Agreement of serum feline pancreas specific lipase and colorimetric lipase assays with pancreatic ultrasonographic findings in cats with suspicion

Transcription

1 SMALL ANIMALS/ Agreement of serum feline pancreas specific lipase and colorimetric lipase assays with pancreatic ultrasonographic findings in cats with suspicion of pancreatitis: 161 cases ( ) Samuel Oppliger, Dr med vet; Sonja Hartnack, Dr med vet; Claudia E. Reusch, Prof Dr med vet; Peter H. Kook, Dr med vet Objective To investigate agreement of a feline pancreas specific lipase assay and a colorimetric lipase assay with a 1,2-o-dilauryl-rac-glycero-3-glutaric acid-(6 -methylresorufin) ester (DGGR) substrate with results of pancreatic ultrasonography in cats with suspicion of pancreatitis. Design Retrospective case series. Animals 161 client-owned cats with suspicion of pancreatitis. Procedures Feline pancreas specific lipase concentration and DGGR lipase activity were measured from the same blood sample in cats undergoing investigation for pancreatitis, with < 24 hours between ultrasonography and lipase determinations. Ultrasonographic variables evaluated were ultrasonographic diagnosis of pancreatitis, enlargement, margins, echogenicity, mesenteric echogenicity, peripancreatic free fluid, cysts, masses, and common bile and pancreatic duct dilation. Agreement was assessed by use of the Cohen κ coefficient. Results Agreement between the lipase assays was substantial (κ = 0.703). An ultrasonographic diagnosis of pancreatitis had fair agreement with feline pancreas specific lipase concentration 5.4 µg/l (κ = 0.264) and DGGR lipase activity > 26 U/L (κ = 0.221). The greatest agreement between feline pancreas specific lipase concentration 5.4 µg/l and DGGR lipase activity > 26 U/L was found for a hypoechoic and mixed-echoic (κ = and 0.266, respectively), hypoechoic (κ = and 0.181, respectively), and enlarged (κ = and 0.223, respectively) pancreas. Conclusions and Clinical Relevance Agreement between pancreatic ultrasonography and lipase assay results was only fair. It remains unknown whether lipase results or pancreatic ultrasonography constitutes the more accurate test for diagnosing pancreatitis; therefore, results of both tests need to be interpreted with caution. (J Am Vet Med Assoc 2014;244: ) Pancreatitis appears to be a common disease in cats, 1 3 yet antemortem diagnosis remains challenging because of vague clinical signs and nonspecific changes in CBC and serum biochemical analysis results. 4 At present, feline pancreas specific lipase 5 concentration determination and pancreatic ultrasonography 6 are the 2 tests generally considered key tools for diagnosing pancreatitis in cats. The feline pancreas specific lipase assay is generally believed to represent the most sensitive test for the diagnosis of pancreatitis in cats and is in fact most often used as a noninvasive biochemical standard for its diagnosis Recently, a newer catalytic lipase assay 12 was validated for use in feline serum and compared with the feline pancreas specific lipase assay in 251 cats with suspicion of pancreatitis. Results of the 2 lipase assays agree substantially, 13 with the DGGR lipase assay appearing to be a more cost-effective alternative pancreatitis test. Compared with histologic evaluation, the sensitivity and specificity of pancreatic ultrasonography as a di- From the Clinic for Small Animal Internal Medicine (Oppliger, Reusch, Kook) and the Section of Epidemiology (Hartnack), Vetsuisse Faculty, University of Zurich, 8057 Zurich, Switzerland. Presented in part at the American College of Veterinary Internal Medicine Forum, Seattle, June Address correspondence to Dr. Kook (peterhendrikkook@gmail.com). AUC CI DGGR fpli ABBREVIATIONS Area under the curve Confidence interval 1,2-o-dilauryl-rac-glycero-3-glutaric acid- (6 -methylresorufin) ester Feline pancreatic lipase immunoreactivity agnostic tool for pancreatitis in cats varies greatly, 5,14 16 and most of the cases reported were evaluated > 10 years ago. However, the technology of available