- Email: [email protected]
Risk of Recurrent Pancreatitis and Progression to Chronic Pancreatitis After a First Episode of Acute Pancreatitis
Clinical Gastroenterology and Hepatology 2016;14:738–746
PANCREAS, BILIARY TRACT, AND LIVER Risk of Recurrent Pancreatitis and Progression to Chronic Pancreatitis After a First Episode of Acute Pancreatitis Usama Ahmed Ali,*,‡ Yama Issa,* Julia C. Hagenaars,‡ Olaf J. Bakker,‡ Harry van Goor,§ Vincent B. Nieuwenhuijs,k Thomas L. Bollen,¶ Bert van Ramshorst,# Ben J. Witteman,** Menno A. Brink,‡‡ Alexander F. Schaapherder,§§ Cornelis H. Dejong,kk B. W. Marcel Spanier,¶¶ Joos Heisterkamp,## Erwin van der Harst,*** Casper H. van Eijck,‡‡‡ Marc G. Besselink,* Hein G. Gooszen,§§§ Hjalmar C. van Santvoort,* and Marja A. Boermeester,* for the Dutch Pancreatitis Study Group *Department of Surgery, Academic Medical Center, Amsterdam, The Netherlands; ‡Department of Surgery, University Medical Center Utrecht, Utrecht, The Netherlands; §Department of Surgery, §§§Department of Evidence Based Surgery, Radboud University Medical Center, Nijmegen, The Netherlands; kDepartment of Surgery, Isala Clinics, Zwolle, The Netherlands; ¶ Department of Radiology, #Department of Surgery, St Antonius Hospital, Nieuwegein, The Netherlands; **Department of Gastroenterology, Gelderse Vallei Hospital, Ede, The Netherlands; ‡‡Department of Gastroenterology, Meander Medical Center, Amersfoort, The Netherlands; §§Department of Surgery, Leiden University Medical Center, Leiden, The Netherlands; kk Department of Surgery and NUTRIM, University Hospital Maastricht, Maastricht, The Netherlands; ¶¶Department of Gastroenterology and Hepatology, Rijnstate Hospital, Arnhem, The Netherlands; ##Department of Surgery, St. Elisabeth Hospital, Tilburg, The Netherlands; ***Department of Surgery, Maasstad Ziekenhuis, Rotterdam, The Netherlands; ‡‡‡ Department of Surgery, Erasmus Medical Center, Rotterdam, The Netherlands BACKGROUND & AIMS:
Patients with a first episode of acute pancreatitis can develop recurrent or chronic pancreatitis (CP). However, little is known about the incidence or risk factors for these events.
METHODS:
We performed a cross-sectional study of 669 patients with a first episode of acute pancreatitis admitted to 15 Dutch hospitals from December 2003 through March 2007. We collected information on disease course, outpatient visits, and hospital readmissions, as well as results from imaging, laboratory, and histology studies. Standardized follow-up questionnaires were sent to all available patients to collect information on hospitalizations and interventions for pancreatic disease, abdominal pain, steatorrhea, diabetes mellitus, medications, and alcohol and tobacco use. Patients were followed up for a median time period of 57 months. Primary end points were recurrent pancreatitis and CP. Risk factors were evaluated using regression analysis. The cumulative risk was assessed using Kaplan–Meier analysis.
RESULTS:
Recurrent pancreatitis developed in 117 patients (17%), and CP occurred in 51 patients (7.6%). Recurrent pancreatitis developed in 12% of patients with biliary disease, 24% of patients with alcoholic etiology, and 25% of patients with disease of idiopathic or other etiologies; CP occurred in 3%, 16%, and 10% of these patients, respectively. Etiology, smoking, and necrotizing pancreatitis were independent risk factors for recurrent pancreatitis and CP. Acute Physiology and Chronic Health Evaluation II scores at admission also were associated independently with recurrent pancreatitis. The cumulative risk for recurrent pancreatitis over 5 years was highest among smokers at 40% (compared with 13% for nonsmokers). For alcohol abusers and current smokers, the cumulative risks for CP were similar—approximately 18%. In contrast, the cumulative risk of CP increased to 30% in patients who smoked and abused alcohol.
Abbreviations used in this paper: APACHE II, Acute Physiology and Chronic Health Evaluation II score; CI, confidence interval; CP, chronic pancreatitis; IQR, interquartile range; OR, odds ratio; RP, recurrent pancreatitis.
Most current article © 2016 by the AGA Institute 1542-3565/$36.00 http://dx.doi.org/10.1016/j.cgh.2015.12.040
May 2016
CONCLUSIONS:
Progression Risk to Chronic Pancreatitis
739
Based on a retrospective analysis of patients admitted to Dutch hospitals, a first episode of acute pancreatitis leads to recurrent pancreatitis in 17% of patients, and almost 8% of patients progress to CP within 5 years. Progression was associated independently with alcoholic etiology, smoking, and a history of pancreatic necrosis. Smoking is the predominant risk factor for recurrent disease, whereas the combination of alcohol abuse and smoking produces the highest cumulative risk for chronic pancreatitis.
Keywords: Longitudinal Study; Pancreatitis Progression; Risk Factors; Chronic Pancreatitis.
he incidence of acute pancreatitis and its impact on health care have been increasing steadily over the past decades.1–3 Although most patients with acute pancreatitis recover completely, a small group may develop recurrent pancreatitis (RP) or progress to chronic pancreatitis (CP).4,5 There are a paucity of data regarding the natural course and risk factors of these events. Studying these aspects may offer ways for an earlier diagnosis and a better understanding of the underlying pathophysiology. This in turn may help to identify strategies for preventing disease progression. Therefore, we designed this study to describe the longterm outcomes in terms of RP and progression to CP in a large cohort of patients with a first acute pancreatitis episode collected during a previous randomized trial.6
T
Methods Study Design We performed a cross-sectional survey of a prospectively collected multicenter patient cohort.
Study Population Between December 2003 and March 2007, patients with a first episode of acute pancreatitis admitted to 15 Dutch hospitals were registered prospectively as part of a randomized controlled trial.6 Patient characteristics and details of the first pancreatitis episode were collected prospectively.6 In this study, all patients who survived the initial acute pancreatitis episode were followed up after a median duration of 57 months. Follow-up evaluation was based on medical records and patient questionnaires. Medical charts were checked for clinical course, outpatient visits, readmissions, imaging reports, laboratory tests, histology reports, surgical reports, and mortality. Standardized follow-up questionnaires were sent to all consenting patients. These included questions about hospitalizations and interventions for pancreatic disease, abdominal pain, steatorrhea, diabetes mellitus, medications, and alcohol and tobacco use. Questionnaires were sent by mail to patients, with a maximum of 2 reminders in case of no response. This study was reviewed by and obtained a waiver from the Medical Ethical Committee of the University Medical Center Utrecht.
Study Outcomes Primary end points were RP or CP after the first acute pancreatitis episode. RP was defined as follows: one or more recurrent pancreatitis episodes, defined as hospitalization for pancreatitis. Pancreatitis was defined as abdominal pain with an increase of serum amylase and/or lipase levels to at least 3 times the upper normal limit, or with computed tomography findings consistent with pancreatitis. With repeated episodes, typical clinical presentation was considered sufficient if during the course of hospitalization other causes were excluded. CP was defined based on the M-ANNHEIM diagnostic criteria for definite CP as follows7: a typical clinical history of chronic pancreatitis: such as recurrent pancreatitis or abdominal pain, and one or more of the following additional criteria: (1) pancreatic calcifications; (2) moderate or marked ductal lesions (according to the Cambridge classification); (3) marked and persistent exocrine insufficiency: pancreatic steatorrhea markedly reduced by enzyme supplementation (patient questionnaires), documented persistent steatorrhea in medical charts with the physician clearly linking this to pancreatic disease or prescription of pancreatic enzymes, or laboratory evidence of exocrine insufficiency (ie, fecal elastase levels 200 mg/g or abnormally increased fat in 24-h feces); and (4) typical histology of an adequate tissue specimen. Secondary outcomes were as follows: (1) abdominal pain of potential pancreatic origin: stated as the reason for an outpatient visit or hospitalization, or at least weekly occurring epigastric pain without other causes (patient questionnaire); (2) exocrine pancreatic insufficiency: as defined previously; (3) diabetes mellitus: clearly stated diagnosis or use of antidiabetic drugs; and (4) imaging abnormalities suggestive of CP: we reviewed all available reports of (endoscopic) ultrasound, computed tomography, magnetic resonance imaging, endoscopic retrograde cholangiopancreatography, and magnetic resonance cholangiopancreatography after the initial pancreatitis episode for abnormalities as defined by the M-ANNHEIM criteria.7
Risk Factors Data regarding the following potential risk factors were collected: age, sex, body mass index, etiology and
740
Ahmed Ali et al
severity of first acute pancreatitis, pancreatic necrosis, organ failure,6 Acute Physiology and Chronic Health Evaluation (APACHE) II score,8 modified Glasgow (Imrie) score, surgical intervention during first episode, number of RP episodes, alcohol consumption, smoking status (current smokers vs never/past smokers), and duration of follow-up evaluation. Biliary etiology of acute pancreatitis was defined as follows: (1) gallstones and/ or sludge on ultrasound or computed tomography, (2) dilated common bile duct, or as (3) 2 laboratory abnormalities: increased bilirubin level (>1.3 mg/dL), alanine aminotransferase levels greater than 100 U/L with an alanine aminotransferase level greater than the aspartate aminotransferase level; or an alkaline phosphatase level greater than 195 U/L with a g-glutamyltransferase level greater than 45 U/L.9 Etiology was classified as alcoholic if medical records documented alcohol abuse, or alcohol consumption of 40 gram or more per day (medical records or questionnaire) before the first episode.10 Patients lacking any of these criteria were classified as idiopathic, unless clear evidence existed for other causes (eg, hypertriglyceridemia or a clear link with medication).
