• Open Access

Phase I Dose Escalation of Doxorubicin Chemotherapy in Tumor-Bearing Equidae


Corresponding author: Dr Alain P. Théon, Department of Surgery and Radiology, University of California, 1 Shields Avenue, Davis, CA 95616; e-mail: aptheon@ucdavis.edu.



There is no information on the use of doxorubicin in horses with tumors.


To determine dose-limiting toxicosis (DLT) and maximum tolerated dose (MTD) of doxorubicin in tumor-bearing horses.


Seventeen horses with 34 localized or multicentric advanced tumors.


Two-stage dose-ranging design involving intrapatient and interpatient dose escalation. Treatment protocol included 6 treatment cycles given at 3-week intervals with dosages ranging from 40 to 85 mg/m2. Clinical signs, hematologic, and nonhematologic changes were evaluated.


Total doses ranged from 1,127 to 2,900 mg in 12 horses that completed the assigned treatment protocols. The MTD was 75 mg/m2. Hypersensitivity reactions and neutropenia were dose limiting. Hypersensitivity was dose-dependent but schedule invariant. Neutropenia was dose- and cycle-dependent but dose-escalation schedule invariant. Cardiotoxicity was not observed.

Conclusion and Clinical Relevance

The recommended dosage of doxorubicin to treat horses is 70 mg/m2 given at 3-week intervals as single agent. Adjunctive treatment with antihistamines and nonsteroidal anti-inflammatory drugs is recommended to control hypersensitivity.


aortic dimension


dose-limiting toxicosis


fractional shortening


interventricular septal wall thickness at end-diastole


interventricular septal wall thickness at end-systole


left ventricular internal dimension at end-diastole


left ventricular internal dimension at end-systole


maximum tolerated dose

There is little information on the use of systemic chemotherapy in horses.[1] Because of the availability of generic forms of several major antineoplastic chemotherapy drugs in recent years, treatment of horses with cancer has become affordable. Anecdotal dosing regimens derived from dog and cat protocols have been described for horses.[2] However, no studies reporting safe and effective drug dosage have been published.

Doxorubicin1 is an anthracycline cytotoxic drug. Its principal mechanism of cytotoxicity is mediated by DNA intercalation and topoisomerase II inhibition, resulting in single- and double-stranded DNA breaks, and deregulation of calcium and sodium transport at the cell membrane level. In addition, doxorubicin redox cycling promotes the generation of oxygen free-radicals that can oxidize nuclear and mitochondrial DNA bases and contribute to cytotoxicity.[3] The clinical application of this drug, however, is limited by its dose-related adverse effects, which include bone marrow toxicity, gastrointestinal and renal toxicity, as well as acute and cumulative cardiotoxicity.[4, 5]

The use of cytotoxic chemotherapy at doses close to the MTD is an established tenet of conventional oncologic practice, and the determination of MTD remains the primary objective of early phase clinical evaluation. In people, doxorubicin has a narrow therapeutic range and displays a steep dose–response curve gradient at doses below the MTD.[6] Determination of the MTD of doxorubicin in horses therefore is a critical step in the development of safe and effective chemotherapy protocols.

The hypothesis of this study was that doxorubicin can be used safely in horses. The objectives of the study were to determine the DLT and MTD of doxorubicin and establish a recommended dose (RD) for use in horses.

Materials and Methods


Horses with advanced or multicentric tumors referred to the Equine Medical Oncology Clinic at the Veterinary Medical Teaching Hospital, UC Davis were enrolled in this study. Horses eligible for this study had histologically confirmed tumors, measurable disease, no evidence of organ failure, and no prior history of local or systemic treatment with corticosteroids or other antineoplastic agents. Inclusion criteria were a life expectancy of ≥3 months, normal renal and liver function tests and hematologic variables, and an echocardiogram (ECHO) that showed a left ventricular ejection fraction within normal limits. Because of the lack of established equine cardiac exclusion criteria for doxorubicin administration, horses were deemed ineligible for doxorubicin if they had systolic impairment with a fractional shortening (FS) <20%, underlying cardiomyopathy or severe valvular insufficiency. Patients with malnutrition or active infection were not eligible. Staging of cutaneous tumors was performed by use of a modified tumor/node/metastasis classification system[7] in which the T category indicated maximum diameter of the tumor, N category indicated the extent of regional nodal involvement, and the M indicated the presence of distant metastases. Staging of horses with lymphoma was performed by use of the World Health Organization system for clinical staging for lymphoma in domestic animals.[8] Because we do not perform bone marrow aspiration routinely in patients with lymphoma, bone marrow examination was required only in horses with unexplained lymphocytosis, neutropenia, and thrombocytopenia and presence of circulating lymphoid cells by microscopic observation of peripheral blood smear. In these horses, cytologic confirmation of bone marrow involvement resulted in exclusion from the study.

