Feline babesiosis : signalment , clinical pathology and concurrent infections

INTRODUCTION To date, 6 piroplasms of the cat family have been reported: Babesia felis in an African wild cat in Sudan (Felis ocreata, syn. Felis sylvestris) in 1929; Babesiella felis in a puma (Felis concolor) in 1934; Nuttallia felis var. domestica in a domestic cat in 1937; B. cati in an Indian wild cat (Felis catus) in 1950; B. herpailuri in a jaguarundi (Herpailurus yaguarundi) in South America in 1967; and B. pantherae in a leopard (Panthera pardus) in Kenya in 1972. Confusion surrounding this nomenclature led to a proposal by Dennig and Brocklesby that all feline piroplasms should be divided into either of 2 small (B. felis and B.cati) or 2 large Babesia spp. (B.herpailuri and B. pantherae). Of the small Babesia spp., B. felis has been reported to occur most commonly in domestic cats, but is believed to have a wide host range within the cat family. However, a small piroplasm recently isolated from lions (Panthera leo) in the Kruger National Park, South Africa, was found to be morphologically similar to, but serologically distinct from B. felis found in domestic cats in South Africa. In addition, the original B. felis parasite isolated from an African wild cat was found to be transmissible to domestic cats, but did not cause clinical illness. By contrast, the B. felis parasite found in domestic cats in South Africa appears to be morphologically similar, but it causes a distinct clinical illness that is potentially fatal. This raises the question whether B. felis of domestic and wild felids is, in fact, a single species. Babesiosis of domestic cats has been reported sporadically from other countries, including France, Germany, Thailand and Zimbabwe, but it does not seem to be a regularly occurring significant clinical disease in any country other than South Africa. In a recent survey in South Africa, feline babesiosis was found to be endemic along most of the South African coast, from KwaZulu-Natal to the south-western Cape. Clinical cases are occasionally seen in non-endemic parts of the country. These usually involve pets that have returned home after coastal holidays with their owners. An isolated focus of naturally occurring feline babesiosis has, however, recently been identified at Kaapschehoop, a village on the escarpment west of Nelspruit, Mpumalanga. In 1980, Futter and Belonje published a series of articles on feline babesiosis. The first dealt with the history and classification of the disease, while the second article described clinical observations in 20 experimentally-infected and 70 naturally-infected cats. Lethargy, anorexia and anaemia were consistent findings in both study groups and icterus was occasionally observed. Pyrexia was not a feature of the disease. The third article described the haematological findings in the same study groups. Cases showed a rapid fall in haematocrit, haemoglobin and erythrocyte count, and the erythrocytes were often macrocytic and hypochromic. No significant changes were seen in total leukocyte counts. The fourth article described the chemical pathology and macroscopic and microscopic post mortem findings. Total serum protein concentration was unchanged, but there was an increase in globulin concentration and a decrease in and globulin concentration. Moderate elevation in hepatic alanine transaminase (ALT) levels was recorded in a number of cases. Renal function was unaffected and venous blood pH remained within the normal range throughout the study. Post mortem findings included extreme pallor of the viscera, thin watery blood and yellow to orange rectal faeces. Bile stasis and hepatic necrosis were evident in some cases, while marked icterus was seen in only 2 cats. As no further studies on naturallyinfected cases of feline babesiosis in South Africa had been performed since 1981, the purpose of this study was to review the signalment and clinico-pathological changes associated with naturallyoccurring disease and to investigate certain parameters that were not studied previously, such as thrombocyte counts, ALP and GGT activities and serum electrolyte status. The study was also undertaken to identify concurrent infections and IMHA and to measure and compare peripheral and central parasitaemias, none of which had previously been investigated in feline babesiosis.