Clinical Gastroenterology and Hepatology Vol. 14, No. 5 Table 1. Patient Characteristics
Median age, y (IQR) Sex, male, n (%) Body mass index, mean (SD) Etiology, n (%) Biliary Alcoholic Idiopathic/unknown Other (medication/ hypertriglyceridemia) Severe pancreatitis, n (%) Predicted Actual Pancreatic necrosis, n (%) Surgery at first episode, n (%) APACHE II at admission, median (IQR) Modified Glasgow score, median (IQR) Follow-up duration, median (IQR) Questionnaires, n (%) Available for questionnaires Filled-in questionnaire Refused questionnaire No reply Unavailable for questionnaires No further contact in prior consent Deceased Moved to unknown address
N
Characteristic
669 669 495 669
57 (42–70) 366 (55%) 27.5 (5.6) 384 153 108 24
(58%) (23%) (15%) (4%)
478 147 122 68 6 2 57
(71%) (22%) (18%) (10%) (4–10) (1–3) (49–65)
494 266 92 136 175 64 85 26
(74%) (40%) (14%) (20%) (26%) (10%) (13%) (4%)
669
669 669 669 669 669 669
Statistical Analysis Potential risk factors were evaluated by logistic regression. Results were presented as odds ratios (ORs) with 95% confidence intervals (CIs). Analyses for the 2 primary outcomes, RP and CP, were performed separately. For RP, a subgroup analysis for patients without progression to CP was performed. For CP, a subgroup analysis for patients with biliary etiology was performed. Risk factors showing a potential association (P < .2) in univariate analysis subsequently were entered into a multivariate model. For inter-related risk factors (eg, severity of pancreatitis and presence of necrosis), only the factor with the strongest association was included (to satisfy the linear independence assumption of regression). When RP was used as a risk factor for CP, associations between some factors and CP can be masked if those factors also are associated with RP. Therefore, were performed 2 analyses for CP: one with and one without RP as a covariate. Kaplan–Meier analysis showing the cumulative hazard risk of developing the primary end points over time was performed. Results were stratified by etiology and smoking. A P value of .05 was used as the threshold for statistical significance.
Results A total of 669 patients who survived a first episode of acute pancreatitis were included. Baseline characteristics are presented in Table 1. The median age was 57 years (interquartile range [IQR], 42–70 y) and 366 (55%) were males. Biliary (58%) and alcoholic (23%)
etiologies were most common. Data on the alcoholic etiology were obtained in all but 49 (7.3%) patients. For these patients the etiologic classification as made at prospective registration (ie, first acute pancreatitis) was used. Detailed data on continuation of alcohol consumption after the first acute pancreatitis episode were available for 259 patients who returned the study questionnaires. Of those, 131 (51%) continued drinking, and 128 (49%) completely stopped use of alcohol. Patients were followed up until October 2010, with a median follow-up period of 57 months (IQR, 49–65 mo). Questionnaires were sent to 494 (74%) patients, and 266 were completed and returned (response rate, 54%). Questionnaires were not sent when patients were deceased (85 patients; 13%), chose not to be contacted for future research at the initial registration (64 patients; 10%), or if no current address was found (26 patients; 4%).
Clinical Outcomes Primary and secondary study outcomes are presented in Table 2. RP episodes were observed in 117 (17%) patients. The median duration until the first RP was 5 months (IQR, 1–15 mo). CP was diagnosed in 51 (7.6%) patients. The median duration to diagnosis of CP was 21 months (IQR, 8–33 mo). CP was diagnosed within the first 3 months of the primary acute pancreatitis episode in 6 (12%) patients. Such a short interval suggests that this acute pancreatitis episode was the first
May 2016
Progression Risk to Chronic Pancreatitis
Risk Factors for Occurrence of Recurrent Pancreatitis
Table 2. Study Outcomes Findings (N ¼ 669)
Outcome Recurrent pancreatitis, n (%) 1 episode of RP 2 episodes of RP 3 episodes of RP Chronic pancreatitis, n (%) M-ANNHEIM criteria Secondary outcomes, n (%) Pain of potential pancreatic origin Exocrine pancreatic insufficiency Diabetes mellitus Imaging abnormalities suggestive of CP, n (%) US available, abnormalities EUS available, abnormalities CT/MRI available, abnormalities ERCP/MRCP available, abnormalities
117 77 14 26
741
A regression analysis of risk factors for RP is presented in Table 3. At univariate analysis, age, etiology, smoking, pancreatic necrosis, and APACHE II score at admission were associated significantly with the development of RP. In addition, the duration of follow-up evaluation, severity of the first episode, and surgical intervention during the first episode were correlated sufficiently (P < .2) to warrant inclusion in the multivariate model. At multivariate analysis, idiopathic/other etiology, smoking, pancreatic necrosis, and APACHE II score were found to be independent risk factors for the development of RP (Table 3). Notably, alcoholic etiology was not found to be associated independently (P ¼ .07). A cumulative risk analysis of the development of RP stratified by etiology and smoking status is presented in Figure 1. The cumulative risk for RP over 5 years was the highest among smokers (approximately 40%). The cumulative risks for patients with alcoholic and other/ idiopathic etiologies were comparable (approximately 30%). When smoking was combined with either alcoholic or idiopathic/other etiology the cumulative risk was approximately 50% after 5 years (not shown). The risk factors for the development of RP without
(17%) (12%) (2%) (4%)
51 (7.6%) 87 (13%) 34 (5.1%) 136 (20%)
628 (94%)/27 (4%) 48 (7.2%)/(1.2%) 487 (73%)/55 (8.2%) 291 (43%)/26 (3.9%)
CT, computed tomography; ERCP, endoscopic retrograde cholangiopancreatography; EUS, endoscopic ultrasound; MRCP, magnetic resonance cholangiopancreatography; MRI, magnetic resonance imaging.
manifestation of an existing CP. An RP episode either preceded CP or was the manifestation leading to the diagnosis of CP in 37 (73%) patients. Of the 117 patients with RP, 37 (32%) were diagnosed with CP.
Table 3. Univariate and Multivariate Analysis of Risk Factors Associated With Recurrent Pancreatitis Univariate analysis Factor
N
Age (per 10 years) Sex Males Females BMI Etiology of first episode Biliary Alcohol Idiopathic/other Smoking Current smoker Never smoked/past smoker Severe pancreatitisa Yes No Necrotizing pancreatitis Yes No Organ failure Yes No APACHE II score Modified Glasgow score Surgical intervention after first episode Yes No Duration of follow-up period, mo
669 669
n/N (%)
Multivariate analysis
OR (95% CI)
P value
OR (95% CI)
P value
0.80 (0.71–0.90)
<.001
-
-
1.00 (0.67–1.49)
1.0
0.99 (0.95–1.03) 1 2.29 (1.42–3.70) 2.39 (1.45–3.93)
.47 <.001 .001 .001
1 1.68 (0.96–2.96) 2.51 (1.44–4.37)
.004 .072 .001
50/158 (31.6%) 56/426 (13.1%)
3.06 (1.98–4.74)
<.001
2.77 (1.69–4.53)
<.001
33/147 (22.4%) 84/522 (16.1%)
1.51 (0.96–2.37)
.07
31/118 (26.3%) 86/551 (15.6%)
1.92 (1.20–3.08)
.006
2.53 (1.48–4.33)
.001
16/75 (21.3%) 101/594 (17.0%) -
1.32 (0.73–2.39)
.35
0.92 (0.88–0.97) 0.99 (0.87–1.12)
.001 .82
0.91 (0.86–0.96)
.001
1.67 (0.93–3.01)
.09
-
-
0.99 (0.97–1.01)
.19
-
-
64/366 (17.5%) 53/303 (17.5%) 495 669 47/384 (12.2%) 37/153 (24.2%) 33/132 (25.0%) 584
669
669
669
669 669 669
17/68 (25.0%) 100/601 (16.6%) 669
BMI, body mass index. a Not included in multivariate analysis due to correlation with another factors (ie, pancreatic necrosis).
742
Ahmed Ali et al
Clinical Gastroenterology and Hepatology Vol. 14, No. 5
Figure 1. Cumulative risk over time for (A and B) first recurrent pancreatitis and (C and D) progression to chronic pancreatitis, stratified by etiology and current smoking.
progression to CP also were analyzed. Age, sex, smoking, and APACHE II score on admission were included in the multivariate analysis. Younger age (OR, 0.80; 95% CI, 0.68–0.94) and smoking (OR, 2.2; 95% CI, 1.19–3.90) were associated independently with an increased risk of recurrences.
