Appraisal of the diagnostic procedures of acute pancreatitis in the guidelines

Four guidelines [2, 3, 7, 8] recommend abdominal pain as a reference indicator for the diagnosis of severe pancreatitis; the other three only mention the symptom. The recommendations of the three guidelines [3, 7, 8] for abdominal pain (sustained attacks, severe and acute episodes of epigastric pain usually radiating to the back) had a high similarity (80–100%) to the reference recommendations. The cited levels of evidence were only grade 5 [2, 7]. Six guidelines [2‐5, 7, 8] recommend serum lipase activity (or amylase) at least three times the upper normal limit as a reference indicator for the diagnosis of severe pancreatitis, among which three guidelines make an explicit recommendation [4, 5, 8], while the remaining three guidelines only casually recommend it [2, 3, 7]. Five guidelines [3‐5, 7, 8] recommending serum lipase activity (or amylase) at least three times the upper normal limit had a high degree of similarity (80–100%) to the reference recommendation, and the reference level of evidence was up to grade 4 [4, 8]. In six guidelines [1‐5, 7, 8], the characteristics of abdominal imaging are recommended as reference indicators for the diagnosis of severe pancreatitis. Three of these guidelines make explicit recommendations [4, 5, 8], and the remaining three guidelines merely mentioned these characteristics in the content [2, 3, 7]. The recommendations of five guidelines [3‐5, 7, 8] for the characteristic findings from abdominal imaging had a high similarity (80–100%) to the reference recommendation. The reference level of evidence, cited in only four of the guidelines, was a low grade 5 [2, 4, 5, 7].

Five guidelines [2‐4, 7, 8] recommend the revised Atlanta classification as a reference indicator for the diagnosis of severe pancreatitis, of which two guidelines make explicit recommendations [2, 4] and the remaining three only mentions the classification in the content [3, 7, 8]. The recommendations of four guidelines [3, 4, 7, 8] for the revised Atlanta classification had a high similarity (80–100%) to the reference recommendation. The reference level of evidence cited in four of the guidelines was grade 5 [2, 4, 7, 8]. Four guidelines [2‐4, 8] recommend CRP as a reference indicator for the diagnosis of severe pancreatitis, of which 3 guidelines make clear recommendations [2, 4, 8], and the remaining guideline only mentions CRP in the content [3]. The recommendations of two of the guidelines [4, 8] for CRP have a high similarity to the reference recommendation (80–100%). One guideline [3] demonstrated moderate similarity to the reference recommendation for CRP (40–60%). The reference level of evidence cited in one set of guidelines was grade 5 [2], and the reference level of evidence cited in another set of guidelines was a poor grade 5 [4]. Five guides [2‐4, 7, 8] recommend the APACHE-II score as a reference indicator for the diagnosis of severe pancreatitis, three of which make explicit recommendations [2, 4, 7], and two of which only mention the score in the content [3, 8]. The recommendations of four of the guidelines [3, 4, 7, 8] for APACHE-II had a high similarity (80–100%) to the reference recommendation. One guideline had a reference level of evidence of only grade 5 [7], two guidelines had only grade 4 [4, 8], and one guideline had up to grade 2b [2].

Seven guidelines [2‐8] recommend US/EUS as the reference index for the diagnosis of the cause of severe pancreatitis, of which three guidelines make explicit recommendations [2, 4, 8], and 4 guidelines only mention US/EUS in the content [3, 5‐7]. The recommendations of four guidelines [3‐5, 8] US/EUS had high similarity (80–100%) to the reference recommendation, one guideline had moderate similarity [6] (40–60%), and one guideline had low similarity [7] (20–40%). The reference levels of evidence for one set of guidelines were grades 1a, 2a, 3a, and 5. Six guidelines [2‐6, 8] recommend magnetic resonance imaging (MRI)/magnetic resonance cholangiopancreatography (MRCP) as a reference indicator for the diagnosis of severe pancreatitis, four of which make explicit recommendations [2‐6, 8], while the remaining guideline only mentions MRI/MRCP in the content [3]. The recommendations for MRI/MRCP of these six guidelines had a high similarity (80–100%) to the reference recommendation. Four guidelines provide reference levels of evidence, the highest of which was grade 2a.

Five guidelines [2‐5, 8] recommend endoscopic retrograde cholangiopancreatography (ERCP) as a reference indicator for the diagnosis of severe pancreatitis, of which three guidelines make explicit recommendations [4, 5, 8], while the remaining two guidelines only briefly mention ERCP in the content [2, 3]. The ERCP recommendations of four studies had high similarity to the reference recommendation (80–100%) [3‐5, 8]. Only two guidelines cited reference levels of evidence for ERCP, one with a grade 5 level of evidence [2], and the other with grade 1a [8].

