- Email: [email protected]
Diagnostic criteria in predicting a biliary origin of acute pancreatitis in the era of endoscopic ultrasound: Multicentre prospective evaluation of 213 patients
Original Paper Pancreatology 2005;5:450–456 DOI: 10.1159/000086547
Received: September 20, 2004 Accepted after revision: February 3, 2005 Published online: June 28, 2005
Diagnostic Criteria in Predicting a Biliary Origin of Acute Pancreatitis in the Era of Endoscopic Ultrasound: Multicentre Prospective Evaluation of 213 Patients Philippe Lévya Arnaud Boruchowiczb Patrick Hastierc Alexandre Pariented Thierry Thévenote Jean Louis Frossardf Louis Buscailg François Mauvaish Jean Claude Duchmanni Alain Courrierj Philippe Buloisk Jean Louis Ginestonl Marc Barthetm Henri Lichtn Dermot O’Toolea Philippe Ruszniewskia a
Fédération médico-chirurgicale d’hépato-gastroentérologie, AP-HP, Hôpital Beaujon, Clichy, Service de Gastroentérologie, Centre Hospitalier Régional, Valenciennes, c Service de Gastroentérologie, Centre Hospitalier Régional, Pau; d Service de Gastroentérologie, Centre Hospitalier Universitaire, Nice, and e Service de Gastroentérologie, Centre Hospitalier Régional, Cambrai, France; f Centre Hospitalier Régional Universitaire, Genève, Switzerland; g Service de Gastroentérologie, Centre Hospitalier Universitaire, Toulouse, h Centre Hospitalier Régional, Beauvais; i Centre Hospitalier Régional, Compiègne, j Centre Hospitalier Régional, Metz, k Centre Hospitalier Universitaire, Lille, l Centre Hospitalier Régional, Rodez, m Centre Hospitalier Universitaire, Marseille, and n Centre Hospitalier Régional, Saint Denis, France b
Key Words Acute pancreatitis Biliary lithiasis Endoscopic ultrasonography
Abstract Background: No study on bioclinical criteria predicting a biliary origin for acute pancreatitis has included endosonography as a reference examination. Re-examination of bioclinical parameters deserves consideration in the era where other causes are known (e.g. hereditary, autoimmune). Aim and Methods: To determine the performance of bioclinical markers in predicting a biliary origin of acute pancreatitis where the diagnosis of biliary lithiasis was established or ruled out using endosonography. Only patients with a first acute episode of pancreatitis were included. Results: 213 patients (male: 55%; median age: 56 years) were prospectively included in 14
© 2005 S. Karger AG, Basel and IAP 1424–3903/05/0055–0450$22.00/0 Fax +41 61 306 12 34 E-Mail [email protected] www.karger.com
Accessible online at: www.karger.com/pan
centres. Causes of acute pancreatitis were: biliary (62%), alcoholic (25%), other (13%). Delay between symptomonset and admission was !48 h in 80%. Endosonography was the sole method establishing the diagnosis of biliary pancreatitis in 15% of patients. At univariate analysis, age, female sex, declared alcohol consumption, elevated aspartate and alanine transaminases on admission, gammaglutamyl transferase, alkaline phosphatase, total bilirubin, lipase, mean corpuscular volume were predictive of a biliary origin. Only age (p ! 0.0001), sex (p ! 0.0008) and alanine transaminase (p ! 0.0004) remained significant at multivariate analysis. At age 50, the respective sensitivity and specificity were 73 and 65%. With an elevated alanine transaminase at 2 times the upper limit of normal range, the respective sensitivity and specificity were 74 and 84%. The probability of a biliary origin of acute pancreatitis could be estimated by the following formula: = 1/1 + exp(4.6967 – 0.0656 x age + 1.1208 x
Prof. Philippe Lévy Fédération médico-chirurgicale d’hépato-gastroentérologie, Hôpital Beaujon FR–92118 Clichy Cedex (France) Tel. +33 1 40 87 52 15, Fax +33 1 42 70 37 84 E-Mail [email protected]
sex – 0.6909 x alanine transaminase). Conclusion: When endosonography is performed to confirm or exclude a biliary origin of acute pancreatitis, age, sex and alanine transaminase at admission are the only factors predictive of a biliary cause. Copyright © 2005 S. Karger AG, Basel and IAP
Introduction
The accurate diagnosis of the cause of acute pancreatitis is of particular importance and modifies the therapeutic strategy. This is especially true in the case of biliary pancreatitis which is severe in 15–20% of patients with a mortality rate of 15% [1]. The therapeutic choices may include emergency or elective endoscopic sphincterotomy [2–4], or cholecystectomy [5]. The diagnosis of biliary lithiasis is easy when gallbladder lithiasis is seen at transabdominal ultrasonography. However, biliary pancreatitis may be due to microlithiasis which is difficult to diagnose and may be responsible for severe forms of pancreatitis [6, 7]. If a biliary cause of pancreatitis is not recognized, the risk of pancreatitis relapse is approximately 30% after 6 months follow-up with unpredictable severity [5, 8, 9]. The importance of early recognition of biliary acute pancreatitis is thus crucial and much attention has been paid to the predictive value of epidemiological and biochemical parameters (age, sex, transaminases, alkaline phosphatase, amylase, etc.). Criteria used in this setting stem from studies performed dating to almost 20 years [10, 11], some of which were retrospective analyses [10]. In addition, over the past decade, other causes of acute pancreatitis have been described including metabolic, hereditary or autoimmune causes [12]. Current advances allow us to recognize such causes of pancreatitis and therefore, former studies probably overestimated the prevalence of so-called idiopathic pancreatitis. In light of this, it is also probable that cases of pancreatitis were erroneously attributed to biliary lithiasis. A further difficulty in applying criteria from older studies was absence of endoscopic ultrasound (EUS) as the ultimate procedure in the diagnosis of biliary lithiasis. EUS has been established as the most accurate procedure in the diagnosis of cholelithiasis or choledocolithiasis in patients with [6, 13–17] or without [6, 18] acute pancreatitis. The sensitivity and specificity of EUS remains excellent even in cases of microlithiasis [6, 19]. This may have lead to an underestimation of biliary causes in the
Prediction of Biliary Pancreatitis
setting of acute pancreatitis. Thus, reliable interpretation concerning the usefulness of bio-clinical markers in the prediction of the cause of acute pancreatitis may be undermined. This study aimed at reassessing the value of clinical and biochemical markers in patients with a first bout of acute pancreatitis. EUS was used as a reference examination for the diagnosis of biliary lithiasis in patients without evidence of biliary lithiasis at transabdominal ultrasonography and CT scan. Particular attention was paid to the delay between the onset of symptoms and the time of the first blood sampling.
Patients Patients were prospectively recruited from 14 French and Swiss centers from 01/2000 to 01/2002. All the recruited patients had to be able to undergo EUS and gave their informed and written consent to participate in the study. Inclusion Criteria All consecutive patients primarily hospitalized in the different centers with a first bout of acute pancreatitis defined by typical abdominal pain associated with one of the two following criteria were included: (a) elevation of serum lipase 13 times the upper limit of normal; (b) unequivocal signs of acute pancreatitis on CT scan (Balthazar’s severity index 63 [20]). The first attack was defined by the first clinical bout of pancreatitis as ascertained by careful appraisal of typical symptoms in all patients. The delay between onset of symptoms and hospitalization was recorded but was not an inclusion criteria. Non-Inclusion Criteria Patients with acute pancreatitis in the following setting were not included in the study: acute pancreatitis in post-operative setting or following endoscopic retrograde cholangio-pancreatography setting; recurrent cases of acute pancreatitis; patients with already established diagnosis of chronic pancreatitis as previously defined in other papers [21]; patients hospitalized in a primary center and then transferred to a secondary or tertiary center; contra-indication to EUS; previous cholecystectomy. Patients with two potential causes of acute pancreatitis (in general alcohol and biliary lithiasis) were included but their data were excluded from analysis since it was impossible to surely classify them in one of the two etiologic groups of the study (biliary, non-biliary). Data Recorded The following characteristics were recorded: age, sex, date and hour of clinical onset of acute episode of pain. Blood sampling was performed at admission and 48 8 2 h later. The following blood tests were performed: aspartate transaminase, alanine transaminase, gammaglutamyl transpeptidase, alkaline phosphatase, lactate deshydrogenase, bilirubin, serum amylase and lipase and mean corpuscular red blood cell volume.
