Accumulations of excessive fluid in the pleural, peritoneal, or pericardial space represent edema of these body cavities.
Accumulations of excessive fluid in the pleural, peritoneal, or pericardial space represent edema of these body cavities. As such, the dynamics of fluid accumulation are governed by Starling's law. In general, pleural, peritoneal, and pericardial effusions can result from either inflammatory or non-inflammatory causes. Non-inflammatory edema results from arterial pressure (a rare cause in veterinary medicine), venous stasis, hypoalbuminemia (reduces plasma oncotic pressure), or lymphatic obstruction. Inflammation causes fluid accumulation through vascular damage and leakage.
Evaluation of abnormal fluid accumulation is based upon integrated assessment of both physical and cytologic characteristics. Collection techniques and sample management are not detailed here.
Physical assessment of effusions includes estimation of total solids (total protein) by refractometry and determination of total nucleated cell count with either an automated particle or manual methods. On the basis of physical characteristics, effusions are classified as transudates, modifies transudates, or exudates. As a general rule (there are exceptions which will be discussed later), transudates are non-inflammatory in origin while exudates are the result of inflammation. Thus, physical characteristics of effusions allow classification according to pathophysiologic mechanism of fluid accumulation but reveal little about specific disease diagnosis.
Cytologic evaluation of effusions augments physical evaluation by allowing the diagnostician to identify specific etiologic agents, to follow the progression of a disease process, and to monitor response to therapy. The decision of whether or not to culture effusion fluid is based upon cytologic findings. Through cytologic examination, inflammatory reactions are classified as to type, hyperplastic responses of the mesothelial lining can be identified, and neoplasia is diagnosed.
Transudates are defined as excessive accumulations of fluid having normal physical characteristics. Transudates therefore have low total proteins (less than 3 g/dl) and low nucleated cell counts (less than 500 nucleated cell/μl). The single exception to these criteria is peritoneal fluid from the horse where normal values of up to 2000 nucleated cells have been reported.
Transudative effusions result most commonly from venous stasis and less frequently from hypoalbuminemia and lymphatic obstruction. For example, they are the typical effusion seen with congestive heart failure (venous stasis), liver failure (hypoalbuminemia, venous stasis, and lymphatic obstruction), the nephrotic syndrome (hypoalbuminemia), and in some cases of neoplasia (venous stasis and lymphatic obstruction). It is emphasized that pleural and peritoneal effusions due solely to hypoalbuminemia will only occur when serum albumin levels fall below 1 g/dl. At values greater than this some other additional contributory mechanism must also be considered.
Transudative fluids are quite clear at the time of collection. Despite the paucity of cells, however, cytologic examination is certainly warranted. While most transudates are nonspecific cytologically, those which are caused by neoplasms may contain malignant cells which allow specific diagnosis. To maximize the likelihood of finding such abnormal cells, sediment smears should always be used when transudates are examined.
The accumulation of transudative fluid in one of the body cavities 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 as they do they release chemoattractants which draw small numbers of phagocytes into the effusion to remove cellular debris. The result is a mild increase in both total protein (3.0-5.0 g/dl) and nucleated cell count (less than 5000/μl). When this occurs, the fluid is known as a modified transudate. Thus, modified transudates are nothing more than transudates which have been present long enough to illicit a mild inflammatory reaction. Eventually, physical characteristics of the fluid may change enough so that the fluid is classified as an exudate.
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 been called "pseudochylous effusions" (a term which in our opinion should be dropped). The gross appearance of these fluids is the result of a high lipid content but is in no way related to a true chylous effusion. The phagocytes attracted to remove cellular debris from transudates are rich in enzymes which digest protein but are virtually devoid of enzymes which 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 fluid. Pseudochylous effusions formed in this manner are easily distinguished cytologically from true chylous effusions.
