Comparison of open with laparoscopic distal pancreatectomy: a single institution's transition over a 7-year period

Abstract

Objectives

Many studies have shown laparoscopic distal pancreatectomy (LDP) to have benefits over open distal pancreatectomy (ODP). This institution made a unique abrupt transition from an exclusively open approach to a preference for the laparoscopic technique. This study aimed to compare outcomes in patients undergoing LDP and ODP, respectively, over the period of transition.

Methods

A retrospective review of all patients undergoing LDP (n = 82) or ODP (n = 90) was performed. Surrogate oncologic markers for the subgroup of patients with malignant disease were also studied.

Results

The ODP and LDP groups were well matched with regard to demographics, comorbidities and tumour characteristics. Significant differences were noted in favour of the LDP group in which decreases were seen in estimated blood loss (<0.001), need for packed red blood cell transfusions (<0.001), length of hospital stay (<0.001) and intensive care unit stay (<0.001). No other significant differences in the occurrence of complications or oncologic outcomes were seen. Rates of Grade B and C fistulae were 10% and 6% in the ODP and LDP groups, respectively. Grade III–V complications occurred in 20% and 13% of the ODP and LDP groups, respectively.

Conclusions

Laparoscopic distal pancreatectomy continues to compare favourably with ODP when well-matched patient series are reviewed. The results show a decreased need for blood transfusions and hospital resources in LDP. Additionally, there may be oncologic advantages associated with LDP compared with ODP in pancreatic malignancies.

Introduction

Recent advances in the field of minimal-access surgery have resulted in the increasing acceptance of the laparoscopic approach to distal pancreas resection over the past decade. Many early publications have shown the feasibility of the procedure^1–5 and described the technique of distal pancreatectomy.^6–8 Smaller single-institution series have shown the laparoscopic approach to offer significant advantages compared with the open approach, but only four of these series reported over 50 patients in the laparoscopic group.^9–12 More recently, comparative meta-analyses comprising larger groups of patients over multiple institutions have demonstrated that the laparoscopic approach is safe and results in significant improvements in hospital stay and morbidity compared with its open counterpart.^13–15 Other series have looked at the oncologic aspects of laparoscopic distal pancreatectomy (LDP) and have found it to have acceptable outcomes compared with open distal pancreatectomy (ODP).¹⁶ Despite this, a recent look at administrative database information on the use of laparoscopy in the resection of lesions in the body or tail of the pancreas implies that LDP is widely underutilized across the USA.¹⁷

The fundamental goal of any change in surgical technique is to improve outcomes. For those undergoing surgery for benign disease, the goal of surgery is to decrease morbidity, recovery time and the occurrence of complications. For those with malignant disease, improved surgery would ideally maintain or enhance oncologic results, including longterm survival. Because of its recent introduction and the short follow-up associated with LDP, no longterm studies are available. However, multiple studies have analysed operative variables in pancreatic and peripancreatic malignancies that serve as markers of improved cancer outcomes, including decreased operative blood loss,¹⁸ decreased lymph node ratio,¹⁹ and negative resection (R0) margin status.²⁰

This study reports a comparison of a single institution's experience in distal pancreatectomy prior to and after the introduction of the laparoscopic approach without the need for a learning curve. Clinical outcomes were analysed to assess any differences between the LDP and ODP groups. Additionally, surrogate markers of oncologic outcomes were compared to assess the non-inferiority of the minimal-access approach and to establish whether LDP might offer any advantages over ODP for patients with pancreatic malignancies.

