Acute pancreatitis (AP) is a potentially lethal disease with

Transcription

1 CLINICAL GASTROENTEROLOGY AND HEPATOLOGY 2010;8: Primary Conservative Treatment Results in Mortality Comparable to Surgery in Patients With Infected Pancreatic Necrosis PRAMOD KUMAR GARG,* MANIK SHARMA,* KAUSHAL MADAN,* PEUSH SAHNI, DEBABRATA BANERJEE,* and ROHIT GOYAL* Departments of *Gastroenterology and Gastrointestinal Surgery, All India Institute of Medical Sciences, New Delhi, India See related article, Gluck M et al, on page 1083 in this issue of CGH; see Editorial on page BACKGROUND & AIMS: The standard treatment for patients with infected pancreatic necrosis (IPN) is surgical necrosectomy. We compared the outcomes of surgical treatment versus primary conservative treatment (patients kept in intensive care unit and treated with antibiotics, organ support, intensive nutritional support, and, if required, percutaneous drainage) among patients with IPN. METHODS: We performed retrospective comparative (with prospectively acquired database) and prospective observational studies; data were collected from all consecutive patients with acute pancreatitis (n 804), and those with IPN formed the study group. Patients with IPN were divided into 2 groups on the basis of diagnosis of IPN during (group 1, n 30) or (group 2, n 50). Eighteen patients in group 1 were treated by surgical necrosectomy, and 40 patients in group 2 were given primary conservative treatment; surgery was performed on patients if conservative treatment failed (n 10). The primary outcome measure was mortality. RESULTS: The mortality was comparable in group 1 versus group 2 (43% vs 28%; P.22). During a period of 10 years, the patients who received primary conservative treatment had significantly higher survival rates than those who received surgery (76.9% vs 46.4%; P.005). In the prospective study during , the mortality from infected necrosis was 29.6% after primary conservative treatment, confirming the results of the comparative study. CONCLU- SIONS: In treating patients with IPN, a primary conservative strategy resulted in mortality that was comparable with that after surgery, and 76% of the patients were able to avoid surgery; 54.5% of IPN patients were successfully managed with the primary conservative strategy. Keywords: Acute Pancreatitis; Infected Pancreatic Necrosis; Medical Therapy; Necrosectomy. View this article s video abstract at Acute pancreatitis (AP) is a potentially lethal disease with considerable morbidity and up to 40% mortality. 1 Two major forms of AP occur, interstitial and necrotizing pancreatitis. Necrotizing pancreatitis usually runs a severe course and is the cause of most of the morbidity and mortality. Although patients with sterile pancreatic necrosis might have severe course and die, infection of the nonviable necrotic pancreatic tissue is an ominous development during the course of AP. 2 We and others have shown that the extent and infection of pancreatic necrosis correlate with the development of organ failure and mortality in AP. 3,4 Although conservative treatment is recommended for sterile necrosis, 5 surgical necrosectomy is the standard of care for infected pancreatic necrosis (IPN) according to various practice guidelines. 6,7 A few case reports and case series have shown that IPN could be treated with conservative treatment There has not been any trial comparing surgery with conservative treatment for infected necrosis; a randomized trial was probably not feasible, given the overwhelming evidence in favor of surgical treatment in the published literature. We treated patients with IPN primarily with surgical treatment as per the standard guidelines. However, our management policy changed after 2002 after we observed that many patients with IPN improved with conservative treatment. From 2003 onwards, we treat patients with IPN primarily with medical conservative treatment, and surgery is performed only in those patients who fail to respond to conservative treatment. The data of all consecutive patients with AP have been recorded prospectively and meticulously since 1997 as a part of many prospective studies, some of which have been published. 4,11 14 The objective of the present study was to compare the mortality of IPN in the era of open necrosectomy versus conservative-first approach during 2 time periods during 10 years, ie, (primary surgical treatment) and (primary conservative treatment). We further evaluated prospectively the strategy of primary conservative treatment for IPN from January 2007 to December Methods The study design was a comparative study. The setting was a tertiary care academic center. Patients All consecutive patients with AP admitted to our hospital were included in the study. The diagnosis of AP was made in the presence of suggestive clinical features, increased serum amylase level ( 3 times the upper limit of normal), and evidence of AP on imaging studies, ie, abdominal ultrasonography Abbreviations used in this paper: AP, acute pancreatitis; APACHE, Acute Physiology, Age, and Chronic Health Evaluation; CT, computed tomography; CECT, contrast-enhanced computed tomography; FNA, fine-needle aspiration; ICU, intensive care unit; IPN, infected pancreatic necrosis; SN, sterile necrosis by the AGA Institute /$36.00 doi: /j.cgh