Exudates are defined as fluid accumulations which are abnormally high in total solids and/or nucleated cell count. Total proteins range between 3.0 and 7.0 g/dl and total cell counts may be as high as 100,000/μl. The vast majority of exudates are caused by inflammation; however, the common denominator of exudate formation is vascular (including lymphatic) damage. Consequently, both hemothorax (and hemoperitoneum) and chylothorax (chyloperitoneum) are classified as exudates on the basis of physical characteristics and pathogenesis.
Cells present in normal pleural, pericardial, and peritoneal fluids include low numbers of mesothelial cells and occasional inflammatory cells. Mesothelial cells are present in small clusters or as individuals. If knocked loose from the cavity lining during the collection process, mesothelial cells resemble squamous cells with a low N/C ratio and abundant faintly basophilic cytoplasm. Cells with this morphology have historically been referred to as "pale" mesothelial cells. Normal mesothelial cells which are present in the fluid prior to collection have rounded up and are quite basophilic ("dark" mesothelial cells). They measure between 25 and 35μ in diameter. Nuclei are centrally located, round, and uniformly granular. Cytoplasm is abundant. The most striking characteristic of these cells is the presence of an eosinophilic peripheral brush border or "skirt".
The inflammatory cells present in normal fluids have the morphology of normal peripheral blood leukocytes. The predominant leukocyte seen varies with the species. In dogs and horses, neutrophils are prevalent. In cats and cattle, lymphocytes predominate.
Cellular morphology in true transudates is similar to that in normal fluid; as emphasized previously, the principal alteration in transudation is increased fluid volume. The principle cellular constituent of the modified transudate (even the pseudochylous effusion) is the reactive mesothelial cell.
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. 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 primary mesothelioma. This is done by critically evaluating the cell populations for criteria of malignancy.
Reactive mesothelial cells also may assume a variety of other morphologies. Individual mesothelial cells sloughed into effusions gradually lose their brush border and may develop phagocytic capabilities, accumulate intracellular debris, and become indistinguishable from macrophages.
As modified transudates mature, the proportion of inflammatory cells they contain increases. In most cases the principle inflammatory cell is the non-degenerate neutrophil, but neutrophils rarely account for more than 30% of the total cell population. As modified transudates gradually become cytologically indistinguishable from nonspecific low grade mixed to macrophagic exudates.
Inflammatory effusions are classified according to the standard rules for inflammation as neutrophilic, mixed, or macrophagic. In neutrophilic reactions, neutrophils (either non-degenerate or degenerate) predominate. Mixed reactions are characterized by a mixture of neutrophils and macrophages, and in macrophagic inflammation, macrophages are the prevalent cell seen.
Inflammatory effusions are by their nature irritative to the lining of involved body cavities. Consequently, virtually all inflammatory effusions are characterized by some degree of reactive mesothelial cell hyperplasia. Because of the tendency of reactive mesothelial cells to cytologically merge with the monocyte/macrophage continuum, they are considered to be a part of the chronic component of any inflammatory effusion.
Most inflammatory effusions are cytologically nonspecific in terms of etiologic diagnosis. However, as with inflammatory responses elsewhere, cytologic morphology provides significant clues as to underlying cause. Neutrophilic inflammatory effusions indicate severe irritation. If neutrophils are degenerate, then an effort should be made to identify bacterial organisms within phagocytes. This is generally easiest at the featheredge of the smear. If organisms are not seen, the fluid should still be cultured. Mixed and macrophagic effusions reflect less severe irritation and are found with resolving neutrophilic effusions or in association with less irritating etiologic agents than bacteria (e.g., fungal organisms or foreign bodies).
While most inflammatory effusions are cytologically nonspecific, some etiologies cause reactions of characteristic morphology. These effusions are discussed below.
Feline Infectious Peritonitis is unique among the causes of inflammatory effusion in that the fluid which accumulates is of low cellularity. Classification as an inflammatory effusion is based exclusively on the presence of an extremely high total protein which is a reflection of a similar elevation in serum protein. Electrophoresis of either the effusion fluid or the serum reveals a polyclonal gammopathy.