Pancreatology 2005;5:450–456
451
Diagnostic Procedures for the Diagnosis of Biliary Lithiasis All patients had transabdominal ultrasonography and abdominal spiral CT scan with and without iodine contrast injection; CT scan thickness was 2.5 mm slice-width. When biliary lithiasis was not found at either trans abdominal ultrasonography or CT scan, EUS was performed by one skilful endoscopist in each center, using a Olympus GFUM20/EUM 20 device. The procedure was performed in patients who had taken at least two or more meals to avoid false-positive EUS results due to the formation of sludge in case of a prolonged fasting period. The examination technique for biliary lithiasis has been previously described in detail [19]. Briefly, the transducer was inserted into the inferior portion of the second duodenum and gradually withdrawn to duodenal bulb and stomach. The acoustic liaison between the transducer and the digestive wall was achieved by a balloon filled with 5–20 ml water. The gallbladder was examined through the duodenum and the stomach depending on anatomical variations using 7.5 and 12 MHz transducer frequencies. The common bile duct was visualized from the ampulla to porta hepaticus at the level of the second portion of the duodenum and duodenal bulb. The examination lasted 10–30 min and patients were sedated using intravenous propofol (ICI Pharma, Cergy, France) as commonly practiced in France. EUS criteria for cholelithiasis were: (a) the presence within the gallbladder of a hyperechoic semi-circular arch or circular shape associated with an acoustic shadow; (b) hyperechoic foci without acoustic shadow; (c) gallbladder sludge defined by mobile, low amplitude echoes seen in the lumen that layered the most dependent part of the gallbladder without acoustic shadowing [6, 22]. Criteria for choledocholithiasis were: (a) hyperechoic semicircular arch or circular shape focus/foci within the common bile duct associated with an acoustic shadow; (b) hyperechoic foci without acoustic shadow [18]. In patients with biliary lithiasis previously revealed by transabdominal ultrasonography or CT scan, additional EUS was left to the responsibility of the physicians in charge of the patients. Alcoholism was considered a potential cause when the daily alcohol consumption was 160 g/day for more than 2 years [21]. Other causes of pancreatitis were searched for by careful and systematic clinical data recording, with estimation of serum calcium and triglyceride levels in all patients. In addition, in patients less than 35 years old, a systematic search for CFTR, cationic trypsinogen and SPINK 1 mutation was performed [23–25]. At the end of etiological screening, patients were classified in the ‘biliary pancreatitis’ group if they had gallbladder or common bile duct lithiasis (as defined above) and no other obvious cause of pancreatitis. When none of the procedures including EUS showed biliary lithiasis, they were classified in the ‘non-biliary pancreatitis’. Clinical or biochemical parameters tested in the present study were not used to determine the cause of the pancreatitis. The study protocol was submitted to the Ethic Committee of Ambroise Paré hospital in Boulogne-Billancourt, France. As the procedure for searching the cause of the pancreatitis including EUS was considered as usual, the study protocol was accepted. Statistics Data are given as median and range. Comparisons were performed between groups of patients with biliary or non-biliary acute pancreatitis. Qualitative and quantitative data were compared by univariate logistic regression. p ! 0.05 was considered statistically significant. Multivariate analysis was performed using logistic re-
452
Pancreatology 2005;5:450–456
Table 1. General characteristics in patients with biliary and non-
biliary pancreatitis
Age, median years range Female sex: n (%) Male sex: n (%)
Biliary acute pancreatitis (n = 132)
Non-biliary acute pancreatitis (n = 81)
p
65.5 17–97 79 (84) 53 (45)
44.6 19.6–90.4 15 (16) 65 (55)
0.0001
0.0001
gression. A formula was established using the weighted factors determined by multivariate analysis to calculate the probability of a biliary origin of acute pancreatitis. We used the SAS version 8.2 software, issued by the SAS Institute, Cary, N.C., USA.
Results
Two hundred and twenty-four patients were prospectively evaluated. The median number of patients per center was 11 (1–43). There were no significant differences between the centers for age of patient, sex ratio, biliary/ non biliary cause ratio. Eleven patients were deemed excluded from analysis because they had two potential causes of acute pancreatitis (alcohol and biliary lithiasis; n = 9) or because EUS was not performed despite a negative transabdominal ultrasound and CT scan (n = 2). The following analysis thus included 213 patients. General Characteristics General characteristics of patients with biliary and non-biliary pancreatitis are given in table 1. There were 119 males and 94 females. The median age was 56 (17– 96) years. The delay between the onset of symptoms and admission was !24 h in 136 (64%) patients, 24–48 h in 34 (16%) and 148 h in 43 (20%). No significant difference for this delay was observed in patients with biliary and non biliary acute pancreatitis (logistic regression). Cause of Pancreatitis The cause of acute pancreatitis was biliary in 132 patients (62%); biliary lithiasis involved the gallbladder alone in 107 patients and both the common bile duct and gallbladder in 25 patients. Non-biliary causes were found in 81 patients (38%) (including alcoholic in 53 (25%), miscellaneous in 14 (7%) and idiopathic in 14 (7%)). All patients had ultrasonography and CT scan. EUS was per-
Lévy et al.
Table 2. Biochemical data at admission in patients with biliary and
non biliary pancreatitis (median and [range]) Biliary acute pancreatitis (n = 132)
AST (N) 4.9 [1–17] ALT (N) 4.2 [1–43] GGT (N) 6.0 [1–50] Alkaline phosphatases (N) 1.1 [1–29] Lacticodeshydrogenase (N) 1.3 [1–3] 27 [5–480] Bilirubin, mol/l Serum amylase (N) 9.6 [1–10] Serum lipase (N) 20.5 [1–50] Mean corpuscular red blood 91 [61–105] cell volume, l3
Non-biliary p acute pancreatitis (n = 81)
1.0 [1–12] 1.0 [1–12] 2.3 [1–42] 1.0 [1–5] 1.0 [1–2] 17 [3–361] 4.0 [1–8] 11.0 [1–15]
0.0001 0.0001 0.02 0.002 NS 0.03 0.0001 0.0003
95 [68–122]
0.0001
N = N times the upper limit of normal range. NS = Not significant.
Table 3. Biochemical data 48 h after admission in patients with
biliary and non-biliary pancreatitis (median and [range]) Biliary acute pancreatitis (n = 132)
AST (N) 1.3 [1–20] ALT (N) 2.5 [1–30] GGT (N) 4.7 [1–42] Alkaline phosphatase (N) 1.0 [1–12] Lacticodeshydrogenase (N) 1.0 [1–4] 19 [5–245] Bilirubin, mol/l Serum amylase (N) 2.0 [1–3] Serum lipase (N) 2.5 [1–6]
Non-biliary acute pancreatitis (n = 81)
p
1.0 [1–5] 1.0 [1–8] 2.2 [1–25] 1.0 [1–4] 1.0 [1–4] 17 [2–396] 1.3 [1–2] 2.5 [1–8]
NS NS NS NS NS NS NS NS
N = N times the upper limit of normal range. NS = Not significant.
Table 4. Diagnostic value of female sex, age and ALT at admission for the diagnosis of the biliary origin of an acute pancreatitis
Female sex Age >50 Age >60 ALT > 2N ALT > 3N
Sensitivity %
Specificity Positive predic- Negative predic% tive value, % tive value, %
60 73 61 74 61
82 65 79 84 91
84 77 82 88 92
N = N times the upper limit of normal range.
Prediction of Biliary Pancreatitis
56 60 52 66 59
formed in 140 patients including all patients (n = 81) with non-biliary and 59 with biliary acute pancreatitis. Among the 107 patients with cholelithiasis without common bile duct stones, the diagnosis was made by ultrasonography or CT scan in 94 (50 by ultrasonography alone, 3 by CT scan alone and 41 using both procedures). Diagnosis was made at EUS alone (negative ultrasonography and CT scan) in 13 patients and confirmed (positive ultrasonography or CT scan) in 32. Among the 25 patients with choledocolithiasis, the diagnosis was made by ultrasonography or CT scan in 18 (ultrasonography alone: 11; CT scan alone: 11; combined: 6). EUS established the presence of CBD stones in 7 patients where ultrasonography and CT scan were negative and confirmed the diagnosis in 7 other patients. In summary, the diagnosis of cholelithiasis or choledocolithiasis was established using EUS alone in 20 patients that is 15% of all the patients with biliary lithiasis. Comparison of General Characteristics and Biochemical Data in Patients with Biliary and Non-Biliary Pancreatitis Univariate Analysis Patients with biliary acute pancreatitis were older than that of those with non biliary pancreatitis. Female patients were more likely to have biliary pancreatitis (table 1). Biochemical data at admission and 48 h later are given in tables 2 and 3, respectively. Liver and pancreatic serum enzyme levels were higher at admission and mean corpuscular red blood cell volume was lower in patients with biliary pancreatitis. Lactate deshydrogenase level was not significantly different between the two groups. There was no significant difference for any biochemical parameters sampled 48 h after admission. Multivariate Analysis Only three parameters were independent variables predicting a biliary origin of an acute pancreatitis: age (p = 0.0001), sex (p = 0.0008), alanine transaminase at admission (p = 0.0004). Table 4 shows the sensitivity and specificity, positive and negative predictive value when using age, sex and alanine transaminase at different thresholds. The probability that an episode of acute pancreatitis was of biliary origin could be calculated using the formula issued from the multivariate analysis: 1/1 + exp(4.6967 – 0.0656 ! age + 1.1208 ! sex – 0.6909 ! alanine transaminase)
with sex = 1 for men, 0 for women, alanine transaminase expressed in X times the upper limit of normal range. For
Pancreatology 2005;5:450–456
453
example, the probabilities that an episode of acute pancreatitis is of biliary origin in a 65-year-old woman with alanine transaminase at 3 times the upper normal value at admission and in a 45-year-old man with 1.5 times the normal alanine transaminase at admission were 87 and 19%, respectively.