Risk Factors for Progression to Chronic Pancreatitis The univariate analysis of risk factors for CP is presented in Table 4. Sex, etiology, smoking, severity of first episode, necrotizing pancreatitis, organ failure, modified Glasgow score, surgical interventions during first episode, and RP all were associated significantly with the development of CP (P < .05). Results of the multivariate analysis are shown in Table 5. When RP was included as a covariate, alcoholic etiology (OR, 4.9; 95% CI, 2.0–11.5), necrotizing pancreatitis (OR, 8.8; 95% CI, 4.1–18.9), and
RP (OR, 2.9 per episode; 95% CI, 2.1–4.1) were found to be independent risk factors. Notably, smoking was not associated independently in this analysis. When RP was not included, alcoholic etiology (OR, 4.2; 95% CI, 1.8–9.7) and necrotizing pancreatitis (OR, 6.7; 95% CI, 3.4–13.0) remained independent risk factors. In addition, idiopathic/other etiology (OR, 4.0; 95% CI, 1.6–9.7) and smoking (OR, 2.9; 95% CI, 1.4–5.9) were identified as independent risk factors. The cumulative risk for progression to CP is plotted in Figure 1. The cumulative risks for patients with alcoholic etiologies and current smokers (both approximately 18%) were high. When an alcoholic etiology and current smoking co-existed the cumulative risk increased substantially to approximately 30% (not shown). Patients with biliary etiology had the lowest risk of developing CP, with a cumulative risk of approximately 3% over 5 years. In the subgroup analysis for patients with a biliary etiology only 13 of 384 patients with a biliary etiology
May 2016
Progression Risk to Chronic Pancreatitis
743
Table 4. Univariate Analysis of Risk Factors for Progression to Chronic Pancreatitis Chronic pancreatitis (N ¼ 51) Factor
N
n/N (%)
OR (95% CI)
P value
Age (per 10 years) Sex Males Females BMI Etiology of first episode Biliary Alcohol Idiopathic/other Smoking Current smoker Never smoked/past smoker Severity of episode (proven severe) Yes No Necrotizing pancreatitis Yes No Organ failure Yes No APACHE II score Modified Glasgow score Surgical intervention at first episode Yes No Recurrent pancreatitis (per episode) Number of recurrent pancreatitis episodes 0 1 2 Duration of follow-up period, mo
669
-
0.91 (0.77–1.07)
.26
669
35/366 (9.6%) 16/303 (5.3%) 13/384 (3.4%) 25/153 (16.3%) 13/132 (9.8%)
1.90 (1.03–3.50)
.04
0.99 (0.93–1.05) 1 5.57 (2.77–11.22) 3.12 (1.41–6.91)
.64 <.001 <.001 .005
26/158 (16.5%) 22/426 (5.2%)
3.62 (1.98–6.60)
<.001
26/147 (17.7%) 25/522 (4.8%)
4.27 (2.38–7.66)
<.001
26/118 (22.0%) 25/551 (4.5%)
5.95 (3.29–10.7)
<.001
14/75 (18.7%) 37/594 (6.2%) -
3.46 (1.77–6.75)
<.001
1.01 (0.95–1.08) 1.21 (1.02–1.43)
.71 .026
4.44 (2.29–8.64)
<.001
2.99 (2.25–3.98) 1 14.74 (6.81–31.9) 45.00 (19.4–104) 0.97 (0.95–1.00)
<.001 <.001 <.001 <.001 .068
495 669
584
669
669
669
669 669 669
15/68 (22.1%) 36/601 (6.0%) 12/552 (2.2%) 19/77 (24.7%) 20/40 (50%) -
669 669
669
BMI, body mass index.
developed CP, limiting the power of this analysis. Multivariate analysis showed that necrotizing pancreatitis (OR, 41.0; 95% CI, 4.30–393) and RP (OR, 4.15; 95% CI, 2.10–8.21) were the only independent risk factors.
Discussion Several important findings were observed in this large multicenter cohort study. First, the incidence rates of RP and CP were 17% and 8%, respectively, a median
Table 5. Multivariate Analysis of Risk Factors for Progression to Chronic Pancreatitis CP with RP as a risk factor
Sex, male Duration of follow-up period Etiology Biliary Alcohol Idiopathic/other Current smoking Necrotizing pancreatitis Modified Glasgow score Surgical intervention after first episode Recurrent pancreatitis (per episode)
CP without RP as a risk factor
OR (95% CI)
P value
OR (95% CI)
P value
-
NS NS .02 <.001 .08 NS <.001 NS NS <.001
-
NS NS .001 .001 .002 .004 <.001 NS NS -
1 4.85 (2.04–11.52) 2.45 (0.90–6.71) 8.78 (4.09–18.86) 2.90 (2.07–4.05)
1 (1.83–9.73) (1.64–9.65) (1.42–5.93) (3.40–13.01) Not included
4.22 3.98 2.90 6.65
744
Ahmed Ali et al
of 5 years after the first episode of acute pancreatitis. Second, smoking was found to be the most important factor for RP, although the combination of smoking and alcohol was most deleterious in terms of progression to CP. Third, pancreatic necrosis was found to be an independent risk factor for development of both RP and CP, which might be helpful in further understanding the pathophysiology of these entities. Few reports have studied the progression of acute pancreatitis to RP and CP.4,5,11 During the Marseille classification of 1985, little evidence existed on this topic, and this led experts to conclude that progression from acute pancreatitis to CP was extremely uncommon.12 It was thought that acute pancreatitis episodes in patients who developed CP were an early manifestation of co-existing CP. Ammann et al13,14 showed that up to 78% of patients with alcoholic pancreatitis progressed to CP over 16 years, none of whom had co-existent CP at the time of the first pancreatitis episode. Recent publications have shown rates of RP varying from 16% to 20% overall, and from 22% to 46% for alcoholic pancreatitis.4,5,11,14–19 For CP, rates varied even more widely from 3.6% to 22%.4,5,11,17–19 Rates of CP in patients with alcoholic, biliary, and idiopathic etiology were reportedly approximately 11% to 32%, 0% to 2%, and 0% to 13%, respectively.4,5,11,14,17–19 In this large study, these rates were 16%, 3%, and 10%, respectively. Although various factors, such as selection criteria, follow-up duration, and outcome definitions, might be responsible for the variations observed, most studies had important shortcomings. Two studies, one by Yasuda et al4 and one by Takeyama et al,5 had high rates of loss to follow-up evaluation of 65% and 72%, respectively.4,5 In addition, the study by Takeyama5 only used survey data and thus had no access to imaging to confirm the diagnosis of CP. Yasuda et al4 had a small sample size of only 45 patients. Population-based studies also have been reported from the United States and Denmark.17–19 The main drawbacks of these studies were the use of hospital registration data for diagnosing RP and CP without use of verifiable criteria, and the lack of detailed patient characteristics. The report by Lankisch et al11 was a well-conducted study and had a large sample size (532 patients) with adequate follow-up evaluation. Patients were, however, accrued over a relatively long period of time (approximately 20 years). During this period many developments took place in terms of imaging quality, classification, and treatment of pancreatitis, which raises a concern about the heterogeneity of the population. In addition, this study defined CP based on a self-developed scoring system that does not take clinical symptoms, such as pancreatitis episodes, abdominal pain, or steatorrhea, into consideration. This is different from how most physicians evaluate CP and also different from other well-known criteria, such as the Buchler et al20 and M-ANNHEIM classifications.7 Establishing the etiology of the first episode of acute pancreatitis can be challenging. In our study, patients
Clinical Gastroenterology and Hepatology Vol. 14, No. 5
were evaluated prospectively for inclusion in a randomized trial.6 Therefore, the first pancreatitis episode was well documented, allowing for high reliability in establishing biliary etiology based on laboratory and imaging findings. For an alcoholic etiology, one still had to rely on patient history. Although no clear accepted threshold exists for the amount or duration of alcohol consumption needed to classify pancreatitis as being of alcoholic etiology, a recent meta-analysis suggested a strong relationship (OR, approximately 3) between drinking 4 to 5 U/d and the development of pancreatitis.10 We defined alcoholic etiology as documented alcohol abuse, or a documented history of average alcohol consumption of more than 40 g/d before the first episode. Patients who did not fulfill these criteria were strictly classified as having idiopathic pancreatitis. This definition, combined with the known under-reporting of alcohol use, probably has led to an underestimation of the rate of alcoholic pancreatitis. This might explain the relatively high rate of disease progression in patients with idiopathic compared with those with biliary etiology, a phenomenon that was observed in most previous studies as well.4,5,11 An interesting finding of our cumulative risk analysis was that smoking played a more important role than etiology for the progression to RP. Cumulative risk for RP was the highest among smokers (approximately 40% over 5 years). This also indirectly was suggested by a recent case-control study in which heavy smoking was found as an independent risk factor for RP (OR, 1.91; P ¼ .01), although this was not the case for heavy alcohol consumption (OR, 1.16; P ¼ .6).21 For progression to CP the combination of smoking and alcohol seems to be most deleterious. The presence of both of these factors resulted in a cumulative risk of approximately 30%, compared with 10% for nonsmoking patients with alcoholic etiology. Previous studies indirectly suggested this additive effect of smoking and alcohol.21–23 Maisonneuve et al22 showed that smoking was associated with an earlier diagnosis of CP of alcoholic etiology by an average of 4.7 years. In a case-control study, Yadav et al21 showed an increase in risk of CP between heavy drinkers who smoked (OR, 4.69; 95% CI, 2.76–7.97) compared with heavy drinkers who did not smoke (OR, 1.37; 95% CI, 1.4–7.2), as well as a high increase in risk in heavy smokers who drank (OR, 8.07; 95% CI, 4.97–13.1) compared with heavy smokers who did not drink (OR, 2.35; 95% CI, 0.71–7.78). Our study identifies the absolute clinical effect of this relationship in a cohort design including unselected patients after their first acute pancreatitis episode. This allows for identification of patients with a higher risk earlier in the disease process, who can be the focus of preventive measures. Although a (randomized) trial studying the preventive effect of smoking cessation in this group of patients would be ideal, its feasibility would be doubtful because of poor patient compliance, the large sample size, and the long follow-up period required. Thus, longitudinal
