Search

Liver resection (LR) is the main modality of treatment for hepatocellular carcinoma (HCC) and colorectal liver metastasis (CRLM). Post-hepatectomy liver failure (PHLF) remains the most dreaded complication. We aim to create a prognostic score for early risk stratification of patients undergoing LR. Clinical and operative data of 472 patients between 2000 and 2016 with HCC or CRLM undergoing major hepatectomy were extracted and analysed from a prospectively maintained database. PHLF was defined using the 50–50 criteria. Liver cirrhosis and fatty liver were histologically confirmed in 35.6% and 53% of patients. 4.7% (n = 22) of patients had PHLF. A 90-day mortality was 5.1% (n = 24). Pre-operative albumin–bilirubin score (p = 0.0385), prothrombin time (p 7 and International Study Group for Liver Surgery criteria, intensive care unit admissions, length of stay, all complications, major complications, re-admissions and mortality (p

Follicular dendritic cell sarcoma (FDCS) of the liver is an extremely rare disease, accounting for a mere 0.4% of all soft tissue sarcomas. FDCS most commonly involves lymph nodes but also affects extranodal sites such as the gastrointestinal system, oral cavity, liver, spleen and pancreas, albeit less commonly. It is widely considered a low-to-intermediate grade malignancy. We report a case of FDCS with metachronous involvement of the liver, small intestines and spleen, its imaging, histological findings and its management.

Backgrounds/Aims To determine the prevalence of post-hepatectomy liver failure/insufficiency (PHLF/I) in patients undergoing extensive hepatic resections for hepatocellular carcinoma (HCC) and to assess the predictive value of preoperative factors for post-hepatectomy liver failure or insufficiency (PHLF/I). Methods A retrospective review of patients who underwent liver resections for HCC between 2001 and 2013 was conducted. Preoperative parameters were assessed and analyzed for their predictive value of PHLF/I. Definitions used included the 50–50, International Study Group of Liver Surgery (ISGLS) and Memorial Sloan Kettering Cancer Centre (MSKCC) criteria. Results Among the 848 patients who underwent liver resections for HCC between 2001 and 2013, 157 underwent right hepatectomy (RH) and extended right hepatectomy (ERH). The prevalence of PHLF/I was 7%, 41% and 28% based on the 50–50, ISGLS and MSKCC criteria, respectively. There were no significant differences in PHLF/I between RH and ERH. Model for End-Stage Liver Disease (MELD) score and bilirubin were the strongest independent predictors of PHLF/I based on the 50–50 and ISGLS/MSKCC criteria, respectively. Predictive models were developed for each of the criteria with multiple logistic regression. Conclusions MELD score, bilirubin, alpha-fetoprotein and platelet count showed significant predictive value for PHLF/I (all p

Background Currently, the management strategy of pancreatic cyst (PC) remains controversial because of the inability to diagnose this type of cyst accurately and the limited knowledge of its natural history. Previously, many clinicians have advocated an aggressive resectional policy. This approach is no longer appropriate, and the number of PCs detected incidentally has increased. This study reviews the present literature and attempts to provide a management algorithm of pancreatic cysts based on currently available evidence. Methods A Medline search was conducted to identify studies investigating PC, with particular emphasis placed on studies addressing its diagnosis and management. Additional articles were obtained from the reference lists of key articles and recent reviews. Comments Based on current evidence, the optimal management of PC remains an art and should be individualized based on the risk–benefit ratio of surgery, which is influenced by multiple factors, such as the patient’s potential life expectancy, surgical risk; and malignant potential of the cyst. Our proposed management algorithm is based on an individual’s predicted risk–benefit ratio of surgery. Prospective evaluation of the algorithm is needed to determine its integrity.

Aim Pancreatic endocrine neoplasms (PENs) may occasionally manifest as cystic lesions of the pancreas. The aim of this study is to report our experience with cystic PENs and to compare their clinico-pathological features with their solid counterparts. Materials and methods From 1990 to 2004, 38 patients with PENs were reviewed. Six of these tumours appeared on radiological imaging as a cystic lesion of the pancreas. Results Of the 38 patients with a PEN, 21 of the patients were female and with a median age of 54.5 (range, 33–83) years. Sixteen patients had functional endocrine tumours of which insulinoma was the most common. The six patients with cystic PEN had a median age of 55.5 (range, 41–70) years and half were female. Cystic PENs were significantly larger [48 (range, 25–170) mm vs 19 (range, 3–120) mm, P =0.013] and were less likely to be benign (0 vs 50%, P =0.017) compared to their solid counterparts. There was no difference between cystic and solid PENs in terms of age, sex, presence of symptoms, proportion of functioning tumours and location of tumours within the pancreas. Conclusion Cystic PENs share many clinico-pathological features with solid PENs. These differ only in the cystic appearance and tend to be of a larger size. Hence, these findings suggest that cystic and solid PENs are unlikely to be distinct pathological entities but are likely to be morphological variants of the same entity.

