Synchronous Organ Failure and Infected Pancreatic Necrosis Dene Genuine Critical Acute Pancreatitis

Background: The determinant-based classication (DBC) of acute pancreatitis (AP) was proposed in 2012. One of the highlights of the DBC was critical acute pancreatitis (CAP), which was supposed to be strongly associated with the highest risk of adverse outcomes. However, the denition of CAP needs to be further claried. Methods: A prospective cohort with consecutive patients of infected pancreatic necrosis (IPN) at a tertiary hospital was analyzed. Patients were assigned to IPN alone, Metachronous-CAP (MCAP) and Synchronous-CAP group (SCAP) according to presence or absence of organ failure (OF) and the crosstalk between OF and IPN. Clinical interventions and outcomes were compared among groups. Results: A total of 248 IPN patients were enrolled and the overall mortality was 25.8%. Compared with MCAP, patients with SCAP were associated with higher mortality (45/68, 66.2% vs. 5/50, 10.0%; OR= 17.6,95% CI, 6.2-50.4; P < 0.001) and morbidity (28/68, 41.2% vs. 9/50, 18.0%; P = 0.013), longer duration of OF (median 35.5 days vs. 12.0 days, P < 0.001), longer ICU length of stay (LOS) (median 28.0 days vs. 16.0 days, P = 0.001), longer hospital LOS (median 67.0 days vs. 60.0 days, P < 0.001) as well as earlier requirement for surgical interventions. The IPN alone and MCAP had comparable mortality (10.8% vs. 10.0%, P = 0.88), morbidity and hospital LOS, except that MCAP patients were characterized with longer duration of OF and ICU LOS (P< 0.05). Conclusions: SCAP, characterized with synchronous persistent OF and IPN, was associated with higher mortality and morbidity and should be dened as genuine CAP.


Introduction
Acute pancreatitis (AP) is a heterogeneous disease with increasing incidence [1]. The clinical course of AP ranges from uneventful and self-limiting course to that with high morbidity and mortality [2]. Owing to the advancement in understanding the pathophysiology and natural history of AP, the determinant-based classi cation (DBC) of AP was proposed by international groups of experts through a multidisciplinary consultative process in 2012 [3]. Prospective validation of DBC has been published by comparing the performance of stratifying outcomes with the Revised Atlanta Classi cation (RAC) and the Atlanta classi cation (AC) in a range of clinical settings [4][5][6][7]. One of the highlights of the DBC classi cation is a newly-de ned category called critical acute pancreatitis (CAP), which includes patients characterized by the presence of both persistent organ failure (POF) and infected pancreatic necrosis (IPN) and has been supposed to be the most ominous severity category of AP [8,9]. Prospective observation studies from different centers reported that CAP accounts for only 2.2-6.6% of AP, while the mortality was as high as 17.7-87.5% [4][5][6][7][8][9]. Although both POF and IPN occur in CAP patients, a considerable portion of cases had already recovered from early organ failure and then came to IPN in the late phase, which was named Metachronous-CAP (MCAP). The prognosis of MCAP seemed comparable with DBC "severe" category.
The rest portion of CAP was called Synchronous-CAP (SCAP), in which organ failure persisted until the pancreatic necrosis got infected. SCAP seemed to have the most fulminant course, leading to an extremely high rate of death and signi cant medical burden. Therefore, we suppose that the critical category based on DBC should be further modi ed and propose that the SCAP, not the MCAP, is the genuine "critical" AP.
Herein, we attempt to validate the hypothesis through an evidence-based approach. The present study is undertaken in a prospectively maintained database of the patients with IPN to identify the most lethal subgroup of AP at a large tertiary hospital, with a speci c focus on the clinical outcomes to determine whether CAP requires further strati cation in a tertiary referral setting.