ultrasonographic devices has markedly improved since then. Also, the perception of the clinical relevance of pancreatitis in cats has clearly increased in the past 10 years, and the routine incorporation of the feline pancreas specific lipase assay into the pancreatitis testing might have additionally changed the selection of patients undergoing pancreatic ultrasonography. The following ultrasonographic variables are usually considered for a diagnosis of pancreatitis in cats: pancreatic size, pancreatic echogenicity, echogenicity of the surrounding mesentery, peripancreatic fluid accumulation, and cystic-appearing lesions. 15,17,18 Although these variables are routinely used in the ultrasonographic evaluation of the pancreas, no study has ever assessed the agreement of ultrasonographic findings compatible with pancreatitis 1060 Scientific Reports JAVMA, Vol 244, No. 9, May 1, 2014

2 with the most useful laboratory pancreatitis markers in cats suspected of having the disease. This is particularly surprising in view of the inherent invasiveness of pancreatic histologic evaluation, which generally renders it unfeasible. Thus, the aim of the study reported here was to assess the agreement between results of pancreatic ultrasonography and the results of 2 lipase tests. Materials and Methods Case selection For the period from December 2008 to August 2012, all cats with a suspicion of pancreatitis, with testing involving abdominal ultrasonography, and in which feline pancreas specific lipase and DGGR lipase activity had been determined from the same blood sample were included in the study. Pancreatitis was suspected when one of the following criteria was present and could not be explained otherwise: anorexia, lethargy, vomiting, or signs of abdominal pain. 3,4,16 Results were included only if the time interval between abdominal ultrasonography and lipase determinations did not exceed 24 hours. Medical records review Repeated measurements in the same cat were excluded from analysis. For results of the feline pancreas specific lipase assay, a only the cutoff value 5.4 µg/l was considered. 19 Serum DGGR lipase activity (reference range, 8 to 26 U/L) was measured with a recently validated in-house assay. 13,b Variables for which values were retrieved from the medical records for each cat were signalment, ultrasonographic findings, feline pancreas specific lipase concentration, and DGGR lipase activity. Ultrasonography was performed by board-certified radiologists or residents in training for board certification supervised by a boardcertified radiologist who were unaware of both lipase results (results of lipase determinations were not available at the time of the ultrasonography study). Pancreatic ultrasonography was performed with state-of-the art devices. c,d The following ultrasonographic features of the pancreas were recorded for each cat: ultrasonographic diagnosis (ie, pancreatitis [yes or no]), enlargement, irregular margins, echogenicity (hypoechoic, hyperechoic, and mixed-echoic), surrounding mesenteric echogenicity, peripancreatic free fluid, cysts, masses, and common bile and pancreatic duct dilation. When available, results of pancreatic histologic evaluation (biopsy samples collected during exploratory surgery or results from necropsy) were also included. The histopathologic diagnoses of acute pancreatitis, chronic pancreatitis, or normal pancreas were available and compared with the results of ultrasonography. Results from histologic evaluation were included only if the time interval between histologic evaluation and determination of both lipase tests did not exceed 9 days. Acute pancreatitis was defined as neutrophilic infiltration and necrosis; chronic pancreatitis was defined as lymphocytic infiltration and fibrosis. Statistical analysis Statistical analysis was performed with the aid of commercial software. e The agreement of pancreatic ultrasonographic findings with feline pancreas specific lipase concentration and DGGR lipase activity was assessed on the basis of Cohen κ coefficients. 