Introduction

Acute pancreatitis is the single most frequent gastrointestinal cause of hospital admissions in the USA, affecting up to 45/100,000 persons and accounting for roughly $2.2 billion of inpatient cost annually.1,2 The incidence of acute pancreatitis is rising and will likely only continue to do so as the population ages and the prevalence of obesity increases.3–5 Approximately 44 to 54 % of these episodes of pancreatitis are caused by gallstones, and another 20 to 34 % are of unclear etiology and may be related to occult biliary disease such as sludge or microlithiasis.3 Without cholecystectomy (CCY), the risk of recurrent biliary pancreatitis and other biliary tract complications is as high as 18 % over a median period of 40 days.6

In order to mitigate these risks, current guidelines recommend CCY during the index admission for mild to moderate biliary pancreatitis.7 Due to a higher rate of perioperative complications during CCY performed soon after severe pancreatitis with peripancreatic fluid collections, it is recommended to delay cholecystectomy in these cases until the collections are resolved or for at least 6 weeks.8 The optimal timing of CCY in patients with necrotizing pancreatitis has not been well studied. We hypothesized that CCY performed at the time of pancreatic debridement would not add significant morbidity to the procedure but would decrease the risk of subsequent biliary tract complications.

Methods

This study was approved by Massachusetts General Hospital’s (MGH) Institutional Review Board (study protocol no. 2011P002679) and was HIPAA compliant. We retrospectively queried our institutional database of patients who underwent pancreatic necrosectomy for necrotizing pancreatitis from 1992 to 2012. Relevant biliary-related variables (preoperative gallbladder imaging, single-stage CCY at time of necrosectomy, final pathology, postoperative biliary complications, need for eventual CCY postnecrosectomy) were supplemented to the database by means of an independent medical chart review.

Definitions and Grading Systems

Infected necrosis was confirmed if there was a positive microbiological culture from fine-needle aspiration or catheter drainage before necrosectomy or positive culture from index necrosectomy. Pancreatic fistula was defined as drainage of amylase-rich fluid (>450 U/ml) either through surgical drains or skin openings including through surgical site closures or cutaneous fistulas. Postoperative bleeding was defined as any decrease in hematocrit levels, with or without hemodynamic instability, that prompts angiographic, surgical, or endoscopic interventions or hemodynamic instability that requires two or more units of blood for treatment.

The computed tomography (CT) severity score index (Balthazar score) was used to characterize the extent of pancreatic parenchymal injury in patients with necrotizing pancreatitis. This scoring system combines a predefined acute pancreatitis grade9 with the degree of pancreatic necrosis stratified according to the proportion of pancreatic parenchyma involved10 and has been validated and shown to have prognostic correlation with disease morbidity and mortality. The Acute Physiology and Chronic Health Evaluation (APACHE) II score was used to stratify the patient’s physiologic state 24 h prior to pancreatic necrosectomy.11

Surgical Technique

Open necrosectomies were performed by utilizing standard techniques. The majority were approached via a midline incision, with the lesser sac entered through the transverse mesocolon or the gastrocolic omentum. After entering the lesser sac, necrotic pancreatic tissues were bluntly dissected, leaving healthy parenchyma. If a single-stage cholecystectomy was to be performed, it was carried out in using the standard dome-down technique. At the conclusion of the procedure, closed suction drains were placed in all patients and, for a significant number of patients, ¾-in. Penrose drains stuffed with gauze were packed into each major extension of the cavity, brought out through separate stab wounds, and secured to the skin with sutures as previously described.12

Statistical Analyses

Statistical analyses were performed using Intercooled Stata software, version 12.0 (StataCorp, College Station, TX) and SAS version 9.2 (SAS Institute Inc., Cary, NC). Continuous variables were summarized as mean ± standard deviation or median, as appropriate based on their distribution. Categorical variables were reported as frequencies and proportions. All reported p values are two sided and p ≤ 0.05 was used to indicate statistical significance.