2 1090 GARG ET AL CLINICAL GASTROENTEROLOGY AND HEPATOLOGY Vol. 8, No. 12 and/or computed tomography (CT) scan. Patients with IPN formed the study group. Patients admitted from 1997 to 2006 were divided into 2 groups depending on the time of their admission to the hospital. Group 1 included patients admitted from January 1997 to December 2002, and group 2 included patients from January 2003 to December From January 2007 to December 2008, patients were evaluated prospectively to validate the results of primary conservative treatment for IPN. Exclusion Criteria Patients with acute exacerbation of chronic pancreatitis, those admitted with late complications of AP such as pancreatic pseudocyst, and those with pancreatic malignancy were excluded. Characterization of Acute Pancreatitis AP was categorized as either interstitial or necrotizing. Pancreatic necrosis was diagnosed as nonenhancing (nonviable) areas of the pancreas on a contrast-enhanced CT (CECT) scan. Pancreatic necrosis was classified as either sterile or infected. IPN was suspected if there was evidence of sepsis in the form of development of fever of 38 C with features of sepsis and deterioration or no improvement in their clinical condition. 4 The diagnosis of IPN was made when pancreatic necrotic tissue showed presence of bacteria on Gram stain or when it grew an organism on culture as described previously 7 (Supplementary Materials and Methods). Severity of Acute Pancreatitis (Supplementary Materials and Methods) Management of Acute Pancreatitis Supportive management. All patients were managed according to a predefined management protocol as described previously by us 4,11 14 (Supplementary Materials and Methods). Management of Pancreatic Necrosis According to the Study Protocol Sterile necrosis. Patients with sterile necrosis (SN) in both groups were treated with intensive conservative therapy with organ support in an intensive care unit (ICU). If a patient continued to deteriorate despite maximum supportive treatment, surgery was offered. Infected Pancreatic Necrosis Management of infected pancreatic necrosis in group 1 (from January 1997 December 2002). Patients were treated primarily with surgical necrosectomy, lavage, and drainage unless they refused surgery or were considered poor surgical risk precluding a major surgical procedure or if they showed unequivocal improvement on conservative management and antibiotics. The initial surgical treatment included debridement (necrosectomy) and if required (eg, intraoperative bleeding necessitating packing or inadequate necrosectomy), planned re-exploration(s) after 48 hours. When intraoperative assessment was considered satisfactory regarding hemostasis/ necrosectomy, the abdomen was closed, multiple drains were placed, and perioperative lavage was carried out. Management of infected pancreatic necrosis in group 2 (from January 2003 December 2006). The management protocol was changed during this period from primary surgery to primary conservative medical treatment, ie, all patients with IPN were treated primarily with an aggressive medical management in an ICU that included combination antibiotics, organ support, intensive nutritional support, and percutaneous drainage if required. Percutaneous drainage was done for IPN that had become organized and walled off. 15 Percutaneous drainage was achieved by placing single or multiple 12F 16F single pigtail/malecot catheters under ultrasonographic or CT guidance. The catheters were flushed with ml of normal saline 3 4 times a day to keep them patent. If clinical improvement was noted, the patient was continued on conservative treatment, and antibiotics were given for a period of 4 weeks. But if there was no improvement or if there was deterioration in the patient s condition, then the patient was subjected to surgery. The clinical criteria of improvement were defined a priori, ie, decline in the signs of sepsis (response in fever and decrease in leukocyte count), decrease in Acute Physiology, Age, and Chronic Health Evaluation (APACHE) II score by 2 points, and improving organ failure at least by 1 grade. Management protocol for patients during January 2007 December The management of patients during this period was similar to that for patients in group 2 to confirm and validate the results of primary conservative