An effusion is the fluid that can accumulate in a body cavity due to various different pathologies. Mechanistically, effusions are caused by transudation, exudation, rupture of a blood/lymph vessel/ hollow organ, or neoplasia. Types of effusions include transudates and exudates, but also haemorrhagic effusions, chyle and uroperitoneum. The composition of the fluid gives clues as to the pathology that caused it, and the goal of fluid cytology is to describe this composition, by looking at cellularity, cell types present, protein concentration, SG and the presence of other substances like bilirubin or creatinine.
Effusions can be examined in-house or sent away to reference laboratory for evaluation. Fluid should be collected into an EDTA tube for cytological analysis, and into a plain tube for culture. Two direct smears (using the blood-smear technique) should be made from the EDTA-fluid. If the samples are to be sent away, these smears should be left unstained, packaged in plastic slide-keepers and stored at room temperature until sent. The EDTA and serum tubes should be stored at 4°C. For in-house analysis, the total nucleated cell count (TNCC) can be determined from the uncentrifuged EDTA fluid sample on an automated cell counter – in most cases the same analyzer used for measuring haematology/ CBC samples. If this is not available, an estimate of cellularity (“low” or “high”) can be made from the direct smears, with some experience. After TNCC measurement, the EDTA tube is centrifuged. Total protein and SG of the supernatant is measured with a refractometer. If the TNCC is < 30 000 cells/ µL, or cellularity is low on the direct smear, smears should be made from the sediment after centrifugation. The supernatant should be saved in case any biochemical tests need to be performed. If the fluid is very bloody, prepare a buffy coat smear to concentrate any nucleated cells of interest.
It is usual to perform a 100-cell differential count, or at least estimated the proportions of the different nucleated cell populations, when examining the smears. Note should be made of any infectious agents, foreign material, background or atypical cells. Normal pleural fluid has a protein concentration of < 25 g/L with very low numbers of macrophages and mesothelial cells. Mesothelial cells line the pleural, pericardial and peritoneal cavities and will be present in most effusions. They are large cells, present singly or in clusters, as seen in Figure 1. Mesothelial cells become hyperplastic very quickly in response to an increased volume of any type of fluid. Numbers increase, and they display many atypical features – multinucleation, mitotic figures, prominentnucleoli, coarse chromatin and nuclear moulding are not unusual (Figure 2).
They also tend to slough off in sheets when reactive. It is almost impossible to differentiate highly reactive mesothelial cells from a neoplastic population, as well as from cells originating from an adenocarcinoma. Diagnosis of a mesothelioma based on cytology alone is not considered accurate, and imaging findings should always be taken into consideration. If cytology is suspicious, fluid sediment/ pellet can be fixed in formalin and processed as a histopathology sample. Evaluation of the arrangement of the cells as well as immunohistochemical staining can then be performed.
Classification of Effusions
Readers may notice the absence of the “modified transudate” in the descriptions below. This classification is only used in veterinary, and not human medicine. The classification of transudate and exudate, based on cut-offs for TNCC, SG and protein was first introduced in a veterinary internal medicine textbook in 1968. The modified transudate appeared in 1971, in order to cover fluids that did not fit the criteria for transudate or exudate. Rather alarmingly, the veterinary classification system, using cellularity and total protein to indicate the mechanism of effusion formation, is not evidence-based and has been shown to be inaccurate. The modified transudate, in particular, is considered by many veterinary clinical pathologists to be a meaningless and unhelpful category. The classification system used here attempts to give information about the underlying aetiology for the effusion. (Regarding units for total nucleated cell counts: 1 x 109/L = 1000 /µL. The former is the SI unit, but the latter is more commonly used.)
Transudates appear when there is increased vascular hydrostatic pressure with or without loss of oncotic pressure due to hypoalbuminaemia (i.e. a disruption of Starling’s forces). Transudates are further subdivided as:
i. Low protein transudate: clear colourless effusions with low protein concentrations and low nucleated cell count (NCC):
- Dogs/ cats: TP < 25 g/L, NCC< 1500 cells/ µL
- Cell types present: mixture of non-degenerate neutrophils, lymphocytes, macrophages and reactive mesothelial cells (all in very low numbers).
- Caused by leakage of low protein fluid from the blood vessels into a body cavity, usually due to hypoalbuminaemia together an increase in hydrostatic pressure.
