Inflammatory bowel disease (IBD) is a chronic inflammatory condition of the gastrointestinal (GI) tract that comprises Crohn disease (CD) and ulcerative colitis (UC). The number of affected patients has been growing in recent decades both in adults and children creating an increasing burden on GI units.1 Inflammatory bowel disease has a multifactorial etiology where genetic predisposition, environmental factors, and altered intestinal microflora could lead to the intestinal immune abnormalities that fuel the mucosal inflammation.2 Although the main symptoms are connected to the GI tract, IBD is a systemic disease that often presents with associated conditions or extraintestinal manifestations (EIM): dermatologic, musculoskeletal, oral, ocular, cardiovascular, neurologic, hepatobiliary, or pancreatic lesions.3,4
The most common pancreatic pathologies associated with IBD are acute pancreatitis (AP) and asymptomatic pancreas enzyme level elevations (for a recent review, see Iida et al5). Acute pancreatitis is a sterile inflammatory condition of the pancreatic tissue characterized by the activation of pancreatic enzymes inside the pancreas for various reasons; the most common etiologies in the general population are biliary obstruction and excessive ethanol consumption.6,7 Although these non–IBD-specific etiologies can be observed in IBD patients as well, several publications suggest that IBD is associated with an elevated risk for AP.8–10 Based on case reports and cohort studies, 2 disease-specific forms of AP can be seen in IBD: one is most likely related to the pathogenesis of IBD and, therefore, can be considered as EIM of IBD, whereas the other form is a consequence of the management of IBD or its associated diseases (eg, biliary stones).4 However, the distribution of these etiologies in IBD associated AP is yet to be explored. Likewise, to our knowledge, no comprehensive synthesis of the large-scale studies reporting the odds or the annual incidence of AP in IBD is available. Therefore, we aimed to explore, analyze, and systematically assess the current literature to provide evidence-based data on the association of AP and IBD.
MATERIALS AND METHODS
This study was reported according to the Preferred Reporting Items for Systematic review and Meta-Analysis (PRISMA) guidelines.11 The review protocol was registered on the PROSPERO International Prospective Register of Systematic Reviews (CRD42017080464).
A systematic search was conducted in 4 major electronic databases (MEDLINE via PubMed, Embase, Web of Science, Scopus) up to June 19, 2019 (date of the last search), without any search restrictions. The search strategy used comprehensive strings of words with variations of the terms “pancreatitis” in combination with term variations for “inflammatory bowel diseases,” “Crohn's disease,” or “ulcerative colitis.” The exact search query used for the search can be found in Supplementary Table 1, http://links.lww.com/MPA/A823. The reference lists of selected articles were also checked.
Study Selection and Eligibility Criteria
Selection of the studies and screening was conducted by 2 investigators (B.T. and B.S.) independently. The screening was performed through a review of the titles and abstracts of the records. Studies that met the inclusion criteria and those with abstracts that lacked crucial information for the decision regarding their exclusion were retrieved for full-text evaluations. Decisions of eligibility and disagreements were resolved through discussion or by a third reviewer (G.V.). To be included in this review, studies had to meet all of the following criteria: (1) completed and published observational studies with human subjects, (2) the use of objective definitions of IBD and AP, (3) the provision of quantitative reports of IBD and AP, and (4) all participants had been counted only once in the published report. To appraise the odds of AP in IBD the following PECO was used: (P) general population, (E) IBD patients, (C) non-IBD population, and (O) odds of AP. Studies that did not report the incidence of AP either in the IBD or the non-IBD population or the reported data were redundant (ie, a participant could be counted more than once in the final report) were excluded. To examine the annual incidence of AP in IBD, we included studies sampling from the IBD population (P) and reporting incidence rates of AP (O) over an observation period given in person-years (PYs). Studies were not included if they reported only mean follow-up data or reported redundant data. Conference abstracts were excluded from both analyses. When studies reported the same population (database) and period, the most recent study was selected for inclusion.
The same authors conducted the data extraction independently, and disagreements were resolved by consensus. Name of the first author, date of publication, geographical location, study type, study period, age-range of included individuals, the subtype(s) of IBD, number of IBD patients, crude incidence of AP cases (for IBD and non-IBD population, respectively), adjusted relative measures, and the observation period of the study in PYs were extracted using a data extraction table, if applicable (Table 1.).
