Management of acute pancreatitis: from surgery to interventional intensive care

Percutaneous drainage

Interventional techniques have become increasingly important in recent years due to the now ubiquitous availability of CT scanning and ultrasonography. In 1998, Freeny et al reported for the first time a series of patients with infected acute necrotising pancreatitis who were exclusively drained by CT guided percutaneous catheter drainage.⁵³ Earlier reports covered other infectious complications of acute pancreatitis,⁵⁴ including infected pancreatic fluid collections, pseudocysts, or abscesses, as classified by the International Symposium on Acute Pancreatitis in Atlanta.⁷

Freeny and colleagues⁵³ developed a technique of percutaneous drainage which not only drained infected necrosis passively but included necrosectomy by adding aggressive irrigation through large bore percutaneous catheters (28 F). Thirty four patients with necrotising pancreatitis and uncontrolled sepsis were treated. An average of three separate catheter sites per patient and four catheter exchanges per patient were necessary for the removal of necrotic material. Pancreatic surgery was avoided in 16 patients (47%), and sepsis was controlled in 25 patients (74%). Although nine of the latter group needed elective surgery, the surgical procedure could be avoided successfully in critically ill patients until stabilisation. Percutaneous drainage was ineffective in nine patients who needed surgery to control sepsis or bleeding (26%). The overall mortality was 12%. These four patients all were critically ill with multiorgan failure, bleeding, or shock. The recipe of success in this series was the commitment of the interventional radiologists (with often daily catheter interventions: 146 catheter exchanges, long duration of drainage of 25–152 days, no complications) and the improvement of the technique which enabled percutaneous necrosectomy. However, patients with central gland necrosis, who often present with disruption of the midsection of the main pancreatic duct resulting in a fistula, responded poorly to percutaneous drainage in the series from Seattle (cure in 4/14 (28%) and control of sepsis in 50%).

The radiological approach was taken to its limits by Gmeinwieser and colleagues.⁵⁵ They combined percutaneous retroperitoneal necrosectomy, fragmentation of necrotic pancreatic and peripancreatic tissue with a snare catheter and Dormia baskets, continuous lavage of the cavity, and repeated bronchoscopic visualisation of the cavity with percutaneous blockade of a pancreatic duct disruption to successfully treat and avoid surgery in a young man with infected necrosis who strictly declined the surgical procedure recommended.

The same group created a technique to treat patients suffering from infected necrosis, primarily percutaneously, when signs of sepsis were present and intensive care treatment was unsuccessful. Large bore 20–28 F catheters were placed under CT and fluoroscopy guidance, routinely with a retroperitoneal approach from the left side (fig 3 ▶). Under fluoroscopy control, necrotic and solid material was fragmented and removed actively with aspiration, snares, and forceps, using peel away sheets (fig 4 ▶). Continuous lavage was performed using up to 12 litres of NaCl solution per day and additionally manual aspiration of solid material was performed on a daily basis.⁵⁶ Twenty five of 29 patients treated by this method were successfully cured of sepsis and only three needed further elective surgery at a later time.⁵⁷ Moreover, the intervention related complication rate was very low (table 4 ▶). Nine patients treated by this technique of percutaneous necrosectomy were re-evaluated after a median follow up period of 30 months with respect to quality of life, morphology, as well as endocrine and exocrine pancreatic function.⁵⁸ All patients were in good general condition with respect to quality of life. Only 2/9 (22%) patients had moderate to marked changes in morphology, as observed on CT. There was mild to moderate exocrine dysfunction in 5/8 (63%) patients, and severe restriction of exocrine pancreatic function in 2/8 (25%). The oral glucose tolerance test was normal in 3/6 (50%) patients tested. One patient with diabetes in the oral glucose tolerance test had pre-existing type II diabetes requiring insulin therapy since the onset of acute pancreatitis. In 3/9 (33%) patients an oral glucose tolerance test was not performed due to known pre-existing diabetes.

