• Open Access

Efficacy of Atovaquone and Azithromycin or Imidocarb Dipropionate in Cats with Acute Cytauxzoonosis


  • This study was designed jointly at the University of Missouri (UMC) and North Carolina State University (NCSU). An abstract of this work was presented at the ACVIM Forum in Anaheim, CA, June 2010. It was managed and data collated at UMC. Genetic testing and PCR testing were conducted at NCSU.

Corresponding author: L. A. Cohn, University of Missouri, 900 E. Campus Drive, Columbia, MO 65211; e-mail: cohnl@missouri.edu.


Background: Imidocarb or a combination of atovaquone and azithromycin (A&A) has been suggested for treatment of cats with cytauxzoonosis, but neither has been prospectively evaluated for efficacy.

Hypothesis/Objectives: That survival to hospital discharge is improved by treatment with A&A as compared with imidocarb.

Animals: Eighty acutely ill cats with Cytauxzoon felis infection treated at one of 18 veterinary clinics in 5 states.

Methods: An open-label, randomized prospective study compared survival in cats treated with atovaquone (15 mg/kg PO q8h) and azithromycin (10 mg/kg PO q24h) or imidocarb (3.5 mg/kg IM). All received heparin, fluids, and supportive care. Clinical and clinicopathologic data from initial presentation were collated. Parasitemia was quantified (n = 79) and pathogens genotyped (n = 60). Logistic regression was used to determine the impact of treatment group on the primary endpoint, survival to hospital discharge or death. Covariants were analyzed by rank-sum testing.

Results: Of 53 cats treated with A&A, 32 (60%) survived to discharge while only 7 of 27 cats (26%) treated with imidocarb survived (P= .0036; odds ratio 7.2, 95% CI 2.2, 24). Cats with a lower parasitemia were more likely to survive, as were cats with higher white blood cell counts and lower total bilirubin. Unique pathogen genotypes were identified from 15 cats, while genotype isolated from 21 cats had been described previously. There were multiple pathogen genotypes identified in 24 cats.

Conclusions and Clinical Importance: Survival to discharge was more likely in cats treated with A&A as compared with imidocarb, although case fatality rate remained high.


atovaquone and azithromycin


cycle threshold


interquartile range


internal transcribed spacer


North Carolina State University


polymerase chain reaction


red blood cells


University of Missouri


Veterinary Medical Teaching Hospital

Cytauxzoonosis is an acute-onset disease of cats caused by the apicomplexan parasite Cytauxzoon felis, which is transmitted by the bite of a tick vector. Historically considered a fatal infection in domestic cats, in recent years a number of cats that survived the disease have been described.1–7 Cats that survive the disease can overcome a severe illness with or without specific antimicrobial drug therapy, or persistent infection can be discovered in healthy cats with no known preceding illness. The severe acute illness is believed to be caused by the schizogenous phase of parasite development during which parasite-distended mononuclear cells cause occlusion of blood vessels. After several days, merozoites are released from ruptured cells and taken up by red blood cells (RBCs), producing the classic signet ring shaped merozoites frequently used to diagnose the infection on blood smear. While the schizogenous phase of infection only appears to last for days to weeks, merozoite infected RBCs can be detected for long periods in cats that survive acute infection.

To date, no antiprotozoal therapy has been demonstrated to alter the disease course in acutely infected cats. Survival was not enhanced in experimentally infected cats treated with parvaquone or buparvaquone.6 Case reports exist of survival in naturally infected cats treated with diminazene aceturate, an antiprotozoal agent used for the treatment of babesiosis and trypanosomiasis, but the drug is not approved for use in the United States.2 Despite a lack of proven efficacy, the standard therapy for cytauxzoonosis has become imidocarb dipropionate.1,8 A combination of an antimalarial drug atovaquone with the antibacterial drug azithromycin has been successful in the management of other haemoprotozoan infections that were notoriously difficult to treat.9,10 This led the authors to investigate the utility of this drug combination for the treatment of cytauxzoonosis. Some efficacy was demonstrated in a preliminary, uncontrolled trial of naturally infected cats.a Further, the combination was more effective than imidocarb dipropionate alone in reducing the parasitemia of healthy but persistently infected cats.b We hypothesized that treatment of naturally infected, ill cats with atovaquone and azithromycin (A&A) would result in improved survival as compared with treatment of an equivalent group of cats with imidocarb dipropionate. An open label, imbalanced randomized block designed prospective clinical trial was conducted to assess the validity of this hypothesis.

