management of gallstone pancreatitis in auckland - Wiley Online Library

4 downloads 3073 Views 184KB Size Report
Conclusion: There has been reasonable compliance with published guidelines and some ... used to define the actual clinical severity of pancreatitis (mild or.
ANZ J. Surg. 2003; 73: 194–199

ORIGINAL ARTICLE ORIGINAL ARTICLE

MANAGEMENT OF GALLSTONE PANCREATITIS IN AUCKLAND: PROGRESS AND COMPLIANCE SOO-KIM ONG, PETER M. CHRISTIE AND JOHN A. WINDSOR Hepatobiliary–Pancreatic/Upper Gastrointestinal Unit, Department of General Surgery, Auckland Public Hospital, Auckland, New Zealand

Background: Recent advances in the management of acute gallstone pancreatitis include the introduction of laparoscopic cholecystectomy, defining the role of endoscopic retrograde cholangiopancreatography (ERCP) and early cholecystectomy to prevent recurrent pancreatitis. The aim of the present study was to review the current management of gallstone pancreatitis in Auckland Hospital, compare findings with a similar study published a decade ago and to determine the extent to which the management is compliant with recently published consensus guidelines. Methods: A retrospective review of consecutive patients admitted with acute pancreatitis during a 39-month study period was undertaken. Data were recorded regarding demographics, diagnosis, predicted and actual severity of gallstone pancreatitis (index and recurrent attacks), the role of ERCP and computed tomography scanning, the timing of cholecystectomy (open and laparoscopic), intraoperative cholangiography, duration of hospital stay, complications and mortality. Results: There were 216 patients admitted with acute pancreatitis, 106 of whom had proven gallstones. An ERCP was performed in 62 (59%) patients with gallstone pancreatitis but not more commonly in patients with severe pancreatitis, and common bile duct stones were identified in 26% of these patients. Of the 70 (66%) patients who had a cholecystectomy, 56 (80%) had it within 3 weeks of admission. Although the proportion of patients with gallstone pancreatitis who had a cholecystectomy is similar to the earlier study, there has been a significant increase in the proportion of patients having a cholecystectomy during the index admission (χ2 = 3.83; P = 0.05). This has resulted in a reduction in recurrent pancreatitis (P < 0.001). Although the overall mortality from gallstone pancreatitis has not significantly decreased, it has for patients with predicted severe gallstone pancreatitis (P = 0.02). Conclusion: There has been reasonable compliance with published guidelines and some progress in the management of gallstone pancreatitis, particularly in relation to performing timely laparoscopic cholecystectomy with a reduction in the incidence of recurrent pancreatitis. Concerns remain regarding the overuse of diagnostic ERCP in patients with mild pancreatitis. Key words: cholecystectomy, endoscopic retrograde cholangiopancreatography, gallstone pancreatitis, recurrent pancreatitis. Abbreviations: ceCT, contrast-enhanced computed tomography; ERCP, endoscopic retrograde cholangiopancreatography;

IOC, intraoperative cholangiography; MRCP, magnetic resonance cholangiopancreatography.

INTRODUCTION During the last 20 years there have been several evidence-based developments in the management of acute gallstone pancreatitis. These include the introduction of laparoscopic cholecystectomy,1,2 early cholecystectomy3,4 and endoscopic retrograde cholangiopancreatography (ERCP) in patients with acute gallstone pancreatitis.5–10 It is not known to what extent these developments have penetrated current practice or improved outcome. A previous study of acute gallstone pancreatitis, published in 1990, noted a high recurrence rate of pancreatitis (53%) while patients awaited cholecystectomy.11 This reflected a policy of delayed open cholecystectomy, which in that study was performed at an average of 79 days after index admission. During that 9-year study period ERCP was rarely undertaken. The aim of the present study was to examine the current management and outcome of acute gallstone pancreatitis in light S-K. Ong MBChB; P. M. Christie MBChB, MD, FRACS; J. A. Windsor BSc, MBChB, MD, DipObst, FRACS, FACS. Correspondence: Associate Professor J. A. Windsor, Department of Surgery, Faculty of Medical and Health Sciences, University of Auckland, 5th Floor, Auckland Hospital, Private Bag 92024, Auckland, New Zealand. Email: [email protected] Accepted for publication 2 December 2002.

of the recently published UK guidelines for the management of acute pancreatitis,3 with particular reference to the role of computed tomography (CT) and ERCP, the timing of cholecystectomy and the recurrence of acute pancreatitis.

