Laparoscopic distal pancreatectomy: the Brisbane experience of forty-six cases

Abstract

Background and aims. Laparoscopic distal pancreatectomy (LDP) is a safe alternative to conventional open distal pancreatectomy, with advantages that include smaller incisions, less pain, and shorter postoperative recovery. Despite these apparent advantages, however, uptake of the procedure has been slow, with only a handful of series published. Material and methods. All LDPs performed in Brisbane, Australia, over a 10-year period (May 1996 to June 2006) were retrospectively reviewed. Results. Forty-six consecutive LDPs were performed. A variety of lesions were resected, including nine cancers. Twelve patients were converted for oncological (6) or technical reasons (6). The spleen was retained in 14/29 patients, either by main splenic vessel preservation (9) or solely supported by the short gastric vessels (5), resulting in inferior pole infarction in 2 patients. Overall morbidity was 39%, including 15% pancreatic fistula. All fistulas resolved after a median of 6 weeks without re-operation. A non-significant trend toward fewer fistulas with stapled rather than sutured stump closure was observed (13% vs 19%; p=0.43). Median operative duration and hospital stay were 157 min and 7 days, respectively. There was no mortality. Conclusion. LDP is a safe alternative to conventional resection for a wide range of lesions. As with open resection, pancreatic fistula is the dominant morbidity, but is generally indolent. While spleen preservation is often possible, care must be taken to avoid infarction of the inferior pole if the Warshaw technique is utilized.

Introduction

Laparoscopic resection of the pancreas remains uncommon despite the widespread use of minimally invasive surgery for other abdominal pathologies. While the current evidence does not support laparoscopic pancreatico-duodenectomy, distal pancreatectomy may well be suited to the laparoscopic approach, because no anastomosis is required and the size of incisions can be greatly reduced compared to the conventional open approach. The published literature concerning laparoscopic distal pancreatectomy (LDP) indicates the usual advantages of minimal access surgery, including reduced postoperative pain, faster recovery, and fewer wound-related problems.

Some of the main concerns about LDP include questions about oncological adequacy, if the resected lesion proves to be malignant, and a perception of a higher rate of pancreatic fistula. Clearly, large comparative trials, ideally prospective and randomized, help clarify these concerns; however, it is surprising how few cases have been reported in the world literature, the majority being in the form of isolated case reports or small retrospective series with fewer than 10 patients. This article reports on the results of 46 patients and, to our knowledge, one of the largest experiences by a single group of surgeons to date. It is intended that these data contribute to the published experience and begin specifically to address some of the controversies including indications, optimal pancreatic stump management, as well as issues surrounding splenic preservation.

Methods

A detailed retrospective review was conducted of all LDPs performed by a group of seven surgeons at five related institutions in and around Brisbane (Queensland, Australia) between 1996 and 2006.

The technique of LDP used was similar to that first published by both Gagner ¹ and Cuscheri in 1996 ². This involved exposure of the pancreatic neck, body, and tail by division of the gastro-colic omentum and mobilization of the splenic flexure. Using a medial to lateral approach, the pancreatic neck was transected by either harmonic scalpel or endoGIA 45×2.5 mm linear stapler and the distal pancreas freed toward the spleen. A variety of techniques was used to address the pancreatic stump, including staples alone, augmented by fibrin sealant, sutures, or omental patch, or sutured closure if transection was carried out using ultrasonic shears. If splenectomy was performed, the splenic artery and vein were divided level with the pancreatic neck by endoGIA 45×2.5 mm, allowing the spleen to be taken en bloc with the pancreas. In spleen preserving LDP, splenic vessels were either preserved with ligation of pancreatic side branches by clips and/or harmonic scalpel, or resected with the distal pancreas using endoGIA, relying on the short gastric vessels alone to maintain splenic perfusion as described by Warshaw ³. Plastic specimen retrieval bags and a 19Fr closed suction silicon drain were used routinely. The drain, which was positioned near the pancreatic stump, was removed by the third postoperative day if the lipase content was less than 1000 IU.

