Abstract
Practical relevance:
Feline coronavirus (FCoV) infection is very common in cats, usually causing only mild intestinal signs such as diarrhoea. Up to 10% of FCoV infections, however, result in the fatal disease feline infectious peritonitis (FIP).
Clinical challenges:
Obtaining a definitive diagnosis of FIP based on non-invasive approaches is difficult. Confirmation of the disease relies on finding appropriate cytological or histopathological changes in association with positive immunostaining for FCoV antigen. In FIP cases with effusions, cytology and immunostaining on effusion samples can be relatively easy to perform; otherwise obtaining diagnostic samples is more challenging and collection of biopsies from tissues with gross lesions is necessary. In the absence of a definitive diagnosis, a high index of suspicion of FIP may be obtained from the cat’s signalment and history, combined with findings on clinical examination and laboratory test results. If largely consistent with FIP, these can be used as a basis for discussion with the owner about whether additional, more invasive, diagnostic tests are warranted. In some cases it may be that euthanasia is discussed as an alternative to pursuing a definitive diagnosis ante-mortem, especially if financial limitations exist or where there are concerns over a cat’s ability to tolerate invasive diagnostic procedures. Ideally, the diagnosis should be confirmed in such patients from samples taken at post-mortem examination.
Global importance:
FIP occurs wherever FCoV infection is present in cats, which equates to most parts of the world.
Evidence base:
This review provides a comprehensive overview of how to approach the diagnosis of FIP, focusing on the tests available to the veterinary practitioner and recently published evidence supporting their use.
What are coronaviruses?
Coronaviruses are large, enveloped, positive-sense, single-stranded RNA viruses with non-segmented genomes of around 30,000 nucleotides in length. 1 Feline coronavirus (FCoV) is a subspecies of the alphacoronavirus 1 species, along with canine coronavirus (CCoV). Coronaviruses exhibit a high rate of mutation during RNA replication and therefore exist as clusters of genetically diverse populations, known as quasispecies.2,3 This genetic diversity, along with a readiness to recombine with other coronavirus strains, is associated with their pathogenicity and cross-species transmission. FCoVs appear to have emerged in the 1950s, possibly due to cross-species transmission, and cats worldwide have now been found to be infected with FCoV, with the exception of those residing in a few areas, such as the Galápagos Islands 4 and Falkland Islands, 5 likely due to the isolation afforded by their remote island habitat.
Schematic diagram of an FCoV particle. The spike protein binds to the feline receptor, mediating host cell entry. The feline receptor is known to be aminopeptidase N for type 2 FCoVs, but is, as yet, unknown for type 1 FCoVs. Modified with permission from Dr Emi Barker
The complex relationship between FCoV and feline infectious peritonitis
FCoV infection is very common in cats. Around 40% of the domestic cat population has been infected with FCoV, and this figure increases to 90% in multi-cat households.11,12 Modern trends in feline husbandry (eg, more cats kept indoors and in multi-cat households) are likely to have led to an increase in FCoV-related disease. 13
Natural infections with FCoV are transient in ~70% of cats, but persistent infections occur in ~13% of cats. 13 These persistently infected animals are sometimes referred to as ‘carrier’ or ‘chronically shedding’ cats. In most cases, FCoV infection is asymptomatic or results in only mild gastrointestinal clinical signs (eg, inappetence, diarrhoea, vomiting), although more severe gastrointestinal disease is sometimes seen. Interestingly, around 5–10% of cats are believed to be resistant to FCoV infection. 13 However, in a small percentage of cases, FCoV infection results in feline infectious peritonitis (FIP), a fatal disease that is a common cause of death in young cats.14,15 Occasionally, and possibly with increasing frequency recently, outbreaks of FIP (when a larger percentage of cats are affected) in multicat households or shelters are reported.16,17 No curative treatments for FIP are currently available, although some novel treatments, including protease inhibitors, show promise.18–20
Asymptomatic FCoV infection was previously believed to be confined to the intestinal tract, but we now know that healthy FCoV-infected cats can have systemic infection, albeit with lower viral loads than cats with FIP.3,21–23
Factors contributing to the development of FIP
Viral factors
Viral factors are important in the pathogenesis of FIP. As described in the box on page 228, the S protein of FCoV mediates host cell entry. The S protein contains a putative fusion peptide that enables fusion of the FCoV envelope with the host cell membrane. 24 Mutations in the S gene can result in amino acid substitutions in the transcribed S protein that influence the tropism of FCoV. 25 Studies have identified mutations in the fusion peptide sequence of the FCoV S gene that were thought to be markers of FIP,26,27 as well as changes in the closely related furin cleavage site that were also thought to be correlated with FIP. 28 Recently, it has been found that the fusion peptide mutations are likely to be markers of systemic FCoV infection, which can occur in both FIP and non-FIP cats, rather than of FIP per se. 29 However, these mutations are still important as it is probably via these, and/or other mutations, that the FCoV acquires its monocyte/ macrophage tropism that allows it to spread systemically, outside of the intestinal tract, and contribute to the development of FIP.
