Small gallstones, preserved gallbladder motility, and fast crystallization are associated with pancreatitis


  • A portion of this study was presented at the annual meeting of the American Gastroenterological Association (New Orleans, 2004) and was published as an abstract (Gastroenterology 2004;126:A232).

  • Conflict of interest: Nothing to report.


Acute pancreatitis is a severe complication of gallstones with considerable mortality. We sought to explore the potential risk factors for biliary pancreatitis. We compared postprandial gallbladder motility (via ultrasonography) and, after subsequent cholecystectomy, numbers, sizes, and types of gallstones; gallbladder bile composition; and cholesterol crystallization in 21 gallstone patients with previous pancreatitis and 30 patients with uncomplicated symptomatic gallstones. Gallbladder motility was stronger in pancreatitis patients than in patients with uncomplicated symptomatic gallstones (minimum postprandial gallbladder volumes: 5.8 ± 1.0 vs. 8.1 ± 0.7 mL; P = .005). Pancreatitis patients had more often sludge (41% vs. 13%; P = .03) and smaller and more gallstones than patients with symptomatic gallstones (smallest stone diameters: 2 ± 1 vs. 8 ± 2 mm; P = .001). Also, crystallization occurred much faster in the bile of pancreatitis patients (1.0 ± 0.0 vs. 2.5 ± 0.4 days; P < .001), possibly because of higher mucin concentrations (3.3 ± 1.9 vs. 0.8 ± 0.2 mg/mL; P = .04). No significant differences were found in types of gallstones, relative biliary lipid contents, cholesterol saturation indexes, bile salt species composition, phospholipid classes, total protein or immunoglobulin (G, M, and A), haptoglobin, and α-1 acid glycoprotein concentrations. In conclusion, patients with small gallbladder stones and/or preserved gallbladder motility are at increased risk of pancreatitis. The potential benefit of prophylactic cholecystectomy in this patient category has yet to be explored. (HEPATOLOGY 2005.)

Acute pancreatitis is a severe complication of gallstone disease, with considerable mortality (5%-10%).1 The annual incidence of acute pancreatitis in gallstone patients falls between 0.05% and 1%.2, 3 According to data from the National Information System on Hospital Care, the incidence of acute pancreatitis in The Netherlands increased by 30% in the period 1985-19954 and further increased by 35% in the period 1997-2002 (W. Hoogen Stoevenbelt, personal communication, 2004). Underlying causes could be increased gallstone prevalence, possibly related to increased prevalence of overweight and/or a higher proportion of elderly subjects in the population.

Limited data are available on the risk factors for pancreatitis in gallstone patients. Patients with small gallbladder stones might be at increased risk for pancreatitis.5, 6 In up to 70% of patients with acute “idiopathic” pancreatitis, cholesterol crystals have been found via polarizing light microscopy of bile obtained during endoscopic retrograde cholangiopancreatography or aspirated bile.7, 8 Cholesterol crystals might cause a functional obstruction at the sphincter of Oddi by inducing papillitis, spasm, or stenosis.9

Potential factors promoting cholesterol crystallization in the aqueous phase include high cholesterol saturation index, hydrophobic bile salts, and pronucleating biliary proteins. Also, patients with low amounts of biliary phospholipids exhibit fast crystallization due to deficient capacity of solubilizing vesicles, which is known as low phospholipidassociated cholelithiasis.10, 11 This patient category is at high risk for recurrence of biliary pain and gallstones after cholecystectomy,10 and possibly for acute pancreatitis.11 Missense mutations in the multidrug resistance protein 3 gene encoding for a phospholipid flippase in the hepatocyte canalicular membrane constitute the underlying defect.10, 11 Relative phospholipid deficiency could theoretically also be caused by high biliary phospholipase A2 activity.

Finally, strong postprandial gallbladder contraction, potentially related to increased release of cholecystokinin (CCK), could lead to pancreatitis via enhanced expulsion of gallbladder stones and crystals. High CCK concentrations could also promote pancreatitis by direct stimulation of the pancreas.

