out, up to 30% of episodes of acute pancreatitis are labeled as idiopathic .... of them were lost to follow-up after biliary drainage (2 with normal findings ..... while they were on the waiting list for surgery, one after endoscopic sphincterotomy, one ...
GASTROENTEROLOGY 1991;101:1701-1709
Occult Microlithiasis in 'Idiopathic' Acute Pancreatitis: Prevention of Relapses by Cholecystectomy or Ursodeoxycholic Acid Therapy EMILIO ROS, SALVADOR NAVARRO, CONXITA BRU, A N T O N I O GARCIA-PUGES, a n d RODRIGO VALDERRAMA Gastroenterology Service and Radiology Department, Hospital Clinic i Provincial, Barcelona School of Medicine, Barcelona, Spain
Gallstone pancreatitis is usually related to small stones, which may not be detected by conventional cholecystographic techniques. In the current study, it was hypothesized that some patients with acute pancreatitis of u n k n o w n cause could harbor occult microstones in the gallbladder. Therefore, evidence was sought prospectively of missed gallstones by biliary drainage and microscopic examination of centrifuged duodenal bile in 51 patients recovering from an attack of acute pancreatitis, including 24 patients with relapsing episodes. Clusters of cholesterol monohydrate crystals, calcium bilirubinate granules, and/or CaCo 3microspheroliths were found in 67% of the patients. Biliary drainage showed no abnormal findings in 12 patients convalescing from a bout of known alcoholic pancreatitis. Examination of gallbladder bile at cholecystectomy and/or serial ultrasonography of the gallbladder for up to 12 months showed that 73% of the patients with unexplained pancreatitis had biliary sludge or microlithiasis; the prior finding of biliary crystal/solid markers predicted their existence with both a sensitivity and a specificity of 86% and a predictive value of 94%. The probability of harboring occult gallstones was also associated with age (P = 0.004), prior recurrent pancreatitis (P = 0.024), and altered liver function tests results during an index episode (P = 0.003). In 13 patients with cholesterol monohydrate crystals in bile, ursodeoxycholic acid (10 mg. kg -1 • day -1) eliminated gallbladder microlithiasis within 3-6 months, and subsequent maintenance treatment with a daily dose of 300 mg prevented both gallstone recurrence and further attacks of pancreatitis over a m e a n follow-up period of 44 months. Cholecystectomy also prevented gallstoneassociated relapses in 17 of 18 patients followed up
for a mean postoperative period of 36 months. This study provides firm evidence showing that in most patients with idiopathic acute pancreatitis, the disease is related to microscopic gallstones, as evidenced by the follow-up development of macroscopic stones or sludge and by the prevention of relapses with either cholecystectomy or a cholelitholytic bile acid. Occult gallstones should be strongly suspected w h e n acute pancreatitis of unknown cause occurs in a relapsing m a n n e r and in aged patients and w h e n it is associated with altered liver function test results. Biliary microscopy and/or follow-up ultrasonography of the gallbladder provide a simple means of uncovering them to institute appropriate therapy and prevent further attacks. cute pancreatitis is a common abdominal emergency that can have serious complications and still has a high mortality rate (1). Once cholelithiasis, alcoholism, and miscellaneous etiologies are ruled out, up to 30% of episodes of acute pancreatitis are labeled as idiopathic (1-3). Gallstones are the most common cause of acute pancreatitis in community hospitals in the United States and in Western Europe (1-4). The pathogenesis of biliary pancreatitis is related to the temporary impaction of migrating gallstones at the ampulla of Vater (5). Indeed, small gallstones are frequently found on surgical exploration of the gallbladder and/or common bile duct in Abbreviations used in this paper: CBG, calcium bilirubinate granules; CMC, cholesterol monohydrate crystals; ERCP, endoscopic retrograde cholangiopancreatography; GGTP, ~oglutamyltranspeptidase; MSL, microspheroliths; US, nltrasonography. © 1991 by the American Gastroenterological Association 0016-5085/91/$3.00
