as with the more chronic cyst ( 10). Thus, endoscopic cystoenterostomy may be more useful in these earlier cases, if clinically mdicated, since surgery is otherwise only infrequently required following acute pancreatitis. Due co the high spontaneous resolution rate, cyst drainage in this group by any method should be limited to cysts persisting beyond severa l months with documented maturation by serial scanning, rapidly enlarging and clinically symptomatic cysts, or cysts with progressive gastrointestinal or biliary compression.
In cases of chronic pancreatitis, the spontaneous pseudocyst remission rate is not as high ( 11) and the long term complications of cyst hemorrhage, perforation or infection are more frequent (33.5%). Thus, elective cyst drainage is less controversial in this group and is often simpler both surgically and nonsurgically due to the cyst wall and con-570 tent characteristics (12). The walls are more clearly defined and adherent to adjacent gastrointestinal structures and the cyst contents more fluid, facilitating adequate drainage. Precise indications for cyst drainage in this group are ill defined because they are frequently asymptomatic or have symptoms inseparable from the primary pancreatic pathology. C learly, progressive enlargement or suspicion of supennfecrion would mandate drainage. Frequent indications for drainage in this group include gastrointestinal or biliary compression, chronicity, risk of future complications of infection, perforation or internal hemorrhage ( 13).
The surgical diversion nf pancreatic cysts has associated severe complications such as infection, bleeding or perforation m 10 to 25% with mortality rates from 5 to 12%. Thus, surgical intervention, while highly successful, does incur significant morbidity in this chronically ill patient population. Recurrence rates following surgery are around 5% (2,10,14). The complication, mortality and recurrence rates in the current study compare favorably with surgical treatment. The long term cyst disappearance rate of 68% may be felt to be low compared to surgical series, but documentation of cyst disappearance fo llowing surgical drainage is not uniform in all series. The true rate and time course of cyst cavity disappearance is often not clear. The bottom line of symptom resolution and stable or decreasing cyst size by serial scanning remains an adequate measure of satisfactory outcome. Three patients ultimately required surgical intervention in the study -two for acute complications of perforation and hemorrhage and one for failure of initial cyst drainage. One might thus quote a nonsurgica l failure rate of 18%. The avoidance of surgery in 82% of the present patients over a mean follow-up period of 26 months was felt to be an important positive outcome.
The first nonsurgical drainage procedures were percutaneous, x-raygu ided, pseudocyst punctures with drain placement (15)(16)(17). Ultrasound or computed tomography scan guidance is now routinely used. One is somewhat limited by the size of drainage catheters that can be placed, and maintenance of catheter patency can be difficult. The viscosity and heterogeneous particulate nature of the cyst fluid in acute pancreatitis make this method of drainage somewhat less <lesirable than endoscopic or surgical drainage. Additionally, although it remains co be conclusively shown, internal drainage is, a priori, preferable to external <lrainage with regard to infection. Immediate success rates of 50 to 100% with needle puncture and drainage are quoted in the literature, but a high recurrence rate of greater than 50% makes external drainage a less than optimal choice if endoscopic techniques are suitable. This technique should be reserved for patients who are higher surgical risks and when endoscopic pseudocyst v1sualizat1on is nm possible.
Laugier et al (18), reporting their experience with the combined results of percutaneous an<l endoscopic drainage of 3 5 pancreatic pseudocysts, note an immediate success rate of 88.5% with no mortality, and a 14.3% complication rate. Cremer et al (19) perforated 13 cystoduodenostomies and five cystogastrostomies withour complications or mortality an<l with complete cyst disappearance in 15 cases (83.3%). There were two recurrences and two incomplete drainages in that series. These figures are comparable to the presem findings m terms of early technical failure and recurrence rates. Thus, overall, experien ce with endoscopic drainage is quite promising and at present the main limitation of this technique is the necessity for visual bulging of the pseudocyst into the stomach or duodenum prior to drainage. The~e early results may improve as imaging techniques continue to advance over the next decade, parttcularl y in the area of endoscopic ultrasound. More accurate pseudocyst localization would increase the number of patients who could benefit from endoscopic treatment.
In conclusion, internal drainage of pancreatic pseuJocysts by endoscopic means has a morbidity and mortality lower than surgical pseudocyst drainage and can be proposed as the treatment of choice when the cyst can he identified as bulging into the stomach or duodenum. Immediate drainage is usually efficient with a minimal long term recurrence rate.