Recurrent Acute Pancreatitis: Clinical Profile and an Approach to Diagnosis
- First Online:
- Cite this article as:
- Sajith, K.G., Chacko, A. & Dutta, A.K. Dig Dis Sci (2010) 55: 3610. doi:10.1007/s10620-010-1175-8
- 346 Views
Background and Aims
Though recurrent acute pancreatitis is often seen in clinical practice, there are few comprehensive articles on this entity. The aim of this study therefore was to assess the etiological and clinical profile as well as diagnostic yield of non-invasive and invasive tests in this group of patients.
All patients with recurrent acute pancreatitis seen from 2002 to 2007 were included in the study, retrospectively. Clinical information, investigation, and treatment data were collected for all patients by a standardized review of medical charts. Diagnostic tests were grouped into level one (non-invasive) and level two (invasive) tests and their yield was assessed. Comparison was made between the group with known etiology and idiopathic group to look for significant differences.
A total of 188 patients with recurrent acute pancreatitis were seen during the study period. Common etiological factors were biliary disease (37%), pancreas divisum (8.5%) and alcohol (6.4%). Multiple etiologies were seen in 7% of cases, and no cause was found in 32.4%. The diagnostic yield of level-one investigation (non-invasive) was 29.3%. Level-two tests (invasive) identified presumptive etiologies in 38.3% of cases. Complications developed in 12.2% and there was no mortality. Clinical features and complications were similar in the idiopathic group and those with known etiology.
Etiological diagnosis was obtained in 67.6% of patients after comprehensive diagnostic work up. Diagnosis in the majority of patients could only be reached after invasive tests (bile crystal analysis, EUS, ERCP). Early diagnosis and etiology-based therapy is the key to optimum patient outcome.
KeywordsRecurrent acute pancreatitisEtiologyNon-invasive and invasive investigation
Acute pancreatitis is a common condition characterized by inflammation of pancreatic tissue in the absence of morphological changes on imaging studies . About 2–10% of patients succumb to the initial illness  and another 10–35% go on to have recurrent episodes of acute pancreatitis [3–5]. Considering the potential burden of recurrent acute pancreatitis (RAP), it is surprising that there are few comprehensive articles on this entity [3–5]. Most papers deal with an etiological and/or therapeutic subset of RAP [6–10]. Studies on etiology show that a cause for RAP can be determined in 70–90% of patients, the remaining being grouped as idiopathic recurrent acute pancreatitis (IRAP) [5–7, 11, 12]. A paucity of data on the etiological profile of RAP in Asia and the recent introduction of new technology necessitates fresh data to be generated to plan efficient and cost-effective strategies for diagnosis and management of this disease. Previous studies suggest that RAP has a lower mortality than a single episode of pancreatitis [13, 14]. IRAP has been found to have a higher morbidity and mortality than the group where an etiology is detected [5, 13]. We do not know whether this is still true. The purpose of this paper is: (1) to report our experience with RAP and provide a more complete and updated picture of etiological and clinical profile of this disease and (2) assess diagnostic yield of invasive and non-invasive tests and generate a diagnostic algorithm.
All patients seen in the Department of Gastrointestinal Sciences, Christian Medical College and Hospital, Vellore, India, from January 2002 to December 2007 were screened retrospectively. Patients with a diagnosis of recurrent acute pancreatitis were included in the study. Clinical information and laboratory and treatment data were collected for all patients by a standardized review of medical charts using uniform structured data forms.
Recurrent acute pancreatitis (RAP) was defined as two or more documented episodes of abdominal pain, typical of acute pancreatitis, more than 2 months apart and at least one of the following: (1) serum amylase or lipase elevation more than three times the upper limit of normal, (2) features of acute pancreatitis on imaging (ultrasound/CECT) . Patients with features suggestive of chronic pancreatitis including calcifications, ductal dilatation, and parenchymal atrophy were excluded . An attack of pancreatitis was considered to be severe if there was multi-organ dysfunction or local complications, namely, fluid collection, necrosis, or abscess .
