Surgical management of acute pancreatitis in Italy: lessons from a prospective multicentre study

Abstract

Objective

This study aimed to evaluate the surgical treatment of acute pancreatitis in Italy and to assess compliance with international guidelines.

Methods

A series of 1173 patients in 56 hospitals were prospectively enrolled and their data analysed.

Results

Twenty-nine patients with severe pancreatitis underwent surgical intervention. Necrosectomy was performed in 26 patients, associated with postoperative lavage in 70% of cases. A feeding jejunostomy was added in 37% of cases. Mortality was 21%. Of the patients with mild pancreatitis, 714 patients with a biliary aetiology were evaluated. Prophylactic treatment of relapses was carried out in 212 patients (36%) by cholecystectomy and in 161 using a laparoscopic approach. Preoperative endoscopic retrograde cholangiopancreatography was associated with cholecystectomy in 83 patients (39%). Forty-seven patients (22%) were treated at a second admission, with a median delay of 31 days from the onset of pancreatitis. Eighteen patients with severe pancreatitis underwent cholecystectomy 37.9 days after the first admission. There were no deaths.

Discussion

The results indicate poor compliance with published guidelines. In severe pancreatitis, early surgical intervention is frequently performed and enteral feeding is seldom used. Only a small number of patients with mild biliary pancreatitis undergo definitive treatment (i.e. cholecystectomy) within 4 weeks of the onset of pancreatitis.

Introduction

The surgical treatment of mild and severe acute pancreatitis (AP) has been established in several international evidence-based guidelines.^1–4

In mild biliary pancreatitis, cholecystectomy on the same admission or within 2 weeks of discharge is recommended for patients who are fit enough to undergo surgery.¹ In those who are too frail to be operated on, a definitive endoscopic sphincterotomy is recommended.¹ In severe AP, current indications for surgery include the presence of infected pancreatic necrosis, extensive sterile necrosis in patients in whom symptoms have failed to resolve despite maximal conservative treatment or in patients who develop catastrophic complications related to pancreatic necrosis such as bleeding, visceral perforation or infarction. In addition, it is considered that delaying surgical intervention in severe AP until at least 2 weeks after presentation is advantageous.^1–4 Several studies from different countries have recently highlighted poor compliance with guidelines for the treatment of AP.^5–9 However, most of the data reported in these studies were based on questionnaires sent to specialists (surgeons, gastroenterologists, etc.) with the aim of evaluating their ideal therapeutic approach to AP. Recently, a large multicentre study was performed in Italy under the auspices of the Associazione Italiana per lo Studio del Pancreas (AISP [Italian Association for the Study of the Pancreas]). This group has previously published its results with regard to conservative, endoscopic and surgical treatments.^10,11

The aim of this study was to assess compliance with international guidelines with respect to indications, timing and type of surgery for patients presenting with both mild biliary pancreatitis and AP of all aetiologies in Italian centres that had contributed to the multicentre prospective study on the treatment of severe AP in Italy.

Materials and methods

The AISP study involved 56 centres distributed uniformly throughout Italy. All patients admitted with AP during the period from December 2001 to November 2003 were included. The diagnosis of AP was based on clinical (pancreatic-type pain) and biochemical (three-fold increase in serum amylase or lipase) findings, and/or morphological pancreatic changes detected by ultrasonography (US), computed tomography (CT) or magnetic resonance imaging (MRI). The disease was classified as mild or severe according to the Atlanta criteria, as well as sequelae of AP (fluid collections, pseudocysts, fistulae).^12,13 Data were collected using a software programme which included 530 items, of which 61 were surgical, designed by members of the study group (gastroenterologists, endoscopists and surgeons) and sent to each participating centre. The data were collected and tabulated centrally, and a thorough monitoring process was carried out during the study. At the end of the recruitment phase, the completeness and congruence of each patient's chart were assessed in order to exclude patients for whom data were incomplete or inconsistent.