Study Design

A 2-stage dose-ranging design involving intrapatient dose escalation and intercohort dose escalation was used to determine acute and cumulative dose-limiting toxicities and establish a safe treatment dosage.[9] An accelerated exploratory titration design was used in the 1st stage in a cohort of 3 horses to quickly determine an entry dosage range of doxorubicin for the second stage. A rapid intrapatient drug dose escalation was conducted with a starting dosage for each horse of 40 and 20 mg/m2 dose increments. The dosing schedule was based on established safe dosages when used as a single agent for dogs[10] and people.[11] Each dose level was used twice in each horse before the dose was increased to the next level.

The second phase was conducted in an additional cohort of horses to determine MTD and estimate RD of doxorubicin in horses. Based on the results of the 1st stage of the study, the dosage to be evaluated ranged from 60 to 85 mg/m2 with increments of 5 mg/m2. Horses were entered in groups of 4, each receiving the same starting dose. In each treatment group, horses received 2 cycles at the same dosage level before the dose was increased to the next dosage level (5 mg/m2 dose increment). The starting dosage level per group (dose-escalation schedule) was 60 mg/m2 in Group 1 (60–65–70 mg/m2), 65 mg/m2 in Group 2 (65–70–75 mg/m2), 70 mg/m2 in Group 3 (70–75–80 mg/m2), and 75 mg/m2 in Group 4 (75–80–85 mg/m2). The protocol was approved for an additional group with a starting dosage of 80 mg/m2 (80–85–90 mg/m2) if the MTD had not been reached. This protocol ensured that a minimum of 3 cycles at each dosage level (60–85 mg/m2) and cycle number (cycle 1–6) would be administered to assess acute and cumulative toxicity.

Protocol approval was obtained from the UCD Institutional Animal Care and Use Committee. Signed informed consents were obtained from owners.

Treatment Procedures

The calculated dose was based on the scheduled dosage (mg/m2) and the conversion formula for kg to body surface area (BSA) in horses: BSA(m2) = 10.5 × W(kg)2/3 × 10−2.[12] Handling and administration of doxorubicin1, as well as disposal of related waste, were done according to guidelines on chemotherapy use in veterinary oncology.[13] Doxorubicin was drawn into 50 cc luer-lock syringes. A jugular IV line was placed under aseptic conditions (14 ga 5¼ in IV catheter,2 76 in solution set,3 33 in extension set,4 and 6 in T-port extension set5). The drug was administered by slow IV push through the injection port (Y-port) of the solution set at a slow, steady rate over a period of approximately 1 hour. As doxorubicin was instilled, a primary IV normal saline solution (0.9% sodium chloride USP) was administered by IV drip at a rate of approximately 400–450 drops/min (2,000–2,500 mL/h) to ensure continued dilution (approximately 0.2 mg/mL at a rate 5 mg/min) of the drug.

Treatment protocol included 6 treatment cycles at 3-week intervals. Prophylactic premedication with diphenhydramine hydrochloride at a dose of 1 mg/kg was administered IM, 30 min before infusion.

Based on published data in dogs[14, 15] on residues in urine and blood and period of risk for potential contamination after injection in a dog or cat,[13] it was recommended that daily excreta and soiled bedding be handled as biohazard and kept on site 7 days before disposal. During that period, owners were instructed to wear gloves, to not handle excreta directly, and to prevent contact with children and pregnant women. In addition, blood or urine samples were not allowed to be submitted to a clinical diagnostic laboratory.

Treatment Toxicosis

Toxicosis was graded using a modified Veterinary Clinical Oncology Group Toxicity grading.[16] Dose-limiting toxicity for any endpoint measured was defined as (1) grade 3 toxicity or (2) treatment delay >2 weeks because of unresolved toxicity. Dose modifications for toxicity were defined for hematologic and nonhematologic toxicities. If a horse had a DLT during a treatment, the dose was decreased by 1 dosage level for the after treatment. If the worst toxicity in the 21 days after a treatment was moderate, defined as grade 2, the dose remained the same for the next cycle. If the worst toxicity was less than moderate (grade 0 or 1) during a cycle, the dose was increased to the next dosage level for the next treatment cycle. If the predominant DLT was of short duration and medically manageable, further dose escalation was allowed.

Treatment was delayed if neutrophil count was <1.5 × 109/L, or the platelet count <100 × 109/L or nonhematologic toxicities had not resolved on the treatment day. If treatment delay was <1 week, no dose modification was implemented and the horse was treated at the scheduled dosage level.