Aim To compare the clinico-pathological features of intraductal papillary mucinous cystic tumours (IPMT) and mucinous cystic tumours (MCT) of the pancreas. Methods Eighteen patients with IPMT and 18 with MCT who underwent surgical resection between 1990 and 2004 were retrospectively reviewed. Their clinico-pathological features were compared using univariate analysis. Statistical analyses of potential predictive factors of malignancy for each of these two groups were also conducted. Results Patients with IPMT were found to be older (64±10 vs 43±18 years, p

Laparoscopic resection of gastric gastrointestinal stromal tumors (GISTs) has been shown by several retrospective studies to be technically feasible and associated with favorable outcomes when compared to the open approach. This study aims to mitigate potential selection bias by performing a case control study of laparoscopic (LWR) versus open wedge resection (OWR) matched by resection type, location and tumor size.We retrospectively identified 50 consecutive patients who underwent LWR for a suspected gastric GIST from a prospective database and matched this cohort with 50 patients who underwent OWR.There was no statistical difference between the key baseline clinicopathological features of patients' who underwent LWR versus OWR. Patients who underwent LWR had longer operating times [150 (range, 65-270) minutes vs 92.5 (25-200) minutes, P.001] but decreased median blood loss [0 (0-300) ml vs 0 (0-1200) ml, P = .015], decreased frequency of intraoperative or postoperative blood transfusion [1 (2%) vs 8 (16%), P = .031], decreased median time to liquid diet [2 (0-5) vs 3 (1-7) days, P.001], decreased median time to solid diet [3 (1-6) vs 5 (2-11) days, P.001] and decreased postoperative stay [4 (2-10) vs 4.5 (3-17), P.001] compared to OWR. There was no difference in oncological outcomes such as frequency of close margins (≤ 1 mm) and recurrence-free survival.This matched case-control study provides supporting evidence that LWR results in superior perioperative outcomes compared to OWR without compromising on oncological outcomes.

Gastrointestinal arterio-venous malformations are a known cause of gastrointestinal bleeding. We present a rare case of persistent rectal bleeding due to a rectal arterio-portal venous fistula in the setting of portal hypertension secondary to portal vein thrombosis. The portal hypertension was initially surgically treated with splenectomy and a proximal splenorenal shunt. However, rectal bleeding persisted even after surgery, presenting us with a diagnostic dilemma. The patient was re-evaluated with a computed tomography mesenteric angiogram which revealed a rectal arterio-portal fistula. Arterio-portal fistulas are a known but rare cause of portal hypertension, and possibly the underlying cause of continued rectal bleeding in this case. This was successfully treated using angiographic localization and super-selective embolization of the rectal arterio-portal venous fistula via the right internal iliac artery.The patient subsequently went on to have a full term pregnancy. Through this case report, we hope to highlight awareness of this unusual condition, discuss the diagnostic workup and our management approach.

Background This study was a retrospective review of a series of patients with sigmoid volvulus to identify risk factors for recurrence and recommend appropriate treatment. Methods Thirty-five patients with sigmoid volvulus were treated over 8 years. Results Six patients had emergency surgery for peritonitis. Twenty-eight of the other 29 patients had successful endoscopic decompression; 15 of these patients had elective surgery during the same admission. Twelve of the 14 patients who refused operation after endoscopic decompression developed recurrent volvulus, a median of 2·8 months later. Eight subsequently agreed to surgery and underwent elective operation following repeat decompression. Of 29 patients who had surgery, 27 had sigmoid colectomy (two were initial Hartmann procedures) and two had subtotal colectomy. Six patients who had sigmoid colectomy developed recurrent volvulus. Concomitant megacolon and megarectum at the time of initial surgery were significant predictors of recurrence. Conclusion Subtotal colectomy, carried out as the primary procedure if there is concomitant megacolon or megarectum, might reduce the risk of recurrent sigmoid volvulus.