In the present study, we compared the potential factors in patients who had recovered from acute biliary pancreatitis and patients with uncomplicated symptomatic gallstones who were scheduled for cholecystectomy.


CCK, cholecystokinin.

Patients and Methods


Twenty-one consecutive patients with previous acute biliary pancreatitis participated in the study. Acute pancreatitis was defined by elevated serum amylase levels (at least three times the normal upper limit) and characteristic clinical findings. Biliary genesis was defined by elevated serum liver enzymes and/or bilirubin, and presence of gallbladder stones or sludge, in absence of alcohol abuse or other factors predisposing to pancreatitis. Thirty consecutive consenting patients with symptomatic but uncomplicated gallbladder stones, scheduled for elective cholecystectomy, served as the disease control group. Exclusion criteria were: history of hereditary pancreatitis, previous use of ursodeoxycholic acid, prior diagnosis of acute cholecystitis (n = 1), postoperative obstructive jaundice (n = 1), cholangitis or pancreatitis, or gallbladder bile infection (n = 3). Four patients in the control group but none in the pancreatitis group were excluded because of very low total lipid concentrations (“white biles”), indicating nonfunctioning gallbladders. Patients with previous endoscopic papillotomy were also excluded, because gallbladder motility and bile composition are significantly altered by this procedure.12 Clinical data from both groups are given in Table 1. Ranson scores13 in pancreatitis patients were 2.2 ± 0.4 (range: 0-4). The actual course of pancreatitis had been severe in one patient.

Table 1. Clinical and Gallstone Characteristics
 Pancreatitis PatientsSymptomatic Gallstone PatientsP
  • NOTE. Data are given as the mean ± SEM and median/range in case of marked nonparametric distributions or in percentage of patients.

  • Abbreviation: NS, not significant.

  • *

    Data available in 15 patients.

Clinical characteristics
 Age (years)47 ± 348 ± 4NS
 Gender (% males)2943NS
 Body mass index (kg/m2)26 ± 127 ± 1NS
Gallstone characteristics
 Sludge (% of patients)4113.03
 Smallest gallstone (mm)2 ± 18 ± 2.001
 (median: 1; range: 0.5–12)(median: 3; range: 0.5–31) 
 Largest gallstone (mm)7 ± 113 ± 1.004
 (median: 5; range: 0.5–18)(median: 11; range: 3–31) 
 Gallstone number (median)337NS
 Gallstone type*  
  Cholesterol stones (% of patients)8680NS
  Mixed stones (% of patients)73 
  Pigment stones (% of patients)717 

Written informed consent was obtained from all patients. Approval of the protocol was obtained from the ethical committees of the participating hospitals.

Gallstones and Bile.

Gallstones were obtained during cholecystectomy, washed with saline, dried, counted, and measured. Gallstones were divided into cholesterol (dry weight > 50% cholesterol), pigment (< 20% cholesterol), and mixed (20% < cholesterol < 50%) gallstone types.14 Gallbladder sludge was defined as sandy debris in the gallbladder visualized after cholecystectomy. We arbitrarily set gallbladder sludge size at 0.5 mm.

Particular care was taken to obtain complete gallbladder bile aspiration to avoid stratification effects.15 One milliliter of bile was immediately mixed with 6 mL chloroform/methanol (1/2 vol/vol), extracted according to Bligh and Dyer,16 and stored at −20°C. The remaining bile was stored immediately at −20°C or processed for nucleation experiments (see Microscopic Detection of Cholesterol Crystals). Cholecystectomy was performed 111 ± 33 days after recovery from pancreatitis. Gallbladder bile was lost during cholecystectomy in 4 pancreatitis and 5 symptomatic gallstone patients; stones were lost in 1 pancreatitis patient. Gallbladder biles of 2 symptomatic gallstone patients were not processed further because of blood contamination. One pancreatitis patient refused to undergo cholecystectomy. As a result, bile composition could be determined in 16 pancreatitis and 23 symptomatic gallstone patients. Sufficient bile for assessment of crystal observation time via polarizing light microscopy remained in 14 pancreatitis and 23 symptomatic gallstone patients. Sufficient gallstone material for determination of cholesterol content was available in 15 pancreatitis and 30 symptomatic gallstone patients.

