Progression of lipase activity and pancreatic lipase immunoreactivity in dogs hospitalized for acute pancreatitis and correlation with clinical features

Abstract Background Lipase activity and pancreatic lipase immunoreactivity (PLI) have not been compared in dogs hospitalized for acute pancreatitis (AP). Objectives To describe the progression of lipase activity and PLI, and correlations with clinicopathologic features in dogs with AP. Animals Thirty‐nine dogs with AP based on clinical signs and lipase activity >350 U/L (reference interval [RI], 24‐108 U/L). Methods Retrospective study. Lipase activity (LIPC Roche), PLI (SpecPL), and clinical signs were recorded daily. Admission (d1) data (clinical, laboratory, and ultrasound [US] findings), and clinical signs during hospitalization (d2‐d3) were assessed for correlation with lipases. Results Median (range) duration of clinical signs before presentation was 2 days (1‐7 days). Median (range) lipase activity and PLI at d1 were 1070 U/L (range, 357‐1500 U/L) and 1111 μg/L (range, 292‐1500 μg/L). Strong correlation between assays at d1 (r s 0.96; P < .0001; n = 39), remained equally strong on d2 (r s 0.964; P < .0001; n = 39), and d3 (r s 0.966; P < .0001; n = 22). On d2, lipase activity and PLI were within RI in 13/39 (33%) and 18/39 (46%) of cases. Lipase activities were minimally increased (median, 124 U/L) in 5 dogs with d2 PLI <200 μg/L. On d3, 4 more dogs had normal lipase activity and PLI, and the nature and magnitude of change were always the same for both assays. Clinical signs were not associated with lipases. Only a hyperechoic mesentery, but not an US diagnosis of AP, correlated significantly with lipase activity and PLI. Conclusions and Clinical Importance Lipase decreases rapidly to near or within RI within 2 days of treatment in the majority of dogs with AP. Both lipase assays yielded virtually identical results. Mesenteric echogenicity may be an early marker of AP in dogs.


| INTRODUCTION
Acute pancreatitis (AP) is a common disorder in dogs. The diagnosis usually is based on a combination of clinical, laboratory, and imaging findings. In absence of histopathology, determination of serum lipase, either as a concentration (pancreatic lipase immunoreactivity [PLI]) or an activity (1,2-o-dilauryl-rac-glycero-3-glutaric acid-(60-methylresorufin) ester [DGGR]-based lipase assays) is considered the laboratory test of choice. [1][2][3][4][5][6] Although both tests correlate highly, 2,3,5,7-12 comparisons of both assays with clinical signs and routine laboratory evaluations performed at initial presentation are lacking. Comparison of lipase assays over time also has not been examined. Besides lipase measurement, pancreatic ultrasonography (US) is regarded the second cornerstone of a diagnosis of AP, although agreement and correlation of lipase activity and PLI concentration with pancreatic US have been repeatedly found to be low. 9,11,[13][14][15][16] Emerging evidence suggests that duration of clinical signs before presentation affects lipase results and US evidence of pancreatitis differently. 11 With shorter disease duration, dogs had significantly higher lipase activities but US was not positive significantly more often for pancreatitis. 11 So far, the relationship between lipase and pancreatic US has not been assessed in dogs with clearly defined acute disease. Therefore, we had the following objectives: First, to describe the time course of DGGR-lipase activity and PLI and correlations with clinical signs in dogs with AP. Second, to determine if correlations exist between lipase activity and PLI and results of routine blood tests (CBC, serum biochemistry), as well as C-reactive protein (CRP) concentration. Third, to determine if a correlation exists between both lipase assay results and US findings in a defined acute setting. We hypothesized that lipase activity and PLI concentration would decrease to results within or minimally above the reference interval (RI) within 1 to 2 days in the majority of dogs with AP and that results of baseline and follow-up lipase activities and PLI concentrations would be interchangeable.