Discussion
Many studies have been devoted to the search for reliable criteria in establishing the diagnosis of a biliary origin in cases of acute pancreatitis. The main shortcoming in the previous studies was the lack of reliable cause of pancreatitis. Several studies were retrospective and included only few patients. In addition, the gold standards for the diagnosis of biliary lithiasis were variable: stones in gallbladder, pathological examination, issue of stone(s) during surgical or endoscopic examination of the common bile duct, etc. No unequivocal gold standard for excluding the diagnosis of microlithiasis was available. The diagnosis of biliary lithiasis is easy when stones can be visualized in the gallbladder or the common bile duct at transabdominal ultrasonography or CT scan. When these procedures do not demonstrate stones, accurate tests should confirm or definitively rule out the diagnosis of biliary lithiasis. While cholangio-MRI is also a useful tool in detecting biliary lithiasis, its performance in the detection of small stones (!5 mm) is still a matter of debate: stones smaller than 5 mm are still often missed by cholangio-MRI [19, 26, 27] and as such, we restricted our study to EUS whose accuracy remains superior to that of cholangio-MRI. It has been shown that EUS is essential in this setting since it can diagnose biliary lithiasis in no less than 45% of patients with biliary symptoms (including pancreatitis) in whom ultrasonography performed by senior radiologists failed to show any biliary lithiasis [6]. In another study by Tandon et al. [16], biliary lithiasis was found in 16% patients with so called idiopathic acute pancreatitis allowing a major change in the therapeutic strategy. In patients with pancreatitis, EUS has been shown to be 16–41% more sensitive than transabdominal ultrasonography [14, 15]. EUS is at least as sensitive with less technical failure than endoscopic retrograde cholangiography [14]. In addition, a high degree of skill is not mandatory to obtain satisfactory results in diagnosing common bile duct stones as operators with less than one year experience perform well with respect to this indication. In the present study, the diagnosis of biliary lithiasis was performed solely by EUS in 15% of patients, under-
454
Pancreatology 2005;5:450–456
scoring the necessity to perform this procedure in such cases. Patients with previous cholecystectomy were excluded because they are much less exposed to the risk of biliary pancreatitis. A further confounding factor in the setting of acute pancreatitis stems from the recent discovery in other etiopathogenic causes. In the present study, pancreatitis was considered as idiopathic in only 14/213 patients (7%). This result suggests that uncertainty regarding the cause of the pancreatitis was minimized as far as possible. The percentage of alcoholic pancreatitis (25%) is lower than in previous studies [28–30]. This is probably due to the inclusion of patients with a first bout of acute pancreatitis and the exclusion of patients with recurrent pancreatitis (many of whom are alcoholic). A bias for inclusion of patients with a potential biliary cause of acute pancreatitis may result from referral to centers with experience in EUS, however, this was avoided in the study design which included general hospital recruitment without specialist referred. In addition, the multicentric nature of the study would also appear to reflect the true incidence of biliary causes of acute pancreatitis. Clinical and biochemical parameters aimed at predicting a biliary origin of acute pancreatitis have been in use for some time. In the retrospective study by Blamey et al. [10], age (using a 50-year threshold) was an independent predictive factor. This was confirmed by Neoptolemos et al. [11] who established a 70-year threshold. In another study assessing the risk of choledocholithiasis before laparoscopic cholecystectomy, 55 years of age was also found to be predictive [31]. Female sex is also predictive of a biliary origin in acute pancreatitis [10]; women have a 2fold increased risk [32]. The present study confirms that age and female sex are predictive factors of the biliary origin of an acute pancreatitis. The accuracy of biochemical parameters has rarely been assessed in series with sufficient numbers of patients. Several blood parameters have been evaluated including total bilirubin, serum amylase, alkaline phosphatase, alanine and aspartate transaminase levels. Overall, results are discrepant and depend on the thresholds used and the delay between the onset of pancreatitis and blood sampling. Bilirubin level was found to be predictive of biliary pancreatitis in only one study [11] but more often in cases of a retained common bile duct stone (threshold: 25–30 mol/l) [33–35]. It is well established however that biliary stones rapidly migrate through the ampulla in patients with acute biliary pancreatitis [8, 36] and thus elevated liver blood tests tend to normalize with time. The significance of elevated alkaline phosphatase level is sim-
Lévy et al.
ilar to that of bilirubin [11, 34] and did not predict biliary pancreatitis except in the study by Blamey et al. [10]. Serum amylase has not been found to be a reliable predictive criterion as conflicting data have emerged in the literature [10, 37]. Application of a serum lipase/amylase ratio has been largely debated but was finally not considered a useful biochemical index in the setting of acute biliary pancreatitis [38, 39]. To date, much confusion has also concerned the diagnostic performance of serum aspartate and alanine transaminase levels as useful predictive criteria. Alanine transaminase was more accurate than aspartate transaminase in one meta-analysis [37] and in a recent retrospective study [40]. The threshold was found to be between 1.2 and 3 times the upper normal value [10, 37, 40]. However, neither aspartate nor alanine transaminase were found to be accurate predictive factors of a biliary cause in the studies by Blamey et al. and Neoptolemos et al. [10, 11]. The results of the present study support previous findings by both Grau et al. [40] and Tenner et al. [37]. Elevated alanine transaminase level was the only useful predictive parameter found in the meta-analysis by Tenner et al. [37]; using 3 times the upper normal value as a threshold, the positive predictive value was excellent (95%) but the sensitivity was low [37]. With transaminases at 1.2 times normal threshold, the sensitivity was 73% with a positive predictive value of 92% [40]. Our results confirm these figures in a large prospective study. One of the major problems in all studies assessing the usefulness of biochemical parameters is the delay between the onset of pancreatitis and blood sampling. In the present study, 80% of admission blood samplings were
taken less than 48 h after the beginning of pain. None of the tested parameters was significantly associated with biliary pancreatitis, 2 days later. This result underlines the necessity of rapid sampling in patients with suspicion of acute pancreatitis. Application of these simple strategies should have a large impact in overall management of such patients and should prompt appropriate biliary investigations including EUS in those with negative transabdominal ultrasound and CT scan. The negative predictive values of the three selected parameters are relatively low whatever the threshold. Therefore, a negative result should not be considered as excluding the diagnosis of biliary pancreatitis. In conclusion, in this large prospective study using EUS as the gold standard examination for the diagnosis of biliary lithiasis, age, female sex and alanine transaminase levels (sampled at admission) were predictive of biliary origin of a first episode of acute pancreatitis. These parameters should be useful in indicating non-invasive or invasive diagnostic procedures. However, a definitive exclusion of biliary lithiasis should still rely on radiological procedures as EUS.
Acknowledgments We would like to thank Dr. F. Dyard (Solvay Pharma) who kindly gave us financial support for statistical analysis, F. Sévenier, Marie-Anne Caillaud, N. Le Gall (Fovéa company) for their help in statistical analysis and ‘L’association des hépato-gastroentérologues des hôpitaux généraux’ (ANGH) for helping in the recruitment of centers.
References 1 Boyer J: Faut-il réaliser une CPRE avec sphinctérotomie en urgence? Gastroenterol Clin Biol 2001;25:1S122–1S127. 2 Folsch UR, Nitsche R, Ludtke R, Hilgers RA, Creutzfeldt W: Early ERCP and papillotomy compared with conservative treatment for acute biliary pancreatitis. The German Study Group on Acute Biliary Pancreatitis. N Engl J Med 1997;336:237–242. 3 Neoptolemos JP, Carr-Locke DL, London NJ, Bailey IA, James D, Fossard DP: Controlled trial of urgent endoscopic retrograde cholangiopancreatography and endoscopic sphincterotomy versus conservative treatment for acute pancreatitis due to gallstones. Lancet 1988;ii: 979–983.