Materials and methods

Patients and data collection

A review of all patients undergoing distal pancreatic resection at a tertiary academic centre between January 2005 and December 2011 was performed. All patient data were collected from a comprehensive retrospective and prospective pancreatic resection database which included clinical, operative and pathologic information on all patients undergoing pancreatic resection. Distal pancreas resection was defined as any resection of the pancreas parenchyma starting at the neck or distal to the neck with or without splenectomy and included subtotal resections up to the level of the gastroduodenal artery and superior mesenteric vein. All surgical procedures carried out at this high-volume hepatopancreaticobiliary centre over this 7-year period were performed by surgeons with significant experience in pancreatic surgery. Prior to August 2008, 99% (74/75) of all distal pancreatic resections were performed using an open technique. Between August 2008 (from the arrival of the senior author, HJA) and the end of the study time period, only 17% (16/97) of distal pancreatic resections were performed using an open technique.

Preoperative variables examined included patient age, gender, body mass index (BMI), comorbidities, Eastern Cooperative Oncology Group (ECOG) status and American Society of Anesthesiologists (ASA) class. Operative data included estimated intraoperative blood loss (EBL; ml), operative time (min), total packed red blood cells (pRBC; units) transfused during the operation or in the subsequent hospitalization, intensive care unit (ICU; days) stay required, total length of stay (LoS; days) including the days of operation and discharge, and discharge destination.

Definitions

Data on inpatient and 30-day postoperative complications were collected and scored according to the Clavien–Dindo complication scale.²¹ Complications of Grades I and II were considered minor; those of Grades III–V were considered major. Pancreatic fistula,²² delayed gastric emptying²³ and post-pancreatectomy haemorrhage²⁴ were defined according to the International Study Group of Pancreatic Surgery (ISGPS) definitions.

Subgroup analysis was performed for those patients who underwent ODP or LDP for pancreatic adenocarcinoma and pancreatic neuroendocrine tumours (PNETs). This analysis was performed to examine oncologic data including tumour size (maximum dimension, cm), total number of lymph nodes, number of positive lymph nodes, lymph node ratio (in N1 patients only), tumour–node (TN) stage and margin status. Margin analysis was performed in each specimen in a radial fashion and an R1 resection was defined as the microscopic presence of tumour cells at any margin.

Operative technique

The steps and technique used in LDP have been previously published.⁶ Briefly, the patient is positioned in a slight right lateral decubitus position. Initial abdominal access is gained through an abdominal Hasson trocar and three additional trocars are placed (Fig. 1). The steps of the operation proceed in a clockwise fashion: the splenic flexure of the colon is taken down first to obtain complete mobilization of the distal transverse and proximal descending colon from the tail of the pancreas. The gastrocolic ligament is incised, completely exposing the tail of the pancreas. If a splenectomy is to be performed, the gastrosplenic ligament and short gastric vessels are transected up to the superior pole of the spleen. Then, an appropriate transection site on the pancreas is chosen, often with the aid of intraoperative ultrasound. Parenchymal transection is accomplished with an endoscopic stapling device with staple line reinforcement material using a slow, stepwise compression technique. Splenic vasculature is either ligated en bloc for pancreatic transection sites away from the coeliac trunk or separately when the level of the parenchymal transection site is in proximity or to the right of the coeliac trunk. The distal pancreas is then elevated and transected from the retroperitoneum, spleen mobilization is completed if splenectomy is performed, and the specimen is placed into a retrieval bag and removed through the umbilical trocar site. The spleen is usually morselized in a piecemeal fashion within an endoscopic retrieval bag. Jackson–Pratt drains are occasionally placed in the operative bed at the discretion of the surgeon.

Figure 1.

Diagram showing abdominal access in laparoscopic distal pancreatectomy through an abdominal Hasson trocar and three additional trocars

In ODP, the steps used varied among surgeons, but the procedure was generally performed through a left subcostal incision using standard techniques. Parenchymal transection was accomplished with staplers and individual pancreatic duct ligation in most patients. Splenic vasculature ligation was generally accomplished separately from parenchymal transection.