INTRODUCTION
To date, 6 piroplasms of the cat family have been reported: Babesia felis in an African wild cat in Sudan (Felis ocreata, syn.Felis sylvestris) 56 in 1929 8 ; Babesiella felis in a puma (Felis concolor) in 1934 5 ; Nuttallia felis var.domestica in a domestic cat in 1937 21 ; B. cati in an Indian wild cat (Felis catus) in 1950 40 ; B. herpailuri in a jaguarundi (Herpailurus yaguarundi) in South America in 1967 9 ; and B. pantherae in a leopard (Panthera pardus) in Kenya in 1972 10 .
Confusion surrounding this nomenclature led to a proposal by Dennig and Brocklesby 10 that all feline piroplasms should be divided into either of 2 small (B.felis and B.cati) or 2 large Babesia spp.(B.herpailuri and B. pantherae).Of the small Babesia spp., B. felis has been reported to occur most commonly in domestic cats, but is believed to have a wide host range within the cat family 26 .However, a small piroplasm recently isolated from lions (Panthera leo) in the Kruger National Park, South Africa, was found to be morphologically similar to, but serologically distinct from B. felis found in domestic cats in South Africa 30 .In addition, the original B. felis parasite isolated from an African wild cat was found to be transmissible to domestic cats, but did not cause clinical illness 8 .By contrast, the B. felis parasite found in domestic cats in South Africa appears to be morphologically similar, but it causes a distinct clinical illness that is potentially fatal 13,47 .This raises the question whether B. felis of domestic and wild felids is, in fact, a single species.
Babesiosis of domestic cats has been reported sporadically from other countries, including France 2,25 , Germany 39 , Thailand 24 and Zimbabwe 51 , but it does not seem to be a regularly occurring significant clinical disease in any country other than South Africa.In a recent survey in South Africa 23 , feline babesiosis was found to be endemic along most of the South African coast, from KwaZulu-Natal to the south-western Cape.Clinical cases are occasionally seen in non-endemic parts of the country.These usually involve pets that have returned home after coastal holidays with their owners 23,47 .An isolated focus of naturally occurring feline babesiosis has, however, recently been identified at Kaapschehoop, a village on the escarpment west of Nelspruit, Mpumalanga 45 .
In 1980, Futter and Belonje published a series of articles on feline babesiosis [12][13][14][15] .The first dealt with the history and classification of the disease 12 , while the second article described clinical observations in 20 experimentally-infected and 70 naturally-infected cats 13 .Lethargy, anorexia and anaemia were consistent findings in both study groups and icterus was occasionally observed.Pyrexia was not a feature of the disease.The third article described the haematological findings in the same study groups 14 .Cases showed a rapid fall in haematocrit, haemoglobin and erythrocyte count, and the erythrocytes were often macrocytic and hypochromic.No significant changes were seen in total leukocyte counts.The fourth article described the chemical pathology and macroscopic and microscopic post mortem findings 15 .Total serum protein concentration was unchanged, but there was an increase in ( globulin concentration and a decrease in " and $ globulin concentration.Moderate elevation in hepatic alanine transaminase (ALT) levels was recorded in a number of cases.Renal function was unaffected and venous blood pH remained within the normal range throughout the study.Post mortem findings included extreme pallor of the viscera, thin watery blood and yellow to orange rectal faeces.Bile stasis and hepatic necrosis were evident in some cases, while marked icterus was seen in only 2 cats 15 .
As no further studies on naturallyinfected cases of feline babesiosis in South Africa had been performed since 1981, the purpose of this study was to review the signalment and clinico-pathological changes associated with naturallyoccurring disease and to investigate certain parameters that were not studied previously, such as thrombocyte counts, ALP and GGT activities and serum electrolyte status.The study was also undertaken to identify concurrent infections and IMHA and to measure and compare peripheral and central parasitaemias, none of which had previously been investigated in feline babesiosis.

Experimental animals
Fifty-six cats with naturally-occurring B. felis infection, presented sequentially to private veterinarians over a 2-month period, were prospectively studied.The study area was concentrated around George and surrounding towns in the southwestern Cape Province.Cats were included, subject to the owners' written consent, if B. felis parasites were identified on a thin peripheral blood smear.Peripheral blood smears were prepared using a drop of capillary blood taken from the ear pinna and then fixed and stained with Cam'sQuick (CA Milsch).Blood samples were collected from the jugular vein into EDTA and serum vacuum tubes (Becton Dickinson Vacutainer Systems).

