• Open Access

Possible Emergence of Drug-Resistant Variants of Babesia gibsoni in Clinical Cases Treated with Atovaquone and Azithromycin


  • Results of this study were presented at the ACVIM Forum, San Antonio, TX, 2008.

Corresponding author: Yasuyuki Endo, Laboratory of Veterinary Internal Medicine, Faculty of Agriculture, Kagoshima University, 1-21-24 Korimoto, Kagoshima 890-0065, Japan; e-mail: yendo@agri.kagoshima-u.ac.jp.


Background: There is no well-established treatment strategy for Babesia gibsoni infection. A new therapeutic protocol using atovaquone (ATV) and azithromycin (AZM) has been proposed, but there is concern about the possible induction of relapse and the emergence of ATV-resistant variants after treatment.

Objective: To evaluate the clinical use of combination therapy with ATV and AZM as a first-line treatment of clinical B. gibsoni infection in dogs, and to investigate the emergence of ATV-resistant variants.

Animals: Eight B. gibsoni naturally infected dogs showing signs of acute onset of disease.

Methods: Retrospective case study. Eight clinical cases received combination therapy with ATV and AZM at Kagoshima University Veterinary Teaching Hospital during 2007–2008, and their clinical courses and clinicopathological parameters were evaluated. In addition, alterations in the cytochrome b (CYTb) gene of B. gibsoni were analyzed by polymerase chain reaction and DNA sequencing techniques.

Results: All of the dogs responded well to the treatment, with rapid improvement in their clinical condition and hematological parameters. However, 5 of the 8 dogs relapsed after treatment. Analysis of the CYTb gene strongly suggested the emergence of ATV-resistant variants after treatment.

Conclusions and Clinical Importance: The combination of ATV and AZM can be used as a first-line treatment for dogs with babesiosis, but relapses occur. Attention should be paid to the possible in vivo selection of drug-resistant variants.






complete blood count




cytochrome b






polymerase chain reaction



Diminazene aceturate has been used widely as a first-line agent for the treatment of Babesia gibsoni infection of dogs in Japan. Although diminazene aceturate has anti-Babesia activity, it often fails to eliminate B. gibsoni from affected dogs and a relapse may occur. Furthermore, diminazene aceturate has a narrow clinical safety margin and can induce severe adverse effects such as cerebellar hemorrhage, hepatotoxicosis, and necrosis at the injection site.1–4 Combination therapy of clindamycin (CLDM), metronidazole (MNZ), and doxycycline (DOXY) as an efficacious alternative treatment strategy for B. gibsoni infection.5 We successfully treated 3 of 4 B. gibsoni infected dogs with no adverse effects, and these 3 dogs remained relapse-free after more than 3 years. However, this treatment takes a relatively long time to show its therapeutic effect.5

Recently, a new treatment strategy has been proposed that uses a combination of 2 drugs, an analog of ubiquinone, atovaquone (ATV); and a macrolide antibiotic, azithromycin (AZM).6–11 The use of ATV alone inhibits the growth of Babesia spp., and ATV is used for the treatment of protozoal diseases caused by Plasmodium and Pneumocystis infections.7–10 ATV is presumed to act through the inhibition of the cytochrome bc complex in protozoa.6,12–17 AZM also demonstrates an anti-Babesia effect, but its mechanism of action against Babesia spp. is unknown.17,18 In addition, the simultaneous use of ATV and AZM produces an additive or synergistic therapeutic effect, while the single use of each drug tends to result in a relapse of signs.8,10,19

Combination therapy with ATV and AZM is very effective in eliminating B. gibsoni infection, but results are not consistent based on molecular analysis.20,21 Furthermore, the emergence of drug-resistant variants can occur after the use of ATV alone, and these variants might be caused by mutations in the cytochrome b (CYTb) gene, resulting in amino acid substitutions at the putative ATV binding site.22

We used combination therapy with ATV and AZM in this study as a first-line therapy in B. gibsoni infected dogs. Our results indicate that this drug combination relieved the acute crisis of babesiosis rapidly, but was associated with a high relapse rate. We provide an overview of the clinical cases treated with ATV and AZM, and present the results of molecular analyses of possible ATV-resistant variants, especially those from relapsed cases.