Prediction of Biliary Pancreatitis
4 Fan ST, Lai EC, Mok FP, Lo CM, Zheng SS, Wong J: Early treatment of acute biliary pancreatitis by endoscopic papillotomy. N Engl J Med 1993;328:228–232. 5 Gambiez L: Quand et comment traiter la lithiase biliaire? Gastroenterol Clin Biol 2001; 25: 1S128–1S139. 6 Dahan P, Andant C, Levy P, Amouyal P, Amouyal G, Dumont M, Erlinger S, Sauvanet A, Belghiti J, Zins M, Vilgrain V, Bernades P: Prospective evaluation of endoscopic ultrasonography and microscopic examination of duodenal bile in the diagnosis of cholecystolithiasis in 45 patients with normal conventional ultrasonography. Gut 1996;38:277–281.
7 Nowak A, Marek TA, Nowakowska-Dulawa E, Rybicka J, Kaczor R: Biliary pancreatitis needs endoscopic retrograde cholangiopancreatography with endoscopic sphincterotomy for cure. Endoscopy 1998;30:A256–A259. 8 Kelly TR: Gallstone pancreatitis: The timing of surgery. Surgery 1980;88:345–350. 9 Osborne DH, Imrie CW, Carter DC: Biliary surgery in the same admission for gallstone-associated acute pancreatitis. Br J Surg 1981;68: 758–761. 10 Blamey SL, Osborne DH, Gilmour WH, O’Neill J, Carter DC, Imrie CW: The early identification of patients with gallstone associated pancreatitis using clinical and biochemical factors only. Ann Surg 1983;198:574–578.
Pancreatology 2005;5:450–456
455
11 Neoptolemos JP, London N, Bailey I, Shaw D, Carr-Locke DL, Fossard DP, Moossa AR: The role of clinical and biochemical criteria and endoscopic retrograde cholangiopancreatography in the urgent diagnosis of common bile duct stones in acute pancreatitis. Surgery 1986; 100:732–742. 12 Conférence de Consensus: Pancréatite aiguë. Gastroenterol Clin Biol 2001;25:177–192. 13 Novotny I, Lata J: Chronic pancreatitis in endoscopic ultrasonography – analysis of newly diagnosed cases during a 28-month period. Bratisl Lek Listy 2000;101:649–653. 14 Liu CL, Lo CM, Chan JK, Poon RT, Lam CM, Fan ST, Wong J: Detection of choledocholithiasis by EUS in acute pancreatitis: A prospective evaluation in 100 consecutive patients. Gastrointest Endosc 2001;54:325–330. 15 Chak A, Hawes RH, Cooper GS, Hoffman B, Catalano MF, Wong RC, Herbener TE, Sivak MV Jr: Prospective assessment of the utility of EUS in the evaluation of gallstone pancreatitis. Gastrointest Endosc 1999;49:599–604. 16 Tandon M, Topazian M: Endoscopic ultrasound in idiopathic acute pancreatitis. Am J Gastroenterol 2001;96:705–709. 17 Rosch T, Mayr P, Kassem MA: Endoscopic ultrasonography in acute biliary pancreatitis. J Gastrointest Surg 2001;5:223–228. 18 Amouyal P, Amouyal G, Levy P, Tuzet S, Palazzo L, Vilgrain V, Gayet B, Belghiti J, Fekete F, Bernades P: Diagnosis of choledocholithiasis by endoscopic ultrasonography. Gastroenterology 1994;106:1062–1067. 19 Palazzo L, O’Toole D: EUS in common bile duct stones. Gastrointest Endosc 2002; 56: S49–S57. 20 Balthazar EJ, Robinson DL, Megibow AJ, Ranson JH: Acute pancreatitis: value of CT in establishing prognosis. Radiology 1990; 174: 331–336. 21 Levy P, Milan C, Pignon JP, Baetz A, Bernades P: Mortality factors associated with chronic pancreatitis: Unidimensional and multidimensional analysis of a medical-surgical series of 240 patients. Gastroenterology 1989; 96: 1165–1172.
456
22 Lee SP, Nicholls JF, Park HZ: Biliary sludge as a cause of acute pancreatitis. N Engl J Med 1992;326:589–593. 23 Maire F, Bienvenu T, Ngukam A, Hammel P, Ruszniewski P, Levy P: Fréquence des mutations du gène CFTR dans la pancréatite chronique idiopathique. Gastroenterol Clin Biol 2003;27:398–402. 24 Chen JM, Piepoli Bis A, Le Bodic L, Ruszniewski P, Robaszkiewicz M, Deprez PH, Raguenes O, Quere I, Andriulli A, Ferec C: Mutational screening of the cationic trypsinogen gene in a large cohort of subjects with idiopathic chronic pancreatitis. Clin Genet 2001;59:189–193. 25 Ferec C, Raguenes O, Salomon R, Roche C, Bernard JP, Guillot M, Quere I, Faure C, Mercier B, Audrezet MP, Guillausseau PJ, Dupont C, Munnich A, Bignon JD, Le Bodic L: Mutations in the cationic trypsinogen gene and evidence for genetic heterogeneity in hereditary pancreatitis. J Med Genet 1999;36:228–232. 26 de Ledinghen V, Lecesne R, Raymond JM, Gense V, Amouretti M, Drouillard J, Couzigou P, Silvain C: Diagnosis of choledocholithiasis: EUS or magnetic resonance cholangiography? A prospective controlled study. Gastrointest Endosc 1999;49:26–31. 27 Zidi SH, Prat F, Le Guen O, Rondeau Y, Rocher L, Fritsch J, Choury AD, Pelletier G: Use of magnetic resonance cholangiography in the diagnosis of choledocholithiasis: Prospective comparison with a reference imaging method. Gut 1999;44:118–122. 28 Lankisch PG, Assmus C, Maisonneuve P, Lowenfels AB: Epidemiology of pancreatic diseases in Luneburg County: A study in a defined german population. Pancreatology 2002; 2: 469–477. 29 Gullo L, Migliori M, Olah A, Farkas G, Levy P, Arvanitakis C, Lankisch P, Beger H: Acute pancreatitis in five European countries: Etiology and mortality. Pancreas 2002; 24: 223– 227. 30 Gullo L, Migliori M, Pezzilli R, Olah A, Farkas G, Levy P, Arvanitakis C, Lankisch P, Beger H: An update on recurrent acute pancreatitis: data from five European countries. Am J Gastroenterol 2002;97:1959–1962.
Pancreatology 2005;5:450–456
31 Barkun AN, Barkun JS, Fried GM, Ghitulescu G, Steinmetz O, Pham C, Meakins JL, Goresky CA: Useful predictors of bile duct stones in patients undergoing laparoscopic cholecystectomy. McGill Gallstone Treatment Group. Ann Surg 1994;220:32–39. 32 Appelros S, Borgstrom A: Incidence, aetiology and mortality rate of acute pancreatitis over 10 years in a defined urban population in Sweden. Br J Surg 1999;86:465–470. 33 Hauer-Jensen M, Karesen R, Nygaard K, Solheim K, Amlie EJ, Havig O, Rosseland AR: Prospective randomized study of routine intraoperative cholangiography during open cholecystectomy: Long-term follow-up and multivariate analysis of predictors of choledocholithiasis. Surgery 1993;113:318–323. 34 Onken JE, Brazer SR, Eisen GM, Williams DM, Bouras EP, DeLong ER, Long TT 3rd, Pancotto FS, Rhodes DL, Cotton PB: Predicting the presence of choledocholithiasis in patients with symptomatic cholelithiasis. Am J Gastroenterol 1996;91:762–767. 35 Abboud PA, Malet PF, Berlin JA, Staroscik R, Cabana MD, Clarke JR, Shea JA, Schwartz JS, Williams SV: Predictors of common bile duct stones prior to cholecystectomy: A meta-analysis. Gastrointest Endosc 1996;44:450–455. 36 Acosta JM, Ledesma CL: Gallstone migration as a cause of acute pancreatitis. N Engl J Med 1974;290:484–487. 37 Tenner S, Dubner H, Steinberg W: Predicting gallstone pancreatitis with laboratory parameters: A meta-analysis. Am J Gastroenterol 1994;89:1863–1866. 38 Pezzilli R, Billi P, Barakat B, Miglio F: Lipaseamylase ratio does not determine the etiology of acute pancreatitis. Another myth bites the dust. J Clin Gastroenterol 1998;26:34–38. 39 Lankisch PG, Petersen M: Lipase/amylase ratio: Not helpful in the early etiological differentiation of acute pancreatitis. Z Gastroenterol 1994;32:8–11. 40 Grau F, Almela P, Aparisi L, Bautista D, Pascual I, Pena A, Rodrigo JM: Usefulness of alanine and aspartate aminotransferases in the diagnosis of microlithiasis in idiopathic acute pancreatitis. Int J Pancreatol 1999;25:107–111.