| Case selection and data collection
The study population of this retrospective study consisted of clientowned dogs. All dogs were presented to the Clinic for Small Animal Internal Medicine of the Vetsuisse faculty, University of Zurich between January 2019 and August 2021. Inclusion criteria were hospitalization for AP, available consecutive measurements of lipase activity, available surplus serum for corresponding PLI determinations from the same samples that were used to measure lipase activity, and owner permission to use leftover samples. A suspicion (US evidence of pancreatic mass lesions) or confirmation of pancreatic carcinoma was an exclusion criterion. Diagnosis of AP was based on at least 1 of the following signs (lethargy, anorexia, vomiting, abdominal pain, diarrhea) of ≤7 day duration, and DGGR-lipase activity >350 U/L (LIPC, Roche on Cobas, Roche Diagnostics, Rotkreuz, Switzerland using DGGR as substrate; RI, 24-108 U/L). 9,11

| Statistical analyses
Spearman's rank correlation coefficients (r s ) between lipase activity and PLI, at baseline (d1) and days 2 to 3, as well as from all 3 days together were determined. Spearman's rank correlation coefficients also were used to assess correlations among lipase activity, PLI, CRP

| Clinical signs at presentation
Median duration of clinical signs before presentation was 2 days (range, 1-7). Fifteen dogs presented after 1, 14 dogs after 2, 5 dogs after 3, 2 dogs after 4 and 5 days, respectively, and 1 dog after 7 days. The presence of clinical signs at d1 did not correlate with lipase activity or PLI (Table S1). Duration of clinical signs before presentation also did not correlate with lipase activity and PLI. Lipase activity and PLI were not significantly higher when recorded clinical signs were present or absent.  (Table S1). Lipase activity and PLI were not significantly different between dogs when individual clinical signs were present or absent.
The nature and magnitude of change within the reported range of results were the same for both assays. In 2 dogs, lipase activity and PLI had decreased at d2 and increased again on d3 (Figure 3), whereas lipase activity/PLI had further increased at d2 and decreased again at d3 in another 2 dogs (Figure 4). In 3 dogs with lipase activity >1500 U/L and PLI >1500 μg/L at days 1 and 2, lipase activity remained >1500 U/L on d3 and PLI concentrations on d3 were 1500, 1452, and 1303 μg/L, respectively.
In 3 dogs, lipase measurements were available beyond d3. In two dogs with persistently increased lipase activity and PLI at d3, results further decreased to 243 U/L and 276 μg/L at d4 and were within RI at d5 in 1 dog, whereas both lipase results were within RI at d4 and d5 in the other dog ( Figure 1A,B). In 1 dog with lipase activity and PLI at d2 and d3 within RI, lipase activity and PLI had returned to the initial level at d7 despite lack of relapsing clinical signs of AP ( Figure 1A,B).  significantly decreased compared with d2 (P = .0002; Figure 5A). The  F I G U R E 4 Lipase activity (blue) and PLI concentration (red) increased at d2 and decreased again at d3 in two dogs. Circles and squares symbolize lipase activity and PLI measurements from one patient, respectively apathy for 1 day and markedly increased lipase activities and PLI concentrations (marked red in Figure 2A,B). Upon first examination, the pancreas was of normal size, shape, and echogenicity. The mesentery was hyperechoic, and the US diagnosis was a hyperechoic mesentery in the cranial abdomen. During the second US examination, the pancreas was enlarged, with an irregular surface, rounded form, and mixedechoic parenchyma with a hyperechoic mesentery. The US diagnosis was AP. During both presentations, lipase activity and PLI completely decreased into RI within 24 hours (Figure 2A,B), and the dog was discharged at d2.

