Abstract
Background
The advantage of a pancreaticogastrostomy (PG) over a pancreaticojejunostomy (PJ) after a pancreaticoduodenectomy (PD) is not clear.
Aim
The aim of the present study was to compare the pancreatic fistula (PF, defined according to the International Study Group for Pancreatic Fistula classification) rate and other complications between both methods.
Methods
Retrospective analysis of prospectively collected data of 424 [median: 65 years (17–83)] patients who underwent PG (239, 56.4%) and PJ (185, 43.6%) reconstruction between January 2005 and December 2009.
Results
PF occurred in 55 (23.5%) in the PG and 30 (16.2%, P = 0.067) patients in the PJ group. Grade A PF occurred in 19 (7.9%), B in 22 (9.2%) and C in 14 (5.8%) in the PG compared with 5 (2.7%), 12 (6.5%) and in 13 (7.0%), respectively, in the PJ group. The median hospital was 10 days in both groups. The morbidity was higher in the PG group (108, 45.2 vs. 62, 33.5%, P = 0.015). However, there was no significant difference in the 90-day mortality between both groups (PG-17, 7.0% vs. PJ-16, 8.6%, P = 0.558).
Conclusion
There was no difference in the overall PF rate, hospital stay and overall mortality between PG and PJ reconstruction methods. However, the grade A PF rate was higher in the PG group.
Keywords: resection, outcomes
Introduction
A pancreaticoduodenectomy (PD) is the only available curative treatment option for patients with pancreatic head and peri-ampullary malignancy. However, PD is associated with significant morbidity and mortality mainly as a result of leakage from pancreatic anastomosis (pancreatic fistula, PF). The rate of PF is highly variable, ranging from 2% to 30%1–9 depending on the type of definition used. A pancreaticojejunostomy (PJ) is the most commonly used reconstructive method after PD and several alternative techniques and of variations of PJ have been described to reduce the PF rate. A pancreaticogastrostomy (PG) is the most widely used alternative reconstructive method. However, the advantage or disadvantages of PG over PJ is not clear.10–20 Evidence from several non-randomized studies suggests that the morbidity and mortality secondary to PF after PG is much less compared with PJ anastomosis. Contrarily, the evidence from randomized controlled studies suggests that there was no difference between both methods.
The major criticism of the previously published studies was that there were wide variations in the definition of PF precluding accurate comparison of the results. To allow the comparison of various techniques and results from various centres, an International Study group for Grading of Pancreatic Fistula (ISGPF) has developed a classification system for the grading of PF based on both clinical and biochemical parameters.21 Since its publication in 2005, there were no major studies which studied the rate of PF (graded according to the ISGPF system) and associated complications between PG and PJ methods after PD. The aim of the present study was to examine the rate of PF and associated morbidity and mortality in a large group of patients who underwent PG and PJ reconstruction after PD.
Material and methods
Between January 2005 and December 2009, retrospective analysis was undertaken of all prospectively collected data of all patients who underwent PD at the Liver Unit, Queen Elizabeth Hospital. Six consultant surgeons (J.B., A.D.M., D.F.M., S.B., J.I. and P.M.) with significant experience in pancreatic resections performed or supervised all PD. Patients with multi-visceral pancreatic resections and distal pancreatectomies were excluded from the present study. The data on pre- per- and post-operative parameters were collected prospectively and maintained on a secure pancreatic database by a full-time database manager. The following data were collected on the database: patient's demographics, presenting symptoms, past medical history, pre-operative radiology and laboratory results, and pre-operative interventions such as endoscopic retrograde cholangio-panctreaticography (ERCP)/percutaneous transhepatic cholangiography (PTC), the details of surgery, post-operative morbidity and mortality, details of all the post-operative interventions, the cause of death and the histology.
Per- and post-operative management
The institution has a pancreaticoduodenectomy care pathway in place and this was used to standardize care to ensure uniform post-operative management. All patients were given prophylactic antibiotics (as per the local antibiotic guidelines) at the time of induction and antibiotics were continued for 24 h after surgery. Prophylactic octreotide (100 µg subcutaneously three times a day) was routinely given to all patients for 5 days. All patients had nasogastric (NG) and nasojejunal (NJ) tubes placed at the time of surgery and were started on enteral feed via a nasal gastric tube on the first post-operative day at 10 ml/hour and increased to 30 ml/hour on day three and the feed continued until day 6 and for longer if necessary. Abdominal drain volumes were measured every 24 h and drain fluid amylase (DFA) levels were measured on day 5. The timing of the drain removal was left to the individual surgeon's discretion. On average, the drain was removed on day six but they were left in for longer if the DFA was more than three times the serum levels, or if the contents were bilious, enteric, milky, water or ‘prune juice’ in appearance. Once the surgeon was satisfied that there was no evidence of PF, bile leak or enteric leak the NG and NJ tubes were removed on day six and patients were allowed to have clear fluids and gradually built up to normal diet over next 48 h. Where appropriate, patients with PF were also treated with antibiotics, total parental nutrition (TPN), prolonged drainage, radiological procedures or surgery as appropriate.
