Biochemical prediction of gallstones in acute pancreatitis: a prospective study of three systems

The Etiologic Evaluation of Acute Pancreatitis in a General Hospital of Seoul-Gyeonggi Province in Korea

Background/Aims

In recent years, the incidence of acute pancreatitis (AP) has been increasing. A better understanding of the etiology is directly linked to more favorable outcomes. Unfortunately, there have been reports suggesting the variation of etiologies of AP across countries. The objective of this study was to determine the etiology of AP in a general hospital of Seoul-Gyeonggi province in Korea during the past decade.

Methods

We retrospectively reviewed the medical records of consecutive patients with AP who were admitted to St. Paul's Hospital (Seoul, Korea) with an affiliation to the Catholic University of Korea between January 2003 and January 2013.

Results

A total of 1,110 patients were enrolled, totaling 1,833 attacks, and the most frequent cause of AP was alcohol consumption. The recurrence rate of AP was 24.5% (272/1,110), and habitual recurrence rate (more than three times) was 12.6% (140/1,110). The rate of severe AP was 4.9% (90/1,833 attacks). The mortality rate of AP was 2.6% (29/1,110 patients). The frequency of an idiopathic cause of AP was 13.3%. The recurrence rate and mortality rate of idiopathic AP were 16.2% and 5.4%, respectively. In 41.7% (10/24) of cases of idiopathic AP, microlithiasis was suspected.

Conclusions

Between 2003 and 2013 in Korea, alcohol was the most frequent cause of AP in the general hospital of Seoul-Gyeonggi province of Korea. It appears that alcohol abstinence program may be necessary. Further nationwide studies would be needed to evaluate the etiologies of AP.

The predictivity of serum biochemical markers in acute biliary pancreatitis

Background and Aim. There are no accurate methods of differentiating acute biliary pancreatitis. Obstructions of biliary ducts, idiopathic pancreatitis may be related with biliary origin which needs identification for acute treatment. We searched for the predictivity of biochemical markers in early acute biliary pancreatitis. Patients and Methods. Serum levels of AST (Aspartate Transaminase),ALT (Alanine Transaminase), ALP (Alkaline Phosphatase), GGT (Gamma Glutamyl Transferase), total bilirubin, direct bilirubin, LDH (Lactate Dehydrogenase), amylase, lipase, CRP (C-Reactive Protein) and WBC (White Blood Cell) were measured in 157 patients with acute pancreatitis. Biliary and nonbiliary pancreatitis were differentiated by Magnetic Resonance Cholangiopancreatography (MRCP), Endoscopic Retrograde Cholangiopancreatography (ERCP), Intraoperative Cholangiopancreatography (IOC). Cut-off points of admission biochemical markers with sensitivity, specifity, positive predictive value and negative predictive value were determined after identification of significant variables. Receiver Operator Curves were plotted for each biochemical marker. Results. Serum Alkaline Phosphatase, total bilirubin, direct bilirubin, amylase and lipase levels were significantly higher in biliary pancreatitis with a positive predictive value of 80.8%, 83.9%, 81.6%, 78.8%, 79.7%. Conclusion. Increased Alkaline Phosphatase,total bilirubin, direct bilirubin, amylase and lipase levels may be used in prediction of biliary pancreatitis.

Etiology and diagnosis of acute biliary pancreatitis

Establishing a biliary etiology in acute pancreatitis is clinically important because of the potential need for invasive treatment, such as endoscopic retrograde cholangiopancreatography. The etiology of acute biliary pancreatitis (ABP) is multifactorial and complex. Passage of small gallbladder stones or biliary sludge through the ampulla of Vater seems to be important in the pathogenesis of ABP. Other factors, such as anatomical variations associated with an increased biliopancreatic reflux, bile and pancreatic juice exclusion from the duodenum, and genetic factors might contribute to the development of ABP. A diagnosis of a biliary etiology in acute pancreatitis is supported by both laboratory and imaging investigations. An increased serum level of alanine aminotransferase (>1.0 microkat/l) is associated with a high probability of gallstone pancreatitis (positive predictive value 80-90%). Confirmation of choledocholithiasis is most accurately obtained using endoscopic ultrasonography or magnetic resonance cholangiopancreatography. This Review discusses the pathogenesis of ABP and the clinical techniques used to predict and establish a biliary origin in patients with suspected ABP.

