ABP is characterized by inflammation of the pancreas and biliary pathogenesis that can result in MODS and death. Considering severity, complication and prognosis, identification of different morbid states in ABP should be developed. In the current study, we have derived a simple clinical classification system for use in ABP. Using disease severity (revised 2012 Atlanta criteria), biliary obstruction, and common bile duct stones, the novel classification system was able to stratify patients within the first 48 h of hospitalization into 4 distinct types with 2 subtypes for guiding the early-definitive treatment for ABP.
Patients with moderately SAP or with SAP have distinct risks for in-hospital mortality. Therefore, as a first step, the severity of ABP (divided into Type I/II or Type III/IV) should be determined by the classification system. The mortality was 0.3% and 15.2% in Type I/II and Type III/IV in the retrospective cohort, respectively (P ˂0.001); 0% and 4.1% in the prospective cohort, respectively (P = 0.008). Since the pattern of surgery in turn depends on biliary obstruction, it is critical for introducing biliary obstruction into the classification system. Therefore, ABP was stratified as 4 distinct types by the classification system: non-obstructive mild ABP, obstructive mild ABP, non-obstructive moderately severe/severe ABP, obstructive moderately severe/severe ABP. Common bile duct stone is the main cause of biliary obstruction. In a few patients, persistent CBDS could lead to ongoing pancreatic duct or bile duct obstructions, in turn leading to SAP. Nevertheless, in some cases, biliary obstruction might be caused by postoperative strictures, inflammatory changes in the pancreas, ampulla or bile ducts and traumas8. Furthermore, CBDS does not always lead to biliary obstruction. In the retrospective cohort of 799 patients with biliary obstruction, 232 (29.0%) patients had no common bile duct stone; 65.2% patients with biliary obstruction had common bile duct stone in the prospective cohort. Therefore, it was important to figure out the association between biliary obstruction and CBDS. Subtype a or b in each type represented as CBDS or non-CBDS should be addressed.
For the convenience of the classification system in clinical application, 4 distinct types were sorted by the severity and prognosis of ABP: Type I, Type II, Type III, and Type IV. Patients with non-obstructive mild ABP or obstructive mild ABP frequently recover within about 7 days of onset of illness. However, patients with pancreatitis who have persistent biliary obstruction are recommended to undergo urgent biliary sphincterotomy and endoscopic stone extraction. Otherwise it is possible to aggravate the state of illness by biliary obstruction. Non-obstructive mild ABP and obstructive mild ABP were therefore defined as Type I and Type II, respectively. In addition, all deaths in the prospective cohort occurred in patients with non-obstructive moderately severe/severe ABP with the standard care provided according to guidelines. Obstructive moderately severe/severe ABP and non-obstructive moderately severe/severe ABP were therefore defined as Type III and Type IV, respectively. As a prognostic scoring system for use in AP, BISAP score was higher in patients with Type III/IV than that in patients with Type I/II. A few parameters that were associated with increased mortality in AP, such as albumin23, serum calcium24, C reactive protein25, and pleural effusion26 also displayed a marked difference between Type I/II and Type III/IV. Moreover, the presence of SIRS and the proportion of two or more positive SIRS criteria increased in Type III/IV. SIRS which has become increasingly widespread in clinical practice for evaluation of the systemic inflammatory response has been demonstrated to have prognostic value in AP27. This indicated the validity and practicality of the proposed ranking methods of the classification system.
The primary advantage of the classification system is simplicity. There were only three procedures to determine the type of a patient. There is no need for additional computation. In addition, the parameters in each step can be easily obtained early on in the course following hospital admission. The parameters involved in the classification system were all objective.
A strength of the current study was that the classification system was prospectively applied to the clinical practice to precisely direct clinical treatment. The complexity of ABP regarding its course, severity, prognosis, accompanying diseases in the biliary system, and the variety of available interventions have raised the need for multidisciplinary management and individualization of every case. However, there is a paucity of data as to which treatment, including cholecystectomy, endoscopic retrograde cholangiopancreatography and sphincterotomy, and what time, which is optimal, especially in severe biliary pancreatitis28. In this classification system, early laparoscopic cholecystectomy is recommended for patients with Type I; early biliary sphincterotomy and endoscopic duct clearance alone is suggested for Type II; for Type III, biliary sphincterotomy, endoscopic duct clearance, or choledochotomy is recommended. Furthermore, it is recommended that patients diagnosed with CBDS are offered stone extraction if possible. Early continuous regional arterial infusion was carried out in patients with Type III/IV. Although widespread consensus regarding the important role of cholecystectomy in the management of ABP, discrepancy between guideline recommendations for early cholecystectomy and actual rates of implementation has been reported, ranging from 32–67%28–32. In a nationwide survey across England, the overall rate of early definitive treatment was only 34.7%18. The rate of early definitive treatment increased significantly with the addition of the guidance by current classification system in patients with Type I/II, from 35.9–77.4% (P < 0.001), by which recurrence rate and associated costs can potentially be reduced33. In the current study, CRAI was recommended as an early definitive treatment for patients with Type III/IV, in which the mortality has clearly descended.
To the best of our knowledge, this was the first classification system for ABP. Physicians could assort an individual into a clinical type accurately through this easy-to-use classification system. Identifying each type of patients had an impact on the treatment or care option. Furthermore, this tool can provide information for the design of clinical studies, gaining better equivalence between study arms. Nevertheless, the data were collected at a single center. Further efforts to test the classification system in clinical application at a multicenter approach with wider geographic recruitment are encouraged.
In summary, we have developed a simple clinical classification system for use in ABP. With the guidance by the classification system, the rate of early definitive treatment in patients with ABP was improved and mortality was reduced. The classification system can strengthen decision making for each patient. Nevertheless, the classification system will have to stand the test of clinical application at a multicenter level.