Surgical technique
The pancreatic reconstruction was performed using either a PG or PJ reconstruction technique depending on the individual surgeon's choice. The techniques of PJ reconstruction were: the end-to-end dunking technique or two-layered duct-to-mucosa anastomosis. For both techniques, a 1.5- to 2-cm cut end of remnant of the pancreas was mobilized from the splenic vein. The proximal end of the retained jejunum is brought through a right transverse mesocolic window. The end-to-end pancreaticojejunostomy was created by invaginating the end of the pancreatic remnant into the end of the jejunum using a single-layer, continuous 3–0 Polypropylene (Prolene) suture (Ethicon, Johnson & Johnson, UK). The two-layered duct-to mucosa PJ reconstruction was created by interrupted 4–0 Poly dioxone (PDS, Ethicon, Johnson & Johnson, UK) for the duct to-mucosa anastomosis and 3–0 PDS for the parenchymal suturing as previously described by Z'graggen et al. in 2002.22
The pancreaticogastrostomy was carried out under direct vision through an anterior gastrostomy. First, a 4–5 cm stump of the pancreas was mobilized from the splenic vessels and retroperitoneum. The stump of the pancreatic remnant brought out through an oblique incision in the posterior wall of the mid-body of the stomach and sutured to the posterior wall of the stomach with continuous 3–0 Prolene ensuring that the pancreatic duct lumen is not obliterated. An anterior wall gastrostomy is closed with a continuous 3–0 PDS suture at the end of the procedure.
After completion of a pancreatic reconstruction a retro-colic hepatico-jejunostomy was performed with a continuous 4–0 or 5–0 double-ended PDS suture as previously described.23 This was followed by a single-layer, end-to-side duodeno-jejunostomy (DJ) or gastrojejunostomy (GJ) with a continuous 3–0 PDS suture approximately 40–50 cm from the biliary anastomosis. The NJ and NG tubes were placed into the efferent jejunal loop and stomach, respectively, before the closure of the anterior layer of the DJ/GJ anastomosis. A 32-F Robinson's portex drain® was placed posterior to the pancreatic and biliary anastomosis and brought out through the lateral abdominal wall on the right-hand side and secured with a 1–0 silk suture and connected to a closed non-vacuum drainage system.
Pancreatic fistula and other complications
A PF was classified into three grades based on ISGPF criteria.21 The full details of the classification are summarized in Table 1. The definitions of various other complications are described in Table 2. PF was defined as drain output of any measurable volume of fluid on or after post-operative day 3 with amylase content greater than three times the serum amylase activity. This definition also includes all peri-pancreatic collections, abscesses, leaks and fistulae thought to manifest from poor anastomotic healing. PF has been classified into three grades: A, B and C.21
Table 1.
The ISGPF criteria used for grading of the pancretic fistula (Adapted with permission from Bassi C, Dervenis C, Butturini G et al. Post-operative pancreatic fistula: an international study group (ISGPF) definition21
| Criteria | No fistula | Grade A fistula | Grade B fistula | Grade C fistula |
|---|---|---|---|---|
| Drain amylase | <3x normal serum amylase | >3x normal serum amylase | >3x normal serum amylase | >3x normal serum amylase |
| Clinical conditions | well | Well | often well | ill appearing/bad |
| Specific treatment | No | No | Yes/No | Yes |
| US/CT (if obtained) | Negative | Negative | Negative/Positive | Positive |
| Persistent drainage > 3 weeks | No | No | Usually Yes | Yes |
| Signs of infection | No | No | Yes | Yes |
| Readmission | No | No | Yes/No | Yes/No |
| Sepsis | No | No | No | Yes |
| Re-operation | No | No | No | Yes |
| Death related to fistula | No | No | No | Yes |
US, ultrasound; CT, computer tomogram
Table 2.
The definitions of various complications
| Complication | Definition |
|---|---|
| Wound infection | Any evidence of infection (i.e. erythaema, purulent discharge, induration) requiring antibiotic treatment, or evidence of dehiscence |
| Post pancreatectomy haemorrhage | Haematemesis or melena and no other source of ongoing blood loss, or the sudden appearance of frank blood either in the NG tube or per rectum, with subsequent fall in haemoglobin of 2 g/dl, and requiring blood transfusion or reoperation or radiological intervention or endoscopic intervention) |
| Delayed gastric emptying | Failure to resume oral liquid intake by post-operative day 10, and/or emesis > 500 ml on or after post-operative day 5, and/or continued nasogastric drainage > 500 ml on after post-operative day 5 |
| Acute renal failure | Serum creatinine doubling of baseline value, and/or need for dialysis |
| Acute respiratory failure | Respiratory distress requiring intubation, or the need for intubation or mechanical ventilation for more than 24 h post-operatively |
| Enteric leak | Leakage of enteric contents from intra-operatively placed drains, and/or radiographically confirmed fluid collection, requiring surgical, endoscopic, or radiographic intervention |
| Bile leak | Bilious drainage from intra-operatively placed drains, and/or radiographically confirmed fluid collection, requiring surgical, endoscopic, or radiographical intervention |
| Chest infection | Presence of body temperature > 38°C, abnormal elevation of white blood count, or positive sputum Gram stain or culture along with presence of clinical or radiological signs of chest infection and requiring intra venous or oral antibiotic treatment |
| 90-day mortality | Death during the initial hospitalization or within 90-days of hospital discharge |
| Hospital stay | The number of days in hospital from the time of initial operation to hospital discharge |
Statistical analysis
Statistical analysis was performed using Statistical Package for Social Sciences (SPPS, Inc., Chicago, IL, USA) version 16. Qualitative data were compared using a χ2-test with Yates' correction or a Fisher's exact test when necessary. Normally distributed data are presented as mean ± standard deviation (SD) and other data presented as median and range. The independent t-test was employed to compare the means of the two normally distributed groups. Non-parametric tests were employed to examine the difference between the two samples when the data had a highly skewed distribution. All P-values were two-tailed, and the P < 0.05 was considered to indicate statistical significance.
