Strategy for treatment of acute necrotizing pancreatitis: Importance of continuous regional arterial infusion of protease inhibitor and antibiotics in the early phase

A new strategy for the treatment of acute necrotizing pancreatitis (ANP) is reported. In this prospective study, all patients received intensive medical support. Surgery was performed in patients with infected pancreatic necrosis and/or sepsis. Continuous regional arterial infusion (CRAI) of the protease inhibitor, nafamostat mesilate, and the antibiotic imipenem was initiated in patients with ANP referred to our hospital within 7 days of the onset of the disease. Sixty patients with ANP were allocated to three groups: group I, no CRAI (n =16); group II, CRAI of nafamostat only (n=22); and group III, CRAI of nafamostat mesilate and imipenem (n=22). The mortality rate was 43.3% in group I, 13.6% in group II (P<0.05 vs group I), and 13.6% in group III (P<0.05 vs group I). The frequency of infected pancreatic necrosis was 50% in group I, 36.4% in group II, and nil in group III (P<0.01 vs group I and II). Combination of the protease inhibitor and the antibiotic infused intraarterially reduced the mortality rate and the frequency of infected pancreatic necrosis. However, 6 patients in the CRAI groups died of multiple organ failure (MOF), although the pancreatic necrosis was sterile. Massive retroperitoneal necrosis and bleeding was observed in these patients. CRAI is a potent mode of treatment in the early phase of necrotizing pancreatitis and most patients respond to this treatment. However, surgical intervention should be considered when the patient does not respond to CRAI and organ failure progresses, even though the pancreatic necrosis is sterile.     

Fungal infection but not type of bacterial infection is associated with a high mortality in primary and secondary infected pancreatic necrosis

INTRODUCTION: Knowledge of microbiology in the prognosis of patients with necrotizing pancreatitis is incomplete. AIM: This study compared outcomes based on primary and secondary infection after surgery for pancreatic necrosis. METHOD: From a limited prospective database of pancreatic necrosectomy, a retrospective case note review was performed (October 1996 to April 2003). RESULTS: 55 of 73 patients had infected pancreatic necrosis at the first necrosectomy. 25 of 47 patients had resistant bacteria to prophylactic antibiotics (n = 21) or did not receive prophylactic antibiotics (n = 4), but this was not associated with a higher mortality (9 of 25) compared to those with sensitive organisms (4 of 22). Patients with fungal infection (n = 6) had a higher initial median (95% CI) APACHE II score compared to those without (11 (9-13) verus 8.5 (7-10), p = 0.027). Five of six patients with fungal infection died compared to 13 of 47 who did not (p = 0.014). With the inclusion of secondary infections 21 (32%) of 66 patients had fungal infection with 10 (48%) deaths compared to 11 (24%) of 45 patients without fungal infection (p = 0.047). CONCLUSION: Whether associated with primary or secondary infected pancreatic necrosis, fungal but not bacterial infection was associated with a high mortality.     

急性坏死性胰腺炎的手术时机及手术指征

急性坏死性胰腺炎（ＡＮＰ）的手术时机及手术指征仍存争议。作者对该院１９８５年以来手术治疗的１１９例ＡＮＰ患者进行分析。将发病２周以内手术者定为早期手术，２周以后手术者为晚期手术。发现早期手术病例术中病理所见多为局灶性坏死（占７５．６％），术后并发症以循环及胰外脏器功能紊乱为主，术后死亡率达２８．２％；而晚期手术者多系全胰坏死型（占５３．７％），术后并发症以胰周局部脏器病变为主，死亡率降至１２．３％。因此作者认为ＡＮＰ应尽可能采用晚期手术，并对早期及晚期手术的指征进行了讨论。     

急性坏死性胰腺炎外科治疗20年经验总结

作者总结了１９７４年８月～１９９４年１１月２４３例急性坏死性胰腺炎外科治疗的经验。按治疗观点及治疗方式分为三个阶段。第一阶段为针对胰腺坏死作早期彻底切除坏死组织手术,辅以“三造瘘”、“创口敞开”、“局部灌洗”等措施。治愈率为６１．３％。第二阶段以坏死是否伴有感染为指征的“个体化治疗”,即对感染者作手术治疗,对非感染者作保守治疗,治愈率提高到６８．５％。第三阶段治疗有二个特点。第一是在手术时间上强调后期手术,如在严密的治疗观察下感染能缓解控制,可延缓到后期手术;对有恶化倾向者,则作早期或急诊手术。第二是建立综合治疗体系,依靠多学科的参与,使手术治愈率提高到８０％,非手术治愈率达到１００％,总治愈率为８３．１％。     