If the dosage resulted in grade 4 toxicity, additional enrollment at that dose was stopped. Criteria for patient withdrawal from the study included declining quality of life because of disease progression, unresolved severe (grade 3) or life-threatening toxicity (grade 4), and noncompliance with the protocol.

MTD was defined as the dose after which one-third of cycles resulted in DLT. The RD was the highest dosage level below the MTD that resulted in no more than 1 DLT out of 6 treatment cycles, including 2 consecutive cycles at that dosage in 3 treatment groups.

Hematologic and nonhematologic toxicoses were assessed during and after treatment, according to the treatment schedule. On the day of treatment, each horse had a thorough history and physical examination to determine any clinically adverse effects of the drug treatment. Horses were hospitalized overnight after treatment to monitor for acute toxicosis. At each visit, owners were asked to provide information regarding any adverse effects observed at home.

Hematologic toxicity was evaluated by complete (absolute and differential) blood cell count (CBC) before and between each treatment. Between treatments, CBC were done either at our institution or a blood sample was collected by the referring veterinarian and sent to a commercial clinical diagnostic laboratory. For 15 intertreatment evaluations, horses had a CBC done between 10 and 14 days after administration at the referring veterinarian for convenience if no hematologic toxicity was observed at the previous cycle with same dosage.

Cardiac evaluation included physical examination, thoracic auscultation, and an ECHO performed at our institution. An electrocardiogram was performed at the discretion of the cardiologist if a murmur or arrhythmia was detected. In stage 1 of the study, ECHOs were performed at baseline and repeated before each dose increment and approximately 3 months after the last chemotherapy cycle. In stage 2 of the study, ECHOs were scheduled at baseline, before cycle 5, and approximately 3 months after the last chemotherapy cycle. Measured variables included left ventricular internal dimension at end-diastole (LVIDd), left ventricular internal dimension at end-systole (LVIDs), FS, interventricular septal wall thickness at end-diastole (IVSd), interventricular septal wall thickness at end-systole (IVSs), left ventricular posterior wall thickness at end-diastole (LVPWd) and end-systole (LVPWs), aortic dimension and left atrial (LA) dimensions. Plasma CK, SDH, and AST activities were used as biomarkers of cardiac toxicity. They were evaluated as part of a standard biochemistry panel before each dose escalation. Serum cardiac troponin I (cTnI) concentrations were used to detect myocardial injury and monitor progression of cardiac damage. Three different troponin I assays were used during the course of the study depending on when and where the analyses were done.678 Serum troponin I concentration was scheduled at baseline, at cycle 4 and 6 (24–36 hours after doxorubicin administration) and at the 3-month re-evaluation, as well as in the event of any clinically recognized cardiac abnormality.

Hepatic and renal toxicoses were evaluated by serum biochemistry panel and urinalysis. Horses entered in stage 1 of the trial had weekly serum biochemistry panels during the course of the treatment. Horses entered in stage 2 of the trial had serum biochemistry panels before each dose-escalation cycle.

Gastrointestinal toxicity was evaluated by monitoring fecal consistency and signs of anorexia, stomatitis, weight loss, colic, and diarrhea. Fecal occult blood testing9 was scheduled 1 week after each dose escalation.

Hypersensitivity reactions – Cutaneous and systemic drug sensitivity reactions, as well as adverse effects caused by drug properties that are not consistent with known toxicity, were evaluated during and after administration.[17]


Analysis of peripheral cell counts and serum biochemistry results was carried out using 2-way repeated measures ANOVA to determine the effects of doxorubicin dose and schedule of administration over time. Both absolute and maximum relative decreases per treatment cycle were analyzed in this study. The effects of schedule of administration were analyzed in terms of dosage levels over time, including cycle number at which the dosage was used and dose-escalation schedule (ie, dose given at the 1st, 2nd, or 3rd step of the dose escalation). The cumulative effects between the 1st and 2nd administration at the same dosage level were analyzed for each dosage level when the drug was given at the starting dosage. ECHO parameters were analyzed using 1-way ANOVA. All posthoc pairwise comparisons of the mean values of the different groups were carried out by Tukey's test. All computations were performed by use of statistical software.10 Statistical significance was set for P < .05.