Other viral factors mediating effective and sustained replication in monocytes, and activation of infected monocytes, are also likely to be important for the subsequent development of FIP following systemic FCoV infection. 30
Host factors
Host factors are additionally very likely to play an important part in FIP development. These include the host immune response (eg, T lymphocyte depletion occurs in cats that develop FIP), the ability of monocytes to sustain FCoV replication, breed and genetics.31–34
Environmental factors
Environmental factors, such as the level of stress and overcrowding in a household, may also play a role. These may act to accelerate the rate of viral replication in an individual cat, 35 which could increase the generation of viral mutants and support FIP development.
Approaching a diagnosis
Making a definitive diagnosis
A definitive diagnosis of FIP traditionally relies on histopathological examination of tissues, usually with detection of the virus within lesions by immunohistochemistry (IHC) for FCoV antigen. Immunostaining of FCoV antigen in effusion samples is also an option in cases with effusions showing biochemical and cytological features consistent with FIP.
Histopathological examination, effusion analysis and FCoV antigen immunostaining are discussed in more detail later.
Obtaining a high index of suspicion
In the absence of a definitive diagnosis, a high index of suspicion of FIP may be obtained from background information, clinical signs and routine clinicopathological results. With experience, this can be used as a basis for discussion with the owner of whether additional, more invasive, diagnostic tests (eg, surgical biopsy of affected tissues) are warranted. In such cases it may be that euthanasia is discussed as an alternative to pursuing a definitive diagnosis ante-mortem; this may be preferable in, for example, shelter cats, where there are financial limitations, or when cats are very sick and concerns exist over a patient’s ability to tolerate further diagnostic procedures.
If euthanasia is performed without a definitive diagnosis, post-mortem examination is strongly advised and relatively simple to perform. 36 This permits both assessment for gross changes consistent with FIP (Figure 2) and sampling for histopathological examination. If financial limitations preclude the latter, it is worth contacting researchers with an interest in FIP (eg, the University of Bristol Feline Coronavirus Research Group and Bristol-Zurich FIP Consortium) to see if samples are being sought for research studies that could allow for analysis at a reduced cost or free of charge.
Typical gross post-mortem findings in cases of FIP. Granulomatous lesions in organs or fibrinous plaques on the serosa of organs may be visible in the abdominal or thoracic cavity; tissues that are good to examine for these are the mesenteric lymph nodes, liver, spleen, kidneys and intestinal surfaces, as well as the peritoneal lining of the abdominal wall and diaphragm. In effusive cases, yellow sticky fluid can be visible in the pleural and/or peritoneal cavities, but the pericardium can also be checked for fluid
Many important differential diagnoses should be considered in cats suspected of having FIP, such as toxoplasmosis, mycobacterial infection and lymphocytic cholangitis. These, and others, are described in Table 1, together with consideration of features distinguishing those diseases from FIP.
Diseases to consider in the differential diagnosis of FIP
Modified from Tasker S and Dowgray N, 36 with permission from BSAVA Publications, Gloucester
Examples of clinical signs seen in cases of FIP. Clinical signs are typically assigned to wet or dry FIP presentations, but there is much overlap between the two presentations and many signs are seen in both forms
Signalment and background evidence for FIP
It should be remembered that FIP is most common in young cats (those less than 3 years of age, and especially those less than 2 years of age) 37 but a smaller peak of cases is seen in cats older than 10 years of age. Male cats are also at a slightly higher risk. 37 Some breeds in some countries may be predisposed to FIP,38,39 but this is likely due to the presence of unknown specific genetic risk factors in those breeds in those countries, and generalised breed predispositions may not exist. 37 A recent history of stress (adoption, being in a shelter, neutering, upper respiratory tract disease, vaccination, etc) may be apparent 37 and may play a part in triggering the development of FIP in an FCoV-infected cat.
Although living in a multi-cat household increases the likelihood of being FCoV seropositive, a recent large study 37 found that the majority of cats presented to a university hospital with FIP were from households containing a small number of cats at the time of diagnosis.
Clinical signs
FIP typically manifests as a vasculopathy resulting in (‘wet’) effusions (up to 80% of FIP cases have effusions), or granuloma formation resulting in (‘dry’) mass lesions, or a combination of the two; indeed, most FIP cases with effusions also have granulomatous lesions visible at post-mortem examination.