May 2016
cohorts, such as this, are likely to be the best obtainable evidence. The mechanism by which smoking contributes to the development of RP and CP is not fully understood. Laboratory studies have indicated that nicotine activates several pathways as well as induces oxidative stress in pancreatic acinar cells.24,25 These changes cause a higher influx of calcium, resulting in cytotoxicity and cell death.24 The increased oxidative stress also increases cell proliferation and cell injury.25 Whether necrosis is associated with progression to RP and CP is a matter of continued debate. In 1994, Ammann and Muellhaupt14 found that progression from alcoholic acute pancreatitis to CP was correlated strongly with higher rates of “pseudocyst” formation, considered indicative of necrosis in that report. This, combined with histologic evaluations showing significant fibrosis after recurring pancreatitis, formed the base for the “necrosis–fibrosis”14 hypothesis for disease progression.26 Only 2 studies have examined this relationship since, with one study showing a higher rate of CP in patients with necrotizing pancreatitis,4 and another study showed an inverse relationship between necrosis and CP.5 Limitations of these studies, especially high rates of loss to follow-up evaluation, have been discussed previously. Our study had a large cohort and provides further reliable proof for the necrosis–fibrosis sequence for pancreatic disease progression. The strengths of this study included the high number of patients from hospitals in a confined geographic area allowing for comprehensive follow-up evaluation. In addition, the case-mix of academic and community hospitals added to the generalizability of study findings. Another strength was the availability of detailed data during the first pancreatitis episode. This study had certain limitations. Data were partly collected retrospectively. Smoking status and exact quantity of alcohol consumption at baseline, for example, had to be obtained from medical charts and questionnaires. This limited our ability to calculate a relationship between dose and outcome because of missing data. Second, it is possible that a small portion of patients chose to have their follow-up evaluation at different hospitals, potentially resulting in incomplete data. However, the included hospitals are acknowledged as the expert centers for pancreatitis in The Netherlands, and are geographically distributed across the country, minimizing chances of such bias. The use of questionnaires to verify outcomes further limited this potential problem. Another limitation was the median follow-up period of 5 years. Although results after a longer follow-up period may provide further insights, this analysis provides interesting results and is comparable with most available studies. In conclusion, after a first episode of acute pancreatitis, 1 of 6 patients will develop RP and 1 of 12 patients will progress to CP within 5 years. Smoking is the predominant risk factor for recurrent disease,
Progression Risk to Chronic Pancreatitis
745
whereas the combination of alcohol and smoking has the highest cumulative risk for chronic disease. Patients should be advised to discontinue alcohol use after a first pancreatitis episode, but also should be made aware of the importance of smoking cessation. The strong and independent association between pancreatic necrosis and both recurrent and chronic pancreatitis provides further support for the necrosis–fibrosis sequence.
References 1. Spanier B, Bruno MJ, Dijkgraaf MG. Incidence and mortality of acute and chronic pancreatitis in the Netherlands: a nationwide record-linked cohort study for the years 1995-2005. World J Gastroenterol 2013;19:3018–3026. 2. Spanier BW, Dijkgraaf MG, Bruno MJ. Trends and forecasts of hospital admissions for acute and chronic pancreatitis in the Netherlands. Eur J Gastroenterol Hepatol 2008;20:653–658. 3. Lowenfels AB, Maisonneuve P, Sullivan T. The changing character of acute pancreatitis: epidemiology, etiology, and prognosis. Curr Gastroenterol Rep 2009;11:97–103. 4. Yasuda T, Ueda T, Takeyama Y, et al. Long-term outcome of severe acute pancreatitis. J Hepatobiliary Pancreat Surg 2008; 15:397–402. 5. Takeyama Y. Long-term prognosis of acute pancreatitis in Japan. Clin Gastroenterol Hepatol 2009;7:S15–S17. 6. Besselink MG, van Santvoort HC, Buskens E, et al. Probiotic prophylaxis in predicted severe acute pancreatitis: a randomised, double-blind, placebo-controlled trial. Lancet 2008; 371:651–659. 7. Schneider A, Lohr JM, Singer MV. The M-ANNHEIM classification of chronic pancreatitis: introduction of a unifying classification system based on a review of previous classifications of the disease. J Gastroenterol 2007;42:101–119. 8. Larvin M, McMahon MJ. APACHE-II score for assessment and monitoring of acute pancreatitis. Lancet 1989;2:201–205. 9. van Santvoort HC, Besselink MG, de Vries AC, et al. Early endoscopic retrograde cholangiopancreatography in predicted severe acute biliary pancreatitis: a prospective multicenter study. Ann Surg 2009;250:68–75. 10. Irving HM, Samokhvalov AV, Rehm J. Alcohol as a risk factor for pancreatitis. A systematic review and meta-analysis. JOP 2009; 10:387–392. 11. Lankisch PG, Breuer N, Bruns A, et al. Natural history of acute pancreatitis: a long-term population-based study. Am J Gastroenterol 2009;104:2797–2805. 12. Singer MV, Gyr K, Sarles H. Revised classification of pancreatitis. Report of the Second International Symposium on the Classification of Pancreatitis in Marseille, France, March 28-30, 1984. Gastroenterology 1985;89:683–685. 13. Ammann RW, Muellhaupt B, Meyenberger C, et al. Alcoholic nonprogressive chronic pancreatitis: prospective long-term study of a large cohort with alcoholic acute pancreatitis (19761992). Pancreas 1994;9:365–373. 14. Ammann RW, Muellhaupt B. Progression of alcoholic acute to chronic pancreatitis. Gut 1994;35:552–556. 15. Lund H, Tonnesen H, Tonnesen MH, et al. Long-term recurrence and death rates after acute pancreatitis. Scand J Gastroenterol 2006;41:234–238. 16. Pelli H, Sand J, Laippala P, et al. Long-term follow-up after the first episode of acute alcoholic pancreatitis: time course and
746
Ahmed Ali et al
Clinical Gastroenterology and Hepatology Vol. 14, No. 5
risk factors for recurrence. Scand J Gastroenterol 2000;35: 552–555.
23. Morton C, Klatsky AL, Udaltsova N. Smoking, coffee, and pancreatitis. Am J Gastroenterol 2004;99:731–738.
17. Yadav D, Lee E, Papachristou GI, et al. A population-based evaluation of readmissions after first hospitalization for acute pancreatitis. Pancreas 2014;43:630–637.
24. Chowdhury P, MacLeod S, Udupa KB, et al. Pathophysiological effects of nicotine on the pancreas: an update. Exp Biol Med (Maywood) 2002;227:445–454.
18. Yadav D, O’Connell M, Papachristou GI. Natural history following the first attack of acute pancreatitis. Am J Gastroenterol 2012;107:1096–1103. 19. Nojgaard C, Becker U, Matzen P, et al. Progression from acute to chronic pancreatitis: prognostic factors, mortality, and natural course. Pancreas 2011;40:1195–1200.
25. Chowdhury P, Walker A. A cell-based approach to study changes in the pancreas following nicotine exposure in an animal model of injury. Langenbecks Arch Surg 2008; 393:547–555.
20. Buchler MW, Martignoni ME, Friess H, et al. A proposal for a new clinical classification of chronic pancreatitis. BMC Gastroenterol 2009;9:93. 21. Yadav D, Hawes RH, Brand RE, et al. Alcohol consumption, cigarette smoking, and the risk of recurrent acute and chronic pancreatitis. Arch Intern Med 2009;169:1035–1045. 22. Maisonneuve P, Lowenfels AB, Mullhaupt B, et al. Cigarette smoking accelerates progression of alcoholic chronic pancreatitis. Gut 2005;54:510–514.
26. Kloppel G, Maillet B. The morphological basis for the evolution of acute pancreatitis into chronic pancreatitis. Virchows Arch A Pathol Anat Histopathol 1992;420:1–4.
Reprint requests Address requests for reprints to: Marja A. Boermeester, MD, Department of Surgery (G4-132.1), Academic Medical Center Amsterdam, Meibergdreef 9, 1105 AZ, Amsterdam, The Netherlands. e-mail: [email protected]; fax: (31) (0) 20-566-9243. Conflicts of interest The authors disclose no conflicts.
