Multiple Hits in Acute Pancreatitis: Components of Metabolic Syndrome Synergize Each Other’s Deteriorating Effects

Introduction The incidence of acute pancreatitis (AP) and the prevalence of metabolic syndrome (MetS) are growing worldwide. Several studies have confirmed that obesity (OB), hyperlipidemia (HL), or diabetes mellitus (DM) can increase severity, mortality, and complications in AP. However, there is no comprehensive information on the independent or joint effect of MetS components on the outcome of AP. Our aims were (1) to understand whether the components of MetS have an independent effect on the outcome of AP and (2) to examine the joint effect of their combinations. Methods From 2012 to 2017, 1435 AP cases from 28 centers were included in the prospective AP Registry. Patient groups were formed retrospectively based on the presence of OB, HL, DM, and hypertension (HT). The primary endpoints were mortality, severity, complications of AP, and length of hospital stay. Odds ratio (OR) with 95% confidence intervals (CIs) were calculated. Results 1257 patients (55.7 ± 17.0 years) were included in the analysis. The presence of OB was an independent predictive factor for renal failure [OR: 2.98 (CI: 1.33–6.66)] and obese patients spent a longer time in hospital compared to non-obese patients (12.1 vs. 10.4 days, p = 0.008). HT increased the risk of severe AP [OR: 3.41 (CI: 1.39–8.37)], renal failure [OR: 7.46 (CI: 1.61–34.49)], and the length of hospitalization (11.8 vs. 10.5 days, p = 0.020). HL increased the risk of local complications [OR: 1.51 (CI: 1.10–2.07)], renal failure [OR: 6.4 (CI: 1.93–21.17)], and the incidence of newly diagnosed DM [OR: 2.55 (CI: 1.26–5.19)]. No relation was found between the presence of DM and the outcome of AP. 906 cases (mean age ± SD: 56.9 ± 16.7 years) had data on all four components of MetS available. The presence of two, three, or four MetS factors increased the incidence of an unfavorable outcome compared to patients with no MetS factors. Conclusion OB, HT, and HL are independent risk factors for a number of complications. HT is an independent risk factor for severity as well. Components of MetS strongly synergize each other’s detrimental effect. It is important to search for and follow up on the components of MetS in AP.

Introduction: The incidence of acute pancreatitis (AP) and the prevalence of metabolic syndrome (MetS) are growing worldwide. Several studies have confirmed that obesity (OB), hyperlipidemia (HL), or diabetes mellitus (DM) can increase severity, mortality, and complications in AP. However, there is no comprehensive information on the independent or joint effect of MetS components on the outcome of AP. Our aims were (1) to understand whether the components of MetS have an independent effect on the outcome of AP and (2) to examine the joint effect of their combinations.
Methods: From 2012 to 2017, 1435 AP cases from 28 centers were included in the prospective AP Registry. Patient groups were formed retrospectively based on the presence of OB, HL, DM, and hypertension (HT). The primary endpoints were mortality, severity, complications of AP, and length of hospital stay. Odds ratio (OR) with 95% confidence intervals (CIs) were calculated.
Results: 1257 patients (55.7 ± 17.0 years) were included in the analysis. The presence of OB was an independent predictive factor for renal failure [OR: 2.98 (CI: 1.33-6.66)] and obese patients spent a longer time in hospital compared to non-obese patients (12.1 vs. 10.4 days, p = 0.008). HT increased the risk of severe AP [OR: 3.41 (CI: 1.39-8.37)], renal failure [OR: 7.46 (CI: 1.61-34.49)], and the length of hospitalization INTRODUCTION Acute pancreatitis is a severe inflammatory condition with increasing incidence and hospitalization worldwide (Forsmark et al., 2016;Garg et al., 2019). AP has a variable severity ranging from mild and self-limited to severe and fatal. The mortality of the disease ranges approximately from 2 to 5% and depends on the development of organ failure and local complications, which are summarized in the revised Atlanta classification (Banks et al., 2013). The major etiological factors are gallstones and alcohol consumption (Forsmark et al., 2016), but hypertriglyceridemia (HTG) and intake of certain medications may also be in the background.
The severity and outcome of AP are influenced by the metabolic comorbidities of the host (Working Group Iap/Apa Acute Pancreatitis Guidelines, 2013; Goodger et al., 2016). Metabolic syndrome is characterized by the clustering of abdominal OB, HTG, low levels of high-density lipoprotein (HDL), elevations in blood pressure and fasting glucose, or diabetes (Alberti et al., 2009). MetS is associated with an increased risk of development of and death from cardiovascular disease and chronic kidney disease (Isomaa et al., 2001). The presence of MetS was previously shown to be associated with a higher risk of severe AP, higher mortality rate, and longer duration of stay in the intensive care unit (Mikolasevic et al., 2016). However, in another study, MetS did not affect the severity of AP (Sawalhi et al., 2014). OB was previously shown to be independently associated with the severity of AP (Sawalhi et al., 2014) and the development of organ failure but not with mortality in AP (Smeets et al., 2019). DM was associated with a higher risk of AP (Yang et al., 2013) and negatively influenced the outcome of AP by raising the incidence of renal failure, intensive care unit admission, and length of hospital stay (LOS) (Miko et al., 2018). The presence of HTG increased severity, complication rate, and mortality in AP (Kiss et al., 2018).
However, there is no data regarding a link between the outcome of AP and the presence of arterial HT. Furthermore, there is a lack of data on how the components of MetS, namely, OB, DM, HT, and HL, influence the outcome of AP individually or in combination. Therefore, in this study, we aimed to analyze how the components of MetS influence the outcome of AP (1) individually and (2) in combination.