Statistical analysis

Descriptive data are reported as medians and ranges to summarize continuous variables, or as frequencies and proportions for categorical data. Continuous data were compared using Wilcoxon signed-rank test analysis; categorical data were analysed using the Fisher's exact test or Cochran–Armitage trend test as appropriate. The statistical significance of between-group differences was assumed at a P-value of <0.05. All calculations were performed using sas Version 9.2 (SAS Institute, Inc., Cary, NC, USA).

Research ethics board approval

Appropriate approval to conduct this retrospective review was obtained through the Mayo Clinic Internal Review Board.

Results

Patient demographics

In the study time period, a total of 172 patients underwent distal pancreatectomy at the study institution. Open and laparoscopic distal pancreatectomies were performed in 90 (52%) and 82 (48%) patients, respectively. Table 1 gives patient information including age, sex, BMI, comorbidities, ECOG and ASA status for the two groups. No significant differences in any of these parameters were found between the groups.

Table 1.

Patient demographics and comorbidities

	Open surgery group (n = 90, 52%)	Laparoscopic group (n = 82, 48%)	P-value
Gender, n (%)			0.486
Male	35 (39)	33 (40)
Female	55 (61)	49 (60)
Hypertension, n (%)	48 (53)	47 (57)	0.652
Diabetes mellitus, n (%)	23 (26)	19 (23)	0.731
Cardiac disease, n (%)	23 (26)	18 (23)	0.603
Pulmonary disease, n (%)	25 (28)	20 (24)	0.732
ECOG grade, n (%)			0.344
0	63 (70)	63 (77)
1	25 (28)	17 (21)
2	1 (1)	2 (3)
3	1 (1)	0
ASA class, n (%)			0.140
2	19 (21)	22 (27)
3	65 (72)	59 (72)
4	6 (7)	1 (1)
Age, years, median (range)	64 (28–85)	65 (17–89)	0.448
BMI, kg/m2, median (range)	27.7 (17.2–62.5)	26.8 (17.4–50.8)	0.616

ECOG, Eastern Cooperative Oncology Group; ASA, American Society of Anesthesiologists; BMI, body mass index.

Operations and indications

Data on pancreatic resection procedures and indications for surgery are given in Table 2. All patients undergoing completion distal pancreatectomy were operated on an elective basis. Intraoperative conversion to a hand-assisted method or to an open approach was required in nine (11%) and six (7%) patients, respectively, and was generally performed for significant inflammation, adhesions or the close proximity of tumour to the coeliac trunk.

Table 2.

Type of resection and pathology

	Open surgery group (n = 90, 52%)	Laparoscopic group (n = 82, 48%)	P-value
Surgical approach, n (%)
Distal	88 (98)	81 (99)	1.0
Completion distal	2 (2)	1 (1)	1.0
Splenic preservation	2 (2)	10 (12)	0.013
Pathology, n (%)
Pancreatic adenocarcinoma	21 (23)	18 (22)	1.0
IPMN	21 (23)	18 (22)	1.0
PNET	13 (14)	13 (16)	1.0
Benign/chronic pancreatitis	15 (17)	13 (16)	1.0
Cystic neoplasm	12 (13)	14 (17)	1.0
Miscellaneous solid neoplasm	8 (9)	6 (7)	1.0

IPMN, intraductal pancreatic mucinous neoplasm; PNET, pancreatic neuroendocrine tumour.

Additional abdominal procedures were performed in 28 (31%) and 18 (22%) patients undergoing ODP and LDP, respectively (P = 0.227). These were most commonly cholecystectomy, partial gastrectomy, partial colectomy, adrenalectomy and nephrectomy.

Operative variables and complications

Table 3 gives data on operative variables and complications in the LDP and ODP groups. Forty (44%) patients undergoing ODP and 10 (12%) patients undergoing LDP required at least one unit of pRBC during the operation or subsequent hospitalization (P < 0.001). Overall morbidity and mortality were similar in the LDP and ODP groups. Clinically significant pancreatic fistula (Grades B and C) occurred in 10% and 6% of patients in the ODP and LDP groups, respectively. Major complications (Grades III–V) occurred in 20% and 13% of patients in the ODP and LDP groups, respectively.