Lipid Analysis.

A colorimetric assay was performed for analysis of biliary cholesterol (Boehringer, Mannheim, Germany).17 Phospholipid concentrations were assayed by determining inorganic phosphate.18 Total bile salt concentrations were measured enzymatically in whole bile using the 3α-hydroxysteroid dehydrogenase method.19 Cholesterol saturation index was derived using Carey's critical tables.20 Total lipid concentration was the sum of cholesterol, bile salt, and phospholipid concentrations expressed in g/dL. Conjugated bile salt species were analyzed in whole bile via isocratic high-performance liquid chromatography.21, 22 Cumulative hydrophobicity index for bile salt species was calculated according to Heuman.23 Phospholipid classes were isolated by thin-layer chromatography on silica gel (60-HR; Merck, Darmstadt, Germany).22

Microscopic Detection of Cholesterol Crystals.

Cholesterol crystal observation time was determined in ultra-filtered bile.24 Various crystal shapes were quantitated via daily examinations with the aid of polarizing microscopy and KOVA plastic slides (Hycor Biomedical, Garden Grove, CA) with 10 standardized examination chambers.25

Protein Analysis.

Protein concentrations in gallbladder bile were determined after delipidation26 via a colorimetric assay (BioRad Laboratories, Hercules, CA).27 Immunoglobulin M, immunoglobulin G, immunoglobulin A, α1-acid glycoprotein, and haptoglobin in gallbladder bile were quantitated using enzyme-linked immunosorbent assay.22

Mucin Determination.

Mucin concentrations in gallbladder bile were determined according to Miquel et al.28 In short, bile was fractionated using Sepharose 4B-Cl gel chromatography (30 × 1.5 cm column; Pharmacia, Uppsala, Sweden). To avoid coelution of cholesterol-phospholipid vesicles with high–molecular weight proteins in the void volume, 25 mmol/L of sodium taurocholate in phosphate-buffered saline was used as elution buffer. Fractions of 0.6 mL were collected. Mucin concentrations in individual fractions were determined using a fluorescent assay.29

Gallbladder Motility Studies.

Postprandial gallbladder motility was studied in all 30 symptomatic gallstone patients and in 18 pancreatitis patients 96 ± 35 days after recovery from pancreatitis. Three pancreatitis patients refused to participate in the gallbladder motility study. Fasting and postprandial gallbladder volumes were measured via ultrasonography using the sum-of-cylinders method.30 The standard semisolid mixed meal consisted of 30 g fat, 30 g protein, and 70 g carbohydrate (2,815 kJ). Fasting gallbladder volumes were defined as the means of three measurements at 5-minute intervals in the fasting state. After subsequent meal ingestion, gallbladder volumes were determined at 15-minute intervals during a 120-minute period. The following indices of gallbladder motor function were determined: fasting volumes (in mL), postprandial volumes (in mL and percentage of fasting volume)—including minimal postprandial volumes—and maximal decrease of gallbladder volumes (in mL). Patients were characterized as strong (minimum postprandial volume ≤6 mL) or weak (minimum postprandial volume >6 mL) contractors.31, 32

Plasma CCK Measurement.

Blood samples for the measurement of CCK concentrations were collected in the fasting state and at 15-minute intervals during a 2-hour period postprandially. Blood samples were collected in ice-chilled tubes and centrifuged at 3,000 rpm for 10 minutes, and the supernatant was stored at −20°C until analysis. CCK concentrations were determined via a sensitive and highly specific radioimmunoassay quantitating the bioactive forms of CCK (CCK-58, -33, -22, and -8) with equimolar potency.33 The following indices of CCK release were determined: fasting and maximal CCK concentrations, maximal increase of CCK concentration, and time to maximal CCK concentration.