急性坏死性胰腺炎外科治疗20年经验总结

总结了1974年8月～1994年11月243例急性坏死性胰腺炎外科治疗的经验.按治疗观点及治疗方式分为三个阶段.第一阶段1974年8月～1987年12月为针对胰腺坏死作早期彻底切除坏死组织手术,辅以“三造瘘”、“创口敞开”、“局部灌洗”等措施.治愈率为61.3%.第二阶段1988年1月～1991年6月以坏死是否伴有感染为指征,对有感染者作手术治疗,治愈率67.1%;对非感染者作保守治疗,治愈率85.7%,总治愈率提高到68.5%.第三阶段1991年7月～1994年11月治疗有二特点.第一是在手术时间上强调后期手术,如在严密的治疗观察下感染能缓解控制、可延缓到后期手术;对于有恶化倾向、或对阻塞性胆源性胰腺炎及胰腺脓肿穿破致腹膜炎等病例则作早期或急诊手术.第二是建立综合治疗体系,依靠多学科的参与使患者渡过诸如全身细菌、霉菌感染、深部脓肿的处理、以及全身严重消耗的纠正等、使手术治愈率提高到80%、非手术治愈率达到100%、总治愈率为83.1%.     

Characteristics of Infection with Candida Species inPatients with Necrotizing Pancreatitis

This study focuses on the relevance of Candida infection (albicans and non-albicans) in patients with necrotizing pancreatitis. Altogether, 92 patients with infected pancreatic necrosis were reviewed for Candida infection. All patients underwent surgical necrosectomy for infected pancreatic necrosis. Data from patients with Candida growth in intraoperative smears were compared to those obtained from patients without Candida infection. There were 22 patients (24%) with Candida infection. Patients with or without Candida infection were comparable regarding age, gender, etiology, and severity scores at admission. Candida patients suffered a higher mortality (64% vs.19%, p = 0.0001) and experienced more systemic complications (3.2 +/- 1.6 vs. 2.1 +/- 1.4; p= 0.004) than patients without Candida. Preoperative antibiotics were given significantly longer prior to Candida infection (19.0 +/- 13.2 vs. 6.4 +/- 10.3 days; p < 0.0001). With regard to the concomitant spectrum of bacteria, solitary gram-negative infection was rare in Candida patients (5% vs. 43%, p =0.0006). The presence of Candida in patients with infected pancreatic necrosis is associated with increased mortality. Our data provide evidence that application of antibiotics contributes to the development of Candida infection and to changes in the bacterial spectrum of infected necrosis with an increase in the incidence of gram-positive infection.     

Surgical treatment of necrotizing pancreatitis by necrosectomy and closed lavage: changing patient characteristics and outcome in a 19-year, single-center series

BACKGROUND: Surgical treatment of necrotizing pancreatitis (NP) has undergone considerable changes during the past 2 decades. In this study, we report our experience of necrosectomy and continuous closed lavage over the past 19 years in an attempt to define changes in patient characteristics and outcome at an academic referral center. METHODS: Among 1520 patients admitted with acute pancreatitis, 392 had NP, 285 of whom underwent operative treatment. The total series was evaluated separately for treatment period A (May 1982 until April 1993) and treatment period B (May 1993 until May 2001). RESULTS: Intraoperative bacteriology revealed sterile necrosis in 145 and infected necrosis in 140 patients. Preoperative disease severity did not differ between the groups; however, the extent of pancreatic parenchymal necrosis was less in patients with sterile necrosis (P < .003). Postoperative complications were more frequent in infected necrosis (78%) than in sterile necrosis (61%) (P < .004), with mortality rates of 27% and 23%, respectively. The analysis of the 2 treatment periods revealed that during period B, there was a decrease in operatively treated patients with sterile necrosis (P < .0005). The preoperative systemic disease severity was significantly higher in these patients than in patients with infected necrosis. CONCLUSIONS: Surgical treatment of NP by necrosectomy and closed lavage carries an overall mortality of 25%. Patients with sterile necrosis and early onset high disease severity may represent a distinct clinical entity in whom the optimal treatment strategy remains to be defined.     

