Feline panleukopenia (FPL) is a highly contagious and often fatal disease characterised by acute severe enteritis, severe dehydration and sepsis due to lymphoid depletion and pancytopenia [1
]. FPL is usually associated with infection by feline parvovirus (FPV), a member of the genus Protoparvovirus
(formerly Feline panleukopenia virus
is one of eight genera of vertebrate viruses within the subfamily Parvovirinae
of the family Parvoviridae.
Collectively, FPV and canine parvovirus (CPV), along with associated variants found in various carnivore species such as mink and raccoons, constitute the species Carnivore protoparvovirus 1
Until the 1980s, FPV was the only reported viral cause of FPL in cats. FPV is able to infect cats by first binding to the feline transferrin receptor (fTfR) expressed on the surface of cells, followed by clathrin-mediated endocytosis to initiate infection [3
]. Canine parvovirus CPV-2 emerged in the late 1970s and was initially unable to infect cats, as it could not bind to the fTfR [3
]. However, infectivity for feline cells was acquired soon after by the genetic variant CPV-2a, which emerged in 1979 and replaced CPV-2 [4
]. The ability to infect cats has also been retained by subsequent antigenic variants of CPV-2a, termed CPV-2b and CPV-2c, which only differ from CPV-2a at a single amino acid position (VP2 426). These and other antigenic variants of CPV can cause FPL in both naturally acquired and experimental infections of cats [5
In contrast to parvoviral enteritis in dogs, estimated to cause 20,000 cases per year in Australia, clinical cases of FPL have rarely been diagnosed in Australia since the mid-1970s, and there have been no reports of FPL outbreaks for over 40 years [9
]. In 2014, FPL re-emerged in eastern Australia, and subsequent outbreaks occurred between 2015 and 2018 in several locations in this region. Similarly, outbreaks of FPL occurred in New Zealand (NZ) between 2016 and 2018, as well as in the United Arab Emirates (UAE) in 2017, with no outbreaks of FPL reported in either country in recent decades, likely due to the widespread use of FPL vaccines.
The contemporaneous re-emergence of FPL in geographically distinct settings long considered FPL-free has raised questions as to whether virus-related or other unknown risk factors played a role in the observed FPL outbreaks. Herein, case data and clinical samples from 989 and 113 cats, respectively, were analysed to identify the lineages of Carnivore protoparvovirus 1 responsible for the outbreaks of FPL in Australia (2014 to 2018), the UAE (2017) and NZ (2017–2018) and evaluate epidemiological factors associated with these outbreaks, including vaccination status.
An understanding of the drivers of the re-emergence of FPL, the oldest known viral disease of cats, is essential to contain this fatal infection. We provide strong evidence of multiple outbreaks of FPL in three countries. There are no previous published reports of FPL outbreaks in any of these regions for comparison, although there is anecdotal first-hand experience of FPL among veterinarians practicing in Australia from the late 1960s to the mid-1970s, before commercial vaccines were used routinely [1
]. Despite the widespread circulation of CPV in Australia, this virus was not detected in any samples tested in these FPL outbreaks. Canine parvovirus causes approximately 5% of FPL cases globally, but these have been confined to sporadic individual cases, and there are no reports of FPL outbreaks caused by CPV in multi-cat environments [8
]. Whether this is due to viral or to host factors is currently unknown and warrants further investigation.
A single FPV lineage (clade A) was responsible for disease in both outbreak regions in Victoria, Australia, even though these are geographically distant (550 km) from each other. Phylogenetic analysis revealed the closest related virus was taken from a domestic cat in Japan in 1994 (GenBank accession no. AB000050), although it is likely that a closer ancestral virus exists but has not been sampled here. The 100% sequence identity of VP2 sequences from cases in Mildura and Melbourne is consistent with the epidemiological investigation, which found evidence supporting direct viral spread through transport of unvaccinated cats for adoption in 2014 and 2015. The nonenveloped virions of FPV, shed in high titres in all excretions of infected animals, are environmentally resilient and remain viable in infected premises such as shelters for over a year [17
]. Indirect transmission by fomites is a major mechanism of parvovirus transmission among cats and dogs and was likely a crucial contributing factor to the canine global CPV pandemic in the late 1970s [19
Unexpectedly, the FPL outbreak in Sydney, NSW, was not caused by spread of the Victorian strain, but rather by a separate clade (C) comprising a relatively distinct virus [20
]. This finding, together with the detection of additional Australian FPVs (Clades B and D), is consistent with multiple independent virus entries into Australia, including the virus from an unvaccinated cat (clade D), which was seemingly derived from an FPV vaccine strain. Three recently vaccinated cats, in which the same vaccine virus was detected (clade D), likely did not have FPL, including two with Giardia
co-infections initially diagnosed with FPL from a positive commercial faecal qPCR assay. MLV vaccine virus shed in faeces can be detected in qPCR assays and CPV faecal antigen tests [22
]. Quantifying viral loads can assist in the discrimination of vaccine and field strains. However, quantification data were not provided by the commercial laboratory for these two cases. CPV vaccine viral loads in recently vaccinated dogs are 4- to 5-fold lower than those of dogs infected with field strains [24
]. Similarly, a recent report found several recently vaccinated cats had low FPV viral loads (<1.38 × 102
copies/mg of faeces) over a 28-day period of surveillance [23
Before this investigation, only one FPV VP2 sequence from Australia had been deposited in the GenBank database (FPV 193/70, GenBank accession no. X55115). The evolutionary relationships with FPV 193/70 cannot be inferred reliably, since the isolate, obtained from a 3-month-old cat in Melbourne with peracute fatal FPL, had been serially passaged in vitro before sequencing in 1990. Interestingly, FPV 193/70 falls notably close to a clade of vaccine viruses in our analysis. In addition, we generated the VP2 sequence of an FPV strain detected in a healthy feral cat from Tasmania in 2010, which was genetically distinct from outbreak strains in this study and also fell into clade D, suggesting a vaccine-related origin. These findings, together with a report of FPV detection in the bone marrow of a healthy feral cat from Melbourne in 2012 [25
], provide evidence that FPV has been circulating widely in Australia, at least since the 1970s, although the viruses are phylogenetically unrelated to the 2015–2018 outbreaks.