Clinical signs (Figure 3) seen in both effusive and non-effusive FIP include lethargy, anorexia, weight loss (or failure to gain weight/stunted growth in younger cats), a fluctuating pyrexia that is usually non-responsive to drugs such as antibiotics or non-steroidal anti-inflammatories, and sometimes jaundice (more common in effusive FIP).
A recent study 40 evaluating referral feline cases with a history of pyrexia found that FIP was the most common diagnosis made (20.8%, 22/106 cats), showing the importance of FIP as a differential diagnosis in pyrexic cats in this population of referred cats. Another study, describing FIP cases only, reported that body temperature exceeded 39.5°C in 81% of cats and 40°C in 39% of cats. 37 Pyrexia was far more common in cats with effusive FIP than those with neurological non-effusive FIP. 37
Lymphadenomegaly can also be present in both the effusive and non-effusive forms of FIP.
Effusive (wet) FIP is associated with abdominal, pleural and/or pericardial effusions (occasionally in the scrotum too of entire male cats), and is often quite acute in nature, progressing within a few days or weeks and severely limiting survival. 41 These cats can present with dyspnoea, tachypnoea and/or abdominal distension.
Non-effusive (dry) FIP is typically associated with neurological signs (which can be focal, multifocal or diffuse in nature, often with central vestibular signs, occasionally as a T3–L3 myelopathy) 42 and/or ocular signs (anterior and/or posterior uveitis) and is more chronic, progressing over a few weeks to months. Dermatological signs (manifesting typically as small, multiple, non-pruritic papules or nodules)43,44 have also been reported in dry FIP.
Renomegaly may occur in non-effusive FIP with renal involvement. Occasionally a diffuse pyogranulomatous pneumonia is reported.
It is important to remember that clinical signs of FIP can change over time, so repeated clinical examinations are important to detect newly apparent signs (eg, development of a small volume of effusion, ocular changes visible on retinal examination).
Focal non-effusive FIP occasionally occurs, presenting typically as a palpable abdominal mass. It can be particularly challenging to diagnose as the lesions can be hard to initially differentiate from neoplasia and myco-bacterial infection. Reports of focal FIP cases have comprised cats presenting with mesenteric lymph node enlargement due to necrogranulomatous lymphadenitis, 45 or solitary mural intestinal lesions of the colon or ileocaecocolic junction with associated regional lymph -adenopathy. 46 The cats with focal intestinal FIP had a history of vomiting and diarrhoea.
Possible findings from routine laboratory tests
Haematology
Haematological changes in FIP are non-specific but there are a number of abnormalities that can be looked for to support a diagnosis. Lymphopenia is particularly common (55–77% of cases, although a recent study 37 found only 49.5% of FIP cases to be lymphopenic), with neutrophilia (39–57%), a left shift, and mild to moderate normocytic, normochromic anaemia (37–54%) also reported.37,44–49 An association between FIP and microcytosis (with or without anaemia) was recently reported. 33 Severe immune-mediated haemolytic anaemia (IMHA), with an associated regenerative anaemia, can occur with FIP, 49 but is uncommon.
Serum biochemistry
The changes in serum biochemistry in FIP cases are varied and often non-specific, but there are a number of important abnormalities that should be looked for to support a diagnosis of FIP.
Hyperglobulinaemia
Hyperglobulinaemia is reported in 89% of cases, often with hypoalbuminaemia or low-normal serum albumin (seen in 64.5% of cases). 37 The presence of hypoalbuminaemia alongside hyperglobulinaemia means that hyperproteinaemia may not always occur; past reports documented hyperproteinaemia in up to 60% of cases, especially in dry FIP cases, but lower prevalences of 17.5% have been reported recently. 37
The combination of hyperglobulinaemia and hypoalbuminaemia or low-normal albumin concentration also means that the albumin:globulin (A:G) ratio is low, and this parameter can be useful to evaluate how likely FIP is in an individual case. Reports of useful cut-off values for A:G ratios in the diagnosis of FIP vary, but it has been suggested that an A:G ratio of <0.4 makes FIP very likely, while an A:G ratio of >0.8 makes FIP very unlikely.47–49 Although these cut-off values are useful to consider, the author does not use a specific value but looks at the A:G ratio in conjunction with other diagnostic test results; the lower the value, the bigger the suspicion for FIP becomes, especially if other findings are consistent with a diagnosis of FIP. Interestingly, a study in a population of cats with a low prevalence of FIP (akin to the situation that is usually encountered in veterinary practice) found that an A:G ratio of >0.6 was useful in ruling out FIP, but that lower ratios were not helpful in ruling in FIP. 50
Additionally, the frequency and extent of hypoalbuminaemia, hyperglobulinaemia, low A:G ratio and serum protein electrophoresis (SPE) abnormalities reported in FIP cases have decreased recently.37,51 With respect to SPE changes, in one study 51 cases diagnosed with FIP from 2013–2014 tended to have elevated alpha (α)2-globulins rather than the elevated gamma (γ)-globulins seen in cases from 2004–2009. This is possibly due to veterinarians diagnosing FIP earlier, meaning that cases have not progressed to show elevated γ-globulins. Polyclonal and monoclonal elevated γ-globulins have been reported with FIP, 52 although polyclonal elevations are far more common.