Gallbladder Histopathology.

Histopathology was examined in hematoxylin-eosin–stained sections by an independent gastrointestinal pathologist unaware of patient history. Thickness was calculated as the mean of three representative measurements perpendicular to gallbladder corpus mucosal surface. A numerical score was used to quantify the extent of muscular coat fibrosis, cholesterolosis, Ashoff-Rokitansky sinuses, and transmural infiltration (Table 2).

Table 2. Gallbladder Pathology Scores
Histopathological ScorePancreatitis Patients (n = 20)Symptomatic Gallstone Patients (n = 30)
  1. NOTE. Muscular coat fibrosis, cholesterolosis, Ashoff-Rokitansky sinuses, and transmural infiltration were scored as follows: absent, 0; rare, 1; regularly present, 2; frequently present, 3.

Total thickness subserosal fibrosis–mucosal crypts (mm)1.4 ± 0.21.8 ± 0.2
Thickness subserosal fibrosis (mm)0.7 ± 0.10.9 ± 0.2
Thickness muscular coat (mm)0.7 ± 0.10.8 ± 0.1
Thickness muscular coat (% of total thickness)50 ± 351 ± 3
Fibrosis muscular coat (0–3)1.3 ± 0.21.3 ± 0.2
Cholesterolosis (0–3)0.6 ± 0.30.4 ± 0.2
Ashoff-Rokitansky sinuses (0–3)0.8 ± 0.30.7 ± 0.2
Transmural infiltration (0–3)1.1 ± 0.20.9 ± 0.1

Statistical Analysis.

Results are shown as means ± SEM and, in case of marked nonparametric distributions, also as medians with ranges. Differences were tested for statistical significance using the Student t test, Mann Whitney U test, and χ2 test as appropriate with the aid of NCSS software (Kaysville, UT). The temporal relationship of postprandial gallbladder volumes and CCK concentrations in pancreatitis versus symptomatic uncomplicated gallstone patients was tested via general linear models ANOVA as required in case of multiple comparisons. Univariate and multivariate regression analyses were performed to identify potential risk factors for pancreatitis. Statistical significance was defined as two-tailed probability less than .05.



The smallest gallbladder stones were smaller in pancreatitis patients than in those patients with symptomatic uncomplicated gallstones (mean: 2 ± 1 mm [median: 1; range: 0.5-12] vs. 8 ± 2 mm [median: 3; range: 0.5-31]; P = .001) (Fig. 1). Also, the largest gallbladder stones were much smaller in pancreatitis patients (mean: 7 ± 1 mm [median: 5; range: 0.5-18] vs. 13 ± 1 mm [median: 11; range: 3-31]; P = .004). Although the numbers of gallbladder stones were higher in pancreatitis patients (median: 33 vs. 7 in symptomatic gallstone patients), significance was not achieved because of a dispersion of data (P = .24). The presence of sludge occurred in 7 pancreatitis and 4 symptomatic gallstone patients (41% vs. 13%; P = .03). The cholesterol contents of gallbladder stones were virtually identical (70% ± 6% vs. 66% ± 6%; P = not significant). Also, distribution into gallstone types was comparable in pancreatitis and symptomatic gallstone patients (86% vs. 80%, 7% vs. 17%, and 7% vs. 3% in cholesterol, pigment, and mixed gallstone patients, respectively; P = not significant) (see Table 1).

Figure 1.

Size of smallest gallbladder stones. The sizes of the smallest gallbladder stones are smaller in patients with acute biliary pancreatitis compared with patients with symptomatic gallstones. Dots represent the smallest stones of individual patients. Horizontal lines represent the medians of the two groups. P = .001. GB, gallbladder, GS, gallstones.