Patient identi cation and de nitions
A total of 248 patients with IPN were prospectively and consecutively enrolled at Xiangya Hospital of Central South University (a tertiary referral center with an average of 600 admissions with AP annually) from January 2010 to December 2020. The prospectively maintained Xiangya-IPN cohort included clinical, radiological, microbiological and follow-up data, which has been introduced and published previously [8, 10,11]. Patients transferred from other hospitals were included in the cohort, medical records from other hospitals were obtained at the time of enrollment, and pertinent data elements were abstracted. The study was conducted in accordance with the principles of the Declaration of Helsinki.
Ethical approval was obtained from the Institutional Review Board of Xiangya Hospital and written informed consent was obtained from all patients or their representatives for publication of data. We followed the STROBE statement for the reporting of data [12].
The diagnosis, classi cation and de nitions associated with AP were based on the DBC [3] and American Gastroenterological Association (AGA) guideline [13], and were also described in our previous studies [8,10]. Brie y, IPN was de ned as the presence of gas bubbles within (peri)pancreatic necrosis on computed tomography or a positive culture of (peri)pancreatic necrotic uid obtained during the rst drainage or necrosectomy. The criteria for organ failure (OF) de ned for three organ systems (respiratory, cardiovascular, or renal) based on the worst measurement over a 24-h period. POF was de ned as OF in the same organ system for 48 h or more. Failure of at least two organ systems on the same day was named multiple OF. Multi-drug resistant organisms (MDROs) were de ned as microorganisms that are resistant to two or more classes of antimicrobial agents [14]. CAP was characterized by the presence of IPN and POF during the course of disease. To further clarify CAP, we separated it into Metachronous-CAP (MCAP) and Synchronous-CAP group (SCAP). MCAP de ned patients who had already recovered from early OF and then came to IPN in the late phase. SCAP described patients whose OF persisted until the pancreatic necrosis got infected. The difference in time course between the two groups was shown in Figure 1.

Standardized management protocol
All patients received standard treatment according to the latest international guidelines [1,2,13,15]. After admission, all patients were assessed by the multidisciplinary team, including pancreatic surgeons, ICU physicians, gastroenterology physicians and radiologists. Patients with organ failure were treated with organ-speci c support as needed, including mechanical ventilation, continuous renal replacement therapy, vasoactive agents, and others. Principles for the intervention of IPN in this cohort were introduced in our previous publications [8,11]. Brie y, a step-up approach consisting of percutaneous catheter drainage (PCD) or endoscopic drainage was generally the rst-line approach. If drainage did not achieve clinical improvement, subsequent minimally invasive or open pancreatic necrosectomy (OPN) would be performed to control the infection.

Clinical outcomes
Primary outcomes included mortality, major morbidity (hemorrhage, intestinal leakage and pancreatic stula), intensive care unit (ICU) admission, ICU length of stay (LOS), duration of OF, overall hospital LOS, time from onset to rst intervention (percutaneous, endoscopic and surgical drainage/necrosectomy) and times needed for interventions. These outcomes have been previously reported in validation studies to classify AP severity [4,6,9,16]. In this study, inter-Groups were compared for their ability to stratify severity in accordance with clinical outcomes.

Statistical analysis
Categorical variables were described using frequencies and percentages. Continuous variables were summarized using median and interquartile range depending on the distribution. Chi-square test or Fisher exact test was used to compare the distribution of categorical variables, and Student t-test or Mann-Whitney U test for continuous variables. All tests were bilateral, and P-values < 0.05 were considered statistically signi cant. The SPSS (22.0) was used for all analyses.

Patient characteristics
A total of 248 patients with IPN were enrolled in the analysis (74.2% males; median age 48 years). The most common etiology was biliary (n = 107, 43.1%), followed by hyperlipidemia (n = 102, 41.1%), others (n = 33, 13.4%), and alcoholic (n = 6, 2.4%). According to DBC categories, 52.4% (n = 130) of patients were classi ed as severe (IPN without OF), and 47.6% (n = 118) as critical category. The median duration of OF, ICU LOS, and hospital LOS were 17 days, 13 days and 57.5 days, respectively, and overall mortality was 25.8% (64/248). Demographics, etiologies, local, major complications, interventions and clinical outcomes are summarized in Table 1. Step-up surgical approach 199 (80.2%) Step  10.0%), morbidity and hospital LOS. However, compared with IPN alone cases, MCAP patients did have a longer duration of OF and ICU LOS, and they also presented more frequently with multiple OF, bloodstream infection and MDROs infection (P < 0.05), as shown in Table 2.