Hyperbilirubinaemia
Hyperbilirubinaemia occurs in 21–63% of FIP cases, and is especially seen with effusive FIP, often without marked elevations in alanine aminotransferase (ALT), alkaline phosphatase (ALP) or γ-glutamyltransferase enzyme activity (although these can be moderately elevated in FIP cases). Hyperbilirubinaemia due to IMHA is uncommonly reported with FIP, 49 and cats are often not severely anaemic. Thus the presence of hyperbilirubinaemia in the absence of elevated hepatic enzyme activities or severe anaemia should raise the index of suspicion for FIP (note that sepsis and pancreatitis can also cause hyperbilirubinaemia in the absence of elevated hepatic enzyme activities [Table 1]). One study 47 found, by sequential assessment of cats with FIP, that hyperbilirubinaemia was more commonly identified in cats just before their death or euthanasia than at first presentation. Additionally, this study also found that the bilirubin levels were more elevated in cats just before their death or euthanasia compared with at first presentation.
Acute phase proteins
Acute phase proteins (APPs) are made in the liver in response to cytokines released from macrophages and monocytes (especially inter-leukins 1 and 6 and tumour necrosis factor α) in many inflammatory and non-inflammatory diseases.
α1-acid glycoprotein (AGP) is an APP, and its measurement can be helpful in the diagnosis of FIP. Although AGP elevations (>0.48 mg/ml) per se are not specific for FIP, markedly elevated AGP levels (>1.5 mg/ml) are often seen in FIP cases. Therefore, the magnitude of the increase may be helpful in aiding diagnosis of FIP, with higher levels being more useful in raising the index of suspicion.53–56 Indeed, a study found that when the pretest probability of FIP was high (ie, history and clinical findings supportive of FIP), moderate serum AGP levels (1.5–2 mg/ml) could discriminate cats with FIP from cats without FIP, but only higher serum AGP levels (>3 mg/ml) could support a diagnosis of FIP in cats with a low pretest probability of disease (ie, history and clinical findings not supportive of FIP). 55 However, another, albeit very small, study of unusual cases of FIP found that modest AGP concentrations (>1.5 mg/ml) were still able to discriminate between FIP and non-FIP cases. 53
FCoV serology in FIP
Serum antibody tests for FCoV are usually ELISAs, indirect immunofluorescence antibody tests or rapid immunomigration tests. 57 Most use CoV-infected swine or feline cells as a substrate and titres are read in distinct multiples of serum dilutions. A positive FCoV antibody test indicates that the cat has been infected with FCoV and has seroconverted (this takes 2–3 weeks from initial infection).
Breed-related differences in median FCoV antibody titre have been detected, and may reflect differences in breed response to FCoV infection.58,59
Although FIP cats tend to have higher FCoV antibody titres than non-FIP cats, there is much overlap, with no difference between median FCoV antibody titres in healthy and suspected FIP cats. Therefore, the titre in an individual animal is of limited use in distinguishing cats with FIP. 59 Many clinically healthy cats (especially those in multi-cat households) have positive, often very high, FCoV antibody titres, while ~10% of cats with FIP are seronegative (this could be due to the presence of virus in the sample binding antibody and rendering it unavailable to the serological test), 60 highlighting difficulties in interpretation. It may be that a negative FCoV antibody result in a suspected dry FIP case is more useful to rule out a diagnosis of FIP; 61 however, negative results have been reported in cases of neurological FIP. 62 As a result, clinicians vary as to whether they perform serology in suspected cases, although a positive result certainly indicates exposure to FCoV.
Analysis of effusion samples
Analysis of any effusion sample in a suspected case of FIP is extremely helpful for diagnosis, so obtaining samples of effusions should always be a priority. Ultrasonography is generally regarded as being more sensitive than radiography for the detection of small volumes of fluid in the thorax and abdomen, but this may depend on where pockets of fluid reside. Repeated ultrasonography to identify any small volume effusion is recommended and, similarly, ultrasonography can be used to guide sampling of small pockets of fluid.