Objective To critically analyze a large single-institution experience with distal pancreatectomy (DP), with particular attention to the risk factors, outcome, and management of the postoperative pancreatic fistula (PF). Design Retrospective study. Setting Tertiary referral center. Patients A total of 232 consecutive patients with pancreatic or extrapancreatic disease necessitating DP over 21 years. Interventions Twenty-one patients underwent spleen-preserving DP, 117 underwent DP with splenectomy, and 94 underwent DP with multiorgan resection. Main Outcome Measures The perioperative and postoperative data of patients who underwent DP were analyzed. This included factors associated with postoperative morbidity with particular attention to the PF (defined by the International Study Group of Pancreatic Fistula) and changing trends in operative and perioperative data during the study period. Results The overall operative morbidity and mortality were 47% (107 patients) and 3% (7 patients), respectively. During the study period, the rates of resection increased from 3 cases to 23 per year, and increasingly these were performed for smaller and incidental lesions. The morbidity rate remained unchanged, but there was a decline in postoperative stay and the need for care in the intensive care unit. Pancreatic fistulas occurred in 72 patients (31%); 41 (18%) were grade A, 13 (6%) grade B, and 18 (8%) grade C. Increased weight, higher American Society of Anesthesiologists score, blood loss greater than 1 L, increased operation time, decreased albumin level, and sutured closure of the stump without main duct ligation were associated with a postoperative PF on univariate analysis. A DP with splenectomy was associated with a higher incidence of grade B or C PF and non–PF-related complications. Ninety-two percent of PFs were successfully managed nonoperatively. Clinical outcomes correlated well with PF grading, as evidenced by the progressive increase in outcome measures such as postoperative stay, readmissions, reoperations, radiologic interventions, and non–PF-related complications from grade A to C PFs. Conclusions Pancreatic fistula is the most common complication after DP and its incidence varies depending on the definition applied. Several risk factors for developing a PF were identified. Splenic preservation after DP is safe. The grade of a PF correlates well with clinical outcomes, and most PFs may be managed nonoperatively.

A 79-year-old Indian male was admitted with upper abdominal discomfort of 1-year duration which was associated with loss of weight and appetite. Serum tumor markers, including carcinoembryonic antigen and carbohydrate antigen 19-9, were within normal limits. A computed tomographic scan demonstrated a cystically dilated and tortuous pancreatic duct measuring 1.9 cm, suggestive of an intraductal papillary mucinous neoplasm. Fusion positron emission tomography/computed tomography with 2-deoxy-2-[18F] fluoro-D-glucose positron emission tomography (FDG-PET/CT) which was subsequently performed confirmed a metabolically active focus within the pancreatic head mass with a standard uptake value (SUVmax) of 3.5 compatible with carcinoma. A total pancreatectomy was performed and the final histology demonstrated a main-duct type intraductal papillary mucinous neoplasm with a focus of high-grade dysplasia compatible with carcinoma-in-situ. These images illustrate the emerging utility of FDG-PET/CT in the preoperative detection of malignancy in intraductal papillary mucinous neoplasm.

Background This study aims to determine the use of preoperative clinical, biochemical, and cross-sectional imaging features for predicting malignancy in cystic lesions of the pancreas (CLP). Study Design Two hundred twenty patients who underwent operations for CLP or suspected CLP were reviewed. Patients were divided into two groups, patients undergoing operations for pseudocysts and patients undergoing operations for suspected cystic neoplasms. The predictive effect of various preoperative factors on the malignant potential of CLP was evaluated. Results Forty-four patients with a preoperative diagnosis of pseudocysts underwent operations for complications of pseudocyst. Forty-two were confirmed pathologically to have pseudocysts, but two were found, unexpectedly, to harbor malignant lesions. One hundred seventy-six patients underwent operations for suspected pancreatic cystic neoplasms. There were 70 benign, 51 potentially malignant, and 55 malignant CLP. On multivariate analysis, three factors, ie, elevated serum carcinoembryonic antigen (CEA) or carbohydrate antigen 19-9; cyst size > 3 cm; and presence of one or more of three morphologic features, such as solid component; peripheral calcification; and main duct dilation on cross-sectional imaging were independent predictors of malignancy. Presence of two or three of these factors had a positive predictive value of 88% in predicting a premalignant or malignant CLP. Conclusions Most pancreatic pseudocysts can be accurately diagnosed preoperatively. In patients with suspected pancreatic cystic neoplasms, elevated serum CEA or carbohydrate antigen 19-9, cyst size > 3 cm, and presence of suspicious morphologic features on imaging are predictors of potentially malignant or malignant CLP. Patients with a high likelihood of a potentially malignant or malignant lesion based on these three factors should undergo operation without additional investigations.