Clinico-pathological laboratory measurements
Within 6 hours of collection, EDTA samples were used to prepare thin central blood smears (as described above), to perform in-saline agglutination tests using a standard technique 55 , and to determine full haematology using a portable COULTER ® A c "T diff™ Haematology Analyzer Veterinary Applications counter (Beckman Coulter S.A.) The veterinary applications software was set for analysis of feline blood.
Serum was separated and frozen at -20 °C for a maximum of 2 months.On completion of the data collection phase, the serum samples were allowed to thaw and reach room temperature.Total serum protein, albumin, urea, creatinine, alanine transaminase (ALT), alkaline phosphatase (ALP), gamma glutamyltransferase (GGT) and total bilirubin were determined on a Technicon RA-1000 system (Technicon Instruments Corporation).Serum globulin was calculated by subtracting albumin from total serum protein.Serum electrolytes (sodium, potassium) were measured using a Rapidlab™ 348 pH/Blood Gas Analyzer (Chiron Diagnostics).Serum protein electrophoresis was performed on a Beckman Microzone ® Electrophoresis System (Beckman Instruments), using a cellulose acetate membrane and Barbital buffer (0.5 ionic strength and pH 8.6).This system identifies alpha ("1 and "2), beta ($1 and $2) and gamma (() peaks.
Relative differential white blood cell counts, nucleated erythrocyte counts and platelet scores were determined manually using central blood smears under ×50 or ×100 oil magnification.Leukocytes were identified until 100 to 200 cells had been classified by type, and expressed as a percentage.The number of nucleated erythrocytes (nRBC) per 100 leukocytes was recorded.The automated white cell count (WCC) was corrected for nRBC using the formula 11 : Corrected WCC = (initial WCC × 100) / (100 + nRBC) Platelet scores were used to verify the automatic platelet count, as feline blood is notably prone to clotting during collection 11 .Platelet numbers were scored as normal (8-10 per ×100 field), increased or decreased (<3-4 per ×100 field) 11 .The presence of platelet clumps on stained smears were also noted and used to refine the score.
All blood smears, both peripheral and central, were prepared by the same investigator (TS), using a consistent technique.Percentage parasitaemia was estimated on central and peripheral blood smears.Parasitised red blood cells (pRBC) in 10 fields (x100 objective), each containing approximately 400 RBC, were counted and the percentage of pRBC calculated by dividing the total number of pRBC by 40 (400 RBC per field, 10 fields examined).The presence of any other haematological parasites was also recorded.

Serology
A commercial ELISA test kit (SNAP Combo Plus Feline Leukemia Virus Antigen/Feline Immunodeficiency Virus Antibody Test Kit, IDEXX Corp.) was used to determine the presence of concurrent infections with feline immunodeficiency virus (FIV) and/or feline leukemia virus (FeLV).The kit detects FeLV groupspecific viral core antigen (p27) and/or specific antibodies directed to FIVgag or FIVenv(gp40) proteins 18 .A very faint colour reaction in the FeLV test well was reported as an equivocal positive result.Where possible, cats with a positive (whether equivocal or clear-cut) FeLV test were retested after 12 weeks to determine whether FeLV antigenaemia was persistent or transient.Cats that were positive for FIV infection only were not retested.

Data analysis
Statistical analysis of the data was performed on a commercial statistical software package, SigmaStat™ v2.0.3 (Jandel Scientific Software).Normality of the data was tested using the Kolmogorov-Smirnov test and equal variance was tested using the Levene median test.Because the data obtained in most instances were either discrete variables or percentage data, or were not normally distributed, non-parametric methods of analysis were used.The Spearman rank order correlation was used to determine the strength of association between selected variables.The Mann-Whitney rank sum test was used for comparisons between 2 groups.For all tests, the probability value for significance was set at P < 0.05.

Signalment
Ages of affected cats ranged from 6 months to 13 years, with a median of 2 years (Fig. 1).Most cats (80 %; 45/56) were less than 3 years old, and most did not represent a specific breed, with 77 % (43/56) being either of the domestic shorthair (32/43) or the domestic longhair (11/43)   females (32/56) and 43 % males (24/56) being affected.Typical clinical signs attributable to feline babesiosis were observed in 77 % (43/56) of cats and consisted mainly of anorexia, depression and anaemia.Less common signs, such as weight loss, icterus, constipation and pica, were also observed in some cats.The other 23 % (13/56) of cats did not display typical clinical signs of the disease and included 9 cats (16 %) that were examined for reasons other than illness, such as routine pre-surgical examinations.As these cats were reported to be healthy at the time of examination, they were likely to be subclinical carriers of the disease.

Parasitaemia
Peripheral parasitaemia was assessed in 55 cats (the other was not possible due to poor quality of the peripheral blood smear) and ranged from 0.3 % to 42.3 %, with a median of 5.9 %.Central parasitaemia was assessed in 56 cats and ranged from 0.2 % to 41.4 %, with a median of 6.4 %.Central and peripheral parasitaemia showed a strong positive correlation (rs = 0.997, P < 0.05) (Fig. 2).A statistically significant, but weak, negative correlation between central parasitaemia and haematocrit was also seen (rs = -0.595,P < 0.05).This negative correlation appeared to be most pronounced when parasitaemia was higher than 20 %, with 90 % (9/10) of these cats having a haematocrit (Ht) of 15 % or less.When parasitaemia was less than 20 %, Ht levels tended to be more randomly distributed.

