Abstract
The safety of fine-needle aspiration (FNA) of the feline pancreas has not been reported. The incidence of complications following ultrasound-guided pancreatic FNA in 73 cats (pancreatic aspirate [PA] cats) with clinical and ultrasonographic evidence of pancreatic disease was compared with complications in two groups of matched control cats also diagnosed with pancreatic disease that either had abdominal organs other than the pancreas aspirated (control FNA, n = 63) or no aspirates performed (control no FNA, n = 61). The complication rate within 48 h of the ultrasound and/or aspirate procedure did not differ among the PA cats (11%), control FNA (14%) or control no FNA (8%) cats. There was no difference in rate of survival to discharge (82%, 84% and 83%, respectively) or length of hospital stay among groups. The cytologic recovery rate for the pancreatic samples was 67%. Correlation with histopathology, available in seven cases, was 86%. Pancreatic FNA in cats is a safe procedure requiring further investigation to establish diagnostic value.
Introduction
The clinical importance of feline pancreatic disease is increasingly being recognized. A pathologic classification has been proposed, including various histopathologic types of pancreatitis, nodular hyperplasia, neoplasia, pancreatic pseudocyst, abscess, amyloid deposition and pancreatic atrophy. 1 In addition to the broad spectrum of diseases, the clinical picture is further complicated by the potential for pancreatic disease to present concurrently with inflammatory bowel disease and/or cholangitis. 2 Given the significance of pancreatic diseases in cats, there is a pressing need for minimally invasive diagnostic tests to distinguish between inflammatory, hyperplastic and neoplastic lesions.
Ultrasound is often used in conjunction with clinical and laboratory findings to identify pancreatic disease. Abdominal ultrasonography is safe and has good specificity for the presence of pancreatitis, but has suboptimal sensitivity, reportedly 30–80% in cats.3–5 Additionally, ultrasonography cannot reliably differentiate between different pathologic processes in the pancreas.5–7 Nodular pancreatic change raises concern for neoplasia, but is not specific, 6 and may also occur with inflammatory disease.8,9
Obtaining samples of abdominal organs by ultrasound-guided fine-needle aspiration (FNA) for cytologic evaluation is a reasonably safe and accurate diagnostic modality that often augments imaging studies. 10 In human medicine, percutaneous pancreatic FNA has a complication rate of 1.5–3.0%.9,11–13 Diagnostic performance is good, with a reported sensitivity of 92.5% and specificity of 68–100% for neoplasia.11,13 Few studies have addressed pancreatic FNA in cats.5,6,14 Although no complications have been reported in the 17 cats evaluated in these studies, speculation about potential risks following pancreatic sampling may contribute to the reluctance to perform pancreatic FNA.
The primary objective of this study was to evaluate the safety of FNA of the feline pancreas with clinical and ultrasonographic evidence of pancreatic disease. Secondary objectives were to evaluate the diagnostic yield of samples obtained by ultrasound-guided pancreatic FNA and the type of diagnostic information obtained. We hypothesized that FNA of the feline pancreas would be a safe procedure warranting further studies on diagnostic value.
Materials and methods
Case selection and matching with controls
The electronic medical record system at the University of Minnesota Veterinary Medical Center was searched for cats that had pancreatic cytology performed between September 2004 and September 2011. Inclusion criteria for pancreatic aspirate (PA) cats consisted of (1) abdominal ultrasound performed by a board-certified veterinary radiologist in cats with clinicopathologic changes compatible with pancreatic disease (anorexia, vomiting, diarrhea or chemistry changes); (2) ultrasonographically identified pancreatic lesions; (3) FNA of the pancreas performed with ultrasound guidance; (4) successful pancreatic aspirate confirmed during imaging or by cytologic evaluation; and (5) a complete medical record with hospitalization for a minimum of 48 h postultrasound and/or aspirate procedure. Cases were excluded if the required information was not available.