CECT is recommended as a reference indicator for the diagnosis of severe pancreatitis in seven guidelines, four of which make explicit recommendations [4‐6, 8], and 3 guidelines only mention CECT in the content [2, 3, 7]. The recommendation of CECT in five guidelines [3, 5‐8] had high similarity to the reference recommendation (80–100%), and that of one guide [4] had a moderate similarity to the reference recommendation (40–60%). Only grade 5 evidence was cited in one guideline [2], grade 4 evidence is cited in one guideline [4], grade 2a in 1 guideline [5], and grade 3a in one guideline [6]. All seven guidelines recommend the collection of infections and fine needle aspiration (FNA) as a reference indicator for the diagnosis of severe pancreatitis after necrosis or infection, of which five guidelines make explicit recommendations [2, 4, 6‐8], while the remaining two guidelines mention this procedure in the content [3, 5]. Four guides have a high similarity to the reference recommendations of FNA (80–100%). The level of evidence cited was grade 5 in one guideline [2], grade 4 in two guidelines [3, 5], and grade 5 in two guidelines [6, 8]. Four guidelines [3, 5, 6, 8] recommend continuous abdominal pressure (IAP) > 20 mmHg for the diagnosis of abdominal syndrome (ACS) as a reference indicator for the diagnosis of severe pancreatitis, two of which make explicit recommendations [3, 6], and the other two guidelines only mention this criterion [5, 8]. Although guideline [2] was the standard literature for similarity evaluation, it did not mention the diagnosis of abdominal cavity syndrome (ACS) with continuous intra-abdominal pressure (IAP) > 20 mmHg, so no similarity scores could be obtained for this recommendation. One guideline cited a level of evidence of grade 1b [6], and one guideline cited a level of evidence of 3a [8].

Of the seven guidelines included, the similarity of the recommendations for abdominal pain, biochemical evidence, and imaging characteristics as criteria for the primary diagnosis of acute pancreatitis (Table 3) was high (80–100%) (Table 4). The main reason for the high similarity is that most guidelines [1, 2, 6, 7] recommend abdominal pain and chemical and imaging evidence as the standards for the primary diagnosis of AP, which were identical to the standards of the Atlanta classification [12]. However, there is obviously much room for improvement for the source of evidence and the standardization and rigor of the initial diagnosis. For example, although three guidelines recommend initial diagnostic criteria for acute pancreatitis, no source of evidence is provided [3, 7, 8], a clear indication of a lack of rigorousness. One of the guidelines focuses on the diagnosis of the comorbidity of severe pancreatitis [6]; however, there is a lack of recommendations for the preliminary diagnosis of acute pancreatitis. Three guidelines explicitly recommend biochemical evidence and imaging criteria for diagnosing acute pancreatitis [4, 5, 8] but lack recommendations for abdominal pain. The Japanese guidelines on biochemical examination suggest that two test strips for urinary pancreatinogen can be used for rapid diagnosis of acute pancreatitis in the future; however, the test strip is not commercialized in Japan, and therefore, it is currently not recommended for use [5].

(1) There are many recommended grades for the severity of acute pancreatitis: of the seven guidelines included, a total of 8 grades are available for reference. The Atlanta classification, C-reactive protein level, and APACHE-II score are recommended by more than 3 guidelines, while other indicators are only recommended or mentioned in 1–3 guidelines. This is related to the lack of high-quality studies comparing the pros and cons of these acute pancreatitis severity indicators. (2) The clinical outcome indicators used in the guidelines to support the evidence of recommendation are significantly different. For example, in guideline [2], the supporting evidence for CRP is a diagnostic study [13], and the researchers use the Atlanta classification (2012) as a reference to determine the accuracy of CRP in the diagnosis of severe pancreatitis. Similarly, in another guideline [4], the original literature supporting the evidence for CRP [14] is a review in which the author refers to the use of the Atlanta classification (1992) in a meta-analysis of the accuracy of CRP in the diagnosis of severe pancreatitis. The Japanese guidelines [5] recommend the JPN severity score and BISAP as the criteria for the classification grade of acute pancreatitis, and the reference is a systematic review [15]. In this review, the authors evaluate the pros and cons of these two indicators with reference to persistent multiple organ failure. The evidence supporting the APACHE-II score in guideline [3] uses mortality as a reference evaluation. Obviously, guideline users and readers cannot distinguish the pros and cons of these indicators given the heterogeneity of evidence and references. (3) The description of a standard for acute pancreatitis grading indicators is not sufficiently clear. For example, one guideline suggests that we can refer to indicators such as the APACHE-II score, CT grading, modified CT severity index, CRP, and Ranson, but no specific reference values are given, and the corresponding evidence to support them [3] is also scarce. Four guidelines recommend C-reactive protein as an indicator of the severity of acute pancreatitis, two of which provide specific reference values for CRP [4, 8], two guidelines [2, 3] do not, and two guidelines do not provide evidence to support CRP [3, 8]. One guideline [7] refers to the revised Atlanta classification, the JPN severity score, urea and APACHE-II scores, but no specific evidence is given to support them. (4) The evidence cited is not sufficiently reasonable. For example, one guideline [4] supporting the recommendation of CRP provides evidence; however, the original reference is secondary literature, and the original literature [14] is a review in which the author refers to the Atlanta classification (1992), which was used for a meta-analysis of the accuracy of CRP in the diagnosis of severe pancreatitis. The original text did not provide the heterogeneity of the meta-analysis, and the stability of the results cannot be judged. The Japanese guidelines recommend the JPN severity score and BISAP as grades for the diagnosis of acute pancreatitis. The reference for this evidence is a systematic review [15], in which the authors used continuous multiple organ failure as a reference and found that the JPN severity score and BISAP had high specificity in the diagnosis of multiple organ failure within 48 hours of admission. However, the combined BISAP data in this review were significantly heterogeneous (I² = 93.2), so the results are unstable.