PANCREAS, BILIARY TRACT, AND LIVER Risk of Recurrent Pancreatitis and Progression to Chronic Pancreatitis After a First Episode of Acute Pancreatitis Usama Ahmed Ali,*,‡ Yama Issa,* Julia C. Hagenaars,‡ Olaf J. Bakker,‡ Harry van Goor,§ Vincent B. Nieuwenhuijs,k Thomas L. Bollen,¶ Bert van Ramshorst,# Ben J. Witteman,** Menno A. Brink,‡‡ Alexander F. Schaapherder,§§ Cornelis H. Dejong,kk B. W. Marcel Spanier,¶¶ Joos Heisterkamp,## Erwin van der Harst,*** Casper H. van Eijck,‡‡‡ Marc G. Besselink,* Hein G. Gooszen,§§§ Hjalmar C. van Santvoort,* and Marja A. Boermeester,* for the Dutch Pancreatitis Study Group *Department of Surgery, Academic Medical Center, Amsterdam, The Netherlands; ‡Department of Surgery, University Medical Center Utrecht, Utrecht, The Netherlands; §Department of Surgery, §§§Department of Evidence Based Surgery, Radboud University Medical Center, Nijmegen, The Netherlands; kDepartment of Surgery, Isala Clinics, Zwolle, The Netherlands; ¶ Department of Radiology, #Department of Surgery, St Antonius Hospital, Nieuwegein, The Netherlands; **Department of Gastroenterology, Gelderse Vallei Hospital, Ede, The Netherlands; ‡‡Department of Gastroenterology, Meander Medical Center, Amersfoort, The Netherlands; §§Department of Surgery, Leiden University Medical Center, Leiden, The Netherlands; kk Department of Surgery and NUTRIM, University Hospital Maastricht, Maastricht, The Netherlands; ¶¶Department of Gastroenterology and Hepatology, Rijnstate Hospital, Arnhem, The Netherlands; ##Department of Surgery, St. Elisabeth Hospital, Tilburg, The Netherlands; ***Department of Surgery, Maasstad Ziekenhuis, Rotterdam, The Netherlands; ‡‡‡ Department of Surgery, Erasmus Medical Center, Rotterdam, The Netherlands BACKGROUND & AIMS:
Patients with a first episode of acute pancreatitis can develop recurrent or chronic pancreatitis (CP). However, little is known about the incidence or risk factors for these events.
METHODS:
We performed a cross-sectional study of 669 patients with a first episode of acute pancreatitis admitted to 15 Dutch hospitals from December 2003 through March 2007. We collected information on disease course, outpatient visits, and hospital readmissions, as well as results from imaging, laboratory, and histology studies. Standardized follow-up questionnaires were sent to all available patients to collect information on hospitalizations and interventions for pancreatic disease, abdominal pain, steatorrhea, diabetes mellitus, medications, and alcohol and tobacco use. Patients were followed up for a median time period of 57 months. Primary end points were recurrent pancreatitis and CP. Risk factors were evaluated using regression analysis. The cumulative risk was assessed using Kaplan–Meier analysis.
RESULTS:
Recurrent pancreatitis developed in 117 patients (17%), and CP occurred in 51 patients (7.6%). Recurrent pancreatitis developed in 12% of patients with biliary disease, 24% of patients with alcoholic etiology, and 25% of patients with disease of idiopathic or other etiologies; CP occurred in 3%, 16%, and 10% of these patients, respectively. Etiology, smoking, and necrotizing pancreatitis were independent risk factors for recurrent pancreatitis and CP. Acute Physiology and Chronic Health Evaluation II scores at admission also were associated independently with recurrent pancreatitis. The cumulative risk for recurrent pancreatitis over 5 years was highest among smokers at 40% (compared with 13% for nonsmokers). For alcohol abusers and current smokers, the cumulative risks for CP were similar—approximately 18%. In contrast, the cumulative risk of CP increased to 30% in patients who smoked and abused alcohol.
Abbreviations used in this paper: APACHE II, Acute Physiology and Chronic Health Evaluation II score; CI, confidence interval; CP, chronic pancreatitis; IQR, interquartile range; OR, odds ratio; RP, recurrent pancreatitis.
Most current article © 2016 by the AGA Institute 1542-3565/$36.00 http://dx.doi.org/10.1016/j.cgh.2015.12.040
May 2016
CONCLUSIONS:
Progression Risk to Chronic Pancreatitis
739
Based on a retrospective analysis of patients admitted to Dutch hospitals, a first episode of acute pancreatitis leads to recurrent pancreatitis in 17% of patients, and almost 8% of patients progress to CP within 5 years. Progression was associated independently with alcoholic etiology, smoking, and a history of pancreatic necrosis. Smoking is the predominant risk factor for recurrent disease, whereas the combination of alcohol abuse and smoking produces the highest cumulative risk for chronic pancreatitis.
Keywords: Longitudinal Study; Pancreatitis Progression; Risk Factors; Chronic Pancreatitis.
he incidence of acute pancreatitis and its impact on health care have been increasing steadily over the past decades.1–3 Although most patients with acute pancreatitis recover completely, a small group may develop recurrent pancreatitis (RP) or progress to chronic pancreatitis (CP).4,5 There are a paucity of data regarding the natural course and risk factors of these events. Studying these aspects may offer ways for an earlier diagnosis and a better understanding of the underlying pathophysiology. This in turn may help to identify strategies for preventing disease progression. Therefore, we designed this study to describe the longterm outcomes in terms of RP and progression to CP in a large cohort of patients with a first acute pancreatitis episode collected during a previous randomized trial.6
T
Methods Study Design We performed a cross-sectional survey of a prospectively collected multicenter patient cohort.
Study Population Between December 2003 and March 2007, patients with a first episode of acute pancreatitis admitted to 15 Dutch hospitals were registered prospectively as part of a randomized controlled trial.6 Patient characteristics and details of the first pancreatitis episode were collected prospectively.6 In this study, all patients who survived the initial acute pancreatitis episode were followed up after a median duration of 57 months. Follow-up evaluation was based on medical records and patient questionnaires. Medical charts were checked for clinical course, outpatient visits, readmissions, imaging reports, laboratory tests, histology reports, surgical reports, and mortality. Standardized follow-up questionnaires were sent to all consenting patients. These included questions about hospitalizations and interventions for pancreatic disease, abdominal pain, steatorrhea, diabetes mellitus, medications, and alcohol and tobacco use. Questionnaires were sent by mail to patients, with a maximum of 2 reminders in case of no response. This study was reviewed by and obtained a waiver from the Medical Ethical Committee of the University Medical Center Utrecht.
Study Outcomes Primary end points were RP or CP after the first acute pancreatitis episode. RP was defined as follows: one or more recurrent pancreatitis episodes, defined as hospitalization for pancreatitis. Pancreatitis was defined as abdominal pain with an increase of serum amylase and/or lipase levels to at least 3 times the upper normal limit, or with computed tomography findings consistent with pancreatitis. With repeated episodes, typical clinical presentation was considered sufficient if during the course of hospitalization other causes were excluded. CP was defined based on the M-ANNHEIM diagnostic criteria for definite CP as follows7: a typical clinical history of chronic pancreatitis: such as recurrent pancreatitis or abdominal pain, and one or more of the following additional criteria: (1) pancreatic calcifications; (2) moderate or marked ductal lesions (according to the Cambridge classification); (3) marked and persistent exocrine insufficiency: pancreatic steatorrhea markedly reduced by enzyme supplementation (patient questionnaires), documented persistent steatorrhea in medical charts with the physician clearly linking this to pancreatic disease or prescription of pancreatic enzymes, or laboratory evidence of exocrine insufficiency (ie, fecal elastase levels 200 mg/g or abnormally increased fat in 24-h feces); and (4) typical histology of an adequate tissue specimen. Secondary outcomes were as follows: (1) abdominal pain of potential pancreatic origin: stated as the reason for an outpatient visit or hospitalization, or at least weekly occurring epigastric pain without other causes (patient questionnaire); (2) exocrine pancreatic insufficiency: as defined previously; (3) diabetes mellitus: clearly stated diagnosis or use of antidiabetic drugs; and (4) imaging abnormalities suggestive of CP: we reviewed all available reports of (endoscopic) ultrasound, computed tomography, magnetic resonance imaging, endoscopic retrograde cholangiopancreatography, and magnetic resonance cholangiopancreatography after the initial pancreatitis episode for abnormalities as defined by the M-ANNHEIM criteria.7
Risk Factors Data regarding the following potential risk factors were collected: age, sex, body mass index, etiology and
740
Ahmed Ali et al
severity of first acute pancreatitis, pancreatic necrosis, organ failure,6 Acute Physiology and Chronic Health Evaluation (APACHE) II score,8 modified Glasgow (Imrie) score, surgical intervention during first episode, number of RP episodes, alcohol consumption, smoking status (current smokers vs never/past smokers), and duration of follow-up evaluation. Biliary etiology of acute pancreatitis was defined as follows: (1) gallstones and/ or sludge on ultrasound or computed tomography, (2) dilated common bile duct, or as (3) 2 laboratory abnormalities: increased bilirubin level (>1.3 mg/dL), alanine aminotransferase levels greater than 100 U/L with an alanine aminotransferase level greater than the aspartate aminotransferase level; or an alkaline phosphatase level greater than 195 U/L with a g-glutamyltransferase level greater than 45 U/L.9 Etiology was classified as alcoholic if medical records documented alcohol abuse, or alcohol consumption of 40 gram or more per day (medical records or questionnaire) before the first episode.10 Patients lacking any of these criteria were classified as idiopathic, unless clear evidence existed for other causes (eg, hypertriglyceridemia or a clear link with medication).