Table 3.

Operative variables and complications

		Open surgery group (n = 90, 52%)	Laparoscopic group (n = 82, 48%)	P-value
Estimated blood loss, ml, median (range)		500 (25–5000)	70 (5–1500)	<0.001
Operative time, min, median (range)		233 (55–568)	188 (83–405)	0.431
pRBCs during hospitalization, units, median (range)		0 (0–24)	0 (0–2)	<0.001
ICU stay, days, median (range)		0 (0–5)	0 (0–3)	<0.001
Overall length of stay, days, median (range)		8 (3–18)	4 (1–10)	<0.001
Cardiac complications, n (%)		10 (11)	3 (4)	0.078
Pulmonary complications, n (%)		13 (14)	4 (5)	0.042
Reoperation, n (%)		2 (2)	2 (2)	1.0
Pancreatic fistula (overall), n (%)		13 (14)	11 (13)	1.0
Fistula grade, n (%)				0.213
Grade A		4 (4)	6 (7)
Grade B		1 (1)	1 (1)
Grade C		8 (9)	4 (5)
Post-pancreatectomy haemorrhage (overall), n (%)		3 (3)	1 (1)	0.624
Haemorrhage grade, n (%)				0.357
Grade A		2 (2)	0
Grade B		1 (1)	0
Grade C		1 (1)	1 (1)
Delayed gastric emptying rate, n (%)		4 (4)	0	0.122
Wound infection, n (%)		5 (6)	2 (2)	0.447
Intra-abdominal abscess, n (%)		5 (6)	4 (5)	1.0
Clavien–Dindo complications (30-day/inpatient), n (%)				0.987
Minor, n (%)	Grade I	15 (17)	11 (14)
Moderate, n (%)	Grade II	20 (22)	5 (6)
Severe, n (%)	Grades III–V	18 (20)	11 (13)
	Grade IIIa	13 (14)	6 (7)
	Grade IIIb	2 (2)	3 (4)
	Grade IVa	1 (1)	2 (3)
	Grade IVb	1 (1)	0
	Grade V	1 (1)	0
Discharged to:				0.193
Home, n (%)		85 (94)	80 (98)
Facility, n (%)		5 (6)	2 (2)

pRBC, packed red blood cells; ICU, intensive care unit.

Two patients in each group required reoperation. Both patients in the ODP group underwent reoperation to control acute arterial haemorrhage. Reoperations in patients in the LDP group included one for a postoperative haematoma evacuation and one for fascial dehiscence at the umbilical site used for specimen extraction.

Oncologic outcomes

Subgroup analysis of oncologic outcomes in patients undergoing ODP or LDP for pancreatic adenocarcinoma or PNET is given in Table 4. Tumour size, TN stage, number of positive lymph nodes, total lymph node harvest, lymph node ratio and margin status did not differ significantly between the two groups.

Table 4.

Oncologic outcomes in patients with pancreatic adenocarcinoma and neuroendocrine tumours undergoing distal pancreatectomy

	Open surgery group (n = 34, 38%)	Laparoscopic group (n = 31, 38%)	P-value
Tumour size, cm, median (range)	2.8 (0.5–15)	2 (0.5–7.5)	0.247
Positive lymph nodes, n, median (range)	0 (0–6)	0 (0–9)	0.793
Total lymph nodes, n, median (range)	11 (1–45)	16.5 (0–48)	0.152
Lymph node ratio (N1 patients only), median (range)	0.130 (0.030–0.670)	0.120 (0.020–0.250)	0.888
Tumour stage, n (%)			1.0
T1	7 (21)	12 (40)
T2	3 (9)	3 (10)
T3	18 (53)	14 (47)
T4	1 (3)	0
Unknown	5 (15)	1 (3)
Node stage, n (%)			0.779
N0	21 (62)	23 (74)
N1	10 (29)	8 (26)
Unknown	3 (9)	0
Surgical margin, n (%)
R0	32 (94)	30 (97)	1.0
R1	2 (6)	1 (3)