Objectives: To determine the outcome of patients undergoing distal pancreatectomy for pancreatic adenocarcinoma. Methods: A retrospective review of 39 patients undergoing distal pancreatectomy for adenocarcinoma. Results: Thirty patients underwent surgery for ductal adenocarcinoma, 5 for malignant intraductal papillary mucinous neoplasm and 4 for mucinous cystadenocarcinoma. Malignant cystic neoplasms were significantly less likely to demonstrate perineural invasion, more likely to be well-differentiated, of lower T stage and of lower AJCC staging compared to ductal adenocarcinoma. These had a longer median disease-specific survival (42 (3–144) vs. 15 (14–16) months, p = 0.002). Eight patients underwent extended resections. These were associated with longer operating times compared to standard resections but there was no difference in surgical morbidity or mortality, blood transfusions, length of hospitalization or long-term survival. Univariate analysis demonstrated that R2 resection, size >30 mm, lymph node involvement, need for perioperative blood transfusion, serum albumin Conclusions: Malignant cystic neoplasms have less aggressive behavior and more favorable outcome compared to ductal adenocarcinoma. R2 resection, larger tumor size, lymph node involvement, perioperative transfusion, decreased serum albumin and low platelet count are factors associated with decreased survival in patients with ductal adenocarcinoma undergoing distal pancreatectomy.

The preoperative distinction between serous cystic neoplasms (SCNs) and mucinous cystic neoplasms (MCNs) is essential, as all MCNs are considered malignant or potentially malignant and should be surgically resected, whereas SCNs are almost always benign. However, the radiologic distinction between SCNs and MCNs is frequently difficult especially with serous oligocystic adenoma (SOA), a morphologic variant of SCN, as both SOA and MCN appear on cross-sectional imaging as a solitary macrocystic lesion in the pancreas. We reviewed all SOAs managed at our institution to determine if any clinicopathologic features would prove useful for establishing a preoperative diagnosis.Over a 15-year period, 64 patients with a pathologically confirmed diagnosis of a pancreatic cystadenoma or cystadenocarcinoma treated at Singapore General Hospital were retrospectively reviewed. There were 27 MCNs and 37 SCNs including 12 SOAs. In addition, 40 cases of SOA previously reported in the literature were reviewed and analyzed together with the 12 patients, making this a series of 52 SOAs.In our experience, SOAs comprised 32.4% of the SCNs, and females predominated (7/12). The median age of the patients was 42.5 years (range 22-74 years), and only 4 of the 12 patients were symptomatic. Most of the cysts were located in the body or tail of the pancreas (9/12), and the median cyst size was 52.5 mm (range 10-190 mm). When the clinicopathologic features of SOAs and serous microcystic adenomas (SMAs) were compared, there was no difference between the patients with SOAs and SMAs in terms of age, sex, presence of symptoms, cyst size, or site of the lesion. However, SOAs occurred in the women less frequently (67.3% vs. 96.3%, P=0.004), were smaller [40 mm (range 10-190 mm) vs. 95 mm (range 25-180 mm), P0.001], and occurred more commonly in the head of the pancreas [25 (48.1%) vs. 2(7.4%)] compared to MCNs. None of the SOAs were frankly malignant compared to the 29.6% of MCNs that were.SOAs and SMAs have similar clinicopathologic features. On the other hand, SOAs differ from MCNs by their relatively higher male/female ratio, higher frequency of tumors occurring in the head of the pancreas, and smaller cyst size. Knowledge of these distinguishing clinical features when used in combination with other diagnostic modalities such as endoscopic ultrasonography/fine-needle aspiration will enable clinicians to better differentiate these two pathologic entities preoperatively.