The pancreatitis was assumed to be biliary in aetiology if gallstones were seen on US or CT or if two of the following abnormal laboratory tests were present: alkaline phosphatase level >125 U/l; alanine aminotrasferase (ALT) level >75 U/l, or bilirubin >2.3 mg/dl.¹⁴ The gallbladder was defined as abnormal if gallstones or signs of inflammation were present on US. Variables assessed for evaluating surgical treatment and adherence to international guidelines are reported in Table 1. In addition, postoperative outcomes including hospital stay, complications and the need for repeated surgery were assessed.

Table 1.

Characteristics of surgical treatment in severe and mild pancreatitis compared with the recommendations of guidelines

Variables	Recommendations
Severe pancreatitis
Indications for surgery	• Infected pancreatic necrosis1,2,4
	•Sterile necrosis with multiple organ failure without improvement despite maximal therapy2
	•Early surgery within 14 days is not recommended unless there are specific indications2
Timing of surgery	After 3–4 weeks2
Type of surgical intervention	Necrosectomy1,2,4
Type of postoperative lavage (intermittent vs. continuous)	Not mentioned
Timing of drainage removal	Not mentioned
Surgical jejunostomy	Not mentioned
Timing of cholecystectomy in severe acute pancreatitis	Delayed after recovery1,2
Preoperative assessment of acute pancreatitis by computed tomography	The surgical indication depends on the clinical picture (evidence of sepsis) and demonstration by computed tomography of pancreatic or peripancreatic necrosis1
Mild pancreatitis
Timing of cholecystectomy	• As soon as the patient has recovered and during the same hospital admission1–4
	• Within 2 weeks of discharge1
Type of cholecystectomy (open vs. laparoscopic)	Not mentioned

Data were presented as medians, ranges and frequencies. Statistical analysis was carried out using the Mann–Whitney U-test and the chi-squared non-parametric test. Data were processed using spss Version 10 (SPSS, Inc., Chicago, IL, USA). P < 0.05 was considered to represent a statistically significant difference.

Results

During the study period, 1173 patients met the inclusion criteria, of whom 167 (14.2%) were diagnosed with severe AP. Twenty-nine (17.4%) of these 167 patients subsequently underwent pancreatic surgery. Of the 1006 (85.8%) patients who presented with mild AP, 802 (79.7%) demonstrated disease of biliary aetiology; their cases were included for further analysis.

Pancreatic surgery

Pancreatic surgery was performed in 29 of the 167 patients with severe pancreatitis (17.4%). The characteristics of these patients were compared with those of the remaining 138 non-operated patients: the aetiology and severity of the pancreatitis were similar. In the operated patients, the rate of cholecystectomy was significantly higher than in patients with severe biliary AP without necrosectomy submitted to delayed gallbladder removal (Table 2).

Table 2.

Characteristics of 167 patients with severe pancreatitis: a comparison of 29 patients undergoing necrosectomy or simple drainage and 138 non-operated patients

	Necrosectomy or drainage patients (n = 29)	Non-operated patients (n = 138)	P-value
Sex, male/female	19/10	74/64	0.241
Median age, years	68.0	69.5	0.166
Aetiological factors, n (%)
Biliary	17 (58.6)	85 (61.6)	0.879
Alcohol	2 (6.9)	12 (8.7)	0.751
Unknown	6 (20.7)	31 (22.5)	0.770
Other	4 (13.8)	10 (7.2)	0.184
Glasgow score	4	4	0.920
Multiple organ failure	10	0
Necrosis	8	0
Endoscopic retrograde cholangiopancreatography	6 (20.7)	30 (21.7)	0.657
Cholecystectomy	18 (62.1)	19 (13.8)	<0.01
Median hospital stay, days	46	22	<0.01
Mortality, n (%)	6 (20.7)	26 (18.8)	0.818
Sequelae,an (%)	13 (44.8)	66 (47.8)	0.952

^aFluid collection, pseudocyst, fistula

Indications for surgery were as follows: acute abdomen in 16 patients; multiple organ failure (MOF) in 10 patients; infected necrosis in eight patients, and sterile necrosis in three patients. More than one indication was present in 18 patients.