Seventeen horses with 34 measurable tumors were enrolled in this study. Mean and median ages were 15.3 and 14 years, respectively (range, 7–32 years), and the male : female ratio was 1.4 (10 : 7). There were 1 stallion, 9 geldings, and 7 mares. Breeds included Quarter Horse (n = 9), Thoroughbred (3), Arabian (1), American Paint Horse (1), and Appaloosa (2) and Pinto Pony (1). Weight ranged between 172 and 600 kg (mean, 461; SD, 113). Histologic tumor types included lymphoma (N = 7), carcinoma (8), melanoma (14), and sarcoid (5). All patients were evaluable for toxicity. There were 3 T-cell-rich B-cell lymphomas (stage IIIa, IIIb and IVb), 2 T-cell and histiocyte-rich B-cell lymphomas (stage IIa and IVb), and 2 T-cell lymphoma (both stage IVb). There were 2 anaplastic carcinomas (stage T4N2M1, and T3N2M0), 5 squamous cell carcinomas (1 stage T2N0M0, 2 stage T3N2M0, 2 stage T3N2M1), and 1 adenocarcinoma (stage T3N2M1). There were 1 subcutaneous anaplastic melanoma (stage T4N2M1), and 1 noncutaneous melanoma (T4N2M1) and dermal melanomatosis (12 lesions) affecting the tail base and penis (stage T3N1M0) in 1 horse. There were 1 fibroblastic sarcoid (stage T3N0M0) and multiple verrucous sarcoids (4 lesions) (T2N0M0) in 1 horse. Follow-up intervals for horses (N = 12) that received a full course of treatment ranged from 1 to 57 months after treatment completion (mean, 17.1 months; median, 9 months).


Ninety-five chemotherapy cycles were administered to 17 horses. Total doses ranged from 1,127 to 2,900 mg in horses that completed the 6 treatment cycles. No death caused by treatment-related toxicity occurred. In stage 1 of the study, 2 of the 3 horses received 50 mg/m2 instead of 40 mg/m2 as a starting dosage because no toxicity was seen in the first horse at a dosage 40 and 60 mg/m2. In stage 2 of the study, 6 dosage levels were evaluated, including 60, 65, 70, 75, 80, and 85 mg/m2. The dosage ranged from 0.93 to 1.41 mg/kg (mean, 1.13 mg/kg; SD, .081). The number of treatment cycles at a given dose and schedule is given in Table 1. Five horses did not complete their assigned protocols. One horse was withdrawn after cycle 2 (60 mg/m2) because of lack of compliance. Four horses were euthanized because of poor quality of life, minimal response, or tumor progression after cycle 4 (1 horse with large B-cell lymphoma stage IIIb at 60 mg/m2, and 1 horse with anaplastic melanoma stage T4N2M1 at 70 mg/m2) and after cycle 5 (1 horse with metastatic intranasal adenocarcinoma stage T3N2M1 at 70 mg/m2 and 1 horse with T-cell and histiocyte-rich B-cell lymphoma stage IV at 70 mg/m2).

Table 1. Most severe toxicoses at all evaluable dosage levels and cycle numbers
No. of CyclesDosage Level (mg/m2)GradeHypersensitivity Reactions: Cycle # (N)Neutropenia: Cycle # (N)Combined DLT Risk
1360Grade 14 (1)

Grade 1

Grade 2

3 (3), 4 (5)

4 (1)


Grade 2

Grade 3

2 (2), 3 (3), 4 (2), 5 (2), 6 (4)

2 (1), 3 (1), 6 (1)



Grade 2

Grade 3

Grade 4

2 (1), 4 (1), 6 (3)

3 (2), 4 (1), 5 (1), 6 (1)

6 (1)

6 (1)



Grade 3

Grade 4

3 (1), 4 (1), 5 (1), 6 (3)

4 (1), 5 (1), 6 (1)

5 (1), 6 (1)


Grade 3

Grade 4

5 (1), 6 (1)

5 (1)

6 (1)


Hematologic Toxicity

The most common hematologic toxicity was mild subclinical, self-limiting, and reversible neutropenia. Drug-induced neutropenia was categorized as grade 0 (count within reference range), grade 1 (2,600–2,000 neutrophils/μL), grade 2 (1,999–1,500/μL), grade 3 (1,499–1,000/μL), and grade 4 (999–500/μL). The severity of neutropenia was dose-related (Fig 1-4). Dose-limiting toxicity for the bone marrow was characterized by grade 3 neutropenia lasting <7 days. Grade 3 and 4 neutropenia were seen in 6% (1/17) of cycles at 75 mg/m2, 20% (2/10) of cycles at 80 mg/m2 and 33% (1/3) of cycles at 85 mg/m2 (Table 1). Neutropenia was not associated with fever or complicated by sepsis and was not judged to be life-threatening in the range of dosages used. Nadir neutrophil counts were seen between 1 and 2 weeks after treatment. In 7 horses that received 65 (N = 4), 70 (5), 75 (5) or 80 mg/m2 (1) CBC were obtained 10–16 days after injection on their second cycle at the same dose. Neutrophil nadir was estimated by extrapolation to be at 10.8 (±2.5 days). In 2 horses that received 80 and 85 mg/m2, bimodal nadir after weeks 1 and 3 was seen in 2 horses (3 cycles) treated with 80 and 85 mg/m2, which required 4–4½ weeks for neutrophils counts to return to >2,000/μL. Grade 2 thrombocytopenia was seen in 1 horse with advanced multicentric T-cell lymphoma after 70 mg/m2 (cycle). The horse also was treated with chloramphenicol for a purulent nonhealing skin fistula.