A randomized controlled trial of enteral versus parenteral feeding in patients with predicted severe acute pancreatitis shows a significant reduction in mortality and in infected pancreatic complications with total enteral nutrition

BACKGROUND: Infectious complications are the main cause of late death in patients with acute pancreatitis. Routine prophylactic antibiotic use following a severe attack has been proposed but remains controversial. On the other hand, nutritional support has recently yielded promising clinical results. The aim of study was to compare enteral vs. parenteral feeding for prevention of infectious complications in patients with predicted severe acute pancreatitis. METHODS: We screened 466 consecutive patients with acute pancreatitis. A total of 70 patients with objectively graded severe acute pancreatitis were randomly allocated to receive either total enteral nutrition (TEN) or total parenteral nutrition (TPN), within 72 h of onset of symptoms. Baseline characteristics were well matched in the two groups. RESULTS: The incidence of pancreatic infectious complications (infected pancreatic necrosis, pancreatic abscess) was significantly lower in the enterally fed group (7 vs. 16, p = 0.02). In the TEN group, 7 patients developed multiple organ failure whereas 17 parenterally fed patients developed multiple organ failure (p = 0.02). Overall mortality was 20% with two deaths in the TEN group and twelve in the TPN group (p < 0.01). CONCLUSION: Early TEN could be used as prophylactic therapy for infected pancreatic necrosis since it significantly decreased the incidence of pancreatic infectious complications as well as the frequency of multiple organ failure and mortality.     

Bacterial analysis of infected pancreatic necrosis and its prevention (Symposium 8: Pancreatobiliary infection (IHPBA))

In severe acute pancreatitis, sepsis mainly due to pancreatic or peripancreatic infection have emerged as the most serious complications and now accounts for more than 80% of deaths. Collective review of organisms associated with secondary pancreatic infection in patients with acute pancreatitis has revealed that most of them are intestinal flora. Several experimental studies including ours have revealed that acute pancreatitis promotes bacterial translocation (BT), which in turn leads to infection of the pancreas and septic complications. Prophylactic antibiotics given intravenously have been demonstrated to be beneficial in reducing the rate of pancreatic infection, but their survival benefit remains unclear. We have demonstrated that continuous regional arterial infusion (CRAI) of an antibiotic is more effective than intravenous administration in preventing pancreatic infection and improving survival, in a canine model of acute necrotizing pancreatitis. Our recent experimental study has revealed that CRAI of an antibiotic via the superior mesenteric artery (SMA) is effective in mitigating intestinal mucosal damage and preventing BT in acute pancreatitis, thereby improving survival. BT aggravates pancreatic necrosis and remote organ damage in acute pancreatitis, and SMA infusion of antibiotics is effective in preventing BT and is practical for clinical use.     

Effects of caffeic acid phenethyl ester on pancreatitis in rats

BACKGROUND: This study investigated the effect of caffeic acid phenethyl ester (CAPE) on acute necrotizing pancreatitis (ANP) induced by glycodeoxycholic acid in rats. CAPE, an active component of honeybee propolis, has previously been determined to have antioxidant, anti-inflammatory, antiviral, and anticancer activities. MATERIALS AND METHODS: Forty-eight rats were divided into four groups of 12. Group 1 animals received intraductal saline and intravenous saline infusion treatment. Group 2 was given intraductal saline and intraperitoneal CAPE infusion treatment. ANP was induced in the animals in group 3 (ANP with saline infusion), and group 4 had induced ANP plus CAPE infusion treatment (ANP with CAPE infusion). Sampling was performed 48 h after treatment. RESULTS: ANP induction significantly increased mortality rate, pancreatic necrosis, and bacterial infection in pancreatic and extrapancreatic organs. ANP also increased levels of amylase and alanine aminotransferase (ALT) in serum, increased levels of urea and lactate dehydrogenase in bronchoalveolar lavage fluid (BAL LDH), increased the activities of myeloperoxidase (MPO) and malondialdehyde (MDA) in pancreas and lung tissue, and decreased the serum calcium levels. The use of CAPE did not significantly reduce the mortality rate but significantly reduced the ALT and BAL LDH levels, the activities of MPO and MDA in the pancreas, the activity of MDA in the lungs, and pancreatic damage. The administration of CAPE did not reduce the bacterial infection. CONCLUSIONS: These results indicate that CAPE had beneficial effects on the course of ANP in rats and suggest that CAPE shows promise as a treatment for ANP.     