Surgery was performed at a median of 2 days (range 0–64 days) after the onset of pancreatitis (Fig. 1). Seven of the 19 patients who underwent an early intervention (before day 5) had not undergone preoperative CT scans. Open necrosectomy was the surgical procedure chosen in 26 patients (89.7%), whereas simple drainage was performed in the remaining three patients; a laparostomy was performed in just one patient. In 21 of the 29 patients necrosectomy was associated with closed postoperative irrigation of the retroperitoneum (continuous in nine patients and intermittent in 12). In 11 patients a feeding jejunostomy was added. The median time to drain removal was 33 days (range 4–148 days) after intervention. Postoperative complications occurred in 14 patients and included: infections (n = 6); MOF (n = 6); pancreatic fistulae (n = 4); haemorrhage (n = 1); occlusion (n = 1); pseudocyst (n = 1), and colonic fistula (n = 1). There were six deaths, of which five resulted from pancreatic causes and one from an extrapancreatic cause (gastrointestinal bleeding). In the patients who underwent early surgery (i.e. before day 5 after the onset of pancreatitis), morbidity was lower and mortality was higher, although the differences were not significant because of the small numbers of patients involved. Eleven patients were discharged as cured and 12 were discharged with sequelae (Table 3).

Figure 1.

Distribution of 29 patients with severe acute pancreatitis according to days between surgery and onset of disease

Table 3.

Relationships between outcome of pancreatitis and timing of surgery (before or after day 5)

	Surgery before day 5, n	Surgery after day 5, n
Recovery	8	2
Morphological sequelae	6	7
Fluid collection	4	2
Pseudocyst	2	2
Fistula	–	4
Mortality	5	1
Total no. of patients	19	10

Elective biliary surgery

Of the 1006 patients with mild pancreatitis, 802 were diagnosed with a biliary aetiology and included for further analysis. Of these, 88 (11.0%) patients had undergone previous cholecystectomy. In 593 (83.1%) of the remaining 714 patients, the gallbladder was reported as abnormal on abdominal US or CT, whereas, in the 121 (16.9%) patients with a normal gallbladder, the aetiology was classified as biliary as a result of enzymatic alterations (none of these patients underwent cholecystectomy). Of the 593 patients with abnormal gallbladder findings, 212 (35.7%) underwent cholecystectomy, which was performed laparoscopically in 161 (75.9%) patients (Fig. 2). In 17 patients with choledocholithiasis, a biliary tree procedure was performed at the same time by open surgery in 11 patients and by laparoscopy in six. Two patients who had undergone a previous cholecystectomy had biliary tract surgery for choledocholithiasis.

Figure 2.

Distribution of 1006 patients with mild biliary acute pancreatitis, according to gallbladder characteristics and type of biliary surgery performed. •Absence of lithiasis or inflammation; ••Presence of lithiasis or inflammation; CBD, common bile duct

Preoperative endoscopic retrograde cholangiopancreatography (ERCP) was associated with cholecystectomy in 83 (39.2%) of 212 patients, 67 (80.7%) of whom subsequently underwent laparoscopic cholecystectomy. ERCP and surgery were performed a median of 6 days (range 1–8 days) and 10 days (range 2–14 days) after admission, respectively.

The median hospital stay was 13 days (range 3–50 days) in patients undergoing cholecystectomy alone and 14 days (range 5–44 days) in those undergoing ERCP + cholecystectomy. In the subgroup of patients subjected to laparoscopic cholecystectomy + preoperative ERCP, the median hospital stay was 13 days (range 4–42 days) vs. 13 days (range 3–50 days) for those undergoing laparoscopic cholecystectomy alone (P = 0.701). In comparison, those undergoing open surgery + preoperative ERCP had a longer median hospital stay of 18 days (range 6–28 days) compared with those undergoing open surgery alone (17 days, range 5–30 days) (P = 0.534).