TABLE 1 -
Basic Characteristics of the Eight Included Studies
||Basis of Diagnosis
||Age Range, y
||Observation Period (PYs)
||No. IBD Patients
||No. Controls (Non-IBD)
|Chen et al, 20169
||All IBD patients
|Kim et al, 201712
||UC patients only
|McAuliffe et al, 201513
||All IBD patients
|Munk et al, 200410
||All IBD patients
ICD-8 and ICD-10 codes
|Rasmussen et al, 199914
||All IBD patients
|Sundström et al, 200615
||All IBD patients
|Thisted et al, 200616
||All IBD patients
ICD-8 and ICD-10 codes
|Yang et al, 201817
||All IBD patients
N/A indicates no data published.
Data Synthesis and Analysis
Odds ratios (ORs) were calculated from the crude incidences and pooled ORs or event rates with 95% confidence intervals (CIs) were calculated. We applied the random-effects model with the DerSimonian-Laird estimation.18 Cochrane Q, I2, and χ2 tests were used to quantify statistical heterogeneity and gain probability values, respectively. Based on Cochrane handbook, I2 = 100% × (Q − df)/Q and represents the magnitude of the heterogeneity (I2 = 30–60% – moderate; 50–90% – substantial; 75–100% – considerable heterogeneity), and P < 0.1 indicated significant heterogeneity.19 All statistical analyses were performed using STATA 16.0 (Stata Corporation, College Station, Tex).
Risk of Bias and Quality of Evidence
Following the Cochrane Prognosis Methods Group recommendation,20,21 the quality assessment of prognostic studies was made using the Quality in Prognosis Studies (QUIPS) tool. First, 6 important domains were critically appraised to evaluate validity and bias in the studies: (1) study participation, (2) study attrition, (3) prognostic factor measurement, (4) outcome measurement, (5) study confounding, and (6) statistical analysis and reporting.20 Each domain contained between 3 and 7 prompting items to be rated on a 4- (yes/partial/no/unsure) or 2-grade scale (yes/no). In a final stage, the overall judgment of the risk of bias (RoB) within each domain was made based on the rated items; all of the responses to the prompting items were taken together when judging a domain's overall RoB, which was expressed on a 3-grade scale (high, moderate, or low RoB). Hence, the QUIPS assessment results in 6 ratings of RoB, 1 for each domain. The final RoB of each study was decided by the number of domains with high and low RoB: studies were considered to have low overall RoB if none of the 6 domains had high and most of the domains had low RoB; high overall RoB was judged when 2 or more domains had high RoB, or less than half of the domains had low RoB; otherwise, the overall RoB was judged to be moderate. To examine small study effects, we used the visual assessment of a funnel plot because tests for funnel plot asymmetry are not advised in analyses with fewer than 10 studies.22
Ethical approval was not required as data is not individualized, and primary data were not collected.
Search and Study Selection
The search of 4 electronic databases resulted in 9178 records (Embase 3540, Scopus 3132, Web of Science 1198, and PubMed 1308), of which 3627 nonduplicate articles were screened by title and abstract. One additional article was found eligible based on reference lists of the studies screened by full text.16 The overview of screening and study selection is shown in Figure 1. The articles selected for full-text screening were handled together for the 2 analyses (143 articles were screened). From a total of 8 articles, 69,10,14–17 and 49,12–14 studies were included in the 2 final analysis, respectively. The basic characteristics and main findings of the 8 articles included in our study are shown in Tables 1 and 2.
TABLE 2 -
Effect Estimates of the Eight Included Studies
||Annual Incidence Per 100,000 PYs
||Adjusted Measures Reported
||Measures Adjusted for
|Chen et al, 20169
||Hazard ratio: CD: 3.4 (95% CI, 2.7–3.26) UC: 2.49 (95% CI, 1.91–3.26)
||Age, sex, alcohol-related disease, biliary stone, hypertension, hyperlipidemia, diabetes mellitus, obesity, hepatitis B and C, COPD, hypertriglyceridemia, cardiovascular disease, chronic kidney disease, hypercalcemia
|Kim et al, 201712
|McAuliffe et al, 201513
|Munk et al, 200410
|Rasmussen et al, 199914
||Standardized incidence ratio: CD: 4.3 (95% CI, 2.9–6.2) UC: 2.1 (1.6–2.8)
|Sundström et al, 200615
||OR: IBD: 4.7 (95% CI, 2.2–10)
|Thisted et al, 200616
|Yang et al, 201817
||Standardized prevalence ratio: CD: 4.94 (95% CI, 4.47–5.40) UC: 2.48 (2.28–2-68)
*Crude ORs, calculated based on published data.