Alanine transaminase rather than abdominal ultrasound alone is an important investigation to justify cholecystectomy in patients presenting with acute pancreatitis

OBJECTIVES

The aims of this study were to investigate the predictive value of an elevated level of alanine transaminase (ALT) for biliary acute pancreatitis (AP) and to reconsider the role of abdominal ultrasound (AUS).

METHODS

All patients admitted to Christchurch Public Hospital with AP between July 2005 and December 2008 were identified from a prospectively collected database. Peak ALT within 48 h of presentation was recorded. Aetiology was determined on the basis of history, AUS and other relevant investigations.

RESULTS

A total of 543 patients met the inclusion criteria. Patients with biliary AP had significantly higher median (range) ALT than those with non-biliary causes (200 units/l [63-421 units/l] vs. 33 units/l [18-84 units/l]; P < 0.001). An ALT level of >300 units/l had a sensitivity of 36%, specificity of 94%, positive predictive value of 87% and positive likelihood ratio of 5.6 for gallstones. An elevated ALT and negative AUS had a probability of 21-80% for gallstones.

CONCLUSIONS

An elevated ALT strongly supports a diagnosis of gallstones in AP. Abdominal ultrasound effectively confirms this diagnosis; however, a negative ultrasound in the presence of a raised ALT does not exclude gallstones. In some patients consideration could be given to proceeding to laparoscopic cholecystectomy based on ALT alone.

Detection of gallstones in acute pancreatitis: when and how?

The detection and management of bile duct stones in acute pancreatitis have improved considerably. Now it is possible to identify non-invasively stones in the duct before definitive treatment. Recently new evidence has confirmed that the indication for early endoscopic sphincterotomy should be severe acute pancreatitis with evidence of bile duct obstruction. This review analyses the evidence that defines current best practice in this area. and IAP.

Clinico-biochemical prediction of biliary cause of acute pancreatitis in the era of endoscopic ultrasonography

BACKGROUND

Occult biliary stones escape detection on conventional investigations, and clinico-biochemical systems proposed for predicting biliary pancreatitis has low predictive values.

AIM

To evaluate the accuracy of clinico-biochemical parameters for prediction of biliary pancreatitis in patients undergoing endoscopic ultrasonography.

METHODS

Early endoscopic ultrasonography was performed on 139 patients presenting with acute pancreatitis within 24 h of admission. The aetiologies of all patients were determined after complete evaluations, and clinico-biochemical characteristics of patients with a biliary cause (biliary group) and non-biliary causes (non-biliary group) were compared.

RESULTS

Biliary pancreatitis was diagnosed in 107 patients and 32 patients had non-biliary causes. The biliary group belonged to a significantly older age group, had a female predominance, significantly more derangement of liver function and a higher incidence of severe attack of acute pancreatitis. On multivariate analysis, female sex, age >58 years and serum alanine aminotransferase >150 U/L were independent predictive factors for biliary cause of acute pancreatitis. Using these three factors for prediction of biliary cause, the sensitivity was 93% and overall accuracy was 85%.

CONCLUSION

Clinico-biochemical prediction for biliary cause of acute pancreatitis improves in the era of endoscopic ultrasonography with a higher sensitivity and overall accuracy. In centres where endoscopic ultrasonography is inaccessible or local expertise is unavailable, clinico-biochemical prediction of biliary cause of acute pancreatitis may provide a useful alternative in the initial management of this group of patients.

Clinical laboratory assessment of acute pancreatitis

BACKGROUND

Several biochemical markers in blood and urine have been investigated to establish their clinical application in patients with acute pancreatitis (AP). The relevant studies are reviewed and critically appraised.

METHODS

Medline and the World Wide Web were searched and the relevant literature was classified under the following categories: (1) diagnosis of AP and (2) prediction of: a) disease severity, b) pancreatic necrosis and its secondary infection, c) organ failure and death, and d) disease etiology.

RESULTS AND CONCLUSIONS

Serum lipase is a more reliable diagnostic marker of AP than serum amylase. Urinary strip tests for trypsinogen activation peptide (TAP) and trypsinogen-2 provide a reliable early diagnosis of AP. Useful predictors of severity may include serum procalcitonin and urinary TAP and trypsinogen-2 on admission, serum interleukins-6 and -8 and polymorphonuclear elastase at 24 h, and serum C-reactive protein (CRP) at 48 h. Other markers such as amyloid A and carboxypeptidase B activation peptide (CAPAP) need further investigation. Biochemical prediction of pancreatic necrosis requires 72 h to reach reliability and is impractical. However, the daily monitoring of serum procalcitonin provides a non-invasive detection of infected necrosis; the promising role of phospholipase A(2) in this regard requires further investigation. Early transient hypertransaminasemia reliably predicts biliary etiology, while serum carbohydrate-deficient transferrin and trypsin may predict an alcoholic etiology.

