Biochemical Analysis of Neoplastic Versus Nonneoplastic Abdominal Effusions in Dogs

Introduction

Abdominal effusion is a common clinical problem in animals. Some causes of abdominal effusion in dogs include portal hypertension, hypoproteinemia, abdominal neoplasia, peritonitis, coagulopathies, trauma, and obstructive lymphatic drainage.¹ Determining the cause of abdominal effusion often involves expensive and invasive diagnostic procedures, and the various causes of abdominal effusion have very different prognoses.

In people, biochemical parameters in abdominal fluid such as cholesterol, albumin, and pH have been helpful in determining the cause of the effusion, particularly when distinguishing malignant effusion from effusion associated with cirrhosis or tuberculosis.^2–6 A study by Bansal et al. found abdominal fluid pH to be lower in malignant and tubercular effusions when compared to sterile cirrhotic ascites.² They also found that albumin was lower in sterile cirrhotic ascites when compared to malignant and tubercular effusions. Attali et al. found that abdominal fluid pH was low in malignant effusions or effusions associated with pancreatitis and tuberculosis, but not in ascites of cardiac origin.³ A higher abdominal fluid albumin level in malignant and tubercular effusions versus cirrhotic ascites was found by Gupta et al.⁴ In that study, an albumin level >2 mg/dL was more consistent with malignant or tubercular effusions. A study by Castaldo et al. found higher cholesterol levels in effusions associated with peritoneal malignancies than in abdominal fluid secondary to cirrhosis or hepatocellular carcinoma.⁵ Sood et al. also evaluated cholesterol in abdominal fluid and found higher values in cases of malignancy than in cases of tuberculosis.⁶ In this latter study, a cholesterol level of >54.5 mg/dL was more consistent with malignancy.

Studies of abdominal effusion in dogs have evaluated the value of serum-to-effusion albumin gradients for transudative effusions and fluid-to-blood creatinine and potassium ratios for cases of uroperitoneum.⁷⁸ One study of pericardial fluid determined that the pH of the fluid helped differentiate neoplastic from nonneoplastic effusions.⁹ In that study, neoplastic effusions were more alkaline (pH 7.0 to 7.5), while inflammatory diseases had a more acidic pH (pH 6.5 to 7.0). This finding was disputed by Fine et al. who did not find a statistical difference in the pH of various types of pericardial effusions.¹⁰ Stewart et al. reported that malignant pleural effusions in cats typically had a pH >7.2.¹¹

Studies in dogs are lacking that evaluate biochemical tests performed on abdominal fluid as an aid in differentiating neoplastic from nonneoplastic forms of effusion. The purpose of this study was to assess a variety of biochemical parameters, including pH, in both neoplastic and nonneoplastic abdominal effusions, to determine if any of the parameters were markers of neoplasia.

Materials and Methods

Dogs that were presented for, or were diagnosed with, abdominal effusion at the University of Illinois Veterinary Teaching Hospital (UI-VTH) between October 2001 and April 2002 were considered for the study and prospectively evaluated. Criteria for inclusion in the study included a determination of the cause of the ascites through cytology (n=4), histopathology (surgical biopsy, n=5), ultrasonography (n=1), or echocardiography (n=4). Dogs that had a diagnostic peritoneal lavage performed 48 hours prior to presentation were not eligible for the study.

Abdominal fluid (3 mL) was collected from each dog by paracentesis under ultrasound guidance. Immediately after collection of the abdominal fluid, a venous blood sample was collected from a cephalic vein into a preheparinized blood gas syringe.^a Both abdominal fluid and venous blood samples were kept at room temperature (22°C) and were analyzed within 10 minutes of sample collection. A blood gas analysis was performed on the venous blood sample. Fluid pH and lactate were determined on the abdominal sample using a point of care chemical analyzer.^b Biochemical analysis was performed on the abdominal fluid using an automated analyzer.^c Biochemical parameters that were measured included urea nitrogen, creatinine, alkaline phosphatase (ALP), alanine aminotransferase (ALT), gamma glutamyl transpeptidase (GGT), total bilirubin, glucose, amylase, and lipase. Total protein and specific gravity were also determined using a refractometer. A hematology analyzer^d was used to determine a nucleated cell count and a red blood cell count on the abdominal fluid.