Results

Between January 1992 and January 2012, we identified 217 consecutive patients who underwent pancreatic necrosectomy for necrotizing pancreatitis with a median follow-up of 32 months. The most common etiologies of pancreatitis were biliary (41 %) and alcoholic (24 %), with a median CT severity index score of 6 ± 1.6 and a 64 % incidence of infected necrosis. Twenty-eight percent of patients experienced organ failure within the 24 h prior to necrosectomy, with a similar proportion of patients requiring ICU admission prior to necrosectomy. The mean APACHE II score in the 24 h prior to necrosectomy was 7 ± 5.7. Thirty-two percent of patients received percutaneous drainage prior to necrosectomy, with the mean time from the onset of pancreatitis to pancreatic necrosectomy being 37 days.

Of the 217 patients undergoing pancreatic necrosectomy, 70 (59 %) patients underwent concomitant CCY (Fig. 1). In the CCY group, an intraoperative cholangiogram (IOC) was performed in 6 (2.8 %) patients. Four IOCs were performed to evaluate for choledocholithiasis, which was not found, and two were performed to confirm biliary anatomy prior to cystic duct transection. The CCY cohort was analyzed against the no CCY group. There was a statistically significant difference in the etiology of pancreatitis between both groups; 83 % of patients in the CCY group had biliary pancreatitis, versus 50 % in the no CCY group (p < 0.001). There were no differences in patient demography, CT severity index score, incidence of infected necrosis, severity of disease, need for percutaneous intervention prior to necrosectomy, and time from admission to necrosectomy (Table 1).

Fig. 1
figure 1

Denominator diagram of patients who underwent single-stage cholecystectomy at the time of pancreatic necrosectomy and those that did not. CCY cholecystectomy

Full size image
Table 1 Patient demographics and disease-related variables of patients with necrotizing pancreatitis undergoing necrosectomy, dichotomized to those that received simultaneous CCY and those that did not
Full size table

The duration of surgery was longer for patients who underwent simultaneous CCY when compared to the group that did not (128 vs 100 min, p = 0.003). The estimated operative blood loss (425 vs 375 cc, p = 0.459) and requirement for intraoperative blood transfusion (40 vs 40 %, p = 1.000), however, did not differ. The postoperative outcome did not differ between groups (Table 2). Postoperative bleeding occurred in 2.9 and 6.1 % of patients in the CCY and no CCY groups, respectively (p = 0.305). Similarly, there was no difference in the incidence of pancreatic fistula between the groups (40 % in the CCY group vs 36 % in the no CCY group, p = 0.574). In the CCY group, there were no incidences of bile duct injury or biliary leaks.

Table 2 Intraoperative and postoperative outcomes of patients with necrotizing pancreatitis undergoing necrosectomy, dichotomized to those receiving simultaneous CCY and those that did not
Full size table

Based on final histopathological analysis, 79 % of patients with gallstones or biliary sludge on preoperative imaging had gallstones or sludge identified on final pathological analysis. Forty-three percent of patients who had no cholelithiasis or biliary sludge on preoperative imaging had gallstones or sludge identified pathologically in the gallbladder specimen after single-stage CCY (Table 3).

Table 3 Correlation between the presenting preoperative imaging of biliary disease with the final postoperative histopathological analysis
Full size table

In the no CCY group (n = 147), 98 patients had undergone CCY prior to the time of necrosectomy. In the remaining 49 patients, CCY was not performed due to a clear non-biliary etiology (35 %), technical difficulty (29 %), intraoperative hemodynamic instability (18 %), or surgeon preference (18 %) (Table 4). Of those who did not receive a single-stage CCY, biliary complications developed in 17 (35 %) patients (21 % cholecystitis, 14 % recurrent gallstone pancreatitis) at a median time to incidence of 10 months (range 0.5 to 112 months). This is in contrast to the simultaneous CCY group where only 3 (4.3 %, p < 0.001) patients developed biliary complications (all were biliary pancreatitis). Of the patients in the no CCY group who eventually developed biliary complications, 37 % had no gallstones or sludge on preoperative imaging. In the no CCY cohort, 17 (35 %) patients eventually received a postnecrosectomy cholecystectomy, of which 75 % required an open procedure.