Results
Overall group
Between January 2005 and December 2009, 424 patients underwent (pylorus-preserving PD (PPPD) = 385; Kausch–Whipple's PD (KWPD) = 39) a pancreaticoduodenectomy. One hundred and seventy patients (40.0%) had post-operative complications. The overall PF rate was 20.0% (85 patients): grade A in 24 (5.6%), B in 34 (8.0%) and grade C in 27 (6.3%) patients. Seventeen patients (4.0%) had an enteric leak, and in 13 of these the leak was associated with PF (grade B = 5 & C = 8). Seven out of 17 patients underwent a re-laparotomy for a leak from DJ (n = 5), leak from PJ (n = 1) and leak from both PJ and DJ anastomosis (n = 1). A bile leak occurred in eight patients and all of them except one were managed conservatively. In total, 18 (4.2%) patients required a re-laparotomy for various indications and 31 (7.3%) patients required radiological intervention. The median hospital stay was 10 days (range: 7–141). The median hospital stay was 21 days (7–141) in the PF group compared with 9 days (range: 7–60, P = 0.001) in the non-fistula group. The 90-day mortality was 7.8% (33 patients). The main causes of death were sepsis with multi-organ failure in 15 (3.5%), bleeding in 9 (2.1%), respiratory failure in 3 (0.7%) and other causes in 6 (1.4%) patients. Sixteen (57.1%) out of 28 patients with a post-pancreatectomy haemorrhage (PPH) died within 90-days. Eight (2.5%) out of 323 patients with no PF and PPH died within 90 days compared with 10 out of 13 (76.9%) (P = 0.001) with a PF and PPH. Nine (12.3%) out of 73 patients with a PF and no PPH died within 90 days compared with six (40.0%) out of 15 patients (P = 0.001) with a PPH alone.
PG and PJ groups
A pancreaticogastrostomy was performed in 239 (56.4%) and a PJ was performed in 185 (43.6%) patients. In the PJ group, 128 (69.2%) had end-to-end dunking type reconstruction and 57 (30.8%) had two-layered duct-to-mucosa reconstruction. Patient demographics, pre-operative, operative, post-operative and histological details of both groups are summarized in Table 3. Two groups were similar with respect to age, pre-operative biliary drainage, site and size of the tumour, indication and type of resection. However, there were significantly more male patients in the PJ group.
Table 3.
Patient characteristics, pre-operative, per-operative and post-operative parameters of patients in the PG and PJ groups
| Varaibles | Panctreatico-gastrostomy (%) | Panctreatico-jejunostomy (%) | P-value |
|---|---|---|---|
| Number of patients | 239 (56.4) | 185 (43.4) | |
| Age in years (median and range) | 65.7 (17–83) | 65.8 (32–83) | 0.839 |
| Gender, n | |||
| Female | 121 (50.6) | 69 (37.6) | 0.006 |
| Male | 118 (50.4) | 116 (62.4) | |
| Pre-operative biliary drainage, n | 100 | 84 | 0.452 |
| Indication for resection, N | |||
| Benign lesions | 26 | 27 | 0.515 |
| Malignant tumours | 204 (85.4) | 151 (81.6) | |
| Neuroendocrine tumours | 9 | 7 | |
| Type of PD, n | |||
| Classic PD | 18 | 21 | 0.181 |
| PPPD | 221 (92.5) | 164 (88.6) | |
| Portal vein resection, n | 27 | 21 | 0.866 |
| Site of the lesion, n | |||
| Ampullary | 57 | 48 | 0.483 |
| Head of pancreas | 138 (57.7%) | 99 (53.5%) | |
| Distal bile duct | 24 | 26 | |
| Duodenum | 13 | 5 | |
| Others | 7 | 7 | |
| Size of the tumour, median | 3 cm (0.5 to 17) | 3 cm (0.9 to 11) | 0.414 |
| Hospital stay (median and range), days | 10 (7–141) | 10 (7–77) | 0.199 |
| Overall pancreatic fistula rate, n | 55 (23.4) | 30 (16.2) | 0.067 |
| Grade of pancreatic fistula, n | |||
| Grade A | 19 (7.9%) | 5 (2.7%) | 0.019 |
| Grade B | 22 (9.2%) | 12 (6.5%) | 0.368 |
| Grade C | 14 (5.8%) | 13 (7%) | 0.690 |
PD, pancreaticoduodenectomy; PPPD, pylorus preserving pancreaticoduodenectomy
Pancreatic fistulae
Pancreatic fistulae occurred in 55 (23.5%) in the PG and 30 (16.2%, P = 0.067) patients in the PJ group. A grade A PF occurred in 19 (7.9%), B in 22 (9.2%) and C in 14 (5.8%) in the PG group compared with a grade A in 5 (2.7%), B in 12 (6.5%) and C in 13 (7.0%) in the PJ group. There was no significant difference in the overall PF rate between the two sub-groups of PJ reconstruction methods.