The postoperative course was uneventful in all patients except two (of whom one underwent laparoscopy and the other an open approach); both developed an abdominal collection.

In 165 of the 212 (77.8%) patients who underwent cholecystectomy, the procedure was performed during the hospital admission for pancreatitis. In the remaining 47 (22.2%) cases, patients were discharged and a second admission was required after a median delay of 31 days (range 12–156 days) after the onset of pancreatitis.

Of the 167 patients with severe pancreatitis, 18 (10.8%) underwent elective cholecystectomy at a median of 28 days (range 3–133 days) after the onset of pancreatitis. The procedure was performed laparoscopically in nine and as open surgery in nine patients. Common duct exploration was required in two patients. No mortality was observed; however, nine longterm sequelae in seven patients (six fluid collections, four pseudocysts, one fistula) were reported. In a further 18 patients with severe pancreatitis, a cholecystectomy was added to a major surgical procedure on the pancreas.

Discussion

In the current study, the high rate of early surgical interventions associated with the absence of a preoperative CT scan in patients with severe AP was an unexpected finding. As a consequence, severe pancreatitis as an indication for surgery was lacking in these patients; in fact, the reported indication for laparotomy was ‘acute abdomen’ in more than 50% of the patients. Despite the fact that only 29 of 167 patients with severe pancreatitis underwent pancreatic surgery, a number too small to allow definitive conclusions, 10 were operated on the same day on which they were admitted. This figure increased to 19 patients in the first 5 days after admission. It is worth noting that, in this group of patients, seven underwent surgery without a preoperative CT scan. Doctors caring for these patients reported that in patients with severe AP, in whom no preoperative CT scan was carried out, laparotomies were performed as a result of diagnostic uncertainties instead of clear surgical indications.

The optimal timing of surgical interventions depends on the pathological characteristics of the infected necrotic tissue. According to the guidelines of the International Association of Pancreatology (IAP),² intervention should be carried out no earlier than 2 weeks after admission, which allows time for the demarcation of the necrotic areas of the pancreas. The later the surgery is performed, the easier and more complete the necrosectomy is.^15–20

Early surgery is indicated only in the very small group of patients with MOF who do not respond to intensive care treatment² or who have severe compartment syndrome.⁴ However, there is still no consensus on the timing of surgical intervention, as demonstrated by a recent European survey. In fact, of the members of the International Hepato-Pancreato-Biliary Association (IHPBA) (i.e. clinicians with a declared interest in AP management), 43% stated that they prefer to perform an intervention within the first 14 days and only 29% choose to wait 21 days.²¹

In this series, a nutritional jejunostomy was placed during surgery in only 11 patients (37.9%); this low number indicates indifference on the part of Italian surgeons to enteral feeding.

In mild biliary pancreatitis, the aim of surgery is to prevent recurrence, as stated by international guidelines^1–4 and confirmed by many studies.^22–25 In this setting, a cholecystectomy should be performed as soon as the patient has recovered from the pancreatitis and preferably during the same hospital admission. In this series, patients were analysed with biliary pancreatitis without any other associated aetiological factor (i.e. alcohol consumption or hypertrigliceridaemia) that might interfere with treatment options. Nevertheless, only 35.7% of all patients with biliary pancreatitis and in situ abnormal gallbladder findings underwent cholecystectomy (212/593). How can this lack of compliance with published guidelines be explained?

Some patients may have been unfit for surgery and may have been treated by sphincterotomy alone.²⁶ Others may have been discharged from medical departments and submitted to delayed surgery. Even if a high rate of drop-out is considered, these reasons do not fully explain the high rate (>60%) of patients who did not undergo surgery. The vast majority of cholecystectomies (76%) were performed using a laparoscopic approach; any recent inflammatory process caused by mild pancreatitis does not contraindicate a minimally invasive intervention.^27,28

The current management of patients with gallstone disease and choledocolithiasis consists of endoscopic stone extraction followed by laparoscopic cholecystectomy.^2,29 In this series, this approach was preferred because local experience in laparoscopic clearance of common bile duct (CBD) stones was poor in most of the centres participating in the study.