Haematology and chemical pathology
Laboratory data for the study population are summarised in Tables 1 and 2. The most frequent red blood cell parameter finding was macrocytic, hypochromic, regenerative anaemia, which occurred in 57 % (32/56) of the cats.Normal Ht values occurred in 43 % (24/56) of the cats, while moderate anaemia (Ht 15-24 %) occurred in 23 % (13/56).Severe anaemia (Ht < 15 %) was seen in 34 % (19/56) of cases.Nucleated red blood cells were present in 70 % (39/56) and counts ranged  from 1 to 814 nRBC/100 white blood cells, with a median count of 10.5.Signs of bone marrow erythropoiesis were commonly evident on blood smear examinations and included the presence of reticulocytes, numerous nucleated red blood cells, marked anisocytosis, polychromasia, an increased number of Howell-Jolly bodies (in excess of 1 % of RBC) and basophilic stippling of some red blood cells.
In-saline agglutination tests showed positive agglutination of RBC in 16 % (9/56) of cats.On examination of blood smears, it was evident that the vast majority of agglutinating RBC were non-parasitised, and that not only mature RBC, but also reticulocytes showed evidence of agglutination.
Ht values for in-saline-positive cats ranged from 8.6 % to 19.1 %, with a median of 13.2 %.Ht values for in-salinenegative cats ranged from 7.9 % to 41.2 %, with a median of 24 %.The difference in Ht values between in-saline-positive and in-saline-negative cats was statistically significant (T = 137.00,P = 0.008).
Corrected WCC were recorded in 54 cats and varied between 2.6 and 41.8 × 10 9 / .Of these counts, 72 % (39/54) were within the normal reference range.Leukocytosis was present in 11 % (6/54) and leukopaenia in 17 % (9/54).These leukocyte changes were caused by changes in absolute neutrophil, lymphocyte and monocyte counts, but no consistent pattern of changes was seen.The abnormal leukograms were classified as a combination of inflammatory, stressinduced and physiologic leukograms.
Automated platelet counts were recorded for 55 cats, with no thrombocyte count given by the Coulter Analyzer for 1 cat.From the automated counts, throm-bocytopaenia was evident in 98 % of the cats (54/55).Using the manual platelet score, only 25 % of the cats (14/56) were thrombocytopaenic, and 71 % (40/56) were scored as normal.Two of the cats (4 %) were thought to have thrombocytosis.Platelet clots were reported in 61 % (34/56).Only 27 % (15/55) of the cats were categorised in the same way by both manual and automated systems.Of the other 40 samples, 73 % (40/55) had visible platelet clots on central smears.
Total serum protein values were elevated in 32 % (18/56) of cats.Hyperalbuminaemia was recorded in 45 % (25/56) of the cats, while hypoalbuminaemia was recorded in only 1 cat.Hyperglobulinaemia was present in 23 % (13/56) of cats, in all cases due to polyclonal gammopathies.Of the 77 % (43/56) of cats with normal total globulin concentrations, abnormal globulin fractions were recorded in 33 % (14/43).These included various combinations of abnormal ", $ and ( globulin concentrations, but no consistent pattern of changes was seen. Increased alanine transaminase (ALT) activity was present in 89 % (50/56) of cats.However, increased alkaline phosphatase (ALP) and gamma glutamyltransferase (GGT) activities were recorded in only 25 % (14/56) and 4 % (2/56) of cats respectively.No correlation could be found between ALT and ALP (rs = 0.002, P = 0.985).A statistically significant, but weak, negative correlation was found between ALT and Ht (rs = -0.497,P < 0.05).Total bilirubin concentrations were elevated in 86 % (48/56) of cats, but clinically visible icterus was observed in only 12.There was a statistically significant difference in ALT values between cats with and without visible icterus (T = 588.00,P < 0.001), with much higher values in the icteric cats (Fig. 3).A significant positive correlation was found between ALT and total bilirubin concentrations (rs = 0.708, P < 0.05).
Renal parameters were within normal limits in most cats.Serum urea concentrations were increased in 25 % (14/56) of cats and decreased in 20 % (11/56), while increased creatinine concentrations were recorded in 25 % (14/56).Of the 14 cats with increased urea concentrations, only 6 had concurrently increased creatinine  concentrations, while creatinine concentrations were normal in the remaining 8 cats.No statistically significant relationship was found between urea and creatinine concentrations (r s = 0.211, P = 0.117).