Lévy et al.
Received: September 20, 2004 Accepted after revision: February 3, 2005 Published online: June 28, 2005
Diagnostic Criteria in Predicting a Biliary Origin of Acute Pancreatitis in the Era of Endoscopic Ultrasound: Multicentre Prospective Evaluation of 213 Patients Philippe Lévya Arnaud Boruchowiczb Patrick Hastierc Alexandre Pariented Thierry Thévenote Jean Louis Frossardf Louis Buscailg François Mauvaish Jean Claude Duchmanni Alain Courrierj Philippe Buloisk Jean Louis Ginestonl Marc Barthetm Henri Lichtn Dermot O’Toolea Philippe Ruszniewskia a
Fédération médico-chirurgicale d’hépato-gastroentérologie, AP-HP, Hôpital Beaujon, Clichy, Service de Gastroentérologie, Centre Hospitalier Régional, Valenciennes, c Service de Gastroentérologie, Centre Hospitalier Régional, Pau; d Service de Gastroentérologie, Centre Hospitalier Universitaire, Nice, and e Service de Gastroentérologie, Centre Hospitalier Régional, Cambrai, France; f Centre Hospitalier Régional Universitaire, Genève, Switzerland; g Service de Gastroentérologie, Centre Hospitalier Universitaire, Toulouse, h Centre Hospitalier Régional, Beauvais; i Centre Hospitalier Régional, Compiègne, j Centre Hospitalier Régional, Metz, k Centre Hospitalier Universitaire, Lille, l Centre Hospitalier Régional, Rodez, m Centre Hospitalier Universitaire, Marseille, and n Centre Hospitalier Régional, Saint Denis, France b
Key Words Acute pancreatitis Biliary lithiasis Endoscopic ultrasonography
Abstract Background: No study on bioclinical criteria predicting a biliary origin for acute pancreatitis has included endosonography as a reference examination. Re-examination of bioclinical parameters deserves consideration in the era where other causes are known (e.g. hereditary, autoimmune). Aim and Methods: To determine the performance of bioclinical markers in predicting a biliary origin of acute pancreatitis where the diagnosis of biliary lithiasis was established or ruled out using endosonography. Only patients with a first acute episode of pancreatitis were included. Results: 213 patients (male: 55%; median age: 56 years) were prospectively included in 14
© 2005 S. Karger AG, Basel and IAP 1424–3903/05/0055–0450$22.00/0 Fax +41 61 306 12 34 E-Mail [email protected] www.karger.com
Accessible online at: www.karger.com/pan
centres. Causes of acute pancreatitis were: biliary (62%), alcoholic (25%), other (13%). Delay between symptomonset and admission was !48 h in 80%. Endosonography was the sole method establishing the diagnosis of biliary pancreatitis in 15% of patients. At univariate analysis, age, female sex, declared alcohol consumption, elevated aspartate and alanine transaminases on admission, gammaglutamyl transferase, alkaline phosphatase, total bilirubin, lipase, mean corpuscular volume were predictive of a biliary origin. Only age (p ! 0.0001), sex (p ! 0.0008) and alanine transaminase (p ! 0.0004) remained significant at multivariate analysis. At age 50, the respective sensitivity and specificity were 73 and 65%. With an elevated alanine transaminase at 2 times the upper limit of normal range, the respective sensitivity and specificity were 74 and 84%. The probability of a biliary origin of acute pancreatitis could be estimated by the following formula: = 1/1 + exp(4.6967 – 0.0656 x age + 1.1208 x
Prof. Philippe Lévy Fédération médico-chirurgicale d’hépato-gastroentérologie, Hôpital Beaujon FR–92118 Clichy Cedex (France) Tel. +33 1 40 87 52 15, Fax +33 1 42 70 37 84 E-Mail [email protected]
sex – 0.6909 x alanine transaminase). Conclusion: When endosonography is performed to confirm or exclude a biliary origin of acute pancreatitis, age, sex and alanine transaminase at admission are the only factors predictive of a biliary cause. Copyright © 2005 S. Karger AG, Basel and IAP
Introduction
The accurate diagnosis of the cause of acute pancreatitis is of particular importance and modifies the therapeutic strategy. This is especially true in the case of biliary pancreatitis which is severe in 15–20% of patients with a mortality rate of 15% [1]. The therapeutic choices may include emergency or elective endoscopic sphincterotomy [2–4], or cholecystectomy [5]. The diagnosis of biliary lithiasis is easy when gallbladder lithiasis is seen at transabdominal ultrasonography. However, biliary pancreatitis may be due to microlithiasis which is difficult to diagnose and may be responsible for severe forms of pancreatitis [6, 7]. If a biliary cause of pancreatitis is not recognized, the risk of pancreatitis relapse is approximately 30% after 6 months follow-up with unpredictable severity [5, 8, 9]. The importance of early recognition of biliary acute pancreatitis is thus crucial and much attention has been paid to the predictive value of epidemiological and biochemical parameters (age, sex, transaminases, alkaline phosphatase, amylase, etc.). Criteria used in this setting stem from studies performed dating to almost 20 years [10, 11], some of which were retrospective analyses [10]. In addition, over the past decade, other causes of acute pancreatitis have been described including metabolic, hereditary or autoimmune causes [12]. Current advances allow us to recognize such causes of pancreatitis and therefore, former studies probably overestimated the prevalence of so-called idiopathic pancreatitis. In light of this, it is also probable that cases of pancreatitis were erroneously attributed to biliary lithiasis. A further difficulty in applying criteria from older studies was absence of endoscopic ultrasound (EUS) as the ultimate procedure in the diagnosis of biliary lithiasis. EUS has been established as the most accurate procedure in the diagnosis of cholelithiasis or choledocolithiasis in patients with [6, 13–17] or without [6, 18] acute pancreatitis. The sensitivity and specificity of EUS remains excellent even in cases of microlithiasis [6, 19]. This may have lead to an underestimation of biliary causes in the
Prediction of Biliary Pancreatitis
setting of acute pancreatitis. Thus, reliable interpretation concerning the usefulness of bio-clinical markers in the prediction of the cause of acute pancreatitis may be undermined. This study aimed at reassessing the value of clinical and biochemical markers in patients with a first bout of acute pancreatitis. EUS was used as a reference examination for the diagnosis of biliary lithiasis in patients without evidence of biliary lithiasis at transabdominal ultrasonography and CT scan. Particular attention was paid to the delay between the onset of symptoms and the time of the first blood sampling.
Patients Patients were prospectively recruited from 14 French and Swiss centers from 01/2000 to 01/2002. All the recruited patients had to be able to undergo EUS and gave their informed and written consent to participate in the study. Inclusion Criteria All consecutive patients primarily hospitalized in the different centers with a first bout of acute pancreatitis defined by typical abdominal pain associated with one of the two following criteria were included: (a) elevation of serum lipase 13 times the upper limit of normal; (b) unequivocal signs of acute pancreatitis on CT scan (Balthazar’s severity index 63 [20]). The first attack was defined by the first clinical bout of pancreatitis as ascertained by careful appraisal of typical symptoms in all patients. The delay between onset of symptoms and hospitalization was recorded but was not an inclusion criteria. Non-Inclusion Criteria Patients with acute pancreatitis in the following setting were not included in the study: acute pancreatitis in post-operative setting or following endoscopic retrograde cholangio-pancreatography setting; recurrent cases of acute pancreatitis; patients with already established diagnosis of chronic pancreatitis as previously defined in other papers [21]; patients hospitalized in a primary center and then transferred to a secondary or tertiary center; contra-indication to EUS; previous cholecystectomy. Patients with two potential causes of acute pancreatitis (in general alcohol and biliary lithiasis) were included but their data were excluded from analysis since it was impossible to surely classify them in one of the two etiologic groups of the study (biliary, non-biliary). Data Recorded The following characteristics were recorded: age, sex, date and hour of clinical onset of acute episode of pain. Blood sampling was performed at admission and 48 8 2 h later. The following blood tests were performed: aspartate transaminase, alanine transaminase, gammaglutamyl transpeptidase, alkaline phosphatase, lactate deshydrogenase, bilirubin, serum amylase and lipase and mean corpuscular red blood cell volume.