More than 20% of the cholecystectomies were performed during a second hospital admission, with a delay of 31 days. This fact, together with the overall low rate of cholecystectomies, contrasts with all published guidelines.^1–4 However, this observed discrepancy has been previously reported by Aly et al.,⁷ Toh et al.³⁰ and Senapati et al.³¹ in the UK and by Lankisch et al.⁹ and Foitzik and Klar⁸ in Germany, and confirmed by the meta-analysis of Barnard and Siriwardena.³²

In 1999, Uhl et al.³³ stated that, after an attack of severe pancreatitis, cholecystectomy should be postponed for at least 3 weeks to limit the risk of infection of pancreatic necrosis. Nealon et al.²⁴ recommended a delayed cholecystectomy in the presence of peripancreatic fluid collection.

In this series, in severe AP cholecystectomy was appropriately delayed^2,34 and the intervention did not seem to affect the course of illness.

This study demonstrated that some aspects of the surgical treatment of biliary lithiasis in AP outlined in published guidelines have not been applied in clinical practice in Italy. This may also be true for European surgery, as suggested by King and Siriwardena.²¹ Conversely, in 2007, Mofidi et al.³⁵ assessed 759 consecutive patients admitted to a regional hepatopancreatobiliary surgery service in Edinburgh covering a population of 1.2 million people and reported a high level of compliance with guidelines, which the authors attributed to the reorganization of general surgical services in Scotland which took place in the 1990s and involved the centralization of teams with an interest in the management of pancreatitis. Connor et al. reported similar results with an audit feedback to surgeons in a tertiary institution in New Zealand.³⁶ However, in the current study only a quarter of all patients with mild gallstone pancreatitis underwent cholecystectomy during their first hospital admission. Even considering the cholecystectomies performed during a second hospital stay, with a 31-day delay, the management goal of adequate treatment was not met in over 60% of patients.

Unlike recently published papers on compliance with guidelines, which report results obtained by questionnaire surveys, the current series does not reflect theoretical treatment patterns (i.e. what specialists would do rather than what they actually do), but reports detailed, prospective clinical data. Thus, it gives a very real picture of clinical practice in Italy.

Although the results of this study call for criticism of non-compliance with published guidelines, they also show low morbidity and mortality in over 1000 patients, similar to published data. What lessons can we learn from this multicentre study? The following simple clinical decisions can probably improve outcomes in pancreatitis. Firstly, patients who are admitted with acute abdominal pain of uncertain cause should undergo appropriate imaging techniques (including early CT) with a higher degree of frequency than is presently the case because, as in this series, the correct differential diagnosis of severe AP from other acute abdominal diseases might avoid the need to perform diagnostic laparotomy on the day of admission. Secondly, in patients operated for severe AP, a jejunostomy should be more diffusely employed because of the proven advantage of enteral feeding over total parenteral nutrition, and finally, according to international guidelines, patients with mild biliary pancreatitis should undergo a cholecystectomy during their initial hospital stay more frequently.

Appendix

Members of the ProInf-AISP (Progetto Informatizzato Pancreatite Acuta, Associazione Italiana per lo Studio del Pancreas) Study Group