Clinical Gastroenterology and Hepatology Vol. 14, No. 5 Table 1. Patient Characteristics
Median age, y (IQR) Sex, male, n (%) Body mass index, mean (SD) Etiology, n (%) Biliary Alcoholic Idiopathic/unknown Other (medication/ hypertriglyceridemia) Severe pancreatitis, n (%) Predicted Actual Pancreatic necrosis, n (%) Surgery at first episode, n (%) APACHE II at admission, median (IQR) Modified Glasgow score, median (IQR) Follow-up duration, median (IQR) Questionnaires, n (%) Available for questionnaires Filled-in questionnaire Refused questionnaire No reply Unavailable for questionnaires No further contact in prior consent Deceased Moved to unknown address
N
Characteristic
669 669 495 669
57 (42–70) 366 (55%) 27.5 (5.6) 384 153 108 24
(58%) (23%) (15%) (4%)
478 147 122 68 6 2 57
(71%) (22%) (18%) (10%) (4–10) (1–3) (49–65)
494 266 92 136 175 64 85 26
(74%) (40%) (14%) (20%) (26%) (10%) (13%) (4%)
669
669 669 669 669 669 669
Statistical Analysis Potential risk factors were evaluated by logistic regression. Results were presented as odds ratios (ORs) with 95% confidence intervals (CIs). Analyses for the 2 primary outcomes, RP and CP, were performed separately. For RP, a subgroup analysis for patients without progression to CP was performed. For CP, a subgroup analysis for patients with biliary etiology was performed. Risk factors showing a potential association (P < .2) in univariate analysis subsequently were entered into a multivariate model. For inter-related risk factors (eg, severity of pancreatitis and presence of necrosis), only the factor with the strongest association was included (to satisfy the linear independence assumption of regression). When RP was used as a risk factor for CP, associations between some factors and CP can be masked if those factors also are associated with RP. Therefore, were performed 2 analyses for CP: one with and one without RP as a covariate. Kaplan–Meier analysis showing the cumulative hazard risk of developing the primary end points over time was performed. Results were stratified by etiology and smoking. A P value of .05 was used as the threshold for statistical significance.
Results A total of 669 patients who survived a first episode of acute pancreatitis were included. Baseline characteristics are presented in Table 1. The median age was 57 years (interquartile range [IQR], 42–70 y) and 366 (55%) were males. Biliary (58%) and alcoholic (23%)
etiologies were most common. Data on the alcoholic etiology were obtained in all but 49 (7.3%) patients. For these patients the etiologic classification as made at prospective registration (ie, first acute pancreatitis) was used. Detailed data on continuation of alcohol consumption after the first acute pancreatitis episode were available for 259 patients who returned the study questionnaires. Of those, 131 (51%) continued drinking, and 128 (49%) completely stopped use of alcohol. Patients were followed up until October 2010, with a median follow-up period of 57 months (IQR, 49–65 mo). Questionnaires were sent to 494 (74%) patients, and 266 were completed and returned (response rate, 54%). Questionnaires were not sent when patients were deceased (85 patients; 13%), chose not to be contacted for future research at the initial registration (64 patients; 10%), or if no current address was found (26 patients; 4%).
Clinical Outcomes Primary and secondary study outcomes are presented in Table 2. RP episodes were observed in 117 (17%) patients. The median duration until the first RP was 5 months (IQR, 1–15 mo). CP was diagnosed in 51 (7.6%) patients. The median duration to diagnosis of CP was 21 months (IQR, 8–33 mo). CP was diagnosed within the first 3 months of the primary acute pancreatitis episode in 6 (12%) patients. Such a short interval suggests that this acute pancreatitis episode was the first
May 2016
Progression Risk to Chronic Pancreatitis
Risk Factors for Occurrence of Recurrent Pancreatitis
Table 2. Study Outcomes Findings (N ¼ 669)
Outcome Recurrent pancreatitis, n (%) 1 episode of RP 2 episodes of RP 3 episodes of RP Chronic pancreatitis, n (%) M-ANNHEIM criteria Secondary outcomes, n (%) Pain of potential pancreatic origin Exocrine pancreatic insufficiency Diabetes mellitus Imaging abnormalities suggestive of CP, n (%) US available, abnormalities EUS available, abnormalities CT/MRI available, abnormalities ERCP/MRCP available, abnormalities
117 77 14 26
741
A regression analysis of risk factors for RP is presented in Table 3. At univariate analysis, age, etiology, smoking, pancreatic necrosis, and APACHE II score at admission were associated significantly with the development of RP. In addition, the duration of follow-up evaluation, severity of the first episode, and surgical intervention during the first episode were correlated sufficiently (P < .2) to warrant inclusion in the multivariate model. At multivariate analysis, idiopathic/other etiology, smoking, pancreatic necrosis, and APACHE II score were found to be independent risk factors for the development of RP (Table 3). Notably, alcoholic etiology was not found to be associated independently (P ¼ .07). A cumulative risk analysis of the development of RP stratified by etiology and smoking status is presented in Figure 1. The cumulative risk for RP over 5 years was the highest among smokers (approximately 40%). The cumulative risks for patients with alcoholic and other/ idiopathic etiologies were comparable (approximately 30%). When smoking was combined with either alcoholic or idiopathic/other etiology the cumulative risk was approximately 50% after 5 years (not shown). The risk factors for the development of RP without
(17%) (12%) (2%) (4%)
51 (7.6%) 87 (13%) 34 (5.1%) 136 (20%)
628 (94%)/27 (4%) 48 (7.2%)/(1.2%) 487 (73%)/55 (8.2%) 291 (43%)/26 (3.9%)
CT, computed tomography; ERCP, endoscopic retrograde cholangiopancreatography; EUS, endoscopic ultrasound; MRCP, magnetic resonance cholangiopancreatography; MRI, magnetic resonance imaging.
manifestation of an existing CP. An RP episode either preceded CP or was the manifestation leading to the diagnosis of CP in 37 (73%) patients. Of the 117 patients with RP, 37 (32%) were diagnosed with CP.
Table 3. Univariate and Multivariate Analysis of Risk Factors Associated With Recurrent Pancreatitis Univariate analysis Factor
N
Age (per 10 years) Sex Males Females BMI Etiology of first episode Biliary Alcohol Idiopathic/other Smoking Current smoker Never smoked/past smoker Severe pancreatitisa Yes No Necrotizing pancreatitis Yes No Organ failure Yes No APACHE II score Modified Glasgow score Surgical intervention after first episode Yes No Duration of follow-up period, mo
669 669
n/N (%)
Multivariate analysis
OR (95% CI)
P value
OR (95% CI)
P value
0.80 (0.71–0.90)
<.001
-
-
1.00 (0.67–1.49)
1.0
0.99 (0.95–1.03) 1 2.29 (1.42–3.70) 2.39 (1.45–3.93)
.47 <.001 .001 .001
1 1.68 (0.96–2.96) 2.51 (1.44–4.37)
.004 .072 .001
50/158 (31.6%) 56/426 (13.1%)
3.06 (1.98–4.74)
<.001
2.77 (1.69–4.53)
<.001
33/147 (22.4%) 84/522 (16.1%)
1.51 (0.96–2.37)
.07
31/118 (26.3%) 86/551 (15.6%)
1.92 (1.20–3.08)
.006
2.53 (1.48–4.33)
.001
16/75 (21.3%) 101/594 (17.0%) -
1.32 (0.73–2.39)
.35
0.92 (0.88–0.97) 0.99 (0.87–1.12)
.001 .82
0.91 (0.86–0.96)
.001
1.67 (0.93–3.01)
.09
-
-
0.99 (0.97–1.01)
.19
-
-
64/366 (17.5%) 53/303 (17.5%) 495 669 47/384 (12.2%) 37/153 (24.2%) 33/132 (25.0%) 584
669
669
669
669 669 669
17/68 (25.0%) 100/601 (16.6%) 669
BMI, body mass index. a Not included in multivariate analysis due to correlation with another factors (ie, pancreatic necrosis).
742
Ahmed Ali et al
Clinical Gastroenterology and Hepatology Vol. 14, No. 5
Figure 1. Cumulative risk over time for (A and B) first recurrent pancreatitis and (C and D) progression to chronic pancreatitis, stratified by etiology and current smoking.
progression to CP also were analyzed. Age, sex, smoking, and APACHE II score on admission were included in the multivariate analysis. Younger age (OR, 0.80; 95% CI, 0.68–0.94) and smoking (OR, 2.2; 95% CI, 1.19–3.90) were associated independently with an increased risk of recurrences.