treatment for IPN. Primary Outcome Measure The difference in the mortality between the 2 groups of patients with IPN was the primary outcome measure. Predictive factors for mortality in patients with infected pancreatic necrosis. Survival analysis with Cox proportional hazard model was performed for predictive factors for mortality, taking into account all patients with IPN in the comparative trial from The following variables were analyzed: age, etiology of AP, interval between onset of disease and admission, admission values of hemoglobin, blood sugar, serum levels of bilirubin, calcium, albumin, and creatinine, extent of necrosis ( 50%, 50%), CT severity score, APACHE II score, organ failure, comorbid conditions, and treatment, ie, surgical or medical. Data Collection The data of patients were collected prospectively and recorded meticulously in a predesigned case record form. The data have been electronically maintained. The data were then anonymized (depersonalized by removing individual patients identifiers) because consent could not be obtained from the patients specifically for the present study. Ethical Clearance Our institutional ethics committee approved the study. Statistical Analysis Intergroup comparisons were done with independent Student t test and 2 test for quantitative and qualitative data, respectively. Survival analysis with Cox proportional hazard model was performed for predictive factors (as described above) for mortality. Variables found significant on univariate analysis

3 December 2010 CONSERVATIVE THERAPY OR SURGERY FOR IPN 1091 Table 2. Bacterial Organisms in Patients With IPN Organism N Escherichia coli 16 Pseudomonas 13 Klebsiella 12 Enterobacter 8 Acinetobacter 2 Proteus 1 Staphylococcus aureus 4 Fungal 13 Polymicrobial 8 N, number of patients. Results A total of 804 consecutive patients were included in the study during the 12-year period. Of them, 683 patients with acute pancreatitis were included in the comparative study from January 1997 to December 2006 (Figure 1). Figure 1. Categorization of patients in the comparative study. with a P value of.1 were entered into a stepwise Cox regression model. Bonferroni correction was applied for multiple comparisons. A P value of.05 was regarded as significant. Table 1. Baseline Characteristics of Patients Group 1 Group 2 P value N Age (y) NS Sex (male/female) 192/ /149 NS Median interval between 2 4 NS onset and admission (d) Mild/severe pancreatitis 190/ /176 NS Etiology Gallstone (%) 135 (43.3) 189 (51.0) NS Alcohol (%) 59 (18.9) 64 (17.2) NS Others (%) 118 (37.8) 118 (31.8) NS CT Severity Index NS APACHE II score NS (admission) Organ failure 65 (20.8%) 117 (31.5%).003 (admission) SN IPN Early severe AP Pancreatic abscess Mortality, n (%) 41 (13.1) 67 (18.1) NS NS, not significant. Comparison of Baseline Data in Group 1 ( ) and Group 2 ( ) There were 312 patients in group 1 and 371 patients in group 2. The 2 groups were comparable with regard to the age and gender of patients, etiology of pancreatitis, median interval between onset of pancreatitis and admission, CT severity index, APACHE II scores at admission, and presence of organ failure (Table 1). Infected Pancreatic Necrosis There were 203 patients with necrotizing pancreatitis in group 1 and group 2; of them, 123 had SN, and 80 had IPN. SN was present in 51 patients in group 1 and 72 patients in group 2. Of the 80 patients with IPN, 30 were in group 1, and 50 were in group 2. Microbiologic confirmation was present in 69 patients (Table 2), and in 11 patients the diagnosis of IPN was made on the basis of presence of air in the pancreatic bed. Patients with IPN were similar between the 2 groups (Table 3). Management of Pancreatic Necrosis in Groups 1 and 2 Of the 51 patients with SN in group 1, 48 underwent medical management, and 3 required surgery. Of the 30 patients with IPN in group 1, 18 patients underwent surgery after a median duration of 25 days (range, 9 48 days), whereas 12 patients were managed conservatively including one who required percutaneous drainage. The reasons for conservative management were multiple comorbid problems in 3 patients precluding surgery, acute coronary events during hospital stay in 2 patients, and unequivocal improvement with medical management in 7 patients. In group 2, of the 72 patients with SN, 67 were managed conservatively (3 of them having required percutaneous drainage for fluid collections), and 5 required surgery. Of the 50 patients with IPN, 40 were managed conservatively, with 24 requiring additional percutaneous drainage. Only 10 patients with IPN required surgery in group 2 after a median period of 46.5 days (range, days). Of these 10 patients, 4 had failed conservative treatment, 2 had complications of pigtail catheter drainage (pericatheter leakage and no improvement, 1 patient; colonic leak, 1 patient), 2 had persis-