Materials and Methods

Study Design

The goal of this multicenter, open-label parallel-group study with imbalanced block randomization (2 : 1) was to compare efficacy of treatment with A&A or imidocarb for acute cytauxzoonosis in naturally infected cats. Cats were treated in a variety of clinical settings ranging from small, single veterinarian rural clinics lacking 24-hour care to university-associated veterinary medical teaching hospitals (VMTH). The vast majority of case recruitment occurred during the spring and summer of 2009, although 5 cases were enrolled before 2009.

Enrollment Criteria

Cats included in the study were presented for evaluation of an acute illness with signs suggestive of cytauxzoonosis, including fever, lethargy, anorexia, or icterus and cytologic demonstration of suspected piroplasms within RBC or schizonts in either blood mononuclear cells or from aspirates of the lymph nodes, liver, or spleen. Cytologic evaluation was conducted by the veterinarian responsible for enrolling the cat.

Participating veterinary clinics were solicited from areas considered endemic for C. felis based on expressed interest in study participation. All participating veterinarians agreed to abide by provided instructions for evaluation and treatment, including compliance with a predetermined order of randomization. Once a veterinarian or veterinary clinic agreed to participate, they were supplied with study drugs adequate to treat 3 cats, detailed instructions, order of randomization, and owner consent forms.

At presentation, the treating veterinarian was responsible for explaining the study to the cat's owner, including potential benefits and harms and the option to withdraw from the study. Study drug was supplied free of charge and additional monetary compensation was provided to offset some of the costs of veterinary care. Owners were invited to contact the lead investigator with questions and were asked to sign an informed consent document before inclusion of their cat in the study.

Exclusion criteria were applied both at the time of enrollment and retrospectively. Cats presented with body temperature <37.8°C (<100.0°F) or in a moribund condition were excluded immediately. Cats were excluded retrospectively if C. felis DNA could not be detected by polymerase chain reaction (PCR), if medical records were not made available to the investigators, or if there was a failure to adhere to randomization order or treatment instructions.

All procedures and consent forms were reviewed and approved by the animal care and use committees of the University of Missouri (UMC), North Carolina State University (NCSU), and Oklahoma State University.

Randomization and Allocation

An imbalanced blocked randomization of 2 : 1 (A&A and imidocarb, respectively) was used. Three pieces of paper, 2 marked with “A&A” and one with “imidocarb,” were folded and placed in a bowl. For most cats, a technician at UMC withdrew the papers one at a time to determine the order of treatment for the 1st 3 cats enrolled at each participating veterinary clinic; the clinic was supplied with this order at the same time they were supplied with study drugs. When the clinic began treating the 3rd cat enrolled, they contacted the technician who repeated the procedure and supplied additional drug and a new randomization order for the next 3 cats enrolled, repeating the procedure as necessary. For 15 cats treated at 1 VMTH (NCSU), randomization was determined in identical fashion onsite.

Test Treatments

Immediately after enrollment treatment was begun with either a combination of A&A or with imidocarb. Atovaquonec was administered PO or by means of enteral feeding tube at a dose of 15 mg/kg q8h for 10 days. Administration along with a fatty meal was suggested but not required. Azithromycind was administered PO or by means of enteral feeding tube at a dose of 10 mg/kg q24h for 10 days. Cats that received imidocarb were pretreated with atropine (0.05 mg/kg) SC 15 minutes before injection of 3.5 mg/kg imidocarb dipropionatee IM. Injection was repeated in the same fashion 7 days later.

Concomitant Treatments

Besides treatment with study drug, therapeutic management was largely left to the discretion of the treating veterinarian, with 2 exceptions. All cats received unfractionated heparin (200 U/kg) SC q8h for the duration of hospitalization, but the duration of hospitalization (and therefore heparin use) was decided by the veterinarian treating the cat based on their clinical judgment. Provided guidelines suggested that hospital discharge was appropriate when the cat was afebrile and eating voluntarily. Similarly, all cats received intravenous crystalloid fluids initially, but fluid type, volume, and duration were left to the treating veterinarian. Guidelines suggested the use of any of several types of balanced saline solutions at a rate of 50–100 mL/kg/d. As with heparin administration, discontinuation of fluid therapy was left to the discretion of the treating veterinarian. Adjunctive and supportive care including blood transfusion, the use of analgesics, antibiotics, antiemetics, and appetite stimulants was allowed and all such treatments were documented. Placement of esophagostomy tubes was specifically approved to ease administration of medication as well as to provide nutritional support.