Table 4 Reasons for why simultaneous CCY was not performed with pancreatic necrosectomy in the no CCY group
Full size table

Discussion

The International Association of Pancreatology/American Pancreatic Association (IAP/APA) evidence-based guidelines for the management of acute pancreatitis recommend that CCY should be performed during index admission for mild biliary pancreatitis, but should be delayed in patients with severe biliary pancreatitis with peripancreatic collections until the collections either resolve or 6 weeks has passed.13 After discussion at the consensus conference, the committee declined to make a recommendation on the advisability of performing simultaneous cholecystectomy at the time of pancreatic necrosectomy, citing a lack of evidence. That discussion provided the impetus for us to perform this study.

The recommendation for index CCY in mild biliary pancreatitis is based on a systematic review of nine studies describing 988 patients with almost equal distribution of patients undergoing CCY at index admission (48 %) versus interval CCY after a median of 40 days. In that study, the interval CCY group experienced increased readmission for biliary events when compared to the CCY at index admission group (18 vs 0 %, p < 0.0001). A similar study from the Dutch Pancreatitis Study Group corroborated these findings (13.7 % readmission for biliary events)14 and, together with the systematic review, led to a strong recommendation (grade 1C) supporting index CCY for mild biliary pancreatitis.6 The recommendation for delayed CCY in severe biliary pancreatitis, however, is solely based on two retrospective studies and is weak (grade 2C). In a study of 151 patients, Nealon et al. found an increased incidence of infected collections in patients undergoing early CCY after severe biliary pancreatitis.8 Patients undergoing pancreatic necrosectomy for biliary pancreatitis routinely underwent cholecystectomy at the author’s institution, but were excluded from the study. In the other study of just 30 patients, Heider et al. reported no recurrent biliary events during the waiting interval prior to CCY if endoscopic sphincterotomy was performed,15 leading to the aforementioned recommendation. Overall, we believe the evidence for the effectiveness of early cholecystectomy for preventing recurrent biliary complications is much stronger than the evidence that early cholecystectomy is harmful in severe pancreatitis. Furthermore, the complication most strongly associated with early cholecystectomy in severe pancreatitis—infection of peripancreatic collections—may not be relevant to the population of patients undergoing pancreatic necrosectomy in which peripancreatic collections are either already infected or will be evacuated at the time of surgery. No prior studies have investigated the safety and feasibility of simultaneous CCY at the time of open necrosectomy.

We audited our institutional experience with single-stage CCY at the time of pancreatic necrosectomy in severe pancreatitis of all etiologies. General concerns about single-stage CCY include the advisability of performing an additional procedure in patients who may be hemodynamically unstable from necrotizing pancreatitis and the safety of CCY in an inflamed abdominal environment which in theory could increase the rates of bleeding, biliary tract injury, or infection. Our data indicate that simultaneous CCY at the time of necrosectomy was safe and usually feasible. We found slightly higher operative times in the CCY group, but no evidence of increased morbidity or mortality. The duration of surgery was longer in the CCY group, but the estimated blood loss and requirement for intraoperative blood transfusion were no different between both groups.