COPD indicates chronic obstructive pulmonary disease; N/A, no data published.
Analysis of the Odds of AP in IBD
We first analyzed the odds of AP in IBD. Of the 6 eligible studies, 1 was cross-sectional, 2 were prospective cohort, and 3 case-control studies; the (International Coding of Diseases [ICD]-based) definition of cases was defined to be AP in all case-control studies,10,15,16 and the control groups had been selected accordingly. The 6 articles contained data of 1,309,278 people from Denmark, Sweden, South Korea, and Taiwan. The pooled odds of AP in IBD was 3 times higher (OR, 3.11; 95% CI, 2.93–3.30; I2 = 0.0%; Fig. 2) compared with the non-IBD population.
Of the 6 studies, 3 reported incidences of AP events broken down to CD and UC subpopulations and, therefore, were eligible for subgroup analysis. The analysis has found that the odds of AP in CD patients was significantly higher than that in UC patients (OR, 4.12; 95% CI, 3.75–4.54; I2 = 0.0% vs OR, 2.61; 95% CI, 2.40–2.83; I2 = 0.0%; P < 0.0001; Fig. 3).
Analysis of the Annual Incidence of AP in IBD
Four studies reported the incidence rates of AP among IBD patients with the time of observation period given in PYs: 2 prospective and 2 retrospective studies. Three studies observed all IBD patients, including both CD and UC patients, whereas 1 study followed up only UC patients. The 4 studies covered a sum of 268,859 PYs observation time. The pooled incidence rate of AP in IBD was 0.21% (95% CI, 0.084%–0.392%), for example, 210/100,000 PYs (95% CI, 84–392/100,000 PYs). The forest plot of the analysis is shown in Figure 4.
Heterogeneity and Quality Assessment of Data
The assessment of the odds of AP in IBD proved to be homogeneous (I2 = 0.0%; P = 0.848), whereas high heterogeneity was detected (I2 = 98.66%; P < 0.001) in the analysis of the annual incidence of AP in IBD. In this latter case, because of the low number of studies, the source of heterogeneity could not be investigated by any further subgroup analysis.
Risk of bias of the included articles in the 2 analyses was assessed by 6 domains, respectively using the QUIPS tool (Supplementary Table 2, http://links.lww.com/MPA/A823).
In the analysis of ORs, the overall quality of included studies was high: RoB was low in 5 and moderate in one of the 6 articles. Study participation, outcome (AP) measurement, and statistical analysis and reporting domains were judged to have low RoB in all included studies. The study attrition domain was assessed only in prospectively recruiting studies9,14,15 and was found to have moderate RoB in all 3 studies. There was 1 study with moderate RoB in the measurement of prognostic factor (IBD) domain and 2 in the Study confounding domains, respectively; all other studies had low RoB in these domains (Supplementary Table 2A, http://links.lww.com/MPA/A823). Visual assessment of the funnel plot suggested no serious small study effects.
In the analysis of the annual incidence, the overall quality of the included articles is moderate: RoB was low in one and moderate in 3 of the 4 studies. The measurement of prognostic factor (IBD) and statistical analysis domains received low RoB judgment for all studies. The study participation and the study confounding domains were judged to be low once and moderate thrice. The RoB of the study attrition domain was assessed only in the prospective studies and was found to be moderate in both cases. The outcome (AP) measurement domain was of moderate RoB in one study and low in the other 3 studies (Supplementary Table 2B, http://links.lww.com/MPA/A823). Because of the type of outcome measure (event rates), the presence of publication bias could not be ruled out.
Acute pancreatitis is a potentially serious inflammatory disorder with a possibly high mortality rate.23,24 The leading symptom of AP is acute abdominal pain. Beside others, AP can occur in association with IBD, where the clinical symptoms of the 2 condition may be difficult to differentiate. In IBD, acute abdominal pain may occur because of a severe relapse of the disease, subileus (as a consequence of strictures), or an intra-abdominal abscess. The overlapping symptoms of AP and these complications may delay the appropriate diagnostic workup such as the measurement of serum lipase and amylase levels. Since the early management of AP is crucially important,25,26 increased surveillance leading to an earlier diagnosis may save patients’ lives.