Grossly, the fluid is generally clear and straw-colored and may be present in either or both the pleural and peritoneal cavities. Even when collected into EDTA, the effusion fluid may contain clots or strands of fibrin.
Cytologically, although these fluids are consistently low in cell number, they are inconsistent in regards to cell type present. In a majority of cases the predominant cell is the non-degenerate neutrophil. However, mixed to macrophagic reactions are also frequently seen. We have even seen rare cases where the lymphocyte is prevalent. Regardless of the predominant cell type, slides in all cases have a purple granular background which results from the high protein content.
Complex bacteria such as Nocardia asteroides and Actinomyces sp. are important causes of both peritoneal and pleural effusions in dogs and cats. Grossly, these effusions are turbid and yellow to blood-tinged; even when collected in EDTA they typically contain visible particulates or granules (the so-called "sulfur granules").
On the basis of physical parameters, these effusions are typical exudates, with high total protein and marked cellularity. Because of the high cellularity, direct smears are generally adequate for cytologic examination. If particles are observed in the fluid, it is important to make squash preparations of these particles in addition to making smears of the fluid alone.
Microscopically, nocardial and actinomycotic infections are characterized by neutrophilic to mixed inflammation, probably dependent upon the duration of the disease. In the more chronic reactions there is generally a significant reactive mesothelial cell component to the response. A striking feature of the inflammatory response is the morphology of the neutrophils. Whereas most cases of septic pleuritis and peritonitis are signaled by the presence of predominantly degenerating neutrophils, in nocardial and actinomycotic effusions the majority of the neutrophils are non-degenerate. Degenerating neutrophils are only seen immediately in the vicinity of the bacterial organisms because these agents, in contrast to most other bacteria, produce only weak local toxins.
The net effect of this phenomenon is that smears in these cases may be easily misinterpreted as noninfectious, particularly if the organisms are not widespread. Because the particles seen grossly often are composed of bacterial colonies, it is extremely important that squash preparations of these particles be examined to insure that the diagnosis is not missed.
Morphology of the organisms microscopically is quite characteristic. Colonies are composed of delicate filamentous often beaded organisms. The most significant diagnostic feature of these organisms is that the filaments are branching. Using standard hematologic stains, Nocardia organisms cannot be differentiated from Actinomyces. However, Nocardia is gram and acid fast whereas Actinomyces sp. are gram and acid fast.
Cytologic diagnosis should be confirmed culturally. Because these species have special cultural requirements, it is important that the bacteriology laboratory be fully aware of the provisional diagnosis at the time of sample submission.
Rupture of the gall bladder or common bile duct may occur in any species secondary to direct trauma. In addition it is an infrequent accompaniment to diaphragmatic hernia from any cause in the dog and cat. When the result of direct trauma to the biliary system, leakage of bile is virtually always restricted to the peritoneal cavity and peritonitis is the result. When associated with diaphragmatic hernia, leakage of bile occurs when the liver is trapped in the diaphragmatic rent and there is necrosis of the gall bladder or common bile duct. In this circumstance, both peritonitis and pleuritis can result.
Because bile is a very irritative substance, its presence very quickly elicits an inflammatory response. Grossly, the fluid may be initially greenish; however, as the response becomes more and more cellular this discoloration may become masked. Large volumes of fluid can usually be obtained.
Cytologically, the striking feature of bile peritonitis/pleuritis is the presence of bile on the slide. In acute reactions, bile is seen as greenish to black granular material scattered in the slide background and in the cytoplasm of reactive mesothelial cells and macrophages. In reactions of greater duration, bile granules may have all been converted to rhomboidal golden crystals of bile pigment. When such crystals are found in the cytoplasm of effusion phagocytes in the absence of evidence of prior hemorrhage, the possibility of bile peritonitis or pleuritis should be strongly considered.