Clinical Evaluation

Available clinical information varied. In all cases, historical information was requested in regards to signalment, housing (indoor/outdoor), duration of illness, ectoparasite application, health of other cats within the household, and clinical signs prompting the veterinary visit. Information requested from the initial physical examination included the cat's temperature, pulse and respiratory rate, body condition score, and any other identified abnormality. Anticoagulated whole blood was collected for attempted PCR detection of C. felis. Although assessment of CBC, serum biochemical profile, coagulation status, and retroviral testing were suggested, they were not required for participation. When tests were repeated, only results from the initial assessment were recorded. Additional recorded information included the primary endpoint (survival to discharge or death), the duration of hospitalization, and all forms of therapy administered. Copies of all medical records were sent to one of the investigators (L.A.C.), who collated and verified the data by dual entry.

Molecular and Genetic Testing

The persons conducting all molecular and genetic testing were unaware of assigned treatment. Anticoagulated blood was sent directly to NCSU for PCR detection of C. felis using previously published methods with minor modifications.11 Briefly, each reaction consisted of 12.5 μL 2 × PCR master mix,f 7 μL water, 50 pmoL of each oligonucleotide primer, and 5 μL sample. Thermal cycling conditions were 98°C for 30 seconds followed by 45 cycles 95°C for 5 seconds and 60°C for 5 seconds. Melting curve analysis was initiated at 75°C and data were captured at increasing increments of 0.5°C for 30 time points. Internal transcribed spacer (ITS) region genotyping was performed as described previously with minor modifications.3,7,12 Each 50 μL reaction contained a 1 × concentration of PCR buffer II, 1.25 U of taq polymerase,g 50 pmoL of each primer, 1.5 mM MgCl, 200 μmoL of each dNTP, 33.5 μL of water, and 5 μL of each sample. Thermal cycling conditions were: 95°C for 5 minutes followed by 45 cycles 95°C for 45 seconds, 54°C for 45 seconds, and 72°C for 45 seconds. ITS amplicons were visualized by ethidium bromide staining and ultraviolet light transillumination after electrophoresis in a 1.5% agarose gel. ITS amplicons were purified and sequenced bidirectionally. Each chromatogram was inspected and contigs were assembled by commercially available software package.h A threshold for calling secondary peaks within each chromatogram was set at 50%. For each PCR assay, positive (previously characterized C. felis samples) and negative (no template) controls were used. Standard precautions were used to prevent amplicon carryover.

Statistical Analysis

The primary endpoint with respect to treatment efficacy was survival to hospital discharge, or death (including euthanasia). Logistic regression was applied to the dichotomous outcome by the LOGISTIC procedure for primary analysis. The Hosmer-Lemeshow statistic was used to assess the model fit. Descriptive statistics were also determined. Two categorical variables were tested for independence by Chi-square tests while categorical groups were compared with each other on numeric variables (temperature, total bilirubin, PCV, white blood cell [WBC] count, cycle threshold [CT]) by Wilcoxon's rank-sum tests. Calculations were performed by standard statistical software,i and significance was set at P < .025, with marginal significance at P < .05 but P > .025.



Eighteen veterinary clinics, including 2 university-associated VMTH and 16 private practice clinics, participated in recruitment of 97 cats with suspected cytauxzoonosis from 5 states. A median of 3 cats per clinic were enrolled (range, 1–20). Of the 97 cats, 17 were later excluded due to either a negative PCR result for C. felis (n = 11) or failure to adhere to study protocol (6). Of the 80 cats included in statistical analysis of outcome, 53 received A&A and 27 received imidocarb. Twenty of these cats were treated at a VMTH and 60 received treatment at a private practice. Cats included in the analysis were recruited from Missouri (n = 28), Tennessee (21), North Carolina (15), Arkansas (10), and Oklahoma (6).