In the last 50 years, an emerging number of case reports and clinical studies suggested that AP is more frequent in IBD; however, the number of population-wide studies and therefore, firm evidence addressing this association is limited. To our knowledge, this current meta-analysis is the first to investigate the association between AP and IBD. We aimed to summarize the currently available findings of large-scale studies. We have concluded that the pooled odds for AP in IBD is 3 times higher (OR, 3.11) than in the non-IBD population (Fig. 2), the odds are higher in CD than in UC (Fig. 3), and the pooled annual incidence is 210/100,000 PYs (Fig. 4). The number of studies and the reported data has proved to be insufficient for a more detailed stratification, including analysis of etiological distribution.
Ball et al27 suggested at first an association between pancreatic involvement and IBD based on autopsy studies of UC patients. In the current guidelines and reviews, AP is the most frequently mentioned pancreatic lesion in IBD4,28–30; however, no clear recommendations for the optimal clinical diagnostic workup are stated. This may be explained by the variety of contributing etiologies and the lack of firm evidence. Because the current hypotheses on the possible background mechanisms causing this association between AP and IBD are strongly connected to the different etiologies of the pancreatic inflammation, and, therefore, are hard to discuss independently, we also wanted to synthesize the available evidence. However, we were unable to find suitable data for the quantitative synthesis of the different etiologies of AP in IBD. Therefore, we can only give a summary of the possible etiologies with some reference to the possible mechanisms of the connection between AP and IBD.
In IBD, the etiology of AP can be divided into 3 groups.4 Some AP cases share common pathogenetic pathways with IBD and, therefore, could be considered as EIMs of IBD: these are the cases of autoimmune, granulomatous, idiopathic, and primary sclerotizing cholangitis associated pancreatitis (provided that they occur in association with IBD). Another major group of AP cases associated with IBD is related to the medical treatment of the disease: drug-induced (especially thiopurine-induced), post-endoscopic retrograde cholangiopancreatography (ERCP), postenteroscopy pancreatitis, and AP secondary to duodenal CD are in this group.4 Besides the previously mentioned causes, the group of “classical” etiologies (eg, biliary obstruction, ethanol abuse) also occurs; however, there is major controversy on the proportion of these 3 groups of etiologies in the published studies.
One of the possible candidates that are presumed to be causing the elevated number of AP cases in IBD is biliary obstruction.5 In the general population, cholelithiasis can be responsible for up to 40% of the cases31 and several studies show an elevated risk for cholelithiasis in IBD.32–34 However, studies focusing on AP in IBD were so far unable to undoubtedly confirm biliary pancreatitis as a major factor behind the elevated chance of AP in IBD. Moolsintong et al35 retrospectively analyzed the clinical features and outcomes of 48 CD patients with AP in the United States between 1976 and 2001 and found that biliary and idiopathic AP were equally frequent, each corresponding for 21% of cases, followed by alcohol abuse (15%), duodenal CD (15%), thiopurine-induced AP (13%), and post-ERCP pancreatitis (10%). Later, in a large Spanish cohort, Bermejo et al8 found that most of the AP cases were attributed to drug exposure (64%), whereas 20% were idiopathic, 12% biliary, and 4% of miscellaneous etiology (duodenal CD, post-ERCP, hypertriglyceridemia, etc). Another team analyzed all the hospital admissions during 2005 and 2011 with a primary diagnosis of AP and a co-diagnosis of IBD in the United States, but they focused only on alcohol and drug abuse, and biliary etiologies.36 They have shown that alcohol abuse was also recorded in approximately 12% of cases (11.6% in CD and 12.1% in UC) and 21% of controls. The rate of cases attributable to medications was found to be significantly higher in CD (6.8% vs 4.9%) but not in UC, and they observed the diagnosis of biliary obstruction less frequently in IBD compared with controls (2.4% vs 4.4%), both in CD and UC. The latest study addressing the etiology of AP in UC from a single referral center in South Korea has shown that 45% of patients had drug induced (of which 55% thiopurine), 25% had autoimmune, 18% idiopathic, and 12% gall stone–induced pancreatitis, but they have not evaluated alcohol abuse.12
In summary, it seems that “classical” risk factors including biliary obstruction, similarly to the general population, have an important role in IBD, but the proportion of the reported etiologies in different studies vary highly. Another important candidate behind the elevated chance of AP in IBD is medical treatment-associated pancreatitis. Accordingly, most of the studies uniformly document a significant proportion (up to 65%) of drug-induced pancreatitis, especially thiopurine-induced cases.