Concurrent infections
Eight cats (14 %) tested positive for FIV infection and 18 cats (32 %) tested clearly positive for FeLV infection.Of these, 5 cats (9 %) had concurrent FIV and FeLV infections.In addition to this, 6 further cats (11 %) had equivocal test results for FeLV.After a period of 12 weeks, an attempt was made to retest all 24 cats that were positive or equivocal for FeLV infection.Of the original 18 cats that were clear-cut FeLV positives, only 8 were known still to be alive after the 12-week period, while 3 were lost to follow-up and the remaining 7 had died.Four of these died naturally while the other 3 were euthanased on account of their moribund state at the time when babesiosis was diagnosed, or shortly thereafter.Of the 6 cats that had equivocal FeLV results, only 4 were still alive after the 12-week period.All 12 remaining cats were retested.Only 2, both of which were clear-cut FeLV positives originally, tested positive for FeLV the second time, while the remaining 10 were negative.
Concurrent infection with Haemobartonella felis was detected on both peripheral and central venous blood smears of 11 % (6/56) of cats.Of the 6 cats with H. felis, 50 % (3/6) also tested positive for concurrent FeLV infection.

DISCUSSION
The high prevalence of babesiosis in young cats is in agreement with previous findings 13 , and supports the possibility that cats in endemic areas contract the infection early in life and become subclinical carriers in a state of so-called premunity.The recommended treatment for feline babesiosis, primaquine phosphate (Primaquine, Kyron), does not sterilise the infection 44,47 , and would thus promote development of premunity 43 .It is possible that clinical infection in older cats in endemic areas would therefore occur either in situations where this premunity never developed or under certain conditions where the pre-existing protective immunity became suppressed.This phenomenon has been shown to exist in dogs 32 .In this study, most affected cats older than 3 years had a concurrent illness or infection that could have influenced the cat's immune system.No breed or sex predisposition was evident, but Siamese cats seemed to be over-represented amongst purebred cats.
The close correlation between central and peripheral parasitaemia in this study indicates that B. felis-parasitised RBC do not sequester in capillary beds, which is different from what has been described in the dog 52 .The high parasitaemia observed in cats in this study is in agreement with previous reports 13,21,49 .The chronicity of feline babesiosis, and its relatively low virulence, are probably related to both of the above phenomena.It has been suggested that Plasmodium falciparum malaria in man, which sequesters, is more pathogenic than P. vivax, which does not, owing to a combination of local tissue hypoxia caused by blockage of small blood vessels, and high local production of inflammatory cytokines near sequestered parasites 7 .The high parasitaemia in cats is probably due to a combination of the relatively low inherent virulence of the parasite, the fact that all parasitised RBC presumably circulate and are thus visible, and the cat's relative resistance to endotoxin 41 .Species that are refractory to endotoxin are also less susceptible to babesiosis and show symptoms of disease at a higher parasitaemia than endotoxin-sensitive species 6 .
Haemolytic anaemia (both intra-and extravascular) is a typical finding in canine babesiosis 52,53 .Based on previous findings of haemoglobinuria 15 , splenomegaly 23 and erythrophagocytosis of both infected and non-infected cells by mononuclear cells 13 , it can be assumed that both intravascular and extravascular haemolysis also occur in feline babesiosis.The cats in this study with severe anaemia (<15 %) were moderately or severely depressed at presentation, but seemed to have an ability to adapt to the severe anaemia.This has been described previously in feline babesiosis 13 , as well as in other forms of anaemia in cats 3,54 .A surprisingly large number of cats in this study presented without anaemia.It is possible that increased owner awareness of the disease in the area could have led to earlier recognition of signs of disease and subsequent presentation for treatment before the disease had progressed very far.This emphasises the importance of a blood smear examination during any routine clinical examination of cats, especially in endemic areas.This group also included subclinical carrier cats, in which babesiosis was diagnosed incidentally at the time of examination.As these cats were not suffering from clinical disease, normal haematocrit values were to be expected.
Auto-agglutination of red blood cells has not previously been described in feline babesiosis, and probably represents secondary immune-mediated haemolytic anaemia (IMHA).The other hallmarks of IMHA, spherocytosis and positive direct Coombs' test 16 , could not be identified, as spherocytes are difficult to identify in feline blood 16 and the feline Coombs' test is not available in South Africa.Agglutination of RBC was differentiated from rouleaux formation by its failure to disappear on dilution with saline 55 .Immune-mediated haemolytic anaemia in cats has rarely been documented, but has been described secondary to Haemobartonella felis or FeLV infections and certain anti-thyroidal medications 4,19,46 .Six of the 9 cats with IMHA were FeLV-positive, while no infections other than babesiosis could be identified as a cause of IMHA in the other 3.Although cats with IMHA had lower Ht values than those without, the clinical significance of the agglutination is unclear, as all these cats responded predictably to antibabesial therapy without addition of immunosuppressive therapy.