Breed, age, and sex were documented. Most cats were domestic short hair or mixed breeds (n = 71), while others were domestic long haired (6), Siamese (2), or Bengal (1). Cats ranged in age from 6 months to 12 years, with a mean of 3.6 ± 2.9 years; for 12 cats, age was recorded simply as “adult.” Forty-three cats were male (10 intact, 33 neutered) and 35 were female (2 intact, 30 neutered, 3 unknown), and the sex of 2 cats was not recorded.

Baseline Clinical Evaluation

Findings on history and physical examination were nonspecific. All infected cats had exposure to the outdoors. The most common historical complaints were lethargy (n = 78), and anorexia (60). Vomiting (usually once or twice) was reported in 6 cats. Other complaints included unsteady gait (3), abnormal behavior (1), abortion (1), and hematuria (1). The most common abnormalities on physical examination included hyperthermia (temperature >39.2°C [102.5°F]; n = 72), icterus (31), elevated nictatans (31), dehydration (22), the presence of ticks (22), tachypnea (respiratory rate >40 breaths per minute; 20), tachycardia (heart rate >200 beats per minute; 13), pallor (9), murmur (8), vocalization (5), discomfort on abdominal palpation (5), lymphadenomegaly (5), and splenomegaly (5). Abortion, stupor, gallop rhythm, muscle wasting, ear mites, abscess, and disorientation were each found in a single cat. Temperature range in the 78 cats with legible recorded temperature was 38.3–41.7°C (101.0–107.0°F). The mean temperature of cats randomized to receive A&A (104.5 ± 1.2°F; 40.3°C) was identical to that of cats randomized to receive imidocarb (104.5 ± 1.1°F; 40.3°C).

Initial laboratory findings, other than piroplasms or schizonts, were also nonspecific and did not differ between treatment groups. Abnormalities on CBC (Table 1) included leukopenia (WBC < 5 × 103/μL; n = 43 of 73) and anemia (PCV < 26%; 40 of 74). While thrombocytopenia was noted on 59 samples, adequate information about platelet clumping was not available to assess the validity of this finding. Common biochemical abnormalities included hyperbilirubinemia (total bilirubin > 0.5 mg/dL; 37 of 50), hyperglycemia (glucose >150 mg/dL; 35 of 55), and hypocalcemia (Ca < 9.0mg/dL; 32 of 47) (Table 1). Total solids or total protein below the lower limits of the reference range (<6.0 g/dL) were detected in 12 of 51 cats. Although frequently assayed, blood urea nitrogen and ALT were rarely outside of reference ranges. Thirty-eight cats underwent retroviral testing; only a single cat was positive for feline leukemia virus.

Table 1.   Initial laboratory findings in cats with cytauxzoonosis treated with either A&A or imidocarb dipropionate.
 All CatsA&AImidocarb
  1. WBC, white blood cell; A&A, atovaquone and azithromycin; ALT, alanine transaminase; BUN, blood urea nitrogen.

  2. Mean ± SD (n).

PCV (%)25.5 ± 5.9 (74)25.95 ± 6.3 (49)24.3 ± 4.5 (25)
WBC × 103/μL5.1 ± 2.8 (73)5.26 ± 2.6 (48)4.53 ± 3.0 (25)
Total solids or protein (g/dL)6.6 ± 1.0 (51)6.7 ± 1.0 (35)6.3 ± 0.7 (16)
Total bilirubin (mg/dL)2.3 ± 2.0 (50)2.15 ± 1.8 (36)2.74 ± 2.4 (14)
Glucose (mg/dL)169 ± 43 (55)169 ± 42 (38)169 ± 45 (17)
Calcium (mg/dL)8.4 ± 0.9 (47)8.4 ± 0.9 (32)8.1 ± 0.8 (15)
BUN (mg/dL)29.0 ± 21.3 (56)27.5 ± 18.4 (38)32.2 ± 26.5 (18)
ALT (U/L)68.3 ± 99.4 (55)58.9 ± 65.2 (37)102 ± 143 (18)

Because of limitations of sample size and number of cats tested for specific variables, only temperature, PCV, WBC count, and total bilirubin were statistically evaluated as potential covariants influencing outcome. Neither temperature nor PCV had a significant relationship to survival or death in hospital. Rank-sum tests comparing treatment relative to WBC (P= .0014) or total bilirubin (P= .031) did indicate significance or marginal significance. Cats that survived were likely to have a higher WBC count (Fig 1) and lower total bilirubin (Fig 2) than cats that died.