The cellular response to bile is generally mixed. Because bile is a severe irritant, large numbers of neutrophils are nearly always present. The degree of degeneration of neutrophils is variable. In addition, bile causes severe irritation to the mesothelial lining of the body cavities, resulting in marked reactive mesothelial cell hyperplasia. The combination of large numbers of neutrophils and reactive mesothelial cells in the fluid leads to the designation as mixed inflammation.
Systemic histoplasmosis is a moderately frequent cause of peritoneal effusion in dogs living in the Ohio River Valley. Because the fungus is ubiquitous in the area, serology cannot be relied upon for diagnosis. In many cases, cytologic identification of the organism is essential.
The effusion fluid of histoplasmosis is relatively unique. The fluid is usually clear and colorless. On the basis of physical characteristics, the fluid is a modified transudate; however, cytologically, it is clearly inflammatory with significantly more neutrophils than are seen in the typical canine modified transudate. Whenever inflammatory modified transudates are seen in dogs, the possibility of histoplasmosis should be considered in the differential diagnosis.
Demonstration of the organism is best done in macrophages at the feathered edge of sediment smears. Histoplasma organisms measure approximately 2-3μ in diameter, are ovoid, and have a single basophilic nucleus surrounded by a thick cell wall. The only other organisms of similar cytologic morphology are Leishmania. However, these protozoans are distinguished by the presence of an internal kinetoplast which gives them the appearance of having two nuclei. Additionally, leishmaniasis is a relatively uncommon cause of peritoneal effusion in the dog.
Only a few comments regarding parasitic effusions in the body cavities are warranted. As expected parasitic infections are often characterized by eosinophilic exudates. In dogs the most common such effusion is the pleural effusion associated with heartworm disease, while in the horse such effusion are most frequently seen in the peritoneal cavity in association with migrating strongyles and colic of verminous arteritis. It should be emphasized that not all cases of parasitic effusion are characterized by eosinophilic exudates; in many cases the exudate is non-specific. Then, too, a number of other diseases can result in eosinophilic peritoneal or pleural effusions. For example, we have seen such responses in association with systemic mastocytosis in both dogs and cats and eosinophilic pneumonitis in a variety of species.
Chylous effusions are the result of leakage of lymph into the body cavity and may involve either the pleural or peritoneal space. Most commonly, the thoracic cavity is involved (chylothorax) as a result of trauma to the thoracic duct.
Chylous effusions are described as opaque milky fluids. In our experience, this description is misleading. The turbidity and white color are a reflection of a high lipid content in the lymph; since significant amounts of lipid are only present in lymph immediately following a meal, chylous effusions formed shortly after feeding are the only ones with this appearance. We have seen numerous proven cases of chylothorax where the fluid was clear to serosanguinous.
Physical characteristics of chylous effusions are quite consistent; these fluids have moderately high protein concentration (3.5-4.5 g/dl) but relatively low cellularity. Cytologically, they are characterized by the presence of large numbers of morphologically normal small lymphocytes. Lesser numbers of reactive lymphocytes are also present. Because these fluids are mildly irritating, long-standing chylous effusions also may contain moderate numbers of reactive mesothelial cells and other inflammatory cells. We have seen several cases of chronic chylothorax where significant numbers of eosinophils were seen. The presence of lipid in the background of the slide, visualized as small unsustained droplets at the periphery of the nucleated cells, is variable.
It is important to note that in the cat, cardiac disease results in pleural effusions which are indistinguishable from chylothorax. The mechanisms behind these effusions have not been clarified; however, it is well established that heart failure causes venous and lymphatic stasis with increases pressure. In the cat it appears that these circumstances predispose to lymphatic leakage and result in a secondary chylous effusion.
True hemorrhagic exudates can occur in any of the major body cavities. Grossly, these effusions are red to serosanguinous depending upon the age of the exudate and the extent of the hemorrhage. Physical evaluation reveals a protein level reflective of but somewhat less than that of peripheral blood. Both nucleated cell counts and red blood cell counts are elevated.