All patients were given preoperative polyvalent pneumococcal, meningococcal, and Haemophilus influenzae vaccines. Routine anti-thromboembolic measures, including intraoperative pneumatic calf compressors and postoperative subcutaneous heparin, were administered. Somatostatin analogs were not routinely used.

Pancreatic fistula was defined as any persistent lipase-rich drainage beyond the 6th postoperative day, or the radiological insertion of a percutaneous drain into a lipase-rich peri-pancreatic collection.

Results

Forty-six consecutive LDPs were performed. Patients had a mean age of 59.8 years (range 25–83 years) and 54% were female. A variety of pancreatic lesions was resected (Table I).

Table I. Final pathology diagnoses (n=46).

Pathology	No.
Serous cystadenoma	10
Ductal adenocarcinoma	9
Mucinous cystadenoma	6
MEN1	3
Insulinoma	2
Non-functioning adenoma	2
Lymphoma	2
IPMN	2
Mucinous cystadenocarcinoma	1
Metastatic lung cancer	1
PanIN	1
Pseudopapillary tumor	1
Gastrinoma	1
Pseudocyst	1
Inflammatory pseudotumor	1
Hemangioma	1
Dermoid	1
Retention cyst	1

The median duration of surgery was 139 min (mean 157 min, range 90–278), with median estimated blood loss of 200 ml (mean 340 ml, range 30–2000 ml). Median length of hospital stay was 7 days (mean 13.5 days, range 3–100).

Conversion to open surgery was performed in a total of 12 patients (26%) (Table II). The need to convert was spread fairly evenly across the 10-year period and was related to the pathology encountered rather than as a consequence of the learning curve. In six patients, a malignant appearance was discovered when the lessor sac was opened, and all of these patients were converted to ensure adequate margins and lymphadenectomy. In the remaining six patients, conversion was performed to deal more safely with dense adhesions from previous surgery (3), splenic vein bleeding (1), and a very bulky lesion (2).

Table II. Conversion to open distal pancreatectomy; 12 cases (26%).

	Reason	No.
Oncological (50%)	Malignant appearance at initial laparoscopy	2
	Concern regarding malignancy during dissection	3
	Contiguous involvement of adjacent organs	1
Technical (50%)	Adhesions from previous surgery	2
	Massive size of lesion	2
	Bleeding	1
	Extraperitoneal insufflation	1

A total of 13 patients had malignant tumors. This became apparent early in the dissection in the 6 cases discussed above and these were converted. The other 7 cases (5 ductal adenocarcinomas, 1 mucinous cystadenocarcinoma, and 1 lymphoma) were completed laparoscopically, with the malignant nature of the lesion not revealed until histopathological analysis. Clear surgical margins were obtained in all 7 patients.

A variety of techniques addressing the pancreatic stump were used according to individual surgeon preference (Table III). Division of the pancreatic neck by linear stapler was preferred by most surgeons. Fibrin sealant was also frequently applied to the stump in the latter part of the series, after this product became available, reflecting a belief that this might lower the rate of pancreatic fistula. One surgeon also routinely added an omental patch after stapled transection.

Table III. Method of stump closure vs fistula incidence (no differences reached statistical significance).

Method	Additional measures	No.	Fistula
Stapled (65%)	Alone	18	2 (11%)
	Fibrin sealant	4	1 (25%)}	average 13%
	Suture oversew	3	1 (33%)
	Suture oversewn, fibrin sealant and omental patch	5	0
Sutured (35%)	Alone	8	1 (13%)}	average 19%
	Fibrin sealant	8	2 (25%)

The overall incidence of pancreatic fistula was 15% (7 patients), 4 in the stapled group (13%) and 3 in whom the pancreas had been divided by ultrasonic shears then oversewn (19%). This difference did not reach statistical significance (p=0.43). The addition of augmented measures, including the use of fibrin sealant to either a stapled or sutured pancreatic stump, did not appear to make any difference to the fistula rate.