Patient Population and Study Design
The APR launched in 2011 by the Hungarian Pancreatic Study Group is an international prospective registry for patients suffering from AP. Besides pancreatic registries, HPSG has already organized five registered clinical trials to investigate AP with the acronyms PREPAST (Dubravcsik et al., 2015), APPLE , PINEAPPLE , GOULASH (Marta et al., 2017), and EASY (Hritz and Hegyi, 2015) and has submitted three further pre-study protocols: GOULASH PLUS (follow-up to the GOULASH study), EMILY (endoscopic sphincterotomy for delaying cholecystectomy in mild acute biliary pancreatitis), and LIFESPAN (lifestyle, prevention, and risk of AP).
From June 2012 to September 2017, 1435 adult patients with AP from 28 community and university hospitals were prospectively enrolled (Supplementary Appendix S1). Demographic and anthropometric data; history of HL, HT, and DM; previous medical therapy and etiology; severity; local and systemic complications; and mortality of AP were collected.
In this study, we aimed to maximize the number of cases for each individual effect analysis. We had information concerning OB from 1257 cases, HT from 1127 cases, DM from 1257 cases, and HL from 1036 cases. Patients were grouped based on the World Health Organization (WHO) classification of BMI (≥30 or <30 kg/m 2 ) and the presence or absence of three other components, HT, HL, and DM. However, in the "joint effect analysis, " we only included cases where data from all four components of MetS, OB, HL, HT, and DM were available (906 cases). We conducted an additional analysis to confirm that the cohorts noted above represent the total cohort of 1435 cases. Importantly, there were no significant differences in demographics or the main outcome parameters between the cohorts (Supplementary Appendix S2).
Data were collected by treating physicians with the help of trained and experienced study administrators on the basis of a standardized case report form and protocol in the prospective APR. Accuracy of data recorded is secured by a four-level quality check system involving both medical administrative personnel and gastroenterologists. Data quality is presented in Supplementary Appendix S3. The study protocol was approved by the Scientific and Research Ethics Committee of the Medical Research Council (22254-1/2012/EKU). All patients provided written informed consent to participate in the registry.