FIP effusions (Figure 4) are usually clear, viscous/sticky, straw-yellow and protein-rich (thick eosinophilic proteinaceous backgrounds are often described on cytology), with a total protein concentration of >35 g/l (>50% globulins). Very occasionally chylous effusions are described.
Typical appearance of effusion seen in cases of FIP. Effusions tend to be clear, viscous and straw-yellow in colour
Typical FIP effusions have similar low A:G ratios (see above) and raised AGP concentrations to those in serum. A recent study found that AGP concentrations (of >1.55 mg/ml) in effusions were more useful (sensitivity and specificity of 93%) in differentiating FIP and non-FIP cases than AGP levels in the serum or other APPs; 56 however, the ‘diagnosis’ of FIP in the cats in this study was not always confirmed by histopathology and immunostaining.
This is a positive Rivalta’s test as the drop has retained its shape with a connection to the surface of the liquid. A positive result indicates that the effusion being tested is an exudate, but is not specific for FIP
FIP effusions are poorly cellular (usually <5 × 109cells/l), and are typically pyogranulomatous in nature with macrophages, non-degenerate neutrophils and very few lymphocytes. The effusions are therefore often described as modified transudates based on cell counts (<5 × 109 cells/l) but as exudates based on protein concentrations (>35 g/l). Rivalta’s test (see box above) is a crude point-of-care assay that can be performed on an effusion sample to allow rapid differentiation of transudate from exudate. A positive result merely means that the effusion is an exudate and thus is not specific for FIP; positive results are reported in non-FIP cases (eg, bacterial/septic peritonitis and lymphoma). 64 Although cytology can sometimes be successful in diagnosing bacterial/septic peritonitis and lymphoma, to help discriminate these from FIP, many vets may not be confident performing cytology in-house.
Serology for FCoV antibodies can also be performed on effusion samples, with very varied results,60,65 so the author does not perform this test in suspected cases of FIP.
Immunostaining for FCoV antigen and RT-PCR for FCoV RNA on effusion samples can also be performed (see later).
Miscellaneous diagnostic tests
In cases with neurological clinical signs, imaging of the brain by MRI may be useful to demonstrate changes. For example, obstructive hydrocephalus, syringomyelia, foramen magnum herniation and marked contrast enhancement of the meninges, third ventricle, mesencephalic aqueduct and brainstem have been reported with FIP.42,66,67
Cerebrospinal fluid (CSF) can be collected from neurological cases, although care should be taken as the risk of brain herniation is significant. CSF may show elevated protein concentrations (>0.30 g/l cisternal samples, >0.46 g/l lumbar samples; occasionally FIP cats show marked elevations of >2 g/l) and an increased cell count (>8 × 106 cells/l lumbar and cisternal samples; FIP cats can have counts of >1000 × 106 cells/l), with the cell type being predominantly neutrophilic, mononuclear or mixed.42,68 Some neurological cases of FIP have unremarkable CSF analysis results. Samples of CSF can also be submitted for RT-PCR for FCoV RNA and immunostaining for FCoV antigen (see below).
RT-PCR for FCoV
RT-PCR assays are available for the detection of FCoV; however, they are not specific for FIP-associated FCoVs. FCoV RT-PCR assays amplify both cell-associated subgenomic mRNA (short lengths of transcriptional RNA produced when the FCoVs replicate), as well as cell-associated or virion-associated genomic RNA, with the relative abundance of each determined by the positioning of primers (ie, where along the FCoV sequence the primers bind during PCR). 69 As viral transcription starts at the 3’ end of the FCoV genome (Figure 6) there are more subgenomic mRNAs containing viral 3’ sequence than those containing viral 5’ sequence, hence quantitative assays (ie, RT-qPCR) directed at the 5’ end of the genome (eg, viral replicase complex genes) are less susceptible to viral load overestimation than those directed at the 3’ end of the genome (eg, 7a/b non-structural protein genes).