Figure 1.

Graph of absolute neutrophil count/μL (mean-SD) in horses receiving 2 cycles at 3-week intervals of doxorubicin at 60 mg/m2 followed by 2 cycles at 65 mg/m2 and then 2 cycles at 70 mg/m2. Time 0 is baseline before 1st infusion, arrows represent infusion times.

Figure 2.

Graph of absolute neutrophil count/μL (mean-SD) in horses receiving 2 cycles at 3-week intervals of doxorubicin at 65 mg/m2 followed by 2 cycles at 70 mg/m2 and then 2 cycles at 75 mg/m2. Time 0 is baseline before 1st infusion, arrows represent infusion times.

Figure 3.

Graph of absolute neutrophil count/μL (mean-SD)in horses receiving 2 cycles at 3-week intervals of doxorubicin at 70 mg/m2 followed by 2 cycles at 75 mg/m2 and then 2 cycles at 80 mg/m2. Time 0 is baseline before 1st infusion, arrows represent infusion times.

Figure 4.

Graph of absolute neutrophil count/μL (mean-SD) in horses receiving 2 cycles at 3-week intervals of doxorubicin at 75 mg/m2 followed by 2 cycles at 80 mg/m2 and then 2 cycles at 85 mg/m2. Time 0 is baseline before 1st infusion, arrows represent infusion times.

There was wide interpatient variability in absolute cell counts (ACC) at all dosage levels. The dosage levels of 40, 80, and 85 mg/m2 were excluded from the multivariate analysis (mixed model ANOVA) of ACC and maximum percent reduction (MPR) because of insufficient number of observations. A statistically significant difference in the values of absolute neutrophil count was found between dosage levels (P < .001) for dosage >60 mg/m2 and cycle number across time (P < .001), independently of dose-escalation schedule (P = .363, power = 0.68). With respect to MPR, there was a statistically significant difference between dosage levels (P < .001) and cycle number across time (cycle 1–6) (P = .003) for neutrophils. The higher the dosage level, the greater the differences in MPR across time. The neutrophil MPR were dose-escalation schedule invariant (P = .255, power = 0.22). No statistical differences were found for lymphocytes, platelets, and red blood cell MPR.

No statistically significant (P > .3, power = 0.39) evidence of a dose-repeat cumulative effect on ACC was found between the 1st and 2nd administration given at the starting dose level, independent of the dosage level (P > .2, power = 0.26). A significant dose-repeat cumulative effect was found for neutrophil MPR only at dosages of 75 mg/m2 (P = .003). No cumulative effect on MPR was found for dosages of 60, 65, and 70 mg/m2. Cumulative effects could not be evaluated for 80 and 85 mg/m2.

Nonhematologic Toxicity

The main clinical signs of drug reactions that were noted after IV administration of doxorubicin consisted of, in decreasing order of frequency, fever, tachypnea, tachycardia, abdominal discomfort, or some combination of the above. In horses that developed hypersensitivity reactions, blood lymphocyte, eosinophil, and basophil cell counts were normal before and 8–10 hours postinfusion.

Drug reactions were categorized as grade 1 (grade 1 fever, heart rate [HR], respiration rate [RR], and colic signs), grade 2 (grade 2 fever, HR, RR and colic signs, muscle fasciculation), grade 3 (grade 3 fever, HR, RR and colic signs, anxiety, requiring cooling and fluid therapy), and grade 4 (grade 3 with hemorrhagic diarrhea, shock, life-threatening consequences such as hemodynamic instability or ventilatory support). Reactions were dose-dependent (P = .01) and developed 4–6 hours postadministration. They were of short duration (4–8 hours) and responsive to medical treatment. The reactions were not life-threatening and all horses recovered fully. Hypersensitivity reactions were schedule invariant (P = .27). A dose-repeat cumulative effect was not found (P = .4). Grade 3 reactions were seen after drug dosages ≥75 mg/m2 (Table 1) and grade 4 reactions were not observed. Corticosteroid medication initially was included for management of grade 3 reactions with no apparent benefit.