20 Values < 0 indicate no agreement, values from 0 to 0.20 indicate slight agreement, values from > 0.20 to 0.40 indicate fair agreement, values from > 0.40 to 0.60 indicate moderate agreement, values from > 0.60 to 0.80 indicate substantial agreement, and values from > 0.80 to 1 indicate almost perfect agreement. 21 Furthermore, the performance of feline pancreas specific lipase and DGGR lipase assays, compared with pancreatic ultrasonography, the hypothetical gold standard, was evaluated through receiver operating characteristic curves and corresponding AUC. Results Animals During the study period, 360 cats underwent diagnostic investigation for pancreatitis, including pancreatic ultrasonography and determination of feline pancreas specific lipase concentration as well as DGGR lipase activity. Of these, 199 cases did not fulfill the inclusion criteria and were excluded, leaving results for 161 cats available for analysis. The study population consisted of 108 (67.1%) male (103 neutered) and 53 (32.9%) female (50 neutered) cats. Ages ranged from < 1 to 18 years (median, 11 years). Breeds included domestic shorthair (n = 120), Persian (9), British Shorthair (5), Maine Coon (5), Birman (3), Abyssinian (2), Balinese (2), domestic longhair (2), domestic shorthair Siamese mix (2), Norwegian Forest Cat (2), Siamese (2), Burman (1), Devon Rex (1), mixed breed (1), Orientale (1), Ragdoll (1), Russian Blue (1), and Turkish Angora (1). Feline pancreas specific lipase and DGGR lipase assay results Seventy-six of 161 (47.2%) cats had high feline pancreas specific lipase concentration, and 86 of 161 (53.4%) cats had high DGGR lipase activity. Agreement between the 2 lipase assays was substantial (κ = 0.703; SE, 0.056); rank correlation between the 2 lipase assays was good (Spearman ρ = 0.832). Results from ultrasonography Seventy-five of 161 (46.6%) cats had an ultrasonographic diagnosis of pancreatitis: 62 (38.5%) had a hypoechoic or a mixedechoic pancreas, 44 (27.3%) had a hypoechoic pancreas, 43 (26.7%) had peripancreatic free fluid, 35 (21.7%) had an enlarged pancreas, 33 (20.5%) had a mixedechoic pancreas, 30 (18.6%) had a hyperechoic mesentery, 26 (16.1%) had irregular pancreatic margins, 14 (8.7%) had evidence of a dilated common bile duct, 13 (8.1%) had evidence of a dilated pancreatic duct, 11 (6.8%) had a hyperechoic pancreas, 6 (3.7%) had a pancreatic mass, and 5 (3.1%) had pancreatic cysts. Ultrasonography versus feline pancreas specific lipase and DGGR lipase assays Of the 75 cats with an ultrasonographic diagnosis of pancreatitis, feline pancreas specific lipase concentration was high in 46 (61.3%). In the remaining 86 cases with an ultrasonographically normal pancreas, feline pancreas specific lipase concentration was high in 30 (34.9%) cats. The DGGR lipase activity was high in 49 of 75 (65.3%) cats with an ultrasonographic diagnosis of pancreatitis. In the remaining 86 cats with an ultrasonographically normal pancreas, DGGR lipase activity was high in 37 (43.0%) cats (Table 1). Pancreatitis was diagnosed ultrasonographically in 46 of 76 (60.5%) cats with a high feline pancreas SMALL ANIMALS/ JAVMA, Vol 244, No. 9, May 1, 2014 Scientific Reports 1061

3 SMALL ANIMALS/ specific lipase concentration and in 29 of 85 (34.1%) cats with a feline pancreas specific lipase concentration within reference range. An ultrasonographic diagnosis of pancreatitis was found in 49 of 86 (60.0%) cats with Table 1 Distribution (number [%]) of cats with high feline pancreas specific lipase concentration or high DGGR lipase activity in 161 cats grouped on the basis of results of ultrasonographic examination of the pancreas. Feline pancreas specific DGGR Ultrasonographic finding lipase lipase Ultrasonographically diagnosed pancreatitis Absent (n = 86) 30 (34.9) 37 (43.0) Present (n = 75) 46 (61.3) 49 (65.3) Pancreatic enlargement Absent (n = 126) 51 (40.5) 58 (46.0) Present (n = 35) 25 (71.4) 28 (80.0) Irregular margins Absent (n = 135) 59 (43.7) 79 (52.7) Present (n = 26) 17 (65.4) 20 (76.9) Hypoechoic pancreas Absent (n = 