Discussion
This is, to our knowledge, the largest study of a prospective cohort of IPN to speci cally evaluate the clinical outcomes of CAP so far. We found that SCAP, characterized with synchronous persistent OF and IPN, had signi cantly higher mortality and morbidity than MCAP. These results justi ed a further classi cation of CAP. For the rst time, we put forward two subgroups of CAP, ie. SCAP and MCAP.
According to the results, SCAP de nes genuine CAP. These results gave us an update to the current knowledge about strati cations and classi cations of AP.
The Atlanta classi cation (AC) reported in 1992 was the most widely accepted classi cation system in the past for the severity in AP, which divided AP into two groups: mild and severe [17]. However, due to substantial progress in understanding the pathophysiological and disease-related complications, most experts realized the necessity for revising the above classi cation system. Thus in 2012, the revised Atlanta classi cation (RAC) [18] and a novel DBC system were proposed [3]. The development of the DBC involved three steps. First, Petrov and Windsor performed a comprehensive review of available evidence and proposed the new classi cation [19]. The second phase consisted of a global Web-based survey of pancreatologists with recent clinical publications regarding AP [3]. Finally, an international consensus symposium was held during the 2011 International Association of Pancreatology meeting to further discuss the proposed classi cation [3]. The DBC was based on the local determinant (pancreatic necrosis) and the systemic determinant (organ failure) which were causally associated with clinical outcomes of AP. As both RAC and DBC systems were adopted, which inevitably raised issues about their validity and practicability. For the past decades, we have witnessed a tremendous effort on the independent external validation of RAC and DBC. By comparing the performance characteristics of each classi cation system in predicting important clinical outcomes, both RAC and DBC were found to be comparable and superior to AC [4-6, 16, 20]. Although there were still some controversies, DBC has become one of the internationally recognized criteria for AP severity classi cation, especially in tertiary medical centers. Hence, we focused on DBC classi cation in this study, especially the most severe category named CAP [3]. The Auckland group [21] performed a meta-analysis of 14 studies comprising 1478 patients demonstrating that both OF and IPN were equivalent determinants of severity with a mortality rate of 30% and 32%, respectively. However, taking them together during the course of AP was associated with even higher mortality (43%). In the multivariate analysis of a Korean study [16], POF was thought of as a continuum, and IPN served as a preceding event pertaining to mortality of late POF, thus concluding that both POF and IPN were independent risk factors for mortality of AP and should be recognized as actual determinants of severity. Among the Spanish [7], Chinese [9], British [6] and Indian [5] cohorts, the mortality of CAP was as high as 54.1%, 57.1%, 87.5% and 87.5%. Also, the death rate of our previous CAP cohort was 31.4% [8]. In that study, 102 CAP patients were enrolled to speci cally evaluate the prognostic factors associated with CAP, especially from three aspects, including organ failure, infection, and surgical interventions. Therefore, identifying patients classi ed as "critical" has de nite implications for prognosis and management choices. However, some controversies still exist about whether the critical category makes sense because of its rarity and whether IPN should be recognized as a key determinant of severity. In a Spanish cohort, only 3 of 543 patients (0.6%) were classi ed as CAP, and mortality among these patients was similar to the severe category. A nationwide multicenter prospective cohort study from the Spain group comprising 1655 AP patients showed that among patients with POF, the presence of infection was solely associated with increased morbidity, but not mortality, which marked POF as the most signi cant determinant of severity [7]. More recent studies concluded that the in uence of OF on mortality was stronger than that of IPN [22,23], and in Scheper's study [23], even no association was found between the duration of organ failure and mortality. It was not surprising that these divergences existed because all the studies mentioned above were based on different settings (community hospital or tertiary referral center) and purposes (clinical management or research). In other words, if the results in these studies all held true, we might note that none of the existing classi cations were based on uniform study designs and outcomes. A Spanish prospective multicenter study [24] proposed a modi ed DBC (MDBC) in ICU patients, where the severe category of DBC was separated into two groups: patients with IPN and transient OF and patients with POF but no IPN, and the two distinct subgroups marked with signi cantly different rates of intervention, morbidity and mortality. Recently, Wu et al. [9] provided further validation for the MDBC proposal to subdivide the DBC "severe" category into two groups in all hospital admissions. Viewing this, during our clinical practice, we noticed that the "critical" category of DBC might also be separated into two groups: the MCAP and SCAP, which might show a signi cant difference in clinical outcomes.