Biliary Lipid Composition.

Total lipid concentrations in bile were similar in pancreatitis and symptomatic gallstone patients (9.3 ± 1.1 vs. 8.3 ± 0.7 g/dL) (Table 3). Also, molar percentages of bile salts, phospholipids, cholesterol, and cholesterol saturation indexes (134% ± 12% vs. 138 ± 10%) did not differ. Biliary bile salt compositions were virtually identical. In particular, the percentages of hydrophobic deoxycholate (25 ± 3% vs. 23 ± 1%) and cumulative bile salt hydrophobicity indexes (0.36 ± 0.01 vs. 0.35 ± 0.01) did not differ. Ratios of phospholipids/(bile salts + phospholipids) were also comparable (0.24 ± 0.01 vs. 0.25 ± 0.01). Biliary phospholipid class compositions were virtually identical (see Table 3). In particular, the percentages of lysophosphatidylcholine did not differ (1.6% ± 0.6% vs. 0.6% ± 0.1%; P = not significant).

Table 3. Biliary Lipid Composition
 Pancreatitis Patients (n = 16)Symptomatic Gallstone Patients (n = 23)P
  1. NOTE. Data are given as the mean ± SEM and median/range in case of marked nonparametric distributions.

  2. Abbreviations: NS, not significant; PL, phospholipids; BS, bile salts.

Total lipid concentration (g/dL)9.3 ± 1.18.3 ± 0.7NS
Bile salts (molar %)69 ± 268 ± 1NS
Phospholipids (molar %)22 ± 122 ± 1NS
Cholesterol (molar %)9 ± 19 ± 1NS
PL/(PL + BS) ratio0.24 ± 0.010.25 ± 0.01NS
 (median: 0.23; range: 0.17–0.37)(median: 0.24; range: 0.18–0.39) 
Cholesterol saturation index134 ± 12138 ± 10NS
 (median: 132; range: 70–264)(median: 122; range: 69–260) 
Taurodeoxycholate (%)6 ± 16 ± 1NS
Glycodeoxycholate (%)20 ± 318 ± 2NS
Hydrophobicity index0.36 ± 0.010.35 ± 0.01NS
Lysophosphatidylcholine (%)1.6 ± 0.60.6 ± 0.1NS
Sphingomyelin (%)0.9 ± 0.10.9 ± 0.1NS
Phosphatidylcholine (%)93 ± 0.894 ± 0.6NS
Phosphatidyl-iosine/serine (%)0.9 ± 0.10.8 ± 0.1NS
Phosphatidylethanolamine (%)2.1 ± 0.32.7 ± 0.3NS
Other phospholipids (%)1.9 ± 0.31.2 ± 0.2NS

Cholesterol Crystallization.

Cholesterol crystallization occurred much faster in pancreatitis patients than in those with symptomatic gallstones (crystal observation time: 1.0 ± 0.0 vs. 2.5 ± 0.4 days; P < .001) (Fig. 2). Crystals were observed within 1 day in all pancreatitis patients but in only 39% of symptomatic gallstone patients. The percentages of patients with non-plate-like (presumably anhydrous) crystals in gallbladder bile during the observation period were similar in pancreatitis and symptomatic gallstone patients (57% vs. 70%; P = not significant). The first occurrence of non-plate-like crystals coincided with the first occurrence of cholesterol monohydrate crystals in 2 pancreatitis and 5 symptomatic gallstone patients. In the remaining patients, non-plate-like crystals were observed only after the occurrence of cholesterol monohydrate crystals. Cumulative numbers of monohydrate crystals during observation were higher in pancreatitis patients (8.7 ± 6.1 vs. 6.9 ± 2.5 per nL), although significance was not achieved due to dispersion of data. Numbers of non-plate-like crystals were not significantly different (data not shown).

Figure 2.