Splenic preservation performed in 14 patients was not considered in 10 patients with lesions very close to the splenic hilum, in 6 patients converted because of malignancy, and in 1 patient with an abnormal appearing spleen. When performed, both splenic and short-gastric vessels were preserved in 9 patients, while in 5 only the short-gastric vessels were preserved (Warshaw technique). Two patients in the Warshaw group subsequently developed painful infarction of the inferior splenic pole (40%) (managed conservatively in both cases).

The total incidence of both major and minor postoperative abdominal complications was 39%, including 7 patients with pancreatic fistula (15%). Early re-operation was performed in 1 patient on day 4 to control postoperative hemorrhage. There were no in-hospital deaths (Table IV).

Table IV. Perioperative complications (39%).

Complication	Management	No.
Pancreatic fistula (15%)	Continued drainage	5
	Percutaneous drainage	2
Bleeding	Post-op transfusion	2
	Laparotomy	1
Intra-abdominal abscess	Percutaneous drainage	3
Wound infection	Antibiotics and wound packing	2
Infarction inferior pole spleen	Analgesia	2
Anastomotic leak from synchronous colonic resection	Percutaneous drainage and antibiotics	1

Two patients (one neuroendocrine tumor, the other IPMN) developed new lesions near the pancreatic stump during follow-up. Fearing recurrent disease, these patients underwent re-operation at 7 and 11 postoperative months. In both cases, the lesions turned out to be benign pseudocysts.

Discussion

Despite the first case of LDP being reported over a decade ago, fewer than 300 cases have been added to the world literature. The largest series to date is the Multicentre European retrospective study, in which 82 LDPs, spread over 25 institutions, were reported ⁴. Only a handful of series deal with more than 10 cases (Table V), serving to highlight that the worldwide experience with the technique still remains in its infancy.

Table V. Laparoscopic distal pancreatectomy series with 10 or more cases.

Author	Year	No.	Converted	Total complications	Fistula	Spleen preservation	LOS (days)
Patterson25	2001	15	13%	26%	13%	75%	6
Fabre27	2002	13	15%	30%	8%	77%	5–22
Park32	2002	25	8%	16%	4%	48%	4
Mabrut7	2005	98	16%	59%	16%	81%	7
Edwin11	2004	17	13%	38%	0%	41%	5.5
Current series		48	26%	39%	17%	57%	7

Comparative analysis of the data that are available, however, does point to several advantages over open distal pancreatectomy. Wound size and complications, blood loss, length of hospital stay, time to return to usual activities, and the rate of spleen preservation rate favor the laparoscopic approach ⁴.

It is generally agreed that the lesions most suited for LDP are benign lesions that are either symptomatic (such as large serous cystadenomas) or carry potential for malignant transformation (particularly mucinous cystadenomas and IPMN). When located in the distal pancreas, neuroendocrine tumors are often benign or low-grade malignancy, and may also be appropriate for LDP ⁵,⁶,⁷. However, for frankly malignant lesions, conventional open distal pancreatectomy with splenectomy and celiac lymphadenectomy remains the preferred approach by our institution. Detecting the true nature of pancreatic lesions preoperatively is reliant on careful work-up that includes fine-slice triple phase computed tomography (CT), serum tumor markers, and selective use of endoscopic ultrasound (EUS). The presence of metastases, nodal involvement, loss of normal tissue planes, or frank local invasion suggests malignancy. Fine-needle aspiration, either percutaneously or during EUS for cytology, mucin, lipase, and tumor markers, may be potentially helpful, but remains controversial despite recent evidence that the risk of tumor seeding may have been overestimated previously ⁸.