Although thiopurine-induced pancreatitis (TIP) is usually considered as a mild complication of IBD because the clinical course of the most cases is mild,37 patient with TIP might require the discontinuation of the ongoing thiopurine therapy and a choice of a more complex treatment.38 Thiopurine-induced pancreatitis usually occurs within the first 30 days after the start of treatment, in 1.5% to 7% of treated patients.8,12,37,39,40 Bermejo et al8 reported that 56% of AP cases in their IBD cohort were related to thiopurine exposure and most cases was seen in CD (65% vs 22% of all AP cases in CD and UC, respectively). In a recent study, pediatric IBD patients during azathioprine treatment had an almost 6 times higher chance for AP compared with no treatment with azathioprine (ratio of incidence rates, 5.82; 95% CI, 2.46–13.72); incidence rate of AP was 49.1 event per 1000 PYs during treatment.39 Another interesting finding is that azathioprine-induced pancreatitis was found more frequently in CD (4.9%) compared with other conditions treated with azathioprine (autoimmune hepatitis (1.5%), renal transplantation (0.5%), and liver transplantation (0.4%) and 0% in UC, rheumatoid arthritis, systemic lupus berythematosus, or Wegener granulomatosis, P < 0.05).41 Although the sex distribution of all AP cases in IBD was found shifted toward males in population-based studies,9,14 a slight female dominance was observed in several reports of TIP in IBD.42–46
In our analysis, we originally aimed to determine the proportion of the certain etiologies of AP associated with IBD; however, we did not reach this goal because of the low number of eligible studies and lack of published data on etiologies in the eligible ones. Furthermore, although we were able to confirm the association between AP and IBD, the causative relationship of the 2 could not be stated. Until we gather firm evidence either on the distribution of etiologies or on the causative connection between IBD and AP, this remains only an interesting association clinicians should be aware of. Therefore, to provide firm evidence, further large-scale studies dedicated to exploring the etiological distribution and the clinical course of AP in IBD are highly needed.
There are several strengths and limitations of this study, and therefore, the results of this meta-analysis should be treated with caution. The main strength of our study is the robust size of the pooled population and the use of the PRISMA guidelines. Another strength is that in our analysis, no heterogeneity was observed when pooling the ORs of AP in IBD (I2 = 0%).
There are also limitations to our study. One is the low number of eligible articles included, which consequently led to a limited number of analyses. Another major limitation is that the ORs reported in this meta-analysis are calculated based on the reported number of events in the investigated publications, and therefore, are crude ORs. Adjusted ORs for major risk factors of AP could not be calculated because of the lack of information. Pooled analysis of adjusted outcome measures was also not available because they were only reported for IBD in general in 1 study and both for CD and UC in 3 studies, and all studies used different outcome measures.
Because IBD is a heterogeneous disease, treating it as one homogeneous patient group, as our main analysis did, could lead to oversimplified and indirect results. However, we were able to conduct a subgroup analysis of 4 studies based on the IBD subtypes and, therefore, provide a clinically more relevant outcome. Unfortunately, none of the eligible studies addressed the occurrence of AP based on anatomic location or clinical behavior of IBD, which would allow further subgroup analyses and help describe the topic more precisely. Similarly, no comparison analysis could be made based on the etiology or behavior of AP, which also simplifies an otherwise complex outcome. In the analysis of annual incidence significant heterogeneity was seen (I2 = 98.67%), however, in prognostic studies with large sample sizes, a high heterogeneity could be observed frequently.21 Another reason for the high heterogeneity observed in this analysis might be the different geographical and temporal distribution of the studies: the studies originated from Taiwan,9 South Korea,12 the United States,13 and Denmark14; the Danish study was conducted in the 1990s, whereas the 3 others in the late 2010s.
In conclusion, in this first meta-analysis on the topic, we showed that the odds of AP in IBD is 3 times higher than that in the non-IBD population. Current clinical guidelines only mention AP as a possible EIM of IBD4 but lack the recommendations for an optimal diagnostic workup to rule out AP in patients with abdominal complaints. Based on this meta-analysis, approximately 21 AP case per 10,000 IBD patient can be anticipated annually, which emphasize the importance of pancreatic enzyme measurements and pancreatic imaging examinations in symptomatic IBD patients. Because CD patients have significantly higher odds of AP than patients with UC, more thorough surveillance of pancreatic involvement in CD is advisable. Unfortunately, the etiology and clinical course of AP in IBD remain a topic, where further large-scale studies are needed.