Background: Cystic lesions of the pancreas consist of a broad range of pathological entities. With the exception of the pancreatic pseudocyst, these are usually caused by pancreatic cystic neoplasms. Non-neoplastic pancreatic cystic and cystic-like lesions are extremely rare. In the present article, the surgical experience with these unusual entities over a 14-year period is reported. Methods: Between 1991 and 2004, all patients who underwent surgical exploration for a cystic lesion of the pancreas were retrospectively reviewed. Patients with a pancreatic pseudocyst were excluded. There were 106 patients of whom 8 (7.5%) had a final pathological diagnosis consistent with a non-neoplastic pancreatic cystic or cystic-like lesion, including 3 patients with a benign epithelial cyst, 2 with a pancreatic abscess (one tuberculous and one foreign body), 2 with mucous retention cysts and 1 with a mucinous non-neoplastic cyst. These eight patients are the focus of this study. Results: There were six female and two male patients with a median age of 61.5 years (range, 41–71 years). All the patients were of Asian origin including seven Chinese and one Indian. Four of the patients were asymptomatic and their pancreatic cysts were discovered incidentally on radiological imaging for other indications. All the patients underwent preoperative radiological investigations, including ultrasonography, computed tomography or magnetic resonance imaging, which showed a cystic lesion of the pancreas. Three patients, all of whom were symptomatic, were diagnosed preoperatively with a malignant cystic neoplasm on the basis of radiological imaging. Two patients were eventually found to have a pancreatic abscess, one tuberculous and the other, secondary to foreign body perforation. The third patient was found on final histology to have chronic pancreatitis with retention cysts. The remaining five patients had a preoperative diagnosis of an indeterminate cyst; on pathological examination, they were found to have a benign epithelial (congenital) cyst (n = 3), retention cyst (n = 1) and mucinous non-neoplastic cyst (n = 1). At a median follow up of 20 months (range, 3–34 months), none of the patients had any evidence of recurrent disease. Conclusion: Non-neoplastic cystic and cystic-like lesions of the pancreas are rare causes of pancreatic cystic lesions that are generally benign and do not require surgery when asymptomatic. However, despite advances in diagnostic investigations such as endoscopic ultrasound with fluid aspirate and magnetic resonance imaging, the preoperative diagnosis remains unreliable. Hence, the challenge for all clinicians is to recognize these lesions preoperatively and to avoid ‘unnecessary' surgery.

Pancreatic carcinoma has a poor prognosis and early detection is essential for potentially curative resection. Despite the wide array of diagnostic tools, preoperative detection of small pancreatic carcinomas remains difficult. We report a case of small pancreatic carcinoma of the head of pancreas with indeterminate findings on US, ERCP, MRI and EUS which was successfully diagnosed via fusion CT-PET. This case illustrates the utility of CT-PET in the diagnosis of patients with small pancreatic carcinoma with equivocal findings on conventional diagnostic modalities.

Biliary-enteric transanastomotic stenting is useful under adverse local conditions or when small-calibre bile ducts are encountered.A commercially available feeding jejunostomy kit, Intestofix, was inserted transjejunally through the blind loop of a Roux-en-Y hepatico-jejunostomy.The stent splinted the anastomosis to reduce biliary leaks and may help to prevent subsequent stricture formation.

Background Oesophagogastroduodenoscopy (OGD) is the diagnostic tool of choice in acute upper gastrointestinal haemorrhage. However, the factors causing diagnostic failure are not well documented or discussed. Methods OGDs performed by a single surgeon for acute upper gastrointestinal haemorrhage were reviewed retrospectively for ‘missed’ diagnosis. Results There were nine cases (1·4 per cent) of ‘missed’ diagnosis in 638 OGDs performed over a 3-year period. Incomplete examination caused by a fundal pool precluded three examinations in which two Dieulafoy's lesions and one chronic gastric ulcer were later found. Other difficult examination sites were the cardia/high lesser curve (three cases), the stomal line of a gastrojejunostomy anastomosis (two) and at the junction of the first and second part of the duodenum (one). Four of the overlooked diagnoses were Dieulafoy's lesions. Conclusion OGD was able to diagnose the cause of bleeding in 98·6 per cent of patients with acute upper gastrointestinal haemorrhage. The factors that caused diagnostic failure were a difficult examination site and failure to recognize lesions (Dieulafoy's).