Grade A fistulae
A total of 19 (7.9%) in the PG group had a PF compared with 5 (2.7%) patients in the PJ group (7.9 vs. 2.7%, P = 0.019). The median hospital stay was 12 days (range: 7–40) in the PG and 13 (range: 10–37) days in the PJ group (P = 0.239). Three patients in the PG and two patients in the PJ group had a computed tomography of the abdomen (CTA) for the investigation of pyrexia associated with raised inflammatory markers. The CTA did not show any evidence of intra-abdominal collection (IABC) in all patients. Three patients in the PG and one patient in the PJ group required antibiotics for a chest and wound infection, respectively. The 90-day mortality was zero in both groups.
Grade B fistulae
Grade B fistulae occurred in 22 (9.2%) in the PG and 12 (6.5%) patients in the PJ group (9.2 vs. 6.5%, P = 0.368). The details of investigations, further complications and the management of patients with grade B fistula are summarized in Table 4. The median (range) hospital stay was 21.5 (8–43) days in the PG and 23 (13–50) days in the PJ group (P = 0.631). Eighteen in the PG and nine patients in the PJ group had signs of infection (P = 0.676). Nine in the PG and three patients in the PJ group had IABC. There were no mortalities in both groups.
Table 4.
The details of further investigations and findings and treatment of patients with grade B fistula
| Variables | PG, n = 22 | PJ, n = 12 | P-value |
|---|---|---|---|
| Investigations, n | 19 | 7 | 0.270 |
| CT | 17 | 7 | |
| US | 2 | 0 | |
| Radiology findings, n | |||
| Free fluid | 1 | 1 | |
| Liver abscess | 1 | 0 | |
| Intra abdominal collection | 9 | 3 | 0.537 |
| Signs of infection, n | 18 | 9 | 0.676 |
| Treatment, n | |||
| Radiological drainage | 5 | 3 | |
| Antibiotics | 17 | 9 | |
| Drain left longer | 3 | 4 | |
| Total parental nutrition | 10 | 8 | |
| Re-operation | None | None | |
| Other complications, n | |||
| Enteric leak | 4 | 1 | 0.635 |
| Bile leak | 1 | 2 | 0.279 |
| Chest infection | 4 | 4 | 0.410 |
| Wound infection | 2 | 2 | 0.602 |
| 90-day mortality | 0 | 0 | |
| Median (range) hospital stay, days | 21.5 (8–43) | 23 (13–50) | 0.631 |
CT, computer tomogram; US, ultrasound
Grade C fistulae
Grade C fistulae occurred in 14 (5.8%) in the PG and 13 (7.0%) patients in the PJ group (5.8 vs. 7.0%, P = 0.690). The median (range) hospital stay was 35.5 days (8–141) in the PG compared with 28 days (10–77, P = 0.458) in the PJ group. Twelve in the PG and nine patients (P = 0.385) in the PJ group had signs of infection. The details of investigations, re-operations and other complications are summarized in Table 5.
Table 5.
The details of further investigations and findings, and treatment of patients with a grade C fistula
| Variables | PG, n = 14 | PJ, n = 13 | P-value |
|---|---|---|---|
| Investigations (CT), n | 10 | 10 | 1.000 |
| CT findings | |||
| Free fluid | 0 | 1 | |
| IABC | 8 | 5 | |
| Anastomotic leak | 0 | 1 | |
| No collection | 2 | 3 | |
| Sepsis | 12 | 9 | 0.385 |
| Treatment | |||
| Radiological drainage | 7 | 4 | |
| Surgery | 5 | 6 | 0.285 |
| Total pancreatectomy | 2 | 1 | |
| DJ repair | 2 | 2 | |
| PPH control | 1 | 3 | |
| TPN | 7 | 8 | |
| Other complications | |||
| Enteric leak | 3 | 3 | 1.000 |
| Bile leak | 2 | 1 | 1.000 |
| Chest infection | 4 | 2 | 0.648 |
| Wound infection | 3 | 0 | 0.222 |
| PPH | 4 | 9 | 0.057 |
| Source of bleed | |||
| Cut surface of pancreas | 2 | 5 | |
| Unknown source | 1 | 1 | |
| DJ anastomosis | 1 | 0 | |
| Pseudoaneurysms | 0 | 3 | |
| 90-day mortality | 9 | 10 | 0.677 |
| Hospital stay (median and range) in days | 35.5 (8–141) | 28 (10–77) | 0.458 |
CT, computer tomogram; IABC, intra abdominal collection; DJ, duodenojejunal; PPH, post pancreatectomy haemorrhage; TPN, total parenteral nutrition
Overall morbidity and mortality between the PG and PJ groups
The details of other complications between the PG and PJ groups are summarized in Table 6. There was no significant difference in the rate of PPH, delayed gastric emptying (DGE), enteric leak, bile leak, chest and wound infection, renal failure, respiratory failure, arrhythmias, sepsis and IABC between both groups. The overall complication rate (including the patients who had PF) was higher in the PG (108, 45.2%) compared with the PJ group (62, 33.5%, P = 0.017). However, after excluding patients with PF that was not associated any other complications, there was no significant difference in the morbidity between both groups (88, 36.8% in PG vs. 58, 31.4% in PJ, P = 0.258). A total of 18 patients required a re-laparotomy, and re-laparotomy rates were similar between both groups: 4.2% (n = 10) in the PG vs. 4.3% (n = 8) (P = 1.000) in the PJ group. There was no significant difference in the 90-day mortality between both groups (17, 7.1% vs. 16, 8.6%, P = 0.558). However, the 90-day mortality after PPH was significantly higher in the PJ group (71.4%, 10 out of 14) compared with the PG group (42.8%, 6 out of 14 patients, P = 0.001).