1702 ROS ET AL.
patients w i t h biliary pancreatitis w h o u n d e r g o surgery during the acute stage of the disease a n d are f o u n d in the feces of > 90% of screened patients (5). In patients t h o u g h t to h a v e idiopathic pancreatitis, small stones m a y not be detected b y c o n v e n t i o n a l gallbladder or ductal imaging t e c h n i q u e s (6); thus, t h e y are e x p o s e d to an u n d u l y high risk of relapse. The m i c r o s c o p i c e x a m i n a t i o n of s t i m u l a t e d duodenal bile, a t e c h n i q u e w i d e l y u s e d in the diagnosis of gallstones before the a d v e n t of m o d e r n imaging procedures (7,8), provides useful i n f o r m a t i o n on b o t h the p r e s e n c e a n d the c h e m i c a l c o m p o s i t i o n of gallstones (9). Further, the finding of cholesterol m o n o h y d r a t e crystals (CMCs) in d u o d e n a l bile is associated w i t h a high p r o b a b i l i t y of successful cholelytholysis w i t h u r s o d e o x y c h o l i c acid (UDCA) t h e r a p y (9). It s e e m e d logical, therefore, to a p p l y this k n o w l e d g e to the p r o b l e m of m i s s e d gallstones in patients w i t h pancreatitis of u n k n o w n cause. The aim of this s t u d y was to seek evidence of m i s s e d gallbladder stones b y biliary m i c r o s c o p y in patients recovering f r o m a recent episode of acute pancreatitis w i t h o u t a p p a r e n t cause. We also sought to confirm the existence of microlithiasis b y follow-up u l t r a s o n o g r a p h y (US) of the gallbladder and/or cholecystectomy. In addition, the efficacy a n d safety of UDCA t h e r a p y in p r e v e n t i n g further attacks of acute pancreatitis were evaluated in a limited n u m b e r of patients.
Patients and Methods
Patients The study group consisted of 64 consecutive patients convalescing from a recent episode of acute pancreatitis of unknown cause at Hospital Clinic i Provincial, Barcelona. All patients were evaluated during a 10-year period starting May 1979. There were 45 women and 19 men with a mean age of 62 years (range, 18-88 years). Most patients (n = 59) had been admitted on an emergency basis to the 30-bed gastroenterology ward from our institution, which is a 900-bed community hospital serving mostly aged, low-middle class population. Another 6 patients were hospitalized at the general medical wards, and 8 patients were referred from other institutions within the Barcelona area. The diagnosis of acute pancreatitis was made when acute abdominal pain was associated to a serum amylase level >900 IU/L (a fourfold elevation above mean _+ 2SD of normal values) and/or a Serum lipase concentration > 180 IU/L (a threefold elevation) in addition to US, computed tomographic, or surgical evidence of pancreatic inflammation. Blood for biochemical analyses, including liver function tests and the plasma concentration of total calcium and triglycerides, was obtained in all patients within 3 days from admission. Excessive alcohol ingestion, defined as average daily consumption of > 30 g ethanol, was ruled out in all patients
GASTROENTEROLOGYVo1. 101, No. 6
by clinical history and measurement of biological markers of alcoholism. Of the 64 patients evaluated, 48 never consumed alcoholic beverages, 8 drank an occasional glass of wine, and 8 had a regular ethanol ingestion < 30 g/day. Other requisites to enter the study were absence of hypercalcemia or hypertriglyceridemia, no recent intake of drugs documented to induce acute pancreatitis (especially steroids, diuretics, and antineoplastic agents), at least two US examinations of the gallbladder showing no abnormality (one during the acute phase of the illness and another after recovery and resumption of oral intake), and an oral cholecystogram showing no abnormality (including erect films). Endoscopic retrograde cholangiopancreatography (ERCP) was also performed in 15 patients with relapsing pancreatitis; no etiological diagnosis was reached. Of 64 patients fulfilling the above criteria, 29 (45%) had at least one episode of idiopathic acute pancreatitis before the current attack (overall, 104 episodes in 64 patients). By Ranson's criteria (10), acute pancreatitis was defined as mild or moderate in all except 2 patients, 1 of whom survived an episode of severe pancreatitis complicated by multiorgan failure and pseudocyst formation. Another 5 patients had a pseudocyst, and 2 of them required surgical drainage.