The biochemical detection of biliary etiology of acute pancreatitis on admission: a revisit in the modern era of biliary imaging

INTRODUCTION

Compared with traditional radiologic methods for the detection of cholelithiasis, early transient hypertransaminasemia had provided a useful prediction of biliary etiology in patients with acute pancreatitis.

AIM

To investigate whether this application remains valid in the modern era of imaging for microlithiasis.

METHODOLOGY

The biochemical detection (LFT) of cholelithiasis was based on an increase in serum alanine transaminase of >or=80 IU/L (normal range, 0-45 IU/L) within 24 hours of admission. We have taken the collective findings of abdominal ultrasound (USS), endoscopic ultrasound (EUS), and postmortem examination to represent the denominator for the diagnosis of cholelithiasis against which comparison with LFT was made.

RESULTS

Of 68 patients with acute pancreatitis who were treated between October 2000 and December 2001, cholelithiasis was the etiological factor in 44 patients (65%). EUS detected microlithiasis in 5 of 10 patients examined. The etiology remained idiopathic in 3 patients (4.4%). The sensitivity, specificity, and positive and negative predictive values for USS were 86%, 100%, 100%, and 80% respectively; for LFT, they were 91%, 100%, 100%, and 86%; and for USS and LFT combined, they were 98%, 100%, 100%, and 96%, respectively.

CONCLUSIONS

In patients with acute pancreatitis, the biochemical analysis within 24 hours of admission provided a simple, rapid, and more accurate prediction of cholelithiasis than USS. The combination of LFT and USS detected or excluded a biliary etiology in almost all patients.

EUS for detection of occult cholelithiasis in patients with idiopathic pancreatitis

BACKGROUND

The aim of this study was to evaluate the use of endoscopic ultrasonography (EUS) in detecting occult cholelithiasis and determining a probable etiology in patients classified as having idiopathic pancreatitis by conventional radiologic methods.

METHODS

A prospective study was performed in 89 consecutive patients with acute pancreatitis. Transcutaneous ultrasonography (US), CT, or both was performed on all patients within 24 hours of admission. Endoscopic retrograde cholangiopancreatography (ERCP) was performed in all patients with confirmed or suspected biliary pancreatitis. EUS was performed in patients classified as having idiopathic pancreatitis.

RESULTS

Cholelithiasis was identified in 64 patients (72%) by conventional radiologic methods. Eighteen patients (20%) were classified as having idiopathic pancreatitis after evaluation by US (all 18 patients), repeated US (9 patients), CT (6 patients) and ERCP (13 patients). EUS performed in these 18 patients revealed small gallbladder stones (1 to 9 mm) in 14 patients; 3 were found to have concomitant choledocholithiasis. All stones were confirmed by subsequent ERCP and cholecystectomy. The remaining 4 patients in whom no etiology was identified had no clinical or radiologic evidence of cholelithiasis at a median follow-up of 22 months.

CONCLUSION

Cholelithiasis is detected by EUS in a large number of patients classified as having idiopathic pancreatitis by conventional radiologic examinations. With identification of a biliary cause of acute pancreatitis, treatment can be initiated early, thereby reducing the risk of recurrent pancreatitis with its associated morbidity and mortality. EUS is a valuable diagnostic modality in the management of patients with acute pancreatitis.

Review article: the treatment of acute biliary pancreatitis

World-wide, gallstones are the commonest cause of acute pancreatitis (AP), a potentially life-threatening condition. Scoring systems based on radiologic findings and serologic markers help predict which patients may have retained bile duct stones. Endoscopic decompression of the biliary tree (by sphincterotomy and stone extraction during ERCP) is safer than surgery, which carries high morbidity and mortality in AP. Four prospective, randomized clinical trials confirm the utility of this approach, however careful patient selection is necessary. Medical therapy for AP is supportive; no agent or medical intervention has been shown consistently to abort or ameliorate the course of AP. Drugs that inhibit inflammatory cytokine production, however, are showing promise.

Usefulness of alanine and aspartate aminotransferases in the diagnosis of microlithiasis in idiopathic acute pancreatitis

CONCLUSION

Serum increases of aminotransferases, especially alanine aminotransferase (ALT), were suggestive of microlithiasis in idiopathic acute pancreatitis, particularly when assessed early after the onset of abdominal pain.