Risk Factors for Progression to Chronic Pancreatitis The univariate analysis of risk factors for CP is presented in Table 4. Sex, etiology, smoking, severity of first episode, necrotizing pancreatitis, organ failure, modified Glasgow score, surgical interventions during first episode, and RP all were associated significantly with the development of CP (P < .05). Results of the multivariate analysis are shown in Table 5. When RP was included as a covariate, alcoholic etiology (OR, 4.9; 95% CI, 2.0–11.5), necrotizing pancreatitis (OR, 8.8; 95% CI, 4.1–18.9), and
RP (OR, 2.9 per episode; 95% CI, 2.1–4.1) were found to be independent risk factors. Notably, smoking was not associated independently in this analysis. When RP was not included, alcoholic etiology (OR, 4.2; 95% CI, 1.8–9.7) and necrotizing pancreatitis (OR, 6.7; 95% CI, 3.4–13.0) remained independent risk factors. In addition, idiopathic/other etiology (OR, 4.0; 95% CI, 1.6–9.7) and smoking (OR, 2.9; 95% CI, 1.4–5.9) were identified as independent risk factors. The cumulative risk for progression to CP is plotted in Figure 1. The cumulative risks for patients with alcoholic etiologies and current smokers (both approximately 18%) were high. When an alcoholic etiology and current smoking co-existed the cumulative risk increased substantially to approximately 30% (not shown). Patients with biliary etiology had the lowest risk of developing CP, with a cumulative risk of approximately 3% over 5 years. In the subgroup analysis for patients with a biliary etiology only 13 of 384 patients with a biliary etiology
May 2016
Progression Risk to Chronic Pancreatitis
743
Table 4. Univariate Analysis of Risk Factors for Progression to Chronic Pancreatitis Chronic pancreatitis (N ¼ 51) Factor
N
n/N (%)
OR (95% CI)
P value
Age (per 10 years) Sex Males Females BMI Etiology of first episode Biliary Alcohol Idiopathic/other Smoking Current smoker Never smoked/past smoker Severity of episode (proven severe) Yes No Necrotizing pancreatitis Yes No Organ failure Yes No APACHE II score Modified Glasgow score Surgical intervention at first episode Yes No Recurrent pancreatitis (per episode) Number of recurrent pancreatitis episodes 0 1 2 Duration of follow-up period, mo
669
-
0.91 (0.77–1.07)
.26
669
35/366 (9.6%) 16/303 (5.3%) 13/384 (3.4%) 25/153 (16.3%) 13/132 (9.8%)
1.90 (1.03–3.50)
.04
0.99 (0.93–1.05) 1 5.57 (2.77–11.22) 3.12 (1.41–6.91)
.64 <.001 <.001 .005
26/158 (16.5%) 22/426 (5.2%)
3.62 (1.98–6.60)
<.001
26/147 (17.7%) 25/522 (4.8%)
4.27 (2.38–7.66)
<.001
26/118 (22.0%) 25/551 (4.5%)
5.95 (3.29–10.7)
<.001
14/75 (18.7%) 37/594 (6.2%) -
3.46 (1.77–6.75)
<.001
1.01 (0.95–1.08) 1.21 (1.02–1.43)
.71 .026
4.44 (2.29–8.64)
<.001
2.99 (2.25–3.98) 1 14.74 (6.81–31.9) 45.00 (19.4–104) 0.97 (0.95–1.00)
<.001 <.001 <.001 <.001 .068
495 669
584
669
669
669
669 669 669
15/68 (22.1%) 36/601 (6.0%) 12/552 (2.2%) 19/77 (24.7%) 20/40 (50%) -
669 669
669
BMI, body mass index.
developed CP, limiting the power of this analysis. Multivariate analysis showed that necrotizing pancreatitis (OR, 41.0; 95% CI, 4.30–393) and RP (OR, 4.15; 95% CI, 2.10–8.21) were the only independent risk factors.
Discussion Several important findings were observed in this large multicenter cohort study. First, the incidence rates of RP and CP were 17% and 8%, respectively, a median
Table 5. Multivariate Analysis of Risk Factors for Progression to Chronic Pancreatitis CP with RP as a risk factor
Sex, male Duration of follow-up period Etiology Biliary Alcohol Idiopathic/other Current smoking Necrotizing pancreatitis Modified Glasgow score Surgical intervention after first episode Recurrent pancreatitis (per episode)
CP without RP as a risk factor
OR (95% CI)
P value
OR (95% CI)
P value
-
NS NS .02 <.001 .08 NS <.001 NS NS <.001
-
NS NS .001 .001 .002 .004 <.001 NS NS -
1 4.85 (2.04–11.52) 2.45 (0.90–6.71) 8.78 (4.09–18.86) 2.90 (2.07–4.05)
1 (1.83–9.73) (1.64–9.65) (1.42–5.93) (3.40–13.01) Not included
4.22 3.98 2.90 6.65
744
Ahmed Ali et al
of 5 years after the first episode of acute pancreatitis. Second, smoking was found to be the most important factor for RP, although the combination of smoking and alcohol was most deleterious in terms of progression to CP. Third, pancreatic necrosis was found to be an independent risk factor for development of both RP and CP, which might be helpful in further understanding the pathophysiology of these entities. Few reports have studied the progression of acute pancreatitis to RP and CP.4,5,11 During the Marseille classification of 1985, little evidence existed on this topic, and this led experts to conclude that progression from acute pancreatitis to CP was extremely uncommon.12 It was thought that acute pancreatitis episodes in patients who developed CP were an early manifestation of co-existing CP. Ammann et al13,14 showed that up to 78% of patients with alcoholic pancreatitis progressed to CP over 16 years, none of whom had co-existent CP at the time of the first pancreatitis episode. Recent publications have shown rates of RP varying from 16% to 20% overall, and from 22% to 46% for alcoholic pancreatitis.4,5,11,14–19 For CP, rates varied even more widely from 3.6% to 22%.4,5,11,17–19 Rates of CP in patients with alcoholic, biliary, and idiopathic etiology were reportedly approximately 11% to 32%, 0% to 2%, and 0% to 13%, respectively.4,5,11,14,17–19 In this large study, these rates were 16%, 3%, and 10%, respectively. Although various factors, such as selection criteria, follow-up duration, and outcome definitions, might be responsible for the variations observed, most studies had important shortcomings. Two studies, one by Yasuda et al4 and one by Takeyama et al,5 had high rates of loss to follow-up evaluation of 65% and 72%, respectively.4,5 In addition, the study by Takeyama5 only used survey data and thus had no access to imaging to confirm the diagnosis of CP. Yasuda et al4 had a small sample size of only 45 patients. Population-based studies also have been reported from the United States and Denmark.17–19 The main drawbacks of these studies were the use of hospital registration data for diagnosing RP and CP without use of verifiable criteria, and the lack of detailed patient characteristics. The report by Lankisch et al11 was a well-conducted study and had a large sample size (532 patients) with adequate follow-up evaluation. Patients were, however, accrued over a relatively long period of time (approximately 20 years). During this period many developments took place in terms of imaging quality, classification, and treatment of pancreatitis, which raises a concern about the heterogeneity of the population. In addition, this study defined CP based on a self-developed scoring system that does not take clinical symptoms, such as pancreatitis episodes, abdominal pain, or steatorrhea, into consideration. This is different from how most physicians evaluate CP and also different from other well-known criteria, such as the Buchler et al20 and M-ANNHEIM classifications.7 Establishing the etiology of the first episode of acute pancreatitis can be challenging. In our study, patients
Clinical Gastroenterology and Hepatology Vol. 14, No. 5
were evaluated prospectively for inclusion in a randomized trial.6 Therefore, the first pancreatitis episode was well documented, allowing for high reliability in establishing biliary etiology based on laboratory and imaging findings. For an alcoholic etiology, one still had to rely on patient history. Although no clear accepted threshold exists for the amount or duration of alcohol consumption needed to classify pancreatitis as being of alcoholic etiology, a recent meta-analysis suggested a strong relationship (OR, approximately 3) between drinking 4 to 5 U/d and the development of pancreatitis.10 We defined alcoholic etiology as documented alcohol abuse, or a documented history of average alcohol consumption of more than 40 g/d before the first episode. Patients who did not fulfill these criteria were strictly classified as having idiopathic pancreatitis. This definition, combined with the known under-reporting of alcohol use, probably has led to an underestimation of the rate of alcoholic pancreatitis. This might explain the relatively high rate of disease progression in patients with idiopathic compared with those with biliary etiology, a phenomenon that was observed in most previous studies as well.4,5,11 An interesting finding of our cumulative risk analysis was that smoking played a more important role than etiology for the progression to RP. Cumulative risk for RP was the highest among smokers (approximately 40% over 5 years). This also indirectly was suggested by a recent case-control study in which heavy smoking was found as an independent risk factor for RP (OR, 1.91; P ¼ .01), although this was not the case for heavy alcohol consumption (OR, 1.16; P ¼ .6).21 For progression to CP the combination of smoking and alcohol seems to be most deleterious. The presence of both of these factors resulted in a cumulative risk of approximately 30%, compared with 10% for nonsmoking patients with alcoholic etiology. Previous studies indirectly suggested this additive effect of smoking and alcohol.21–23 Maisonneuve et al22 showed that smoking was associated with an earlier diagnosis of CP of alcoholic etiology by an average of 4.7 years. In a case-control study, Yadav et al21 showed an increase in risk of CP between heavy drinkers who smoked (OR, 4.69; 95% CI, 2.76–7.97) compared with heavy drinkers who did not smoke (OR, 1.37; 95% CI, 1.4–7.2), as well as a high increase in risk in heavy smokers who drank (OR, 8.07; 95% CI, 4.97–13.1) compared with heavy smokers who did not drink (OR, 2.35; 95% CI, 0.71–7.78). Our study identifies the absolute clinical effect of this relationship in a cohort design including unselected patients after their first acute pancreatitis episode. This allows for identification of patients with a higher risk earlier in the disease process, who can be the focus of preventive measures. Although a (randomized) trial studying the preventive effect of smoking cessation in this group of patients would be ideal, its feasibility would be doubtful because of poor patient compliance, the large sample size, and the long follow-up period required. Thus, longitudinal