METHODS The clinical records of consecutive patients admitted with firstonset acute pancreatitis to Auckland Public Hospital, a large metropolitan teaching hospital, were retrospectively reviewed. The study period was May 1999–October 2001 and the patients were managed in the Department of Surgery, with an increasing proportion within the Hepatobiliary-Pancreatic/Upper Gastrointestinal Unit. The patients were identified using an electronic database search system and were those given a primary diagnosis of acute pancreatitis according to the International Classification of Diseases, 9th edition, coding classification system (ICD-9CM, code 577.0). The diagnosis of acute pancreatitis was made in accordance with the criteria laid down by the UK pancreatitis guidelines.3 This consisted of a consistent clinical picture of upper abdominal pain and serum amylase activity four times above normal (>540 U/L). The diagnosis of acute gallstone pancreatitis required the presence of gallstones or biliary sludge on ultrasound or common bile duct stone(s) on ERCP. The common bile duct was considered dilated if it exceeded 6 mm in diameter on ultrasound.12,13 If an ultrasound was not performed or the

GALLSTONE PANCREATITIS

195

findings were equivocal, the diagnosis of gallstone pancreatitis required three or more independent positive predictive factors.14 The predicted severity of pancreatitis was determined using the modified Glasgow criteria.15 A score of 3 or more constituted severe disease. Ranson’s criteria16 were used when comparing the results with the earlier study. The Atlanta classification17 was used to define the actual clinical severity of pancreatitis (mild or severe). In order to define medical comorbidity, all patients were assigned an American Society of Anesthesiologists (ASA)18 score at the time of cholecystectomy (or at the time of index admission if a cholecystectomy was not performed). The first hospital admission with acute pancreatitis was designated the index admission. An elective cholecystectomy was defined as a cholecystectomy that was performed after the patient had been discharged from hospital. The length of stay in hospital during the index admission, timing of ERCP and cholecystectomy (index or elective) and the duration (in days) from the index

admission to cholecystectomy were recorded. Attempts were made by telephone to follow up all patients who were placed on the waiting list for an elective cholecystectomy. Only patients who had gallstone pancreatitis were included in the analyses. The means and standard deviations were calculated for normally distributed continuous variables while the medians and ranges were calculated for continuous variables with a skewed distribution. Student’s t-test (two-sided) was used to compare the means between groups with continuous variables. The chi-squared test (with the Yates continuity correction) was used for categorical variables while the Fisher’s exact test (twotailed) was used when the number of observations in a group was less than 10. Cohen’s kappa statistic (κ) was used to measure the degree of agreement between two different rating systems. Correlation analysis was reported using Spearman’s coefficient (ρ). The statistical software used was SPSS (SPSS, Chicago, IL, USA, Release 11.0.1). P < 0.05 was considered significant.

RESULTS Table 1. Aetiology of acute pancreatitis Aetiology

Features of pancreatitis in Auckland n

%

Gallstones Idiopathic Alcohol Post-ERCP Drugs (2 × ranitidine, 1 × prednisone) Pancreatic malignancy Pancreaticojejunostomy stricture Hypercalcaemia Hypercholesterolaemia Pancreatic divisum Post-CABG ischaemia Presumed viral cause Trauma

106 35 34 26 3 3 3 1 1 1 1 1 1

49.1 16.1 15.7 12.0 1.4 1.4 1.4 0.5 0.5 0.5 0.5 0.5 0.5

Total

216

100.0

ERCP, endoscopic retrograde cholangiopancreatography; CABG, coronary artery bypass grafting.

Fig. 1. Distribution of the length of hospital stay during the index admission for patients admitted with acute gallstone pancreatitis.

There were 216 patients admitted with acute pancreatitis during the 39-month study period. Of these, 106 (49%) patients (40 men, 66 women; median age: 57 years; range: 15–91 years) had acute gallstone pancreatitis (Table 1). Male patients were on average 11 years older (mean: 63 vs 52 years; t = 2.84; P = 0.03) than female patients and were also more likely to have severe disease (χ2 = 4.06; P = 0.04). Severe pancreatitis was predicted in 28 (26%) patients using the modified Glasgow criteria. According to the Atlanta classification, 19 patients (18%) actually developed severe pancreatitis. There was a reasonable degree of agreement between the predicted severity of pancreatitis and the actual severity of pancreatitis (κ = 0.70; P < 0.001). The median length of hospital stay for the index admission of patients with gallstone pancreatitis was 8 days (range: 1–129 days; Fig. 1). The overall mortality was 4.7% (5/106). The mortality of those with predicted severe gallstone pancreatitis was 3.6% (1/28) and the mortality of those with actual severe gallstone pancreatitis was 11% (2/19).

196

ONG ET AL.