The population from which control cats were matched to PA cats was identified by searching the ultrasound log from September 2004 to September 2011 for consecutive feline cases having ultrasonographic abnormalities of the pancreas or peripancreatic tissue, with no pancreatic FNA performed. Inclusion criteria 1, 2 and 5 were otherwise identical to PA cats. Two control groups were formed: control FNA cats had ultrasound-guided FNA or biopsies of abdominal organs other than the pancreas, and control no FNA cats did not have any sampling of intra-abdominal organs or tissues performed during the ultrasound procedure. We attempted to match PA cats to one control from each group. PA cats and controls were first matched based on age group (0–8 years, 9–15 years and >16 years of age) and date of ultrasound (within 6 months of the PA case). To attempt to control for illness severity, controls were then matched with PA cases based on location of hospitalization (intensive care unit or general wards), and, finally, on type of pancreatic and peripancreatic tissue changes described in ultrasound reports. Cases and controls were matched first on pancreatic size (normal, or mildly, moderately or severely enlarged), echogenicity (hyperechoic, hypoechoic, mixed or normal) and margin irregularity, then on presence or absence of masses or nodules, and peritoneal effusion (scant, moderate, marked). Ultrasound reports for cats in all groups were evaluated and the following information about the pancreas was recorded: pancreatic size, margin irregularity, echogenicity, and the presence or absence of masses or nodular lesions (aggregated owing to the subjective difference between masses and nodules). Mesenteric and hepatic echogenicity changes were also recorded.
Data collection
Signalment, historical illnesses, pre-existing diagnoses, serum liver enzyme activities, and serum bilirubin and creatinine concentrations prior to imaging were recorded for all cats in all groups. Length of hospital stay, discharge status (alive, died or euthanized/discharged to be euthanized within 24 h) and clinical diagnosis listed by the attending clinician or from record review were noted. Records were evaluated to determine if any of the following, designated as ‘complications’, occurred within 48 h of the abdominal ultrasound: hemoabdomen, hypotension, need for corrective clinical action (transfusion, diagnostic abdominocentesis or peritoneal lavage) or respiratory distress. We also recorded if suspicion of any of the above complications led to recheck abdominal ultrasound within 48 h of the original procedure.
Pancreatic FNA procedure and cytologic classification
Aspirates were obtained with a 20 or 22 G hypodermic or spinal needle of an adequate length to reach the pancreas or other organ or tissue using either aspiration or fenestration techniques, or a combination of both. 8 This is the procedure typically used for intra-abdominal aspirates of solid tissues. Multiple FNA attempts, usually two or three, are standardly made for tissue sampling in our hospital; one attempt is typically performed for fluid. Pancreatic cytology reports generated by a board-certified veterinary clinical pathologist were reviewed, and the cellularity of the cytologic samples and cytologic diagnoses for the pancreas noted. Cytologic recovery rate was defined as the percentage of samples from the total submitted that were diagnostic. Cytologic diagnoses were categorized as normal exocrine tissue, cyst, necrosis/inflammation, hyperplasia, neoplasia or non-diagnostic. For a diagnosis of neoplasia, the cell of origin or specific tumor type was recorded, if determined. Histopathologic diagnosis from surgical or necropsy samples was recorded when available, and the interval between cytologic and histopathologic evaluation recorded.
Statistical analysis
The χ2 test of association was used to assess differences among groups in sex, frequency of complications, location of hospitalization and discharge status. Cats that died or were discharged to be euthanized were combined for analysis. Factors used for matching cases to controls were not evaluated for differences, except for ultrasonographic features of the pancreas, as matching for these criteria was incomplete. The data were tested for normality, and ANOVA was used to evaluate differences among the groups in preultrasound serum liver enzyme activities, bilirubin, creatinine and frequency of concurrent diseases. Differences in length of hospital stay were evaluated using a linear mixed model with group (PA, control FNA or control no FNA), location of hospitalization, age group and discharge status as fixed effects and the matched cats as a random effect. P values <0.05 were considered significant.