Materials and Methods


Eight dogs with acute onset of babesiosis, exhibiting anemia, thrombocytopenia, or both, and with microscopic or molecular evidence of parasitemia, were enrolled in this study. The 8 dogs represent all the cases of babesiosis at Kagoshima University Veterinary Teaching Hospital (KUVTH) during 2007–2008. The dogs included were 2 Golden Retrievers, 2 Toy Poodles, 1 Miniature Dachshund, 1 Miniature Schnauzer, 1 Labrador Retriever, and 1 mixed breed dog. There were 5 males (1 neutered) and 3 females. The dogs' age ranged from 1.5 to 14 years old.

The 8 dogs were initially examined for diagnosis and treatment, except for 1 dog (dog 5), which was treated with diminazene aceturatea at the referral hospital 5 days before its admission to KUVTH. The other dogs had not received any antiprotozoal drugs before their admission to KUVTH. All dogs had anorexia, depression, pale mucus membranes, and hepatomegaly, splenomegaly, or both at the time of their first visit. In addition, all dogs had mild to severe regenerative anemia and severe thrombocytopenia on CBC. Blood smear specimens stained with modified Wright-Giemsa staining were prepared for all dogs at the first visit. B. gibsoni parasitemia was detected in 7 dogs by examination of blood smears. The remaining dog (dog 3) could not be diagnosed on the basis of examination of a blood smear, because of a low parasitemia. In this dog, 2 types of polymerase chain reaction (PCR) techniques, which amplify the B. gibsoni-derived p18 gene and detect the 18S ribosomal RNA gene derived from Babesia spp., were used to make a diagnosis.23,24 After diagnosis, the 8 dogs received combination therapy with ATV and AZM, and the changes in clinical parameters were evaluated.

Treatment Protocol

Combination therapy with ATV and AZM was initiated in all 8 dogs from the day of diagnosis. The dosages and duration of the combination therapy were based on previous report20 (ATV,b 13.3 mg/kg PO q8h; AZM,c 10 mg/kg PO q24h). ATV was administered for 10 days to all dogs, but some received AZM for more than 10 days, if PCV did not increase to 35% during this period.

Relapse of babesiosis was defined in this study as the reappearance of B. gibsoni in blood smear specimens or the progression of anemia, thrombocytopenia, or both after ATV and AZM treatment. At the time of relapse, combination therapy with ATV and AZM, at the same dosages mentioned above, was reinitiated, but the treatment period was extended, depending on the clinical signs and laboratory test results. In case of a 2nd relapse during or after the 2nd therapy with ATV and AZM, the treatment protocol was changed to a 3-drug combination therapy with CLDMd (25 mg/kg PO q12h), MNZe (15 mg/kg PO q12h) and DOXYf (5 mg/kg PO q12h). The owners were instructed to continue this 3-drug combination therapy for at least 3 months.5

Evaluation of Clinical Parameters

After diagnosis and initiation of therapy, each animal's clinical signs, hematological parameters, and B. gibsoni infection status were monitored. Blood samples were collected at initial examination. Half of each blood sample was anticoagulated with EDTA and used for CBC, determination of the percent parasitemia from blood smear, and PCR analysis. The other half of the samples were treated with heparin and the plasma was used for biochemical analysis. CBC was performed with an automatic blood calculator.g Blood smear specimens were stained with modified Wright-Giemsa staining. The parasitemia of B. gibsoni was calculated in blood smears as the percentage of parasite-infected erythrocytes/1,000 erythrocytes. Detection of B. gibsoni-derived genomic DNA by PCR was performed as mentioned above.23,24 Biochemical parameters were analyzed with a FUJI DRICHEM 3500V.h