One of the most important purposes of AP reclassi cation was to focus on the population with the highest mortality. There was a considerable portion of DBC "critical" cases whose clinical courses were not that bad and might belong to the DBC "severe" category. This provided evidence that CAP could be further modi ed. Due to the limited number of cases with CAP in previous studies [4][5][6]9], it was not easy to analyze the outcomes of CAP, let alone to rede ned it. The present study made full use of a large cohort of CAP and found a signi cant increase in adverse effects in patients with synchronous persistent OF and IPN compared with MCAP. Thus, theoretically, SCAP should de ne genuine "critical" AP. It was acknowledged that mortality rates in severe pancreatitis were believed to have two peaks, that was, following "early" organ failure and following "late" organ failure due to secondary infection of pancreatic or peripancreatic necrosis [1,2]. It was also generally considered that IPN developed over time, peaking at approximately three weeks or later after disease onset [1,2]. With regard to SCAP in this study, the median time from onset to the diagnosis of IPN was only 16 days, which showed early intervention was preferable in this setting. In 2020, the Chinese Acute Pancreatitis Clinical Trials Group (CAPCTG) designed a multi-center randomized controlled trial protocol to compare the clinical outcomes of early on-demand drainage (EOD) with the current standard approach among acute necrotizing pancreatitis (ANP) patients complicated by POF [25]. Recently, the preliminary result showed that the EOD approach might have potential clinical bene ts [26], which was similar to our conclusion.
Another strength of the present study was that MCAP patients did not have higher mortality compared with those with IPN alone, and this was consistent with a recent multicenter study [23], which found early persistent OF was not associated with increased mortality in patients with necrotizing pancreatitis. The pathophysiology of OF may differ according to whether OF occurs early or late in the disease [22]. In the early phase, the release of in ammatory chemokines and cytokines during the systemic in ammatory response syndrome (SIRS) plays an important role in the development of OF. Fortunately, with the improvement of ICU treatment modality, a majority of severe patients could come through the early OF period. Frequently after several weeks, infection of pancreatic necrosis may be accompanied by sepsis and leads to the late OF phase. Choi et al. [16] demonstrated that patients with late persistent OF were associated with higher mortality and morbidity than those with early persistent OF. IPN-induced late OF rather than SIRS-induced early OF is the key determinant of prognosis. In other words, MCAP might be regarded as "severe" AP according to DBC, just as IPN alone.
There are still several limitations in this study. One of the limitations is that this cohort was conducted at a tertiary care center, 43% of enrolled patients were transferred from other institutions. This could have resulted in selection bias and potential confounding effects related to heterogeneity in initial treatment during the critical early phase of AP, and the results of the current study are less relevant to community hospital settings. Another limitation is that the prospectively maintained cohort all consist of IPN patients, thus, we can't evaluate the pooled in uence of OF on mortality regardless of the presence or absence of IPN, which is also one of the key determinants of DBC [3]. Lastly, the results of the present study, which were derived from a single center, could not be generalized to other hospitals indiscriminately. Further multicenter, prospective original studies would provide more precise data to reduce potential confounding results.

Conclusions
Our present ndings suggested that SCAP characterized with synchronous POF and IPN, had the most fulminant course, leading to an extremely high mortality and morbidity and should be de ned as genuine CAP. While MCAP de ned patients who had already recovered from early OF and then came to IPN in the late phase, had comparably favorable prognosis compared with IPN alone cases.

Declarations
This study was supported nancially by National Natural Science Foundation of China (8180103017) and the Natural Science Foundation of Hunan Province (2020JJ4133).

Availability of data and materials
The datasets used during the current study are available from the corresponding author on reasonable request.
Ethics approval and consent to participate The study was conducted in accordance with the principles of the Declaration of Helsinki. Ethical approval was obtained from the Institutional Review Board of Xiangya Hospital and written informed consent was obtained from all patients or their representatives for publication of data.

Consent for publication
Not applicable Figure 1 The difference between MCAP and SCAP in time course of organ failure and infected pancreatic necrosis.