Table 6.
The details of other complications and overall morbidity and mortality between the PG and PJ groups
| Variables | PG, n = 239 (%) | PJ, n = 185 (%) | P-value |
|---|---|---|---|
| Other complications | |||
| PPH | 14 (5.9) | 14 (7.6) | 0.556 |
| Chest infection | 26 (11) | 18 (9.7) | 0.750 |
| Respiratory failure | 21(8.8) | 9 (5) | 0.130 |
| Wound infection | 17 (7) | 11 (6) | 0.696 |
| Enteric leak | 12 (5) | 5 (2.7) | 0.319 |
| Bile leak | 5 (2.3) | 3 (1.6) | 1.000 |
| Renal failure | 8 (3.3) | 10 (5.4) | 0.573 |
| Delayed gastric emptying | 2 (0.8) | 5 (2.7) | 0.248 |
| AF | 12 (5) | 9 (5) | 1.000 |
| Re-laparotomy | 10 (4.2) | 8 (4.3) | 1.000 |
| Others | 6 (0.2) | 5 (2.7) | 0.685 |
| Overall morbidity, n | 108 (45.2) | 62 (33.5) | 0.017 |
| Overall morbidity after excluding patients with PF alone | 88 (36.8) | 58 (31.4) | 0.258 |
| 90-day mortality | 17 (7.1) | 16 (8.6) | 0.558 |
PPH, post pancreatectomy haemorrhage; AF, atrial fibrillation; PF, pancreatic fistula
Discussion
The main cause of post-operative morbidity and mortality after PD is a pancreatic fistula. Before to July 2005, there was no universally accepted definition of PF and the diagnosis of PF was mainly based on the presence of amylase-rich drain fluid (three times the normal serum activity). However, not all biochemical fistulae were clinically significant and defining the presence or absence of a fistula based on the DFA and drain volume can alone underestimate a clinically significant fistula. In July 2005, the International Study Group for Pancreatic Fistula (ISGPF) developed a universal definition of PF and classified PF into three grades based on clinical parameters.21 After the publication of the ISGPF classification, Pratt et al. studied the incidence of PF in 176 consecutive patients who underwent PD and observed an incidence of 30%, much higher than many of the previously published studies.24 The majority of patients developed grade A (15%) and grade B (12%) fistulae. Grade C fistulae only occurred in 3% of cases. However, when they re-classified PF based on five previously used PF definitions, they noted wide variations in the incidence of PF (9% to 19%).24 Fuks et al. studied the incidence of PF in 680 patients from five centres in France and observed an overall PF rate of 16.3%. The grade C fistula was observed in 5.6% of patients.25
As a part of the meta-analysis, Wente et al. performed a pooled analysis of 12 observation studies (published before the introduction of ISGPF classification), and found a PF rate of 3.7% in the PG group compared with 16.5% (P = 0.001) in the PJ group. There was no significant difference in the overall hospital stay between both groups.26 The overall mortality rate was 3.0% in the PG group compared with 9.1% (P = 0.001) in the PJ group. As part of the same study, Wente et al. also performed a pooled analysis of three randomized controlled trials (published before the introduction of ISGPF classification), and found no difference in the incidence of PF (13.9% vs. 15.8%, P = 0.54), DGE (12.7 vs. 16.9%, P = 0.54) and bile leak (3.1 vs. 5.4%, P = 0.53) between the PG and PJ groups. The length of stay was shorter in the PG group, after excluding the results from the French series.26
In the current series, the ISGPF classification scheme was applied retrospectively to study the rate and the severity of PF in patients who underwent PG and PJ reconstruction methods after PD. This is the first major study that applied the ISGPF criteria to compare the PF rate between two of most commonly used reconstruction techniques. The overall fistula rate of 20% was similar to many published series.1,3,4,25,26 The grade C PF rate of 6.4% in the current series was also similar to many published series.24,25 The overall PF rate was higher in the PG group because of a higher incidence of grade A fistulae. However, more patients in the PJ group had clinically significant (not statistically significant) pancreatic fistulae, supporting the findings of previously published observational studies.26 There was no difference in the overall PF rate and grade of PF between two sub-types of PJ reconstruction. The higher rate of a grade A PF in the PG group suggest that although the risk of a biochemical fistula is higher in PG the clinical consequences are less severe compared with the pancreatic leak after PJ reconstruction. This could be because of several reasons: inhibition of pancreatic enzyme activation by acid environment, an absence of enterokinase in the stomach (necessary for activation of trypsinogen),27 a thick-walled stomach with an excellent blood supply,17 prevention of marginal ulceration by neutralization of acid by pancreatic juices, decompression of the stomach by the nasal gastric tube preventing stasis of gastric and pancreatic contents, and the absence of the long jejunal loop with the accumulation of biliary and pancreatic contents.17
Similar to the previously published studies,26 the hospital stay was shorter in the PG group compared with the PJ group. The hospital stay was significantly longer in the PF group compared with the non-fistula group. However, subgroup analysis showed that the length of hospital stay was shorter in the PJ compared with the PG group in patients with a grade C PF. This could be as a result of the fact that more patients in the PJ group with a grade C PF had a higher early hospital mortality. The overall morbidity was significantly higher in the PG group when all post-operative complications including the PF were included. However, when patients with a grade A PF that were not associated with any other complications were excluded the morbidity was similar between both groups.