Cytology is needed to differentiate true hemorrhagic exudates from sample contamination at the time of collection. Hemorrhagic exudates contain predominantly red blood cells with lesser numbers of nucleated cells. The most significant indicator of true hemorrhage is the presence of activated macrophages containing phagocytized red cells (erythrophagocytosis) or hemosiderin. These cells are best observed at the featheredge of sediment smears. Erythrophagocytosis is not seen if hemorrhage is strictly a collection artifact. A second significant observation is whether or not platelets are seen. True hemorrhagic exudates are devoid of platelets but they are commonly observed in contaminated samples.
Neoplastic processes, both primary and metastatic, are relatively common causes of both abdominal and thoracic effusions in dogs and cats. They are less frequent but still represent an important cause of effusion in the horse.
Neoplastic effusions may be accompanied by significant hemorrhage and/or inflammation but generally they are non-inflammatory. Grossly, the fluid may be clear to cloudy and hemorrhagic. Total protein levels are elevated but nucleated cell counts are highly variable.
In dogs and cats the common causes of neoplastic effusions are thymic lymphosarcoma (pleural), and adenocarcinoma or carcinoma (either pleural or peritoneal). In horses, squamous cell carcinoma of the stomach (peritoneal) is by far the most frequent cause, with lymphosarcoma occurring only occasionally. Mesothelioma can be a rare cause of effusion in any species. In the following paragraphs the cytologic features of the principal neoplastic effusions are summarized and illustrated.
Lymphosarcomatous effusions are generally highly cellular and contain a pleomorphic population of discrete round cells which are morphologically consistent with lymphocytes. The neoplastic cells have high nuclear/cytoplasmic ratios, scant cytoplasm, and often scattered small cytoplasmic vacuoles. Moderate numbers of mitoses are seen. Scattered among the neoplastic cells are red blood cells, reactive mesothelial cells, and inflammatory cells.
The principle differential diagnosis for lymphosarcomatous effusion is chylous effusion. Usually differentiating the two processes is not a major problem. Most cases of lymphosarcoma are lymphoblastic and the malignant cells are much larger than normal with nuclei containing prominent often bizarre and angular nucleoli. Only rare cases of lymphosarcoma have cells which are morphologically the same as normal small lymphocytes. In these cases the principal differentiating feature of the neoplastic disease is a much greater cellularity than what is generally seen with chylous effusion. Unfortunately, this finding is not totally consistent and the differential diagnosis must include both possibilities whenever effusions containing predominantly normal lymphocytes are observed.
Carcinomatous effusions in dogs and cats may be the result of either primary or secondary neoplastic processes. In the chest, the principal primary neoplasm is pulmonary adenocarcinoma. In order for neoplastic cells from this tumor to be present in pleural effusions, the neoplasm must have invaded either into pulmonary vessels and lymphatics or directly through the pleural surface of the lung and into the pleural cavity.
Pulmonary adenocarcinoma is a relatively uncommon tumor and extension into the pleural cavity is rarer still. As a consequence, most carcinomatous pleural effusions are the result of metastatic disease. In females, by far the most common cause of such effusions is metastatic mammary carcinoma; in males, the two most common metastatic carcinomas we have seen are prostatic carcinoma and transitional cell carcinoma.
In the peritoneal cavity the principal causes of carcinomatous effusions are those neoplasms which spread by implantation on the peritoneal surface. Significant among these are bile duct carcinoma, pancreatic adenocarcinoma, ovarian adenocarcinoma in females, and prostatic carcinoma in males. Metastatic mammary carcinoma is also a common cause of peritoneal carcinomatosis.
Cytologically, all of these tumors are morphologically similar and cannot be readily differentiated. Carcinomatous effusions are characterized by the presence of rafts and acinar arrays of round to polygonal cells with variable amounts of often extremely basophilic cytoplasm. Cytoplasmic basophilia may be so intense as to obscure nuclear detail. Inflammation may or may not be present, but reactive mesothelial cell hyperplasia is a constant feature.