Initial etiological evaluation of the patients (level 1) consisted of liver function tests, fasting serum calcium and lipid profile, CA19-9 and noninvasive imaging (ultrasound and/or CECT abdomen). Patients who remained undiagnosed after level 1 investigations were subjected to level 2 evaluations. This consisted of duodenal bile examination for microliths in patients with intact gall bladder and at least one of the sophisticated/invasive imaging techniques (MRCP, ERCP, EUS). The duodenal bile (5–10 ml) was collected during endoscopy, centrifuged at 2,000 × g for 10 min, and then the sediment examined under direct and polarizing microscope. Microliths were diagnosed as a cause of RAP if more than five cholesterol or calcium bilirubinate crystals were present in a patient with intact gall bladder . Genetic tests for RAP were not performed. Patients were labeled as having IRAP if the above investigations were normal. Biliary pancreatitis was considered the cause of RAP if there was jaundice and/or abnormal LFT with cholelithiasis, gall bladder sludge, choledocholithiasis, or biliary microliths . Hypertriglyceridemia was considered the cause when serum triglyceride was more than 500 mg/dl . Hypercalcemia was considered the cause when the fasting serum calcium was elevated (normal range, 8.5–10.5 mg/dl) . Pancreatitis was attributed to alcohol use if the patient consumed an average of 80 g of alcohol daily for more than 5 years or had an alcoholic binge within a week prior to the acute attack . Anomalous union of pancreaticobiliary duct was diagnosed when the junction of the bile duct and the pancreatic duct was at a proximal site with a long (>15 mm) common pancreaticobiliary channel .
The etiology and clinicoepidemiological profile of patients were analyzed. The diagnostic yield of level-one and level-two investigations were also assessed. A comparison was made between the group with known etiology and the idiopathic group to look for significant differences.
The descriptive data are presented as mean values with standard deviations or median with range for continuous variables and as number or proportions for categorical variables. Comparison between the idiopathic group and the rest of the patients was done by Fischer’s exact test for categorical variables and Student’s t test or Mann–Whitney U test for continuous variables. A two-tailed p value of ≤0.05 was considered significant. All analysis was performed in SPSS for Windows Version 11.
Demographic and clinical profile of patients (n = 188)
Age (mean ± SD) (years)
32.6 ± 14
Number of episodes of AP (median, range)
Number of hospital admissions for AP (median, range)
Splanchnic venous thrombosis
Splenic artery pseudoaneurysm
An etiological diagnosis was obtained in 72 of the remaining 133 patients after application of level-two investigations (Fig. 1). Biliary microlithiasis, the commonest etiology was detected in 46 patients followed by structural pancreaticobiliary anomalies in 19 patients (pancreas divisum in 16, anomalous pancreaticobiliary duct union in one, choledochal cyst in one, and duodenal diverticulum in one). Seven patients had multiple etiological factors. Overall, 13 patients (after level-one and level-two investigations) had multiple etiologies. Clinical profile of patients with multiple etiologies (mean age: 32.2 ± 14.5 years; male: 84.6%; east India: 61.5%) and complications were similar to patients with a single etiology.