Pancreatology 2005;5:450–456
451
Diagnostic Procedures for the Diagnosis of Biliary Lithiasis All patients had transabdominal ultrasonography and abdominal spiral CT scan with and without iodine contrast injection; CT scan thickness was 2.5 mm slice-width. When biliary lithiasis was not found at either trans abdominal ultrasonography or CT scan, EUS was performed by one skilful endoscopist in each center, using a Olympus GFUM20/EUM 20 device. The procedure was performed in patients who had taken at least two or more meals to avoid false-positive EUS results due to the formation of sludge in case of a prolonged fasting period. The examination technique for biliary lithiasis has been previously described in detail [19]. Briefly, the transducer was inserted into the inferior portion of the second duodenum and gradually withdrawn to duodenal bulb and stomach. The acoustic liaison between the transducer and the digestive wall was achieved by a balloon filled with 5–20 ml water. The gallbladder was examined through the duodenum and the stomach depending on anatomical variations using 7.5 and 12 MHz transducer frequencies. The common bile duct was visualized from the ampulla to porta hepaticus at the level of the second portion of the duodenum and duodenal bulb. The examination lasted 10–30 min and patients were sedated using intravenous propofol (ICI Pharma, Cergy, France) as commonly practiced in France. EUS criteria for cholelithiasis were: (a) the presence within the gallbladder of a hyperechoic semi-circular arch or circular shape associated with an acoustic shadow; (b) hyperechoic foci without acoustic shadow; (c) gallbladder sludge defined by mobile, low amplitude echoes seen in the lumen that layered the most dependent part of the gallbladder without acoustic shadowing [6, 22]. Criteria for choledocholithiasis were: (a) hyperechoic semicircular arch or circular shape focus/foci within the common bile duct associated with an acoustic shadow; (b) hyperechoic foci without acoustic shadow [18]. In patients with biliary lithiasis previously revealed by transabdominal ultrasonography or CT scan, additional EUS was left to the responsibility of the physicians in charge of the patients. Alcoholism was considered a potential cause when the daily alcohol consumption was 160 g/day for more than 2 years [21]. Other causes of pancreatitis were searched for by careful and systematic clinical data recording, with estimation of serum calcium and triglyceride levels in all patients. In addition, in patients less than 35 years old, a systematic search for CFTR, cationic trypsinogen and SPINK 1 mutation was performed [23–25]. At the end of etiological screening, patients were classified in the ‘biliary pancreatitis’ group if they had gallbladder or common bile duct lithiasis (as defined above) and no other obvious cause of pancreatitis. When none of the procedures including EUS showed biliary lithiasis, they were classified in the ‘non-biliary pancreatitis’. Clinical or biochemical parameters tested in the present study were not used to determine the cause of the pancreatitis. The study protocol was submitted to the Ethic Committee of Ambroise Paré hospital in Boulogne-Billancourt, France. As the procedure for searching the cause of the pancreatitis including EUS was considered as usual, the study protocol was accepted. Statistics Data are given as median and range. Comparisons were performed between groups of patients with biliary or non-biliary acute pancreatitis. Qualitative and quantitative data were compared by univariate logistic regression. p ! 0.05 was considered statistically significant. Multivariate analysis was performed using logistic re-
452
Pancreatology 2005;5:450–456
Table 1. General characteristics in patients with biliary and non-
biliary pancreatitis
Age, median years range Female sex: n (%) Male sex: n (%)
Biliary acute pancreatitis (n = 132)
Non-biliary acute pancreatitis (n = 81)
p
65.5 17–97 79 (84) 53 (45)
44.6 19.6–90.4 15 (16) 65 (55)
0.0001
0.0001
gression. A formula was established using the weighted factors determined by multivariate analysis to calculate the probability of a biliary origin of acute pancreatitis. We used the SAS version 8.2 software, issued by the SAS Institute, Cary, N.C., USA.
Results
Two hundred and twenty-four patients were prospectively evaluated. The median number of patients per center was 11 (1–43). There were no significant differences between the centers for age of patient, sex ratio, biliary/ non biliary cause ratio. Eleven patients were deemed excluded from analysis because they had two potential causes of acute pancreatitis (alcohol and biliary lithiasis; n = 9) or because EUS was not performed despite a negative transabdominal ultrasound and CT scan (n = 2). The following analysis thus included 213 patients. General Characteristics General characteristics of patients with biliary and non-biliary pancreatitis are given in table 1. There were 119 males and 94 females. The median age was 56 (17– 96) years. The delay between the onset of symptoms and admission was !24 h in 136 (64%) patients, 24–48 h in 34 (16%) and 148 h in 43 (20%). No significant difference for this delay was observed in patients with biliary and non biliary acute pancreatitis (logistic regression). Cause of Pancreatitis The cause of acute pancreatitis was biliary in 132 patients (62%); biliary lithiasis involved the gallbladder alone in 107 patients and both the common bile duct and gallbladder in 25 patients. Non-biliary causes were found in 81 patients (38%) (including alcoholic in 53 (25%), miscellaneous in 14 (7%) and idiopathic in 14 (7%)). All patients had ultrasonography and CT scan. EUS was per-
Lévy et al.
Table 2. Biochemical data at admission in patients with biliary and
non biliary pancreatitis (median and [range]) Biliary acute pancreatitis (n = 132)
AST (N) 4.9 [1–17] ALT (N) 4.2 [1–43] GGT (N) 6.0 [1–50] Alkaline phosphatases (N) 1.1 [1–29] Lacticodeshydrogenase (N) 1.3 [1–3] 27 [5–480] Bilirubin, mol/l Serum amylase (N) 9.6 [1–10] Serum lipase (N) 20.5 [1–50] Mean corpuscular red blood 91 [61–105] cell volume, l3
Non-biliary p acute pancreatitis (n = 81)
1.0 [1–12] 1.0 [1–12] 2.3 [1–42] 1.0 [1–5] 1.0 [1–2] 17 [3–361] 4.0 [1–8] 11.0 [1–15]
0.0001 0.0001 0.02 0.002 NS 0.03 0.0001 0.0003
95 [68–122]
0.0001
N = N times the upper limit of normal range. NS = Not significant.
Table 3. Biochemical data 48 h after admission in patients with
biliary and non-biliary pancreatitis (median and [range]) Biliary acute pancreatitis (n = 132)
AST (N) 1.3 [1–20] ALT (N) 2.5 [1–30] GGT (N) 4.7 [1–42] Alkaline phosphatase (N) 1.0 [1–12] Lacticodeshydrogenase (N) 1.0 [1–4] 19 [5–245] Bilirubin, mol/l Serum amylase (N) 2.0 [1–3] Serum lipase (N) 2.5 [1–6]
Non-biliary acute pancreatitis (n = 81)
p
1.0 [1–5] 1.0 [1–8] 2.2 [1–25] 1.0 [1–4] 1.0 [1–4] 17 [2–396] 1.3 [1–2] 2.5 [1–8]
NS NS NS NS NS NS NS NS
N = N times the upper limit of normal range. NS = Not significant.
Table 4. Diagnostic value of female sex, age and ALT at admission for the diagnosis of the biliary origin of an acute pancreatitis
Female sex Age >50 Age >60 ALT > 2N ALT > 3N
Sensitivity %
Specificity Positive predic- Negative predic% tive value, % tive value, %
60 73 61 74 61
82 65 79 84 91
84 77 82 88 92
N = N times the upper limit of normal range.
Prediction of Biliary Pancreatitis
56 60 52 66 59
formed in 140 patients including all patients (n = 81) with non-biliary and 59 with biliary acute pancreatitis. Among the 107 patients with cholelithiasis without common bile duct stones, the diagnosis was made by ultrasonography or CT scan in 94 (50 by ultrasonography alone, 3 by CT scan alone and 41 using both procedures). Diagnosis was made at EUS alone (negative ultrasonography and CT scan) in 13 patients and confirmed (positive ultrasonography or CT scan) in 32. Among the 25 patients with choledocolithiasis, the diagnosis was made by ultrasonography or CT scan in 18 (ultrasonography alone: 11; CT scan alone: 11; combined: 6). EUS established the presence of CBD stones in 7 patients where ultrasonography and CT scan were negative and confirmed the diagnosis in 7 other patients. In summary, the diagnosis of cholelithiasis or choledocolithiasis was established using EUS alone in 20 patients that is 15% of all the patients with biliary lithiasis. Comparison of General Characteristics and Biochemical Data in Patients with Biliary and Non-Biliary Pancreatitis Univariate Analysis Patients with biliary acute pancreatitis were older than that of those with non biliary pancreatitis. Female patients were more likely to have biliary pancreatitis (table 1). Biochemical data at admission and 48 h later are given in tables 2 and 3, respectively. Liver and pancreatic serum enzyme levels were higher at admission and mean corpuscular red blood cell volume was lower in patients with biliary pancreatitis. Lactate deshydrogenase level was not significantly different between the two groups. There was no significant difference for any biochemical parameters sampled 48 h after admission. Multivariate Analysis Only three parameters were independent variables predicting a biliary origin of an acute pancreatitis: age (p = 0.0001), sex (p = 0.0008), alanine transaminase at admission (p = 0.0004). Table 4 shows the sensitivity and specificity, positive and negative predictive value when using age, sex and alanine transaminase at different thresholds. The probability that an episode of acute pancreatitis was of biliary origin could be calculated using the formula issued from the multivariate analysis: 1/1 + exp(4.6967 – 0.0656 ! age + 1.1208 ! sex – 0.6909 ! alanine transaminase)
with sex = 1 for men, 0 for women, alanine transaminase expressed in X times the upper limit of normal range. For
Pancreatology 2005;5:450–456
453
example, the probabilities that an episode of acute pancreatitis is of biliary origin in a 65-year-old woman with alanine transaminase at 3 times the upper normal value at admission and in a 45-year-old man with 1.5 times the normal alanine transaminase at admission were 87 and 19%, respectively.