Figure 1.

 Box-plot of WBC counts in cats that survived to hospital discharge or died in hospital. Each box represents the interquartile range (IQR) from the 25th to 75th percentile. The line inside the box represents the median, “+” represents the mean, and “*” represent statistical outliers (values more than 1.5 IQR beyond the 1st and 3rd quartiles). Whiskers extend to the maximum and minimum values within 1.5 IQR above or below the 75th and 25th percentile. WBC count was significantly higher in cats, which survived to discharge (P= .0014).

Figure 2.

 Box-plot of total bilirubin in cats that survived to hospital discharge or died in hospital. Each box represents 25th–75th percentile of values, while the line represents the median and the “+” represents the mean. Range is indicated by the whiskers. Total bilirubin was marginally lower in cats that survived to discharge (P= .031).

Concomitant and Adjunctive Therapies

A variety of supportive and adjunctive therapies were administered. Thirty cats received only the drugs required for study participation. Blood transfusions were administered to 9 cats. Antimicrobials administered in addition to the study drugs included cefazolin (n = 11), enrofloxacin (9), ampicillin-sulbactam (7), doxycycline (5), amoxicillin-clavulanate (3), amoxicillin (2), ampicillin (2), and penicillin (2), and 1 cat each received marbofloxacin, clindamycin, and gentamicin. Several cats received analgesic, antiinflammatory, or antipyretic drugs. These included buprenorphine (14), meloxicam (10), ketoprofen (5), fentanyl (1), dexamethasone (3), prednisone/prednisolone (3), and metamizole (1). Some cats additionally received antiemetic medications, gastric protectants, or appetite stimulants. These included odansetron (5), dolasetron (3), maropitant (3), metoclopramide (3), famotidine (3), ranitidine (1), sucralfate (2), mirtazpine (2), cyproheptadine (2), or intravenous diazepam (1). Other concomitant drug therapies included B vitamins (3), and lactulose (3), and to 1 cat each glucosamine, SAMe, drontal, fenbendazole, nitenpyram, flumazenil, furosemide, glucosamine, enalapril, and Delta albeplex.j It was rare for cats treated at private practices to receive more than a single adjunctive therapy (usually an analgesic or antibiotic), while all cats treated at VMTH received from 2 to 14 additional therapies during hospitalization. Esophagostomy feeding tubes were placed in 8 cats treated at a VMTH.

Effect of Therapy on Outcome

Of 80 cats included in data analysis, 39 survived and 41 died. Of the 41 that died, 5 were euthanized because of severe clinical deterioration and moribund condition. Twenty-four of 41 cats that died did so the day of or the day after presentation for care; only 3 cats died or were euthanized more than 3 days after presentation. Of the 53 cats treated with A&A, 32 (60%) survived and 21 (40%) died before hospital discharge. Of the 27 cats treated with imidocarb, 7 survived (26%) and 20 died (74%). Survival was greater in cats treated with A&A than in cats treated with imidocarb (P= .0036; odds ratio 7.2; 95% CI 2.2, 24.0).


CT determined from real-time PCR was used as an estimate of parasitemia in 79 cats. There was no difference between CT from cats randomized to treatment with A&A as compared with treatment with imidocarb (Table 2; P= .79). Parasitemia was lower in survivors as compared with nonsurvivors for cats receiving either treatment (Table 2; P= .0012). The odds ratios for this continuous variable suggest that for each unit of increase in CT, odds for survival were 1.3 times higher (95% CI 1.1, 1.5). Sample size limitations prevented statistical analysis of treatment and CT interaction, but mean CT was lower (therefore parasitemia higher) in survivors treated with A&A than in survivors treated with imidocarb.

Table 2.   Cycle threshold from real-time PCR in 79 cats with cytauxzoonosis treated with either A&A or imidocarb dipropionate.
 A&AImidocarbAll Treatments
  • A&A, atovaquone and azithromycin.

  • Mean ± SD (range).

  • *

    P= .0012.