Comparison of clinical and demographic profiles of the idiopathic group with the rest of the patients
Idiopathic (n = 61)
Others (n = 127)
32.5 + 15.6
32.7 + 12.9
Number of episodes of AP (median, range)
Number of hospital admissions for AP (median, range)
The treatment of patients with RAP was tailored according to the etiology. Patients with a stone in the bile duct (n = 6) underwent stone extraction preoperatively by ERCP, and if unsuccessful (n = 1), were advised intraoperative stone clearance. Cholecystectomy was performed for patients with cholelithiasis. Strict abstinence from alcohol was advised in the group with alcohol-related RAP. Lipid-lowering therapy was started in all patients with dyslipidemia. Deworming with albendazole was done for patients with ascariasis. Forty-one patients with biliary microlithiasis were started on ursodeoxycholic acid (UDCA). Twenty-nine (70.7%) of these patients were asymptomatic during a follow-up period ranging from 6 months to 1 year at which time either a biliary sphincterotomy (n = 25) or cholecystectomy (n = 4) was performed. Among the 25 patients who had biliary sphincterotomy, 23 (92%) were asymptomatic at a median follow-up of 2 years. Three of the patients who had cholecystectomy were followed for a median of 3 years during which period two were symptom free. Of the remaining 12 patients started on UDCA, three developed chronic pancreatitis, two continued to have RAP, one underwent lap cholecystectomy at another hospital and six were lost to follow-up. Eleven patients with pancreas divisum underwent accessory papilla sphincterotomy, of whom eight (73%) were symptom free at a median follow-up of 2.5 years. The remaining five patients with biliary microlithiasis and five patients with PD were not followed-up after the acute episode for etiology-directed therapy. A follow-up of all the patients is currently ongoing, and a definite comment on outcome can be made only after long-term follow-up.
During the past 6 years we have managed 188 patients with RAP at our center. Though a frequently encountered entity in clinical practice, few studies have focused on recurrent acute pancreatitis [3–5]. Gullo et al. [3–5] described RAP in patients from five European countries while Gao et al. and Zhang et al. [3–5] described this entity from China. Comparison of age shows that most of our patients were younger adults (third and fourth decade) when compared to other reports where patients presented 10–20 years later. Prevalence of RAP was higher in males in our study as in other studies [3, 5]. The majority of our patients were seen during the third episode of acute pancreatitis. In contrast, most patients in the European and Chinese studies were seen during the second episode [3, 5]. This may be due to either delayed health care seeking behavior of Indian patients or a delay in referral by primary-care physicians. Delayed presentation in the current study suggests the need to educate the patients and medical fraternity about the nature and morbidity of RAP.
The etiology of recurrent acute pancreatitis has been classified as: (1) toxic-metabolic: alcohol, hypertriglyceridemia, hypercalcemia, drugs; (2) mechanical-obstructive: biliary stones/microlithiasis, structural abnormalities (congenital or acquired), trauma; and (3) miscellaneous . In the current study, 94 (50%) patients had mechanical-obstructive pancreatitis, 20 (10.6%) had toxic-metabolic pancreatitis, 13 (7%) had multiple etiologies and 61 (32.4%) had idiopathic pancreatitis. Similar to most of the earlier studies, biliary pancreatitis (due to cholelithiasis, choledocholithiasis, and microlithiasis) was the most common etiology of RAP (37%) in our study [3, 4, 12, 21–24]. As in other studies, biliary microliths was the predominant cause of biliary pancreatitis [23, 24]. Microlithiasis as a cause of RAP has been debated . Reduction of the risk of RAP after cholecystectomy and UDCA however suggest that microliths can cause RAP . Pancreas divisum (8.5%) was the next most common cause of RAP in the current study. Pancreas divisum (PD) as a cause of acute pancreatitis is controversial, as some studies show that the prevalence of PD in acute pancreatitis is similar to the general population . However, the high prevalence of PD in patients with acute pancreatitis in other studies and the favorable response after accessory papilla sphincterotomy favor PD as a cause of RAP [28–34]. Alcohol was the etiology in a small proportion of our patients (6.4%). This is in contrast to a recent multicenter European study where alcohol was the most common etiology in 57% of the patients . A reason for the reduced frequency of biliary pancreatitis in recent Western studies may be the tendency to perform early cholecystectomy or biliary sphincterotomy after an episode of acute pancreatitis. Other structural causes like anomalous pancreato-biliary union, choledochal cyst, duodenal diverticulum, and metabolic causes like hypertriglyceridemia and hypercalcemia were seen in a small number of patients [4, 8, 19, 20]. Multiple etiologies were detected in 13 (7%) patients. As there are no criteria to determine the dominant etiology responsible for RAP, it was difficult to plan therapy for these patients.