117) 45 (38.5) 55 (47.0) Present (n = 44) 31 (70.5) 31 (70.5) Hyperechoic pancreas Absent (n = 150) 72 (48.0) 72 (48.0) Present (n = 11) 4 (36.4) 7 (63.6) Mixed-echoic pancreas Absent (n = 128) 55 (43.0) 62 (48.4) Present (n = 33) 21 (63.6) 24 (72.7) Hypoechoic or mixed-echoic pancreas Absent (n = 99) 36 (36.4) 42 (42.4) Present (n = 62) 40 (64.5) 44 (71.0) Hyperechoic mesentery Absent (n = 131) 57 (43.5) 67 (51.1) Present (n = 30) 19 (63.3) 19 (63.3) Peripancreatic free fluid Absent (n = 118) 48 (40.7) 59 (50.0) Present (n = 43) 28 (65.1) 27 (62.8) Pancreatic cysts Absent (n = 156) 73 (46.8) 84 (53.8) Present (n = 5) 3 (60.0) 2 (40.0) Pancreatic masses Absent (n = 155) 74 (47.7) 83 (53.5) Present (n = 6) 2 (33.3) 3 (50.0) Dilated common bile duct Absent (n = 147) 68 (46.3) 77 (52.4) Present (n = 14) 8 (57.1) 9 (64.3) Dilated pancreatic duct Absent (n = 148) 67 (45.3) 76 (51.4) Present (n = 13) 9 (69.2) 10 (76.9) a high DGGR lipase activity and in 26 of 75 (34.7%) cats with a DGGR lipase activity within reference range (Table 2). Contingency tables of the results were developed (Tables 3 and 4). Twenty-three cats had a serum albumin concentration < 25 g/l (median, 21 g/l; range, 15 to 24 g/l; reference range, 29 to 37 g/l), and 6 of 23 cats had an enlarged pancreas. Five of these 6 cats had high feline pancreas specific lipase concentration, and all 6 had increased DGGR lipase activity. Agreement of ultrasonography versus feline pancreas specific lipase and DGGR lipase assays Agreement between an ultrasonographic diagnosis of pancreatitis and high feline pancreas specific lipase concentration was slight (κ = 0.264; SE, 0.076). Agreement between an ultrasonographic diagnosis of pancreatitis and high DGGR lipase activity was slight (κ = 0.221; SE, 0.076). The highest agreement between an ultrasonographic diagnosis of pancreatitis and high feline pancreas specific lipase concentration was found when the feline pancreas specific lipase concentration cutoff was set at > 16 µg/l (κ = 0.332; SE, 0.068). The highest agreement between an ultrasonographic diagnosis of pancreatitis and high DGGR lipase activity was found Table 3 Contingency table of the distribution of cats with an ultrasonographic diagnosis of pancreatitis (absent or present) and feline pancreas specific lipase assay results in 2 ranges. Pancreatitis Feline pancreas specific lipase Absent Present Total < 5.4 µg/l µg/l Total Table 4 Contingency table of the distribution of cats with an ultrasonographic diagnosis of pancreatitis (absent or present) and DGGR lipase assay results in 2 ranges. Pancreatitis DGGR lipase Absent Present Total 26 U/L > 26 U/L Total Table 2 Distribution (number [%]) of cats (n = 161) with feline pancreas specific lipase concentration and DGGR lipase activity in various ranges with various abnormal pancreatic ultrasonographic findings. Feline pancreas specific lipase concentration DGGR lipase activity < 5.4 µg/l 5.4 µg/l 26 U/L > 26 U/L Ultrasonographic finding (n = 85) (n = 76) (n = 75) (n = 86) Diagnosis of pancreatitis 29 (34.1) 46 (60.5) 26 (34.7) 49 (60.0) Pancreatic enlargement 10 (11.8) 25 (32.9) 7 (9.3) 28 (32.6) Irregular margins 9 (10.6) 17 (22.4) 6 (8.0) 20 (22.4) Hypoechoic pancreas 13 (15.3) 31 (40.8) 13 (17.3) 31 (36.0) Hyperechoic pancreas 7 (8.2) 4 (5.3) 4 (5.3) 7 (8.1) Mixed-echoic pancreas 12 (14.1) 21 (27.6) 9 (12.0) 24 (27.9) Hypoechoic or mixed-echoic pancreas 22 (25.9) 40 (52.6) 18 (24) 44 (51.2) Hyperechoic mesentery 11 (12.9) 19 (25.0) 11 (14.7) 19 (22.1) Peripancreatic free fluid 15 (17.6) 28 (36.8) 16 (21.3) 27 (31.4) Pancreatic cysts 2 (2.4) 3 (3.9) 3 (4.0) 2 (2.3) Pancreatic masses 4 (4.7) 2 (2.6) 3 (4.0) 3 (3.5) Dilated common bile duct 6 (7.1) 8 (10.5) 5 (6.7) 9 (10.5) Dilated pancreatic duct 4 (4.7) 9 (11.8) 3 (4.0) 10 (11.6) 1062 Scientific Reports JAVMA, Vol 244, No. 9, May 1, 2014