For transudative effusions to be present solely due to hypoalbuminaemia, it has been suggested (but not proven in animals) that the serum albumin level has to be <15 g/L. The 2 main disorders causing the formation of a low protein transudate are
- protein-losing nephropathy (protein lost through the kidney – hypoalbuminaemia – decreased oncotic pressure + abnormal regulation of blood volume – water retention – increased hydrostatic pressure)
- hepatic cirrhosis (liver cannot produce proteins – hypoalbuminaemia – decreased oncotic pressure + abnormal regulation of blood volume – water retention – increased hydrostatic pressure)
Other rarer disorders than can cause a low-protein transudate in the absence of hypoalbuminaemia are: portal vein anomalies, pulmonary hypertension. Low protein transudates are rare.ii .High protein transudate: clear to cloudy, yellow, orange or red effusions with medium to high protein concentrations and low to medium nucleated cell count (NCC):
- Dogs/ cats: TP ≥ 25 g/L, NCC< 5000 cells/ µL
- Cell types present: mixture of non-degenerate neutrophils and macrophages and reactive mesothelial cells.
Caused by increased hydrostatic pressure in the lungs or liver (sinusoidal/ post-sinusoidal) due to venous congestion. The increased hydrostatic pressure causes fluid to leak out into the pleural/ peritoneal space. The protein concentration of the fluid is not low.The two main disorders causing venous congestion in the lungs or liver, leading to a high protein transudate are:
- congestive heart failure
- chronic hepatic disease
An exudate forms due to increased vascular permeability caused by inflammation. The exudation of protein-rich fluid is accompanied by migration of inflammatory cells into the effusion:
- Dogs/ cats: TP >25 g/L, NCC >5000 cells/ µL
- Exudates are hazy to cloudy, yellow, tan, cream, orange, and may have a putrid smell if bacteria are involved.
Neutrophils are the dominant cell present in most exudates. With inflammation due to bacterial infections, degenerate neutrophils will be the predominant cells present, unless the bacterial toxin is weak or produced in small amounts. Lack of bacteria and/or organisms does not rule out an infectious cause. A septic exudate refers to the presence of intra- and extracellular bacteria, and a non-septic exudate refers to the absence of bacteria and a negative culture. The classification of an effusion as non-septic should be based on a negative culture. Non-septic exudates are caused by acute pancreatitis, necrosis associated with intracavitatory neoplasia or secondary to intracavitatory organ inflammation. Septic exudates are caused by disorders like GIT rupture or penetrating wounds (Figure 3). If acute pancreatitis is suspected to be the cause of a non-septic exudate, fluid lipase may be helpful. A DGGR-lipase activity in fluid of > 500 U/L, or twice serum lipase; or a SpecCPL of > 500 µg/L is highly specific for pancreatitis as the cause of the exudate.
Exception: FIP exudate: Caused by the FIP virus, has a very specific appearance – straw-coloured, viscous, TP > 25 g/L, NCC < 5000 cells/ µL. Cell population is a mixture of neutrophils and macrophages. (The reason for the low cellularity is that although there is a vasculitis in FIP, the focus of inflammation is inside the blood vessels and not in the body cavity – so vessels are leaky to plasma proteins, but inflammatory cells remain in the blood vessel walls)
3. Haemorrhagic Effusion
Causes include haemostatic defects, trauma or neoplasia invading blood vessel walls. The fluid is red and bloody in appearance. Dominated by red blood cells. PCV >1% All species: TP >25 g/L, NCC >2000 cells/ µL
With haemorrhage of more than 24 hours duration or chronic persistent haemorrhage, erythrophagocytosis and/or haemosiderin or haematoidin should be present. With peracute haemorrhage or iatrogenic blood contamination, platelets could be present. If both erythrophagocytosis and platelets are present, chronic persistent haemorrhage or previous haemorrhage and iatrogenic blood contamination are present. A fluid PCV of > 1% indicates that haemorrhage is contributing to the effusion. Inadvertent aspiration of the liver or spleen could also cause a cytological appearance similar to haemorrhage. The PCV of this fluid will then be higher or similar to the peripheral blood PCV. The most common causes of haemoperitoneum in dogs is splenic disease, specifically haemangiosarcoma. However, fluid cytology has a very low sensitivity for detecting cells from haemangiosarcomas in haemorrhagic fluids, due to the dilution effect of the blood and poor exfoliation of cells from these tumours. In other words, neoplastic cells from haemangiosarcoma are rarely seen in haemoperitoneum caused by this tumour. The absence of neoplastic cells does not at all rule out a haemangiosarcoma. In cats, the most common cause of haemorrhagic effusions is trauma, followed by various types of neoplasia.
Caused by a ruptured bladder. TP and NCC are variable – both very low initially (because urine has a very low protein and cell content), then increasing with time as urine is irritant to the peritoneum and causes a low-grade inflammation, so that the fluid becomes an exudate with time. There may be a urine odour. If uroperitoneum is suspected, fluid creatinine concentrations must be measured.