The authors offer this study to the memory of Prof. Gábor Veres.
1. Molodecky NA, Soon IS, Rabi DM, et al. Increasing incidence and prevalence of the inflammatory bowel diseases with time, based on systematic review. Gastroenterology
. 2012;142:46–54.e42; quiz e30.
2. Shouval DS, Rufo PA. The role of environmental factors in the pathogenesis of inflammatory bowel diseases: a review. JAMA Pediatr
3. Greuter T, Vavricka SR. Extraintestinal manifestations in inflammatory bowel disease - epidemiology, genetics, and pathogenesis. Expert Rev Gastroenterol Hepatol
4. Harbord M, Annese V, Vavricka SR, et al. The first european evidence-based consensus on extra-intestinal manifestations in inflammatory bowel disease. J Crohns Colitis
5. Iida T, Wagatsuma K, Hirayama D, et al. The etiology of pancreatic manifestations in patients with inflammatory bowel disease. J Clin Med
6. Gullo L, Migliori M, Oláh A, et al. Acute pancreatitis in five European countries: etiology and mortality. Pancreas
7. DiMagno MJ, DiMagno EP. New advances in acute pancreatitis. Curr Opin Gastroenterol
8. Bermejo F, Lopez-Sanroman A, Taxonera C, et al. Acute pancreatitis in inflammatory bowel disease, with special reference to azathioprine-induced pancreatitis. Aliment Pharmacol Ther
9. Chen YT, Su JS, Tseng CW, et al. Inflammatory bowel disease on the risk of acute pancreatitis: a population-based cohort study. J Gastroenterol Hepatol
10. Munk EM, Pedersen L, Floyd A, et al. Inflammatory bowel diseases, 5-aminosalicylic acid and sulfasalazine treatment and risk of acute pancreatitis: a population-based case-control study. Am J Gastroenterol
11. Moher D, Liberati A, Tetzlaff J, et al. Preferred reporting items for systematic reviews and meta-analyses: the PRISMA statement. BMJ
12. Kim JW, Hwang SW, Park SH, et al. Clinical course of ulcerative colitis patients who develop acute pancreatitis. World J Gastroenterol
13. McAuliffe ME, Lanes S, Leach T, et al. Occurrence of adverse events among patients with inflammatory bowel disease in the HealthCore Integrated Research Database. Curr Med Res Opin
14. Rasmussen HH, Fonager K, Sørensen HT, et al. Risk of acute pancreatitis in patients with chronic inflammatory bowel disease. A Danish 16-year nationwide follow-up study. Scand J Gastroenterol
15. Sundström A, Blomgren K, Alfredsson L, et al. Acid-suppressing drugs and gastroesophageal reflux disease as risk factors for acute pancreatitis—results from a Swedish case-control study. Pharmacoepidemiol Drug Saf
16. Thisted H, Jacobsen J, Munk EM, et al. Statins and the risk of acute pancreatitis: a population-based case-control study. Aliment Pharmacol Ther
17. Yang BR, Choi NK, Kim MS, et al. Prevalence of extraintestinal manifestations in Korean inflammatory bowel disease patients. PLoS One
18. DerSimonian R, Laird N. Meta-analysis in clinical trials. Control Clin Trials
19. Deeks JJ, Higgins JPT, Altman DG. Chapter 10: analysing data and undertaking meta-analyses. In: Higgins JPT, Thomas J, Chandler J, et al., eds. Cochrane Handbook for Systematic Reviews of Interventions Version 6.0 (Updated July 2019)
. Available at: www.training.cochrane.org/handbook
. Accessed January 15, 2020.