Neutrophilia is a consistent feature of canine babesiosis 48 , and a leukaemoid response has also been described 28 .Neutrophilia was not a consistent finding in feline babesiosis.In fact, the leukocyte abnormalities in this study were more likely to be a result of concurrent problems or diseases than the babesiosis itself, as has been previously reported 14 .
Certain problems can occur when counting cat platelets electronically.Blood samples taken from cats have a tendency to form platelet clumps easily and this will potentially generate grossly inaccurate (low) counts when counted electronically 11 .This study confirmed that automated platelet counts in cats should be treated with caution, and an additional manual platelet score should always be peformed.Although thrombocytopaenia is a consistent feature of canine babesiosis 20,53 , it was not a consistent finding in feline babesiosis.
Serum protein values tended to be either normal or increased in most cats.Hyperalbuminaemia, a relatively common finding in this study, was most likely due to dehydration, as it was mostly associated with concurrent hyperglobulinaemia and a normal albumin:globulin ratio 11 .Apart from 1 cat, hypoalbuminaemia was not seen in the cats in this study, which is in contrast to what has been reported for canine babesiosis 29 .A polyclonal gammopathy was identified in all cases with hyperglobulinaemia.Various combinations of abnormal ", $ and ( globulin concentrations were seen in these cases, but no consistent pattern of changes could be identified.Elevated " globulin concentrations are usually associated with an increase in acute-phase proteins, which are synthesised by the liver as part of the acute-phase response to tissue injury, infection or immunological disorders 11 .The increase in " globulins that was seen in this study could therefore have been an indication of an acute-phase response to babesiosis, but the potential effect of concurrent diseases on " globulin concentrations must also be considered.Elevated $ and ( globulin concentrations are usually indicative of increased production of immunoglobulins (chronic phase proteins) 11 and could be ascribed to the antibody response of the reticuloendothelial system of the patient to Babesia antigens.The increase in ( globulin concentrations corresponds with previous findings 15 .
Marked increases in hepatic cytosolic enzyme activities were seen in most cats in this study, whereas hepatic membrane-associated enzymes were not commonly affected.Feline babesiosis therefore appears to be associated with significant primary hepatocellular involvement or damage 11 .Repeat testing would have been necessary to determine whether this hepatic injury was reversible.In a previous study, ALT activities returned to normal during the recovery stage in most cases of feline babesiosis, but a few cases showed an increase in activity over time 15 .The hepatocellular damage is likely to result primarily from anaemia, as anaemia can lead to hepatocellular hypoxia, which in turn causes progressive hepatic centrilobular necrosis and hepatocellular cytosol leakage with increased enzyme activity 11 .Centrilobular hepatic necrosis is a consistent pathological finding in both feline 15 and canine babesiosis 17 .However, the somewhat inconsistent relationship between ALT and haematocrit, despite a statistically significant negative correlation, indicates that anaemia is probably not the only cause of hepatocellular injury and that other factors, such as inflammatory cytokines, could also play a role 38 .Any primary hepatocellular injury can potentially be accompanied by cellular swelling that can compress bile canaliculi and cause some secondary cholestasis 11 .This is likely to be the reason for cholestasis and resultant increases in ALP and GGT activity in a few cases in this study.
Even though hyperbilirubinaemia was seen in most cats in this study, clinical icterus was observed in only 12 cats.The hyperbilirubinaemia was most likely a result of haemolysis (both intra-and extravascular), but secondary hepatocellular disease and intrahepatic cholestasis (as evidenced by increased ALP and GGT activities in some cases) were also likely to contribute.A combination of unconjugated and conjugated bilirubin would thus be expected in these cases 11 .In a previous study of feline babesiosis, hyperbilirubinaemia was also described in most of the cats 15 .At total bilir ubin concentrations of about 20 µmol/ the unconjugated bilirubin values were higher than conjugated, whereas at higher total bilir ubin concentrations the unconjugated and conjugated values were about equal 1 5 .Hyperbilirubinaemia has also been reported during acute disease caused by B. canis in dogs 20 .As in cats, it seems to be a reflection of both erythrocyte destruction and intrahepatic cholestasis 35 .The higher ALT values in icteric cats, compared to those in non-icteric cats, support the concept that clinical icterus requires hepatocellular damage to be present and is not merely a function of haemolysis 50 .Similar to what has been described for dogs in general 50 and observed for cats with babesiosis in this study, it is also believed that haemolysis alone will not lead to clinical icterus in dogs with babesiosis 35 .Instead, the degree of icterus observed in dogs with babesiosis will correlate with the degree of functional impairment, hepatocellular damage and bile stasis of the liver 33,34 .