4 1092 GARG ET AL CLINICAL GASTROENTEROLOGY AND HEPATOLOGY Vol. 8, No. 12 Table 3. Comparison of Patients With IPN in the 2 Groups Group 1 Group 2 P value N Age, y (mean standard deviation) NS Male/female 19/11 33/17 NS Etiology Gallstone (%) 16 (53.3) 32 (64) NS Alcohol (%) 3 (10) 7 (14) NS Others (%) 11 (36.7) 11 (22) NS Interval between onset and NS admission (d) Median Interquartile range APACHE II (admission) NS Organ failure (at 14 (46.7%) 28 (56%) NS admission) 50% Necrosis 16 (53.3%) 30 (60%) NS Hospital stay (median) NS (range) NS, not significant. tent sepsis with hypotension, and 2 had gastrointestinal bleeding with failed endoscopic and angiographic therapy. Outcome of Patients According to the Type of Treatment for Pancreatic Necrosis Patients with infected pancreatic necrosis. In group 1, of the 18 patients with IPN who underwent surgery, 12 (66%) died, whereas only 1 of the 12 patients (8%) treated conservatively died. In group 2, of the 10 patients treated surgically, 3 (33%) died, and of the 40 patients treated conservatively, 11 (27.5%) died (Table 4). In group 1, the causes of mortality in surgically treated patients were various postoperative complications in 8 patients and worsening of previous organ failure after surgery in 4 patients. The only death in the medically managed patient was due to multiorgan failure. In group 2, the 3 patients who died after surgery had postoperative complications in 2 patients and continued bleeding in the abdominal cavity in 1 patient. Among the 11 medically treated patients who died in this group, 9 (81%) died as a result of persistent organ failure, 1 patient with morbid obesity died of pulmonary thromboembolism, and the remaining patient died as a result of operative wound dehiscence and sepsis (pancreatitis was detected during exploratory laparotomy for acute abdomen). there was a lower mortality in group 2 compared with that in group 1 (28% vs 43.3%), although the difference was not statistically significant (P.22). However, of the total 80 patients with IPN during the 10-year period, there was a significantly higher survival among patients treated medically compared with those treated surgically (40/52 [77%] vs 13/28 [46%]; P.005). On intention-to-treat analysis when patients who underwent drainage or surgery for suspected infected necrosis but were found to have negative cultures were included as infected necrosis, there was no appreciable difference in the results; the mortality of patients with infected necrosis was 14 of 33 (42.42%) in group 1 and 19 of 58 (32.75%) in group 2. Patients with sterile necrosis. The mortality of patients with SN on the basis of the type of treatment received in the 2 groups is given in Table 4. Outcome of Patients During January 2007 December 2008 Of the 121 patients with AP during this period, 23 had SN, and 27 had IPN. Of the 23 patients with SN, all but one were managed conservatively. Of the 27 patients with IPN, 19 were managed medically (11 requiring percutaneous drainage as well), and 8 required surgery. Of the 27 patients, 8 patients died, 2 after surgery and 6 after conservative treatment. Thus, the mortality in patients with IPN was 29.8%, which was similar to the mortality (28%) in patients with IPN managed with the same strategy during , thus confirming the results of conservative management for IPN (Table 4). surgery could be avoided in 59 of 77 (76.6%) patients with IPN, and 42 of 77 (54.5%) patients could be managed successfully with primary conservative therapy from January 2003 December Predictive Factors for Mortality in Patients With Infected Pancreatic Necrosis Because organ failure might be a confounding variable, mortality was analyzed in respect of organ failure. On comparing patients with single organ failure and those with multisystem organ failure, there was no survival difference (17/33 vs 3/9), but the mortality was lower in 28 patients without organ failure (4 died; 3 of the 9 who were operated died, and only 1 of the 19 treated conservatively died). Serum creatinine, APACHE II score, surgical management, and comorbid condition were predictors of mortality on univariate analysis. However, APACHE II score and serum creatinine were found to be 2 independent predictors of survival on Cox proportional hazard model (Table 5). Table 4. Outcome of Patients on the Basis of the Type of Management of Pancreatic Necrosis (group 1) (group 2) IPN Treatment Medical, N 12 N 18 N 30 Medical, N 40 N 10 N 50 Medical, N 19 N 8 N 27 Mortality (%) 1 (8) 12 (66.6) 13 (43.3) a 11 (27.5) 3 (33.3) 14 (28) a 6 (31.5) 2 (25) 8 (29.6) SN Treatment Medical, N 48 N 3 N 51 Medical, N 67 N 5 N 72 Medical, N 22 N 1 N 23 Mortality (%) 3 (6.2) 1 (33.3) 4 (7) 12 (17.9) 4 (80) 16 (22.2) 6 (27.2) 0 6 (26) a P.22.