Agugiaro S, Turri L	Ospedale Santa Chiara	Trento
Bartoli A, Barberini F, Cavazzoni G	Policlinico Monte Luce	Perugia
Bartolo F, Della Papa D	Ospedale San Luca	Vallo della Lucania
Bassi N, Massani M	Ospedale Regionale Cà Foncello	Treviso
Benedetti A, Macarri G, Piergallini L	Ospedale Torrette	Ancona
Briani G, Bartolasi L	Ospedale di Schio	Schio
Bugnano L	Ospedale Civile	Locri
Buonanno GM, Esposito C	Azienda Ospedaliera Moscati	Avellino
Cordovana A	Ospedale Fatebenefratelli	Milan
Cavina E, Seccia M, Lippolis P, Musco B, Barletta M	Ospedale Santa Chiara	Pisa
Chilovi E, De Guelfi A	Ospedale Generale Regionale	Bolzano
Chirletti P, Caronna R, Scozzafava S, Cardi M	Ospedale Policlinico Umberto I	Rome
Cirino E, Buffone A	Ospedale Vittorio Emanuele I	Catania
Colangelo E, Caracino V	Ospedale Civile	Pescara
Cortese F	Ospedale S Filippo Neri	Rome
Casentini A	Ospedale Civile	Catanzaro
Costamagna G, Trincali A	Policlinico A Gemelli	Rome
Curzio M, Clivio S, Segato S	Ospedale di Circolo	Varese
D'Alessandro A, Ambrosiani V	Ospedale San Bortolo	Vicenza
D'Ambrosio B, Chiodo C	Azienda Ospedaliera	Cosenza
Dicillo M, Reale L, Grandolfo A	Ospedale San Paolo	Bari
Fabbrucci P, Bruscino A, Mugnaini P	Ospedale S Maria Annunziata	Florence
Ferrarese S, Ugenti I	Policlinico Consorziale	Bari
Forte GB, Rocco P	Ospedale Civile	Caserta
Franzè A, Bertelè A, Sereni G	Ospedali Riuniti di Parma	Parma
Friedman D, Mariani L, Morelli F	Ospedale San Martino	Genoa
Gai V, Antro C	Az Ospedaliera S Giovanni Battista	Turin
Garcea D, Gardini A, Lucci E	Ospedale GB Morgagni	Forlì
Giulianotti PC, Sbrana F, Balestracci T	Ospedale della Misericordia	Grosseto
Giulini SM, Pellizzari A, Ronconi M, Cimaschi S	Ospedale Policlinico	Brescia
Grassini M	Presidi Ospedalieri	Asti
Lacignola S, Calandro L	Ospedale Civile	Martina Franca
Mazzitelli L, Costarella SM, Egidio A	Ospedali Riuniti	Reggio Calabria
Mello Teggia P, Stefano E, Cassini P	Ospedale San Luigi Gonzaga	Orbassano
Modica G, Lupo F, Giraci G	Policlinico P Giaccone	Palermo
Mosca F, Del Chiaro M	Ospedale Cisanello	Pisa
Mosella G, Benassai G	Policlinico Universitario Federico II	Naples
Nanni M, D'Aristotile A	Ospedale di Pescara	Pescara
Negro P	Ospedale Policlinico Umberto I	Rome
Pirazzoli A	Ospedale degli Infermi	Rimini
Rabitti PG	Azienda Ospedaliera Cardarelli	Naples
Romano C, Gerardi G, Troianello B	Ospedale Assalesi	Naples
Russello D, Di Stefano A, Avelli S	Ospedale Cannizzaro	Catania
Salval N, Bellini N	Ospedale Generale Regionale	Aosta
Scalon P	Ospedale Bassano del Grappa	Bassano del Grappa
Staudacher C, Parolini D	Ospedale San Raffaele	Milan
Strazzabosco M, Signorelli S	Ospedali Riuniti	Bergamo
Tedeschi U	Ospedale San Martino	Belluno
Testoni PA, Masci E, Mariani A	Ospedale San Raffaele	Milan
Torelli E, Garcea MR, Lombardi V, Lecconi L	Ospedale del Ceppo	Pistoia
Valeri L, Presenti L, Alessio F	Azienda Ospedaliera Careggi	Florence
Ventrucci M, Virzi S, Cipolla A	Ospedale di Bentivoglio	Bentivoglio

Conflicts of interest

None declared.