Combined pancreaticoduodenectomy and colon resection for locally advanced peri-ampullary tumours: analysis of peri-operative morbidity and mortality

Abstract

Background

Combined pancreaticoduodenectomy (PD) and colonic resection may be necessary to achieve an R0 resection of peri-ampullary tumours. The aim of this study was to examine the morbidity and mortality associated with this procedure.

Methods

A retrospective cohort study was performed comparing 607 patients who underwent a standard pancreaticoduodenectomy (S-PD) to 28 patients who had a concomitant colon resection and PD (PD-colon) over a 10-year period at an academic centre.

Results

Patients in the PD-colon group were more likely to have received neoadjuvant chemotherapy ± radiation (3/28, 11% versus 14/607, 2%, P = 0.024). Operative time was also longer (530 versus 410 min, P < 0.001) and they were more likely to have had portal vein resections (9/28, 32% versus 76/607, 13%, P = 0.007). There was no difference in the intra-operative blood loss, length of stay, or overall complication rates. The PD-colon group had a higher rate of severe post-operative bleeding (4/28, 11% versus 8/607, 1%, P = 0.002). The post-operative mortality rates for the PD-colon and PD groups were 2/28 (7%) and 8/607 (1%), respectively (P = 0.068).

Conclusions

PD-colon has an acceptable risk of peri-operative morbidity compared with S-PD in well-selected patients.

Introduction

Tumours involving the pancreatic head may arise from several locations including the pancreas, duodenum, ampulla and colon, and the surgical approach for these lesions involves a pancreaticoduodenectomy (PD). Historically this procedure has been fraught with high complications and mortality rates,¹ but over the past decade mortality rates have decreased substantially to less than 5% in experienced high-volume centres.^2–4 As a result there has been increased interest to extend the resection limits beyond the traditional PD in order to achieve negative margins.^5,6 This is of particular importance in the case of pancreatic adenocarcinoma as up to one-third of patients present with locally advanced disease.⁷ Several studies have shown that PD can be safely combined with a vascular resection with no significant difference in outcome.^8–10 Furthermore, several studies have reported that PD can be combined with resection of other organs without a significant increase in morbidity or mortality.^10–17 This paper study outlines the peri-operative outcomes for 28 patients who had PD with colon resection (PD-colon) at a high-volume academic centre, representing the largest series of patients reported in the literature to date.

Patients and methods

Study design

The institution's research ethics board approved this study. A retrospective chart review was conducted of all patients who underwent PD ± other procedures between 1 January 1 2000 and 30 December 2010 at the University Health Network in Toronto Canada. Demographic, operative and peri-operative data were collected and analysed to determine differences in morbidity and 90-day mortality.

Surgical technique

The standard PD (S-PD) was performed in a classic manner. Pylorus-sparing techniques were not used. All colon anastomoses were side-to-side functional end-to-end ileocolic anastamoses created using either a stapled or hand-sewn technique, depending on surgeon preference. All colonic resections were performed by the hepato-pancreatico-biliary (HPB) surgeon. An ‘en bloc’ resection was utilized when tumours were directly invading the colon whereas separate resections were done for secondary vascular compromise. In these instances, the mesentery was taken with the initial specimen and the colon was observed for the remainder of the case to determine viability. Intra-operative drain placement was used selectively, according to surgeon preference.

Post-operative complications

A positive margin was defined as any positive margin on the PD. Post-operative complications were defined using the Clavien–Dindo¹⁸ classification system and ‘major’ classifications were those with a score ≥3. A pancreatic leak or fistula was determined using the definition from the International Study Group on Pancreatic Fistula¹⁹ and delayed gastric emptying (DGE) as per the International Study Group of Pancreatic Surgery.²⁰ A post-operative haemorrhage was defined as per the International Study Group of Pancreatic Surgery.²¹ Intra-abdominal sepsis was defined as symptomatic post-operative fluid collections requiring percutaneous or operative drainage. Post-operative mortality was defined by any death occurring during the index hospital admission or within 90 days of surgery.

Statistical analysis

Data were analysed using a statistical software package (SPSS Version 21; SPSS Inc., Chicago, IL, USA). Results were reported as median (range) unless otherwise specified. Univariate analysis was performed using the Mann–Whitney U-test for continuous variables and Fisher's exact test for categorical values. P-values are two-tailed with P < 0.05 considered significant.

Results

Of the 635 patients who underwent a PD during this period, 28 patients were identified who had a concomitant colon resection (PD-colon) and these were compared with the remaining 607 patients who had a standard PD (S-PD). Demographics of the two groups are shown in Table 1. Comparison of peri-operative factors are shown in Table 2. There was no difference in major complication (7/28 versus 103/607, 17%; P = 0.304), pancreatic fistulae (2/28 versus 77/607, 13%; P = 0.472) or DGE (3/28 versus 49/607, 8%; P = 0.575) rates. For patients who had a concomitant bowel resection the anastomotic leak rate was 7% (2/28). The rate of positive margins was the same in both groups (2, 7% PD-colon; 49, 8% S-PD; P = 1). The positive margins were of the PD specimen, there were no positive colonic margins.

Table 1.