Despite thorough preoperative work-up, the malignant nature of a pancreatic lesion may not be revealed until resection is attempted. Characteristics, including infiltration into the Lessor Sac, may be more clearly revealed operatively. Such operative findings prompted conversion to open resection in 6 of 13 malignant lesions resected in this series. However, there are lesions appearing benign on both preoperative and operative assessment that turn out to be malignant, or have a malignant focus, when examined by the pathologist. We encountered this situation in seven patients (15%), similar to the experience of other researchers ⁹,¹⁰. In our view, further management of such patients (including consideration of re-operation) is best determined on a case-by-case basis within a multidisciplinary setting.

The incidence of pancreatic fistula in this series was 15% by our definition of persistent lipase-rich drainage beyond 6 days or radiological drainage of lipase-rich fluid collection. Rates of between 15% and 20% are commonly reported after distal pancreatectomy, regardless of whether performed laparoscopically or open ⁴,⁹,¹⁰,¹¹,¹²,¹³,¹⁴,¹⁵,¹⁶,¹⁷,¹⁸. The risk appears inherently related to distal pancreatomy rather than to the manner of surgical access. Although relatively common, it is rarely serious or life-threatening. In our series, no patients required re-operation or developed serious complications from pancreatic fistula, and all resolved with simple drainage alone.

The risk of pancreatic fistula did not appear to relate clearly to how the pancreatic stump was managed, although we did notice a slight trend in favor of stapled transection. Nearly identical findings were reported in a much larger review by Knaebel et al. which involved 1080 patients ¹⁹. The use of larger staples (4.8 mm open height) has been postulated as a way of further reducing the risk ⁹. Additional measures, such as applying fibrin sealant to the pancreatic stump, did not appear to help either after stapled or sutured closure in our series. Postoperative somatostatin use has no proven benefit ⁴. However, the consistency of the pancreas may play a role, as fewer leaks seem to occur when affected by chronic pancreatitis ²⁰.

En bloc splenectomy is usually performed during conventional open distal pancreatectomy. As the important role of the spleen becomes better appreciated, there is increasing enthusiasm for splenic preservation whenever possible. Much higher rates of splenic preservation, up to 85% in some series, have been reported after LDP ⁴,⁸,⁹,¹¹,¹³,¹⁵,¹⁶,¹⁷,²¹,²², perhaps indicating better preservation of the spleen by the laparoscopic technique. Two methods have been described. The first involves preservation of both the splenic artery and vein by careful dissection of the pancreatic tail away from these vessels with ligation of side branches. The alternative method, first described by Warshaw in 1988, maintains splenic perfusion solely by the short gastric vessels. The main splenic vessels are divided twice: first at the pancreatic neck, and second near the splenic hilum, resulting in their en bloc excision with the pancreas ³. The choice of technique utilized in our series was influenced by several considerations; the ease of separating the pancreas from the adjacent vessels, the degree of confidence in the lesion being benign, and individual surgeon belief about the relative merits of each method. In our experience with Warshaw's technique, two out of five patients developed symptomatic infarction of their inferior splenic poles. Other authors have reported more serious problems, including splenic abscess or infarction of the entire spleen ⁴,¹⁰,¹². Despite the risk of complications, the Warshaw technique may permit preservation of the spleen when otherwise impossible. When now using Warshaw's technique, we take particular care to preserve the lowermost short gastric vessels when opening the gastrocolic ligament, and preserve any extra hilar anastomotic branches when dividing the main vessels near the splenic hilum.

Conclusions

LDP provides patients with the advantages of minimal access surgery and is appropriate for benign and pre-malignant lesions as well as neuroendocrine tumors. Lesions exhibiting a malignant appearance on preoperative imaging or during initial laparoscopy should still be managed by open resection until enough long-term data accumulate. The incidence of pancreatic fistula was 15% in this series and did not differ significantly between stapled or oversewn stump closure or with the application of fibrin sealant. Spleen preservation is frequently possible, but care needs to be taken to preserve blood supply to the inferior pole when the Warshaw technique is used.