Study Protocol All patients were requested to undergo biliary drainage, which was usually performed within 4-8 weeks from recovery of acute pancreatitis and discharge from hospital. The subsequent approach was mainly dictated by the findings of biliary microscopy, but age, clinical status, history of relapsing pancreatitis, and expected compliance with follow-up were also considered. When patients were submitted to cholecystectomy, gallbladder contents were retrieved. All patients who did not undergo surgery were followed up with periodic gallbladder US. Ursodeoxycholic acid treatment was offered to a selected group of 17 patients who were either unwilling to undergo cholecystectomy or who were considered as poor surgical candidates for reasons of age and/or concomitant disease. The bile acid was given in 150-rag tablets to provide ~ 10 mg• kg -1 • day -1 with ingestion of two thirds of the total dose with the evening meal (11). The efficacy of UDCA therapy was assessed by follow-up gallbladder US in all patients and by repeated biliary drainage in 8 patients. After 3-12 months of uninterrupted full-dose treatment, maintenance UDCA (300 mg/day) was continued indefinitely under periodic supervision. Clinical events during follow-up were recorded in all patients, with particular attention to eventual readmissions for relapsing pancreatitis and subsequent clinical course. To learn whether abnormal biliary solids were present in patients with acute pancreatitis of well-identified etiology, 12 patients with known alcoholic pancreatitis and with no abnormality on gallbladder US (11 men and 1 woman, aged 29-54 years) and 2 patients with pancreatitis associated to massive hypertriglyceridemia (a 32-year-old man and a 25-year-old woman) also underwent biliary drainage during convalescence from the attack. Informed consent was obtained from all subjects. The protocol was approved by the Ethics Committee of the Hospital Clfnic i Provincial of Barcelona in May 1979.
December 1991
OCCULT MICROLITHIASISIN UNEXPLAINEDPANCREATITIS 1703
Methods Duodenal bile was obtained by nasoduodenal intubation after cholecystokinin stimulation of gallbladder contraction. The darkest samples were pooled and thoroughly mixed, and the pellet obtained after centrifugation of a 5-mL aliquot at 3000g for 10 minutes was immediately observed at 37°C under a polarizing microscope fitted with a heating stage for the presence of microcrystalline and other solids. Birefringent CMC and calcium carbonate microspheroliths (MSL) were identified as previously described (9), and brick-red aggregates of any size were recorded as calcium bilirubinate granules (CBG) (8). Their number per slide was estimated as absent, few (< 10 per slide), moderate (2-10 per high-power field), and abundant (7 10 per high-power field, clumps). Duplicate aliquots of uncentrifuged, unextracted duodenal bile samples were used to measure bile acid, phospholipid, and cholesterol composition by enzymatic methods (9). Compositional data were converted to molar percentage for each lipid component according to Admirand and Small (12) to calculate the percentage saturation of cholesterol in bile by the criteria of Carey and Small (13). The composition of gallstones obtained at cholecystectomy was determined by quantitative infrared spectroscopy (9) using a Perkin-Elmer model 681 infrared spectrophotometer (Perkin-Elmer Corp., Norwalk, CT). Gallstones were classified as cholesterol stones when they had 50%-100% cholesterol content and as noncholesterol stones when cholesterol content was < 20% (9). Gallbladder US was performed on fasting patients with a real time scanner and transducer frequencies of 3.5 and 3.7 MHz. Scans were performed with the patient in the supine position using sagittal, transverse, and subcostal projections; when appropriate, the left lateral decubitus position was also used. The criterion for the diagnosis of gallstones was the presence of intraluminal echoes moving with gravity and disclosing distal acoustic shadowing. Hyperechoic, nonshadowing, mobile images were considered as sludge.
Expression of the Results and Statistical Analyses All results are expressed as means _ SD. The Mann-Whitney rank sum test, unpaired Student's t test (two-tailed), and X2 test with Yates' correction, when necessary, were used for statistical treatment of the data. In addition, a stepwise logistic regression analysis (14) was performed with the demonstration of gallstones as dependent variable and other variables with possible predictive value as independent ones. BMDP statistical software (15) was used for the multivariate analysis.