Cholesterol crystal observation time. Crystal observation time is shorter in pancreatitis than in symptomatic gallstone patients. Dots represent the crystal observation times of individual patients. P < .001. GS, gallstones.

Biliary Protein Concentrations.

The concentrations of various pronucleating proteins (immunoglobulin M, immunoglobulin G, immunoglobulin A, α1-acid glycoprotein, and haptoglobin) were comparable (Table 4). However, mucin concentrations in gallbladder bile were higher in pancreatitis patients than in those patients with symptomatic gallstones (3.3 ± 1.9 vs. 0.8 ± 0.2 mg/mL; P = .04) (see Table 4). Univariate analysis revealed that cholesterol saturation index (r = 0.44; P = .005) or number of gallbladder stones (r = 0.85; P < .001), but not percentage of deoxycholate (r = −0.02; P = not significant) or bile salt hydrophobicity index (r = −0.11; P = not significant), were associated with mucin concentrations. However, multivariate analysis identified only number of gallbladder stones (P < .001) as an independent factor correlated to mucin concentrations.

Table 4. Biliary Protein Composition
 Pancreatitis Patients (n = 16)Symptomatic Gallstone Patients (n = 23)P
  1. NOTE. Data are given as the mean ± SEM, and also as the median and range (median; range).

Total protein (mg/mL)10.7 ± 3.65.3 ± 0.9.63
 (3.3; 1.8–49.6)(3.6; 0.6–18.8) 
Immunoglobulin M (mg/L)117 ± 2673 ± 52.39
 (103; 12–341)(57; 14–241) 
Immunoglobulin G (mg/L)260 ± 79277 ± 100.71
 (118; 12–974)(155; 12–2,394) 
Immunoglobulin A (mg/L)202 ± 41171 ± 17.98
 (150; 38–636)(150; 47–371) 
α1-Acid glycoprotein (mg/L)77 ± 2658 ± 7.41
 (43; 15–438)(51; 9–130) 
Haptoglobin (mg/L)44 ± 1627 ± 5.91
 (21; 5–239)(21; 2–85) 
Mucin (mg/mL)3.3 ± 1.90.8 ± 0.2.04
 (0.7; 0.3–30.1)(0.5; 0.2–2.8) 

Gallbladder Motility.

Fasting gallbladder volumes tended to be smaller in pancreatitis patients, without reaching significance (19.1 ± 2.1 vs. 24.0 ± 1.9 mL; P = .07) (Table 5). The differences in minimal postprandial gallbladder volumes between both groups were much more pronounced, with markedly lower values in pancreatitis patients (5.8 ± 1.0 vs. 8.1 ± 0.7 mL; P = .005). When patients were divided into strong (minimum postprandial volume ≤6 mL) or weak (minimum postprandial volume >6 mL)31, 32 contractors, 72% of pancreatitis patients but only 30% of symptomatic gallstone patients were strong contractors (P = .004). Gallbladder volumes as a function of time after meal ingestion were significantly lower in pancreatitis patients (P < .001) (Fig. 3). However, minimal postprandial gallbladder volumes as a percentage of fasting volume and maximal decrease of gallbladder volume were not significantly different (see Table 5).

Table 5. Fasting and Postprandial Gallbladder Volumes and CCK Concentrations
 Pancreatitis Patients (n = 18)Symptomatic Gallstone Patients (n = 30)P
  1. NOTE. Data are given as the mean ± SEM.

  2. Abbreviation: NS, not significant.

Fasting gallbladder volume (mL)19.1 ± 2.124.0 ± 1.9.07
Minimal gallbladder volume (mL)5.8 ± 1.08.1 ± 0.7.005
Strong contractors
 Minimal volume ≤6 mL (% of patients)7230.004
Minimal volume (% of fasting volume)33 ± 538 ± 3NS
Maximal decrease gallbladder volume (mL)13.4 ± 2.115.8 ± 1.8NS
Fasting CCK (pmol/L)1.2 ± 0.11.2 ± 0.1NS
Maximal postprandial CCK (pmol/L)5.2 ± 0.55.7 ± 0.9NS
Maximal CCK increase (pmol/L)3.9 ± 0.54.6 ± 0.8NS
CCK peak time after meal (min)72 ± 1078 ± 7NS
Figure 3.

