Serum apolipoprotein B-to-apolipoprotein A1 ratio is independently associated with disease severity in patients with acute pancreatitis

Early identification of severe acute pancreatitis (SAP) is critical for clinical decision-making. The apolipoprotein B-to-apolipoprotein A1 ratio (ApoB/A1 ratio) reflects the balance between pro-inflammation and anti-inflammation in vivo. This study investigated the association between serum ApoB/A1 ratio at admission and acute pancreatitis (AP) severity. A total of 375 patients with first attack of AP were retrospectively recruited from January 2014 to December 2017. The severity of AP was assessed at admission based on the 2012 revised Atlanta Classification. Serum lipids levels were tested on the first 24 h of hospitalization, of which the correlations with clinical features or scoring systems were also measured. The ApoB/A1 ratio markedly increased across disease severity of AP. The ApoB/A1 ratio, expressed as both quartile and continuous variables, was significantly associated with a high risk of SAP, even after adjustment for other conventional SAP risk factors. The ApoB/A1 ratio positively correlated with the revised 2012 Atlanta Classification, Ranson score, Bedside Index for Severity in AP score, Modified Computed Tomography Severity Index score, and Acute Physiology and Chronic Health Evaluation II score for AP severity. The optimal cut-off value of ApoB/A1 ratio for detecting SAP was 0.88, with a sensitivity of 83.08% and a specificity of 69.03%. Serum ApoB/A1 ratio at admission is closely correlated with disease severity in patients with AP and can serve as a reliable indicator for SAP in clinical setting.

Association between serum lipid parameters and AP severity. The association between serum lipid parameters and disease severity of AP was also demonstrated in Table 1 and Fig. 1. The levels of TC and LDL-C showed no difference among the AP patients with different levels of disease severity based on the 2012 revised Atlanta Classification. The level of ApoB/A1 ratio markedly increased across disease severity. The serum levels of TG, HDL-C, ApoA1, and ApoB were significantly higher in the patients with SAP than in those with MAP. The level of ApoA1 in the SAP group was much higher than that in the MSAP group. Simultaneously, the MSAP group had higher levels of TG and ApoB compared with the MAP group.

Discussion
Due to the clinical features of AP, potential patients with SAP should be recognized to ensure that they receive intensive care and adequate treatment in a timely manner. Earlier efforts to classify the AP severity resulted in the original 1992 Atlanta Classification, which had identified two subgroups of AP, "mild" and "severe", and recommended the clinical treatment for each category 18 . However, emerging evidence has revealed that most AP patients   www.nature.com/scientificreports www.nature.com/scientificreports/ who fell in between the two categories based on the 1992 Severity Classification often suffered relatively good outcomes and positive response to less aggressive treatment than those with severe disease, thus necessitating the substantial revision of this classification system 19 . In 2012, a revised version of the Atlanta Classification has been established by adding a third category defined as "moderately severe" 2 . According to a single-center research in Chinese, 649 patients with "severe" disease defined by the 1992 Atlanta Classification were included, and subsequently regrouped as "MSAP" and "SAP" according to the revised 2012 Atlanta Classification 20 . When compared with the MSAP group, the Ranson score, APACHEII score, BISAP score, MCTSI score, Marshall score, incidence of organ failure, average length of stay, and hospital mortality of the SAP group significantly increased, which indicated that the revised 2012 Atlanta Classification could more precisely reflect AP severity. Therefore, our study explored the value of ApoB/A1 ratio in predicting AP severity based on the revised 2012 Atlanta Classification.
The present study found a close association between high ApoB/A1 ratio and the worsening of AP severity, regardless of conventional risk factor. The mechanisms underlying the association between ApoB/A1 ratio and AP severity have been not yet completely understood and are probably multi-factorial. Systematic inflammatory response syndrome (SIRS) and organ dysfunction are the main features in the early phase of AP, whereas pancreatic necrosis and infection in the late stage 21 . In the traditional view, the main cause of death in AP patients is the   www.nature.com/scientificreports www.nature.com/scientificreports/ complications in the late stage. However, a recent growing number of studies have confirmed that early-phase SIRSs are the main reasons for organ failure and mortality in AP patients 22 . Thus, the balance between pro-inflammatory and anti-inflammatory status is the key point of severity prediction and clinical therapy 23 . ApoB facilitates, whereas ApoA1 suppresses systemic inflammatory states. Hence, an increased ApoB/A1 ratio might reflect a predominance of pro-inflammatory effects over anti-inflammatory of lipoprotein lipids, thus contributing to the progress of inflammation and AP severity. Huh 24 also found that ApoB/A1 ratio is strongly predictive of SAP independent  www.nature.com/scientificreports www.nature.com/scientificreports/ of the etiology of AP, of which the optimal cut-off value was 1.16 with 53% sensitivity and 93% specificity. Notable, the optimal cut-off value in our research was 0.88 with a relatively high sensitivity of 83.08% and a relatively low specificity of 69.03%, which differed considerably from the previous study. Given that SAP is deeply harmful to the human body and needs prompt treatment, false negative results should be reduced and misdiagnose of SAP should be avoided 25 . Therefore, we applied a low cutoff value to achieve a high sensitivity.