Discussion

Over the past decade, substantial evidence for the feasibility and safety of the laparoscopic approach in resections of the left-sided pancreas has emerged. In addition, recent meta-analyses performed on the world literature^13–15 have demonstrated significant advantages over the open approach. Despite this, a recent look at administrative databases shows that LDP appears to be substantially underused across the USA.¹⁷ At the reporting institution, the technique of LDP was not introduced until August 2008, with the arrival of the senior author. Prior to this, with the exception of one patient, all distal pancreatic resections had been performed using an open technique. From August 2008, only 16 (17%) of 97 distal pancreatic resections were performed using an open approach. Eight of these ODPs in the more recent period were performed for benign disease in the presence of severe inflammation that precluded laparoscopic dissection. The other half were performed for malignant disease that involved large bulky tumours or significant prior upper abdominal surgery or multivisceral organ resection.

To the authors' knowledge, there have been only four comparative series with more than 50 patients.^9–12 This study analyses the changes in outcomes when a surgeon with prior expertise in the procedure introduced LDP into a well-established open pancreatic practice at a tertiary teaching institution. Hence, the data are not affected by a learning curve. There were no differences in patient demographics, comorbidities, BMI or underlying pathology among patients undergoing either approach over the whole study period. No differences were anticipated because referral patterns were stable over the course of the study period and the transition from an exclusively open approach to a minimally invasive approach in the majority of patients implies the absence of any significant selection bias. However, it is conceivable that a potential selection bias might emerge over a longer study period and in a larger patient sample. However, it is unclear which patient group would be affected by such a bias because patients who required ODP in the more recent period of the study were often younger (mean age: 56.6 years versus 62.5 years; P = 0.121) and more often had benign chronic pancreatitis (50% versus 11%; P = 0.001) compared with patients undergoing ODP in the earlier part of the study period. Furthermore, the greatest advantages of LDP are often seen in patients of advanced age, higher BMI and with more comorbidities. However, it is clear that patients who underwent ODP in the more recent period were at greater risk for complications either because the magnitude of surgery was greater or because of the presence of chronic pancreatitis or more advanced tumour characteristics, such as attachment to the coeliac trunk or encasement of the superior mesenteric–portal vein trunk. Non-statistically relevant increases in major morbidity, EBL and operative time were observed in the 16 patients who underwent ODP after August 2008 compared with those who underwent ODP in the earlier part of the study period. Rates of pancreatic fistula (31% versus 11%; P = 0.048), mean pRBC transfusions (4.9 units versus 2.0 units; P = 0.022) and mean LoS (9.4 days versus 7.8 days; P = 0.035) were significantly higher in ODP patients in the more recent period compared with the overall group of ODP patients in the earlier part of the study period.

A higher rate of splenic preservation was seen in patients undergoing LDP in this series, similarly to that seen in a Korean study.⁹ Although much of this discrepancy can be attributed to differences in the individual management and aggressiveness of splenic salvage among surgeons performing distal pancreatectomy, the authors feel that splenic vessel preservation is much easier under the magnification afforded by laparoscopy. The underlying indications for ODP and LDP did not change throughout the study period and therefore do not represent a cause for this difference.