Survivors26.5 ± 4.331.3 ± 2.727.4 ± 4.4*
Nonsurvivors24.6 ± 4.124.4 ± 2.724.5 ± 3.4*
All outcomes25.8 ± 4.2426.2 ± 4.025.9 ± 4.2

ITS Genotype of Pathogen

ITS genotyping of the pathogen was attempted in 64 of 80 cats sampled. A complete genotype (ie both ITS1 and ITS2) could be unambiguously resolved for 36 of 64 cats. For 15 of 36, unique ITS genotypes not previously described were detected while 21 of 36 corresponded with previously described ITS genotypes.3,7,12 Of these, 14 corresponded with ITSA/ITSc, (Genbank accession EU450802/EU450804), 5 corresponded to ITSj, (Genebank accession FJ536426/FJ536418) and 2 corresponded with ITSC (Genbank accession EU450803/EU450804). Of the cats infected with pathogen of the ITSA/ITSc genotype, 8 received A&A (7 survived) and 6 received imidocarb (2 survived). Of the cats infected with the ITSj genotype 3 received A&A (all survived) and 2 received imidocarb (1 survived). Both cats infected with ITSC genotype received A&A; 1 survived and 1 died.

For 24 of 67 cases, the sequence chromatograms were consistent with the presence of 2 or more genotypes with insertions or deletions in 1 or both ITS regions. For 4 of 67 cases resolution of the ITS regions was not possible because of poor sequence quality. Attempts to further classify the ITS genotypes (ie, cloning and sequencing of multiple clones) from these samples were not pursued.


This study demonstrates that cats with acute cytauxzoonosis treated with A&A had improved survival to hospital discharge (60%) as compared with a similar population treated with imidocarb dipropionate (26%).

The relatively small study size precluded statistical evaluation of all potential clinical or laboratory predictors of treatment outcome. Subjective evaluation of demographic data, historical, and physical findings did not suggest important or obvious differences between cats that survived infection and those that died. Most cats, regardless of outcome, were young adults with exposure to the outdoors and vague clinical signs of lethargy and anorexia. Similarly, the mean initial temperature of survivors was identical to that of nonsurvivors, and cats with either outcome demonstrated dehydration or icterus. While initial PCV was unrelated to outcome, cats with higher initial WBC count and lower total bilirubin were statistically more likely to survive. Unfortunately, important overlap and modest biological differences in values between survivors and nonsurvivors invalidate these parameters as outcome predictors.

Parasitemia, as determined by real-time PCR, did not differ in cats treated with A&A or imidocarb. However, cats administered either treatment protocol were more likely to survive in the face of a lower parasitemia. The effect of parasitemia was partially mitigated by treatment with A&A, suggesting that cats with a higher parasitemia may be more likely to survive if treated with A&A. The contributions of the life stages (merozoites or schizonts) to the overall parasitemia cannot be determined by a PCR assay detecting genomic DNA. Each infected RBC typically contains a single merozoite with a single genome copy, while each infected macrophage contains a multinucleated schizont that may contain thousands of genome copies. Therefore, clinicians should not attempt to determine a prognosis based on the number of merozoite infected RBCs.

Reports have suggested differences in survival based on geographic region, prompting speculation that genetic differences in pathogen virulence occur in geographic proximity.1,3,12 While our study was not designed to identify geographic associations with outcome, similar proportions of cats lived or died in each state (data not shown). Genotypic variation in pathogens was investigated by analysis of noncoding regions of the rRNA operons known as ITS regions 1 and 2. Prior investigation of pathogens in domestic cats and bobcats has identified distinct pathogen genotypes. In our study, 24 pathogens could not be resolved to a single genotype because of the presence of insertions or deletions in 1 or both ITS regions. These results might reflect polymorphisms in multiple copies of the rRNA operons or coinfection with multiple genotypes.7 Although initial studies seemed to suggest that certain genotypes (ITSa/ITSc) were more likely to be associated with a less pathogenic variant of C. felis than other genotypes (ITSC), it now seems unlikely that ITS genotype can predict pathogenicity.3,7,12

Cats in this study received treatment in any of several veterinary clinics using a number of different adjunctive therapies. As predicted before the study, adjunctive therapies tended to be similar between all cats treated at a given clinic. For instance, clinics tended to give all cats or no cats adjunctive antibiotics, and when antibiotics were used the same type of antibiotic was usually used in all cats from a single clinic. Similarly, clinics tended to use the same type of analgesic (or no analgesic) in treated cats. Randomization of cats at each clinic to a given treatment protocol by block was designed to minimize impact of treatment variation by ensuring that no single clinic would treat large numbers of cats with only 1 of the 2 study protocols. The variety of treatments applied in a variety of clinic types resulted in a study that more closely mimicked a realistic application of treatment than would a study conducted entirely within a single practice. However, nonuniformity of both data collection and treatments can also be viewed as a limitation of the study in that effects of concomitant treatments may have altered outcome.