Results
Study population
Ninety-five possible PA cats were identified; 22 were excluded owing to the inability to confirm pancreatic aspiration. Of the 73 remaining PA cats, 51 were matched to two controls (21 with both controls from the same control group and nine with one from each control group). Twenty-two PA cats had a single control. Of the 162 cats identified as potential controls, 38 were excluded for inability to be satisfactorily matched to a PA cat.
Overall, 197 cats were included in this study, of which 77 were spayed females and 120 were neutered males. The mean age was 12.2 years (range 3–19 years). The PA group (n = 73) consisted of 28 spayed females and 45 neutered males; mean age was 13.4 years (range 3–18 years). Of 63 control FNA cats, 20 were spayed females and 43 were neutered males. The mean age was 12.0 years (range 6–18 years). Of 61 control no FNA cats, 29 were spayed females and 32 were neutered males. The mean age was 11.1 years (range 3–19). Sex distribution did not differ among groups (P = 0.24) (Table 1).
Distribution of cats in each group in terms of location of hospitalization (intensive care unit [ICU] vs general wards), discharge status and preultrasound chemistry values expressed as median (95% confidence interval for the mean) in each group
Values are given as n (%) unless otherwise stated
Control no FNA cats were more likely to be hospitalized in the wards than in the ICU (P <0.01). There was no significant difference in the discharge status among cats
Control FNA cats had a higher total bilirubin than the other two groups (P <0.01)
PA = pancreatic aspirate; FNA = fine-needle aspirate; ALP = alkaline phosphatase; ALT = alanine transaminase; GGT = gamma-glutamyl transferase
Hospitalization
The means for length of hospital stay were as follows: PA 2.77 days (range 1–7 days); control FNA 2.49 days (range 1–11 days); control no FNA 3.00 days (range 1–9 days), with no significant differences among the three groups (P = 0.70). Control no FNA cats were three times as likely to be hospitalized in the general wards than the intensive care unit (P <0.01), while the other groups were more evenly distributed between locations (Table 1).
Biochemical analyses
There were no significant differences in serum activities of alanine transaminase (P = 0.17), alkaline phosphatase (P = 0.68), gamma-glutamyl transferase (P = 0.69) or serum creatinine concentration (P = 0.15) (Table 1). The control FNA group had a higher mean bilirubin than the PA and control no FNA cats (P <0.01).
Concurrent and pre-existing diagnoses
In the PA group, 50 cats (70%) were diagnosed with active pancreatitis, 13 (18%) with hepatic lipidosis and four (6%) with cholangitis. All of the cats with a diagnosis of hepatic lipidosis had a liver aspirate performed. In the control FNA cats, 22 (35%) were diagnosed with pancreatitis, nine (14%) with hepatic lipidosis and four (6%) with cholangitis. In the control no FNA cats, 24 (39%) were diagnosed with pancreatitis, four (7%) with hepatic lipidosis and two (3%) with cholangitis.
In the PA group, 23 (32%) had no pre-existing diseases recorded in their medical records, 26 (36%) had one, 14 (19%) had two, seven (10%) had three, and three (4%) had four pre-existing diseases. Of control FNA cats, 29 (46%) had no pre-existing diseases, 27 (43%) had one, six (10%) had two and one had four (2%). Of control no FNA cats, 27 (44%) had none, 25 (41%) had one, six (10%) had two, two (3%) had three and one (2%) had four pre-existing diseases. The most common diseases were diabetes mellitus, present in 10/73 (14%) PA cats, 3/63 (5%) control FNA cats and 8/61 (18%) control no FNA cats; hyperthyroidism in 7/73 (10%), 4/63 (6%) and 3/61 (6%), respectively; and chronic kidney disease in 14/73 (19%), 13/63 (21%) and 10/61 (16%), respectively. There was no significant difference in the frequency of pre-existing or concurrent diagnoses among the groups (P = 0.60).