In patients with a gallbladder who did not receive a single-stage CCY, biliary complications developed in 35 % of patients (21 % cholecystitis, 14 % recurrent gallstone pancreatitis) at a median time to incidence of 10 months, versus 4 % in the single-stage CCY group. It is important to note that while endoscopic sphincterotomy reduces subsequent biliary complications (from 24 to 10 % at a median follow-up of 40 days in one study), CCY confers superior protection against biliary events (0 to 4 % in the literature at 31 weeks, 4.3 % in our series).6,14 Additionally, within the no CCY cohort, 17 (12 %) patients eventually received a postnecrosectomy cholecystectomy, of which 75 % required an open procedure, well known to carry a higher perioperative morbidity, mortality, and prolonged length of hospital stay.16 It is also worth noting that all the aforementioned studies and recommendations from guidelines specifically referred to gallstone-induced pancreatitis. Within our study, 43 % of patients who had no radiographic evidence of cholelithiasis preoperatively were found to have gallstones or sludge identified within the gallbladder specimen on final histopathological analysis. In fact, 37 % of patients who did not receive single-stage CCY and developed biliary complications did not have evidence of gallstones or biliary sludge prior to their pancreatic necrosectomy. This suggests that CCY should possibly be considered even in pancreatitis of unclear etiology.

The utility of intraoperative cholangiography (IOC) at the time of CCY for gallstone pancreatitis is not well defined. Small retrospective studies have reported no difference in the rate of recurrent pancreatitis or biliary complications when IOC was utilized at the time of CCY.17,18 One even suggested that IOC resulted in a longer operative time and a prolonged postoperative course with no effect on the incidence of retained common bile duct stones.19 One quarter of patients with gallstone pancreatitis may have stones in the common bile duct at the time of IOC, a proportion of which would pass spontaneously.18–21 At our institution, we do not routinely perform IOC at the time of cholecystectomy unless technical difficulty is encountered at the time of CCY or if there is concern for persistent choledocholithiasis. As reported above, IOC was performed in only six (2.8 %) patients. No choledocholithiasis was found at the time of pancreatic necrosectomy, though a number of patients had undergone preoperative endoscopic retrograde cholangiopancreatography (ERCP) and sphincterotomy. If choledocholithiasis were found at the time of debridement, the choice of whether to perform surgical common bile duct exploration or postoperative ERCP and sphincterotomy would need to be made on a case-by-case basis.

This study has a number of limitations. Only patients in whom CCY was judged intraoperatively to be reasonable and feasible based on the physiologic status of the patient and the perceived technical difficulty underwent single-stage CCY. This study should not be interpreted as an endorsement of single-stage CCY at the time of pancreatic necrosectomy in all cases of biliary pancreatitis. It is entirely possible that if this were undertaken, significant complications related to CCY would arise. Rather, our data show that with appropriate intraoperative decision-making, single-stage CCY can be safely performed in selected patients and that this effectively reduced subsequent morbidity related to the biliary tract. This study analyzed patients undergoing open necrosectomies and simultaneous cholecystectomies by experienced pancreatic surgeons in a high-volume referral center. Whether these results are generalizable to laparoscopic approaches during minimally invasive necrosectomy or in the general community is unknown. This study also has important implications regarding the timing of CCY for biliary pancreatitis patients debrided by endoscopic and retroperitoneal routes. As video-assisted retroperitoneal22 and endoscopic debridement23 gain popularity, it is likely that more patients will retain their gallbladders after necrosectomy. Our data strongly suggests that cholecystectomy should be performed in these patients to reduce their subsequent risk of biliary complications. We also report a relatively high rate of open CCY in patients undergoing CCY after pancreatic necrosectomy. Our data encompass two decades of clinical practice, and it is likely that as experience with laparoscopic CCY has grown, our data may overestimate the current need for open CCY after necrosectomy. Finally, due to our retrospective cohort design, we may fail to capture late complications that were not managed at our institution.

In summary, this is the first study investigating outcomes after single-stage CCY at the time of pancreatic necrosectomy and showing that it is safe in selected patients. Single-stage CCY should be performed if technically feasible to prevent future biliary complications and reduce the need for a subsequent separate, often open, abdominal operation. We propose that future consensus statements and evidence-based guidelines consider revising their recommendations to support performing CCY at the time of necrosectomy for biliary pancreatitis if feasible.