Abdominal Fluid

Place the patient in left lateral recumbency and restrain. Clip and surgically prepare an area (e.g., 4 to 6 inches square) with the umbilicus in the center. The urinary bladder should be emptied before performing paracentesis. Infiltrate a small area with local anesthetic, if desired. Use a 20-to 22-gauge needle or over-the-needle catheter to penetrate the abdomen. Attempt to obtain fluid in four quadrants, allowing the fluid to flow freely by gravity and capillary action. If needed, gentle suction with a 3- or 6-ml syringe can be employed. For a complete description of the technique, the reader is referred elsewhere (Ford and Mazzaferro, 2006). Allow the animal to rest quietly while fluid is being removed. Moving the animal or allowing the patient to move while the needle is in the abdomen can result in laceration of the organs. Some investigators prefer to have the patient standing for fluid removal; however, it is more likely that the omentum will occlude the needle if the patient is in a standing position.

Pleural Fluid

For removal of fluid from the thorax, the patient should be in a standing or in ventral/sternal recumbency. Clip the hair and surgically prepare the thoracic wall from the 5th to the 11th intercostal space. Infiltrate a small area at the 7th to 8th intercostal space at the level of the costochondral junction with local anesthetic. It is best to attach extension tubing to the hub of the needle or over-the-needle catheter and a three-way stopcock for removal of pleural fluid. Insert the needle or catheter into the chest wall at the surgically prepared site taking care to avoid the intercostal vessels located just caudal to each rib. For a complete description of this technique, the reader is referred elsewhere (Ford and Mazzaferro, 2006).

Pericardial Fluid

For removal of fluid from the pericardial sac, sedate the patient if necessary. Surgically prepare an area over the lower to mid 5th to 7th intercostal space bilaterally. Place the patient in lateral or sternal recumbency. Attach ECG leads to monitor for dysrhythmias during the procedure. Infiltrate an area at the costochondral junction, or approximately where the lower and mid thorax meet, with local anesthetic. Use an over-the-needle catheter or Intrafusor system with a three-way valve to which a 30-ml syringe is attached. Always maintain negative pressure on the syringe as the chest wall is punctured. Carefully advance the needle into the 4th intercostal space through a nick incision in the direction of the heart. Advance the needle until resistance is met (from the pericardium). A release will be felt as the needle enters the pericardial sac and a flash of blood is often seen. Thread the tubing or catheter so that it is securely within the pericardial sac. For a complete description of this technique, the reader is referred elsewhere (Ford and Mazzaferro, 2006).

Protein Quantitation

Protein quantitation is typically done via refractometry; however, some institutions will determine protein via spectrophotometry or automated analysis. Both methods offer accurate readings in a wide range of protein concentrations (George, 2001; George and O’Neill, 2001). It has been shown with canine effusions that refractometry underestimates the protein content when < 2.0 g/dl and that the spectrophotometry using the biuret method is more accurate when there is high protein content (Braun et al., 2001). Others have found that refractometry can be used accurately down to 1.0 g/dl (George and O’Neill, 2001). A similar finding of underestimating protein content in feline effusions with refractometry may occur (Papasouliotis et al., 2002). In that same study, a dry chemistry analyzer produced increased globulin concentration and therefore lower Albumin:Globulin (A:G) ratios when compared with a reference wet analyzer using the same biuret and bromocresol green methodologies (Papasouliotis et al., 2002). This finding is particularly important since decreased A:G ratios support a diagnosis of feline infectious peritonitis (Hartmann et al., 2003).

For cloudy or turbid samples and bloody samples, the fluid should be centrifuged and the protein measured on the supernatant. Turbidity may interfere with evaluation of protein by either refractometry or spectrophotometry. The protein content is used with the nucleated cell count to classify the effusion and help formulate a list of possible causes.