| DISCUSSION
We described the course of disease and progression of lipase activity and PLI results in dogs hospitalized for AP. Lipase decreased rapidly to minimally increased activity or results were within RI in approximately half of the dogs within 1 day.
The strong correlation between both lipase assays at admission  20 The only data matching this statement arise from a previous abstract. 21 The PLI concentrations (measured using the original PLI ELISA) were increased above an "empirical cutoff value of 250 μg/L" in 9/11 dogs with histologically-confirmed pancreatitis. 21 Dogs with pancreatitis had a median PLI concentration of 676.8 μg/L (no range reported). 21 The empirical cutoff was not described in more detail, but based on the RI at the time, was 2.5Â the upper RI. Form (acute vs chronic) and severity of pancreatitis was not reported. 21 These early data refer to the original PLI assay This information means the above-mentioned empirical 250 μg/L cutoff 21 amounts to approximately 375 μg/L PLI (Spec cPL), which is close to 400 μg/L. In our study, a lipase activity cutoff of >348 U/L corresponded to a PLI of >400 μg/L in the regression analysis. Interestingly, this >348 U/L cutoff turned out to be approximately in the region of 3Â the upper limit of the RI, a standard concept used to diagnose AP in humans and dogs. 8,23,24 We had originally also created a preliminary equivocal range for the interpretation of lipase activity (109-216 U/L 9 ), but experience has taught us that the cutoff for a diagnosis of AP is somewhere between 300 and 350 U/L. Therefore we included only dogs with lipase activity >350 U/L in our study.
However, we prefer to be more cautious with the use of rigid cutoffs because no peer-reviewed information is available about how the 400 μg/L PLI cutoff was established, and it is virtually impossible to definitely diagnose or rule out milder forms of AP without highly invasive biopsies. Moreover, our results imply that disease duration before lipase measurement should be included in the interpretation of lipase in acute cases. For this reason, we refrain from wording such as "consistent with pancreatitis" and generally prefer "suspicious for pancreatitis." It is precisely for these reasons that the concept of a 3Â the upper limit of the RI is applied in human medicine. Lower diagnostic cutoffs have been described recently for DGGR-based lipase activities. 5,12 Different assay methodology, 5 different statistical approaches, 12 and very likely inclusion of dogs with longer disease duration before the presentation may have played a role.
Comparative data evaluating lipase activity and PLI versus standardized histologic examination of the entire pancreas as reported in cats, 25 are not available for dogs. Our findings illustrate that lipase activity results parallel PLI assay results that are widely regarded as highly specific for the diagnosis of pancreatitis. It has been claimed that DGGR-based lipase activity assays are not specific for pancreatic lipase. 26,27 Recently, lipase activity was measured in six healthy dogs using a DGGR-based assay (Diazyme Laboratories) at baseline and 10, 20, 30, 60, and 120 minutes after IV heparin administration. 27 A significant difference (median difference of 4.3 U/L) was found between baseline and 10 minutes after IV heparin. This difference was most likely within RI, but no RI was given. 27 The authors discussed that the substrate DGGR is not only hydrolyzed by pancreatic lipase but also by hepatic lipase and thus not pancreas-specific. 27 This is possible, but does not allow assessment of whether or not such a minimal difference has any clinical relevance.
No clinical sign correlated significantly with lipase activity, PLI, or CRP measurements, neither at admission nor during hospitalization.
Lipase activity, PLI, and CRP concentration also were not significantly higher when clinical signs were present or absent. We were specifically interested in whether individual clinical signs rather than disease activity scores were associated with lipase activity or PLI. Scores may reflect disease severity but do not indicate the clinical problem.
Because dogs usually are discharged when they have improved clinically and lipase results have been shown to decrease after treatment, 28 it is not surprising to find an association between a clinical score and lipase result when all results from admission to discharge are taken together. 29 A single study describes clinical signs in dogs with AP where duration of signs before the presentation was specified. 4 In that study, 109 dogs had been sick for a median of 3 days (range, 0-15 days). Inappetence or anorexia (94%) and abdominal pain (60%) were more common compared with our findings, but other clinical signs were approximately equally frequent. 4  Disease duration before presentation often is not described in dogs with pancreatitis. Most publications do not mention duration of clinical signs before the presentation. 1,3,5,9,12,15,16,[29][30][31][32][33][34][35][36][37][38][39][40] Some studies included dogs with a predefined disease duration, but did not report the actual duration of clinical signs. 6,41,42 Some studies consider acute as short as 2 days, 43 others consider 10 days, 6 2 weeks, 41 or 3 weeks 44 as acute. Lipase activities were significantly higher with shorter durations of clinical signs before presentation in a recent study, 11 whereas the number of sick days before presentation did not correlate with lipase activity and PLI in our study. Delineating the time-lipase relationship in the very acute phase of disease appears difficult, and it might be necessary to record the time factor in hours rather than days to detect a difference in the very acute process when considering the short half-life of pancreatic lipase. 45,46 Two dogs had been treated with corticosteroids. Prednisolone can increase PLI concentrations into a diagnostic range for pancreatitis in healthy dogs, 47 whereas effects on DGGR-based lipase activity results are minimal and mostly within the RI. 48 We purposely did not exclude these dogs, because we were curious if lipase activity and PLI would be affected differently. As can be seen from Figure 2A,B (corticosteroid-treated dogs are marked orange), such was not the case.