Gallbladder emptying. Fasting and postprandial gallbladder volumes as determined by ultrasonography in patients with pancreatitis (○) and symptomatic gallstones (●). P < .001.

CCK Concentrations.

CCK concentrations in the fasting state were similar in pancreatitis and symptomatic gallstone patients (1.2 vs. 1.2 pmol/L) (see Table 5). Also, postprandial CCK concentrations as a function of time after meal ingestion were comparable in both groups as depicted in Fig. 4 (P = not significant). No differences existed between pancreatitis and symptomatic gallstone patients concerning maximal postprandial CCK concentrations, maximal increase of plasma CCK, or time to peak CCK concentrations.

Figure 4.

CCK concentrations. Fasting and postprandial plasma CCK concentrations are comparable in patients with pancreatitis (○) and symptomatic gallstones (●). P = not significant. CCK, cholecystokinin.

Gallbladder Histopathology.

Gallbladder histopathology did not reveal any difference between pancreatitis and symptomatic gallstone patients (see Table 2). Gallbladder inflammation was mild or absent in all patients.


The major finding of the present study is the association between acute pancreatitis and small gallbladder stones or sludge, preserved gallbladder motility, and fast cholesterol crystallization.

In this study, we excluded patients with previous endoscopic papillotomy, because gallbladder motility and bile composition are significantly altered by this procedure.12 Because papillotomy particularly benefits patients with severe pancreatitis, a potential bias of our patient group toward mild pancreatitis cannot be excluded.

The association between pancreatitis and small gallstones is not unexpected: small gallbladder stones might migrate into the common bile duct more easily than larger stones. Furthermore, small stones might cause obstruction in the common bile duct at the sphincter of Oddi, whereas larger stones might rather lead to more proximal obstruction with only obstructive jaundice. Another explanation for the association between small gallstones and pancreatitis could be increased cholesterol crystallization in the aqueous phase in case of small gallstones. Indeed, in supersaturated in vitro systems, competition for precipitation of cholesterol exists between the gallstone surface and the surrounding aqueous phase.34, 35 Further preliminary experiments indicate that the extent of cholesterol crystallization in the aqueous phase is low in supersaturated model systems containing large stones and high in systems containing small stones (Venneman et al., unpublished data).

Cholesterol crystals may cause functional obstruction at the sphincter of Oddi with subsequent reflux into the pancreatic duct.9 Indeed, preliminary experiments by our group in a rat model reveal markedly increased severity of pancreatitis when cholesterol crystals are added to a physiological bile salt mixture that is infused into the pancreatic duct.36

In the present study, pancreatitis patients exhibited much faster cholesterol crystallization than symptomatic gallstone patients. Because cholesterol saturation index, bile salt composition, biliary lipid concentrations, and levels of various pronucleating proteins were comparable in patients with and without pancreatitis, these factors appear not to be responsible for different crystallization speed. On the other hand, gallbladder mucin concentrations were higher in pancreatitis patients. In prairie dog and inbred mouse models for gallstone formation, mucin is the most prominent pronucleating protein in in vivo gallstone formation.37, 38 Apart from promoting crystallization, mucin is also an important component of biliary sludge (a potential risk factor for biliary pancreatitis).7, 39 Several MUC genes (MUC1, MUC3, MUC4, MUC5B, MUC6) are expressed in human gallbladder mucosa.40, 41 Upregulation of these genes could lead to increased gallbladder mucin concentrations. Although mucin secretion has been suggested to be stimulated by cholesterol supersaturation,22 hydrophobic bile salts,42 or gallbladder wall inflammation,43 we found no differences between pancreatitis and symptomatic gallstone patients in this respect. On the other hand, pathobiological effects of gallstones on the gallbladder mucosa could affect mucin secretion. Using multivariate analyses, we found an independent and significant correlation between the number of gallbladder stones and gallbladder mucin concentrations.