Schematic diagram of the FCoV genome. FCoV RT-PCR assays detect FCoV RNA. The section of the genome amplified by different RT-PCRs varies depending on the position of the primers used in the assays. As viral transcription starts at the 3’ end of the FCoV genome, with the production of multiple subgenomic RNAs at this 3’ end, PCR assays with primers located at the 3’ end of the genome (eg, in the M or N regions) will be susceptible to viral load overestimation as these will amplify these subgenomic RNAs, as well as the genomic RNA present in the FCoV. Conversely, PCR assays with primers located at the 5’ end of the genome (eg, the RNA polymerase) will amplify primarily genomic RNA and will be less prone to viral load overestimation. Assays directed at the 3’ end of the FCoV genome will tend to be more sensitive in detecting the presence of FCoV, due to their ability to amplify both subgenomic and genomic RNA. Coronaviruses, including FCoV, frequently undergo mutations and recombinations, meaning that PCRs designed to be specific for particular sequences may not amplify all FCoVs. PCRs can be designed to target conserved regions of the genome to minimise this, but elimination of FCoV sequence variability as a cause of non-amplification is impossible. Modified with permission from Dr Emi Barker
Laboratories should be able to report the sensitivity and specificity of the RT-PCRs they are using to detect FCoV RNA, and the binding site of the primers they use can give some indication as to whether the assay will be prone to viral load overestimation (see above). As an RNA virus, FCoV shows a high rate of errors during replication and any viral mutations at the site of primer and/or probe binding can result in loss of PCR assay efficiency, and ultimately sensitivity. PCR conditions may be altered to tolerate such mutations, but this can result in a loss of specificity. 70
The results of RT-PCR for FCoV RNA can be reported rapidly if the laboratory used has a fast turnaround time; however, once time taken to submit the sample to the laboratory is factored in, reporting of results can still take a few days. This is usually quicker than immunostaining on tissue samples and may be quicker than immunostaining on effusion samples, but immunostaining can provide a definitive diagnosis whereas RT-PCR does not. Recently a rapid molecular technique (loop-mediated isothermal amplification) for detecting FCoV RNA in-house has been described, 71 although it suffered from poor sensitivity.
Which samples can be tested?
RT-PCR can be used to detect FCoV RNA in tissue, effusion, blood, CSF or aqueous humour samples from suspected cases of FIP (see box below). Tissue samples should not be formalin fixed, as formalin degrades the target RNA and can decrease PCR sensitivity; indeed, RNA is very sensitive to degradation and samples for research purposes are often collected into RNA preservation fluids. However, the need for special collection conditions for other samples destined for routine diagnostic purposes by RT-PCR is unproven. The presence (particularly of high levels) of FCoV RNA in blood, effusion, tissue, CSF and/or aqueous humour samples can be highly supportive of a diagnosis of FIP but, in the author’s opinion, cannot be regarded as delivering a definitive diagnosis.
FCoV RT-PCR can also be performed on faecal samples, but this is primarily used to identify cats that are shedding the virus for the management of infection in a multi-cat household. Faecal FCoV RT-PCR is not used to aid in the diagnosis of FIP, but interestingly recent studies have found that cats with FIP are more likely to be shedding FCoV, 69 and have higher amounts of FCoV RNA as determined by RT-qPCR, 29 in their faeces than cats without FIP.
Molecular techniques characterising FCoV S gene mutations
What techniques can be used?
Following the detection of FCoV RNA by RT-PCR, it may be possible to then characterise targeted sections of FCoV genomic sequences present in a sample using molecular techniques such as pyrosequencing, Sanger sequencing or PCR with sequence-specific hydrolysis probes. Such techniques are not always successful; for example, if only low levels of FCoV are present (this can preclude sequence analysis) or if FCoV sequence variability means that targeted sequencing techniques cannot generate sequence results. Characterisation of FCoV genomic sequences would be most useful if FIP-specific mutations existed, as the detection of these mutations would be diagnostic for FIP.
Is mutation analysis useful?
A 2012 paper has described amino acid differences in the fusion peptide encoded by the FCoV S gene as being markers of FCoVs associated with FIP, 26 raising the possibility that detection of the underlying S gene mutations could be used to definitively diagnose FIP. Similarly, amino acid differences in the furin cleavage motif, also encoded by the S gene, have been correlated with FIP disease. 28 However, these S gene markers were identified by comparing the sequences of FCoVs found in the tissues of FIP cats with those found in the faeces of healthy non-FIP cats.
Researchers in the author’s group at Bristol hypothesised that the fusion peptide sequence mutations could reflect the cellular tropism of the FCoV (ie, systemic monocyte/ macrophage-associated FCoV or intestinal epithelium-associated FCoV) rather than being specific for FIP, knowing that non-FIP cats can have systemic FCoV infection. They compared the S gene sequences, from the region of the previously described fusion peptide mutations, of FCoV detected in the tissues of FIP cats with those detected in the tissues of non-FIP cats. 29 This allowed evaluation of the S gene sequences of FCoVs associated with systemic FCoV infection in both non-FIP and FIP cases. They found that the S gene mutations present in most of the FIP tissues were also present in most of the tissues of non-FIP cats that had systemic FCoV infection.