Red Blood Cell and Total Nucleated Cell Count

Although an initial impression of the cellularity and amount of blood can usually be made by visual inspection of the sample (see Fig. 6-1), knowledge of the actual cell counts for erythrocytes and nucleated cells is important for further classification of the type of fluid. With this information, one can begin to narrow down the list of possible causes for the abnormal fluid accumulation. For samples being submitted to a reference laboratory, placing some of the sample in a lavender-top tube (EDTA) and some in a red-top tube is recommended. The lavender-top tube contains anticoagulant, which prevents the sample from clotting if there is a high protein content. EDTA is bactericidal, however, and is contraindicated if a sample is to be cultured (Songer and Post, 2005). Thus, some of the fluid should also be put into a red-top tube or sterile tube without additives. The cell counts will be done either with a hemocytometer or an automated cell-counting instrument. If the amount of fibrinogen in the fluid sample is high, then the sample in the red-top tube is likely to clot, producing erroneous results.

Note: Do not use gel-containing serum separator tubes (SST) for submission of fluid to a reference laboratory. Cells may bind to the gel in these tubes and result in an artifactually lowered cell count.

Nucleated Cell Differential

Standard procedures for performing a differential of the nucleated cells vary among laboratories. Some laboratories do no differential, others a three-part differential of 100 cells (large mononuclear cells, small mononuclear cells, and neutrophils), while others will provide a 100-cell differential of all cell types observed. The differential provides a relative picture of the types and numbers of cells and aids in establishing a list of potential causes for the fluid accumulation. A differential is not a substitute for a cytologic evaluation. The cytologic evaluation is performed in an attempt to determine a specific diagnosis.

Mesothelial Cells

In most cases, the cytologist will find reactive mesothelial cells in body cavity fluids. These are considered as large mononuclear cells for the purpose of the three-part cell differential. Mesothelial cells may be seen as individualized cells or in variably sized clusters. They contain a moderate amount of medium-blue cytoplasm (Fig. 6-2). Hyperplastic mesothelial cells are large (12 to 30 μm) with deep-blue cytoplasm and may display a pink to red “fringed” cytoplasmic border (Fig. 6-3A&B). This feature helps identify these cells as mesothelial cells. These cells may contain one or more nuclei of equal size (see Fig. 6-3B). Nucleoli may be visible and occasional mitotic figures may be evident.

FIGURE 6-2 Normal mesothelial cell. Exfoliated cell in an effusion with its characteristic pink fringe along the cytoplasmic border. (Wright-Giemsa; HP oil.)

(From Meyer DJ, Franks PT: Classification and cytologic examination, Compend Contin Educ Pract Vet 9:123–29, 1987.)

FIGURE 6-3 Reactive mesothelial cell. A, Exfoliated binucleate mesothelial cell (upper right) and mildly vacuolated and basophilic macrophage in an effusion. The mesothelial cell has a characteristic pink fringe along the cytoplasmic border. (Modified Wright; HP oil.) B, A loose group of variably reactive mesothlelial cells at the feathered edge of a smear made from an effusion. These cells may contain one or more nuclei. Note the presence of the “fringe” (glycocalyx) on the mesothelial cells. Several cells contain paranuclear dark granules, the significance of which is unknown. (Modified Wright; HP oil.)

Macrophages

Macrophages are large mononuclear cells with abundant pale-gray to light-blue cytoplasm and a round to kidney bean–shaped nucleus (Figs. 6-3A and 6-4). The chromatin may be fine and nucleoli may be visible. Macrophages often contain vacuoles or previously phagocytosed cells and/or debris if there is inflammation or if the fluid has been present for a long time (Fig. 6-5). Macrophages are considered as large mononuclear cells for the purpose of the three-part cell differential.

FIGURE 6-4 Macrophages. Three unremarkable to mildly basophilic and vacuolated macrophages from an effusion are present. (Modified Wright; HP oil.)