5 December 2010 CONSERVATIVE THERAPY OR SURGERY FOR IPN 1093 Table 5. Results of Survival Analysis With Cox Proportional Hazard Model for Predictors of Mortality Factors Hazard ratio P value 95% Confidence interval APACHE II score Serum creatinine Discussion IPN accounts for most of the mortality in patients with AP. 1,4 The standard and accepted treatment for patients with IPN has been surgical necrosectomy. 6,7 There has been no randomized comparative trial between conservative and surgical therapy in patients with IPN, primarily because conservative management was never considered a viable treatment option. The present study, which compares conservative and surgical therapies in patients with IPN, has shown that a management strategy of primary conservative treatment, ie, conservative first approach, resulted in mortality comparable to surgery. Although the mortality was slightly lower in the conservative group, this was not statistically significant. A type II, ie,, error could be the reason for statistical insignificance. the mortality in patients treated conservatively was significantly lower than that in those treated surgically in both groups combined, and 77% of patients with IPN who were treated conservatively had a successful outcome. There were certain limitations of our study. (1) The most important was that it was not a randomized trial. However, in the absence of convincing data about the success of conservative treatment for IPN, such a randomized trial would have been considered unethical, and thus, perhaps was not possible. Even though ours is not a randomized trial, the patient characteristics were comparable in the 2 groups, thus minimizing selection bias. (2) It was a retrospective study. However, the data were acquired prospectively as a part of many ongoing studies. 4,11 14 Furthermore, we confirmed the results of the primary conservative treatment prospectively as well. (3) The mortality in patients treated surgically in the primary surgery group was high. This was possibly due to several adverse factors that included presence of preoperative organ failure, early surgery, and sicker patients in our study as a result of a referral bias; ours is a tertiary care center. (4) The diagnosis of IPN was made on the basis of air in the pancreatic bed in 11 of 80 patients. Although air in the pancreatic bed might be due to small bowel fistulization, it is generally taken as indicative of IPN. 1 We diagnosed IPN in the presence of air in the pancreatic bed when there was a strong suspicion, ie, fever, leukocytosis, and worsening disease. The results of medical conservative treatment for IPN in earlier series have been reported to be dismal. 6,16 One earlier study had in fact shown that conservative management of patients with IPN and organ failure was associated with 100% mortality. 17 However, a few anecdotal reports have shown that patients with IPN could be treated successfully with conservative treatment alone The results of the present study of successful outcome of IPN with conservative treatment are consistent with 2 recent studies. Runzi et al 9 showed that 16 of 28 patients with IPN could be treated medically with 12% mortality; only 3 patients required percutaneous drainage. Lee et al 10 showed that 31 patients with IPN could be treated medically with 3.2% mortality; 23 patients required percutaneous or endoscopic drainage, and 8 recovered with antibiotics alone. There has been an overwhelming surgical bias in treating patients with IPN. The strongest argument in favor of surgical treatment for IPN is the basic surgical principle of removing nonviable infected tissue. However, the surgery is marred by many problems, such as intraoperative tissue damage and bleeding, worsening organ failure as a result of surgical trauma, multiple operations, stormy postoperative course, lengthy stay in ICU, and a high mortality of up to 75%, particularly if operated early. 18,19 In fact, surgical treatment was found to be a predictor of mortality on univariate analysis in the present study. The issue of medical versus surgical treatment for IPN has been intensely debated of late. 20 The recent reports of patients with IPN having been treated successfully with a combination of conservative treatment and percutaneous drainage support the contention that surgical treatment might not always be required. 21 Management of patients with SN has taken a turnaround from primary surgical to conservative approach. 