Establishing a diagnosis of pleural or peritoneal carcinomatosis is probably the most difficult challenge in diagnostic cytology. Carcinoma cells strongly resemble reactive mesothelial cells and cytoplasmic basophilia in both of these populations makes evaluation of nuclear criteria of malignancy particularly difficult. The problems become even more exaggerated when significant inflammation is present because reactive mesothelial cells can become quite dysplastic.
For these reasons it is particularly important to search diligently for nuclear criteria of malignancy in suspect populations of cells. Areas where nuclear detail can be seen must be found. Whereas in most circumstances four nuclear criteria of malignancy are sufficient to allow a diagnosis of cancer, in the case of pleural or peritoneal carcinomatosis at least five distinct nuclear criteria should be demonstrated.
Once the diagnosis of malignancy has been established, differentiation of carcinoma from adenocarcinoma is much easier. Adenocarcinomas are secretory tumors; as such, most instances these vacuoles are unstained. In some cells the amount of secretory product is sufficient to displace the nucleus peripherally. Simple carcinomas are devoid of secretory vacuoles.
Squamous cells carcinoma of the stomach is the most common cause of neoplastic peritoneal effusions in the horse. This is an old age disease and may be manifested clinically as a wasting syndrome, recurrent colic, or sudden onset acute colic. Early in the course of the disease, peritoneal fluid cytology is unremarkable because the tumor is contained totally within the stomach. However, with time the neoplasm can erode through the stomach wall resulting in the release of gastric contents and cells into the peritoneal cavity.
Gastric contents are extremely irritating and peritonitis results. Although the reaction is generally non-septic, because of the toxic nature of gastric fluid, the neutrophils are usually degenerate. In addition to the inflammatory reaction, variable numbers of squamous epithelial cells, both benign and malignant, are scattered throughout the slide.
Usually, normal squamous cells are more numerous than malignant cells in these effusions and represent the first cytologic clue that neoplasia is present. Normal squamous cells in the peritoneal fluid can come from only four sources: contamination (in which case the cells will be rare), vaginal laceration (found at the time of breeding or foaling), uncomplicated stomach rupture (rare), or squamous cell carcinoma of the stomach. Whenever normal squamous cells are seen in equine peritoneal effusions malignant cells should be carefully sought.
Malignant squamous cells are often quite distinctive. The most striking feature is nuclear/cytoplasmic asynchrony, or the presence of a primitive nucleus in a cell with well differentiated often keratinized cytoplasm. These cells are generally quite large with angulated cellular margins and glassy cytoplasm. Nuclei are large, round to oval, and have striking abnormally clumped chromatin patterns. Nucleoli may or may not be observed. If present they are usually large and angular an often multiple. Malignant cells may be found either as individuals or in small aggregates.
Mesothelioma is a rare tumor which can occur in any species. The neoplasm arises from the mesothelial lining of the body cavities. Cytologically the tumor is very difficult to differentiate from carcinoma and reactive mesothelial hyperplasia. Tumor cells are round to polygonal and are arranged primarily in clusters. Nuclei are hyperchromic and located centrally. Often nuclei of adjacent cells within a cluster appear to press against each other resulting in triangulation of nuclei (nuclear coning).
Because of the difficulty of differentiating mesothelioma from reactive mesothelial hyperplasia, caution must be used when evaluating suspect populations for criteria of malignancy. As with carcinomatous effusions, at least 5 definite nuclear criteria of malignancy must be seen before a presumptive diagnosis is made.
Once the diagnosis of malignancy has been established, an attempt to differentiate mesothelioma from carcinoma can be made. There are no well established criteria for differentiating these tumors; consequently, such attempts are often an exercise in futility. Based upon relative frequency of occurrence, most of these effusions are simply considered to be carcinomatous until proven otherwise.