Another series of catheter directed debridement of infected necrosis with a slightly different technique was published by Echenique and colleagues.⁵⁹ The catheters used were smaller (14–16 F) but stone retrieval baskets and floppy tipped guidewires were used to break debris into smaller peaces. All 20 patients with infected necrosis were treated successfully with regard to clinical course and lesion appearance. Surgery was avoided, and no death occurred. Seven to 32 (average 17) episodes of debridement were necessary which were started on an inpatient basis and continued on an outpatient basis. Fistulas developed in 10 (50%; three to the bowel, seven to the pancreatic duct). However, all patients included in this trial were haemodynamically stable despite the presence of infected necrosis. Several other series have been published since, reporting different success rates (table 4 ▶).^60–62 Again, it is almost impossible to compare these results with available data on surgical results secondary to the inhomogenity of the patient selection in the different series. Considering the publication bias for so called “negative studies”, it has to be concluded that percutaneous drainage of infected pancreatic necrosis is an option that works in the hands of some very dedicated specialists but is probably not a technique which can be recommended for routine use everywhere. Thus “can do” must be separated from “should do”. It may have a role as a temporary measure to bridge the critical early time after the onset of acute pancreatitis to a later optimal time point for definite intervention.^61,63,64

Endoscopic therapy

Successful endoscopic drainage of symptomatic sterile or infected pancreatic necrosis was reported by Baron et al as early as 1996.⁶⁵ The technique applied was originally described for uncomplicated pseudocysts. Several transgastric or transduodenal drainage catheters (10 F) and a nasopancreatic irrigation tube were endoscopically inserted into the retroperitoneum to perform necrosectomy. Lavage was continued until resolution of the collection. In this first report, 2–4 procedures were required for resolution and the mean duration of catheter placement was 19 days. Successful removal of necrosis was achieved in over 80% with no mortality. However, the majority of patients treated had no infected necrosis but residual fluid collections with debris. Almost 40% were iatrogenically infected secondarily by endoscopy. Moreover, there were serious complications in 45% of patients, including serious bleeding and gastric perforation (table 4 ▶). Additionally, it is worthy of note that up to 60% of those patients successfully drained developed further collections in the subsequent two years.⁶⁶ Thus this series confirmed that in the presence of necrosis, drainage must be combined with some form of surgical removal of necrotic material. In 1999, Baron and Morgan described successful placement of percutaneous endoscopic jejunostomy tubes through a PEG tube and subsequently through a transgastric track into the necrotic pancreatic collections for irrigations in two cases.⁶⁷ The theoretical advantages of this technique are that on the one hand it avoids the need for uncomfortable nasopancreatic catheters and on the other does not produce the side effects observed after percutaneous drainage, including skin irritation and external pancreatic fistulas.⁶⁷

Although endoscopic drainage might be applicable in some patients with necrotising pancreatitis, only a few centres have used this technique. The results in infected necrosis are only anecdotal, experience with this method is limited, and no interdisciplinary comparative data exist.

Minimally invasive procedures

Advances in laparoscopic technology and instrumentation allow the utilisation of minimally invasive techniques for management in pancreatic disease, and theoretically lessen the surgical stress in the already compromised patient.

As early as 1996, Gagner described laparoscopic debridement and necrosectomy for the treatment of necrotising pancreatitis with three different minimally invasive approaches: (1) transgastric drainage, (2) retrogastric retrocolic debridement, and (3) a full retroperitoneoscopic technique.⁶⁸ Since then, many different techniques have been applied by several groups in the search for the easiest access to the necrotic masses in the retroperitoneum.

The laparoscopic assisted transgastric approach is similar to the endoscopic approach and somehow identical to the approach using a PEG to access the stomach.⁶⁷ Several case reports exist which describe successful laparoscopic transgastric pancreatic necrosectomy for infected necrosis, suggesting effective debridement and internal drainage in selected patients with this minimally invasive approach.⁶⁹ However, no larger series have been reported.

Others have approached the infected necrosis with standard laparoscopy, combining necrosectomy with splenectomy and cholecystectomy.⁷⁰ Zhu et al published their experience of the laparoscopic approach in 10 patients. Despite the fact that they included patients with acute haemorrhagic and necrotising pancreatitis without infection which do not need surgical intervention at all, their mortality was 10%.⁷¹ As almost every procedure can be performed laparoscopically, this approach has not been evaluated in any larger study or in prospective randomised trials. A theoretical risk is the spread of infection into the abdominal cavity, further intraoperative difficulties in case of reoperations, and enhanced risk of erosions of the intestinal tract.

