Non-pancreatic hyperlipasemia: A puzzling clinical entity

Abstract

BACKGROUND

Increased lipase level is a serological hallmark of the diagnosis of acute pancreatitis (AP) but can be detected in various other diseases associated with lipase leakage due to inflammation of organs surrounding the pancreas or reduced renal clearance and/or hepatic metabolism. This non-pancreatic hyperlipasemia (NPHL) is puzzling for attending physicians during the diagnostic procedure for AP. It would be clinically beneficial to identify the clinical and laboratory variables that hinder the accuracy of lipase diagnosis with the aim of improve it. A more precise description of the NPHL condition could potentially provide prognostic factors for adverse outcomes which is currently lacking.

AIM

To perform a detailed clinical and laboratory characterization of NPHL in a large prospective patient cohort with an assessment of parameters determining disease outcomes.

METHODS

A Hungarian patient cohort with serum lipase levels at least three times higher than the upper limit of normal (ULN) was prospectively evaluated over 31 months. Patients were identified using daily electronic laboratory reports developed to support an ongoing observational, multicenter, prospective cohort study called the EASY trial (ISRCTN10525246) to establish a simple, easy, and accurate clinical scoring system for early prognostication of AP. Diagnosis of NPHL was established based on ≥ 3 × ULN serum lipase level in the absence of abdominal pain or abdominal imaging results characteristic of pancreatitis.

RESULTS

A total of 808 patients [male, n = 420 (52%); median age (IQR): 65 (51-75) years] were diagnosed with ≥ 3 × ULN serum lipase levels. A total of 392 patients had AP, whereas 401 had NPHL with more than 20 different etiologies. Sepsis and acute kidney injury (AKI) were the most prevalent etiologies of NPHL (27.7% and 33.2%, respectively). The best discriminative cut-off value for lipase was ≥ 666 U/L (sensitivity, 71.4%; specificity, 88.8%). The presence of AKI or sepsis negatively affected the diagnostic performance of lipase. NPHL was associated with a higher in-hospital mortality than AP (22.4% vs 5.1%, P < 0.001). In multivariate binary logistic regression, not lipase but increased amylase level (> 244 U/L) and neutrophil-to-lymphocyte ratio (NLR) (> 10.37, OR: 3.71, 95%CI: 2.006-6.863, P < 0.001), decreased albumin level, age, and presence of sepsis were independent risk factors for in-hospital mortality in NPHL.

CONCLUSION

NPHL is a common cause of lipase elevation and is associated with high mortality rates. Increased NLR value was associated with the highest mortality risk. The presence of sepsis/AKI significantly deteriorates the serological differentiation of AP from NPHL.

Key Words: Non-pancreatic hyperlipasemia, Acute pancreatitis, Glycoprotein 2, Acute kidney injury, Sepsis, Mortality

Core Tip: Non-pancreatic hyperlipasemia is a common and puzzling clinical entity in the differential diagnosis of acute pancreatitis (AP). The etiology of NPHL varies; however, sepsis and acute kidney injury (AKI) are the most prevalent causes. NPHL was associated with a high in-hospital mortality rate (22.4%). A readily available laboratory marker, the neutrophil-to-lymphocyte ratio, with a cut-off value > 10.37, has been the best independent laboratory predictor of mortality, with a nearly 4-fold increased risk. In the laboratory diagnosis of AP, the presence of complications, especially sepsis and AKI, warrants clinical attention, as it has a significant negative impact on the diagnostic accuracy of lipase.

INTRODUCTION

Increased lipase level is the serologic hallmark of the diagnosis of acute pancreatitis (AP) but is not specific to the disease. Hyperlipasemia can be detected in various diseases associated with lipase leakage due to inflammation of the surrounding abdominal organs of the pancreas or reduced renal clearance and/or hepatic metabolism[1]. This non-pancreatic hyperlipasemia (NPHL) is puzzling for treating physicians to distinguish it from AP. Improving the accuracy of serological diagnosis, including the more precise description of the NPHL condition, is an important goal. This latter can aid to identify the prognostic factors of this high mortality disease which is currently lacking.

The current definition of AP is based on the presence of two of the following three criteria: (1) Upper abdominal pain; (2) serum amylase and/or lipase elevated at least three times the upper limit of normal (ULN); and/or (3) characteristic findings from abdominal imaging [computer tomography (CT), magnetic resonance imaging (MRI), and ultrasonography][2]. In the presence of ≥ 3 × ULN serum lipase level but the absence of characteristic abdominal pain and imaging results, the diagnosis is NPHL.

Considering the serologic diagnosis of AP, recent United Kingdom guidelines suggested that serum lipase should be preferred over serum amylase, wherever available, due to its greater sensitivity and specificity[3]. In various studies, the sensitivity and specificity of lipase were reported to be 64%–100% and 87%–99.4%, respectively, whereas those of amylase were 35%–93% and 87%–99.1%, respectively[4]. Hence, persistent initiatives are being made to increase the accuracy of serological diagnosis of AP either by enhancing the diagnostic accuracy of lipase or by selecting and validating new and more specific biomarkers.

Increasing the cut-off levels of lipase could be a solution to improve the serological diagnosis[5], especially if certain conditions are associated. There are several reports in the literature where higher than ≥ 3 × ULN cut-off values were suggested for use in the presence of the following comorbidities: Renal insufficiency, malignant tumors, cholecystitis, esophagitis, hypertriglyceridemia, and in patients with diabetes[6]. Cohen et al[7] studied the optimal cut-off value in critically ill patients to diagnose AP. They reported a cut-off level of 532 U/L, which corresponded to almost 9 × ULN serum lipase elevation, and was associated with a specificity, sensitivity, negative predictive value (NPV), and positive predictive value (PPV) of 78%, 77.4%, 84.9%, and 67%, respectively. Although it would be reasonable, it has not been studied whether a renal functionadjusted lipase threshold is able to increase the sensitivity of lipase for diagnosing AP. Acute kidney injury (AKI) is common in patients with AP and even more common in NPHL conditions. Lipase is a small molecular weight (MW) protein of 48 kDa, and renal elimination is thought to be one of the major pathways for its elimination[8]. Accordingly, false or inappropriate increases in lipase levels have been reported in patients with end-stage renal disease owing to the prolonged elimination rate[9]. Potentially negative impact of AKI or sepsis that are common disease specific complications during an AP episode has not been evaluated so far.

Regrading validated new biomarkers, glycoprotein 2 (GP2), a 75 kDa MW protein is a potential contender among others. It is the major membrane protein in pancreatic exocrine zymogen granules and is a pancreatic tissue-specific marker[10]. Up to 40% of the total protein in the pancreatic cell membrane is composed of GP2. GP2 is located in the basal membrane of the acinear cell via a glycosylphosphatidylinositol linkage[11]. When pancreatic secretion is stimulated, GP2 is transported to the apical plasma membrane, cleaved, and secreted in a polarized manner into the pancreatic duct and intestinal lumen[12]. GP2 is also located in intestinal microfold (M) cells[13]. Inflammatory conditions often disrupt the polarity of epithelial cells, causing proteins to be released into the bloodstream via the serosal space, instead of entering the intestinal lumen. This process can be evaluated by measuring levels of serum GP2[12]. GP2 plays a role in modulating innate and adaptive immunity in the intestine[14]. In a well-characterized rat experiment, elevated GP2 serum levels were detected in AP[11]. Human studies are controversial whether serum GP2 level is a superior indicator compared to serum lipase level. Hao et al[12] reported that GP2 has more discriminatory power on diagnosing AP than lipase [the area under the curve (AUC): 0.96 vs 0.81]. According to another study by Roggenbuck et al[15], serum GP2 is a specific marker of AP and contributes to the prognosis of its severe form.

In the present study, we aimed to: (1) Determine NPHL prevalence, etiology, clinical course, and predictors of mortality in a large prospective Hungarian patient cohort; (2) assess diagnostic accuracy of lipase and amylase for AP in the presence of disease specific complications such as sepsis or AKI in order to improve the serologic diagnosis of AP potentially; and (3) to evaluate the diagnostic performance of GP2 in comparison with lipase.

MATERIALS AND METHODS

Patient population

We prospectively evaluated adult (≥ 18 years old) patients with serum lipase levels at least three times higher than the ULN (≥ 3 × ULN) consecutively admitted to the Clinical Centre of the University of Debrecen over 31 months, between June 2017 and December 2019. Patients were identified by daily electronic laboratory reports from the E-Medsolution Data Basis. These reports were developed to support an ongoing observational, multicentre, prospective cohort study (EASY trial, ISRCTN10525246) to establish a simple, easy, and accurate clinical scoring system for early prognostication of AP and were provided twice a day by the Directorate for Health Financing and Control of the University of Debrecen. The definition of AP was based on the presence of two of the following three criteria: (1) Upper abdominal pain; (2) elevation of serum lipase ≥ 3 × ULN (210 U/L, reference range is 0-70 U/L); and/or (3) characteristic findings on abdominal imaging (CT, MRI, or ultrasonography). If either two of the three specified criteria were met, a diagnosis of AP was established. NPHL was established based on ≥ 3 × ULN serum lipase level in the absence of characteristic abdominal pain and imaging results. Patients diagnosed with AP were either admitted to the Division of Gastroenterology or the Intensive Care Unit (ICU) on the Department of Internal Medicine, depending on the severity of their condition. Those with NPHL, on the other hand, were admitted to different departments based on the pathology requiring acute care. After admission, patients were followed-up prospectively until hospital discharge and/or death. Patients previously diagnosed with chronic pancreatitis (CP) based on clinical, morphological, and functional data according to international criteria[16] were excluded from the study.

Data collections

Data collection was based on a standardized protocol. Demographic, clinical, and laboratory characteristics at enrolment of AP and NPHL patients are presented in Table 1, along with the outcome data. Demographic data included age and gender, while clinical data included the Charlson Comorbidity Index (CCI) (including cardio- and cerebrovascular, respiratory, connective tissue, peptic ulcer, liver and renal disorders, diabetes mellitus, solid cancers, leukemia, lymphoma, and acquired immune deficiency syndrome)[17] as well as the presence of sepsis or AKI. Laboratory data included serum lipase, amylase, sodium, albumin, blood urea nitrogen, creatinine, glomerular filtration rate, white blood cell count (WBC), neutrophil granulocyte and lymphocyte counts, neutrophil-to-lymphocyte ratio (NLR), C-reactive protein level (CRP), and procalcitonin (PCT). PCT was measured only in cases of suspected infections (n = 512). The diagnosis of sepsis was made by the treating physician based on the Sepsis-2 consensus definition[18], which requires the presence of a proven or suspected infection along with at least two systemic inflammatory response syndrome (SIRS) criteria including body temperature > 38 or < 36 degrees Celsius, heart rate > 90 beats/min, respiratory rate > 20 breaths/min or partial pressure of CO₂ < 32 mmHg, leukocyte count > 12000 or < 4000/mL or > 10% immature forms or bands. The following bacterial infections were considered based on conventional criteria: Infections of skin and soft tissue, orocavital region, upper and lower respiratory tract (acute bronchitis, pneumonia), biliary tract (cholecystitis, cholangitis, liver abscess), intestinal tract (gastroenteritis), urinary tract (uncomplicated cystitis were excluded), osteomyelitis, and endocarditis. The following laboratory data were considered: Elevation of the WBC [absolute: > 10.8 × 10⁹/L or relative (in patients with leukopenia): Double of count at former visit] with an elevated neutrophil rate (> 76%) and elevated serum levels of high-sensitivity CRP (> 10.0 mg/L) and/or PCT (> 0.15 μg/L), including microbiological culture results, where available. Bacteriaemia was considered when clinical symptoms and signs of infection were present and confirmed by microbiological demonstration of the causative organism from blood culture in the absence of site-specific infection. AKI was defined if the serum creatinine (sCr) level increased more than 0.3 mg/dL (≥ 26.5 μmol/L) within 48 h or increased in sCr ≥ 1.5 times baseline, which is known or presumed to have occurred within the prior 7 days according to Kidney Disease: Improving Global Outcomes criteria[19].

Table 1 Demographic, clinical and laboratory characteristics at enrolment and outcome data of the acute pancreatitis and non-pancreatic hyperlipasemia patients, n (%).

	AP (n = 392)	NPHL (n = 401)	P value
Gender (male)	188 (48)	221 (55.1)	0.047
Age (yr)	62 (47-73)	66 (57-76)	< 0.001
Comorbidities any	310 (79)	369 (92)	< 0.001
CCI	3 (1-5)	5 (3-8)	< 0.001
Sepsis (yes)	56 (14.3)	111 (27.7)	< 0.001
Acute kidney injury (yes)	41 (10.5)	133 (33.2)	< 0.001
Sepsis and AKI (yes)	20 (5.1)	82 (20.4)	< 0.001
Lipase level (U/L)	1331 (579-2837)	278 (236-485)	< 0.001
Amylase level (U/L)	424 (172-1038)	170 (97-249)	< 0.001
Sodium (mmol/L)	139 (136-141)	139 (136-142)	0.568
Albumin (g/L)	39 (33-44)	35 (29-41)	< 0.001
BUN (mmol/L)	5.3 (3.7-7.8)	8.4 (5-20.35)	< 0.001
Creatinine (μmol/L)	73 (60-94)	92 (63-217)	< 0.001
GFR (mL/p/1.73m2)	87 (61-91)	61 (24-91)	< 0.001
WBC (G/L)	10.79 (7.81-14.45)	10.39 (7.63-14.61)	0.805
Neutrophil granulocyte (G/L)	8.46 (5.49-11.97)	7.87 (5.20-11.83)	0.584
Lymphocyte (G/L)	1.27 (0.83-1.82)	1.29 (0.82-1.89)	0.008
NLR	6.52 (3.54-12.93)	6.03 (3.21-12.40)	< 0.001
CRP (mg/L)	19.5 (5.8-95.34)	36.85 (10.88-101.93)	0.008
PCT (μg/L)	0.24 (0.09-0.69)	0.76 (0.25-3.13)	< 0.001
Total length of hospitalization (d)	7 (4-12)	8 (3-18)	0.126
ICU (yes)	174 (44.4)	182 (45.4)	0.830
ICU stay (d)	4 (2-7)	7 (3-16)	< 0.001
In-hospital mortality	20 (5.1)	90 (22.4)	< 0.001

Data are presented as median (interquartile range) if not otherwise indicated. For procalcitonin, the number of patients was 274 in non-pancreatic hyperlipasemia group and 238 in acute pancreatitis group. AP: Acute pancreatitis; NPHL: Non-pancreatic hyperlipasemia; IQR: Interquartile range; BUN: Blood urea nitrogen; GFR: Glomerular filtration rate; WBC: White blood cell; NLR: Neutrophil-to-lymphocyte ratio; CRP: C-reactive protein; PCT: Procalcitonin; ICU: Intensive care unit; CCI: Charlson Comorbidity Index; AKI: Acute kidney injury.

Outcome

Data regarding the length of hospitalization (general ward and ICU) and in-hospital mortality were considered disease outcomes. Prospective evaluation of clinical data during the total length of hospitalization enabled etiological classification of NPHL according to the method published by Hameed et al[20]. Two additional etiological categories were included based on a literature search: Cardiac surgery and obstetric diseases (Table 2).

Table 2 Etiologic classification of hyperlipasemia without acute pancreatitis.

	Patients number (n = 401)	Percentage (%)
Gastric and bowel disorders
Peptic ulcer	2	0.5
GI bleeding	4	1
Bowel necrosis/perforation	4	1
Bowel obstruction	8	2
IBD	5	1.2
Hepatobiliary disorders
Cholecystitis	8	2
Cholangitis	13	3.2
Liver disease	14	3.5
Post-ERCP lipasemia	16	4
Cancers
Pancreatic cancers	22	5.5
Non-pancreatic cancers	27	6.7
Haematological malignancies	6	1.5
Diabetes
Type 2 diabetes mellitus	19	4.7
Diabetic ketoacidosis	6	1.5
Neurosurgical pathology	16	4
Cardiovascular disease	17	4.2
Drugs side effect	17	4.2
Other	5	1.2
Infection without sepsis	30	7.5
Infection with sepsis
Sepsis without AKI	29	7.2
Sepsis with AKI	82	20.4
AKI without sepsis	51	12.7

All patients had elevated serum lipase level (> 210 U/L). GI: Gastrointestinal; IBD: Inflammatory bowel disease; ERCP: Endoscopic retrograde cholangiopancreatography; AKI: Acute kidney injury.

Serum glycoprotein 2 measurements

Blood samples were obtained at enrolment from each patient, and the sera were isolated after half an hour by centrifugation at 1800 × g, transferred to Eppendorf tubes and kept frozen at −70 °C until testing. Serum levels of GP2 were determined by solid-phase enzyme-linked immunoassay (Product code: 3950) according to the manufacturer’s instructions (GA Generic Assays GmbH, Dahlewitz, Germany). Samples were measured undiluted in duplicate on the same plate, and the mean values were used. In case a sample had higher than 10% coefficient of variation (CV) the sample was remeasured. Between each run, the CV was up to 7.8%. The lower limit of detection was 0.2 ng/mL. Serological assays were performed by a qualified analyst (Eva Tomori) at the Department of Laboratory Medicine in a blinded fashion without prior knowledge of the patients’ clinical information.

Statistical analysis

Variables were tested for normality using the Shapiro-Wilk’s W test. Continuous variables are summarized as mean ± SD or as medians [interquartile range (IQR)] according to their homogeneity. Categorical variables were compared using the χ² test or the χ² test with Yates correction, as appropriate. Continuous variables were compared using the Mann-Whitney U test or Student’s t-test. The relationship between continuous variables was assessed using non-parametric Spearman’s correlation. The diagnostic accuracy of lipase and GP2 for AP was estimated using receiver operating curve analysis by plotting sensitivity vs 1-specificity in all patients and sensitivity analyses were performed in patients with sepsis and/or AKI. The area under the receiver operating characteristic (AUC-ROC) and the corresponding 95%CI were calculated. Receiver operating characteristic (ROC) curves were compared using the method described by DeLong et al[21] in MedCalc. The Youden index was chosen, calculated as the maximum (sensitivity + specificity – 1) value, to estimate the best discriminate thresholds. The sensitivity, specificity, PPV, and NPV were calculated to determine the predictive power of lipase and GP2 or their combinations in both clinical settings. Binary logistic regression was used to assess the relationship between lipase level or different clinical and laboratory variables and in-hospital mortality in NPHL. Associations are presented as odds ratios (OR) or likelihood ratios (LR) with 95%CI. A 2-sided probability value < 0.05 was considered as significant. Patients with missing laboratory values (PCT) were excluded from the corresponding analysis. For statistical analysis and graphical presentation SPSS 29.0 (SPSS Inc., Chicago, IL, United States), GraphPad Prism 10 (San Diego, CA, United States), and MedCalc were used. The statistical methods used in this study were reviewed by Professor Elek Dinya, PhD, DSc, Semmelweis University, Institute of Health Informatics, Budapest, Hungary.

Ethical permission:

The study was reviewed and approved by the Committee of Science and Research Ethics of the Hungarian Medical Research Council (ETT TUKEB), and the Regional and Institutional Committee of Science and Research Ethics of the University of Debrecen (RKEB) (305/2014, 30595-1/2014 EKU, 55961-2/2016/EKU, 5753-2/2018 EKU). Study management strictly adhered to the Ethical Guidelines for Observational Studies.

RESULTS

Study population

Eight hundred and eight patients had elevated serum lipase levels, at least three times the ULN, assessed consecutively and prospectively over 31 months. Of these patients 392 had AP, 401 had NPHL, and 15 had CP. Patients in the latter group was excluded from the study. Table 1 summarizes the demographic, clinical, and laboratory characteristics of NPHL and AP patients at enrolment. The NPHL and AP patient groups differed in their demographic and clinical characteristics. There were slightly more male in the NPHL patient group than in the AP patient group (55.1% vs 48.0%, P = 0.047). NPHL patients were significantly older and had more comorbidities, as indicated by higher CCI scores (median, IQR: 5[3-8] vs 3[1-5], P < 0.001). Ninety-two percent of NPHL patients had at least one comorbidity, in contrast to 79% of AP patients (P < 0.001). Regarding laboratory parameters, patients with NPHL had significantly lower serum lipase [median, IQR: 278 (236-485) vs 1331 (579-2837) U/L, P < 0.001] and amylase levels [median, IQR: 170 (97-249) vs 424 (172-1038) U/L, P < 0.001]. Laboratory parameters reflecting the severity of the acute condition (such as albumin level and kidney function) were significantly impaired, whereas the degree of inflammatory response (such as CRP and PCT) was significantly higher in the NPHL group (Table 1). Interestingly, the lymphocyte count was notably higher in the NPHL group, resulting in a lower NLR compared to the AP group [median, IQR: 6.03 (3.21-12.40) vs 6.52 (3.54 vs 12.93), P < 0.001]. There was no difference in sodium levels and total WBC or neutrophil counts between the two groups. More severe clinical conditions in NPHL patients were also reflected in the higher prevalence of complications (sepsis and/or AKI) at enrolment (40.4% vs 19.6%, P < 0.001).

Etiological causes of NPHL

In the NPHL patient group, more than 20 different etiologies were identified with different prevalence rates, as detailed in Table 2. When creating disease etiologic groups, if the patient had sepsis or AKI, regardless of the other identified diseases, the patient was assigned to these groups. In addition to sepsis and AKI, the following disease categories were considered: Gastrointestinal (GI) diseases (gastric diseases, GI bleeding, large and small bowel diseases, hepatobiliary disorders, conditions related to the pancreas but beyond AP), GI and non-GI malignancies, abnormalities related to diabetes mellitus, neurosurgical pathologies, cardiovascular diseases, drug side effects, infections with or without sepsis, and AKI. The most prevalent etiologies were AKI (33.2%) and sepsis (27.7%).

Clinical and laboratory parameters associated with in-hospital mortality in NPHL patients

There was no difference in the total length of hospital stay between the two groups, but NPHL patients needed nearly twice as long ICU treatment as AP patients (median, IQR: 7 (3-16) vs 4 (2-7) d, P < 0.001). The diagnosis of NPHL was associated with a higher in-hospital mortality rate compared to AP (22.4% vs 5.1%, P < 0.001). The levels of different laboratory parameters upon admission were compared between survivors and non-survivors in the NPHL patient group (summarized in Table 3). There was no difference in serum lipase levels between the two groups, whereas amylase levels were significantly higher in non-survivors than in survivors [median, IQR: 221 (137-354) vs 175 (108-240) U/L, P = 0.001]. Moreover, significantly lower albumin levels and impaired kidney function were observed in the non-survivors. Regarding systemic inflammatory markers, NLR had the best discriminative ability [AUC-ROC, 95%CI: 0.747 (0.691-0.803), P < 0.001] with a cut-off value of > 10.37 to differentiate between survivors and non-survivors. The associated sensitivity, specificity, PPV, and NPV for this threshold value were: 63.95, 78.62, 45.8, and 88.5, respectively. The in-hospital mortality rate was significantly higher among patients with NLR above this threshold (45.8% vs 11.5%, P < 0.001). The discriminative ability of NLR was considerably higher than that of CRP [0.696 (0.637-0.756), P < 0.001; comparison P = 0.226] or PCT [0.620 (0.546-0.694), P = 0.002; comparison P = 0.238]. In multivariate binary logistic regression, patients’ age, presence of sepsis, high level of NLR (> 10.37), and amylase level (> 244 U/L), as well as low albumin level (≤ 34 g/L) were independent predictors of in-hospital mortality. Among independent mortality predictors, NLR had the highest OR (3.71, 95%CI: 2.006-6.863, P < 0.001; Table 4).

Table 3 Mortality as a function of different laboratory parameter in the group with non-pancreatic hyperlipasemia.

	Survivors (n = 311)	Non-survivors (n = 90)	P value	AUROC	95%CI
Lipase (U/L)	278 (236-477)	288 (240-561)	0.311	0.535	0.464-0.606
Amylase (U/L)	175 (108-240)	221 (137-354)	0.001	0.618	0.543-0.693
Sodium (mmol/L)	139 (136-142)	139 (133-144)	0.78	0.510	0.430-0.589
Albumin (g/L)	36 (30-42)	29 (25-35)	< 0.001	0.725	0.665-0.784
BUN (mmol/L)	7 (4.6-15.7)	19.6 (9.2-35.4)	< 0.001	0.724	0.666-0.783
Creatinine (µmol/L)	83 (62-170)	166 (78-334)	< 0.001	0.638	0.570-0.705
GFR (mL/p/1.73 m2)	74 (32-91)	29 (13-79)	< 0.001	0.653	0.588-0.719
WBC (G/L)	9.63 (7.23-13.37)	13.19 (9.34-20.98)	< 0.001	0.679	0.615-0.743
Neutrophil granulocyte (G/L)	6.80 (4.80-10.83)	11.11 (7.89-17.76)	< 0.001	0.715	0.655-0.775
Lymphocyte (G/L)	1.39 (0.89-1.97)	0.98 (0.62-1.56)	< 0.001	0.634	0.564-0.703
NLR	5.24 (2.84-9.66)	12.64 (6.64-19.50)	< 0.001	0.747	0.691-0.803
CRP (mg/L)	27.35 (9.23-81.36)	84.74 (34.13-135.34)	< 0.001	0.696	0.637-0.756
PCT (µg/L)	0.6 (0.21-2.04)	2.16 (0.41-4.23)	0.002	0.620	0.546-0.694

For procalcitonin, the number of patients was 197 in survivor group and 77 in non- survivor group. BUN: Blood urea nitrogen; GFR: Glomerular filtration rate; WBC: White blood cell; NLR: Neutrophil-to-lymphocyte ratio; CRP: C-reactive protein; PCT: Procalcitonin. AUROC: Area under the receiver operating characteristic; CI: Confidence interval.

Table 4 Multivariable logistic regression of in-hospital mortality in patients with non-pancreatic hyperlipasemia.

	OR	95%CI	P value
Age	1.024	1.004-1.044	0.019
Sepsis	2.087	1.115-3.906	0.021
NLR (> 10.37)	3.71	2.006-6.863	< 0.001
Amylase (> 244 U/L)	2.593	1.393-4.829	0.003
Albumin (≤ 34 g/L)	3.466	1.826-6.577	< 0.001
Constant	0.009		< 0.001

NLR: Neutrophil-to-lymphocyte ratio; OR: Odds ratio; CI: Confidence interval.

Association between lipase or amylase level and the presence of complications

In the NPHL patient group, the median lipase value was significantly higher in patients with disease complications than in those without any complications, regardless of the type (sepsis or AKI) or the cumulative effect (any or both) of the complication. The amylase values showed a similar range of variation as the lipase values in various clinical settings (Table 5). In the AP patient group, no similar association was observed for lipase or amylase (Table 6).

Table 5 Serum lipase and amylase levels according to the presence of complications in patients with non-pancreatic hyperlipasemia (n = 401).

	No. of patients	Lipase (U/L)		Amylase (U/L)
		Median	IQR	Median	IQR
No complications	239	264	229-428	159	106-226
Sepsis at all	111	425	263-651	216	112-343
AKI at all	133	398	259-625	190	99-335
AKI or sepsis	80	282	242-527	209	125-353
Sepsis and AKI	82	447	275-667	220	128-344
		P value		P value
None/sepsis at all		< 0.001		< 0.001
None/AKI at all		< 0.001		0.003
None/one (AKI or sepsis)		0.042		0.009
None/both (sepsis and AKI)		< 0.001		0.001

AKI: Acute kidney injury; AP: Acute pancreatitis; NPHL: Non-pancreatic hyperlipasemia; IQR: Interquartile range.

Table 6 Serum lipase and amylase levels according to the presence of complications in patients with acute pancreatitis (n = 392).

	No. of patients	Lipase (U/L)		Amylase (U/L)
		Median	IQR	Median	IQR
No complications	315	1287	577-2781	460	184-1064
Sepsis at all	56	1572	613-4588	485	199-1019
AKI at all	41	1370	511-4393	406	178-1381
AKI or sepsis	57	1529	611-4430	512	198-1123
Sepsis and AKI	20	1293	566-5175	447	188-1706
		P value		P value
None/sepsis at all		0.386		0.696
None/AKI at all		0.862		0.698
None/one (AKI or sepsis)		1.000		1.000
None/both (sepsis and AKI)		1.000		1.000

AKI: Acute kidney injury; AP: Acute pancreatitis; NPHL: Non-pancreatic hyperlipasemia; IQR: Interquartile range.

Accuracy of serum lipase and amylase level in the diagnosis of acute pancreatitis according to the presence of complications

The diagnostic accuracy of lipase for identifying patients with AP was established using ROC analysis and compared to that of amylase. Lipase identified patients with AP more accurately than amylase (AUC-ROC: 0.866 vs 0.756, P < 0.001). The best discriminative cut-off was defined as > 666 U/L (sensitivity and specificity of 71.4% and 88.8%, respectively) for lipase and > 365 U/L (sensitivity and specificity of 54.6% and 88.0%, respectively) for amylase.

Figure 1 shows the ROC curves and AUC values of lipase (Figure 1A) and amylase (Figure 1B) for identifying patients with AP according to the presence or absence of disease complications. The presence of AKI or sepsis as a complication of the disease negatively affects the diagnostic performance of lipase. No similar association was found for amylase levels. The diagnostic accuracy of lipase for identifying patients with AP in the presence of AKI or sepsis was obviously lower, represented by specificity (77.5% or 78.8%), LR+ values (3.25 or 3.33), and PPV (62.1% vs 50.9%) for the given > 666 U/L cut-off level compared with the no complication group (specificity: 94.9%; LR+: 14.16; PPV: 94.9%). There was practically no difference in the sensitivity values among the three groups (70.7%-73.2%). The cumulative presence of both complications at the same time yielded the worst diagnostic performance (specificity: 75.6%, LR+: 2.87, and PPV: 42.2%) without affecting the sensitivity (70.0%).

Figure 1 Receiver operating characteristic curve of lipase and amylase levels for identification of patients with acute pancreatitis according to presence of disease complications. A: Receiver operating characteristic curve (ROC) of lipase; B: ROC of amylase; C: Summery table of area under the receiver operating characteristic values (95%CI) and statistic differences. NPHL: Non-pancreatic hyperlipasemia; AP: Acute pancreatitis; AKI: Acute kidney injury; AUROC: Area under the receiver operating characteristic; CI: Confidence interval.

The optimum diagnostic thresholds for lipase based on ROC analysis were higher in the presence of AKI (> 951 U/L) or sepsis (> 861 U/L) and even higher when both complications occurred simultaneously (> 1587 U/L), compared to the absence of complications (> 666 U/L). Tables 7 and 8 summarizes the diagnostic performance of lipase with optimal cut-off values in different clinical settings.

Table 7 Performance characteristics of lipase in identifying acute pancreatitis in various clinical settings.

	Cut-off values (U/L)	Sensitivity (%)	Specificity (%)	LR+	LR-	PPV (%)	NPV (%)
AP overall	> 666	71.4	88.8	6.37	0.32	86.2	76.1
AP without complications	> 666	71.1	95.0	14.16	0.30	94.9	71.4
AP with sepsis	> 861	64.3	88.3	5.49	0.40	73.5	83.1
AP with AKI	> 951	58.5	92.5	7.79	0.45	70.6	87.9
AP with AKI or sepsis	> 722	71.9	90.0	7.19	0.31	83.7	81.8
AP with both complications	> 1587	50.0	97.6	20.5	0.51	83.3	88.9

AP: Acute pancreatitis; NPHL: Non-pancreatic hyperlipasemia; LR+: Positive likelihood ratio; LR-: Negative likelihood ratio; PPV: Positive predictive value; NPV: Negative predictive value; AKI: Acute kidney injury.

Table 8 Performance characteristics of amylase in identifying acute pancreatitis in various clinical settings.

	Cut-off values (U/L)	Sensitivity (%)	Specificity (%)	LR+	LR-	PPV (%)	NPV (%)
AP overall	> 365	54.6	88.0	4.54	0.52	82.6	64.9
AP without complications	> 354	55.6	93.8	8.93	0.47	92.6	60.1
AP with sepsis	> 560	46.4	92.3	6.04	0.58	76.5	76.2
AP with AKI	> 502	46.3	91.8	5.65	0.58	65.5	83.6
AP with AKI or sepsis	> 271	68.5	70.8	2.35	0.45	65.0	73.9
AP with both complications	> 542	45.0	93.5	6.93	0.59	64.3	86.7

AP: Acute pancreatitis; NPHL: Non-pancreatic hyperlipasemia; LR+: Positive likelihood ratio; LR-: Negative likelihood ratio; PPV: Positive predictive value; NPV: Negative predictive value; AKI: Acute kidney injury.

Serum glycoprotein 2 levels

One hundred and seventy-seven samples from patients (AP, n = 138; NPHL, n = 39) and 20 healthy controls (HC) were available for serum GP2 level measurements. Results showed that serum GP2 levels were significantly higher in this sub cohort than in the HC group, however there was no difference in the serum GP2 levels between the NPHL and AP patients’ groups (median, IQR: 1.5 ng/mL, 1,34 ng/mL and 0.34 ng/mL for AP, NPHL, and HC, respectively, P = 0.004). The presence or absence of complications did not affect the GP2 levels in either the AP or the NPHL groups. Moreover, in the NPHL group, there was no difference in serum GP2 levels between survivors and non-survivors (data not shown).

DISCUSSION

A decade ago, a landmark study by Hameed et al[20] was the first systematic review that extensively explored alternative causes of lipase level ≥ 3 × ULN than AP. By comparing 58 studies, they described numerous etiologies of NPHL. This systematic review drew clinicians' attention to the possibility of an incorrect diagnosis of AP and the great need to consider an alternative diagnosis in the interpretation and management of asymptomatic patients with serum lipase ≥ 3 × ULN. Few additional etiologies have been described in case reports and AP series, such as obstetric diseases or cardiac surgery[22,23]. So far, there has been only one study examining the clinical course of NPHL in relation to AP. Compared with the previous small-scale prospective study by Da et al[24] (NPHL, n = 51; AP, n = 170), the present Hungarian patient cohort comprised four times the number of patients (NPHL, n = 401; AP, n = 392). Similar to the method of Da et al[24], our patients were identified by a twice-daily electronic laboratory report from E-Medsolution Database. These reports were developed to support an ongoing observational, multicentre, prospective AP cohort study (EASY trial, ISRCTN10525246). The present study had a total length of 31 months. The primary aim of our study was to determine the prevalence, etiology, and clinical course of NPHL. According to the report of Da et al[24], patients with NPHL were predominantly male (76%), significantly older (age, mean ± SEM: 52.2 ± 1.9 years), and had more comorbidities (CCI, mean ± SEM: 4.6 ± 0.4) than patients with AP. These findings are consistent with our results. In their cohort, the leading etiologies for NPHL were decompensated cirrhosis (25.5%) and renal failure (15.7%), whereas our cohort showed a higher prevalence of sepsis (27.7%) and AKI (33.2%). NPHL patients had remarkably lower serum lipase levels compared to AP patients in both studies, with very similar numerical values [mean ± SEM: 360 ± 36 vs 1453 ± 135 IU/L, P < 0.001 vs median, IQR: 278 (236-485) vs 1331 (579-2837) U/L, P < 0.001]. The same was true for amylase values in our study; however no amylase data were reported in the study by Da et al[24] Laboratory parameters reflecting the severity of the acute condition (such as albumin level and kidney function) were significantly impaired in both NPHL and AP cohorts. Interestingly, in the study by Da et al[24] there were no major differences in length of hospitalization, ICU admission, or mortality between NPHL and AP patients, which was distinctly different from what we found. In our cohort, NPHL patients needed nearly twice as long ICU treatment as AP patients and had markedly increased in-hospital mortality rate (22.4% vs 5.1%, P < 0.001), underscoring the importance of NPHL as an adverse clinical feature. A plausible explanation for this could be the differences in baseline characteristics that reflect disease severity. While the degree of inflammatory response indicated by CRP [median, IQR: 36.85 (10.88-101.93) vs 19.5 (5.8-95.34) mg/L, P = 0.008] and PCT [median, IQR: 0.76 (0.25-3.13) vs 0.24 (0.09-0.69) mg/L, P < 0.001] values, and the presence of sepsis on admission (27.7% vs 14.3%, P < 0.001), was significantly higher in our NPHL group than in the AP group; SIRS on admission did not differ between the NPHL and AP patient groups (25% vs 28%, P = 0.86) in Da et al’s cohort[24]. Sepsis is known to be associated with increased mortality in excess of 10% and greater than 40% in patients with septic shock[18]. For confirmation in our NPHL patient cohort, the presence of sepsis was verified as an independent clinical predictor of in-hospital mortality in the multivariate binary logistic regression model [OR (95%CI): 2.087 (1.115-3.906), P = 0.021].

One of the novel findings of our study is that a simple, cheap, and easily available laboratory marker, the NLR, was the best independent predictor of mortality in NPHL patients. Although NLR is widely used across almost all medical disciplines as a reliable marker of immune response to various infectious and non-infectious stimuli, its clinical significance has not yet been evaluated in the NPHL population. In endotoxemia, there is significant neutrophilia and a parallel decrease in lymphocyte count[25]. In 2001, Zahorec[26] established NLR as a new immune-inflammatory response biomarker that shows the balance between the initial innate (neutrophils) and adaptive (lymphocytes) immune responses. The normal range of the NLR in a healthy adult population was reported between 1-2. However, this is influenced by several other factors, such as age, race, medication, smoking, obesity, anemia, stress, and the presence of various chronic diseases[27]. NLR is a very sensitive diagnostic tool for inflammation, infection, and sepsis. NLR values were significantly higher in patients with sepsis than in those without [median, IQR: 12.62 (7.81-21.58) vs 4.7 (2.67-8.68), P < 0.001] in our NPHL cohort. Pathologic cut-off values for NLR have been reported to be above 5 and below 0.7[28]. Beyond this, both thresholds were significantly associated with mortality under various conditions. NLR with a cut-off value of above 10.37 had a robust prognostic value in our NPHL patients, independent of the presence of sepsis and was helpful in differentiating more severe disease vs milder one. NLR values were notably higher in non-survivors than in survivors [median, IQR: 12.64 (6.64-19.50) vs 5.24 (2.84-9.66), P < 0.001]. An NLR value above 10.37 threshold was associated with a nearly 4-times increased risk of death. In line with this, NLR values above 11 were reported to be associated with a dramatic increase in the severity of illness, stress levels, and serious inflammation in various diseases[29].

Making the distinction between AP and NPHL is fundamental in everyday clinical practice for early and effective management of patients with acute deterioration and to avoid time-consuming but unnecessary examinations. Therefore, it is of vital importance especially for gastroenterologists, emergency physicians, and intensivists to be aware of all the possible etiologic factors of NPHL. In our large NPHL cohort (n = 401), 14.4% of the patients had acute gastric, bowel, or hepatobiliary disorders. Pancreatic tissue has a 50 to 100 fold greater lipase activity than other gastrointestinal organ[30]. Based on Apple et al[31], who examined lipase and pancreatic amylase activities in the pancreas and in six different parts of the human digestive system, the highest activities in the pancreatic tissue were followed by in the corpus of the stomach and duodenal bulb. The occurrence of peptic ulcers and upper GI bleeding was 1.5% in our NPHL cohort. Several authors have reported increased enzyme activity in inflammatory bowel diseases (IBD) with various prevalences. In general, elevation of pancreatic enzymes is more common in patients with Crohn’s disease than in those with ulcerative colitis[32,33]. IBD-associated NPHL was 1.2%, whereas abdominal surgery-related hyperlipasemia was 3% in our cohort. Bowel necrosis, perforation or obstruction, peritonitis, and appendicitis can cause lipase elevation[34]. Lipase elevation is the most common GI disorder in patients with hepatobiliary disorders. In our NPHL cohort, 5.2% and 3.5% of patients had cholangitis/cholecystitis and liver disease, respectively. In patients with liver disease, decreased hepatic metabolism and macro enzyme formation may be responsible for lipase elevation[35]. Transient lipase elevation after endoscopic retrograde cholangiopancreatography (ERCP) can occur without clinical evidence of AP. The incidence of post-ERCP lipasemia was 4% in our NPHL cohort. Gottlieb et al[36] conducted a comprehensive study on the prevalence of pancreatitis followed by ERCP, with a prevalence of 1%-3%. Gottlieb et al[36] developed a discharge algorithm for patients undergoing ERCP. A lipase cut-off value of 1000 U/L, which is about four times higher than that of ULN, was determined to rule out AP 2 h after ERCP with an NPV of 98%. Solid and hematologic malignancies are a major type of NPHL, and 13.7% of our patients had some type of malignancy. The differential diagnosis of hyperlipasemia is particularly difficult for clinicians in critically ill patients. Hyperlipasemia is very common among patients treated in the ICU (40%-57%)[37-39]. Among non-surgical ICU patients, hyperlipasemia is a typical laboratory finding in sepsis, renal failure, and diabetic ketoacidosis (DKA). Sepsis and AKI were the most prevalent NPHL etiologies (27.7% and 33.2%, respectively) in our cohort, while DKA was found in 1.5% of patients. Gut ischemia and severe hypotension are the main hallmarks of the pathogenesis of septic shock[40]. The pancreas is very sensitive to tissue hypoxia, so any condition that causes hypoperfusion of the splanchnic circulation can elevate lipase levels. Pezzilli et al[41] reported that lipase levels significantly increased without AP in septic shock. This result was confirmed by post-mortem evaluation of pancreatic tissue. Denz et al[38] found that only 35% of critically ill patients with increased serum lipase levels had positive CT findings for AP. There are several additional hypotheses underlying elevated lipase levels in sepsis, such as decreased renal excretion or hepatic metabolism and release of lipase from other extra-pancreatic organs because of obstruction or inflammation[20]. In AKI, impaired renal clearance is the main cause of lipase elevation, as in chronic renal failure. Patients with hemodialysis have very high pancreatic enzymes levels[42,43]. Finally, in DKA, the symptoms can be the same as those in AP (abdominal pain, nausea, vomiting), which makes the differential diagnosis more difficult. Yadav et al[44] found that 16%-20% of patients with DKA had elevated lipase levels without the presence of AP. Certain drugs, such as morphine and immune checkpoint inhibitors (PD1 and PD-L-1 inhibitors[45]), as front-line treatment for multiple types of cancer, can be associated with lipase elevation, which was verified in our drug-induced NPHL cases. In our cohort, drug side effects as a cause of hyperlipasemia was 4%. Seven cases of PD1 inhibitor treatment (1 nivolumab, 6 pembrolizumab), 4 patients with morphine derivatives, 3 with nonsteroidal anti-inflammatory drugs, and 3 with various chemotherapeutic agents used in gastric and colorectal cancer (darvulumab, cepacitabine, and oxaliplatin) were responsible for lipase elevation.

Interestingly, increased intracranial pressure, including intracerebral hemorrhage, edema, and tumors, is associated with elevated lipase levels. In 1932, Cushing observed that an increase in intracranial pressure by stimulation of the vagus was associated with hypermotility of the stomach by hypersecretion. This acid secretion causes pancreatic polypeptide efflux which is responsible for the elevated lipase levels[46].

Pancreatic cell damage is a well-known concept, but many studies have shown that in a small proportion of these cases actual pancreatic inflammation is found[39]. After cardiac surgery and cardiopulmonary bypass AP as a complication may appear but in a low percentage. Fernández-del Castillo et al[23] reported the incidence of pancreatic cell injury after cardiac surgery in 300 patients. Only 1% of the patients found to have severe pancreatitis and other study reported similar incidence[47]. Post operative pancreatitis is difficult to diagnose. In 24%-85% of cases asymptomatic hyperamylasemia was seen postoperatively[48,49]. Although clinically pancreatitis does not develop in the majority of cases, but higher amylase levels alone increase post operative mortality[23,49]. Pancreatic cell damage may be caused by perioperatively administered CaCl2 which increases cell damage in a dose-related manner, hence the importance of monitoring Ca during surgery[23].