Molecular Analysis of B. gibsoni CYTb Gene

Nucleotide and deduced amino acid sequences of the CYTb gene were analyzed with the blood samples collected before and after treatment or at relapse. In order to amplify and determine the nucleotide sequence efficiently, we divided the CYTb gene to 3 parts and constructed oligonucleotide primers for nested-PCR analysis based on the reported CYTb gene sequence (DDBJ/GenBank/EMBL accession number, AB215096).22 In the 1st-round PCR, the primers 5′-TTT AGT GAA GGA ACT TGA CAG GTA-3′ (cytb-F, nt −119 to −96) and 5′-ATA TGC AAA CTT CCC GGC TAA AC-3′ (cytb-R, nt 1,105–1,086) were used. The 2nd round PCR was carried out with 3 sets of primers: 5′-GGA AAC AGG GCT TTA ACC AA-3′ (cytb-1F, nt − 57 to −35) and 5′-CCG GAA TCC AAT AAA ACA GG-3′ (cytb-1R, nt 412–393), 5′-CCT TGG TCA TGG TAT TCT GGA-3′ (cytb-2F, nt 277–297) and 5′-AAC ATC TCC CTG AAA CAA TGG TA-3′ (cytb-2R, nt 702–680), and 5′-ATT TGC TGC TTT GGG TGT TC-3′ (cytb-3F, nt 642–661) and 5′-AAA CTT CCC GGC TAA ACT CC-3′ (cytb-3R, nt 1,099–1,080). The 1st-round PCR product was used as template DNA for all 2nd round reactions. PCR amplification was performed under the following conditions for both 1st and 2nd reactions: 1 cycle of predenaturation (5 minutes, 95°C); 35 cycles of denaturation (30 seconds at 95°C); annealing (1 minute at 54°C), and polymerization (1 minute at 72°C); and 1 cycle of complete elongation (10 minutes at 72°C). The nucleotide sequence of the amplified DNA fragments was determined by direct sequencing using the dideoxy chain termination method.i GENETYX Version 8.0 softwarej was used to characterize the obtained nucleotide sequence data.


Clinical Courses of Cases

The clinical courses and laboratory parameters of all 8 dogs are summarized in Figure 1. After diagnosis, all dogs received combination therapy with ATV and AZM, and had rapid recovery of clinical signs and laboratory parameters. Dogs 1 and 2 showed ideal responses to the treatment and experienced no relapses during the observation period. Multiple mast cell tumors were found in the skin and spleen in dog 1 at day 333. These were treated by surgical resection of the skin lesions and by splenectomy, followed by chemotherapy with vinblastinek and prednisolone.l Despite of immunosuppression caused by splenectomy and chemotherapy, the animal had no signs of relapse of babesiosis. Dogs 1 and 2 remained alive, with no clinical signs, for the duration of the study.

Figure 1.

 Clinical course and change in hematological parameters in Babesia gibsoni naturally infected dogs treated with atovaquone and azithromycin. ATV, atovaquone; AZM, azithromycin; CLDM, clindamycin; DOXY, doxycycline; MNZ, metronidazole; VBL, vinblastine.

Five dogs (dogs 3, 4, 6, 7, and 8) relapsed after cessation of the 1st 10-day ATV/AZM treatment. These dogs had continuous positive results on PCR analysis, even after the 10-day ATV/AZM treatment. These dogs received a 2nd ATV/AZM treatment, and the duration of administration was extended to 20, 40, and 24 days in dogs 3, 4, and 7, respectively. Three (dogs 3, 6, and 8) of the 5 relapsed dogs responded well to the 2nd ATV/AZM treatment, with improved clinical signs and laboratory test results. The treatment protocol in dog 8 was changed to 3-drug combination therapy with CLDM, MNZ, and DOXY on day 72, because the owner said that the dog would not accept a liquid drug.

The remaining 2 dogs (dogs 4 and 7) did not respond to the 2nd ATV/AZM treatment and had a 2nd relapse. Notably, dog 4 had a progression of anemia and parasitemia, even during the ATV and AZM administration. ATV/AZM treatment was replaced by 3-drug combination therapy in these dogs, with good responses and no signs of subsequent relapse. The 3-drug combination therapy was applied to dogs 4 and 7 for 106 and 42 days, respectively. We recommended continuing this therapy to the owner of dog 7; however, the owner decided to stop treatment halfway through the recommend treatment period.

In total, 5 of the 8 dogs experienced either 1 or 2 relapses, and 2 of the 5 dogs demonstrated resistance to the ATV and AZM treatment. However, no adverse effects due to ATV and AZM were detected in any animals, including those with relapses. Two dogs (dogs 5 and 8) died during the observation period, but neither the ATV/AZM treatment nor the Babesia infection itself seemed to be the cause of death. Dog 5 had been treated with diminazene aceturate just before admission to KUVTH, and died on day 7 after the sudden development of neurological signs. Furthermore, this dog had a tendency toward recovery based on PCV and platelet counts, although these could only be evaluated at 2 points. Death in this case was possibly caused by an adverse effect of diminazene. Dog 8 had an uncontrollable extramedullary plasma cell tumor, and died on day 170. The total plasma protein level was very high at the time of death, and, therefore, hyperviscosity syndrome was thought to be the likely cause of death.