Results

Fifteen dogs qualified for the study. Seven had effusions associated with nonneoplastic causes. Diagnoses in these seven dogs included heart failure (n=4), portal hypertension secondary to nonneoplastic hepatic disease (n=2), and pancreatitis (n=1). Both cases of hepatic disease were diagnosed by histopathology of surgically obtained biopsy samples. The average age of dogs in the nonneoplastic group was 8.4 years, with a range of 2 to 14 years. The breeds in the nonneoplastic group included the Labrador retriever (n=2), mixed-breed dogs (n=2), and one each of the Akita, Doberman pinscher, and German shepherd dog.

Eight dogs had abdominal effusion secondary to neoplasia. Diagnoses included hemangiosarcoma (n=3), carcinomatosis (n=2), and one each of a neuroendocrine tumor, metastatic adenocarcinoma of the liver, and pancreatic carcinoma. For the metastatic carcinoma in the liver, a primary tumor was not identified. The average age of dogs in the neoplastic group was 12.1 years, with a range of 8 to 17 years. Breeds in the neoplastic group included mixed-breed dogs (n=2) and one each of the American cocker spaniel, golden retriever, German shepherd dog, rottweiler, Scottish terrier, and basset hound.

All four dogs in which cytology provided the diagnosis were in the neoplastic group (i.e., neuroendocrine tumor [n=1], hemangiosarcoma [n=1], and carcinomatosis [n=2]). In the dog with pancreatitis in the nonneoplastic group, the diagnosis was based on suggestive clinical signs, a painful abdomen, an inflammatory leukogram, ultrasonographic findings consistent with pancreatitis, and response to treatment. The dog also had an elevated abdominal fluid lipase, which has been used as a diagnostic aid for pancreatitis in humans.¹²¹³

The Table lists the mean ± standard deviations for the cell counts, fluid specific gravity, and biochemical parameters for both the neoplastic and nonneoplastic effusions. Statistically significant differences were found in fluid lactate (P=0.0377) and glucose (P=0.0431). A difference was also noted in the fluid pH, but the result was not statistically significant. The differences between venous blood pH and abdominal fluid pH were statistically significant for the nonneoplastic group (P=0.011) but not for the neoplastic group.

Discussion

Despite the large number of variables evaluated, only the differences between the lactate and glucose levels in the abdominal fluid samples were statistically significant. The preferential use of glucose by neoplastic cells and increased metabolism by those cells may have accounted for the decreased glucose seen in the neoplastic effusions.¹⁴ One study in humans also documented lower glucose levels in neoplastic effusions.¹⁵ In that particular study, fluid glucose was lower in people with liver metastases when compared to cases that had ascites secondary to cirrhosis.¹⁵

Neoplastic cells have been shown to use anaerobic glycolysis for energy, thereby producing increased levels of lactate.¹⁴ This increased production of lactate may explain the higher levels of lactate observed in the neoplastic effusions in the study reported here.

While there was a difference in the pH of abdominal fluid between the two groups, this difference was not statistically significant. The pH was lower in the neoplastic effusions of this study, which correlated with the findings of Bansal et al. and Attali et al.²³ The results reported here, however, conflicted with the previous findings in dogs and cats for both pericardial and pleural effusions.⁹¹¹ Because neoplastic cells increase the production of lactate, lower pH would be expected in neoplastic effusions when compared to nonneoplastic effusions. Further studies involving a larger study population might detect a difference in the pH between neoplastic and nonneoplastic effusions. In the study reported here, the difference between fluid pH and blood pH was statistically significant for the dogs in the nonneoplastic group, which was similar to the findings of Edwards.⁹

Conclusion

The findings of this study indicated that the measurement of abdominal fluid levels of lactate and glucose might be beneficial in determining whether the effusion is from neoplastic or nonneoplastic causes. Further studies with larger populations are needed to validate these findings and to establish specific diagnostic values for each assay. It may also be beneficial to evaluate other biochemical parameters such as cholesterol, which has proven value in humans with neoplastic abdominal effusions.^4–6 Although biochemical tests cannot be recommended as the only tests used to differentiate malignant and nonmalignant effusions, further studies are indicated to determine their clinical applicability.