5,6 It is quite possible that many patients with presumed SN might have had IPN because infection was not ruled out in most of them by fine-needle aspiration (FNA). Multiple factors possibly contributed to the success of conservative management of patients with IPN in the present study. These include full organ support, the use of sensitive antibiotics including third-generation cephalosporins with -lactamase inhibitors and carbapenems, aggressive nutritional support, and judicious percutaneous intervention. It is quite possible that in patients with infected necrosis, the initial inflammation might subside over time as a result of the body s own healing and reparative processes, similar to what happens in patients with SN. Once the infected necrosis becomes organized and better defined, it can be drained subsequently with less invasive methods if required. It has been shown that early surgery results in high mortality in patients with IPN. 19,22 Most surgeons thus prefer to operate on patients with IPN beyond fourth week of the illness, so that the necrosis becomes organized and surgical planes are well-defined. 22 In the present study also, there was a difference between the timing of surgery in group 1 and group 2 (25 and 46 days) that might have contributed to the better results in group 2. The key element is to support the patient in the interim with organ assist devices until such time that the systemic inflammation ameliorates and organ failure improves. It is quite possible that many such patients with IPN might improve on conservative treatment alone and require percutaneous drainage of an organized necrosis/fluid/pus collection later. 20 In that case, surgery will be truly indicated only in those remaining patients who do not improve with such conservative approach. The method of surgical treatment could be either open or laparoscopic technique, depending on the expertise available. Recent studies have shown feasibility and success of endoscopic transgastric necrosectomy and combined endoscopic and percutaneous necrosectomy in selected patients with infected necrosis. 23,24 The present study has substantiated the emerging concept of primary conservative treatment in patients with IPN but in no way undermines the importance of surgery, preferably minimally invasive, in those patients who really merit surgical necrosectomy. Thus, we conclude that a primary conservative strategy resulted in mortality that was comparable with that after surgery,

6 1094 GARG ET AL CLINICAL GASTROENTEROLOGY AND HEPATOLOGY Vol. 8, No. 12 and 76% of patients were able to avoid surgery; 54.5% of IPN patients were successfully managed with the primary conservative strategy. Patients with infected necrosis treated medically had a better survival than those treated surgically. Supplementary Material Note: To access the supplementary material accompanying this article, visit the online version of Clinical Gastroenterology and Hepatology at and at doi: / j.cgh References 1. Whitcomb DC. Clinical practice: acute pancreatitis. N Engl J Med 2006;354: Bradley EL 3rd. A clinically based classification system for acute pancreatitis: summary of the International Symposium on Acute Pancreatitis, Atlanta, GA, September 11 through 13, Arch Surg 1993;128: Isenmann R, Rau B, Beger HG. Bacterial infection and extent of necrosis are determinants of organ failure in patients with acute necrotizing pancreatitis. Br J Surg 1999;86: Garg PK, Madan K, Pande GK, et al. Association of extent and infection of pancreatic necrosis with organ failure and death in acute necrotizing pancreatitis. Clin Gastroenterol Hepatol 2005; 3: Werner J, Feuerbach S, Uhl W, et al. Management of acute pancreatitis: from surgery to interventional intensive care. Gut 2005;54: Uhl W, Warshaw A, Imrie C, et al. IAP guidelines for the surgical management of acute pancreatitis. Pancreatology 2002;2: UK Working Party on Acute Pancreatitis. UK guidelines for the management of acute pancreatitis. Gut 2005;54(Suppl III):iii1 iii9. 