Characterization of B. gibsoni CYTb Gene

This study found 5 and 2 dogs demonstrating relapse and ATV resistance, respectively. Accordingly, we evaluated the possible emergence of ATV-resistant strains of B. gibsoni by determining the nucleotide sequence of the CYTb gene.20,21,25

Compared with the standard sequence of the B. gibsoni CYTb gene, single nucleotide substitutions were detected at 7 locations (nt 51, nt 322, nt 361, nt 363, nt 492, nt 558, and nt 676) in our cases. Four of these substitutions could give rise to nonsynonymous amino acid substitutions, including A108T, M121I, M121V, and I226V (Table 1). The CYTb gene sequences derived from dog 1, which did not experience a relapse, were identical to the standard sequences throughout the observation period. The CYTb gene from the other dog without a relapse (dog 2) had a substitution at nt 676 resulting in an amino acid substitution at position 226, from isoleucine to valine (I226V). This type of substitution was also observed in other dogs, both with and without relapses. A108T was observed in dog 7, and the CYTb gene associated with this amino acid substitution was detectable even at the first visit. Therefore, I226V and A108T do not appear to be directly related to the drug-resistance phenotype (Table 1). The M121I and M121V substitutions, however, were suspected to be related to the drug-resistance phenotype, because all dogs possessed B. gibsoni with M121I or M121V amino acid substitutions at the time of relapse (Table 1). B. gibsoni with the M121I substitution was dominant in dog 3, even at the time of the primary onset of the disease.

Table 1.   Changes in Babesia gibsoni CYTb gene before and after treatment with ATV and AZM.
DogTime of IsolationSubstitutions
  1. ATV, atovaquone; AZM, azithromycin; nt, nucleotide Rx, treatment.

1Pre Rx (day 1)None
Post Rx (day 22)None
2Pre Rx (day 1)nt 676ATA→GTAI226V
Post Rx (day 7)nt 676ATA→GTAI226V
3Pre Rx (day 1)nt 363ATG→GTGM121I
 nt 676ATA→GTAI226V
Relapse (day 41)nt 363ATG→GTGM121I
 nt 676ATA→GTAI226V
4Pre Rx (day 1)nt 676ATA→GTAI226V
1st relapse (day 54)nt 363ATG→GTGM121I
 nt 676ATA→GTAI226V
2nd relapse (day 74)nt 361ATG→ATTM121V
 nt 676ATA→GTAI226V
5Pre Rx (day 1)None
6Pre Rx (day 1)nt 676ATA→GTAI226V
Relapse (day 64)nt 363ATG→GTGM121I
 nt 676ATA→GTAI226V
7Pre Rx (day 1)nt 322GCT→ACTA108T
 nt 676ATA→GTAI226V
1st relapse (day 47)nt 322GCT→ACTA108T
 nt 363ATG→GTGM121I
 nt 676ATA→GTAI226V
2nd relapse (day 88)nt 322GCT→ACTA108T
 nt 363ATG→GTGM121I
 nt 676ATA→GTAI226V
8Pre Rx (day 1)None
Relapse (day 55)nt 363ATG→GTGM121I


We used combination therapy with ATV and AZM in this study as first-line treatment in 8 dogs that were naturally infected with B. gibsoni, and showed clinical signs of the acute onset of babesiosis. Clinical signs and hematological parameters improved soon after the initiation of the treatment in all cases. No obvious adverse effects related to this treatment protocol were detected in any of our dogs. To date, diminazene aceturate or CLDM-based treatment protocols have been widely used to treat B. gibsoni infection in Japan, but they are associated with adverse effects and are slow to have an effect.1–4 In our previous study using a 3-drug combination therapy, it took approximately 2 weeks to end the acute crisis of babesiosis, and supportive therapies, including blood transfusion, were required.5 However, in the current study, clinical signs in 8 dogs were well controlled by the administration of ATV and AZM, without the need for any special supportive therapies. Therefore, this combination therapy with ATV and AZM seems to be a potent, rapid-acting, and clinically useful strategy for treating canine babesiosis.

Some concerns were raised by this protocol. Although 2 dogs showed rapid clinical improvements after the 10-day treatment and did not experience any relapse during the observation period, 5 dogs did relapse at approximately 1 month after the cessation of treatment. Possible reasons for the relapses include inadequate drug dosages or medication periods. The 10-day treatment protocol was proposed, but a recent report revealed that this 10-day treatment protocol may be inadequate for controlling the parasitemia in the experimentally infected dogs.20,21 Our findings support this latter report. The appropriate dosages and duration of treatment must be determined. We did not evaluate the immunological status of the animals in this study. The future use of convenient and sensitive methods to evaluate immunological function (eg, evaluation of antibody titer by enzyme-linked immunosorbent assays or immunochromatography) or levels of parasitemia (eg, quantification of Babesia-derived DNA by real-time PCR) in the hosts may help to identify suitable dosages and treatment durations.26–29 On the other hand, the results from longitudinal PCR analyses in this study did not always match the clinical course. This might have been due to the sensitivity of PCR itself and the possible detection of dead parasite-derived DNA. Thus, establishing a simple method to distinguish reinfection from relapse in dogs with positive PCR results continuously or intermittently is required.