An important observation from our study is that after noninvasive investigations (level one), available at most centers, an etiology for RAP was identified in only 30% of patients (Fig 1). Invasive investigations (ERCP, EUS, bile crystal analysis in patients with intact gall bladder) and/or MRCP (level-two investigations), usually available at tertiary care centers, were needed to establish presumptive etiology in an additional 38% of patients. This suggests that the algorithm to evaluate etiology for RAP should initially be noninvasive tests, which must then be followed by invasive tests if no etiology is detected. The above also suggests that prompt referral to a center where level-two investigations are available is the key to establishing an early etiological diagnosis and planning effective treatment. This will prevent recurrent attacks of pancreatitis and reduce morbidity, mortality, and progression to chronic pancreatitis [5, 35]. Many studies have reported the diagnostic utility of ERCP, bile crystal analysis, endoscopic ultrasound (EUS), and MRCP in patients with RAP [6, 7, 36, 37]. Feller et al.  obtained a diagnosis in 32% of patients labeled ‘idiopathic’ RAP after ERCP. Coyle and colleagues investigated 90 patients with ‘idiopathic’ acute or recurrent acute pancreatitis using ERCP with sphincter of Oddi manometry (SOM), bile analysis, and EUS, and achieved a diagnosis in about 70% patients . In an American study of 126 patients with ‘idiopathic’ recurrent acute pancreatitis, etiology was identified in 79% after ERCP, bile crystal analysis, and SOM . As ERCP is associated with serious complications, including pancreatitis, bleeding, and perforation, its use as a purely diagnostic modality is on the decline . EUS with its potential to detect small CBD stones, biliary sludge, congenital structural anomalies, and early chronic pancreatitis, is emerging as an important less-invasive alternative to ERCP [39, 40]. MRCP, a non-invasive modality has been shown to be effective in the diagnosis of CBD stones and congenital structural abnormalities .
Sixty-one patients (32%) were labeled as having idiopathic recurrent acute pancreatitis (IRAP) as no etiology was detected after level-one and level-two investigations (Fig. 1). The percentage of patients with IRAP in our study is similar to studies from China, but higher than a European study where 10.4% of patients had IRAP [3, 5]. Studies on patients with IRAP have shown that SOD is responsible for RAP in 15–30% of patients [11, 20, 42]. Genetic studies suggest that CFTR and cationic trypsinogen mutations are associated with RAP [43, 44]. Thomas et al.  have shown that 33% of patients labeled as IRAP on long-term follow up developed chronic pancreatitis. What could the etiology be in our patients with IRAP? It may be due to SOD or genetic factors not evaluated in this study, or early chronic pancreatitis not detected by imaging studies.
Twenty-seven patients (14.4%) in our study had severe pancreatitis. Complications were detected in 23 (12.2%) patients. Pancreatic pseudocyst (8%) was the most common complication followed by pleural effusion (7%), renal failure (3%), and splanchnic venous thrombosis (2%). There was no mortality. These data are similar to the study by Zhang et al.  where 17% of patients with RAP had severe pancreatitis. Similar to previous studies, our study also suggests that RAP has a lower mortality that a single episode of pancreatitis [3, 14]. Table 2 compares the clinical profile of patients with IRAP and other etiologies and shows that the rate of severe pancreatitis and complications is similar in both groups. This is in contrast to data from other studies that show IRAP to be a more severe disease . A possible explanation for this is that these studies had a large number of patients with alcoholic RAP where severity of disease and mortality are lower than biliary or idiopathic RAP . The potential limitation of our study is its retrospective nature. However, a standardized approach to evaluate and treat patients with pancreatitis followed at our center minimizes the chance of missing data.
In conclusion, this study demonstrates that biliary stone disease, pancreas divisum, and alcoholism are common causes of RAP. Diagnosis could only be reached in the majority of patients after use of invasive tests (ERCP, EUS, and bile crystal analysis) available at tertiary care centers. Early diagnosis and etiology-based therapy are the keys to prevent recurrent attacks and to obtain an optimum patient outcome.
Conflict of Interest Statement
The authors declare that there are no conflicts of interests. No grants were received for this study.