Severe acute pancreatitis--diagnostic and therapeutic strategy

To establish the optimal diagnosis and therapeutical strategy in severe acute pancreatitis. 94 (56.9%) severe acute pancreatitis (79 males and 15 females, aged between 26 and 81), selected from 165 acute pancreatitis admitted in the last 5 years (2000-2004) were analyzed. The disease was assigned as severe when one or more of the following criteria were present: Ranson score >3 on admission or at 48 hours, APACHE II score >8, visceral failures, Balthazar CT score C, D or E and local complications (infected necrosis, pseudocyst or pancreatic abscess). Medical treatment (aggressive supportive intensive care therapy, minimizing pancreatic secretion and antibiotic therapy) was the first therapeutical step in all cases. 49 (52.1%) patients were operated on: 20 as early surgery imposed by biliary sepsis (16 cases) or by an acute abdomen with uncertain etiology and unfavourable evolution, and 22 as late surgery (at least 12 days after onset), imposed by the presence of the infected pancreatic necrosis, visceral failures or other local complications, the necrosectomy being the main surgical procedure for infected necrosis. 77 (81.9%) cases had a fair evolution. The conservative treatment led to a complete recovery in 37 (37.2%) cases. We registered an overall mortality rate of 12.7% and postoperative mortality rate of 14%; we also registered 5 (10.2%) postoperative complications: 4 pancreatic and 1 colonic fistulae. (1) The treatment of the severe acute pancreatitis must be performed only in the specialized multidisciplinary well equipped centers with very well trained staff. (2) Medical conservative treatment (aggressive supportive intensive care therapy and antibiotic therapy) is the main therapeutical method within the acute phase (first two weeks). (3) Very restrictive surgical indications within the acute phase. (4) Necrosectomy is the main surgical procedure for the infected necrosis.     

急性坏死性胰腺炎的手术时机及手术指征

探讨急性坏死性胰腺炎(ANP)的手术时机及手术指征。方法:对1985年以来手术治疗ANP119例进行了回顾性分析。结果:将发病2周以内手术者定为早期手术,2周以后手术者为晚期手术。发现早期手术病例中病理所见多为局灶性坏死(占75.6％),术后并发症以循环及胰外脏器功能紊乱为主,术后死亡率达28.2％;而晚期手术者系全胰坏死型(占53.7％),术后并发症以胰周局部脏器病变为主,死亡率降至12.3％。结论:对ANP应尽可能采用晚期手术。     

继发感染致急性坏死性胰腺炎病变加重的动物模型研究

目的 提供继发感染致急性坏死性胰腺炎胰腺坏死病变加重的实验室依据.方法 24只SD大鼠平均随机分成3组:对照组、以逆行胆胰管注射法制备急性坏死性胰腺炎(acute necrotizing pancreatitis,ANP)组与感染性坏死性胰腺炎(infected necrotizing pancreatitis,INP)组.8 h后抽血测定血清淀粉酶、C反应蛋白,并取胰腺组织行光镜、电镜检查和细菌培养.结果 胰腺组织细菌培养阳性率:对照组为0(0/8);ANP组为12.5%(1/8);INP组为100%(8/8).光镜检查结果:对照组未见出血、坏死改变;ANP组可见出血与灶状坏死,中等量炎细胞浸润;INP组出血常见,多见大片状的坏死,大量炎细胞浸润.电镜检查结果:对照组未见异常改变;ANP组可见分泌颗粒增多,粗面内质网轻度扩张;INP组可见分泌颗粒明显增多,粗面内质网扩张明显.ANP组血清淀粉酶、C反应蛋白水平明显升高,而在INP组升高较ANP组更加明显.结论 继发感染确实可导致急性坏死性胰腺炎坏死病变明显加重.     

胆源性胰腺炎的外科治疗

目的　探讨胆源性胰腺炎 (GP)的外科治疗原则。方法　回顾分析本院 1991年 1月至2 0 0 0年 1月收治的GP12 0例 ,其中轻型 95例 ,重型 2 5例。结果　 (1)GP早期手术组 (入院 4 8小时内 )与延期手术组 (入院 4 8小时后 )其住院日并无差别 ,但早期手术组胆总管探查率 (10 0 % )、多脏器功能障碍综合征 (MODS)发生率 (40 % )、胰腺感染率 (2 8% )及死亡率 (2 0 % )均明显高于延期手术组 ,分别为P <0 .0 5、P <0 .0 1、P <0 .0 1、P <0 .0 1。 (2 )重型GP的MODS发生率和胰腺感染率分别为 4 8%和 32 % ,死亡原因与MODS及胰腺感染呈明显的正相关。结论　轻型GP经综合治疗愈后 ,应在同一住院期间手术去除胆道病因以防复发。重型GP综合治疗的同时 ,根据“个体化”原则选取适当的手术时机和手术方式。防治MODS及胰腺感染是降低GP死亡率的关键。     