Another major finding of our study is that the ApoB/A1 ratio was the strongest indicator for SAP compared with the other lipid parameters. Several studies have showed that lipid profiles, including HDL-C, and LDL-C at admission, could predict the development of severe disease and organ failure in AP patients 26,27 . However, these traditional lipid indexes are not always adequate indicators of dyslipidemia due to the wide alternation of measurement, whereas the lipid-transporting apolipoprotein is relatively more stable. First, each lipid particle, including LDL, VLDL, and IDL, contains only one unit of ApoB. Considering that atherogenic dyslipidemia promotes inflammatory response and causes oxidative stress in pancreas, the predictive value is more related to the number of atherogenic particles but not LDL-C alone, which only measures the amount of cholesterol in LDL particles 28 . Thus, the presence of a few but large particles is more favorable than the presence of a large number of small particlesat a given LDL-C concentration 29 . Second, the removal processes of ApoA1 and HDL-C are prone to be affected by plasma transferrin and transferase 30 . During the alteration of lipid metabolism, the change in HDL-C was more significant than that in ApoA1. Three large cohort studies also exhibited a consistent outcome that both ApoB and ApoA1 had independent and equal predictive values, and ApoB/A1 ratio was the strongest and most specific indicator for cardiovascular disease that was superior to the cholesterol ratios [31][32][33] . Therefore, ApoB/A1 ratio demonstrates better superiority over the cholesterol ratios in terms of predictive ability. However, these findings were mainly based on research concerning cardiovascular diseases. Medical evidence for ApoB/ A1 ratio in the prediction of AP severity is insufficient. Thus, additional studies are warranted to further confirm the advantage of this parameter.
To date, several scoring systems including Ranson score, BISAP score, MCTSI score, and APACHE II score, have been widely applied to predict AP severity at an early stage. The scoring systems have played a valuable role in the early prediction of AP severity. In our study, the ApoB/A1 ratio was significantly positively correlated with the above-mentioned scoring systems rather than other lipid indexes, which indicated the preponderance of ApoB/A1 ratio on predicting AP severity from another perspective. However, abundant evidence has proven that these scoring systems are unsuitable in clinical practice because of their complex and time-consuming applications 34 . For example, Ranson score should be obtained within 48 h after admission, and some variables, such as residual alkali and fluid isolation, are not the routine assessment in hospital 35 . BISAP score is established based on mortality, of which the predicting value for AP severity is relatively low 36 . MCTSI score is superior to other scoring systems on the assessment of the extent of pancreatic necrosis but fails in some SAP patients with delayed pancreatic necrosis 37 . APACHE II score was initially designed as an ICU instrument and therefore comprises many variables of medical history and medication details, which may be unavailable if the patients are unconscious, intubated, or transferred fromother medical institutions with insufficient medical records 38 . By contrast, the examinations of ApoA1 and ApoB are routine blood tests on admission, of which the operation is simple, convenient, and can be carried out in most hospitals. Therefore, the ApoB/A1 ratio may be more applicable to predict AP severity than the scoring systems.
Several limitations of this study should be considered when interpreting the results. First, this study was originated from a single institution in China, and the number of patients was relatively small. Thus, the findings of this study cannot be generalized. Further multicentric studies with large sample sizes are needed to validate the results. Second, given its cross-sectional design, this study could not establish a causal relationship between ApoB/A1 ratio and AP severity. Third, we did not compare ApoB/A1 ratio with other conventional inflammatory biomarkers, such as procalcitonin and IL-6. Fourth, outcomes related to SAP, such as mortality and organ failure, were not examined. AP patients with adverse outcomes would likely benefit from early classification of disease severity. Fifth, our study was conducted in the largest hospital in the region, which was committed to the treatment of critical illness, and thus could lead to a disproportional inclusion of patients with MSAP or SAP. Such selection bias might result in an overestimation of the predictive value of ApoB/A1 ratio.