FIGURE 6-5 Macrophage. Neutrophils. Shown are a moderately vacuolated and basophilic macrophage and two nondegenerate neutrophils. (Modified Wright; HP oil.)

Lymphoid Cells

Small and medium lymphocytes found in effusions appear similar to those found in peripheral blood. These nucleated cells often have a thin rim of lightly basophilic cytoplasm and a round nucleus. Lymphocytes are considered as small, mononuclear cells for the purpose of the three-part cell differential. In normal fluids they are present in higher proportions in cats and cattle than in dogs and horses. The nucleus nearly fills the cell, producing a uniformly high nuclear-to-cytoplasmic (N:C) ratio. The chromatin is finely stippled to evenly clumped; nucleoli are not visible (Fig. 6-6A&B).

FIGURE 6-6 A, Normal fluid/transudate. Note the macrophage and small lymphocyte with several erythrocytes. Normal fluid and transudates contain very low nucleated cell counts (<1000/μl) and low protein content (<2.5 g/dl). (Romanowsky; HP oil.) B, Modified transudate. Pleural. Cat. The effusion cells have been concentrated in this animal with cardiomyopathy. Note a large macrophage, many small lymphocytes, and nondegenerate neutrophils. This fluid had a protein of 2.5 g/dl and an increased nucleated cell count of 4000 cells/μl. The macrophage has phagocytized a red blood cell. (Romanowsky; HP oil.)

Neutrophils

Neutrophils appear similar to those found in peripheral blood. They are medium-sized cells with pale to clear cytoplasm and a segmented nucleus. Neutrophils should be absent or present in very low numbers in normal fluid but they will be found in increased numbers with chronic fluid accumulation or with inflammation (see Figs. 6-5 and 6-6B).

Transudate

Fluids are classified as transudates when they have a low protein content and a low cell count (protein < 2.5 g/dl and cells < 1000/μl). These fluids increase in volume in response to physiologic mechanisms, such as increased hydrostatic vascular pressure or decreased colloidal osmotic pressure, that cause the normal homeostatic mechanisms of fluid production and resorption to be overwhelmed. Some causes for transudate accumulation include severe hypoalbuminemia, portal hypertension, hepatic insufficiency, portosystemic shunt, and early myocardial insufficiency. The cells commonly found in transudates are similar to those in normal fluid, which are mostly mononuclear cells consisting of macrophages, small lymphocytes, and mesothelial cells (see Fig. 6-6A). Neutrophils may compose a small proportion of the population.

Modified Transudate

A fluid is classified as a modified transudate when a transudate changes its physical features. The accumulation of transudative fluid in a body cavity causes increased pressure, which is irritating to the mesothelial cells lining the space. They respond by proliferating and sloughing into the effusion. With time, the sloughed mesothelial cells die and in so doing release chemoattractants that draw small numbers of phagocytes into the effusion to remove cellular debris. The result is a mild increase in both total protein (≥2.5 g/dl) and nucleated cell count (less than 5,000/μl). Thus, modified transudates are generally transudates that have been present long enough to elicit a mild inflammatory reaction (see Fig. 6-6B). It is most often associated with cardiovascular disease or neoplastic conditions.

In cases of extended duration, modified transudates can have a cloudy to almost milky gross appearance. Such fluids strongly resemble chyle and, in fact, have in the past been called “pseudochylous effusions” (a term no longer used). The gross appearance of these fluids is the result of high lipid content (due to higher cholesterol content than serum) but is in no way related to a true chylous effusion, as there are no triglycerides or chylomicrons present (Hillerdal, 1997; Tyler and Cowell, 1989). The phagocytes attracted to remove cellular debris from transudates are rich in enzymes that digest protein but are virtually devoid of enzymes that will break down complex lipids. Consequently, while most of the constituents of dying cells are removed by phagocytosis, lipid content of the cells simply accumulates in the effusion. Effusions formed in this manner are easily distinguished cytologically from true chylous effusions.