Management of infection in acute pancreatitis

The clinical course of acute pancreatitis varies from a mild, transitory illness to a severe, rapidly fatal disease. In about 80% to 90% of cases pancreatitis presents as a mild, self-limiting disease with low morbidity and mortality. Unlike mild pancreatitis, necrotizing pancreatitis develops in about 15% of patients, with infection of pancreatic and peripancreatic necrosis representing the single most important risk factor for a fatal outcome. Infection of pancreatic necrosis in the natural course develops in the second and third week after onset of the disease and is reported in 40% to 70% of patients with necrotizing pancreatitis. Just recently, prevention of infection by prophylactic antibiotic treatment and assessment of the infection status of pancreatic necrosis by fine-needle aspiration have been established in the management of severe pancreatitis. Because medical treatment alone will result in a mortality rate of almost 100% in patients with signs of local and systemic septic complications, patients with infected necrosis must undergo surgical intervention, which consists of an organ-preserving necrosectomy combined with a postoperative closed lavage concept that maximizes further evacuation of infected debris and exudate. However, intensive care treatment, including prophylactic antibiotics, reduces the infection rate and delays the need for surgery in most patients until the third or fourth week after the onset of symptoms. At that time, debridement of necrosis is technically easier to perform, due to better demarcation between viable and necrotic tissue compared with necrosectomy earlier in the disease. In contrast, surgery is rarely needed in the presence of sterile pancreatic necrosis. In those patients the conservative approach is supported by the present data. Received: March 20, 2002 / Accepted: April 15, 2002 Offprint requests to: W. Uhl     

Emergence of antibiotic resistance in infected pancreatic necrosis

BACKGROUND: Overall, the use of antibiotics in the treatment of patients with severe acute pancreatitis has increased owing to the use of antibiotic prophylaxis. HYPOTHESIS: The incidence of antibiotic-resistant (AB-R) bacteria in infected pancreatitis is related to prolonged antibiotic treatment and may affect outcome. DESIGN: Case series. SETTING: Fifty-six-bed intensive care unit of a tertiary care center. PATIENTS: Forty-six consecutive patients with infected pancreatic necrosis. MAIN OUTCOME MEASURES: Occurrence rate of AB-R organisms in pancreatic infection, overall duration of antibiotic treatment prior to infection, and mortality, defined as inhospital mortality. RESULTS: Infection with AB-R microorganisms was found in 24 (52%) of 46 patients. Primary infection was present in 7 patients; in 21 patients, nosocomial surinfection with AB-R organisms occurred. Patients with AB-R infections were treated with antibiotics for a longer period (24 vs 15 days, P<.05), while disease severity and the incidence of organ failure were not statistically significantly different. The intensive care unit stay was significantly longer in patients with AB-R infections (23 vs 31 days, P = .02). Mortality was not statistically significantly different in patients with AB-R infections (37% vs 28%, P = .23). CONCLUSIONS: The occurrence rate of infections with AB-R organisms in our patients with severe acute pancreatitis was high and was associated with a longer intensive care unit stay, but no increased mortality could be demonstrated. The duration of antibiotic treatment was increased in patients in whom AB-R infections developed.     

急性坏死性胰腺炎211例治疗经验

为确立急性坏死性胰腺炎（ＡＮＰ）的治疗方案，作者对比研究了以早期手术为主的Ａ组（１９９２年以前的１２２例）和以早期非手术治疗为主的Ｂ组（１９９２年以后７５例，１９９２年以前１４例，共计８９例）患者的并发症发生率和死亡率。结果Ｂ组并发症（ＡＲＤＳ、肾衰、心衰）的发生率和死亡率明显低于Ａ组（Ｐ＜０．０１）。对ＡＮＰ的治疗采取早期非手术治疗明显优于早期手术治疗。作者讨论了非手术治疗的方法、非手术治疗期中转手术的指征、及ＡＮＰ的治疗方案。     