Because of the observed increased risk of hypersensitivity reaction to doxorubicin in horses receiving dosages >65 mg/m2, premedication was modified to include flunixin meglumine (1 mg/kg IV) in addition to diphenhydramine hydrochloride (1 mg/kg IM), 30 min before infusion followed by prophylactic administration of flunixin meglumine (0.5 mg/kg IV), and diphenhydramine hydrochloride (0.5 mg/kg IM) administered 4 hours after infusion. For grade 2 and 3 drug reactions after infusion, diphenhydramine hydrochloride (0.5 mg/kg IM) and flunixin meglumine (0.5 mg/kg IV) were repeated every 4 hours, along with supportive care as needed, until resolution of clinical signs. Supportive care included fluid therapy (Normosol-R11 or Plasma-Lyte A,12 1–2 L/h for 2 hours), cooling treatment (if T ≥ 102.5F) including alcohol baths q1h, iced towel, and ice boots.

Of the 17 horses evaluated, 3 horses were noted to have a heart murmur at initial presentation and 1 horse was diagnosed with a Mobitz type 1 atrioventricular (AV) block. Baseline ECHO identified trace mitral and tricuspid regurgitation (n = 1) and mild tricuspid regurgitation (n = 1). The murmurs were classified (on a scale of 1–6) as grade 1 (n = 1), grade 2 (n = 1), and grade 3–4 (n = 1). The detection of a heart murmur and arrhythmias after variable doses of doxorubicin was not noted in any horses during treatment and follow-up period. Six horses had complete ECHO data, including baseline, before cycle 5 and 3 months after treatment completion. Seven horses had only 2 ECHOs performed at baseline and before cycle 5 and 4 horses had recheck ECHOs at 6, 12, 24, and 22 months after treatment. In these horses, LVIDd during the treatment course was the only parameter found to be significantly affected by treatment in horses that received dosages ≥70 mg/m2 (P = .002). LVIDs, FS, IVSd, and LA dimensions were not found to change significantly over time. These results should be interpreted cautiously because of the low power of the tests (ranging from 0.2 to 0.6). One horse (70–75–80 dosage-escalation schedule) that received a cumulative dosage of 450 mg/m2 had a reduction in FS (22%) measured 21 months after treatment. There were no clinical signs of heart failure in this horse and chamber dimensions and wall thickness were within normal limits. In this horse, no changes in FS were documented during treatment (42% baseline, 44% at cycle 5) and at 3-month re-evaluation (40%).

Serum CK, SDH, and AST activities remained within reference range in all horses during treatment. During routine evaluation, 5 of the 12 horses that received 6 treatments with dosages ranging from 60 to 80 mg/m2 had a complete set of cTnI test results during the course of treatments. Only 3 horses had the test performed approximately 3 months after treatment completion. All routine cTnI tests were negative (within reference range) throughout treatment and at the 3-month re-evaluation. In addition, 6 cTnI tests were performed on 4 horses that developed grade 3 hypersensitivity drug reactions and 2 test results were positive (0.08 and 0.21 ng/mL, above the normal reference range provided by the test manufacturer). In these 2 horses, the cTnI results returned to within the reference range within a day.

Overall, body weight was maintained in treated horses. A difference in average body weight of all horses between the first and last treatment cycles was not found. Maximum weight loss during the course of treatments, expressed as a percentage of initial body weight, varied from 6.7 to 15% (mean, 7.3%). Transient mild anorexia lasting <4 hours was observed in horses that experienced grade 3 hypersensitivity reactions after dosage >70 mg/m2 (2 cycles at 75 mg/m2, 3 cycles at 80 mg/m2, and 1 cycle at 85 mg/m2). Two horses became anorectic for 8 and 18 hours after treatment. These horses developed a lasting (6 and 10 hours) grade 3 hypersensitivity after treatment (70 mg/m2, cycle 4 and 80 mg/m2, cycle 4). No diarrhea or signs or gastrointestinal (GI) bleeding were observed. Occult blood was not detected in feces and the BUN : Cr ratio remained unchanged.

Serum concentrations of creatinine and BUN remained within the reference range throughout the study and after treatment at follow-up re-evaluation. Urine color changed to red or brown immediately at the first urination after administration. Urine color change remained visually apparent for up to 1 day after administration. Abnormalities in urinalysis and serum biochemistry results were not detected on the urinalyses of any horse at any dosage level.

One horse developed multifocal dermatitis after the 1st treatment cycle at a dosage of 50 mg/m2. Treatment with trimethoprim-sulfamethoxazole (5 mg of trimethoprim/kg PO q12h) in this horse also was started on the day of doxorubicin administration. A biopsy of the lesions indicated a neutrophilic, necrotizing, and ulcerative epidermitis consistent with a drug reaction. Two horses developed patchy crusting and scaling of the skin in areas of alopecia after the 3rd (75 mg/m2) and 5th cycles (75 mg/m2). Interpretation of biopsies indicated perivascular eosinophilic and lymphocytic dermatitis consistent with a drug reaction. Lesions were self-limiting and typically resolved in 1–2 weeks. Patchy hair loss was seen in 12 horses that were treated during the active hair growth season. Tail and mane were always affected to various degrees. Alopecia usually started after the third administration. Hair regrowth of the same color was observed in all horses.