BACKGROUND

It has been shown that biochemical laboratory values only are useful parameters in distinguishing gallstone from nongallstone acute pancreatitis. We assessed the diagnostic usefulness of alanine aminotransferase (ALT) and aspartate aminotransferase (AST) for identification of occult microlithiasis in idiopathic acute pancreatitis.

METHODS

Ninety-one patients with idiopathic acute pancreatitis who underwent microscopic examination of stimulated duodenal bile sediments were retrospectively studied. According to earliness of ALT and AST assay after the onset of abdominal pain, patients were divided into two groups: group A, within the first 24 h (n = 56) and group B, between 24 and 72 h (n = 35).

RESULTS

ALT and AST values expressed as number of elevations of the upper limits of normal were higher in group A patients with positive biliary drainage than in group B. Median (range) ALT and AST values were 2.5 (0.1-18.1) vs 0.4 (0.1-8.6) and 3 (0.3-17.4) vs 0.5 (0.3-11.9), respectively. In the univariate analysis and receiver operating characteristic (ROC) curves, ALT within the first 24 h showed a sensitivity of 73%, specificity of 86%, and positive predictive value of 92% for a cutoff of 1.2 elevations of the upper limit of normal. These values were slightly higher, although without statistically significant differences, than those of AST (73, 80, and 89%, respectively).

Effect of endoscopic sphincterotomy and interval cholecystectomy on late outcome after gallstone pancreatitis

BACKGROUND

Endoscopic sphincterotomy alone, or followed by cholecystectomy, are options in patients with gallstone pancreatitis.

METHODS

Ninety-six patients of median age 74 (range 30-93) years with gallstone pancreatitis had endoscopic retrograde cholangiography and were followed for a median of 84 (range 33-168) months. Forty-eight of 49 patients with, and nine of 47 without, common bile duct (CBD) stones had urgent endoscopic sphincterotomy. One patient with, and six without, CBD stones had delayed endoscopic sphincterotomy a median of 35 (range 12-111) days after acute pancreatitis. Thus, 64 patients had endoscopic sphincterotomy (group 1) and 32 did not (group 2). Fifteen and 16 patients in each group respectively had interval cholecystectomy after a median of 3 months and 1 month.

RESULTS

Patients in groups 1 and 2 had similar rates of interval cholecystectomy (15 of 64 versus 16 of 32 patients respectively) or required cholecystectomy (15 of 49 versus five of 16 patients), recurrent CBD calculi (three of 64 versus three of 32 patients) or total length of hospitalization after interval cholecystectomy (median 15.5 and 15 days) or required (median 22 and 24 days) cholecystectomy. The overall incidence of recurrent pancreatitis was one of 64 patients in group 1 and five of 32 in group 2 (P = 0.02), but after interval cholecystectomy the recurrence rate of biliopancreatic symptoms was similar (one of 15 patients versus three of 16 patients respectively).

CONCLUSION

Endoscopic sphincterotomy, but not interval cholecystectomy, reduced the overall incidence of recurrent pancreatitis, but not of late biliary complications. Some 31 per cent of the patients required cholecystectomy, suggesting that routine cholecystectomy should be considered in fit patients following acute pancreatitis.

Choledocholithiasis and gallstone pancreatitis

Gallstones are commonly found within the main bile duct (MBD) of patients undergoing cholecystectomy. Retained MBD stones are a common cause of obstructive symptoms and complications. Endoscopic retrograde cholangiopancreatography (ERCP) and sphincterotomy (ES) is the recommended modality for both the detection of such stones and their extraction. Recent trials of ERCP in conjunction with laparoscopic cholecystectomy suggest that it should be reserved for use post-operatively. Gallstones within the MBD are the most common single cause of acute pancreatitis. Initial treatment is supportive, although new agents designed to suppress the systemic inflammatory response are under development and have proved beneficial in clinical trials. Severe cases should be treated with systemic antibiotics and early removal of the obstructing stones by ERCP and ES. Prophylactic cholecystectomy is recommended to prevent further attacks of gallstone pancreatitis.