The results of this study raise some possible doubts concerning the therapeutic efficacy of AZM, as some of the relapsed dogs had been continuously treated with AZM alone, after the withdrawal of ATV administration. Two relapsed dogs (dogs 3 and 4) responded after the resumption of ATV administration. This suggests that the anti-B. gibsoni activity of the ATV/AZM combination therapy was due mainly to the action of ATV. However, previous reports demonstrated an additive effect of ATV and AZM against B. microti and B. divergens infections, while the relapse rate was high in cases treated with either ATV or AZM alone, suggesting that AZM was required for efficient control of the disease.8,10,19 The mechanism of AZM action against Babesia spp. is unknown, and further studies are required to clarify this point.

Two of the 5 cases in this study (dogs 4 and 7) experienced 2 relapses, and thrombocytopenia, progression of anemia, and reappearance of parasitemia were detected in these animals, even after the 2nd treatment or during the period of treatment with ATV and AZM. Previous reports have also suggested that ATV sensitivity was reduced in experimentally infected dogs that relapsed after receiving ATV treatment.21,22,25 Furthermore, drug-resistant variants were suspected to occur in naturally infected, as well as in experimentally infected dogs.21,22 Therefore, we performed molecular analysis of the CYTb gene, which has previously been reported as being potentially responsible for the development of the ATV-resistant phenotype in B. gibsoni.21,22 Their results demonstrated that ATV-resistant strains had 3 patterns of amino acid substitutions in the CYTb gene, 1 of which, M121I, was located at a putative ATV-binding site. Some nucleotide and amino acid substitutions were detected in the CYTb gene in our cases, most of which did not appear to be related to the ATV-resistant phenotype. However, nucleotide substitutions resulting in the substitution of the amino acid residue at 121 were also detected in our study. Our results, in common with previous report, also suggested that these substitutions at residue 121 may be responsible for the development of ATV-resistance, because this substitution was only observed in the animals that experienced relapses.22 However, these B. gibsoni strains have not yet been tested for in vitro ATV resistance.13–17,21,22 Our findings also suggest that ATV-resistance in naturally infected dogs could have been induced by the same mechanism as that acting in experimentally infected cases.21,22,25

Surprisingly, the possibly drug-resistant M121I variant was dominant in 1 of our dogs, even at the onset of the disease. This suggests that ATV-resistant variants may exist in nature, and that these ATV-resistant variants have emerged mainly through in vivo selection, although mutation of the CYTb gene during ATV treatment cannot be ruled out. Epidemiological surveys of B. gibsoni CYTb genotypes are required to determine the distribution of B. gibsoni variants in nature, and to establish appropriate treatment strategies based on CYTb genotype. Our preliminary epidemiological study found that some clinical cases possessed a possible ATV-resistant variant, although its frequency was relatively low (data not shown).


a Ganaseq, Novartis, Tokyo, Japan

b Mepron, Glaxo SmithKline, Middlesex, UK

c Zithromac, Pfizer Japan Inc, Tokyo, Japan

d Dalacin, Dainippon Sumitomo Pharma Co Ltd, Osaka, Japan

e Flagyl, Shionogi & Co Ltd, Osaka, Japan

f Vibramycin, Pfizer Japan Inc

g pocH-100i, Sysmex, Kobe, Japan

h FUJI DRICHEM 3500V, Fujifilm Medical Co Ltd, Tokyo, Japan

i ABI Prism Big Dye Primer Cycle Sequencing Kit, Applied Biosystems, Foster City, CA

j GENETYX Version 8.0 software, Software Development Co Ltd, Tokyo, Japan

k Exal, Nippon Kayaku Co Ltd, Tokyo, Japan

l Predonine, Shionogi & Co Ltd


This work was supported by grants from the Japan Society for the Promotion of Science.

This work was performed at the Laboratory of Veterinary Internal Medicine, Faculty of Agriculture, Kagoshima University.