4 when the DGGR lipase activity cutoff was set at > 57 U/L (κ = 0.371; SE, 0.056). Receiver operating characteristic curves of feline pancreas specific lipase concentration and DGGR lipase activity against an ultrasonographic diagnosis of pancreatitis as the hypothetical gold standard were plotted (Figure 1). The AUC was (SE, 0.041; 95% CI, to 0.786) and (SE, 0.042; 95% CI, to 0.769) for feline pancreas specific lipase concentration and DGGR lipase activity, respectively (Table 5). Figure 1 Receiver operating characteristic curve for diagnosis of pancreatitis in 161 cats by use of feline pancreas specific lipase (solid line) and DGGR lipase (dashed line) assays with an ultrasonographic diagnosis of pancreatitis as the hypothetical gold standard. Histologic evaluation Histologic evaluation (8 necropsies and 3 biopsies were performed) was available in 11 cats, and pancreatitis was found in 9 (acute pancreatitis was found in 4, and chronic pancreatitis was found in 5). In the 2 cats without histologic evidence of pancreatitis, pancreatic biopsy specimens were collected during exploratory laparotomy (1 cat had gastric foreign body and an ultrasonographic suspicion of pancreatitis; 1 cat was icteric, and laparotomy was performed because of suspicion of feline infectious peritonitis). Ultrasonographic findings in cats with histologically confirmed acute pancreatitis (n = 4) were peripancreatic free fluid (3), irregular margins (2), enlarged pancreas (1), hypoechoic pancreas (1), hyperechoic pancreas (1), mixed-echoic pancreas (1), hyperechoic mesentery (1), dilated common bile duct (1), pancreatic cyst (1), and pancreatic mass (1). Three of 4 cats with a histologically confirmed diagnosis of acute pancreatitis had an ultrasonographic diagnosis of pancreatitis. In these 3 cats, feline pancreas specific lipase concentration and DGGR lipase activity were increased. In the remaining cat, in which an ultrasonographic diagnosis of pancreatitis could not be established, the ultrasonographic finding was a dilated common bile duct. In this cat, feline pancreas specific lipase concentration was between 3.6 and 5.3 µg/l, and DGGR lipase activity was within reference range. Ultrasonographic findings in cats with histologically confirmed chronic pancreatitis (n = 5) were hyperechoic pancreas (2), peripancreatic free fluid (2), dilated common bile duct (2), enlarged pancreas (1), mixed-echoic pancreas (1), and irregular margins (1). One of 5 cats with a histologically confirmed diagnosis of chronic pancreatitis had an ultrasonographic diagnosis of pancreatitis. In 4 cats in which an ultrasonographic diagnosis of pancreatitis could not be established, feline pancreas specific lipase concentration was 5.4 µg/l in 3 of 4 cats and 3.5 µg/l in 1 of 4, whereas DGGR lipase activity was high in 2 of 4. Ultrasonographic findings in the cat with histologically confirmed chronic pancreatitis in which an ultrasonographic diagnosis of pancreatitis was made were an enlarged pancreas, mixed echogenicity, irregular margins, and peripancreatic free fluid. In this cat, feline pancreas specific lipase concentration and DGGR lipase activity were high. In the 2 cats with a histologically normal pancreas, ultrasonographic findings were hyperechoic pancreas (n = 1), mixed-echoic pancreas (1), and hyperechoic mesentery and peripancreatic free fluid (1). Both cats had an ultrasonographic diagnosis of pancreatitis. Feline pancreas specific lipase concentration and DGGR lipase activity were high (n = 1) and within reference range (1). SMALL ANIMALS/ Table 5 Correlations (Cohen κ ± SE) and AUC (95% CI) values for the relationship between pancreatic ultrasonographic findings and results of pancreatic-specific lipase and DGGR lipase assays. Cohen κ Feline pancreas specific DGGR Feline pancreas specific DGGR Ultrasonographic finding lipase assay lipase assay lipase assay lipase assay Diagnosis of pancreatitis ± ± ( ) ( ) Pancreatic enlargement ± ± ( ) ( ) Irregular margins ± ± ( ) ( ) Hypoechoic pancreas ± ± ( ) ( ) Hyperechoic pancreas ± ± ( ) ( ) Mixed-echoic pancreas ± ± ( ) ( ) Hypoechoic or mixed-echoic pancreas ± ± ( ) ( ) Hyperechoic mesentery ± ± ( ) ( ) Peripancreatic free fluid ± ± ( ) ( ) Pancreatic cysts ± ± ( ) ( ) Pancreatic masses ± ± ( ) ( ) Dilated common bile duct ± ± ( ) ( ) Dilated pancreatic duct ± ± ( ) ( ) AUC JAVMA, Vol 244, No. 9, May 1, 2014 Scientific Reports 1063