Acute biliary pancreatitis: diagnosis and management

Acute biliary pancreatitis is a serious complication of biliary calculous disease and is associated with significant morbidity and mortality. The role of endoscopic retrograde cholangiopancreatography (ERCP) in the management of acute biliary pancreatitis has been the focus of discussion in recent years. In addition, the exact role of laparoscopic cholecystectomy (LC) in the management of acute biliary pancreatitis has not yet been fully defined. In this report, we evaluated a protocol of emergency ERCP (within 24 hours) for predicted severe attacks, early ERCP (within 72 hours) for predicted mild attacks, and interval LC for management of acute biliary pancreatitis. Between January 1992 and June 1995 a total of 75 patients with acute biliary pancreatitis were managed according to the protocol. Bedside ultrasonography at admission diagnosed 94% of all 64 patients with gallbladder stones, but the sensitivity of visualizing choledocholithiasis was low (19%). Forty-five (60%) of them were predicted to have a severe attack by either Ranson or glucose/urea criteria. Emergency ERCP and endoscopic sphincterotomy (ES) for identifiable common bile duct (CBD) or ampullary stones were performed on all patients predicted to have a severe attack within 24 hours from presentation. An early endoscopic procedure was performed on all patients predicted to have a mild attack within 72 hours from presentation. ERCP was successful in 95% of all patients, and CBD stones were detected in 52 (69%) of them. ES and stone clearance were successful in all of these 52 patients. The morbidity associated with the endoscopic procedure was 3%, and there were no deaths. All except one patient survived the attack of acute pancreatitis, resulting in an overall mortality of 1%. Interval LC was performed on 46 patients with a conversion rate of 4%. The median postoperative hospital stay after LC was 2 days, and there was no major intraoperative or postoperative morbidity or mortality. Our experience suggests that the policy of emergency ERCP for patients with predicted severe disease, early ERCP for patients with predicted mild disease, and interval LC are associated with favorable outcomes in patients with acute biliary pancreatitis. Acute biliary pancreatitis can be managed safely and effectively by a combined endoscopic and laparoscopic approach.

Should ERCP be routine after an episode of "idiopathic" pancreatitis? A cost-utility analysis

BACKGROUND

Patients often recover from an episode of acute pancreatitis with conservative therapy and without an identified cause. The options include proceeding with ERCP to identify and treat an occult common bile duct stone or performing the procedure only after a second episode of idiopathic pancreatitis occurs.

METHODS

Decision analysis (SMLTREE software) was used to determine incremental cost-utility. Variables were estimated from a search of the literature, a utility analysis involving health professionals familiar with the question, and a retrospective review of hospital charts and costs.

RESULTS

This model estimates an incremental utility gain for the prompt ERCP approach of 1.0 quality-adjusted life weeks per patient at an incremental cost of $245 (Canadian). This yields a cost-utility ratio of $12,740 (Canadian) per quality-adjusted life year. The result was highly sensitive to the probability of finding an occult common bile duct stone.

CONCLUSION

Routine ERCP is of marginal overall benefit, but is of more substantial benefit and is more cost-effective in a subgroup of patients with a greater probability of having an occult common duct stone.

Biochemical models as early predictors of the etiology of acute pancreatitis

In this study we observed the discriminative ability of five commonly measured laboratory tests to distinguish between gallstone- and non-gallstone-associated pancreatitis. We also assessed the ability of the lipase-amylase ratio to discriminate between alcohol- and non-alcohol-induced pancreatitis. One hundred sixty-two patients with acute pancreatitis were included in the study. Group A consisted of patients presenting to our hospital in 1988 and 1989. Group B consisted of patients presenting in 1992. Models developed using group A patients were validated using group B patients. For gallstone pancreatitis, AST (threshold value 80 IU/liter) alone and a three-factor model, AST, ALP and bilirubin (threshold values of 80 IU/liter, 115 IU/liter, and 15 mumol/liter, respectively) were the best predictors, correctly classifying at least 80% of cases in group A and B. A lipase-amylase ratio of two correctly classified only 48% of cases in group A and 54% in group B. We conclude that biochemical models are useful in predicting the presence of gallstone pancreatitis but not alcoholic pancreatitis.

Predictability of clinicobiochemical scoring systems for early identification of severe gallstone-associated pancreatitis