8. Adler DG, Chari ST, Dahl TJ, et al. Conservative management of infected necrosis complicating severe acute pancreatitis. Am J Gastroenterol 2003;98: Runzi M, Niebel W, Goebell H, et al. Severe acute pancreatitis: nonsurgical treatment of infected necroses. Pancreas 2005;30: Lee JK, Kwak KK, Park JK, et al. The efficacy of nonsurgical treatment of infected pancreatic necrosis. Pancreas 2007;8: Garg PK, Khanna S, Bohidar NP, et al. Incidence, spectrum, and antibiotic sensitivity pattern of bacterial infections among patients with acute pancreatitis. J Gastroenterol Hepatol 2001;16: Bohidar NP, Garg PK, Khanna S, et al. Incidence, etiology and impact of fever in patients with acute pancreatitis. Pancreatology 2003;3: Sharma M, Banerjee D, Garg PK. Characterization of newer subgroups of fulminant and subfulminant pancreatitis associated with a high early mortality. Am J Gastroenterol 2007;102: Sharma PK, Madan K, Garg PK. Hemorrhage in acute pancreatitis: should gastrointestinal bleeding be considered an organ failure? Pancreas 2008;36: Baron TH, Morgan DE, Vickers SM, et al. Organized pancreatic necrosis: endoscopic, radiologic, and pathologic features of a distinct clinical entity. Pancreas 1999;19: Rotman N, Methieu D, Anglade MC, et al. Failure of percutaneous drainage of pancreatic abscesses complicating severe acute pancreatitis. Surg Gynecol Obstet 1992;174: Widdison AL, Karanjia ND. Pancreatic infection complicating acute pancreatitis. Br J Surg 1993;80: Connor S, Alexakis N, Raraty MG, et al. Early and late complications after pancreatic necrosectomy. Surgery 2005;137: Basslink MG, Verwer TJ, Schoenmaeckers EJ, et al. Timing of surgical intervention in necrotizing pancreatitis. Arch Surg 2007; 142: Connor S, Raraty MG, Neoptolemos JP, et al. Does infected pancreatic necrosis require immediate or emergency debridement? Pancreas 2006;33: Baril NB, Ralls PW, Wren SM, et al. Does an infected peripancreatic fluid collection or abscess mandate operation? Ann Surg 2000;231: Mier J, Luque-de-Leon E, Castillo A, et al. Early versus late necrosectomy in severe necrotizing pancreatitis. Am J Surg 1997:173: Seifert H, Biermer M, Schmitt W, et al. Transluminal endoscopic necrosectomy after acute pancreatitis: a multicentre study with long-term follow-up (the GEPARD Study). Gut 2009;58: Ross A, Gluck M, Irani S, et al. Combined endoscopic and percutaneous drainage of organized pancreatic necrosis. Gastrointest Endosc 2010;71: Reprint requests Address requests for reprints to: Pramod Kumar Garg, Associate Professor, Department of Gastroenterology, All India Institute of Medical Sciences, New Delhi , India. pkgarg@aiims.ac.in; fax: (91) Acknowledgments The authors thank Dr V Sreeniwas, Department of Biostatistics, All India Institute of Medical Sciences, New Delhi, India, for his great help in this study. Conflicts of interest The authors disclose no conflicts.

7 1094.e1 GARG ET AL CLINICAL GASTROENTEROLOGY AND HEPATOLOGY Vol. 8, No. 12 Supplementary Materials and Methods Computed Tomography Scan in Acute Pancreatitis A dynamic CECT scan of the abdomen was done by using a helical CT machine with rapid acquisition of images. This was done in patients with AP usually at the end of first week of the onset of pancreatitis except in patients with clinically mild pancreatitis who improved with conservative treatment within the first week itself. CECT scan was also done at the time of admission if the diagnosis of AP was not certain. It was repeated during the course of the illness if the clinical condition of the patient so warranted, eg, assessing the CT signs of AP if the patient was deteriorating and looking for the development of complications such as infected necrosis, fluid collections, pseudocyst, and abscess, especially if any intervention such as drainage or surgical treatment was contemplated. CECT scan was not done in the presence of renal failure. The extent of pancreatic necrosis was graded as 30%, 30% 50%, and 50%. A CT severity score was calculated according to Balthazar et al. 1 An experienced consultant radiologist interpreted and reported the CT findings. Diagnosis of Infected Pancreatic Necrosis The diagnosis of IPN was made when pancreatic necrotic tissue showed presence of bacteria on Gram stain or when it grew an organism on culture as described previously. It was cultured for aerobic and anaerobic bacteria and fungal organisms. Pancreatic tissue was obtained by an FNA done under ultrasonography guidance by using a 20-gauge to 22-gauge needle, taking all aseptic precautions. 