Triglyceride concentration was the only laboratory variable significantly correlated with both lipase assays. We could not identify a correlation between CRP and triglyceride concentrations. This finding is in contrast to a recent study consisting of 31 dogs with AP where triglycerides were significantly correlated with CRP but not with PLI concentration at admission. 6 This difference may have resulted from longer disease duration before presentation (up to 10 days) in the previous study 6 when considering how rapidly lipase can decrease.
Another reason might have been different study designs. Acute pancreatitis was diagnosed if PLI was ≥400 μg/L either on the day of presentation or the next day, but PLI results were not reported. 6 The prevalence of hyperglycemia in our study was 37%, which was somewhat higher than the 30% reported in a previous study. 31 Hyperglycemia also has been reported in dogs with experimentallyinduced AP. 49 Data in humans suggest that hyperglucagonemia contributes to the hyperglycemia of AP. 50 Other explanations are decreased glucose tolerance because of metabolic stress associated with systemic inflammation. 51 We could not identify a correlation between glucose concentrations and lipase assays nor between glucose and CRP concentrations, and neither lipase activities nor PLI were higher when compared between dogs with normal and those with increased glucose concentrations. Stress hyperglycemia may have masked AP-induced hyperglycemia in dogs and larger studies are needed to examine this relationship.
A recent study suggested that an increased NLR ratio provides useful information regarding the course of AP in dogs. 52 In people, a recent meta-analysis suggested an increased NLR has diagnostic value in predicting the severity of AP. 53 6 Recently, PLI and CRP were measured daily during hospitalization in 13 dogs with AP. 29 No comparisons were made between days but, similar to our results, CRP increased and decreased during the first days of hospitalization. 29 Other authors suggested CRP might be useful for monitoring recovery from AP at the fifth day after treatment. 32 Concentration of CRP only was measured at admission and after 5 days, and not in between. 32 Our results suggest that CRP is less useful in the first 2 days after admission. Perhaps no correlation was observed on d1 because CRP takes longer to peak, and no correlation may have been observed on d3 because the lipase peak was already over. Similarly, data in humans with AP indicate that CRP at admission is unpredictable of disease severity. 54 The authors discuss that CRP is usually low in humans with AP if presentation is within a few hours of the onset of clinical signs because the hepatic synthesis of CRP peaks at 36 to 50 hours in people. 54 We also were interested in how pancreatic US compared with lipase measurements in a clearly defined acute setting. Data in humans suggest that it takes ≥3 episodes of AP without morphological changes in the pancreas until morphological changes are detectable. 55 Similarly, prevalence of imaging signs of AP in emergency department patients with lipase ≥3Â upper RI is low in people. 56 Also in dogs, US pancreatic changes consistent with AP may lag behind and occur later during hospitalization. 16 Individual pancreatic US variables were not separately assessed in that study. 16 We found that only a hyperechoic mesentery correlated significantly with lipase activity and PLI when dogs were acutely sick. Also, significantly higher lipase activity and PLI concentrations only were found when compared between dogs with and without a hyperechoic mesentery. Possibly, a hyperechoic mesentery represents an early marker for AP when the pancreas itself still appears unremarkable. Prospective studies are needed to verify this hypothesis.
Our study had some limitations. It would have been ideal to measure lipase activity, PLI, and CRP in all patients until all results decreased into the RI, but doing so was not possible in cases where clinical improvement preceded decreases in lipase. Also, in some cases there was not enough residual serum for CRP measurements.
Documentation of a renewed sharp increase of lipase activity and PLI on d7 in one dog was only possible because the owner preferred to have the dog hospitalized longer than recommended and requested another laboratory assessment at discharge when the dog had no clinical signs. Furthermore, we could have included more cases without concurrent PLI measurement. However a goal of the study was to characterize the relationship between the 2 lipase assays. To better study the relationship of lipase and CRP, it would be ideal to include dogs with identical disease duration before presentation in future studies. Furthermore, all dogs in our study survived to discharge, thus it is possible that findings of our study are not necessarily applicable to more severe cases.
In conclusion, lipase activity and PLI can decrease rapidly in dogs with AP and reach near-normal or normal results within 1 day in approximately 50% of cases even when markedly increased at first measurement. Thus, a diagnosis of AP can be missed if lipase determination is delayed by only 1 day. This observation is especially concerning when considering that pancreatic US can be normal in acute

ACKNOWLEDGMENT
No funding was received for this study.