One horse with penile anaplastic carcinoma (T3N2M0), 2 horses with T-cell lymphoma (stage IVb), 1 horse with perirectal carcinoma (stage T3N2M0), and 1 horse with T- and histiocyte-rich B-cell lymphoma (stage IVb) were euthanized because of tumor relapse after treatment. One horse with T-cell lymphoma (stage IVb) and 1 horse with a fibroblastic sarcoid (T3N0M0) and T-cell-rich B-cell lymphoma (stage IIIa) were euthanized 5 and 57 months after treatment, respectively, for colic and chronic lameness. Four of these 6 horses were available for necropsy. The time interval between treatment completion and necropsy ranged from 5 to 57 months (mean, 28.5 months; median, 15 months) and dosages ranged from 70 to 80 mg/m2. Microscopic examination of heart muscle (left ventricular myocardium and interventricular septum) did not disclose cytoplasmic vacuolization or loss of myofibrils. The horse that had been diagnosed with FS reduction had evidence of multifocal, subacute, tubular necrosis with multifocal, tubular regeneration, atrophy, and cellular casts. This horse was 16 years and had a metastatic penile squamous cell carcinoma that recurred 21 months after treatment. The horse was treated with palliative intent with piroxicam (0.3 mg/kg once daily PO) for 6 weeks with no response before euthanasia. One week before euthanasia, serum concentrations of creatinine and BUN were within the reference range.

Maximum Tolerated Dose

The NOAEL (no observable adverse effect level) for doxorubicin was 60 mg/m2 in this study. Although this study was not designed to evaluate the antitumor activity of doxorubicin, no tumor response was observed after 2 cycles at that dosage in any treated horse. At a dosage of 75 mg/m2, DLT was achieved in 7 of 18 cycles, which met the MTD criterion in this study. Thus, the maximum tolerated dosage of doxorubicin was 75 mg/m2. There were 12 dose reductions because of grade 3 toxicity. In 2 horses that did have dosage escalation from 70 and 75 mg/m2 at cycle 5, the dosage was escalated at cycle 6 without attendant increased toxicity. Dose-limiting toxicoses were drug hypersensitivity and neutropenia (Table 1). The combined (ie, hypersensitivity and neutropenia) MTD was achieved in 4 cycles (15%) at 70 mg/m2, in 7 cycles (39%) at 75 mg/m2, in 6 cycles (60%) at 80 mg/m2 and in 2 cycles (67%) at 85 mg/m2. Hypersensitivity and neutropenia were not associated with horse size (median cutpoint, 310 kg) and age (median cutpoint, 9 year old). Neutropenia with fever did not develop in any patient.


Doxorubicin1 is one of the most clinically effective and widely used anticancer agents in veterinary[5, 18] and human medicine.[3] Safe dosages of doxorubicin have been established for dogs and cats to treat lymphoma, osteosarcoma, hemangiosarcoma, and a variety of carcinomas. Maximum tolerated doses of antineoplastic drugs have not been established for horses. This is a critical step in the development of safe and effective chemotherapy protocols. When chemotherapy drug dosage regimens recommended for dogs are used in horses, remission usually is of short duration and toxicity is not expected.[2] To determine the role of doxorubicin in equine oncology, a safe dosage must be determined.

The acute and short-term toxicoses associated with doxorubicin administration to horses with a variety of spontaneously arising tumors were evaluated in a dose-escalating fashion. Interpatient dosage escalation historically had been considered the standard in a phase I trial. However, intrapatient dosage escalation dose-ranging designs have several advantages, including faster definition of MTD, less concern about delivering a suboptimal dose at entry level, and greater likelihood of providing more representative data about the extent of variation in individual MTD.[19] In this study, a rapid intrapatient drug dosage escalation in a cohort of 3 horses was used to determine a starting dosage and effectively decrease the number of patients that may be undertreated. This approach allowed a reduction in the number of horses required to reach MTD, and provided a margin of safety, insofar as an excessive entry level dose would never be more than one dosage level from a safe dosage. One drawback of intrapatient dose escalation is the difficulty in distinguishing acute from cumulative dose-limiting toxicities that may affect the recommended dosage.[20] In this study, however, acute and dose-limiting toxicities were easily dissociated because there was no intrapatient cumulative toxicity at the same dosage and the toxicity associated with a dosage level was schedule invariant and independent of interpatient variability. As a result, the MTD could easily be determined and the RD reliably estimated.