2 The necrotic areas were localized with the help of CECT scan. The FNA was performed in the second or third week of the illness. FNA was repeated in the following week if the first FNA was negative and the signs of sepsis persisted. Necrotic tissue and fluid were also obtained for culture during percutaneous drainage or at surgery in patients who were operated on. In patients with suspected IPN, presence of extraintestinal gas in the pancreatic bed on a CT scan was taken as another evidence of infected necrosis. Other causes of sepsis were also looked for, eg, cholangitis, intravenous line sepsis, and chest infection; appropriate cultures of blood and fluids such as bile (if available), sputum, and urine and other material such as an intravenous catheter were obtained periodically. Organ Failure Organ failure was defined according to the Atlanta classification 3 as follows: respiratory failure (pao 2 60 mm), acute renal failure (serum creatinine 2.0 mg/dl), cardiovascular failure (systolic blood pressure 90 mm Hg), or severe gastrointestinal bleeding ( 500 ml/24 h). Severity of Acute Pancreatitis Severe AP was defined by 1 or more of the following criteria: (1) Presence of organ failure as per the Atlanta classification (2) CT severity: a CT score of 7 was indicative of severe pancreatitis. 1 (3) APACHE II score: an APACHE II score of 8 was taken as indicative of severe pancreatitis. 4 Early Severe Acute Pancreatitis Early severe AP was diagnosed if there was development of severe organ failure defined as either single grade 3-4 organ failure or 2 organ failures within 7 days of the onset of AP as reported by us. 5 Supportive Management of Acute Pancreatitis All patients were managed according to a predefined management protocol as described previously by us. 2, 5 8 They were treated conservatively with nil by mouth, analgesics, intravenous fluids, and supportive treatment. Antibiotics were prescribed if (1) patients had infected necrosis, (2) there was documented infection at extrapancreatic sites such as cholangitis, (3) patients had severe acute necrotizing pancreatitis with organ failure, and (4) patients had signs of sepsis in the form of fever and leukocytosis even in the absence of documented infection, provided the fever (temperature 38 C) persisted for more than 2 days. The antibiotics chosen were according to the culture and sensitivity report whenever available. In the absence of a sensitivity report, a combination of ceftazidime, ofloxacin, and metronidazole was used. Patients were treated with newer antibiotics with better coverage that included cefoperazone with sulbactam, piperacillin with tazobactum, and carbapenems if there was no response or if the culture sensitivity report so suggested. In patients with mild pancreatitis, oral feeding was resumed soon after abdominal pain subsided. Feeding was started with clear liquids. In patients with severe pancreatitis, a nasojejunal tube was placed for enteral feeding unless the patient had persistent ileus or active gastrointestinal bleeding, in which case parenteral nutrition was instituted. The calorie intake was increased gradually to Kcal/day. Patients with severe AP were treated in an ICU, and all possible organ support systems were used including ventilatory support, vasopressors, and dialysis as and when required. Patients with biliary obstruction, as evidenced by abnormal liver function tests and biliary ductal dilatation on ultrasonography, underwent endoscopic retrograde cholangiopancreatography, biliary decompression with insertion of a 7-F nasobiliary drainage tube, and extraction of any bile duct stone if detected. Further reading 1. Balthazar EJ, Robinson DL, Megibow AJ, et al. Acute pancreatitis: value of CT in establishing prognosis. Radiology 1990;174: Garg PK, Madan K, Pande GK, et al. Association of extent and infection of pancreatic necrosis with organ failure and death in acute necrotizing pancreatitis. Clin Gastroenterol Hepatol 2005; 3: Bradley EL 3rd. A clinically based classification system for acute pancreatitis: summary of the International Symposium on Acute Pancreatitis, Atlanta, GA, September 11 through 13, Arch Surg 1993;128: Wilson C, Health DI, Imrie CW. Prediction of outcome in acute pancreatitis: a comparative study of APACHE II, clinical assessment and multiple factor scoring systems. Br J Surg 1990;77: Sharma M, Banerjee D, Garg PK. Characterization of newer sub-

8 December 2010 CONSERVATIVE THERAPY OR SURGERY FOR IPN 1094.e2 groups of fulminant and subfulminant pancreatitis associated with a high early mortality. Am J Gastroenterol 2007;102: Garg PK, Khanna S, Bohidar NP, et al. Incidence, spectrum, and antibiotic sensitivity pattern of bacterial infections among patients with acute pancreatitis. J Gastroenterol Hepatol 2001;16: Bohidar NP, Garg PK, Khanna S, et al. Incidence, etiology and impact of fever in patients with acute pancreatitis. Pancreatology 2003;3: Sharma PK, Madan K, Garg PK. Hemorrhage in acute pancreatitis: should gastrointestinal bleeding be considered an organ failure? Pancreas 2008;36: