Received 00th January 20xx,

Department of Materials, Department of Bioengineering and Institute for Biomedical Engineering, Imperial College London, London SW7 2AZ, UK

Mologic Ltd, Bedford Technology Park, Thurleigh, Bedfordshire, MK44 2YP, UK

Hepatology and Gastroenterology Section, Department of Medicine, St Mary's Hospital Campus, Imperial College London, London, W2 1NY, UK

* Address correspondence to: m.stevens@imperial.ac.uk

Electronic Supplementary Information (ESI) available with additional characterisation and statistical analysis. See DOI: 10.1039/x0xx00000x

Accepted 00th January 20xx

DOI: 10.1039/x0xx00000x

www.rsc.org/

Phospholipase A2 as a point of care alternative to serum amylase and pancreatic lipase

Nathan J. Liu,a Robert Chapman,a Yiyang Lin,a Andrew Bentham,b Matthew Tyreman,b Natalie Philips,c Shahid A. Khan,c and Molly M. Stevens.a,*

Acute pancreatitis is a relatively common and potentially fatal condition, but the presenting symptoms are non-specific and diagnosis relies largely on the measruement of amylase activity by the hospital pathology unit. In this work we develop a point of care test for pancreatitis using secretory phospholipase A2 group IB (sPLA2-IB). Novel antibodies for sPLA2-IB were raised and used to design an ELISA and a lateral flow device (LFD) for the point of care measurement of sPLA2-IB, which was compared to pancreatic amylase activity, lipase activity, and sPLA2-IB activity in 153 serum samples. 98 of these samples were obtained from the pathology unit of a major hospital and classified retrospectively according to presence or absence of pancreatitis, and the remaining 55 were obtained from commercial sources to serve as high lipase (n=20), CA19-9 positive (n=15), and healthy (n=20) controls. sPLA2-IB concentration correlated well with the serum activity of both amylase and lipase, and performed at least as well as either markers in the differentiation of pancreatitis from controls.

Nanoscale

Please do not adjust margins

ARTICLE

ARTICLENanoscale

Please do not adjust margins

Please do not adjust margins

Nanoscale ARTICLE

Please do not adjust margins

Please do not adjust margins

This journal is © The Royal Society of Chemistry 20xxJ. Name., 2013, 00, 1-3 | 1

Please do not adjust margins

2 | J. Name., 2012, 00, 1-3This journal is © The Royal Society of Chemistry 20xx

This journal is © The Royal Society of Chemistry 20xxJ. Name., 2013, 00, 1-3 | 3

Introduction

In the UK, approximately 10,000 cases of acute pancreatitis occur annually [1]. While most cases of pancreatitis pass within a week from clinical presentation, 2-15% of cases may lead to death [1–3]. Diagnosis generally follows patient presentation with severe abdominal pain, nausea, and/or vomiting, and begins by laboratory measurement of serum amylase or lipase activity [4], followed by confirmation of inflammation in the pancreas using imaging modalities such as ultrasound (US) and computer-aided tomography (CT). The use of amylase and lipase activity as enzyme biomarkers, though valuable, are hardly conclusive in the diagnosis of pancreatitis, and may remain at normal levels or be only slightly elevated even in severe cases of acute pancreatitis [5,6]. The serum levels of amylase and lipase are dependent both on production and clearance, and may be elevated in other conditions, including cholecystitis, intestinal obstruction, ischemia, peptic ulcer disease, appendicitis, and gynecologic presentations, as well as in patients where renal function is compromised [4,7]. Furthermore, the turnaround time of amylase and lipase testing is several hours, limiting their utility for timely physician decision-making and potentially resulting in unnecessary hospital admission, for example following endoscopic retrograde cholangio-pancreatography (ERCP). As such, there is a need for validation of additional biomarkers for the rapid diagnosis of pancreatitis.

To address this issue, we recently developed a point of care activity assay for secretory phospholipase A2 group IB (sPLA2-IB) and group IIA (sPLA2-IIA) based on lateral flow technology [8,9]. The role of sPLA2 in pancreatitis was first described nearly three decades ago by Eskola et al., who identified several types of sPLA2 not pancreatic in origin, which we now know as other isoforms of the sPLA2 enzyme superfamily [10,11]. Particularly, both the enzymatic activity and mRNA expression of sPLA2 group enzymes were documented to increase from the onset of acute pancreatitis, suggesting serum sPLA2 may be tied to inflammatory mechanisms in pancreatitis [12,13]. Other studies also reported similar elevations in serum sPLA2-IB and IIA concentrations beyond normal reference limits in patients with acute pancreatitis [14,15]. However, contradictory findings which describe a decrease in sPLA2-IB mRNA expression and immunoreactivity in patients with pancreatitis versus controls have been reported [13]. Nonspecific sPLA2 activity has been shown to perform worse than amylase or lipase as a diagnostic biomarker for pancreatitis [16], possibly as a result of using assay formats that do not distinguish between the various confounding forms of PLA2. Indeed, we have found a number of commercially-available antibodies to be non-specific to the endogenous form of the enzyme having been raised against a peptide immunogen. In this work, we describe the development of a novel enzyme-linked immunosorbent assay (ELISA) and corresponding lateral flow device (LFD) for specific measurement of sPLA2-IB concentration. We then compare these assays to the sPLA2-IB activity assay previously described using a large range of clinical samples, to determine which format gives the most suitable test for pancreatitis.

Experimental

Sample collection: 98 serum samples, from 66 unique patients, constituting of high (>250 U/L), middle-range (100-250 U/L), and low (<100 U/L) amylase activity, were obtained from the Division of Blood Sciences of the Pathology Service at Hammersmith Hospital, via an ethical approval through the Human Tissue Bank at the Charing Cross Hospital (R14134, REC 12/WA/0196). Using container numbers recorded during sample collection, anonymized patient age, ethnicity, admission date, procedures performed, medical histories, and radiographic imaging were retrieved and examined retrospectively to determine a diagnosis. Diagnosis of pancreatitis was confirmed by CT imaging or established by the physician if laboratory results matched the clinical evaluation of the patient. The samples were then anonymized prior to testing. In addition, 20 serum samples with high lipase activity and 15 serum samples with elevated serum levels of pancreatic cancer marker CA19-9 were commercially obtained from BBI Solutions (Cardiff, UK), and 20 serum samples from healthy controls were obtained from SeraLabs (Haywards Heath, UK). All samples were stored up to one month prior to use at -80 °C.

Amylase activity: Amylase activity was measured using a commercial assay from Randox Laboratories (Antrim, United Kingdom), based on the substrate ethylidene-blocked pnitrophenyl maltoheptaoside (Ethylidene PNPG7).

Lipase activity: Lipase activity was measured using 1,2-o-dilauryl-rac-glycero glutaric acid-(6'-methylresorufin) ester (DGGR, Sigma-Aldrich) as a substrate. Assays were performed in a 96 well plate by mixing lipase standard / serum (5 μL) and assay buffer (150 μL, 41 mM TRIS + 1 μg/mL colipase + 1.8 mM sodium deoxycholate + 0.2 mM CaCl2, pH 8.4). The rate of fluorescence increase at 600 nm (ex = 529 nm) between 10 and 20 minutes after addition of the substrate solution (15 μL, 1.6 mM tartrate + 0.2 mM DGGR substrate + 7.2 mM taurodeoxycholate + 0.2 mM CaCl2, pH 4.0) was then monitored, and lipase values were calculated from a standard curve.

sPLA2-IB concentration: sPLA2-IB was measured using Mologic’s enzyme-linked immunosorbent assay (ELISA) and lateral flow devices (LFDs) (see http://mologic.co.uk/). The ELISA and LFD are based on a typical sandwich assay using mouse monoclonal Fab antibodies (1BC1, 1BD1, 1BD2) and a recombinant human sPLA2-IB standard (Alere). Recombinant enzymes were also purchased from R&D systems (PLA2-IB, PLA2-IIA, PLA2-VII), or Sinobiological (PLA2-IID). Briefly, for the ELISA, high-binding plates (Costar) were sensitized to sPLA2-IB with recombinant monoclonal anti sPLA2-IB capture Fab 1BC1 overnight. After washing , plates were blocked with 1% (w/v) BSA (Sigma-Aldrich #7906) in PBS for one hour, and then incubated with recombinant sPLA2-IB standard or serum at a dilution of 1:19 in sample buffer for two hours. Plates were washed and incubated for another hour with the detection antibody, which was prepared by conjugation of the desired Fab to horseradish peroxidase (HRP) using Lightning Link from Innova Biosciences (Babraham, UK). Following washing, TMB substrate solution (Biopanda Diagnostics) was added and developed for 20 minutes. Sulfuric acid was added to stop the reaction, and absorbance in the wells was read at 450 and 570 nm.

Gold nanoparticles for the lateral flow devices were prepared by immobilising the detection Fab 1BC1 on 40 nm gold nanoparticles (BBI,) and used together with an anti-biotin (control) gold colloid purchased from BBI . A mixture of test and control gold particles were sprayed onto a glass fibre release pad (Millipore) . Test lines of the detection Fab 1BD1 or 1BD2) were plotted on CN140 nitrocellulose membrane (Sartorious) along with a BSA-biotin control line. Gold conjugate pad and nitrocellulose membrane were mounted on plastic backing material together with a sample separator material and sink pad before finally being assembled within device housings. The lateral flow device was run by adding 80 µL of sample (typically diluted 1:20 in sample running buffer) or 80 µL standard to the device. The strength of the test line signal was read at 10 minutes using a lateral flow device reader (Abingdon Health, York, UK).

sPLA2-IB activity: Measurement of sPLA2-IB activity was performed according to our previous work, on lateral flow devices prepared as above, but with streptavidin coated gold nanoparticles and test line [8]. Liposomes were prepared from a suspension of 1-palmitoyl-2-oleoyl-sn-glycero-3-phosphoglycerol (POPG, 2.5 mg) in HEPES / NaCl (50 / 150 mM, 500 μL) buffer containing 4-arm polyethylene glycol (PEG) biotin (0.11 mM). The suspension was freeze-thawed twice, mixed, and extruded at 200 nm. 300 μL of the extruded liposomes were purified by column containing Sephadex G100, and the liposome fractionate was collected and diluted up to 1.5 mL in HEPES (50 mM) / NaCl (150 mM) buffer solution. The serum / standard (5 µL) was diluted with 15 µL of assay buffer (50 mM HEPES / 150 mM NaCl / 1% (wt) BSA / 20 mM CaCl2) containing 2 μL 100 mM sPLA2-IIA inhibitor in HEPES / NaCl buffer (LY315920, Generon, Berkshire, UK) to ensure specificity for sPLA2-IB. After incubating for 5 minutes, 3 µL of POPG liposomes containing biotinylated 4-arm polyethylene glycol (PEG) and 55 µL of assay buffer was added and the solution was allowed to flow up the laminar flow strips by capillary action. Signal was read after 10 minutes.

Statistical analysis: Statistical analyses were performed in R 3.0.3. All retrieved and measured characteristics for each subject were tabulated. Correlations between amylase, lipase, GIB concentration and activity were assessed using both the Spearman rank and Pearson correlation test. Receiver operating characteristic (ROC) curves were plotted and area under the ROC curve (AUC) analysis was performed to calculate optimal diagnostic thresholds and corresponding sensitivity and specificity using the methods provided by Robin et al.[17] Optimal thresholds for marker diagnostics were determined using Youden optimization. The size of all cohorts

Figure 1. Assays used in this work. a) Schematic of the lateral flow device (LFD) design. Sample, containing phospholipase A2 group IB (sPLA2-IB) is introduced to the sample pad and drawn through the device by capillary pressure. In the presence of the enzyme signal is generated by adhesion of immunogold to the test line through binding of de-novo fragment antibodies. b) ELISA standard curves and c) LFD standard curves, showing good sensitivity and specificity towards sPLA2-IB. d) Correlation between the sPLA2-IB concentration as measured by ELISA and LFD formats in the clinical serum samples (n=153).

Table 1. Baseline characteristics and sample numbers for the study set

Samples, n (unique patients)

Mean age, yrs

Male, n (%)

Pancreatitis

17 (8)

54

7 (88%)

Disease controls

81 (58)

61

36 (62%)

High lipase

20 (20)

NR

NR

CA19-9 positive

15 (15)

NR

NR

Healthy controls

20 (20)

NR

NR

NR = not reported (commercially obtained).

was not large enough to enable statistical comparison of ROC curves using DeLong’s test. All group differences in markers were assessed using Welch’s one-way analysis of variance (ANOVA), using a significance level of α=0.05.

Results

Novel ELISA and LFD assays for phospholipase A2 (PLA2IB) have been developed by Mologic Ltd, both of which showed good selectivity for the group IB form over other isoforms of PLA2 including group IIA, group VII and group IID (Figure 1b). Two different Fab antibody sandwich pairings were tested in the development of the PLA2IB ELISA. As can be seen the limit of detection (<0.025 ng/ml) of the first antibody pair (1BC1 / 1BD1was roughly twice that of the second antibody pair (1BC1 / 1BD2). The more sensitivity antibody pairing was therefore used in the ELISA for measurement of patient samples. In order to translate the assay to lateral flow device (LFD) format, we first immobilised the ELISA capture antibody 1BC1, on gold nanoparticles (AuNPs, 40 nm), and then dried these onto a glass fibre pad at the base of the Device. Test lines of the detection Fab (1BD1 or 1BD2) printed on the nitrocellulose membrane and the sensitivity of the two antibody pairings were compared (Figure 1a). Although more sensitive in ELISA format, control strips with the first antibody pairing (1BC1 / 1BD1) showed non-specific background signal in the absence of PLA2-IB (supporting information, Figure S1). This non-specific background was not observed when the second antibody pairing (1BC1 / 1BD2) was used, and given the signal for both formats at 100 ng/mL PLA2-IB was equally strong, the second antibody pairing was used in all future LFDs. These devices showed good sensitivity and specificity towards sPLA2-IB, with a limit of detection of < 0.25 ng/mL. The ten-fold loss of sensitivity relative to the ELISA is unsurprising given the reduction in incubation time from hours to minutes, and the use of the less sensitive antibody pairing. However even this limit of detection is well within the clinically important range for sPLA2-IB of 1-10 ng/mL. The LFD assay demonstrated good reproducibility and a wide dynamic range of at least one order of magnitude. Figure 1d shows a strong 1:1 correlation between the sPLA2-IB concentration as measured by ELISA and LFD in all of the clinical samples collected for this study, indicating that the LFD assay is highly robust and is not affected by anything normally present in serum. Although incorporation of a red blood cell filter into the LFD will allow the test to be run in whole blood, this was not necessary in this work as all of the clinical samples used were serum samples.

As clinical samples were collected randomly from the pathology unit of a major hospital, the range of conditions was extremely heterogeneous. 17 of the samples were subsequently assigned as originating from patients with pancreatitis and the 81 (classified as ‘disease controls’) from patients with a range of gastro- and nephro- related conditions but without pancreatitis. Table 1 summarises the clinical data on these samples, as well as for the serum samples obtained commercially to act as controls. This sample set provided a diverse range of both diseased and healthy samples with which to test how traditional biomarkers such as amylase and lipase correlate with sPLA2-IB. Amylase and lipase activity were measured using standard commercial assays used in the pathology unit at Hammersmith hospital, sPLA2-IB concentration was measured using our in-house ELISA and LFDs described above, and sPLA2-IB activity was measured by the LFD activity assay we have described in our previous work [8]. As expected, amylase and lipase activity were strongly associated with a Spearman correlation coefficient of 0.71 (Figure 2a). The parametric Pearson coefficient, showed a slightly weaker correlation indicating the relationship is monotonic but not necessarily linear (see supporting information, Table S1). sPLA2-IB concentration was correlated to both pancreatic amylase and lipase activities, with coefficients of 0.68 in both cases (Figure 2b-c) but only weakly with sPLA2-IB activity, with a Spearman correlation coefficient of 0.36 (Figure 2d). In order to estimate the effect of the differences in circulation time for each biomarker, the correlations were also performed on the first sample taken from each unique patient. In all cases, the correlations improved slightly with such selection, indicating some differences in circulation times exist, but the trends remained the same.

All biomarkers were significantly raised in the serum of patients with pancreatitis relative to the healthy controls (Figure 3), as other studies have reported. The sPLA2-IB LFD showed similar results to the other assays (Figure 3d),

Figure 2. Correlations between a) amylase and lipase activity, b) amylase activity and sPLA2-IB concentration (ELISA), c) lipase activity and sPLA2-IB concentration (ELISA), and d) sPLA2-IB concentration (ELISA) and activity (LFD), across the clinical sample set. Spearman correlation coefficients for all samples (n=153), and for the first sample from each unique patient (n=121) are reported (see Table S1 for full statistics).

Figure 3. Beeswarm and boxplots for the healthy control (n=20), disease control (n=81), and pancreatitis clinical samples (n=17), showing the levels of a) amylase b) lipase c) sPLA2-IB concentration (as measured by ELISA), d) sPLA2-IB concentration (as measured by LFD), and e) sPLA2-IB activity (LFD). Red markers indicated points above the optimum threshold value determined from the ROC curves. Notes: ns p > 0.05, * p < 0.05, ** p < 0.005.

Figure 4. Beeswarm and boxplots for the healthy control (n=20), high CA19-9 (n=20) and high lipase samples (n=20), showing the levels of a) amylase b) lipase c) sPLA2-IB concentration (as measured by ELISA), d) sPLA2-IB concentration (as measured by LFD), and e) sPLA2-IB activity (LFD). Notes: ns p > 0.05, * p < 0.05, ** p < 0.005.

ARTICLENanoscale

Please do not adjust margins

Please do not adjust margins

confirming the utility of this point of care test to complement the more cumbersome amylase and lipase assays. The heterogeneity of the data set provided an opportunity to also test the efficiency of sPLA2-IB to diagnose pancreatitis relative to disease controls. Amylase activity was not significantly overexpressed in the serum of pancreatitis patients relative to these controls, but both lipase and sPLA2-IB were (p < 0.05, Figure 3a-c). The point of care LFD for sPLA2-IB concentration similarly gave good discrimination between pancreatitis and disease control samples (Figure 3d), but the activity LFD did not (Figure 3e). Similar statistical significance was observed whether all samples were considered or only the first sample from each unique patient (see supporting information, Tables S2-3). No significant differences were observed in amylase, lipase, or sPLA2-IB activity between CA19-9 positive serum samples and healthy controls (Figure 4). The receiver operator curves (ROC) for each marker were also determined for the diagnosis of pancreatitis relative to the healthy and diseased controls (Figure 5). These show sPLA2-IB concentration was at least as good a biomarker as amylase, and almost as good as lipase, for pancreatitis. The area under the curve (AUC) for amylase was only 0.61 against the disease controls, but this rose to 0.70 for lipase and 0.72 for sPLA2-IB as measured by the point of care LFD (see supporting information, Tables S4-S7).

Figure 5. Receiver operator curves (ROC) showing sensitivity vs specificity for the diagnosis of pancreatitis (n=17) relative to a) healthy controls (n=20) and b) diseased controls (n=81), using amylase, lipase, sPLA2-IB concentration (ELISA) and activity.

Discussion

The assays developed in this study are highly specific and sensitive for the group IB form of phospholipase A2 (sPLA2-IB), and this allows for the first time the development of a point of care lateral flow test for the enzyme concentration. Our data reveals a strong correlation exists between serum sPLA2-IB concentration, serum amylase and serum lipase. Pancreatic lipase performed much better than amylase as a diagnostic biomarker for pancreatitis in this set, a confirmation that it is more specific and more resistant to changes over time. However, sPLA2-IB concentration performed at least as well as both of these more traditional biomarkers relative to both healthy and disease controls. Because the concentration of sPLA2-IB can be measured at point of care with considerably less effort than laboratory tests for these other pancreatic enzymes, we propose that it may serve as a useful analogue for amylase in pancreatic conditions. Interestingly, the concentration and activity of GIB also were not found to correlate, a surprising finding that we hypothesize results from the confounding action of an endogenous inhibitor of sPLA2-IB activity. When eliminating all serum activity using chemical inhibition, and spiking the inhibited serum with recombinant sPLA2-IB, we observed varying amounts of spike recovery (supporting information, Figure S2), which supports this hypothesis. Consequently, sPLA2-IB concentration and activity are not comparable or equivalent biomarkers.

The ability to measure sPLA2-IB at the point of care may justify its use in parallel with existing clinical laboratory testing. An area of clinical need for a point of care analogue for pancreatic enzymes is the prediction of acute pancreatitis following endoscopic retrograde cholangio-pancreatography (ERCP). Approximately 5-10% of patients undergoing the procedure will develop severe, sometimes life-threatening pancreatitis[18,19]. Amylase and lipase measurements taken at 4 hours are the current standard of care for diagnosis of pancreatitis within this population[18,19]. As a result of cumulative delay from the time course increase of the pancreatic enzymes and the extended turnaround times of laboratory testing, nearly all patients undergoing ERCP are often admitted to hospital overnight, incurring excess costs and resources for a large majority who do not develop clinically significant pancreatitis.

Conclusions

We have developed an accurate and sensitive assay for the measurement of sPLA2-IB concentration at the point of care. sPLA2-IB concentration correlates well with both amylase and lipase and is able to differentiate patients diagnosed with pancreatitis over diseased and healthy controls at least as well as these markers. Because the assay can be performed from a pinprick of blood at the bedside, we expect phospholipase A2 will prove to be a powerful clinical test for pancreatitis.

Acknowledgements

This work was carried out under a grant from the Technology Strategy Board (TSB) and the UK Engineering and Physical Sciences Research Council (EPSRC) (EP/K502352/1). NJL was supported by the US-UK Fulbright Commission and the Whitaker International Program for this work. The authors are also indebted to Sarah Chilcott-Burns and the Imperial College Healthcare NHS Trust Tissue Bank, and Dr. Beigun-Laroy at the Hammersmith Pathology Department for facilitating the transfer of the clinical samples and the NIHR Biomedical Research Centre Funding scheme at Imperial College London for support. MMS is an author on the patent application WO 2011/113813 which relates to the liposome amplification strategy utilised in this work for the PLA2 activity assay.

Notes and references

1. Toh SK, Phillips S, Johnson CD. A prospective audit against national standards of the presentation and management of acute pancreatitis in the South of England. Gut. 46, 239–243 (2000).

2. Frey CF, Zhou H, Harvey DJ, White RH. The incidence and case-fatality rates of acute biliary, alcoholic, and idiopathic pancreatitis in California, 1994-2001. Pancreas. 33(4), 336–344 (2006).

3. Lund H, Tønnesen H, Tønnesen MH, Olsen O. Long-term recurrence and death rates after acute pancreatitis. Scand. J. Gastroenterol. 41(2), 234–238 (2006).

4. Viljoen A, Patrick JT. In search for a better marker of acute pancreatitis: third time lucky? Clin. Chem. 57(11), 1471–3 (2011).

5. Clavien P, Robert J, Meyer P, et al. Acute pancreatitis and normoamylasemia: Not an uncommon combination. Ann. Surg. 210(5), 614–620 (1989).

6. Lankisch PG, Burchard-Reckert S, Lehnick D. Underestimation of acute pancreatitis: patients with only a small increase in amylase/lipase levels can also have or develop severe acute pancreatitis. Gut. 44(4), 542–544 (1999).

7. Muniraj T, Dang S, Pitchumoni CS. Pancreatitis or not? – Elevated lipase and amylase in ICU patients. J. Crit. Care. 30(6), 1370–1375 (2015).

8. Chapman R, Lin Y, Burnapp M, et al. Multivalent Nanoparticle Networks Enable Point-of-Care Detection of Human Phospholipase-A2 in Serum. ACS Nano. , DOI: 10.1021/nn5057595 (2015).

9. Liu NJ, Chapman R, Lin Y, et al. Point of care testing of phospholipase A2 group IIA for serological diagnosis of rheumatoid arthritis. Nanoscale [Internet]. , 10.1039/C5NR08423G (2016). Available from: http://pubs.rsc.org/en/Content/ArticleLanding/2016/NR/C5NR08423G.

10. Eskola JU, Nevalainen TJ, Kortesuo P. Immunoreactive pancreatic phospholipase A2 and catalytically active phospholipases A2 in serum from patients with acute pancreatitis. Clin. Chem. 34(6), 1052–1054 (1988).

11. Six D, Dennis E. The expanding superfamily of phospholipase A2 enzymes: Classification and characterization. Biochim. Biophys. Acta. 1488(1-2), 1–19 (2000).

12. Büchler M, Malfertheiner P. Prognostic value of serum phospholipase A in acute pancreatitis. Klin. Wochenschr. 67(3), 186–189 (1989).

13. Friess H, Shrikhande S, Riesle E, et al. Phospholipase A2 isoforms in acute pancreatitis. Ann. Surg. 233(2), 204–212 (2001).

14. Kitagawa M, Hayakawa T, Kondo T, Shibata T, Sakai Y. The diagnostic value of serum pancreatic phospholipase A2 (PLA2) in pancreatic diseases. Gastroenterol. Jpn. 26(1), 62–68 (1991).

15. Nevalainen TJ, Eerola LI, Rintala E, Laine VJO, Lambeau G, Gelb MH. Time-resolved fluoroimmunoassays of the complete set of secreted phospholipases A2 in human serum. Biochim. Biophys. Acta. 1733(2-3), 210–23 (2005).

16. Clavé P, Guillaumes S, Blanco I, et al. Amylase, lipase, pancreatic isoamylase, and phospholipase A in diagnosis of acute pancreatitis. Clin. Chem. 41(8), 1129–34 (1995).

17. Robin X, Turck N, Hainard A, et al. pROC: an open-source package for R and S+ to analyze and compare ROC curves. BMC Bioinformatics. 12(1), 77 (2011).

18. Nishino T, Toki F, Oyama H, Shiratori K. More accurate prediction of post-ERCP pancreatitis by 4-h serum lipase levels than amylase levels. Dig. Endosc. 20(4), 169–177 (2008).

19. Sutton VR, Hong MKY, Thomas PR. Using the 4-hour post-ERCP amylase level to predict post-ERCP pancreatitis. J. Pancreas. 12(4), 372–376 (2011)

.