Definitions
Diagnosis of AP was made according to the recommendations in the IAP/APA guidelines. At least two criteria of the following three were present: upper abdominal pain, pancreatic enzyme levels exceeding more than three times the upper normal level, and features of pancreatitis on imaging (Working Group Iap/Apa Acute Pancreatitis Guidelines, 2013). Severity and complications of AP were determined according to the revised Atlanta classification (Banks et al., 2013). OB was determined if BMI was ≥30 kg/m 2 (Jensen et al., 2014). HT was determined if blood pressure was ≥140/90 mmHg or if the patient was on anti-hypertensive medication. HL was defined by the presence of either hypercholesterolemia or a low level of HDL or HTG. The condition was regarded as HL when fasting cholesterol level >200 mg/dL (5.2 mmol/L), HDL < 44 mg/dL (1.15 mmol/L; female) or <35 mg/dL (0.9 mmol/L; male), triglyceride level exceeded 150 mg/dL (1.7 mmol/L), or the patient was receiving drug therapy for HL. The diagnosis of DM was made in accordance with the American Diabetes Association Criteria (American Diabetes Association, 2010) or if the patient was receiving drug therapy for hyperglycemia.
The primary endpoints were mortality, severity, and complications of AP and LOS.

Statistical Analyses
Case numbers and percentages were calculated for categorical variables, mean with SD, and medians with 25 and 75% quartiles (Q1 and Q3, respectively) and ranges were computed for numerical variables in descriptive analysis.
The t-test was used for normally distributed data and the Mann-Whitney U-test for non-normally distributed data to compare two groups of independent samples. The relation between categorical variables was inspected by the Chi-square test and Z-test with the Bonferroni correction and ORs with 95% CIs.
Logistic regression was used to define the independent effect of the MetS factors and age. A two-sided p-value of <0.05 was regarded as statistically significant. The available-case analysis was used for missing data. Statistical analyses were performed with SPSS 25.0 software (IBM Corporation).
The major etiologies of AP were biliary stones in 37.8% of the cases of the total cohort, alcohol in 18.5%, and HL in 3.7%. OB increased the risk of biliary etiology [OR: 2.06 (CI: 1.61-2.64)]. Meanwhile, HTG-induced AP was more frequent in the presence of HL (12.9 vs. 0.1%, p < 0.001) compared to the non-HL group and in the presence of DM compared to the non-DM patient group [OR: 2.34 (CI: 1.39-4.00)], respectively ( Table 1)

Independent effect
Logistic regression revealed that HT was a predictive factor for severity [OR: 3.41
For patients with HL, the chance of having mild AP was lower [OR: 0.64 (CI: 0.49-0.85)], but HL had no significant effect

Independent effect
Logistic regression revealed that HL was an independent predictive factor for local complications   one by one (Supplementary Appendix S5). The presence of two, three, or four MetS factors significantly increased the rate of worse outcome parameters by 9.5, 24.1, and 66.7%, respectively (Figure 5). Older age was demonstrated to be independently associated with respiratory and heart failure in our study. Statistically significant values (ORs with CIs) are marked in bold digits.