Two out of the following three criteria are diagnostic for uroperitoneum:
- Fluid creatinine > 2x serum creatinine
- Fluid creatinine > 4x normal serum creatinine (upper reference limit)
- Fluid K+ > 1.4 x serum K+
5. Chylus Effusion / Chyle
Chylus effusions occur due to leakage of lymphatic fluid into body cavities due to physical or functional obstructions to lymphatic ducts or to trauma (rare). These effusions therefore have characteristics similar to lymphatic fluid in the lymph ducts draining the GIT.
- They are milky and white, TP >25 g/L, NCC < 10 000 cells/ µL.
- The dominant cell population is small lymphocytes, but cytology may vary if the chylus fluid causes irritation of and inflammation is present.
6. Bile Peritoneum
Caused by a rupture in the biliary system due to gall bladder disease, choleliths or cholangitis. Starts off as a low protein transudate, but as bile is extremely irritant to the peritoneum, the protein and cellularity increase and the process becomes exudative.The fluid may have a greenish or orange tinge. Bilirubin concentrations in the fluid are >2x bilirubin in serum. Bilirubin crystals may be seen on cytology.
7. Neoplastic Effusion
Neoplasia may cause:
- obstruction to lymphatics resulting in a chylus effusion
- obstruction to blood vessels resulting in a high-protein transudate
- invasion of blood vessels resulting in a haemorrhagic effusion
- inflammation resulting in an exudate
A neoplastic effusion is identified as such when malignant cells are seen in the effusion. (Figure 4) The most common tumours associated with effusions are lymphoma and carcinomas. A cytological diagnosis of neoplasia in an effusion (excluding mesothelioma) has a high specificity and high positive predictive value. Sensitivity is however low, and a negative diagnosis does not rule out neoplasia.
8. Pericardial Effusions
In dogs, the most common causes are neoplasia (haemangiosarcoma) and idiopathic. Rarely, a pericardial effusion may be caused by left atrial rupture, coagulopathy or bacteria. Cytology usually reveals a haemorrhagic effusion with highly reactive, atypical looking mesothelial cells. Sensitivity for identifying HAS is very low. Around 75% of pericardial effusions in cats are caused by chronic heart failure (i.e. will have cellularity and protein concentration consistent with high protein transudate). However, periocardiocentesis is performed extremely rarely in cats.
1. Buob, S., Johnston, A.N. and Webster, C.R.L. (2011), Portal Hypertension: Pathophysiology, Diagnosis, and Treatment. Journal of Veterinary Internal Medicine, 25: 169–186.
2. Cagle, L.A., Epstein, S.E., Owens, S.D., Mellema, M.S., Hopper, K. and Burton, A.G. (2014), Diagnostic Yield of Cytologic Analysis of Pericardial Effusion in Dogs. J Vet Intern Med, 28: 66–71
3. Chartier MA, Hill SL, Sunico S, Suchodolski JS, Robertson JE, Steiner JM. Pancreas-specific lipase concentrations and amylase and lipase activities in the peritoneal fluid of dogs with suspected pancreatitis. Vet J. 2014 Sep;201(3):385-9.
4. Cornell University College of Veterinary Medicine: http://www.eclinpath.com/cytology/effusions-2/
5. Dempsey SM, Ewing PJ. A review of the pathophysiology, classification, and analysis of canine and feline cavitary effusions. J Am Anim Hosp Assoc. 2011 Jan-Feb;47(1):1-11.
6. Epstein SE. Exudative pleural diseases in small animals. Vet Clin North Am Small Anim Pract. 2014 Jan;44(1):161-80.
7. Hirschberger J, DeNicola DB, Hermanns W, Kraft W. Sensitivity and specificity of cytologic evaluation in the diagnosis of neoplasia in body fluids from dogs and cats. Vet Clin Pathol. 1999;28(4):142-146.
8. Hall DJ, Shofer F, Meier CK, Sleeper MM. Pericardial effusion in cats: a retrospective study of clinical findings and outcome in 146 cats. J Vet Intern Med. 2007 Sep-Oct;21(5):1002-7.
9. Schmiedt, C., Tobias, K. M. and Otto, C. M. (2001), Evaluation of Abdominal Fluid: Peripheral Blood Creatinine and Potassium Ratios for Diagnosis of Uroperitoneum in Dogs. Journal of Veterinary Emergency and Critical Care, 11: 275–280.
10. Zoia, A., Drigo, M. Diagnostic value of Light’s criteria and albumin gradient in classifying the pathophysiology of pleural effusion formation in cats. J Fel Med Surg, 2016;18:666-672