20. Hayden JA, van der Windt DA, Cartwright JL, et al. Assessing bias in studies of prognostic factors. Ann Intern Med
21. Iorio A, Spencer FA, Falavigna M, et al. Use of GRADE for assessment of evidence about prognosis: rating confidence in estimates of event rates in broad categories of patients. BMJ
22. Sterne JA, Sutton AJ, Ioannidis JP, et al. Recommendations for examining and interpreting funnel plot asymmetry in meta-analyses of randomised controlled trials. BMJ
23. Párniczky A, Kui B, Szentesi A, et al. Prospective, multicentre, nationwide clinical data from 600 cases of acute pancreatitis. PLoS One
24. Mikó A, Vigh É, Mátrai P, et al. Computed tomography severity index vs. other indices in the prediction of severity and mortality in acute pancreatitis: a predictive accuracy meta-analysis. Front Physiol
25. Hritz I, Czakó L, Dubravcsik Z, et al. [Acute pancreatitis. Evidence-based practice guidelines, prepared by the Hungarian Pancreatic Study Group]. [Article in Hungarian]. Orv Hetil
26. Hines OJ, Pandol SJ. Management of severe acute pancreatitis. BMJ
27. Ball WP, Baggenstoss AH, Bargen JA. Pancreatic lesions associated with chronic ulcerative colitis. Arch Pathol (Chic)
28. Ramos LR, Sachar DB, DiMaio CJ, et al. Inflammatory bowel disease and pancreatitis: a review. J Crohns Colitis
29. Bregenzer N, Hartmann A, Strauch U, et al. Increased insulin resistance and beta cell activity in patients with Crohn's disease. Inflamm Bowel Dis
30. Seyrig JA, Jian R, Modigliani R, et al. Idiopathic pancreatitis associated with inflammatory bowel disease. Dig Dis Sci
31. Zilio MB, Eyff TF, Azeredo-Da-Silva ALF, et al. A systematic review and meta-analysis of the aetiology of acute pancreatitis. HPB (Oxford)
32. Parente F, Pastore L, Bargiggia S, et al. Incidence and risk factors for gallstones in patients with inflammatory bowel disease: a large case-control study. Hepatology
33. Fagagnini S, Heinrich H, Rossel JB, et al. Risk factors for gallstones and kidney stones in a cohort of patients with inflammatory bowel diseases. PLoS One
34. Chen CH, Lin CL, Kao CH. Association between inflammatory bowel disease and cholelithiasis: a nationwide population-based cohort study. Int J Environ Res Public Health
35. Moolsintong P, Loftus EV Jr, Chari ST, et al. Acute pancreatitis in patients with Crohn's disease: clinical features and outcomes. Inflamm Bowel Dis
36. Alexoff A, Roginsky G, Zhou Y, et al. Inpatient costs for patients with inflammatory bowel disease and acute pancreatitis. Inflamm Bowel Dis
37. Ledder O, Lemberg DA, Day AS. Thiopurine-induced pancreatitis in inflammatory bowel diseases. Expert Rev Gastroenterol Hepatol
38. Lee LYW, Gardezi AS, Santharam V, et al. Effect of azathioprine intolerance on outcomes of inflammatory bowel disease: a cross-sectional study. Frontline Gastroenterol
39. Wintzell V, Svanström H, Olén O, et al. Association between use of azathioprine and risk of acute pancreatitis in children with inflammatory bowel disease: a Swedish-Danish nationwide cohort study. Lancet Child Adolesc Health
40. Teich N, Mohl W, Bokemeyer B, et al. Azathioprine-induced acute pancreatitis in patients with inflammatory bowel diseases—a prospective study on incidence and severity. J Crohns Colitis
41. Weersma RK, Peters FT, Oostenbrug LE, et al. Increased incidence of azathioprine-induced pancreatitis in Crohn's disease compared with other diseases. Aliment Pharmacol Ther
42. López-Martín C, Chaparro M, Espinosa L, et al. Adverse events of thiopurine immunomodulators in patients with inflammatory bowel disease. Gastroenterol Hepatol
43. Moran GW, Dubeau MF, Kaplan GG, et al. Clinical predictors of thiopurine-related adverse events in Crohn's disease. World J Gastroenterol
44. Schwab M, Schäffeler E, Marx C, et al. Azathioprine therapy and adverse drug reactions in patients with inflammatory bowel disease: impact of thiopurine S-methyltransferase polymorphism. Pharmacogenetics
45. van Geenen EJ, de Boer NK, Stassen P, et al. Azathioprine or mercaptopurine-induced acute pancreatitis is not a disease-specific phenomenon. Aliment Pharmacol Ther
46. Zelinkova Z, Bultman E, Vogelaar L, et al. Sex-dimorphic adverse drug reactions to immune suppressive agents in inflammatory bowel disease. World J Gastroenterol