Early identification of severe gallstone-associated acute pancreatitis (GAAP) is a prerequisite for treatment with urgent endoscopic sphincterotomy. This study assesses the value of two clinicobiochemical scoring systems to this end. Over the 7-year period from 1983 to 1989, 100 consecutive patients with acute pancreatitis (45 related to gallstones, 36 to alcohol, and 19 of undetermined etiologies) had clinicobiochemical analysis within 48 hours of admission. The final diagnosis and outcome were retrospectively compared with the prediction achieved by the scoring systems. With regard to Blamey's criteria for early identification of gallstones, significant differences were found between the biliary and nonbiliary groups with respect to female sex, serum amylase concentration greater than or equal to 4,000 IU/L, alkaline phosphatase level greater than or equal to 300 IU/L, and alanine aminotransferase level greater than or equal to 100 IU/L (all p values less than 0.001). Age greater than or equal to 50 years was found to be significant (p less than 0.02) only in differentiating gallstone- versus alcohol-associated acute pancreatitis. When three or more positive factors were present, the sensitivity and specificity for predicting gallstones were 60% and 87%, respectively; the predictive value of a positive result was 79%, of a negative result 74%, and the overall accuracy was 75%. At a cutoff level of five, rather than three or more prognostic factors, the modified Ranson's criteria for patients known as having GAAP allowed a suitable discrimination of patients with an expected high risk of complications and mortality. When the two scoring systems (Blamey greater than or equal to 3 and Ranson greater than or equal to 3) were combined, 17 patients were predicted as having severe GAAP: 6 of these 17 patients were misdiagnosed as having biliary pancreatitis, whereas 9 patients with definite severe GAAP were not selected because of a Blamey score less than 3. More specific diagnostic tools are needed, and higher cutoff levels for prognostic scores are required for the prediction of severe GAAP, particularly in view of selecting patients for potentially dangerous approaches such as urgent endoscopic sphincterotomy.

Acute pancreatitis: the role of early surgery

Conservative treatment for an attack of acute pancreatitis still takes priority. The treatment of choice in biliary pancreatitis is endoscopic papillotomy with extraction of any bile duct stones. After this procedure the patient usually recovers quickly and cholecystectomy should be done as an interval operation. An early operation has to be done when conservative treatment fails and organ failure occurs. The 'gold standard' of surgical therapy today is the opening of the lesser sac with continuous postoperative lavage and drainage, digital elimination of necrotic tissue and drainage of the paracolic areas. Other described surgical procedures do not produce significantly better results. Necrotizing pancreatitis still has a high mortality in contrast to mild oedematous pancreatitis, which is rarely a problem.

Endoscopic management of pancreatic disease

Since its introduction in 1968, ERCP has developed from being a purely diagnostic method, mostly used in the investigation of unexplained upper abdominal pain, to an invaluable tool for the management of patients with pancreatic disorders. In cases with severe gallstone pancreatitis, the biliary obstruction is disclosed and relieved by ERCP and ES. In patients with severe acute pancreatitis of other aetiologies, as well as in post-traumatic pancreatitis, ERCP is indispensable for revealing complications (e.g. pancreatic duct rupture) and/or for planning the treatment strategy. Furthermore, in cases of pancreatitis not related to alcohol or gallstones, it often demonstrates causes which may be treatable, and it is also useful for evaluation of the gland after massive pancreatic necrosis. Moreover, ERCP is helpful in establishing the diagnosis of chronic pancreatitis and its complications as well as in demonstrating morphological grounds for therapeutic intervention. Although the indications, limitations, and practicability of the different techniques of therapeutic ERCP in various pancreatic diseases still remain to be defined, the method appears to offer an alternative to surgery, particularly in cases in which operative treatment is technically difficult and the results are less favourable. Frequency and severity of complications associated with both diagnostic and therapeutic ERCP seem to be, at least in the hands of experts, reasonably low.

Management of gallstone pancreatitis

The clinical course of 143 patients with gallstone pancreatitis is reviewed. Thirty-one patients (22%) had three or more positive prognostic factors on admission and 24 (77%) of these had a complicated course. Thirteen patients died, giving an overall mortality rate of 9%. Patients were divided into three groups on the basis of performance and timing of surgery. In group 1 (n = 56), surgery was undertaken during the first admission with acute pancreatitis; eight of these patients had a complicated course and three died. In group 2 (n = 40), biliary surgery was deferred to a subsequent admission; none of these patients died but 10 experienced further attacks of pancreatitis while awaiting reoperation. Group 3 patients (n = 47) did not undergo surgery; nine patients were diagnosed as having gallstone pancreatitis for the first time at autopsy, five refused operation, seven were lost to follow-up, six were dealt with by endoscopic sphincterotomy, and in 20 cases surgery was not considered appropriate because of general debility or advanced age. Despite the zero mortality rate in group 2, it is advocated that biliary surgery be carried out during the index hospital admission. Endoscopic sphincterotomy can now be considered as an alternative to cholecystectomy and duct clearance in the elderly and unfit, and may be used as a preliminary manoeuvre when severe acute pancreatitis fails to settle promptly on conservative management.