In the late 1980s, the open retroperitoneal approach with lumbotomy was used to avoid these drawbacks.^72–74 This open approach allowed exploration of the pancreas and manual removal of infected necrosis. Although the mortality rates for this retroperitoneal approach compared favourably with historical controls in the same hospitals, the technique has not gained popularity because of its high complication rates. In three series, mortality varied between 20% and 33%, and local complications, mainly enteric fistulas and bleeding, were observed in 20–50%. Fagniez et al reported major morbidity, including enteric fistula (45%), haemorrhage (40%), and colonic necrosis (15%). Thus the main drawback of this open retroperitoneal approach is its narrowness and blindness about the risk of damage to visceral organs and vessels. However, the improvement in CT in recent years with its exact information about the location of necrotic tissue eases the direct retroperitoneal approach. In fact, Nakasaki et al reported successful treatment of infected necrosis by retroperitoneal laparotomy even in patients who needed reoperations after laparotomy (table 5 ▶).⁷⁵

In recent years, minimally invasive techniques using the theoretical advantages of retroperitoneal access have been developed.^68,76 Most groups which apply minimally invasive surgery for infected necrosis today use this retroperitoneal approach.^76–81 Despite small variations in the different techniques applied, they have in common that the infected necrosis of the retroperitoneum is accessed under endoscopic visualisation with subsequent necrosectomy and lavage. The techniques involve either intraoperative dilatation of a percutaneous drain tract which was applied by ultrasound or CT guidance preoperatively, or a direct approach of the infection with a retroperitoneoscope. Depending on the localisation of the infectious tissue, access can be gained from the left or right flank, and over one or more routes. While at first the technique was used in patients after laparotomy as a reoperation, it is now even advocated as a primary approach in some patients.^77,79,80

Several reports show that the technique is applicable and that it can successfully treat some patients with infected necrosis. The results with regard to mortality and morbidity of the larger series published to date are presented in table 6 ▶. Morbidity ranges between 30% and 60%, the success rate for complete necrosectomy between 60% and 100%, and the mortality in these series between 0% and 27%. Although this seems to support the view that laparoscopic assisted necrosectomy is a safe alternative to open necrosectomy, the data have to be interpreted with caution. At first it is obvious that a careful case selection was performed in all trials and it is evident that the patients cannot be compared with those treated in series with open surgery. In the largest series published so far,⁸⁰ only 24 of 45 patients were treated by minimally invasive retroperitoneal pancreatic necrosectomy. In the others, open surgery was performed for the following reasons: lack of a safe retroperitoneal guidewire access route, presence of an additional intra-abdominal complication, extensive paracolic necrotic extension, or multifocal necrotic foci. Moreover, three of the 24 attempted minimally invasive procedures could not be undertaken for technical reasons and another five required open surgery secondary to sepsis or bleeding after the initial minimally invasive approach. Thus only 35% (16 of 45) of all patients treated with necrosectomy were the right candidates for the retroperitoneoscopic technique. Although the proportion of patients undergoing minimally invasive necrosectomy increases over time with experience in the centres,^77,80 many patients will not be treated successfully with this method. An additional disadvantage is the need for repeated procedures and subsequently an increase in the length of hospital stay with the minimally invasive procedures.^76,80 This contrasts with the results of the open techniques which imply a postoperative method to continuously remove residual pancreatic necrosis to achieve complete removal of the infected necrosis with a single operation in the majority of cases. During laparoscopic assisted necrosectomy there is significant potential for major injury to intra-abdominal organs or vascular structures. Indeed, all reports show quite a high incidence of serious complications, including fistula (20%–60%) and bleeding (15%), despite the preselection of patients.

Proponents of minimally invasive technologies in this clinical setting cite a desire to minimise the physiological insult in patients who are already critically ill. Although some investigators^77,80 believe that a reduction in the deterioration in organ function and the need for postoperative intensive care management can be observed in patients treated with the minimally invasive techniques, no data exist to clearly demonstrate that minimally invasive procedures are less prone to morbidity than open surgery.

In conclusion, techniques for laparoscopic necrosectomy are still evolving. Safe retroperitoneal access and necrosectomy is possible in patients with infected necrosis but not all patients, depending on size and localisation of the infectious foci, can be treated successfully. No randomised studies exist comparing one management technique with another. All studies reported involved small numbers of patients, were analysed retrospectively, and comprised selected patients with great variation in comorbidities and disease severity. Consequently, it is not possible to draw any conclusions from the literature with regard to the best interventional or surgical approach for infected necrosis. However, in the absence of well designed clinical trials, we must be cautious in the application of new technologies. Technical feasibility does not obviate the need for scientific rigor and sound clinical judgement. Thus today, outside clinical trials, minimally invasive surgery should be limited to specific indications and to those patients who are critically ill and otherwise unfit for conventional surgery.

“Minimally invasive procedures, including percutaneous drainage, endoscopic drainage, or minimally invasive surgery (that is, retroperitoneoscopy) for infected pancreatic necrosis may play a role as a temporary measure to bridge the critical early time after onset of acute pancreatitis to a later optimal time point for definite intervention. Otherwise, they should be limited to specific indications in patients who are critically ill and unfit for conventional surgery or clinical trials”