Summary of Findings
Our results demonstrated in a large database of prospectively collected cases that the components of MetS deteriorate the outcome of AP. OB was shown to be an independent risk factor for renal failure and was associated with a longer hospital stay. HT was proved to be an independent risk factor for severity of AP and increased the risk of renal failure, while patients with HT spent a longer time in hospital. HL increased the risk of local complications, renal failure, and the new diagnosis of DM. Preexisting DM did not change the outcome of AP. Our study demonstrated that the more components of MetS the patients had, the higher the rate of worse outcome parameters was observed. The incidence of AP is increasing, and this is partly due to the rising prevalence of OB, which stimulates gallstone formation and increases HL, both causing AP (Yadav and Lowenfels, 2013;Bonfrate et al., 2014). Indeed, biliary AP was more frequent in obese patients compared to the total cohort in our study.
To date, several cohort studies and a systematic review have reported that OB increases the severity, mortality, and occurrence of local and systemic complications in AP. However, these results are conflicting on the link between OB and outcomes in AP (Dobszai et al., 2019). The reason behind this conflict may be that most of the included studies reported unadjusted analysis; therefore, it cannot be clarified whether OB is an independent prognostic factor in AP or not (Dobszai et al., 2019). In a recent individual patient data meta-analysis, where confounders were adjusted, OB was independently associated with the development of organ failure and multiple organ failure in AP; however, there was no relation between OB and mortality, necrosis, and intervention (Premkumar et al., 2015). These data are in agreement with our results, where OB was demonstrated to be an independent predictive factor for renal failure but did not modify the mortality rate ( Table 2).
A possible mechanism by which OB is associated with a higher risk of renal failure is lipotoxicity.  Obesity is associated with elevated levels of intrapancreatic fat and with elevated visceral fat surrounding the pancreas (Smeets et al., 2019). This hypothesis is also supported by experimental data. A long-term high-fat diet caused acinar cell injury and pancreatic fibrosis via fat accumulation in pancreatic acinar cells (Matsuda et al., 2014). It has also been suggested that intrapancreatic fat, which may cause metabolic and inflammatory processes, is associated with OB (Majumder et al., 2017). In addition, in the presence of intrapancreatic fat, pancreatic lipases are released in AP digest adipocytes, resulting in an outflow of unsaturated fatty acids into the circulation; they are toxic and can act as proinflammatory mediators and are implicated in the development of systemic inflammation and organ failure (Navina et al., 2011).
Hypertension was independently associated with the severity of AP and the rate of renal failure in our study. To the best of our knowledge, no study has ever analyzed the effect of arterial HT on the outcome of AP. The underlying mechanisms by which HT deteriorates the outcome of AP is unclear. It has been suggested that the sympathetic nervous system may act as an amplifier of the blood pressure elevation and may be involved in the development of HT-related complications. Sympathetic activation favors the development and progression of vascular hypertrophy and remodeling and contributes to impairing arterial distensibility and vascular compliance (Seravalle et al., 2014). The presence of a hyperadrenergic state and microvascular and macrovascular structural changes in the arteries may be responsible for the deteriorative effects of HT (Smits and van Geenen, 2011).
Preexisting HL was shown to be independently associated with local complications and renal failure in our study. Our results are in line with those of a recent meta-analysis, which reported that the presence of HTG significantly elevated the risk of renal failure but did not increase the risk of mortality in AP (Kiss et al., 2018). However, HTG also significantly elevated the risk of severe AP in this meta-analysis (Kiss et al., 2018), while HL did not increase the risk of severe AP in our study. This discrepancy can be explained by the fact that (1) most of the studies included in the meta-analysis reported an unadjusted analysis, and, therefore, the independent effect of HTG in AP cannot be elucidated; and (2) the HL group in our study included patients with either hypercholesterinemia and/or HTG, while patients with HTG only were included in the meta-analysis. One possible mechanism by which HL increases local and systemic complications in AP is the formation and toxic effect of unsaturated fatty acid by pancreatic lipases. In addition, in the case of HTG, the chylomicron concentration is elevated. As a result, blood viscosity increases, thus impairing blood flow and causing pancreatic ischemia and acidosis (Pedersen et al., 2016).
There is a special relationship between the exocrine and endocrine pancreas. Experimental data suggest that insulin has a local protective effect on acinar cells during pancreatitis. Pancreatitis evoked by L-arginine causes severe acinar cell necrosis in most of the territory of the exocrine pancreas. However, acinar cells located around the islets of Langerhans remain totally intact (Hegyi et al., 1997). In addition, we also confirmed that if the beta cells are destroyed by streptozotocin treatment prior to the induction of AP, this locally visible protective effect disappears irrespectively of exogenous insulin administration (Takacs et al., 2001). Unfortunately, in our registry analysis, we could not investigate the local effects of insulin. Here we showed that preexisting DM does not significantly influence severity, mortality, or rate of complications in AP in our cohort. We hypothesized that our cohort was not sufficiently large to determine a significant difference. We have recently published a meta-analysis in which DM significantly elevated both local and systemic complications when an analysis was conducted of 354,880 cases (Miko et al., 2018). However, it is clearly impossible to collect this number of patients in a single cohort. Furthermore, intensive care unit mortality only grew significantly with higher mean blood glucose concentration in non-DM patients but not in DM patients (Egi et al., 2008;Pedersen et al., 2016). In agreement with our results, critically ill patients with DM did not have higher mortality compared to non-DM patients (Whitcomb et al., 2005).
Older age was demonstrated to be independently associated with pulmonary and heart failure in our study (Table 2B). Older age has been investigated extensively as a marker of severity and mortality in AP and is included in the APACHE II score, Ranson score, Bedside Index of Severity in AP (BISAP) score, and Japanese Severity Score (JSS) as a marker of severity (Graham et al., 2010). However, after adjusting for comorbid disease, only the very extreme age (>85 years old) was associated with 30day in-patient mortality and persistent organ failure in a recent prospective, multicenter study (Mounzer et al., 2012). Our results are in line with a recent cohort analysis that found that elderly patients had a significantly higher risk of developing systemic complications, while high mortality in this group is due to the effect of severe comorbidities (Szakacs et al., 2018).
Patients with AP often develop diabetes during and after the attack of AP (Moran et al., 2018); however, the risk of DM was not fully evaluated. The severity of AP, its etiology, and individuals' age and sex had a minimal effect on the development of newly diagnosed diabetes in AP (Moran et al., 2018). We showed that HL is an independent risk factor for the development of newly diagnosed DM in AP. High cholesterol and triglyceride levels increase the risk of DM, a finding supported by earlier studies (von Eckardstein and Sibler, 2011;Das et al., 2014). We can hypothesize that the predisposition to DM caused by dyslipidemia was manifested during AP. This finding emphasizes the need for a thorough screening for DM in AP patients with HL. Moreover, all AP patients should be followed and screened for DM as hyperglycemia stimulates the proliferation of pancreatic stellate cells and collagen secretion, while hypoinsulinemia inhibits acinar cell growth and synthesis of pancreatic enzymes and therefore facilitates fibrosis of the pancreas and might cause chronic pancreatitis (Czako et al., 2009).