5 SMALL ANIMALS/ Discussion In the present study, the agreement between the most commonly performed tests (ultrasonography and serum lipase concentration or activity) in cats with suspicion of pancreatitis was assessed. Only slight to fair agreement was found between pancreatic ultrasonography and serum lipase determinations, with the best evaluated ultrasonographic variables being a combined hypoechoic and mixed-echoic pancreas, hypoechoic pancreas, and enlarged pancreas. We also chose to include the DGGR lipase assay in our investigation because it represented a considerably less expensive test. The high level of agreement between both assays suggested that the DGGR lipase assay was a viable alternative to the feline pancreas specific lipase assay. One explanation for the poor agreement between ultrasonography and serum lipase assay results could be too many false-positive ultrasonographic diagnoses. Previous pancreatitis might have gone undetected or been misdiagnosed, but pancreatic ultrasonography abnormalities could still be detectable. Similarly, even with an in-depth medical history, it is virtually impossible to rule out previous episodes of pancreatitis in feline patients. Furthermore, we cannot rule out individual age-related pancreatic parenchymal changes. Another explanation could be silent chronic pancreatitis without relevant enzyme release but with sufficient ultrasonographically detectable pancreatic parenchymal remodeling. Also, there could have been too many false-negative results of pancreatic ultrasonography. Even with the newest ultrasonography technology, there could be insufficient differences among acoustic impedances of the abnormal and normal pancreas to permit clinically applicable characterization of pancreatic tissue. 22 Few studies have evaluated single ultrasonographic pancreatic variables in cats. Saunders et al 15 retrospectively investigated performance of ultrasonography in 20 cats with acute pancreatic necrosis. Frequent ultrasonographic findings in these cats were peritoneal effusion (n = 9 cats), hypoechoic pancreas (7), and hyperechoic mesentery (7). An ultrasonographic diagnosis of pancreatitis could be established in only 7 of 20 cats. 15 Ferreri et al 16 performed ultrasonography in 28 cats with acute necrotizing pancreatitis and 24 cats with chronic nonsuppurative pancreatitis before necropsy. An ultrasonographically normal pancreas was found in approximately 50% of cats with acute and chronic pancreatitis. Pancreatic echogenicity appeared to be the most sensitive variable for detecting pancreatitis. 16 Understandably, the design of the studies comparing ultrasonography with histologic evaluation is not clinically applicable to patients undergoing investigation of suspected pancreatitis, and comparisons with the results of noninvasive tests are needed. Most recently, a study 11 has assessed the ultrasonographic pancreatic findings in 35 cats with suspicion of pancreatitis and compared these with results of fpli (based on radioimmunoassay) and feline pancreas specific lipase tests as their gold standard. Single ultrasonographic variables evaluated in stored still images were compared with pancreatic lipase results determined within 3 days after ultrasonographic examination. In our study, a maximum of 24 hours between performance of serum lipase tests and pancreatic ultrasonographic examination was chosen because the biological half-life of circulating feline pancreas specific lipase and DGGR lipase is presently unknown. However, with this short interval, we believed it unlikely that ultrasonographic abnormalities may have resolved themselves in the period between serum analysis and ultrasonography. The most recent study 11 comparing ultrasonographic results with feline pancreas specific lipase values did not reveal whether the patients actually also had an ultrasonographic diagnosis of pancreatitis at the time of