Discussion
Many studies have been devoted to the search for reliable criteria in establishing the diagnosis of a biliary origin in cases of acute pancreatitis. The main shortcoming in the previous studies was the lack of reliable cause of pancreatitis. Several studies were retrospective and included only few patients. In addition, the gold standards for the diagnosis of biliary lithiasis were variable: stones in gallbladder, pathological examination, issue of stone(s) during surgical or endoscopic examination of the common bile duct, etc. No unequivocal gold standard for excluding the diagnosis of microlithiasis was available. The diagnosis of biliary lithiasis is easy when stones can be visualized in the gallbladder or the common bile duct at transabdominal ultrasonography or CT scan. When these procedures do not demonstrate stones, accurate tests should confirm or definitively rule out the diagnosis of biliary lithiasis. While cholangio-MRI is also a useful tool in detecting biliary lithiasis, its performance in the detection of small stones (!5 mm) is still a matter of debate: stones smaller than 5 mm are still often missed by cholangio-MRI [19, 26, 27] and as such, we restricted our study to EUS whose accuracy remains superior to that of cholangio-MRI. It has been shown that EUS is essential in this setting since it can diagnose biliary lithiasis in no less than 45% of patients with biliary symptoms (including pancreatitis) in whom ultrasonography performed by senior radiologists failed to show any biliary lithiasis [6]. In another study by Tandon et al. [16], biliary lithiasis was found in 16% patients with so called idiopathic acute pancreatitis allowing a major change in the therapeutic strategy. In patients with pancreatitis, EUS has been shown to be 16–41% more sensitive than transabdominal ultrasonography [14, 15]. EUS is at least as sensitive with less technical failure than endoscopic retrograde cholangiography [14]. In addition, a high degree of skill is not mandatory to obtain satisfactory results in diagnosing common bile duct stones as operators with less than one year experience perform well with respect to this indication. In the present study, the diagnosis of biliary lithiasis was performed solely by EUS in 15% of patients, under-
454
Pancreatology 2005;5:450–456
scoring the necessity to perform this procedure in such cases. Patients with previous cholecystectomy were excluded because they are much less exposed to the risk of biliary pancreatitis. A further confounding factor in the setting of acute pancreatitis stems from the recent discovery in other etiopathogenic causes. In the present study, pancreatitis was considered as idiopathic in only 14/213 patients (7%). This result suggests that uncertainty regarding the cause of the pancreatitis was minimized as far as possible. The percentage of alcoholic pancreatitis (25%) is lower than in previous studies [28–30]. This is probably due to the inclusion of patients with a first bout of acute pancreatitis and the exclusion of patients with recurrent pancreatitis (many of whom are alcoholic). A bias for inclusion of patients with a potential biliary cause of acute pancreatitis may result from referral to centers with experience in EUS, however, this was avoided in the study design which included general hospital recruitment without specialist referred. In addition, the multicentric nature of the study would also appear to reflect the true incidence of biliary causes of acute pancreatitis. Clinical and biochemical parameters aimed at predicting a biliary origin of acute pancreatitis have been in use for some time. In the retrospective study by Blamey et al. [10], age (using a 50-year threshold) was an independent predictive factor. This was confirmed by Neoptolemos et al. [11] who established a 70-year threshold. In another study assessing the risk of choledocholithiasis before laparoscopic cholecystectomy, 55 years of age was also found to be predictive [31]. Female sex is also predictive of a biliary origin in acute pancreatitis [10]; women have a 2fold increased risk [32]. The present study confirms that age and female sex are predictive factors of the biliary origin of an acute pancreatitis. The accuracy of biochemical parameters has rarely been assessed in series with sufficient numbers of patients. Several blood parameters have been evaluated including total bilirubin, serum amylase, alkaline phosphatase, alanine and aspartate transaminase levels. Overall, results are discrepant and depend on the thresholds used and the delay between the onset of pancreatitis and blood sampling. Bilirubin level was found to be predictive of biliary pancreatitis in only one study [11] but more often in cases of a retained common bile duct stone (threshold: 25–30 mol/l) [33–35]. It is well established however that biliary stones rapidly migrate through the ampulla in patients with acute biliary pancreatitis [8, 36] and thus elevated liver blood tests tend to normalize with time. The significance of elevated alkaline phosphatase level is sim-
Lévy et al.
ilar to that of bilirubin [11, 34] and did not predict biliary pancreatitis except in the study by Blamey et al. [10]. Serum amylase has not been found to be a reliable predictive criterion as conflicting data have emerged in the literature [10, 37]. Application of a serum lipase/amylase ratio has been largely debated but was finally not considered a useful biochemical index in the setting of acute biliary pancreatitis [38, 39]. To date, much confusion has also concerned the diagnostic performance of serum aspartate and alanine transaminase levels as useful predictive criteria. Alanine transaminase was more accurate than aspartate transaminase in one meta-analysis [37] and in a recent retrospective study [40]. The threshold was found to be between 1.2 and 3 times the upper normal value [10, 37, 40]. However, neither aspartate nor alanine transaminase were found to be accurate predictive factors of a biliary cause in the studies by Blamey et al. and Neoptolemos et al. [10, 11]. The results of the present study support previous findings by both Grau et al. [40] and Tenner et al. [37]. Elevated alanine transaminase level was the only useful predictive parameter found in the meta-analysis by Tenner et al. [37]; using 3 times the upper normal value as a threshold, the positive predictive value was excellent (95%) but the sensitivity was low [37]. With transaminases at 1.2 times normal threshold, the sensitivity was 73% with a positive predictive value of 92% [40]. Our results confirm these figures in a large prospective study. One of the major problems in all studies assessing the usefulness of biochemical parameters is the delay between the onset of pancreatitis and blood sampling. In the present study, 80% of admission blood samplings were
taken less than 48 h after the beginning of pain. None of the tested parameters was significantly associated with biliary pancreatitis, 2 days later. This result underlines the necessity of rapid sampling in patients with suspicion of acute pancreatitis. Application of these simple strategies should have a large impact in overall management of such patients and should prompt appropriate biliary investigations including EUS in those with negative transabdominal ultrasound and CT scan. The negative predictive values of the three selected parameters are relatively low whatever the threshold. Therefore, a negative result should not be considered as excluding the diagnosis of biliary pancreatitis. In conclusion, in this large prospective study using EUS as the gold standard examination for the diagnosis of biliary lithiasis, age, female sex and alanine transaminase levels (sampled at admission) were predictive of biliary origin of a first episode of acute pancreatitis. These parameters should be useful in indicating non-invasive or invasive diagnostic procedures. However, a definitive exclusion of biliary lithiasis should still rely on radiological procedures as EUS.
Acknowledgments We would like to thank Dr. F. Dyard (Solvay Pharma) who kindly gave us financial support for statistical analysis, F. Sévenier, Marie-Anne Caillaud, N. Le Gall (Fovéa company) for their help in statistical analysis and ‘L’association des hépato-gastroentérologues des hôpitaux généraux’ (ANGH) for helping in the recruitment of centers.
References 1 Boyer J: Faut-il réaliser une CPRE avec sphinctérotomie en urgence? Gastroenterol Clin Biol 2001;25:1S122–1S127. 2 Folsch UR, Nitsche R, Ludtke R, Hilgers RA, Creutzfeldt W: Early ERCP and papillotomy compared with conservative treatment for acute biliary pancreatitis. The German Study Group on Acute Biliary Pancreatitis. N Engl J Med 1997;336:237–242. 3 Neoptolemos JP, Carr-Locke DL, London NJ, Bailey IA, James D, Fossard DP: Controlled trial of urgent endoscopic retrograde cholangiopancreatography and endoscopic sphincterotomy versus conservative treatment for acute pancreatitis due to gallstones. Lancet 1988;ii: 979–983.