The open label nature of the study allowed the introduction of bias. Masking was considered impractical because 1 protocol involved daily oral medication whereas the other involved injections. Additionally, pretreatment of cats with atropine would be easily distinguished from pretreatment with placebo based on observable reactions to the medication. Outcome bias was minimized by using the simple end point of survival or death during hospitalization. Investigators who conducted genetic analysis of pathogens from sampled cats were unaware of treatment arm, thereby minimizing bias regarding genotypic results or pathogen burden in any given cat.

Because no placebo treatment was included in the trial, it is impossible to speculate whether treatment with imidocarb dipropionate offered any advantage over supportive care alone. All of the veterinarians participating in our trial had routinely treated infected cats with imidocarb in years past. Most of these veterinarians espoused a belief that very few cats treated with imidocarb survived the acute illness, a belief that prompted their participation in the trial. While both our study and multiple case reports document survival of cats treated with imidocarb, anecdotal mention is made in a retrospective report of 18 surviving infected cats (only 1 of which was treated with imidocarb) that imidocarb lacked efficacy in the treatment of experimentally induced cytauxzoonosis.1,2,5

This randomized prospective treatment trial demonstrated improved survival in cats with acute cytauxzoonosis treated with A&A as compared with treatment with imidocarb dipropionate. Nonetheless, mortality remained high with ∼40% of A&A treated cats dying in hospital. The course of illness with cytauxzoonosis is swift, and most cats that succumbed did so shortly after presentation for care. It may well be that even effective antiprotozoal therapy cannot reverse the course of clinical illness. Efforts must be focused on prevention of disease, either through minimization of exposure to tick vectors or through development of chemoprophylaxis or vaccination.


a Birkenheuer AJ, Cohn LA, Levy MG, et al. Atovaquone and azithromycin for the treatment of Cytauxzoon felis. J Vet Intern Med 2008;22:703 (abstract)

b Cohn LA, Birkenheuer AJ, Ratcliff E. Comparison of two drug protocols for clearance of Cytauxzoon felis infections. J Vet Intern Med 2008;22:704 (abstract)

c Mepron, GlaxoSmithKline, Research Triangle Park, NC

d Azithromycin suspension, Greenstone LLC, Peapack, NJ

e Imizol, Intervet Schering-Plough Animal Health, Summit, NJ

f SsoFast EvaGreen Supermix Bio-Rad, Hercules, CA

g Amplitaq gold, Applied Biosystems by Life Technologies, Carlsbad, CA

h ContigExpress, Vector NTI Advance 11.0, Invitrogen by Life Technologies, Carlsbad, CA

i SAS v9; SAS Institute Inc, Cary, NC

j Delta albeplex, Pfizer Animal Health, New York, NY


Cats were enrolled and received treatment at 18 animal clinics in Missouri (Fair Grove Veterinary Services; All Creatures Veterinary Hospital; Animal Clinic of West Plains; Best Friends Animal Hospital; Cross Creek Animal Hospital; Emergency Veterinary Clinic of Southwest Missouri; Gainesville Veterinary Clinic; Mid Ozark Animal Health Center), Arkansas (Russellville Animal Clinic; Eureka Springs Animal Hospital; Southwind Animal Hospital; Sugar Creek Animal Hospital), Oklahoma (Animal Care Central; Oklahoma State University-VMTH), Tennessee (Lebanon Animal Hospital; DeKalb Animal Clinic; VCA Animal Care Center), and North Carolina (NCSU).

The work was supported by a charitable entity, which prefers to remain anonymous.

The authors acknowledge all of the veterinarians participating in the care of these cats, as well as the expert technical assistance of Ms Victoria Catto and Mrs Ginny Dodam.