The principle cellular constituent of the modified transudate is the reactive mesothelial cell (see Fig. 6-3A). Because of the ability of mesothelial cells to respond to irritation by proliferation, the presence of increased numbers of mesothelial cell clusters and rafts is a common finding in reactivity (see Fig. 6-3B). Mitoses are increased and occasional multinucleated reactive mesothelial cells are seen. Reactive mesothelial cells in clusters are capable of imbibing lipid from the effusion fluid and when they do, they take on the characteristics of secretory cells. In this form they must be differentiated from metastatic adenocarcinoma or mesothelioma. This may be done by critically evaluating the cell populations for criteria of malignancy.

As modified transudates mature, the proportion of inflammatory cells they contain will increase. In most cases the principal inflammatory cell is the nondegenerate neutrophil, but neutrophils rarely account for more than 30% of the total cell population. Over time, modified transudates gradually become cytologically indistinguishable from nonspecific exudates. One method used to distinguish modified transudates from exudates or transudates from exudates is measurement of C-reactive protein concentration in canine effusions (Parra et al., 2006).

Exudate

Exudates are the result of either increased vascular permeability secondary to inflammation or vessel injury/leakage (hemorrhagic effusion, chylous effusion). An exudative fluid usually contains both increased protein and an increased nucleated cell count. The total protein concentration is usually greater than 3.0 g/dl along with cell counts greater than 5,000/μl. Infectious causes for exudates include bacteria (Fig. 6-7A&B), fungi, viruses, protozoa such as Toxoplasma (Toomey et al., 1995), or helminthes such as Mesocestoides sp. (Caruso et al., 2003). Noninfectious causes involve organ inflammation such as pancreatitis, steatitis, inflammatory neoplasia and irritants such as bile or urine. Cytologic evaluation is useful to determine an underlying cause in cases of exudative effusions.

FIGURE 6-7 A, Septic peritonitis, dog. Numerous degenerate neutrophils are noted with a large number of pleomorphic bacteria seen, both in the background and within neutrophils. This concentrated smear was made from a dog with a ruptured pancreatic abscess. (Modified Wright; HP oil.) B, Septic exudate. Pleural. Cat. Degenerate neutrophils are bloated, with foamy or vacuolated cytoplasm and also swollen lytic nuclei. Note presence of small gram-positive, pleomorphic bacterial rods. Aerobic and anaerobic cultures are recommended in cases of pyothorax. This sample contains many nucleated cells (>100,000 cells/μl) and >3.0 g/dl protein. (Gram; HP oil.)

Inflammatory effusions are classified according to the standard rules for inflammation as neutrophilic, mixed, or macrophagic. In neutrophilic reactions, neutrophils (either nondegenerate or degenerate) comprise >70% of the inflammatory cells seen. Mixed reactions are characterized by a mixture of neutrophils and macrophages. In histiocytic inflammation, macrophages are the prevalent cell seen.

Most inflammatory effusions are cytologically nonspecific in terms of etiologic diagnosis. However, as with inflammatory responses elsewhere, cytomorphology provides significant clues as to the underlying cause. Neutrophilic inflammatory effusions indicate severe active irritation (Fig. 6-8). If neutrophils are degenerate, an effort should be made to identify bacterial organisms within phagocytes (primarily neutrophils). This is generally easiest at the feathered edge of the smear. If organisms are not seen, the fluid should still be cultured. Mixed and macrophagic inflammatory effusions reflect less severe irritation and are found with resolving acute effusions or in association with less irritating etiologic agents than bacteria (e.g., fungal organisms or foreign bodies). Chemical evaluation of effusion fluid is also useful in recognizing sepsis.

FIGURE 6-8 Nonseptic exudate. Peritoneal. Dog. Concentrated fluid from an animal with pancreatitis showing one cluster of reactive mesothelial cells and neutrophils. The protein content in this sample was 3.0 g/dl with a cell count of 8000 cells/μl. Most of the cells are neutrophils, suggesting an underlying inflammatory condition. Further testing (e.g., fluid and serum lipase or ultrasonography) is required to determine the specific cause for the fluid accumulation. (Romanowsky; HP oil.)