May 2016
cohorts, such as this, are likely to be the best obtainable evidence. The mechanism by which smoking contributes to the development of RP and CP is not fully understood. Laboratory studies have indicated that nicotine activates several pathways as well as induces oxidative stress in pancreatic acinar cells.24,25 These changes cause a higher influx of calcium, resulting in cytotoxicity and cell death.24 The increased oxidative stress also increases cell proliferation and cell injury.25 Whether necrosis is associated with progression to RP and CP is a matter of continued debate. In 1994, Ammann and Muellhaupt14 found that progression from alcoholic acute pancreatitis to CP was correlated strongly with higher rates of “pseudocyst” formation, considered indicative of necrosis in that report. This, combined with histologic evaluations showing significant fibrosis after recurring pancreatitis, formed the base for the “necrosis–fibrosis”14 hypothesis for disease progression.26 Only 2 studies have examined this relationship since, with one study showing a higher rate of CP in patients with necrotizing pancreatitis,4 and another study showed an inverse relationship between necrosis and CP.5 Limitations of these studies, especially high rates of loss to follow-up evaluation, have been discussed previously. Our study had a large cohort and provides further reliable proof for the necrosis–fibrosis sequence for pancreatic disease progression. The strengths of this study included the high number of patients from hospitals in a confined geographic area allowing for comprehensive follow-up evaluation. In addition, the case-mix of academic and community hospitals added to the generalizability of study findings. Another strength was the availability of detailed data during the first pancreatitis episode. This study had certain limitations. Data were partly collected retrospectively. Smoking status and exact quantity of alcohol consumption at baseline, for example, had to be obtained from medical charts and questionnaires. This limited our ability to calculate a relationship between dose and outcome because of missing data. Second, it is possible that a small portion of patients chose to have their follow-up evaluation at different hospitals, potentially resulting in incomplete data. However, the included hospitals are acknowledged as the expert centers for pancreatitis in The Netherlands, and are geographically distributed across the country, minimizing chances of such bias. The use of questionnaires to verify outcomes further limited this potential problem. Another limitation was the median follow-up period of 5 years. Although results after a longer follow-up period may provide further insights, this analysis provides interesting results and is comparable with most available studies. In conclusion, after a first episode of acute pancreatitis, 1 of 6 patients will develop RP and 1 of 12 patients will progress to CP within 5 years. Smoking is the predominant risk factor for recurrent disease,
Progression Risk to Chronic Pancreatitis
745
whereas the combination of alcohol and smoking has the highest cumulative risk for chronic disease. Patients should be advised to discontinue alcohol use after a first pancreatitis episode, but also should be made aware of the importance of smoking cessation. The strong and independent association between pancreatic necrosis and both recurrent and chronic pancreatitis provides further support for the necrosis–fibrosis sequence.
References 1. Spanier B, Bruno MJ, Dijkgraaf MG. Incidence and mortality of acute and chronic pancreatitis in the Netherlands: a nationwide record-linked cohort study for the years 1995-2005. World J Gastroenterol 2013;19:3018–3026. 2. Spanier BW, Dijkgraaf MG, Bruno MJ. Trends and forecasts of hospital admissions for acute and chronic pancreatitis in the Netherlands. Eur J Gastroenterol Hepatol 2008;20:653–658. 3. Lowenfels AB, Maisonneuve P, Sullivan T. The changing character of acute pancreatitis: epidemiology, etiology, and prognosis. Curr Gastroenterol Rep 2009;11:97–103. 4. Yasuda T, Ueda T, Takeyama Y, et al. Long-term outcome of severe acute pancreatitis. J Hepatobiliary Pancreat Surg 2008; 15:397–402. 5. Takeyama Y. Long-term prognosis of acute pancreatitis in Japan. Clin Gastroenterol Hepatol 2009;7:S15–S17. 6. Besselink MG, van Santvoort HC, Buskens E, et al. Probiotic prophylaxis in predicted severe acute pancreatitis: a randomised, double-blind, placebo-controlled trial. Lancet 2008; 371:651–659. 7. Schneider A, Lohr JM, Singer MV. The M-ANNHEIM classification of chronic pancreatitis: introduction of a unifying classification system based on a review of previous classifications of the disease. J Gastroenterol 2007;42:101–119. 8. Larvin M, McMahon MJ. APACHE-II score for assessment and monitoring of acute pancreatitis. Lancet 1989;2:201–205. 9. van Santvoort HC, Besselink MG, de Vries AC, et al. Early endoscopic retrograde cholangiopancreatography in predicted severe acute biliary pancreatitis: a prospective multicenter study. Ann Surg 2009;250:68–75. 10. Irving HM, Samokhvalov AV, Rehm J. Alcohol as a risk factor for pancreatitis. A systematic review and meta-analysis. JOP 2009; 10:387–392. 11. Lankisch PG, Breuer N, Bruns A, et al. Natural history of acute pancreatitis: a long-term population-based study. Am J Gastroenterol 2009;104:2797–2805. 12. Singer MV, Gyr K, Sarles H. Revised classification of pancreatitis. Report of the Second International Symposium on the Classification of Pancreatitis in Marseille, France, March 28-30, 1984. Gastroenterology 1985;89:683–685. 13. Ammann RW, Muellhaupt B, Meyenberger C, et al. Alcoholic nonprogressive chronic pancreatitis: prospective long-term study of a large cohort with alcoholic acute pancreatitis (19761992). Pancreas 1994;9:365–373. 14. Ammann RW, Muellhaupt B. Progression of alcoholic acute to chronic pancreatitis. Gut 1994;35:552–556. 15. Lund H, Tonnesen H, Tonnesen MH, et al. Long-term recurrence and death rates after acute pancreatitis. Scand J Gastroenterol 2006;41:234–238. 16. Pelli H, Sand J, Laippala P, et al. Long-term follow-up after the first episode of acute alcoholic pancreatitis: time course and
746
Ahmed Ali et al
Clinical Gastroenterology and Hepatology Vol. 14, No. 5
risk factors for recurrence. Scand J Gastroenterol 2000;35: 552–555.
23. Morton C, Klatsky AL, Udaltsova N. Smoking, coffee, and pancreatitis. Am J Gastroenterol 2004;99:731–738.
17. Yadav D, Lee E, Papachristou GI, et al. A population-based evaluation of readmissions after first hospitalization for acute pancreatitis. Pancreas 2014;43:630–637.
24. Chowdhury P, MacLeod S, Udupa KB, et al. Pathophysiological effects of nicotine on the pancreas: an update. Exp Biol Med (Maywood) 2002;227:445–454.
18. Yadav D, O’Connell M, Papachristou GI. Natural history following the first attack of acute pancreatitis. Am J Gastroenterol 2012;107:1096–1103. 19. Nojgaard C, Becker U, Matzen P, et al. Progression from acute to chronic pancreatitis: prognostic factors, mortality, and natural course. Pancreas 2011;40:1195–1200.
25. Chowdhury P, Walker A. A cell-based approach to study changes in the pancreas following nicotine exposure in an animal model of injury. Langenbecks Arch Surg 2008; 393:547–555.
20. Buchler MW, Martignoni ME, Friess H, et al. A proposal for a new clinical classification of chronic pancreatitis. BMC Gastroenterol 2009;9:93. 21. Yadav D, Hawes RH, Brand RE, et al. Alcohol consumption, cigarette smoking, and the risk of recurrent acute and chronic pancreatitis. Arch Intern Med 2009;169:1035–1045. 22. Maisonneuve P, Lowenfels AB, Mullhaupt B, et al. Cigarette smoking accelerates progression of alcoholic chronic pancreatitis. Gut 2005;54:510–514.
26. Kloppel G, Maillet B. The morphological basis for the evolution of acute pancreatitis into chronic pancreatitis. Virchows Arch A Pathol Anat Histopathol 1992;420:1–4.
Reprint requests Address requests for reprints to: Marja A. Boermeester, MD, Department of Surgery (G4-132.1), Academic Medical Center Amsterdam, Meibergdreef 9, 1105 AZ, Amsterdam, The Netherlands. e-mail: [email protected]; fax: (31) (0) 20-566-9243. Conflicts of interest The authors disclose no conflicts.