Ultrasound and pathology results
The most common ultrasonographic abnormalities of the pancreas observed are reported in Table 2. All cats had multiple ultrasonographic abnormalities. Despite attempts to match on all major ultrasonographic findings, the PA cats were significantly more likely to have nodular or mass-like lesions (P <0.01) than cats in either control group. There were no differences in the prevalence of hepatic echogenicity changes among the three groups of cats (P = 0.70). Of PA cats, 12 had pancreatic aspirates only, while 61 had multiple organs sampled, including the liver (n = 40), spleen (n = 9), abdominal lymph nodes (n = 14), mesentery (n = 10), kidney (n = 5), cystocentesis (n = 2) and abdominal effusion (n = 15). One PA cat had an ultrasound-guided liver biopsy in addition to pancreatic FNA. Of control FNA cats, 39 had one organ aspirated, 22 had two, and one cat each had aspirates of three and four organs. Organs sampled included the liver (n = 40), spleen (n = 9), intestine (n = 3), abdominal lymph nodes (n = 6), mesentery (n = 7), kidney (n = 2), cystocentesis (n = 4) and abdominal effusion (n = 15). Four cats also had liver biopsies performed.
Prevalence of the most common ultrasonographic abnormalities among cats within each study group
Values given as n (%)
PA cats were significantly more likely to have mass lesions or nodular change to their pancreas (P <0.01)
PA = pancreatic aspirate; FNA = fine-needle aspirate
Twenty-four of 73 cytologic samples of the pancreas were non-diagnostic for a cytologic recovery rate of 67%. Seventeen cats had two pathologic processes noted. The cytologic distribution of pathologic processes is reported in Table 3. Nine cats also had pancreatic histopathology performed 2 days to 1 year postaspiration (one surgical, eight at necropsy). Histopathologic and cytologic diagnoses are indicated in Table 4. Pancreatic cytology and histopathology were in agreement in 7/9 cases.
Distribution of pathologic pancreatic cytology diagnoses from pancreatic aspirate cats
Agreement between fine-needle aspirate pancreatic samples and pancreatic histopathology
Complications
There was no difference among groups in complication incidence (P = 0.90) or in the number of cats with complications that survived to discharge (Tables 1 and 5). Eight (11%) PA cats, nine (14%) control FNA cats and five (8%) control no FNA cats developed complications within 48 h of their pancreatic aspirate procedure or ultrasound. Complications occurred in 7/62 PA cats that had the pancreas and other tissues sampled (three liver, three peritoneal effusion, and one both), and in 1/11 that had only pancreatic aspiration. Two PA cats, one control FNA cat and two control no FNA cats had recheck ultrasound exams while hospitalized to evaluate for complications. Overall, 3/8 (37%) of the PA cats and 5/9 (56%) of the control FNA cats experiencing complications survived to discharge compared with 57/65 (88%) and 49/54 (91%), respectively, without clinical complications. There was a 40% survival rate (2/5) in the control no FNA cats developing complications after ultrasound compared with 48/56 (86%) without complications.
Complications encountered within 48 h of abdominal ultrasound, listed as number of cats and percentage of cats in each group
OPK Biotech
Ten complications occurred in eight cats. One cat in this group experienced hypotension and received whole blood, and a second cat developed hemoabdomen and received a packed red blood cell transfusion
Discussion
This study failed to demonstrate any increase in complication or mortality rates in cats undergoing pancreatic aspiration compared with those that did not. The incidence of complications in our study for cats undergoing FNA of any intra-abdominal organ is higher than that in one previous report (about 5%). 8 We evaluated changes in clinical status up to 48 h postultrasound to ensure that we accounted for all potential complications. Many of the complications we report are unlikely to be due to the aspiration procedure. This is supported by the lack of difference in complication rate experienced by cats that did or did not have aspiration procedure(s) performed. The majority of complications in the PA cats were noted when a second organ, typically the liver, was aspirated. The major complications encountered in PA and control FNA cats were hemorrhage and hypotension, which together occurred in 3% (PA cats) and 6% (control FNA) of the cats.