保留导管区域灌注在治疗重症脂源性胰腺炎中的应用

【摘要】〓目的〓评价经供胰腺血管留置导管持续灌注疗法(CRAI)治疗重症脂源性胰腺炎的疗效。方法〓选择2013年1月至2016年5月期间纳入的43例患者重症脂源性胰腺炎患者,包括区域灌注组(23例)和常规治疗组(20例),分析对比白细胞计数﹑血清淀粉酶﹑尿淀粉酶﹑APACHE-Ⅱ、血糖值﹑血钙值变化情况;比较两组患者腹痛缓解时间﹑血清淀粉酶恢复时间﹑总住院天数,总治疗费用;经不同治疗后,两组患者胰腺及胰周感染胰腺假性囊肿的发生率;对两组临床效果效率进行评价。结果〓血清淀粉酶、尿淀粉酶、血糖、血钙水平以及APACHE-Ⅱ评分第六天、第十天的恢复结果,区域灌注组优于常规组(P<0.05),血淀粉酶恢复时间7.2±1.6 d,区域灌注组优于常规组,腹痛缓解时间3.1±1.3 d,区域灌注组优于常规组(P<0.05)。接受区域灌注治疗的患者的平均住院时间明显缩短,腹膜后感染、胰腺假性囊肿、行胰肠内引流手术例数明显低于常规组,总临床效果好于常规组(P<0.05)。结论〓区域灌注在治疗重症急性胰腺炎方面有其优势,可结合其他手段作为针对坏死性胰腺炎的联合治疗方案。     

Prophylactic antibiotics alter the bacteriology of infected necrosis in severe acute pancreatitis

BACKGROUND: Use of appropriate prophylactic antibiotics has been shown to decrease infectious complications and mortality rate in patients with severe acute pancreatitis, but its influence on the bacteriology of secondary pancreatic infection is poorly defined. STUDY DESIGN: Operative cultures from 61 consecutive patients with pancreatic necrosis treated during routine prophylactic antibiotic use (1993-2001) were compared with 34 consecutive patients with necrosis treated before routine antibiotic use (1977-1992). RESULTS: The two groups of patients were similar in demographics, etiology of pancreatitis, and severity of illness. All patients in the antibiotic group received prophylactic antibiotics compared with only 38% (13 of 34) in the control group. Routine broad-spectrum prophylactic antibiotics altered the bacteriology of secondary pancreatic infection in severe acute pancreatitis from predominantly gram-negative coliforms (56% versus 26%, p = 0.005) to predominately gram-positive organisms (23% versus 52%, p = 0.009) without a significant increase in either the rate of beta-lactam resistance or fungal infections. The overall hospital stay in patients treated with prophylactic antibiotics was significantly reduced (61 +/- 24 days versus 41 +/- 28 days, p = 0.002), and there was a trend toward a decline in mortality rate in the antibiotic treatment group. CONCLUSION: Routine broad-spectrum prophylactic antibiotic use has altered the bacteriology of secondary pancreatic infection in severe acute pancreatitis from predominantly gram-negative coliforms to predominantly gram-positive organisms without altering the rate of beta-lactam resistance or fungal superinfection.     

Percutaneous ultrasound-guided drainage in the surgical treatment of acute severe pancreatitis

AIM OF THE STUDY: To report results of percutaneous ultrasound-guided drainage, performed by a surgeon, in the treatment of complications of acute pancreatitis (AP), and to determine the role of this technique in the therapeutic armamentarium of severe AP. PATIENTS AND METHODS: From 1986 to 2001, 59 patients were included in this retrospective study. All patients initially had severe necrotizing AP (mean Ranson score = 4.1 ; range : 2-7). Anatomical lesions included pancreatic abscess in 6 patients and necrosis in 53 (17 stage D and 36 stage E according to Balthazar's classification). Necrosis was infected in 42 and sterile in 11 respectively. Drainage was performed under ultrasound guidance and local anaesthesia using small-diameter drains (7-14 French). RESULTS: Drainage was performed on average 23 days after onset of AP. Infection was proven by fine-needle aspiration in 47 (80 %) patients (41 infected necrosis and 6 localized abscess). In one patient, culture of aspirated fluid was negative but necrosis was infected (one false negative). Culture of aspirated fluid was negative and necrosis was sterile in 11 patients. Nineteen (32%) patients healed without subsequent surgery: 7 (16%) in the infected necrosis group, 6(55%) in the sterile necrosis group, and 6 (100%) in the abscess group. Forty (68%) patients had subsequent necrosectomy including 8 (14%) who died. Twenty (34 %) digestive fistulas healed spontaneously, except one treated by diversion stomia. Of the 16 (27 %) pancreatic fistulas, 6 needed subsequent interventional treatment. CONCLUSION: In selected patients, percutaneous drainage can represent an alternative to surgery with a 14% mortality rate. The high rate of subsequent necrosectomy suggests that drains with larger diameter, possibly associated with continuous irrigation, should be used.