A recent, more extensive, study confirmed these findings, and calculated that if the identification of S gene mutated FCoVs was included as an additional confirmatory step to the detection of FCoV by RT-PCR, this only slightly increased specificity for the diagnosis of FIP in tissue samples (from 92.6 to 94.6%). 69 Sensitivity was moderately decreased (from 89.8% to 80.9%) as non-mutated FCoVs were sometimes identified and mutation analysis was not possible in all tissue samples (eg, due to low FCoV copy numbers or the presence of type 2 FCoVs). These results question the value of S gene mutation analysis over and above the detection of FCoV by RT-PCR, particularly in view of the extra financial expense and time required to perform this additional analysis.
Analysis for S gene mutations has also been performed on effusions in recently published studies.74,75 The majority of FCoVs in the effusions of FIP cats do indeed have the mutations described. 26 In one study, 75 12/17 FCoV-positive FIP effusion samples had S gene mutations, while one did not have a mutation and four could not be sequenced due to the low levels of FCoV present. In another study, 74 32/36 FCoV-positive FIP effusion samples had S gene mutations, while three did not have mutations and one could not be sequenced.
The recent extensive study by the author’s group 69 calculated that the identification of S gene mutated FCoVs as an additional step to the detection of FCoV alone by RT-PCR did not increase specificity for the diagnosis of FIP in fluid samples (primarily effusions but also CSF and aqueous humour) – specificity stayed at 97.9% – but markedly decreased sensitivity (from 78.4% to 60%), for the same reasons as described for the tissue samples. In this study 69 all FCoV-positive samples from cats with FIP had S gene mutations in CSF samples, while no non-FIP samples were positive for FCoV. Therefore, S gene mutation analysis in FCoVs does not substantially improve the ability to diagnose FIP in effusion or fluid samples as compared with detection of FCoV RNA alone by RT-PCR.
Histopathological examination
Samples of affected tissues, such as liver, kidney or mesenteric lymph nodes, can be collected ante-mortem by ultrasound-guided percutaneous needle-core biopsy, laparoscopy or laparotomy, although the invasive nature of collection may preclude this in sick cats. Samples are often collected at post-mortem examination following euthanasia due to a high index of suspicion of FIP. Samples are evaluated for characteristic histopathological changes of FIP, which, when present, are generally regarded as being reliable for diagnosis. Immunostaining for FCoV antigen is usually also recommended to confirm the diagnosis. However, a lack of histopathological lesions is more difficult to interpret, especially in cases with a high index of suspicion of FIP, as absence of gross lesions to guide biopsy could lead to sampling of non-affected organs or tissue. 78 A small study recently documented that 5/8 FIP cases did not have histopathology changes typically consistent with FIP, even though large representative biopsies were taken; 53 diagnosis in these cases was based on positive FCoV antigen immunostaining.
Immunostaining of FCoV antigen
Immunostaining is performed on formalin-fixed tissues using IHC, or on cytological (typically effusion) samples using immuno-cytochemistry (ICC) or immunofluorescence (IF). These techniques exploit the binding of antibodies to host cell-associated FCoV antigens, which are subsequently visualised by enzymatic reactions producing a colour change (ICC) or by fluorescence (IF).
Positive FCoV antigen immunostaining of tissues is said to confirm a diagnosis of FIP (ie, it is very specific); a negative result, however, does not exclude FIP as a diagnosis as FCoV antigens may be variably distributed within lesions 78 and thus are not detected in all histopathological sections prepared from lesions from FIP cases. 30 This somewhat contradicts the suggestion by some that immunostaining is mandatory to confirm/ exclude FIP in doubtful cases, 53 though the problem may be overcome by taking multiple and/or large samples with confirmed pathology, as well as possibly requesting additional sections of biopsies with pathology to be cut and stained.
Immunostaining of effusion samples has shown variable sensitivity (ranging from 57–100%).79–84 Since this technique relies on staining FCoV antigen within macrophages in the effusion, and the effusion is often cell-poor and/or the FCoV antigen is masked by FCoV antibodies in the effusion, a false negative result may be obtained. Immunostaining was thought to be very specific; however, two (heart failure and cholangio-carcinoma cases) of seven non-FIP effusions were positive by IF in one study, 82 and eight (including two cats with heart failure and two cats with neoplasia) of 29 non-FIP effusions were positive by ICC in another, 79 raising questions about the specificity of ICC. The reported poorer specificity may be due to the methodology used in one study (ie, double staining for both FCoV antigen and macrophages [via major histocompatibility complex II staining] was used), and the suboptimal storage of slides in the other, which could cause non-specific staining and false positive results. Some have suggested that using cell pellets produced from centrifuged effusion samples to prepare formalin-fixed, paraffin-embedded samples that can then be treated like a tissue specimen for IHC, can improve the reliability of detection of FCoV antigen, 30 although the processing time required for this would be longer than for ICC.