examination. In contrast, we focused on the agreement between ultrasonographic and lipase assay results rather than declaring the feline pancreas specific lipase assay a gold standard, given that details on its validation and diagnostic accuracy are still unknown. Moreover, we excluded all cats for which the earlier fpli had been measured because the agreement and correlation between fpli and feline pancreas specific lipase concentration also remains unknown. We only investigated the agreement between pancreatic ultrasonographic results and serum lipase assays without knowing the true state of disease. With the present study design, we could not address whether ultrasonography or determination of serum lipase is more accurate in diagnosing pancreatitis. Theoretically, this would require a standardized histologic examination of the entire pancreas immediately after pancreatic ultrasonography. However, such an experimental approach would be ethically unjustifiable. There were some limitations to this study. Because of the retrospective study design, cats were examined by different ultrasonographers. Although only complete ultrasonographic reports containing all variables were included, it remains unclear how much weight different radiologists ascribed to the single pancreatic findings when a diagnosis of pancreatitis was made. In the present study, the ultrasonography report was written by the radiologist after completion of the examination. This approach seems more appropriate, better reflecting the completeness of the pancreatic evaluation than reviewing single stored ultrasonographic images, in view of the fact that radiologists differ dramatically in their assessment (even between normal and abnormal) of static ultrasonography images. 22 Ultrasonography reports in the present study did not specify how much of both limbs and the body of the pancreas was actually examined or whether body weight affected visibility. This might be a limitation, especially in cats with an unremarkable ultrasonographic appearance of the scanned pancreas. Only a small number of histologic samples were available in the present study. Also, because most histologic results were from necropsy reports, there might be a bias toward cats that died or were euthanized. Nonetheless, there is presently debate as to whether histologic evaluation represents the ideal gold standard, considering that mild pathological changes in the pancreas might not be clinically relevant and localized inflammatory disease can easily be missed when surgical or laparoscopic biopsy samples are obtained. Concern has been raised that hypoproteinemia could affect assessment of pancreatic size when evaluat Scientific Reports JAVMA, Vol 244, No. 9, May 1, 2014

6 ed by use of ultrasonography in dogs, 23 but this has not yet been critically evaluated in cats. We do not believe that decreased oncotic pressure leading to pancreatic edema was a confounding factor in our study because only 6 of the 23 cats with a serum albumin concentration < 25 g/l had an ultrasonographically enlarged pancreas and nearly all of them had increased lipase values (6 times DGGR lipase activity and 5 times feline pancreas specific lipase concentration). Overall, pancreatic ultrasonographic results and the 2 lipase assay results had only poor agreement. Even if patients with clinical signs attributable to pancreatitis will still benefit from a full abdominal ultrasonographic examination to exclude other differential diagnoses, pancreatic ultrasonography findings need to be interpreted with caution, especially considering the low level of agreement between serum lipase assays and ultrasonographic variables commonly assumed to be reliable (ie, peripancreatic fluid accumulation and mesentery echogenicity) for diagnosing pancreatitis. On the basis of the present results, pancreatic enlargement, a hypoechoic pancreas, and irregular margins seem most reliable. Further prospective studies investigating ultrasonographic characteristics of pancreatitis in cats are warranted. a. IDEXX