Results
Patients With Pancreatitis of Well-Identified Etiology Duodenal intubation and biliary drainage were successfully a c c o m p l i s h e d in the 12 patients with
alcoholic pancreatitis and the 2 patients with hypertriglyceridemia-associated pancreatitis; the microscopic examination of the bile sediment s h o w e d no abnormalities in all of them.
Patients With Idiopathic Acute Pancreatitis Of the 64 patients with acute pancreatitis of u n k n o w n cause initially evaluated, 13 did not complete the study protocol for a variety of reasons. Five of t h e m were lost to follow-up after biliary drainage (2 with normal findings on bile microscopy, 2 with CMC, and 1 with MSL); 1 patient with normal bile microscopy results had an insufficient follow-up ( < 12 months); in 3 patients duodenal intubation was unsuccessful; and 4 patients refused the procedure. The latter 7 patients were followed up with periodic gallbladder US. Four of these patients had evidence of gallstones at US after 3-9 m o n t h s of follow-up and u n d e r w e n t elective c h o l e c y s t e c t o m y [the gallbladder c o n t a i n e d small cholesterol stones in 2 patients and noncholesterol stones in the other 2). One patient in this group of 4 had mild relapsing pancreatitis before surgery. Another had US evidence of sludge at 6 m o n t h s but refused intervention and further followup. The other 2 patients have r e m a i n e d well and with no abnormalities in gallbladder US for periods of 18 and 24 months, respectively. Fifty-one patients (37 w o m e n and 14 men, aged 59 ± 12 years) u n d e r w e n t successful biliary drainage and have a follow-up of at least 12 months. Subseq u e n t data apply only to these patients, w h o constitute the main b o d y of this report. Findings in bile. Microscopic examination of stimulated d u o d e n a l bile sediments s h o w e d CMC in 20 patients (estimated as few in 6 cases, moderate in 6 cases, and a b u n d a n t in 8 cases), CBG in 12 patients (estimated as moderate in 3 cases and a b u n d a n t in 9 cases), and MSL in 8 patients (estimated as few in 6 cases and moderate in 2 cases). Various associations of biliary crystal/solid markers were f o u n d in 5 patients (Table 1). The bile sediment was normal in 17 patients. Thus, the overall p r o p o r t i o n of abnormal biliary microscopic findings was 34 of 51 or 67%. The p e r c e n t biliary cholesterol saturation was 149 ± 41 (range, 112-218) in patients w i t h CMC, 111 _ 21 (range, 75-142) in patients with other solids. in bile, and 95 ± 34 (range, 55-184) in patients with normal biliary microscopic findings. The bile was u n s a t u r a t e d with cholesterol in 8 of 15 patients with abnormal biliary solids other t h a n CMC and in 8 of 17 patients with a normal biliary sediment. Demonstration of cholelithiasis. Gallstone disease was eventually p r o v e n in 35 patients. A cholecystectomy was performed in 18 patients, showing gallstones in 15 patients, sludge in I patient, and chronic
1704
ROS ET AL.
GASTROENTEROLOGY Vol. 101, No. 6
Table 1. Results of Microscopic Examination of Stimulated Duodenal Bile, Subsequent Ultrasonography of the Gallbladder, and Cholecystectomy in 51 Patients With Acute Pancreatitis of Unknown Cause Biliary microscopy Findings
Ultrasonography No.
Normal
17 17
12 2c
CBG
8
MSL
4
No crystals CMC
CBG + MSL CMC + CBG CMC + MSL CMC + CBG + MSL Total
2 1 1 1 51
Sludge
Cholecystectomy Gallstones
No.
1 3
4 (+1) a'b 10
2 7
ia
3
4 (+2) b
4
Ia
2
1 (+1) b
2
0 0 0 0 16
0 0 0 1 10
2 1 1 0 27
2 0 0 1 18
Gallstone type Noncholesterol 5 cholesterol 2 no gallstones d 3 noncholesterol 1 cholesterol I noncholesterol 1 MSL sludge 2 noncholesterol
Cholesterol 8 noncholesterol 7cholesterol 1 MSL sludge 2 no gallstones
NOTE. Follow-up US was performed in all patients except two in the CMC group w h o u n d e r w e n t emergency cholecystectomy (cholesterol microstones and sludge were present in the gallbladder of both patients). aFive patients with normal bile sediment in w h o m gallstones were found on follow-up US (false negative results of biliary microscopy) and two patients with abnormal findings w h o had normal gallbladders on follow-up US (false positives). bUltrasonography s h o w e d gallstones after prior documentation of sludge. Tollow-up ultrasonographies performed after early initiation of UDCA therapy. ~Cholecystectomies performed during UDCA treatment after prior US documentation of gallstones; in both cases the gallbladder s h o w e d pathological changes of chronic cholecystitis.