Septic effusions may be diagnosed by use of biochemical parameters in dogs and cats. It was shown that effusion fluid in dogs and cats with a pH <7.2, pCO₂ >55 mm Hg, glucose concentration <50 mg/dl, and lactate concentration >5.5 mmol/L is highly likely to have bacterial infection (Swann et al., 1996). A study involving peritoneal fluid in dogs and cats evaluated blood and effusion fluid glucose concentrations. This study found that a difference of blood-effusion glucose >20 mg/dl provides a rapid and reliable means to differentiate septic and nonseptic effusion fluid (Bonczynski et al., 2003). In this study, a difference of blood-effusion lactate concentration in dogs <−2.0 mmol/L was 100% sensitive and 100% specific for the diagnosis of septic peritonitis.

Feline Infectious Peritonitis

Feline infectious peritonitis (FIP) is unique among the causes of inflammatory effusion in that the fluid that accumulates is usually high in protein, yet has a low cellularity (McReynolds et al., 1997). Classification as an inflammatory effusion is based primarily on the presence of high total protein (often >4.5 g/dl) which is a reflection of a similar elevation in serum protein. Electrophoresis of either the effusion fluid or the serum reveals a polyclonal gammopathy. An albumin-to-globulin ratio of less than 0.8 on the fluid is very suggestive for FIP. It has been reported that if gamma globulin is greater than 32% of the protein in the effusion fluid, this is suggestive of FIP (Shelly et al., 1988). An additional test that can help rule in or rule out FIP is the Rivalta test. To perform this relatively simple test, place one drop of 98% acetic acid into 5 ml distilled water and mix well in a reagent tube. Then add slowly one drop of the effusion fluid to the surface of the acetic acid solution. A positive test requires that the drop of effusion retain its form on the surface or slowly sink to the bottom as a droplet or jellyfish-like shape (Fig. 6-9A). This test indicates a high protein content as well as fibrin and inflammatory mediators. This test has a positive predictive value of 86% and a negative predictive value of 97% (Hartmann et al., 2003). A more specific cytologic test for FIP involves immunofluorescence of intracellular feline coronavirus (FCo V) within effusion macrophages. When positive (Fig 6-9B), this test is diagnostic for FIP; however, the negative predictive value is 57%, so a negative test may miss cases of FIP (Hartmann et al., 2003).

FIGURE 6-9 A-C, Abdominal effusion, cat. A, Rivalta test. Positive test results are indicated by a layer of gel on top of the acetic acid solution. Fluid is from a cat with PCR confirmed FIP. The cat was moderately icteric. Note the yellow streaks of gel in the middle of the tube from partially floating material. B, FCo V immunofluorescence test. A specific cytologic test for FIP involves immunofluorescence of intracellular feline coronavirus (FCo V). Shown are three infected and intact macrophages (green) present in abdominal fluid from a cat with FIP. (FCo V; HP oil.) C, This concentrated smear comes from a cat diagnosed with FIP. Note the moderately basophilic macrophages and nondegenerate neutrophils. The background contains basophilic, coarsely granular protein as well as basophilic protein crescents and strands of fibrin. (Modified Wright; HP oil.) D, Nonseptic exudate. Peritoneal. Cat. This animal was diagnosed with FIP. Fluid contains foamy, vacuolated macrophages, mildly degenerate neutrophils, and intermediate to large lymphoid cells. Lymphocytes may be intermediate in size and appear reactive in some cases of FIP. Also note the granular precipitated protein throughout the slide. (Romanowsky; HP oil.)

(A, Photo by Sam Royer, Purdue University. B, Courtesy of Jacqueline Norris, University of Sydney, Australia.)