One possible explanation for the multiple independent FPV introductions described here is that FPV viruses circulating in the feral cat population may occasionally spill over into domestic cats. In comparison to the owned cat population, estimated at 3.8 million [26
], some 2.1 to 6.3 million feral cats (free-roaming, unowned, unsocialised cats) inhabit Australia [27
]. An FPV seroprevalence of 79% among feral cats in Victoria and NSW, documented in 1981, provides evidence of widespread FPV exposure [28
]. Viral introductions could also have been anthropogenic, for example through fomite spread during international travel.
We were also able to analyse sequence data from a contemporaneous FPL outbreak in the UAE in 2018. This is the first study to characterise full-length VP2 sequences from the Middle East. Viruses most closely related to the UAE viruses, which formed a phylogenetically distinct lineage, were from a range of locations in Europe, Asia and the Americas, but the long lineage leading to the UAE viruses suggests that we have not yet sampled their ultimate ancestor.
The VP2 sequence data from cases analysed from a third country with contemporaneous FPL outbreaks showed that the NZ FPV strains segregated with FPV vaccine viruses. The virological findings from these cases yet again highlight the difficulty of diagnosing FPL in cats that have recently received MLV vaccines, as well as raising important questions about the potential capacity of reversion to virulence of MLV vaccines. The diagnosis of FPL in cases from NZ was initially based on consistent clinical signs and a positive faecal antigen test. Point-of-care immunochromatographic assays designed to detect CPV faecal antigen also detect FPV, so are used in practice for the diagnostic investigation of cats with clinical signs consistent with FPL [22
]. However, faecal antigen tests detect both field and vaccine strains of FPV and CPV. In one study evaluating three commercial CPV faecal antigen kits, 20% of cats vaccinated against FPV in the preceding 14 days returned positive test results [29
The sensitivity of faecal antigen tests for the detection of parvoviruses is influenced by viral load in faeces, as well as by the presence of gut antibodies that can sequestrate viral particles and prevent binding to the test antibody [30
]. CPV vaccine viral loads in recently vaccinated dogs are usually below the limit of detection of antigen tests, ranging from 1.48 × 100
to 2.5 × 104
copies/mg of faeces in one study [24
] and from ≈100
to 7.50 × 105
copies/mg of faeces in another [31
]. The lower limit of detection of faecal antigen tests for CPV in dogs is 105
viral DNA copies per mg of faeces [32
]. Low titres of gut FPV antibodies in the New Zealand cats could explain the positive faecal antigen test results in those cats with low faecal viral loads.
All but one of the VP2 sequences obtained from the New Zealand cases were identical to the FPV vaccine strain that the cats had been inoculated with. Co-infection with other pathogens including Feline coronavirus or Giardia
spp. could account for the clinical signs in five of these cases, which, although consistent with FPL, were non-specific. However, in one cat, high loads of vaccine strain virus were detected concurrently in lymph node and faeces (FPV_221, Table 3
), which is more consistent with loads attained by pathogenic virus strains than with those attained by vaccine virus strains. Further, unequivocal support of the diagnosis of FPL in this case was obtained from histological examination of the intestine, where lesions pathognomonic for FPL were identified (Figure 5
). An FPV-vaccine strain was detected in this cat, and the sequencing chromatogram showed no evidence of a coinfecting FPV or CPV. However, deep sequencing of the sample with examination of the metatranscriptome or metagenome would be required to rule out co-infection with other pathogenic viruses not included in the qPCR test panel. An alternative explanation is that there was reversion to virulence of the vaccine strain, although this has not been described previously in cats.