The incidence of AP in pregnancy is rare, with one case per 1500-4500 pregnancies, but can be associated with high maternal-fetal mortality[50] that occurs most often in the second or third trimester or in the postpartum period. With increasing maternal age and associated metabolic abnormalities, such as obesity, high blood fat levels, and calcium levels, the incidence of the disease increases[51]. Gallstone formation (progesterone effect) and hypertriglyceridemia are the leading etiological factors to AP during pregnancy[52].

In our present cohort, we found two cases of transient lipase elevation after caesarean section, but the patients had no history of new abdominal pain and no evidence of gallstone disease. They were rather young, the serum triglyceride and calcium levels were within the normal range, and their pancreas was intact on abdominal ultrasonography.

The secondary aim of our study was to assess the diagnostic accuracy of lipase and amylase for AP in the presence of disease complications, such as sepsis or AKI, which are known to affect enzyme levels via different pathways. The large and equal number of patients in the AP (n = 392) and NPHL (n = 401) cohorts enabled the inclusion of sufficient individuals with disease complications in both cohorts. Fourteen point three percent of the patients with AP (n = 56) and 27.7% of NPHL (n = 111) patients had sepsis, while 10.5% (n = 41) and 33.2% (n = 133) had AKI, respectively. In line with the literature, lipase was found to be a more reliable indicator of AP than amylase[53], which is explained by the fact that pancreatic lipase is approximately four times more active than amylase and is not affected by loss of exocrine pancreatic function[31]. Moreover, lipase has a longer serum presence than amylase (8–14 d vs 1 wk)[54]. The sensitivity and specificity of lipase were reported to be 64%–100% and 87%–99.4%, respectively, while those of amylase were 35%–93% and 87%–99.1%, respectively in various studies. In our study, these values were 71.4% and 54.6% for sensitivity, and 88.8% and 88.0% for specificity, respectively. Various guidelines have shifted their preference to lipase tests instead of amylase. It is also recommended to use lipase as the only diagnostic marker instead of co-ordering it with amylase, which would eliminate unnecessary expenditure[4]. Currently, most guidelines for the laboratory diagnosis of AP stipulate serum amylase and/or lipase to be elevated three or more times the ULN. We also used this criterion to diagnose AP in the present cohort. The appropriate cut-off values for lipase and amylase in the laboratory diagnosis of AP are still under discussion. A greater than 5-fold increase in lipase has also been suggested[55], while a Japanese consensus[56] conference could not reach a consensus on appropriate cut-off values for amylase and lipase. In our study, the best discriminative cut-off was defined as > 666 U/L for lipase and > 365 U/L for amylase, which corresponds to 10 × ULN and 5 × ULN, respectively. Finally, in our cohort, serum GP2 protein levels were not useful for discriminating between the AP and NPHL patient groups. Although patients had significantly higher serum values than healthy subjects, no difference was observed between the two patient groups.

A recent study of Cohen et al[7] has drawn attention to the need for caution in interpreting elevated lipase ≥ 3 × ULN in critically ill patients due to its relatively low specificity and PPV. In the overall cohort of 417 ICU patients, the optimum lipase cut-off for the diagnosis of AP was 532 U/L. This cut-off value corresponded to 9 × ULN in their study. Accordingly, both the PPV (from 38.1% to 67.0%) and NPV (from 72.7% to 84.9%) improved the diagnostic value of lipase in AP, which helped to reduce unnecessary cross-sectional imaging in ICU patients. There are also common clinical conditions outside the ICU, where lipase levels rise above the threefold threshold, even in the absence of AP. AKI and sepsis are frequent complications of various diseases and common causes of NPHL. In our study, we consecutively evaluated a large patient population who arrived at the Clinical Centre of the University of Debrecen because of acute deterioration of various underlying conditions with ≥ 3 × ULN. A total of 30% of our patients had AKI and/or sepsis. The diagnostic performance of lipase was significantly hindered by the presence of either sepsis or AKI, and even more so when both conditions were present. The PPV was 94.9% in the absence of any complications, while 62.1% or 50.9% in the presence of sepsis or AKI, respectively. The PPV was 41.2% in the presence of both complications. One strategy to increase the specificity and PPV of the lipase assay is to increase the cut-off values for the diagnosis of AP that worked in our cohort as well. We describe novel lipase thresholds for certain clinical settings that might be directly applicable to assist clinicians in triaging the significance of elevated lipase levels in everyday clinical practice outside the ICU as well. The cut-off values for sepsis or AKI were > 861 U/L with PPV of 73.5% and > 951 U/L with PPV of 70.6%, respectively. When both complications were present at the same time, the best cut-off value was > 1587 U/L with a PPV of 83.3%. However, this strategy obviously results in a decreased sensitivity from 73.2% to 64.3% or from 70.7% to 58.5% in cases of sepsis or AKI, respectively. Similarly, when both complications were present simultaneously sensitivity decreased from 70% to 50%.

Limitations of the present study include the fact that it is a single-centre investigation. Furthermore, the results are presented in only one patient population, without a validation cohort, which may to some extent limit the generalisability of the results. However, the large size of the study population ensures the robustness of these findings. Furthermore, an extensive literature review has been performed that provides several supporting data and explanations for the presented results, which further increases the reliability of our findings.

CONCLUSION

NPHL is puzzling for attending physicians during the diagnostic procedure for AP. It would be clinically beneficial to identify the clinical and laboratory variables that hinder the accuracy of lipase diagnosis, as this would provide a more precise description of the NPHL condition, including its etiology, clinical significance, and prognostic factors. NPHL is associated with a high in-hospital mortality rate in our large-scale Hungarian prospective patient cohort. Sepsis and AKI were the most prevalent etiologies of NPHL (27.7% and 33.2%, respectively). NLR is the best laboratory predictor of mortality with a nearly 4-fold increased risk at the cut-off value of > 10.37. Physicians should maintain caution when interpreting hyperlipasemia in patients with acute deterioration but presented with sepsis and/or AKI. Traditionally used ≥ 3 × ULN lipase cut-off is associated with relatively low PPV in the presence of these complications. Increase of lipase cut-off improves diagnostic value of lipase for AP and helps to reduce unnecessary imaging in the absence of typical abdominal pain.