Prediction of Biliary Pancreatitis
4 Fan ST, Lai EC, Mok FP, Lo CM, Zheng SS, Wong J: Early treatment of acute biliary pancreatitis by endoscopic papillotomy. N Engl J Med 1993;328:228–232. 5 Gambiez L: Quand et comment traiter la lithiase biliaire? Gastroenterol Clin Biol 2001; 25: 1S128–1S139. 6 Dahan P, Andant C, Levy P, Amouyal P, Amouyal G, Dumont M, Erlinger S, Sauvanet A, Belghiti J, Zins M, Vilgrain V, Bernades P: Prospective evaluation of endoscopic ultrasonography and microscopic examination of duodenal bile in the diagnosis of cholecystolithiasis in 45 patients with normal conventional ultrasonography. Gut 1996;38:277–281.
7 Nowak A, Marek TA, Nowakowska-Dulawa E, Rybicka J, Kaczor R: Biliary pancreatitis needs endoscopic retrograde cholangiopancreatography with endoscopic sphincterotomy for cure. Endoscopy 1998;30:A256–A259. 8 Kelly TR: Gallstone pancreatitis: The timing of surgery. Surgery 1980;88:345–350. 9 Osborne DH, Imrie CW, Carter DC: Biliary surgery in the same admission for gallstone-associated acute pancreatitis. Br J Surg 1981;68: 758–761. 10 Blamey SL, Osborne DH, Gilmour WH, O’Neill J, Carter DC, Imrie CW: The early identification of patients with gallstone associated pancreatitis using clinical and biochemical factors only. Ann Surg 1983;198:574–578.
Pancreatology 2005;5:450–456
455
11 Neoptolemos JP, London N, Bailey I, Shaw D, Carr-Locke DL, Fossard DP, Moossa AR: The role of clinical and biochemical criteria and endoscopic retrograde cholangiopancreatography in the urgent diagnosis of common bile duct stones in acute pancreatitis. Surgery 1986; 100:732–742. 12 Conférence de Consensus: Pancréatite aiguë. Gastroenterol Clin Biol 2001;25:177–192. 13 Novotny I, Lata J: Chronic pancreatitis in endoscopic ultrasonography – analysis of newly diagnosed cases during a 28-month period. Bratisl Lek Listy 2000;101:649–653. 14 Liu CL, Lo CM, Chan JK, Poon RT, Lam CM, Fan ST, Wong J: Detection of choledocholithiasis by EUS in acute pancreatitis: A prospective evaluation in 100 consecutive patients. Gastrointest Endosc 2001;54:325–330. 15 Chak A, Hawes RH, Cooper GS, Hoffman B, Catalano MF, Wong RC, Herbener TE, Sivak MV Jr: Prospective assessment of the utility of EUS in the evaluation of gallstone pancreatitis. Gastrointest Endosc 1999;49:599–604. 16 Tandon M, Topazian M: Endoscopic ultrasound in idiopathic acute pancreatitis. Am J Gastroenterol 2001;96:705–709. 17 Rosch T, Mayr P, Kassem MA: Endoscopic ultrasonography in acute biliary pancreatitis. J Gastrointest Surg 2001;5:223–228. 18 Amouyal P, Amouyal G, Levy P, Tuzet S, Palazzo L, Vilgrain V, Gayet B, Belghiti J, Fekete F, Bernades P: Diagnosis of choledocholithiasis by endoscopic ultrasonography. Gastroenterology 1994;106:1062–1067. 19 Palazzo L, O’Toole D: EUS in common bile duct stones. Gastrointest Endosc 2002; 56: S49–S57. 20 Balthazar EJ, Robinson DL, Megibow AJ, Ranson JH: Acute pancreatitis: value of CT in establishing prognosis. Radiology 1990; 174: 331–336. 21 Levy P, Milan C, Pignon JP, Baetz A, Bernades P: Mortality factors associated with chronic pancreatitis: Unidimensional and multidimensional analysis of a medical-surgical series of 240 patients. Gastroenterology 1989; 96: 1165–1172.
456
22 Lee SP, Nicholls JF, Park HZ: Biliary sludge as a cause of acute pancreatitis. N Engl J Med 1992;326:589–593. 23 Maire F, Bienvenu T, Ngukam A, Hammel P, Ruszniewski P, Levy P: Fréquence des mutations du gène CFTR dans la pancréatite chronique idiopathique. Gastroenterol Clin Biol 2003;27:398–402. 24 Chen JM, Piepoli Bis A, Le Bodic L, Ruszniewski P, Robaszkiewicz M, Deprez PH, Raguenes O, Quere I, Andriulli A, Ferec C: Mutational screening of the cationic trypsinogen gene in a large cohort of subjects with idiopathic chronic pancreatitis. Clin Genet 2001;59:189–193. 25 Ferec C, Raguenes O, Salomon R, Roche C, Bernard JP, Guillot M, Quere I, Faure C, Mercier B, Audrezet MP, Guillausseau PJ, Dupont C, Munnich A, Bignon JD, Le Bodic L: Mutations in the cationic trypsinogen gene and evidence for genetic heterogeneity in hereditary pancreatitis. J Med Genet 1999;36:228–232. 26 de Ledinghen V, Lecesne R, Raymond JM, Gense V, Amouretti M, Drouillard J, Couzigou P, Silvain C: Diagnosis of choledocholithiasis: EUS or magnetic resonance cholangiography? A prospective controlled study. Gastrointest Endosc 1999;49:26–31. 27 Zidi SH, Prat F, Le Guen O, Rondeau Y, Rocher L, Fritsch J, Choury AD, Pelletier G: Use of magnetic resonance cholangiography in the diagnosis of choledocholithiasis: Prospective comparison with a reference imaging method. Gut 1999;44:118–122. 28 Lankisch PG, Assmus C, Maisonneuve P, Lowenfels AB: Epidemiology of pancreatic diseases in Luneburg County: A study in a defined german population. Pancreatology 2002; 2: 469–477. 29 Gullo L, Migliori M, Olah A, Farkas G, Levy P, Arvanitakis C, Lankisch P, Beger H: Acute pancreatitis in five European countries: Etiology and mortality. Pancreas 2002; 24: 223– 227. 30 Gullo L, Migliori M, Pezzilli R, Olah A, Farkas G, Levy P, Arvanitakis C, Lankisch P, Beger H: An update on recurrent acute pancreatitis: data from five European countries. Am J Gastroenterol 2002;97:1959–1962.
Pancreatology 2005;5:450–456
31 Barkun AN, Barkun JS, Fried GM, Ghitulescu G, Steinmetz O, Pham C, Meakins JL, Goresky CA: Useful predictors of bile duct stones in patients undergoing laparoscopic cholecystectomy. McGill Gallstone Treatment Group. Ann Surg 1994;220:32–39. 32 Appelros S, Borgstrom A: Incidence, aetiology and mortality rate of acute pancreatitis over 10 years in a defined urban population in Sweden. Br J Surg 1999;86:465–470. 33 Hauer-Jensen M, Karesen R, Nygaard K, Solheim K, Amlie EJ, Havig O, Rosseland AR: Prospective randomized study of routine intraoperative cholangiography during open cholecystectomy: Long-term follow-up and multivariate analysis of predictors of choledocholithiasis. Surgery 1993;113:318–323. 34 Onken JE, Brazer SR, Eisen GM, Williams DM, Bouras EP, DeLong ER, Long TT 3rd, Pancotto FS, Rhodes DL, Cotton PB: Predicting the presence of choledocholithiasis in patients with symptomatic cholelithiasis. Am J Gastroenterol 1996;91:762–767. 35 Abboud PA, Malet PF, Berlin JA, Staroscik R, Cabana MD, Clarke JR, Shea JA, Schwartz JS, Williams SV: Predictors of common bile duct stones prior to cholecystectomy: A meta-analysis. Gastrointest Endosc 1996;44:450–455. 36 Acosta JM, Ledesma CL: Gallstone migration as a cause of acute pancreatitis. N Engl J Med 1974;290:484–487. 37 Tenner S, Dubner H, Steinberg W: Predicting gallstone pancreatitis with laboratory parameters: A meta-analysis. Am J Gastroenterol 1994;89:1863–1866. 38 Pezzilli R, Billi P, Barakat B, Miglio F: Lipaseamylase ratio does not determine the etiology of acute pancreatitis. Another myth bites the dust. J Clin Gastroenterol 1998;26:34–38. 39 Lankisch PG, Petersen M: Lipase/amylase ratio: Not helpful in the early etiological differentiation of acute pancreatitis. Z Gastroenterol 1994;32:8–11. 40 Grau F, Almela P, Aparisi L, Bautista D, Pascual I, Pena A, Rodrigo JM: Usefulness of alanine and aspartate aminotransferases in the diagnosis of microlithiasis in idiopathic acute pancreatitis. Int J Pancreatol 1999;25:107–111.
Lévy et al.