Molecular analysis of faeces from dogs displaying signs of acute gastroenteritis shortly after CPV vaccination ruled out reversion to virulence of CPV MLV, since co-infection with field strains of CPV or other pathogens including Canine coronavirus, Canine distemper virus and Isospora canis
was detected [33
]. In three pups in which only the vaccine virus was detected, CPV vaccine loads were lower than those associated with enteritis from CPV field strains, ranging from 1.03 × 103
to 1.78 × 105
DNA copies/mg of faeces [33
The re-emergence of FPL in two Australian states in relatively quick succession over a three-year period, caused by two distinct FPV lineages, raises the question as to why this disease is re-emerging now, after decades of apparent quiescence. Animal shelters are conducive environments for pathogen emergence or re-emergence because of a large number of susceptible hosts living in a confined area. This is exacerbated by factors including young age, immunological naivety or immunosuppression, close contact and co-morbidities such as heavy parasite burdens. Similarly, suboptimal biosecurity protocols favour pathogen persistence and fomite transmission [34
Failure to vaccinate or inadequate or inappropriate vaccination in shelter-housed cats were major contributors to the re-emergence of FPL in Australia. Before these outbreaks, over half of the shelters from which cases were derived did not administer FPV vaccinations. In general, FPV vaccines are a highly effective tool in parvovirus control. For example, protective FPV antibody titres were measured in 95% of vaccinated cats in a field study of adult cats after MLV vaccination, of which 64% had pre-existing protective titres [35
]. In addition, viral challenge studies of FPV-vaccinated cats usually show vaccine efficacy >90% [36
]. Although state legislation mandates vaccination of owned cats in commercial boarding facilities in NSW and Victoria, vaccination is not enforced for cats admitted to private or municipal animal shelters, until they are sold for rehoming. Most shelters that did vaccinate, completed the primary course of vaccinations in kittens by 12 to 14 weeks of age and used inactivated vaccines, in contravention of the current guidelines. The World Small Animal Veterinary Association (WSAVA) Vaccination Guidelines recommend the use of MLV vaccines in shelter-housed cats for rapid induction of long-lasting humoral immunity [22
]. A minimum age of 16 weeks for the final vaccination in the primary course is recommended, based on the failure of 45% of kittens to seroconvert after vaccinations at 8 and 12 weeks of age with the commercial FPV MLV vaccination [37
]. This immunity gap relates to persistent maternally derived antibodies (MDA) in kittens, which decline below protective titres yet effectively neutralize the vaccine virus and thereby prevent seroconversion [38
]. Persistent MDA would explain the presence of vaccination non-responders in our study that were <16 weeks of age at the time of diagnosis of FPL, highlighting that vaccination alone cannot prevent disease in environments where susceptible kittens are exposed to FPV.
Among the 21% of vaccination non-responders that were older than 16 weeks, some were likely exposed to FPV prior to seroconversion, either before or after admission to the shelter, as indicated by the median vaccination-to-diagnosis interval of 7.5 days. The incubation period for FPL is between 2 and 10 days [17
]. Other reasons for apparent vaccine failure include persistence of MDA for >16 weeks in kittens born to queens with high anti-FPV antibody titres or failure to mount an immune response to vaccination because of immunodeficiency (acquired or genetic), incorrect vaccine administration or storage errors [35
The median time from shelter admission to diagnosis of FPL was 9 days. The minimum holding time for unidentified cats in municipal shelters in the states of Victoria and NSW is 7 days, after which cats can be euthanised or rehomed (Domestic Animals Act 1994, Companion Animals Act 1998). Changes in euthanasia practices resulting in longer length of stay (LOS) may have contributed to a cumulative increase in selection pressure for FPL re-emergence, especially in shelters where vaccination was not practiced. The likelihood of exposure to infectious diseases increases within a shelter environment with increasing LOS [39
]. National data from Royal Society of Prevention to Cruelty (RSPCA) shelters show a progressive decline in euthanasia rates from 74% in 1995 to 27% in 2018 [40
]. Similar declines in euthanasia have occurred in shelters in other countries such as Canada [42
A seasonal pattern of FPL was identified in Australia and New Zealand, with an annual peak in case numbers from December to May. This pattern may reflect waning maternal immunity lagging two to three months behind the peak kitten births. Similar seasonality of FPL outbreaks was reported in the Northern Hemisphere, where seasons are inverted; case numbers peaked from July to December and were correlated with an influx of kittens born annually in spring and waning of maternal immunity over the subsequent two to three months [43
The duration of faecal shedding of FPV, which is relevant to infection control, has been inadequately studied. This study demonstrates that virus can be shed in faeces for an extended period of up to 14 weeks from diagnosis. Faecal shedding of vaccine virus was detected by qPCR post vaccination for the duration of one four-week study [23
]. Shedding of field CPV strains in dogs has been detected by qPCR for 45–50 days [44
], while CPV vaccine virus was shed for 19–22 days on average [31
]. Further studies are warranted in cats to further define the duration of faecal shedding of both field and vaccine strains of FPV detectable by qPCR.