GmbH, Ludwigsburg, Germany. b. Lipase colorimetric for Roche Cobas Integra 800, Roche Diagnostics, Rotkreuz, Switzerland. c. Philips iu22, Philips AG, Zurich, Switzerland. d. Aloka Prosound Alpha 10, Hitachi Medical Systems Europe Holding AG, Zug, Switzerland. e. SPSS, version 20 for Mac OS X, IBM Corp, New York, NY. References 1. De Cock HEV, Forman MA, Farver TB, et al. Prevalence and histopathologic characteristics of pancreatitis in cats. Vet Pathol 2007;44: Owens JM, Drazner FH, Gilbertson SR. Pancreatic disease in the cat. J Am Anim Hosp Assoc 1975;11: Steiner JM, Williams DA. Feline exocrine pancreatic disorders. Vet Clin North Am Small Anim Pract 1999;29: Steiner JM. Diagnosis of pancreatitis. Vet Clin North Am Small Anim Pract 2003;33: Forman MA, Marks SL, De Cock HEV, et al. Evaluation of serum feline pancreatic lipase immunoreactivity and helical computed tomography versus conventional testing for the diagnosis of feline pancreatitis. J Vet Intern Med 2004;18: Armstrong PJ, Williams DA. Pancreatitis in cats. Top Companion Anim Med 2012;27: Forcada Y, German AJ, Noble PJM, et al. Determination of serum fpli concentrations in cats with diabetes mellitus. J Feline Med Surg 2008;10: Bayliss DB, Steiner JM, Suchodolski JS, et al. Serum feline pancreatic lipase immunoreactivity concentration and seroprevalence of antibodies against Toxoplasma gondii and Bartonella species in client-owned cats. J Feline Med Surg 2009;11: Zimmermann E, Hittmair KM, Suchodolski JS, et al. Serum feline-specific pancreatic lipase immunoreactivity concentrations and abdominal ultrasonographic findings in cats with trauma resulting from high-rise syndrome. J Am Vet Med Assoc 2013;242: Schweighauser A, Gaschen F, Steiner J, et al. Evaluation of endosonography as a new diagnostic tool for feline pancreatitis. J Feline Med Surg 2009;11: Williams JM, Panciera DL, Larson MM, et al. Ultrasonographic findings of the pancreas in cats with elevated serum pancreatic lipase immunoreactivity. J Vet Intern Med 2013;27: Panteghini M, Bonora R, Pagani F. Measurement of pancreatic lipase activity in serum by a kinetic colorimetric assay using a new chromogenic substrate. Ann Clin Biochem 2001;38: Oppliger S, Hartnack S, Riond B, et al. Agreement of the serum Spec fpl and 1,2-O-dilauryl-rac-glycero-3-glutaric acid- (6 -methylresorufin) ester lipase assay for the determination of serum lipase in cats with suspicion of pancreatitis. J Vet Intern Med 2013;27: Gerhardt A, Steiner JM, Williams DA, et al. Comparison of the sensitivity of different diagnostic tests for pancreatitis in cats. J Vet Intern Med 2001;15: Saunders HM, Van Winkle TJ, Drobatz K, et al. Ultrasonographic findings in cats with clinical, gross pathologic, and histologic evidence of acute pancreatic necrosis: 20 cases ( ). J Am Vet Med Assoc 2002;221: Ferreri JA, Hardam E, Kimmel SE, et al. Clinical differentiation of acute necrotizing from chronic nonsuppurative pancreatitis in cats: 63 cases ( ). J Am Vet Med Assoc 2003;223: Wall M, Biller DS, Schoning P, et al. Pancreatitis in a cat demonstrating pancreatic duct dilatation ultrasonographically. J Am Anim Hosp Assoc 2001;37: Hecht S, Henry G. Sonographic evaluation of the normal and abnormal pancreas. Clin Tech Small Anim Pract 2007;22: Texas A&M University College of Veterinary Medicine and Biomedical Sciences Department of Small Animal Clinical Sciences Gastrointestinal Laboratory. Pancreatic lipase immunoreactivity (PLI). Available at: vetmed.tamu.edu/gilab/service/assays/pli. Accessed Nov 4, Cohen J. A coefficient of agreement for nominal scales. Educ Psychol Meas 1960;20: Landis JR, Koch GG. The measurement of observer agreement for categorical data. Biometrics 1977;33: Feeney DA, Anderson KL, Ziegler LE, et al. Statistical relevance of ultrasonographic criteria in the assessment of diffuse liver disease in dogs and cats. Am J Vet Res 2008;69: Lamb CR. Pancreatic edema in dogs with hypoalbuminemia or portal hypertension. J Vet Intern Med 1999;13: SMALL ANIMALS/ JAVMA, Vol 244, No. 9, May 1, 2014 Scientific Reports 1065