No remarkable changes in renal parameters occurred in most cats in this study, indicating that, as in the dog 22 , and as reported previously 15 , marked renal damage and/or renal failure is a rare event in feline babesiosis.Only 2 cats (10 and 3 years old) had severe azotaemia and concurrent hyperkalaemia, which were indicative of renal failure.Although babesiosis could have been the primary cause for renal failure in these cats, it is likely that underlying renal disease was present and that acute decompensation with subsequent renal failure was precipitated by the babesiosis, as has been described in dogs 36 .
Although no consistent changes in serum electrolyte concentrations were identified in cats with babesiosis, a variety of electrolyte disturbances did occur in a substantial number of cases.It is therefore advisable to measure electrolytes in any cat affected with babesiosis, as some individual cases might require specific treatment.
A surprisingly high prevalence of FeLV and FIV was seen in this study population, but without background data it is not clear whether this reflects the epidemiology of these diseases in the study area or increased susceptibility of positive cats to feline babesiosis.The overwhelming majority of cats that seroconvert to FIV following infection remain infected 1,42 , and the cats that tested FIVpositive in this study were therefore considered to be persistently infected.However, only about 30 % of all cats that are exposed to FeLV will become persistently viraemic, while the other 70 % will seroconvert to a negative state 27,42 .As a direct relationship exists between a sample's FeLV antigen concentration and the degree of colour change in the test well, equivocal FeLV test results probably reflect low antigen concentration 31 .The interpretation of an equivocal test result is difficult and current recommendations are that the cat should be retested after at least 1 month 31,37 .As equivocal test results were obtained for 6 cats in this study, proper interpretation of those results was not possible unless the cats were retested.Because the FeLV ELISA test detects the presence of antigen as early as the primary viraemic stage, it can neither indicate nor predict whether the cat is transiently or persistently infected.To determine whether a cat is persistently infected, an IFA test can be done immediately (a positive IFA result is highly predictive of persistent infection), or the cat should be retested with ELISA after a minimum of 6-8 weeks 42 .A second positive test result is indicative of persistent infection.Most of the cats in this study that were retested after 12 weeks were negative for FeLV the second time.These cats had probably seroconverted to a negative FeLV state, but it is also possible that the initial positive FeLV results included some false positives.The fact that most of the cats in this study population were young might have artificially increased prevalence of FeLV (which is more prevalent in young cats 1 ) and decreased that of FIV (which is more prevalent in older cats 1,37 ).However, the mean ages of infected cats, 3.1 and 5.2 years for FeLV and FIV respectively, correspond with those documented 1,37 .
The differentiation between feline babesiosis and haemobartonellosis is difficult, as both diseases cause clinical signs of anorexia, depression, regenerative anaemia, weakness, weight loss and occasional icterus during the acute phase of the disease 4,13,19 .Also, in both diseases the total and differential white blood cell counts are quite variable and of limited diagnostic value.Autoagglutination of erythrocytes in cases with haemobartonellosis has been well described, but, as in cats with babesiosis in this study, the clinical significance is not clear 4,19  always, be diagnosed on stained blood smears and appear as small rings, cocci and rods that are attached to the erythrocytes.Differentiation of H. felis parasites from B. felis parasites, Howell-Jolly bodies and basophilic stippling on a blood smear is problematic, as all of these can have a very similar appearance 19 .Additionally, H. felis parasites only appear in the blood in a cyclical manner within discrete parasitaemic episodes 19 .Generally, these organisms are present in adequate numbers to be easily recognised on stained blood smears only about 50 % of the time during the acute phase of the disease, thus the absence of H. felis organisms from the blood does not definitively rule out a diagnosis of haemobartonellosis 4,19 .It is therefore possible that more than 6 cats in this study had concurrent babesiosis and haemobartonellosis.In the light of the above, it may be advisable to treat for both diseases simultaneously whenever co-infection is suspected in a case of feline babesiosis.
In conclusion, this study indicated that feline babesiosis is typically a subacute to chronic disease, often of young adult cats, that results in haemolytic anaemia and hepatocellular damage.It is frequently associated with concurrent infections including FeLV, FIV and H. felis, and an attempt should be made to identify such concurrent infections when chronic carrier adult cats become symptomatic, when cats are pyrexic at presentation, or when cats have recurrent infections that do not respond predictably to therapy.