Overall hospital LoS and ICU utilization were significantly lowered by the use of LDP. Pulmonary complications were also lower in LDP, but operative time, overall morbidity and the occurrence of pancreatic fistula were not significantly affected by the approach. A tendency towards a rate of decreased fistula in LDP was nevertheless noted. The 6% rate of fistulae of Grades B and C compares favourably with rates reported in the majority of the literature.^14,15 Although a definitive conclusion cannot be drawn from this particular study, it may be that this lower rate is attributable to the specific technique of stapled division using slow compression, as previously described.⁶

Estimated blood loss and transfusion requirements in patients undergoing LDP were dramatically lower than in patients undergoing ODP over the study period. As the two groups were similar in underlying indications and other patient comparison measurements, it appears that this disparity is almost entirely related to the differences between ODP and LDP in dissection technique. Laparoscopic distal pancreatectomy is associated with more precise dissection performed under magnification using a ‘no-touch’ technique. The tumour is generally not manipulated and the vasculature is surrounded and transected early in the procedure using a right-to-left approach. Additionally, the short gastric vessels and diaphragmatic attachments of the spleen are all controlled under direct visualization during LDP, which is not always possible in ODP.

Comparisons of oncologic outcomes in patients undergoing LDP and ODP, respectively, for pancreatic adenocarcinoma or PNET were analysed to assess whether there was any compromise in the extent of oncologic resection in the laparoscopic approach. Laparoscopic distal pancreatectomy is still used selectively by some surgeons at other institutions who consider the presence of malignancy as a contraindication to laparoscopy. The comparison showed a smaller tumour size, but a higher total lymph node harvest and lower lymph node ratio in patients undergoing LDP. However, none of these differences reached statistical significance. Tumour stage, nodal status and margins were also similar between the groups. Longterm outcomes for LDP are not yet available, but surrogate markers for improved survival, including need for blood transfusion, lymph node harvest and ratio, and margin status show that the LDP approach is at the least not inferior to ODP and may even provide benefits in patients undergoing distal pancreatic resection for malignancy. Strasberg et al.²⁵ have pointed out that modification of the dissection plane to the posterior of the adrenal gland and Gerota's fascia, also known as posterior radical antegrade modular pancreatosplenectomy (RAMPS), will occasionally be necessary to achieve negative tangential margins in carcinoma of the body or tail of the pancreas in order to obtain satisfactory survival rates. Although the procedures reported by Strasberg et al.²⁵ were performed using an open approach, the present authors believe that this modification can be readily performed when completing LDP. In fact, four patients undergoing LDP for malignancy (13%) underwent a successful RAMPS for tumours extending posteriorly into this plane. None of the conversions were performed for this reason, although one patient required conversion for close tumour involvement of the coeliac trunk vessels.

At present, the evidence for the treatment of pancreatic malignancy of the body and tail, namely adenocarcinoma, using LDP remains uncertain. Nonetheless, the present authors consider that a bloodless, margin-negative resection with adequate or improved peripancreatic lymph node harvest is attainable using this approach and that the presence of malignancy should not contraindicate its use. However, as previously reported,⁶ one patient was found to have a trocar site recurrence 2 months after undergoing an uncomplicated LDP. The recurrence was successfully excised and the patient remained disease-free for >1 year, but suffered eventual disease recurrence and died at 23 months. Longterm follow-up with matched comparisons are required before any definitive conclusions on differences in survival for pancreatic malignancy can be made.

At present, it appears that there is enough evidence in the literature to postulate that LDP should play a more standard role and should be considered as the procedure of choice in distal pancreatic resection. An open procedure, however, may still be preferred for selected patients, particularly in the presence of significant chronic inflammation or a large lesion in direct proximity to the coeliac trunk.

Conclusions

This series suggests that the laparoscopic approach can be safely introduced into an experienced open pancreatic practice without negative effect when there is no need for a learning curve. As in other studies, the advantages of the laparoscopic approach relate to decreases in blood loss, number of transfusions required, hospital LoS and time to overall recovery. Findings for other outcomes tended to favour LDP, but no significant differences emerged between the procedures except in the rate of pulmonary complications. Oncologic principles as measured by surrogate markers appear not to be compromised and, in fact, some were improved in the laparoscopic group. However, longterm survival data and larger series are required to support any definitive conclusions on the potential advantages of the laparoscopic approach in patients with cancer.

Conflicts of interest

None declared.