Other factors that may account for the prevalence of complications in these cats include concurrent pathology or diagnoses aside from their pancreatic changes. Diabetes mellitus, hyperthyroidism and chronic kidney disease were the most common pre-existing diagnoses; these were not correlated to the prevalence of complications owing to the overall low numbers of each complication encountered. The underlying causes for the higher complication rates in cats that had liver aspirates are unknown; the retrospective nature of the study precluded complete diagnostic evaluation of each patient. The presence of other subclinical or undiagnosed conditions cannot be excluded, particularly liver disease, which requires biopsy to fully exclude. In both groups, roughly equal numbers of cats had hepatic aspirates performed, and similar numbers of cats in both the PA group and the control FNA group had a diagnosis of hepatic lipidosis. Ultrasonographic evidence of cholangitis was found in only a few cats, but this does not exclude this as a possible undiagnosed condition in some patients. The retrospective nature of the study limits the certainty with which concurrent conditions can be excluded. Additionally, coagulation testing was not performed prior to ultrasound in all cases that experienced complications.
The control no FNA cats were more likely to be hospitalized in the general wards than the intensive care unit, which suggests that they may not have been as critically ill as the cats in the other groups. We attempted to match for hospitalization between groups, but this was difficult owing to the numerous matching criteria. However, their complication rate was not different from the rate of the other groups. This supports that the complication risk is not greater for more critically ill cats.
The control FNA group cats had higher serum bilirubin concentrations relative to the other groups of cats. This may reflect a propensity for this group of cats to have primary hepatic or biliary pathology, likely resulting in a decision to aspirate the liver rather than the pancreas, despite ultrasonographic pancreatic abnormalities. PA cats had a higher proportion of pancreatic nodules, which may have led to clinician bias towards aspirating the pancreas owing to suspicion of neoplasia. One previous study demonstrated significant overlap in ultrasonographic abnormalities between cats with pancreatic neoplasia and nodular hyperplasia; the only finding unique to neoplasia was a single pancreatic nodule or mass >2 cm. 9
The pancreatic cytologic recovery rate of 67% is slightly lower than the 86% recovery that is reported for aspiration of abdominal masses in dogs and cats. 15 There was good agreement with histopathology in the small number of cases available for comparison, suggesting that additional evaluation of diagnostic accuracy should be pursued in future studies.16,17
There are several limitations to this study. As a retrospective study, we were unable to control for variations in clinician bias for or against pancreatic aspiration, clinical pathologist and radiologist interpretation, radiologist aspiration technique, and client permission for FNA to be performed. Ideally, each PA cat would have been matched to one cat from each control group; this was not possible based on the multiple matching criteria established. There was also some variation in matching by ultrasound lesions, as each cat had multiple lesions and it was not possible to match on all of them.
Conclusions
Our data support previous smaller studies demonstrating that ultrasound-guided aspiration of the feline pancreas is a safe diagnostic procedure. The combination of acceptable risk and good diagnostic yield, with preliminary indications of concordance with histopathology in a limited number of cases, suggests that future investigations are warranted to evaluate diagnostic applications in cats.
Footnotes
Conflict of interest
The authors do not have any potential conflicts of interest to declare.
Funding
This research received no grant from any funding agency in the public, commercial or not-for-profit sectors
Work performed at the University of Minnesota, College of Veterinary Medicine, Veterinary Medical Center, Veterinary Diagnostic Laboratory, and Veterinary Clinical Sciences Department, St Paul, Minnesota. This research was presented, in part, as an abstract at the Annual Forum of the American College of Veterinary Internal Medicine, 2013.