The overall mortality in the current series was much higher compared with the randomized Italian4 and John Hopkin's series3,4,28 and similar to the French1 and many observational studies.13–18,29 The 90-day mortality rate was zero among patients with grade A & B compared with 70% (19 out of 27) in patients with a grade C leak. There was no significant difference between the PG and PJ groups and among the two different sub-types of PJ reconstruction supporting the notion that it is the experience and familiarity of the surgeon with one particular pancreatic reconstruction technique that is the most important factor in reducing the PF rate and not the type of reconstruction.30–32
The high mortality among patients with a grade C PF is as a result of a high incidence of associated intra-abdominal abscesses, bleeding and sepsis. In a large retrospective series by Fuks et al., 35 out of 36 patients with a grade C PF required re-operation for either a pancreatic leak or intra-abdominal haemorrhage.25 In the current series, the 6.6% prevalence of a PPH was similar to 5–12% prevalence reported in the literature.33–36 In a large retrospective series, Yekebas et al. found 5.7% prevalence of PPH and in 39% of patients a PPH was associated with PF. They also observed that the PPH–related mortality was closely associated with the occurrence of PF, vascular pathologies, and soft pancreas texture.33 In the present study, PPH was associated with a grade C PF in 46% of patients but none of the patients with a grade A and B PF had PPH. Furthermore, the prevalence of PPH was significantly higher among patients with a grade C PF after PJ reconstruction. The 90-day mortality in patients with a PPH associated with a PF was 77% compared with 40% in patients with a PPH but with no PF (P = 0.001). The higher prevalence of PPH and associated mortality among grade C patients in the PJ group could be because of erosion of the major blood vessels and the surrounding tissue by the activated pancreatic enzymes and high volume pancreatic juice. This is less of a problem in patients with PG reconstruction as the stomach is decompressed by the nasogastric tube and trypsinogen is not activated by the gastric juice. Another reason for the high mortality among patients with a grade C PF in this series could be because of inclusion of late mortalities that occurred after patients were discharged from the hospital. Our approach for management of grade C fistulae includes urgent laparotomy for patients who have an acute abdomen or are unstable from severe bleeding. For stable patients our approach is one of radiological drainage for collections escalating to surgical intervention if there is no improvement. For bleeding an attempt at radiological stenting or embolization is tried and if this fails a laparotomy is carried out. Overall, 11 (40.3%) out of 27 patients with a grade C leak underwent a re-laparotomy but only three had a completion pancreatectomy. The rest had an attempt to control the leak or haemorrhage. The majority with a grade C leak were managed radiological or medically but did not reach a laparotomy – either they were judged to be responding or were too ill for a major intervention. As the mortality of our group of grade C leak patients was higher than some published series it could be postulated that a more aggressive approach with more patients having earlier surgical intervention, and more use of completion pancreatectomy in the surgical group, could have improved the outcome.
The major criticism of the present study would be the retrospective nature of the study. It has all the disadvantages that are associated with any retrospective series. However, patients in both groups were well matched for every parameter except the gender ratio. Furthermore, this should not influence either the incidence of PF or any other complications between both groups. There were no data on the consistency of the pancreas and the pancreatic ductal diameter but there was no reason to believe that there were more patients in either group with a soft pancreas as the number of patients who underwent pre-operative biliary drainage, indications and the diagnosis was similar between both groups.
Conclusions
In the current series, the ISGPF classification system has been successfully applied to a large group of patients who underwent a PD followed by either PG or PJ reconstruction. The grade PF rate was significantly higher in the PG group compared with the PJ group. There was no significant difference in the overall PF rate, 90-day mortality and length of hospital stay between both methods supporting the notion that the experience and familiarity of the surgeon with one particular reconstruction technique is perhaps more important than the type of reconstruction itself.
Acknowledgments
The authors would like to thank Mr Chris Coldham, the Liver Unit Research Manager and David Bodonisky and Jason Foster (two medical students) for helping with the data collection.
Conflicts of interest
None.