cholecystitis in 2 patients; findings in the gallbladders obtained at surgery are shown in Table 1. Gallstone size did not exceed 5 mm in 8 patients who underwent surgery early ( < 2 months) after acute pancreatitis. Besides microstones, the gallbladders usually contained thick bile, rich with sludge composed of CMC and/or CBG, with or without MSL. As a general rule, there were no qualitative differences in the results of microscopic examination between bile samples collected before and during surgery. Follow-up gallbladder US showed abnormalities in 33 patients, including 15 patients who subsequently underwent cholecystectomy. Table 1 shows the resuits of US examination in each category of biliary microscopic findings. Sequential US scans were available in all except 2 patients (Table 1). Unequivocal US evidence of gallstones was found in 27 patients, whereas only sludge was shown in the other 6 patients. With a single exception, all sludge-positive patients disclosed abnormal biliary solids (Table 1); 3 of them underwent cholecystectomy, with small gallstones and sludge found in the gallbladder at surgery. Sequential US showed the transformation of sludge into unequivocal gallstones in 4 additional patients (Table 1). With the exception of a single patient who had US evidence of gallstones 4 years after acute pancreatitis, US positivity occurred between 2 and 12 months in all untreated patients. In patients not submitted to intervention before US examination, abnormal biliary microscopic findings
were highly predictive of the subsequent demonstration of gallstones. Thus, 24 of 26 (92%) patients with any abnormality in the bile sediment and 13 of 13 (100%) with CMC had US evidence of gallstones, as opposed to 5 of 17 (29%) with a normal bile sediment (P < 0.001). Besides the 35 patients with surgical and/or US evidence of gallstones, 2 additional patients with CMC in bile who were treated early with UDCA and in whom follow-up US showed no abnormality, as described below, were considered as gallstone-positive cases on the basis of available evidence (8,9,16). Thus, gallstone disease as the most likely etiology of idiopathic acute pancreatitis was eventually documented in 37 of 51 (73%) patients. Factors associated with gallstones. By univariate analysis, age, relapsing pancreatitis, abnormal liver function test results during the index episode, and the finding of abnormal solids in bile were all associated with occult gallstones as the etiology of pancreatitis (Table 2). On multivariate analysis, the finding of biliary crystal/solid markers remained as a strong predictor of the eventual evidence of gallstones (P < 0.001), whereas relapsing pancreatitis fell short of statistical significance (P = 0.074). For the diagnosis of gallstone disease as the missed cause of idiopathic acute pancreatitis, the finding of abnormal biliary solids (CMC, CBG, MSL, or any combination thereof) had a sensitivity of 86%, a specificity of 86%, and a positive predictive value of 92%. During acute pancreatitis, the liver function test
December 1991
OCCULT MICROLITHIASIS IN UNEXPLAINED PANCREATITIS
1705
Table 2. Demographic, Clinical, Biochemical, and Bile Microscopic Features in Relation to the Eventual Diagnosis of Gallstones in 51 Patients With Presumably Idiopathic Acute Pancreatitis Patients w i t h gallstones
No. of m e n (%) Age (yr) No. of r e l a p s i n g pancreatitis a (%) S e r u m amylase, m e a n ± SD (mlU/mL) b No. of altered liver f u n c t i o n tests b (%) No. of a b n o r m a l bile m i c r o s c o p y (%)
Yes (n = 37)
No (n = 14)
12 (32) 62 ± 13 27 (57) 3293 ± 2233 34 (92) 32 (86)
2 49 3 1943 7 2
[14) ± 17 (21) ± 2117 (50) (14)
X2
t
0.592 2.770 5.088 1.722 8.806 20.688
P NS 0.004 0.024 NS 0.003