Demographic data for patients undergoing a pancreaticoduodenectomy (PD) with colon resection or standard PD

	PD-colon n = 28	%; range	PD n = 607	%; range	P
Female : Male	9:16		254:356		0.563
Female		36		42
Age (years)	59	23, 75	64	17, 84	0.014a
Size (cm)	4	1, 11.5	2.8	0, 23	<0.001a
Primary diagnosis
Pancreatic adenocarcinoma	9	32	249	41	<0.001a
Ampullary cancer	6	21	129	21
Duodenal cancer	5	18	23	4
Colon cancer	2	7	2b	<1
Cholangiocarcinoma	0	0	76	13
Neuroendocrine tumour	1	3	40	7
Other	5	18	88	15

^aDenotes statistical significance.

^bThese patients were found to have colon cancer metastases to the peri-ampullary region requiring a pancreaticoduodenectomy without a colon resection.

Table 2.

Peri-operative comparisons between patients undergoing a pancreaticoduodenectomy (PD) and colon resection versus a standard PD

	PD-colon n = 28	%; range	PD n = 607	%; range	P
Neoadjuvant chemotherapy	3	11	14	2	0.024a
Neoadjuvant radiotherapy	3	11	10	2	0.016a
Pre-operative biliary drainage	14	50	368	63	0.324
Length of stay (days)	9	5, 332	10	4,144	0.530
Operative time (min)	530	405, 675	410	232, 875	<0.001a
Estimated blood loss (ml)	800	400, 5000	700	100, 6500	0.089
Portal vein resection	9	32	76	13	0.007a
Margin positivity	2	7	49	8	1.00
Lymph node positivity	20	71	328b	57	0.296
Major complications	7	25	103	17	0.304
Post-operative bleed	6	21	31	5	0.004a
Intra-abdominal	5		18
Retroperitoneal	0		2
Gastrointestinal	1		14
Haematuria	0		1
GI + intra-abdominal	0		1
Severe bleed	4	14	8	1	0.002a
Pancreatic fistula	2	7	77	13	0.472
Gastrojejunostomy leak	0		6	1	0.621
Biliary leak	0		4	<1	0.637
Intra-abdominal sepsis	7	28	81	14	0.208
Delayed gastric emptying	3	11	49	8	0.575
EnBloc resection	16	57
Bowel leak	2	7
Reason for resection
Tumour involvement	6	21
Vascular compromise	19	68
Other	3	11
Post-operative mortality	2	7	8	1	0.068
Median follow-up	19	0, 131	25	0, 255	0.042
Median survival (months)	15		37		0.082
Overall survival					0.015a
1 year	54%		79%
3 years	42%		61%
5 years	35%		29%

^aDenotes statistical significance.

^bDenotes data missing from 27 patients.

There was a significant difference in the rate of severe post-operative haemorrhage in the groups (4/28 PD-colon; 8/607, 1% S-PD; P = 0.002). Further examination revealed that four of the PD-colon patients who required a post-operative transfusion also had a portal vein resection as part of their procedure. Of the two patients in the PD-colon group who died, one died from sepsis from an ischaemic leg, the cause of death from the other was unknown. Both patients developed a leak at the ileocolic anastomosis. Of the eight patients who died in the S-PD group, two were a result of multi-system organ failure after intra-abdominal sepsis, and one patient died after a stroke when the family decided to withdraw care. The cause of death for the remaining patients is unknown. The median 1-, 3- and 5-year survivals for both groups are shown in Table 2.

Discussion

This study reports a cohort study of patients undergoing a PD-colon resection for peri-ampullary tumours. The results suggest that patients who underwent a simultaneous PD and colon resection were more likely to have received neoadjuvant treatment than those who underwent a standard PD. As expected, this complex multi-visceral procedure is technically complex and is often combined with a vascular resection (PV resection rate 32% versus 13%). Importantly, post-operative complications including pancreatic leak, biliary leak, intra-abdominal sepsis and DGE as well as length of stay were similar between the PD and PD-colon groups. These findings are in keeping with other reported studies.^16–20 Major post-operative haemorrhage and anastomotic leak were higher in patients who underwent PD-colon. This may be as a result of the complexity of the performed procedure. It was not our practice to routinely divert ileocolic anastomosis in elective operations. Although there was no statistically significant difference in lymph node positivity or median survival between groups there was a trend towards the increased number of involved nodes and decreased survival in the PD-colon group which may be reflective of more aggressive tumour biology.

To date there are few studies that have evaluated the outcomes for patients undergoing a simultaneous colon resection and PD. Several smaller studies have reported similar findings to the present study: mortality and morbidity rates were comparable between PD and the extended resection groups. These studies also consistently show longer operative times for the more complex procedure and a higher rate of portal vein resection.

The limitations of this study include the retrospective nature over a relatively long (11 years) study period. As a result of the small numbers of patients in the study group, this study may be underpowered to detect small differences in outcomes, resulting in a type II error. Furthermore, the need for a vascular resection in many of the patients undergoing PD-colon makes it difficult to differentiate between the morbidity of the vascular resection as compared with the intestinal resection. It is likely that the finding of higher peri-operative bleeding risk in patients who underwent a PD-colon resection is primarily attributable to the additive need for PV in this patient population. As a result, it is important to validate the results of the present findings in further studies.

This is the largest reported series to date of patients undergoing a combined colon resection and pancreaticoduodenectomy for peri-ampullary tumours. The results suggest that while there is no significant difference in most post-operative complications, there is a higher risk of post-operative bleeding.

This study reports that an extended resection with PD-colon may be required to achieve a R0 margin in patients with locally advanced disease. It is associated with an acceptable but increased risk of peri-operative morbidity compared with PD alone. As a result, careful patient selection should be taken before recommending patients undergo this complex procedure.

Conflicts of interest

None declared.