Strengths and Limitations
The main strength of the present study is that it has a large sample size of prospectively collected cases from hospitals in multiple countries, including tertiary and non-tertiary centers. Furthermore, a logistic regression analysis was applied to control confounding variables, and the independent prognostic factors of the components of MetS were analyzed for AP. Finally, our study is the first to report the relation between the outcome of AP and the presence of arterial HT and to analyze the influence of the combined presence of the components of MetS on the outcome of AP.
The present study has limitations. First, since APR is a multicenter prospective registry and not an observational trial, our findings are affected by confounding factors or selection bias. Second, our study design is cross-sectional, thus precluding any causal interferences about the directionality of the relations observed in our study; therefore, long-term clinical outcomes could not be evaluated. Accordingly, long-term prospective trials are needed in the future. Third, our study assessed the effect of HL, not HTG, thus not fully suiting the definition of MetS. Fourth, peripancreatic fluid accumulations could not always be adequately defined according to the modified Atlanta classification. Acute fluid collection and acute necrotic fluid collection, pseudocysts, and walled-off pancreatic necrosis could not always be differentiated because abdominal CT was not performed in all cases. Therefore, peripancreatic fluid collections without a definitive wall were named as acute fluid collections and with a wall as pseudocysts.

CONCLUSION
In conclusion, the components of MetS deteriorate the outcome of AP. OB, HT, and HL are independent risk factors for a number of complications. HT is an independent risk factor for severity as well. The more elements of MetS are present, the higher the risk for complications. It is important to search for and follow up on the components of MetS in AP.

DATA AVAILABILITY
All datasets generated for this study are included in the manuscript and/or the Supplementary Files.

ETHICS STATEMENT
The study protocol was approved by the Scientific and Research Ethics Committee of the Medical Research Council (22254-1/2012/EKU). All patients provided written informed consent to participate in the study.

AUTHOR CONTRIBUTIONS
AS, AP, and PH contributed to the design of the research. AP, ÁV, JB, SG, PS, FI, AH, IT, HF, MP, MV, JH, JN, AM, EM, VS, LC, and TT collected the data. AP, KM, DD, DI, and BK assessed the data quality. NG and AS processed the data and conducted the analysis. AS and PH designed the figures. AS, LC, and BE drafted the manuscript. PH supervised and coordinated the work. All the authors discussed the results and commented on the manuscript.