Obviously, in the future, international organizations are expected to establish a global consensus on the severity classification of pancreatitis, so as to end the current confusion.

(1) The recommendations for the diagnosis of the aetiology are not comprehensive; it is important to identify the cause of acute pancreatitis for its early treatment. The main cause of acute pancreatitis is gallbladder or bile duct stones, followed by alcohol use and smoking, other rare drug use-related causes, and factors such as hypercalcemia, hypertriglyceridemia, surgery, and trauma [16, 17]. Therefore, the recommendations in the guidelines for the diagnosis of acute pancreatitis aetiology focus more on the examination of the biliary tract and gallbladder (Table 3), including US/EUS, MRI/MRCP, ERCP, and liver function tests. Only two recommendations are aimed at other rare factors, including hypertriglyceridemia and family genetic factors. The possible reason for this lack of focus may be that other rare factors that cause acute pancreatitis are relatively scarcely encountered in clinical studies, and the guideline developers did not systematically document the factors related to acute pancreatitis. (2) The evidence cited is unreasonable. For example, for the MRCP recommendations, one guideline [4] cites a high-quality meta-analysis [18], while two guidelines [3, 8] recommend MRCP to check the biliary system; however, no evidence was provided. One guideline [5] provided 3 pieces of evidence in support of MRCP, but one piece of evidence focused on endoscopic ultrasound and was not related to MRCP [18], likely a citation error. In two guidelines [4, 5], the ultrasound evidence is a retrospective case analysis, one guideline [2] uses a review, and three guidelines that recommend ultrasound or ultrasound endoscopy provide no evidence [3, 7, 8].

For acute pancreatitis comorbidities such as pancreatic necrosis, peritoneal effusion, pseudocyst, and wall necrosis (WoN), the 7 included guidelines recommend the selection of CECT for a clearer diagnosis. If the effusion or necrotic tissue is considered to be infected, all guidelines recommend fine needle aspiration (FNA) to confirm the diagnosis of bacterial culture. Although the evidence of the recommendations for the diagnosis of acute pancreatitis comorbidities lacks high-quality studies, the evidence provided by the guidelines is very different. For example, for the CECT diagnosis of pancreatic necrosis, the evidence of three guidelines [4‐6] is a retrospective case analysis, and the evidence of one guideline [2] is a review; three guidelines [3, 7, 8] recommending CECT do not provide evidence. The evidence of one guideline [2] recommending FNA is a prospective study, two guidelines [4, 5] use a retrospective observational study, one guideline [6] uses expert consensus, and one guideline [8] directly quotes another guideline. All of this evidence was published before 2014. The main reason for this evidence heterogeneity is that the guideline authors did not conduct a systematic search of evidence.

In summary, this study puts forward suggestions for improving the procedures in the diagnostic guidelines for acute pancreatitis: (1) It is recommended that in the future guidelines, the concept of diagnostic procedures should be introduced, which should be standardized to logically help guide users to diagnose acute pancreatitis more clearly; (2) greater attention should be paid to the issue of evidence citation in the diagnosis of acute pancreatitis, with a focus on a comprehensive and systematic retrieval of existing evidence and fair citations; (3) the graded indicators of acute pancreatitis should be clearly described so that readers and clinical practitioners can use the guidelines accurately; and (4) the recommendations for the diagnosis of acute pancreatitis generally lack high-quality research support, and it is recommended that research teams participating in the development of guidelines pay attention to and carry out corresponding high-quality research.