Background There is still considerable controversy and debate regarding the features and prognosis of colorectal cancer in young patients. Methods One hundred and ten patients (5·1 per cent) under the age of 40 years with colorectal cancer (group Y; male: female ratio 48:62) were compared with 2064 patients with colorectal cancer aged 40 years or more (group O; 917 women, 1147 men). Mode of presentation, stage at diagnosis, tumour characteristics and survival were analysed. Results Predisposing malignant conditions and family history of colorectal cancer were present in 20·9 per cent of patients in group Y versus 2·2 per cent in group O (P < 0·001). Common chief complaints included change in bowel habits, bleeding from the rectum and a significantly higher incidence of abdominal pain in group Y. There was no difference in stage at presentation between the two groups (the proportion of Dukes stage A, B, C and ‘D’ lesions in group Y was 8·2, 24·5, 37·3 and 30·0 per cent respectively versus 10·5, 27·9, 33·4 and 28·1 per cent in group O). Tumour site and characteristics were similar in both groups. The incidence of mucinous/signet ring cell and poor grade tumours was 6·9 and 11·8 per cent respectively in group Y and 4·5 and 10·5 per cent in group O. With a mean follow-up of 31·8 months, the overall 5-year survival rate was 54·8 per cent in group Y and 54·1 per cent in group O. Comparing stage for stage, survival was not significantly different in the two groups. However, the adjusted hazard ratios of the age groups Y, M (40–59 years), S (60–79 years), and E (80 years and above) were 1·3, 1 (baseline for calculations), 1·4 and 2·4 respectively, suggesting an adverse outcome for patients in group Y compared with patients aged 40–59 years. Conclusion This study revealed no difference in tumour characteristics and survival in patients with colorectal cancer aged less than 40 years compared with those aged above 40 years. However, a higher hazard ratio in the youngest group may connote a worse prognosis than that for those aged 40–59 years. A significant family history of colorectal cancer and predisposing conditions in the young warrants aggressive screening, surveillance and treatment of the underlying conditions. The detection of colorectal cancer in young patients should be no different from that in the old but demands a high index of suspicion.

Background: Biliary-enteric transanastomotic stenting is useful under adverse local conditions or when small-calibre bile ducts are encountered. Methods: A commercially available feeding jejunostomy kit, Intestofix, was inserted transjejunally through the blind loop of a Roux-en-Y hepatico-jejunostomy. Conclusions: The stent splinted the anastomosis to reduce biliary leaks and may help to prevent subsequent stricture formation.

To evaluate the influence of preoperative biliary drainage on morbidity and mortality after surgical resection for ampullary carcinoma.We analyzed retrospectively data for 82 patients who underwent potentially curative surgery for ampullary carcinoma between September 1993 and July 2007 at the Singapore General Hospital, a tertiary referral hospital. Diagnosis of ampullary carcinoma was confirmed histologically. Thirty-five patients underwent preoperative biliary drainage (PBD group), and 47 were not drained (non-PBD group). The mode of biliary drainage was endoscopic retrograde cholangiopancreatography (n = 33) or percutaneous biliary drainage (n = 2). The following parameters were analyzed: wound infection, intra-abdominal abscess, intra-abdominal or gastrointestinal bleeding, septicemia, biliary or pancreatic leakage, pancreatitis, gastroparesis, and re-operation rate. Mortality was assessed at 30 d (hospital mortality) and also long-term. The statistical endpoint of this study was patient survival after surgery.The groups were well matched for demographic criteria, clinical presentation and operative characteristics, except for lower hemoglobin in the non-PBD group (10.9 +/- 1.6 vs 11.8 +/- 1.6 in the PBD group). Of the parameters assessing postoperative morbidity, incidence of wound infection was significantly less in the PBD than the non-PBD group [1 (2.9%) vs 12 (25.5%)]. However, the rest of the parameters did not differ significantly between the groups, i.e. sepsis [10 (28.6%) vs 14 (29.8%)], intra-abdominal bleeding [1 (2.9%) vs 5 (10.6%)], intra-abdominal abscess [1 (2.9%) vs 8 (17%)], gastrointestinal bleeding [3 (8.6%) vs 5 (10.6%)], pancreatic leakage [2 (5.7%) vs 3 (6.4%)], biliary leakage [2 (5.7%) vs 3 (6.4%)], pancreatitis [2 (5.7%) vs 2 (4.3%)], gastroparesis [6 (17.1%) vs 10 (21.3%)], need for blood transfusion [10 (28.6%) vs 17 (36.2%)] and re-operation rate [1 (2.9%) vs 5 (10.6%)]. There was no early mortality in either group. Median survival was 44 mo (95% CI: 34.2-53.8) in the PBD group and 41 mo (95% CI: 27.7-54.3; P = 0.86) in the non-PBD group.Biliary drainage before surgery for ampullary cancer significantly reduced postoperative wound infection. Overall mortality was not influenced by preoperative drainage.