Investigations Ultrasound scanning was done during the index admission in 93 (88%) patients. An abnormality was found in the biliary tree in all but eight patients (three previous cholecystectomies, three had a subsequent cholecystectomy and two were lost to follow up). These abnormalities were gallstones (n = 76), biliary sludge (n = 6) and a dilated common bile duct (n = 38). A contrast-enhanced computed tomography (ceCT) scan was performed in 24 (23%) patients during the index admission. Only 8/28 (29%) patients with predicted severe pancreatitis had a ceCT during the index admission. Patients with more serious comorbidities (ASA 3, 4, or 5, 17/24, 71%) were more likely to undergo a ceCT scan during the index admission compared with those with less serious comorbidities (ASA 1 or 2, 23/82, 28%; χ2 = 12.70; P < 0.001). Sixty-two (59%) patients had an ERCP during the index admission. There was a trend towards patients with predicted severe pancreatitis being more likely to be referred for ERCP (20/ 28, 71% patients with severe disease vs 42/78, 54% with mild pancreatitis; Fisher’s exact test, P = 0.08). A higher proportion of patients with predicted mild pancreatitis and a dilated common bile duct on ultrasound were more likely to have preoperative ERCP (17/25, 68%) compared with those who did not have a dilated duct (19/45, 42%; Fisher’s exact test, P = 0.05). The median interval to ERCP following admission was 4 days (1–18 days) in patients with predicted severe pancreatitis and 5 days (0–46 days) in patients with predicted mild pancreatitis. A sphincterotomy was performed in 42/62 (68%) patients who had an ERCP. Choledocholithiasis was detected in 16/62 (26%) of patients who had an ERCP and in 38% (16/42) of those who had a sphincterotomy. The patients with predicted severe pancreatitis were not more likely to have choledocholithiasis detected at ERCP (5/20 (25%) patients vs 11/42 (26%) with mild pancreatitis; Fisher’s exact test, P = 1.0). Patients who had an ERCP were just as likely to have a cholecystectomy (40/62, 65%) as those who did not have an ERCP (30/44, 68%; χ2 = 0.03, P = 0.85). But those who had an ERCP were more likely to have delayed surgery (i.e. 68% (27/40) who had a cholecystectomy performed more than 7 days after admission) compared with those who did not have ERCP (40%, 12/30; χ2 = 4.20, P = 0.04). Patients who had an ERCP were also more likely to have a delay in discharge from hospital (median length of stay 9 days (range: 3–129 days) vs 6 days (range: 1–14 days), Mann–Whitney U-test; P < 0.001). Magnetic resonance cholangiopancreatography (MRCP) was performed in only three (2.8%) patients. Surgical management of acute gallstone pancreatitis A cholecystectomy was performed in two-thirds of the patients with gallstone pancreatitis (70/106; 66%). Of those who had a cholecystectomy, 44/70 (63%) had it during the index admission and 56/70 (80%) had it within 21 days of the index admission. Twenty (36%) patients with predicted mild pancreatitis were discharged for an elective cholecystectomy compared with 6/15 (49%) patients with predicted severe disease (Fisher’s exact test, P = 1.0). The reasons why a cholecystectomy was not performed in some patients with gallstone pancreatitis are shown in Table 2. It can be seen that patient refusal (n = 9) was the most common reason followed by medical contraindications (n = 7), patients being lost to follow up (n = 6) and patients who already had a cholecystectomy (n = 5). Five patients died before an operation could be performed. Of these, three patients died from pancreati-

Table 2. Reasons why a cholecystectomy was not performed in 36 patients with gallstone pancreatitis Reason

n

Refused cholecystectomy Medical contraindications Lost to follow up Died before operation Previous cholecystectomy Waiting for other operation first Gallstone pancreatitis not suspected; identified with Glasgow criteria