6 0038 - 11 Fig. 2 :
Fig. 2: Correlation of central and peripheral parasite counts in 56 cats with Babesia felis infection.A statistically significant strong positive correlation is evident, indicating that central and peripheral parasitaemias tend to increase together (rs = 0.997, P < 0.05).

Fig. 3 :
Fig. 3: Comparison of ALT values between icteric and non-icteric cats in 56 cats with Babesia felis infection.The ends of the boxes indicate the 25th and 75th percentiles, with a line at the median.Error bars indicate the 10th and 90th percentiles.Circles represent outliers.There is a statistically significant difference between the 2 groups (T = 588.00,P < 0.001).

Table 1 : Laboratory data of 56 cats with Babesia felis infection. Summary data are shown as median and 25th and 75th percentiles.
*Normal values are those used at the Onderstepoort Veterinary Academic Hospital.

Table 2 : Proportions of normal and abnormal laboratory findings in 56 cats with Babesia felis infection. Summary data are shown as the percentage of total.
Jacobson L S, Schoeman T, Lobetti R G 2000 A survey of feline babesiosis in South Africa.Journal of the South African Veterinary Association 71: 222-228 24.Jittapalapong S, Jansawan W 1993 Preliminary survey on blood parasites of cats in Bangkhen District area.Kasetsart Journal, Natural Sciences 27: 330-335 25.Leger N, Ferte H, Berthelot P, Nourry D, Brocvielle P 1992 Un cas de babesiose feline en Haute-Saone France.Sciences Veterinaires Medecine Comparee 94: 249-252 26.Levine N D 1971 Taxonomy of the piroplasms.Transactions of the American Microscopical Society 90: 2-33 27.Loar A S 1993 Feline leukemia virus: immunization and prevention.Veterinary Clinics of North America: Small Animal Practice 23: 193-200 28.Lobetti R G 1995 Leukaemoid response in two dogs with Babesia canis infection.Journal of the South African Veterinary Association 66: 182-184 29.Lobetti R G, Möhr A J, Dippenaar T, Myburgh E 2000 A preliminary study on the serum protein response in canine babesiosis.Journal of the South African Veterinary Association 71: 38-42 30.Lopez-Rebollar L M, Penzhorn B L, De Waal D T, Lewis B D 1999 A possible new piroplasm in lions from the Republic of South Africa.Journal of Wildlife Diseases 35: 82-85 31.Lopez N A, Jacobson R H, Scarlett J M, Center S A, Randolph J F, Scott F W 1989 Sensitivity and specificity of blood test kits for feline leukemia virus antigen.Journal of the American Veterinary Medical Association 195: 747-751 32.Malherbe W D 1950 The management of drugfast biliary fever in dogs.Journal of the South African Veterinary Medical Association 21: 158-159 33.Malherbe W D 1965 Clinico-pathological studies of Babesia canis infection in dogs.I: The influence of the infection on bromsulphalein retention in the blood.Journal of the South African Veterinary Medical Association 36: 25-29 34.Malherbe W D 1965 Clinico-pathological studies of Babesia canis infection in dogs.II: The influence of the infection on plasma transaminase activity.Journal of the South African Veterinary Medical Association 36: 173-176 35.Malherbe W D 1965 Clinico-pathological studies of Babesia canis infection in dogs.IV: The effect on bilirubin metabolism.Journal of the South African Veterinary Medical Association 36: 569-573 36.Malherbe W D 1965 Clinico-pathological studies of Babesia canis infection in dogs.V: The influence of the infection on kidney function.Journal of the South African Veterinary Medical Association 37: 261-264 37. Malik R, Kendall K, Cridland J, Coulston S, Stuart A J, Snow D, Love D N 1997 Prevalences of feline leukaemia virus and feline immunodeficiency virus infections in cats in Sydney.Australian Veterinary Journal 75: 323-327 38.Miller D 1999 The yellow patient.Proceedings of a Symposium on Canine Babesiosis and Ehrlichiosis, Faculty of Veterinary Science, Onderstepoort, University of Pretoria, South Africa, 6 November 1999: 95-97 39.Moik K, Gothe R 1997 Babesien-infektionen der Feliden und Fallbeschreibung bei einer Katze in Deutschland.Tierärztliche Praxis 25: 532-535 40.Mudaliar S V, Achary G R, Alwar V S 1950 On a species of Babesia in an Indian wild cat (Felis catus).Indian Veterinary Journal 26: 391-395 41.Otto C M, Rawlings C A 1995 Tumor necrosis factor production in cats in response to