Reference
- 1.Duffas JP, Suc B, Msika S, Fourtanier G, Muscari F, Hay JM, et al. A controlled randomized multicenter trial of pancreatogastrostomy or pancreatojejunostomy after pancreatoduodenectomy. Am J Surg. 2005;189:720–729. doi: 10.1016/j.amjsurg.2005.03.015. [DOI] [PubMed] [Google Scholar]
- 2.Buchler MW, Friess H, Wagner M, Kulli C, Wagener V, Z'Graggen K. Pancreatic fistula after pancreatic head resection. Br J Surg. 2000;87:883–889. doi: 10.1046/j.1365-2168.2000.01465.x. [DOI] [PubMed] [Google Scholar]
- 3.Yeo CJ, Cameron JL, Maher MM, Sauter PK, Zahurak ML, Talamini MA, et al. A prospective randomized trial of pancreaticogastrostomy versus pancreaticojejunostomy after pancreaticoduodenectomy. Ann Surg. 1995;222:580–588. doi: 10.1097/00000658-199510000-00014. Discussion 8-92. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 4.Bassi C, Falconi M, Molinari E, Salvia R, Butturini G, Sartori N, et al. Reconstruction by pancreaticojejunostomy versus pancreaticogastrostomy following pancreatectomy: results of a comparative study. Ann Surg. 2005;242:767–771. doi: 10.1097/01.sla.0000189124.47589.6d. discussion 71-3. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 5.Lin JW, Cameron JL, Yeo CJ, Riall TS, Lillemoe KD. Risk factors and outcomes in postpancreaticoduodenectomy pancreaticocutaneous fistula. J Gastrointest Surg. 2004;8:951–959. doi: 10.1016/j.gassur.2004.09.044. [DOI] [PubMed] [Google Scholar]
- 6.Poon RT, Lo SH, Fong D, Fan ST, Wong J. Prevention of pancreatic anastomotic leakage after pancreaticoduodenectomy. Am J Surg. 2002;183:42–52. doi: 10.1016/s0002-9610(01)00829-7. [DOI] [PubMed] [Google Scholar]
- 7.Bassi C, Falconi M, Salvia R, Mascetta G, Molinari E, Pederzoli P. Management of complications after pancreaticoduodenectomy in a high volume centre: results on 150 consecutive patients. Dig Surg. 2001;18:453–457. doi: 10.1159/000050193. [DOI] [PubMed] [Google Scholar]
- 8.Balcom JHT, Rattner DW, Warshaw AL, Chang Y, Fernandez-del Castillo C. Ten-year experience with 733 pancreatic resections: changing indications, older patients, and decreasing length of hospitalization. Arch Surg. 2001;136:391–398. doi: 10.1001/archsurg.136.4.391. [DOI] [PubMed] [Google Scholar]
- 9.Roder JD, Stein HJ, Bottcher KA, Busch R, Heidecke CD, Siewert JR. Stented versus nonstented pancreaticojejunostomy after pancreatoduodenectomy: a prospective study. Ann Surg. 1999;229:41–48. doi: 10.1097/00000658-199901000-00005. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 10.Miyagawa S, Makuuchi M, Lygidakis NJ, Noguchi T, Nishimaki K, Hashikura Y, et al. A retrospective comparative study of reconstructive methods following pancreaticoduodenectomy–pancreaticojejunostomy vs. pancreaticogastrostomy. Hepatogastroenterology. 1992;39:381–384. [PubMed] [Google Scholar]
- 11.Morris DM, Ford RS. Pancreaticogastrostomy: preferred reconstruction for Whipple resection. J Surg Res. 1993;54:122–125. doi: 10.1006/jsre.1993.1018. [DOI] [PubMed] [Google Scholar]
- 12.Mason GR, Freeark RJ. Current experience with pancreatogastrostomy. Am J Surg. 1995;169:217–219. doi: 10.1016/S0002-9610(99)80140-8. [DOI] [PubMed] [Google Scholar]
- 13.Kim SW, Youk EG, Park YH. Comparison of pancreatogastrostomy and pancreatojejunostomy after pancreatoduodenectomy performed by one surgeon. World J Surg. 1997;21:640–643. doi: 10.1007/s002689900286. [DOI] [PubMed] [Google Scholar]
- 14.Takano S, Ito Y, Watanabe Y, Yokoyama T, Kubota N, Iwai S. Pancreaticojejunostomy versus pancreaticogastrostomy in reconstruction following pancreaticoduodenectomy. Br J Surg. 2000;87:423–427. doi: 10.1046/j.1365-2168.2000.01395.x. [DOI] [PubMed] [Google Scholar]
- 15.Fujino Y, Suzuki Y, Ajiki T, Tanioka Y, Ku Y, Kuroda Y. Risk factors influencing pancreatic leakage and the mortality after pancreaticoduodenectomy in a medium-volume hospital. Hepatogastroenterology. 2002;49:1124–1129. [PubMed] [Google Scholar]
- 16.Schlitt HJ, Schmidt U, Simunec D, Jager M, Aselmann H, Neipp M, et al. Morbidity and mortality associated with pancreatogastrostomy and pancreatojejunostomy following partial pancreatoduodenectomy. Br J Surg. 2002;89:1245–1251. doi: 10.1046/j.1365-2168.2002.02202.x. [DOI] [PubMed] [Google Scholar]
- 17.Aranha GV, Hodul P, Golts E, Oh D, Pickleman J, Creech S. A comparison of pancreaticogastrostomy and pancreaticojejunostomy following pancreaticoduodenectomy. J Gastrointest Surg. 2003;7:672–682. doi: 10.1016/s1091-255x(02)00432-8. [DOI] [PubMed] [Google Scholar]
- 18.Hoshal VL, Jr, Benedict MB, David LR, Kulick J. Personal experience with the Whipple operation: outcomes and lessons learned. Am Surg. 2004;70:121–125. Discussion 6. [PubMed] [Google Scholar]