Gallstone patients generally have larger fasting and minimal postprandial gallbladder volumes compared with normal subjects.31, 32 Furthermore, in various conditions such as pregnancy,44 obesity,45 and treatment with somatostatin analogues,46 increased fasting and minimal postprandial volumes are associated with increased risk of gallstone formation. Similarly, in gallstone-susceptible male C57L inbred mice, both fasting and residual gallbladder volumes increase markedly during lithogenic diet, which is not the case in gallstone-resistant male AKR mice (van Erpecum and Wang, unpublished observations). Although fasting and minimal postprandial gallbladder volumes are correlated, minimal postprandial volumes appear to be particularly relevant for gallstone formation and recurrence after extracorporeal shock wave lithotripsy.31, 32 In one study, the risk for gallstone recurrence was increased ninefold if minimal residual volumes exceeded 5 mL.47 Although there was a nonsignificant trend toward smaller fasting volumes in pancreatitis patients in the current study, the differences in minimal postprandial volumes were much more pronounced (5.8 ± 1.0 vs. 8.1 ± 0.7 mL in the pancreatitis and symptomatic gallstone group, respectively [P = .005], which was 3.2 ± 0.7 mL in 23 asymptomatic gallstone-free subjects previously studied in our laboratory32). Apparently, gallbladder motility is more preserved in the pancreatitis group, though gallbladder volumes per se could theoretically be smaller in pancreatitis patients.

Gallbladder stones could migrate more easily from the gallbladder to the cystic or common bile duct if gallbladder contraction is preserved. Because postprandial CCK release and gallbladder histology were comparable in gallstone patients with or without pancreatitis, these factors do not explain gallbladder motility differences. Whether the number or sensitivity of CCK-A receptors in the gallbladder wall are different has yet to be elucidated.

Ursodeoxycholic acid may decrease the risk of recurrence after biliary pancreatitis.8 Increased fasting and residual postprandial gallbladder volumes,48, 49 decreased gallbladder mucin contents,22 and inhibition of cholesterol crystallization may be the underlying mechanisms. Since the landmark study of Gracie and Ransohoff,50 prophylactic cholecystectomy is generally not performed in asymptomatic gallstone patients. One may hypothesize that patients with small gallbladder stones and/or preserved gallbladder motility could benefit from prophylactic cholecystectomy, considering their apparently increased risk of pancreatitis. Although ursodeoxycholic acid could also be beneficial under these circumstances, long-term therapy would lead to considerable costs.

In conclusion, the present study shows that patients with small gallbladder stones, preserved gallbladder motility, and fast cholesterol crystallization (possibly related to high mucin concentrations) are at increased risk of acute pancreatitis. The potential benefit of prophylactic cholecystectomy or ursodeoxycholic acid in these patients has yet to be elucidated.


We are most grateful to the employees of the Departments of Surgery of the University Medical Center Utrecht, St. Antonius Hospital Nieuwegein, Academic Medical Center Amsterdam, and Jeroen Bosch Hospital's Hertogenbosch for their various contributions, and especially Drs. M. G. H. Besselink (University Medical Center Utrecht), M. I. van Berge Henegouwen (Academic Medical Center), and K. Bosscha (Jeroen Bosch Hospital's Hertogenbosch) for recruiting patients with biliary pancreatitis. We also thank Dr. M. E. I. Schipper of the Department of Pathology of the UMC Utrecht for histopathologic scoring of gallbladder tissues; M. B. M. de Smet of the Department of Surgery of the UMC Utrecht for measurement of bile salt species; and W. Hoogen Stoevenbeld of Prismant, The Netherlands, for supplying data on the incidence of pancreatitis.