In conclusion, this study revealed that serum ApoB/A1 ratio at admission is independently associated with disease severity in patients with AP. Pretreatment serum ApoB/A1 ratio can serve as a reliable indicator for SAP in clinical setting, and its application at admission may improve clinical management strategies for patients with AP.

Methods
Patient population. This retrospective observational study included 375 AP patients aged 18-75 years who were consecutively admitted to the Department of Gastroenterology between January 2014 to December 2017. All these patients were diagnosed with a first attack of AP according to typical symptoms, including acute abdominal pain and serum amylase level that was at least three times higher than the upper normal limit or a confirmation of pancreatitis by radiologic findings. The interval between the occurrence of symptoms and on admission was within 24 h. We excluded patients with chronic diseases, such as hypertension, diabetes, liver disease, renal disease, and malignant tumors. We also excluded patients who were pregnant or have taken lipid-lowering drugs within the last 6 months. The study protocol was approved by the Ethics Committee of the Affiliated Hospital of Guangdong Medical University (No: PJ2018-041). The study methods were carried out in accordance with relevant guidelines and regulations. Informed consent was obtained from each patient prior to study enrolment.
Definition. The disease severity of AP patients was measured at admission based on the 2012 revised Atlanta Classification and was divided into three groups, namely, MAP, MSAP, and SAP 2 . MAP was defined as the absence of organ failure and local or systemic complications. MSAP was defined as the presence of organ failure that resolved within 48 h and/or local or systemic complications without persistent organ failure. SAP was defined www.nature.com/scientificreports www.nature.com/scientificreports/ as persistent organ failure (lasting > 48 h). The Bedside Index for Severity in AP (BISAP) was determined within 24 h of admission. The Acute Physiology and Chronic Health Evaluation II (APACHE II) score and the Ranson score were also measured at 24 and 48 h after admission, respectively. Contrast-enhanced computed tomography (CT) was performed within 24 h after admission, and the Modified CT Severity Index (MCTSI) scores were subsequently assessed. Imaging examinations were performed by the same two senior radiologists. Body mass index (BMI) was calculated by dividing weight in kilograms by the square of a person's height in meters. Data collection. Demographics (i.e., age, sex, and BMI), etiology (i.e., gallstones or not), and lifestyle factor (smoking habit) were recorded for each patient on admission. Blood samples were collected from patients after overnight fasting on the first 24 h of hospitalization and then analyzed immediately for laboratory tests, including serum lipids [i.e., TC, triglyceride (TG), HDL-C, LDL-C, ApoA1, and ApoB], complete blood count [i.e., white blood cell (WBC), neutrophil, lymphocyte, red blood cell (RBC), hemoglobin (HGB), and platelets (PLT)], and serum biochemical indexes [i.e., CRP, alanine aminotransferase (ALT), aspartate aminotransferase (AST), lactate dehydrogenase (LDH), amylase, albumin (ALB), blood urea nitrogen (Bun), and calcium]. The ApoB/A1 ratio was calculated according to the ratio between serum ApoB and ApoA1 concentrations on the first day of hospitalization. Therapeutic regimen and ICU admission were decided by the attending physician, independently from the participants of this study. The length of hospital stays and the number of death with 28 days after admission were also recorded. At the time of data collection, the participants (YF Wang and WK Tan) were blinded to the disease severity and clinical outcomes of patients.
Statistical analysis. Data were recorded into a Microsoft Excel database. Continuous data accorded withnormal distribution were presented as means ± standard deviations (SDs); otherwise, they were presented as medians and interquartile ranges. Categorical variables were described as frequencies and proportions (%). Differences amongmultiple groups were evaluated by one-way variance analysis for continuous variables with normal distribution, nonparametric Kruskall -Wallis test for continuous variables with abnormal distribution, and by Pearson's χ 2 test or Fisher's exact test for categorical variables. Spearman rank correlation was performed to evaluate the correlation between serum lipid levels and scoring systems. We calculated the odds ratios (ORs) for predicting the risk of SAP by using logistic regression analysis after adjustment for confounding factors. All statistical analyses were conducted by SPSS version 25.0 for windows (SPSS Inc., Chicago, IL), except for the analyses with receiver operating characteristic (ROC) curve.
The area under the ROC curve (AUC) was applied to determine the discriminative values of serum lipid parameters for SAP. It was also used to assess the optimal cut-off by showing the trade-off between sensitivity and specificity. The AUCs were compared using Z test to identify the difference in predictive capability for SAP between serum lipid parameters 39 . The ROC was analyzed with the STATA 15.0 software (STATA Corp., College Station, TX). Two-sided P values less than 0.05 were regarded as statistically significant.