- 19.Ohwada S, Ogawa T, Kawate S, Tanahashi Y, Iwazaki S, Tomizawa N, et al. Results of duct-to-mucosa pancreaticojejunostomy for pancreaticoduodenectomy Billroth I type reconstruction in 100 consecutive patients. J Am Coll Surg. 2001;193:29–35. doi: 10.1016/s1072-7515(01)00869-9. [DOI] [PubMed] [Google Scholar]
- 20.Arnaud JP, Tuech JJ, Cervi C, Bergamaschi R. Pancreaticogastrostomy compared with pancreaticojejunostomy after pancreaticoduodenectomy. Eur J Surg. 1999;165:357–362. doi: 10.1080/110241599750006901. [DOI] [PubMed] [Google Scholar]
- 21.Bassi C, Dervenis C, Butturini G, Fingerhut A, Yeo C, Izbicki J, et al. Postoperative pancreatic fistula: an international study group (ISGPF) definition. Surgery. 2005;138:8–13. doi: 10.1016/j.surg.2005.05.001. [DOI] [PubMed] [Google Scholar]
- 22.Z'Graggen K, Uhl W, Friess H, Buchler MW. How to do a safe pancreatic anastomosis. J Hepatobiliary Pancreat Surg. 2002;9:733–737. doi: 10.1007/s005340200101. [DOI] [PubMed] [Google Scholar]
- 23.Dmitrewski J, Buckels JA. A simple modification of choledochoenterostomy using continuous pull through suture. Surg Gynecol Obstet. 1992;175:573–574. [PubMed] [Google Scholar]
- 24.Pratt WB, Maithel SK, Vanounou T, Huang ZS, Callery MP, Vollmer CM., Jr Clinical and economic validation of the International Study Group of Pancreatic Fistula (ISGPF) classification scheme. Ann Surg. 2007;245:443–451. doi: 10.1097/01.sla.0000251708.70219.d2. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 25.Fuks D, Piessen G, Huet E, Tavernier M, Zerbib P, Michot F, et al. Life-threatening postoperative pancreatic fistula (grade C) after pancreaticoduodenectomy: incidence, prognosis, and risk factors. Am J Surg. 2009;197:702–709. doi: 10.1016/j.amjsurg.2008.03.004. [DOI] [PubMed] [Google Scholar]
- 26.Wente MN, Shrikhande SV, Muller MW, Diener MK, Seiler CM, Friess H, et al. Pancreaticojejunostomy versus pancreaticogastrostomy: systematic review and meta-analysis. Am J Surg. 2007;193:171–183. doi: 10.1016/j.amjsurg.2006.10.010. [DOI] [PubMed] [Google Scholar]
- 27.Pikarsky AJ, Muggia-Sullam M, Eid A, Lyass S, Bloom AI, Durst AL, et al. Pancreaticogastrostomy after pancreatoduodenectomy. A retrospective study of 28 patients. Arch Surg. 1997;132:296–299. doi: 10.1001/archsurg.1997.01430270082016. [DOI] [PubMed] [Google Scholar]
- 28.Yeo CJ, Cameron JL, Lillemoe KD, Sitzmann JV, Hruban RH, Goodman SN, et al. Pancreaticoduodenectomy for cancer of the head of the pancreas. 201 patients. Ann Surg. 1995;221:721–731. doi: 10.1097/00000658-199506000-00011. discussion 31-3. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 29.Oussoultzoglou E, Bachellier P, Bigourdan JM, Weber JC, Nakano H, Jaeck D. Pancreaticogastrostomy decreased relaparotomy caused by pancreatic fistula after pancreaticoduodenectomy compared with pancreaticojejunostomy. Arch Surg. 2004;139:327–335. doi: 10.1001/archsurg.139.3.327. [DOI] [PubMed] [Google Scholar]
- 30.Bassi C, Falconi M, Molinari E, Mantovani W, Butturini G, Gumbs AA, et al. Duct-to-mucosa versus end-to-side pancreaticojejunostomy reconstruction after pancreaticoduodenectomy: results of a prospective randomized trial. Surgery. 2003;134:766–771. doi: 10.1016/s0039-6060(03)00345-3. [DOI] [PubMed] [Google Scholar]
- 31.Sakorafas GH, Friess H, Balsiger BM, Buchler MW, Sarr MG. Problems of reconstruction during pancreatoduodenectomy. Dig Surg. 2001;18:363–369. doi: 10.1159/000050174. [DOI] [PubMed] [Google Scholar]
- 32.Warshaw AL, Swanson RS. Pancreatic cancer in 1988. Possibilities and probabilities. Ann Surg. 1988;208:541–553. doi: 10.1097/00000658-198811000-00001. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 33.Yekebas EF, Wolfram L, Cataldegirmen G, Habermann CR, Bogoevski D, Koenig AM, et al. Postpancreatectomy hemorrhage: diagnosis and treatment: an analysis in 1669 consecutive pancreatic resections. Ann Surg. 2007;246:269–280. doi: 10.1097/01.sla.0000262953.77735.db. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 34.Sato N, Yamaguchi K, Shimizu S, Morisaki T, Yokohata K, Chijiiwa K, et al. Coil embolization of bleeding visceral pseudoaneurysms following pancreatectomy: the importance of early angiography. Arch Surg. 1998;133:1099–1102. doi: 10.1001/archsurg.133.10.1099. [DOI] [PubMed] [Google Scholar]
- 35.Balladur P, Christophe M, Tiret E, Parc R. Bleeding of the pancreatic stump following pancreatoduodenectomy for cancer. Hepatogastroenterology. 1996;43:268–270. [PubMed] [Google Scholar]
- 36.Halloran CM, Ghaneh P, Bosonnet L, Hartley MN, Sutton R, Neoptolemos JP. Complications of pancreatic cancer resection. Dig Surg. 2002;19:138–146. doi: 10.1159/000052029. [DOI] [PubMed] [Google Scholar]