FCoV antigen ICC staining has been reported as being successful in detecting FCoV in the CSF of a cat with neurological FIP. 85 A recent study evaluated ICC in the CSF (collected at post-mortem examination) of cats with and without FIP that presented with and without neurological signs. 86 This study found that 17/20 cats with FIP gave positive results, as did 3/18 cats without FIP, limiting the test’s specificity, although methodology may again have been an issue. These analyses excluded cases in which cellularity was inadequate for ICC to be performed. Application of ICC to CSF samples collected ante-mortem from a larger number of cats with neurological signs due to FIP and non-FIP causes would be desirable to further evaluate the usefulness of this technique.
The use of FCoV antigen immunostaining has recently been described in aqueous humour samples collected at post-mortem examination from cats with and without FIP. 87 The study evaluated FCoV ICC in aqueous humour samples from 25 cats with FIP (interestingly the majority were effusive FIP cases, and did not present with uveitis) and 11 non-FIP cats. A sensitivity of 64% and specificity of 81.8% were demonstrated. Positive results were obtained in two of the 11 control cats, one with lymphoma and one with pulmonary adenocarcinoma; neither had aqueous humour cytological features consistent with FIP (pyogranulomatous inflammation). Being able to use aqueous humour for reliable diagnostic investigations in cases of FIP would be especially valuable as it would be possible to collect this in non-effusive cases, although the sample collection technique used in this study would need to be modified (eg, smaller gauge needle) for use ante-mortem. Further evaluation of ICC on aqueous humour samples collected antemortem from cats with uveitis due to FIP and non-FIP causes would be helpful to further evaluate the usefulness of this technique.
It is possible that fine needle aspirates could also be used as samples for FCoV antigen immunostaining, but sensitivity may be poor due to difficulties in targeting lesions. Further studies would be necessary to evaluate their utility in the diagnosis of FIP and no evidence currently exists to support this.
Key Points
Look for features that could be suggestive of FIP in the history and on clinical examination: young cat, originally from a multi-cat household (including shelters or catteries), fluctuating non-responsive pyrexia, evidence of an effusion, and ocular or neurological signs.
Look for a lymphopenia on haematology.
Look for hyperglobulinaemia, hyperbilirubinaemia (in the absence of moderate to severe increases in ALT and ALP enzyme activity, or anaemia), a reduced A:G ratio and elevated AGP (>1.5 mg/ml) on serum biochemistry.
Prioritise finding and sampling any effusion, whether pleural, peritoneal or pericardial in type. Effusions due to FIP are typically clear, viscous, sticky and straw-yellow, with a total protein concentration of >35 g/l and a low A:G ratio of <0.4. AGP concentrations can also be raised as in the serum. FIP effusions are poorly cellular (usually <5 × 109 cells/l) and typically pyogranulomatous in nature.
Be aware of the value and limitations of currently available tests in achieving a definitive diagnosis of FIP. In particular: - Demonstration of FCoV antigen by immunostaining in effusions or biopsies, in association with typical cytological or histopathological features of FIP, provides for a definitive diagnosis of FIP - Detection of FCoV RNA by RT-PCR, especially if present at high levels, in diagnostic samples such as effusions, CSF and biopsies, is highly supportive of a diagnosis of FIP. - Detection of FCoV spike gene mutations following a positive FCoV RT-PCR result does not appear to offer additional information for the diagnosis of FIP.
Footnotes
Acknowledgements
ST would like to acknowledge the many contributions made by the University of Bristol Feline Coronavirus Research Group and Bristol-Zurich FIP Consortium to the viewpoints and discussions described in this review. Special thanks go to Emi Barker and Samantha Saunders for their helpful comments on this manuscript, and to Jacqui Norris and Anja Kipar for helpful discussions on FCoV antigen staining. Andrew Davidson, Anja Kipar and Stuart Siddell are also thanked for their valued contributions to past and current FCoV research. Additional thanks go to the veterinary practices, cat owners, cat breeders and rescue centres that helped in the acquisition of samples used in these research studies and to colleagues, current and past, at the Feline Centre and Veterinary Pathology Unit, Langford Vets, University of Bristol, who have assisted in obtaining samples.
Conflict of interest
ST is a member of the World Forum for Companion Animal Vector Borne Diseases, supported by Bayer Animal Health, and of the European Advisory Board on Cat Diseases, supported by Boehringer Ingelheim. She also does work for the Molecular Diagnostic Unit, Langford Vets, University of Bristol.
Funding
ST receives financial support for current infectious disease research from BSAVA Petsavers, Journal of Comparative Pathology Educational Trust, Langford Trust, Langford Vets Clinical Research Fund, Morris Animal Foundation, NERC/BBSRC/MRC, Petplan Charitable Trust, South West Biosciences DTP and Zoetis Animal Health.