9 7 6 5 5 3 1

tis during the index admission; another patient died while on the waiting list for cholecystectomy (cause unknown); and the other patient died from other causes while waiting for another operation which took priority over a cholecystectomy. The median overall time from admission to cholecystectomy was 9 days (range: 0–368 days) and for those who had a cholecystectomy during the index admission it was 6 days (range: 0–21 days). The median interval to cholecystectomy was 8 days (range: 0–317 days) in patients with predicted mild pancreatitis and 11 days (range: 0–368 days) in those with predicted severe pancreatitis. The median waiting time following discharge for patients scheduled for an elective cholecystectomy was 21 days (range: 3–362 days). Patients with less serious comorbidity (ASA 1 or 2, 51/66, 77%) were more likely to have a cholecystectomy than those with more serious comorbidity (ASA 3, 4 or 5, 19/40, 48%; χ2 = 8.56, P = 0.003; Fig. 2). Patients were also more likely to have a cholecystectomy if they had a longer hospital stay during the index admission (Spearman’s correlation coefficient = 0.54, P = 0.013; Fig. 3). A laparoscopic cholecystectomy was attempted in 63/70 (90%) patients and was completed in 60 patients, giving a conversion rate of 5%. The total number of open cholecystectomies was 10 (seven from the outset and three from conversion). Intraoperative cholangiography (IOC) was attempted in less than half (31/70, 44%) of the patients undergoing cholecystectomy and was successful in 30 patients (97%). An IOC was more likely to be performed in patients who did not have preoperative ERCP (22/30 (73%) vs 9/40 (23%); Fisher’s exact test, P < 0.001). An IOC was performed more often during laparoscopic cholecystectomy (29/60, 48%) than open cholecystectomy (2/10, 20%) although this did not reach statistical significance (Fisher’s exact test, P = 0.17). Overall, there were 22 (21%) patients who had no imaging of the common bile duct by ERCP or IOC. Eight patients (11%) who had a cholecystectomy had no imaging (ERCP or IOC) of the common bile duct. The median length of postoperative stay for patients who had a cholecystectomy during the index admission was 3 days (range: 1–110 days). Patients with mild pancreatitis had a marginally shorter median postoperative stay than those with severe disease (median: 3 days (range: 1–39 days) vs 4 days (range: 1–110 days); Mann–Whitney U-test, P = 0.56). Recurrence rate Five (7%) patients (four women, one man) had a recurrence of pancreatitis while waiting for a cholecystectomy, all of them in the elective cholecystectomy group (5/26, 19%). Three patients were on the waiting list and the intervals to recurrence were 7,

GALLSTONE PANCREATITIS

197

severity of the recurrent pancreatitis was worse than the attack during the index admission in only one patient. An ERCP was performed during the index admission in two patients. Following recurrence of pancreatitis, all five patients had a cholecystectomy during the second hospital admission.

DISCUSSION

Fig. 2. Distribution of American Society of Anesthesiologists (ASA) scores for () cholecystectomy patients and () those who had no operation.

Fig. 3. Proportion of patients having a cholecystectomy during index admission (), and elective admission ().

16 and 17 days from admission. One patient was thought to have acute gallstone pancreatitis on a background of chronic alcoholic pancreatitis and following a normal ERCP was discharged with no plans for a cholecystectomy. Another patient presented with mild gallstone pancreatitis but a cholecystectomy was not considered an option due to significant comorbidities. The predicted

The present study has reviewed the current management of acute pancreatitis in a large teaching hospital. The purpose of the study was to compare current management with a previous study involving patients from the same catchment area in order to determine whether there had been any progress and also to determine the extent to which current management complied with the recently published UK guidelines.3 In the present study it was found that only two-thirds of patients admitted with gallstone pancreatitis had a cholecystectomy. Of these, 63% were performed during the same admission and 80% within 3 weeks. A cholecystectomy was more likely to be performed during the index admission if the patient had low comorbidity (ASA 1 or 2) and had a longer hospital admission but was less likely in patients with severe pancreatitis. The rate of ERCP was not significantly higher in patients with severe disease. The overall rate of sphincterotomy was 68% despite a common duct stone detection rate of 26%; in other words, there were three sphincterotomies performed for every stone detected. There was no evidence that ERCP and sphincterotomy was used as an alternative to cholecystectomy in high-risk patients presenting with gallstone pancreatitis. Cholecystectomy was performed laparoscopically in 90% of patients and there was a conversion rate of 5%. Intraoperative cholangiogram was attempted in only 44% of patients with a 97% success rate. The overall incidence of recurrent gallstone pancreatitis while awaiting a cholecystectomy was 7%. The recurrence increased almost threefold (19%) in those patients discharged for an elective cholecystectomy. These results have been compared with the previous study (Table 3), which examined the management of patients with gallstone pancreatitis from the same catchment area during the years 1979–1987.11 There was a decrease in the percentage of patients who were lost to follow up (from 36% to 17%). Although gallstones accounted for almost half (49%) of the admissions with acute pancreatitis, there was a significant decrease in cases attributable to alcohol (from 39% to 16%; χ2 = 26.45; P < 0.001). There was a significant decrease (from 53% to 31%) in the proportion of patients presenting with predicted severe pancreatitis by Ranson’s criteria. There appeared to be a trend towards patients presenting at a younger age (median: 57 vs 71 years). There was no significant change in overall mortality of gallstone pancreatitis between the two studies (13% vs 4.7%; Fisher’s exact test, P = 0.11) but there was a significant decrease in mortality from predicted severe gallstone pancreatitis (from 25% to 3%; Fisher’s exact test, P = 0.02). The UK pancreatitis guidelines set a target mortality of