The dose-limiting toxicities of doxorubicin in dogs and people include myelosuppression and cardiotoxicity.[21, 22] In this study, dose-limiting toxicoses were hypersensitivity reactions and myelosuppression. No other drug toxicoses were observed in treated horses. The MTD of doxorubicin was 75 mg/m2. The recommended dosage of doxorubicin as single agent chemotherapy was 70 mg/m2. This dosage was comparable to that used in people, both as determined on a mg/m2 and a mg/kg basis.[23, 24] However, this dosage is more than twice than that used in dogs on a mg/m2 basis, but similar on a mg/kg basis.[10, 12] The level of toxicity at the RD was considered acceptable with no more than 1 DLT in 6 treatment cycles and in agreement with the standards of care in veterinary oncology. Age and patient size in this study sample were not found to affect the MTD and drug toxicity profile. Lower body weight in dogs has been shown to be associated with increased toxicity when doxorubicin dosage is based on body surface area (mg/m2).[25]

Mild bone marrow suppression was a common and self-limiting adverse effect of chemotherapy and did not require any treatment. In this study, dose dependence for blood cell count was found for neutrophils but not for other leukocytes, platelets, or red blood cells. The neutrophil nadir consistently was identified between 1 and 2 weeks after doxorubicin administration and was estimated to be 10.8 ± 2.5 days. The neutrophil count recovered to within the reference range by week 3. Based on these results, the recommended dosing interval for single-agent usage is 3 weeks.

Hypersensitivity reaction was a dose-limiting toxicity. The reactions in horses were similar to those in people and dogs, but the timing was different. Although reactions are observed during and immediately after administration of doxorubicin in dogs and people, they developed 4–5 hours after administration in horses. Signs described in people and dogs include flushing, alterations in heart rate and blood pressure, abdominal discomfort, dyspnea, bronchospasm, fever, pruritus, panting, shaking, and anaphylaxis.[22, 26] The effects have been reported to be worsened with the use of generic formulations.[27]

Hypersensitivity reactions occurring after doxorubicin administration in dogs and humans have been associated with drug-induced histamine release. Doxorubicin causes a dose-related increase in peripheral tissue histamine release and a secondary catecholamine release in response to histamine and histamine-mediated hemodynamic effects in the dog.[28]

Most of the hypersensitivity reactions in dogs and people treated with doxorubicin are not immunologically mediated.[17, 29] The reactions observed in horses in this study were not typical of a classic allergic drug reaction. They were unlikely to have been immune-mediated because subsequent reactions were not more severe and treatment with dexamethasone was not beneficial. Reactions were associated with a latency period and were dose-dependent but schedule invariant. Rechallenge at the same dose or lower dose did not result in subsequent reactions that were more severe. All reactions subsided within a few hours without short-term complications or leukocyte cell count changes that would be indicative of an allergic reaction. Although comedication with antihistamines and NSAIDs used in this study may have lessened the severity and duration of the reactions, it did not completely obviate the signs of drug toxicity.

In dogs and cats, clinical use of doxorubicin is limited by the development of a progressive dose-dependent cardiomyopathy that irreversibly evolves toward congestive heart failure that usually is refractory to conventional treatment. In dogs, the reported incidence of doxorubicin-induced cardiomyopathy with subsequent congestive heart failure ranges from 2.3 to 8.6% with a median cumulative dose of 150–155 mg/m2.[30, 31] In people, the incidence of cardiomyopathy ranges between 0.14 and 7%, with a median cumulative dose of 390 mg/m2.[32]

In this study, the risk of cardiomyopathy could not be determined with cumulative dosages ranging from 380 to 480 mg/m2 (median, 420 mg/m2). No horse developed clinical signs such as exercise intolerance, cough, respiratory distress, or labored breathing. Cardiac injury assessed during treatment and at the first re-evaluation by ongoing monitoring tests including ECHOs and blood biomarkers (CK, SDH, and AST activities and cTnI concentration) was not observed. Only 1 horse in this study had a reduction in FS 22 months after treatment without any change in ECHO parameters. This horse had no clinical signs of heart failure and no histological evidence of heart muscle injury. Histological evaluation in 5 other horses did not identify evidence of cardiac injury. Transient 24-hour increase in cTnI concentrations in 2 horses with grade 3 hypersensitivity drug reactions may have reflected intense muscular exertion while experiencing a drug reaction rather than cardiac muscle injury.[33]

The lack of doxorubicin-induced cardiotoxicity may have reflected a low plasma concentration of doxorubicin and metabolites, increased tolerance in the horse, or study design flaws. However, diet and physical conditions of the horses also may have alleviated doxorubicin-induced cardiac muscle damage. Doxorubicin cardiomyopathy has been shown to be associated with antioxidant deficit as well as increased plasma lipid concentrations.[34, 35] The substantially lower plasma triglyceride and cholesterol concentrations in horses when compared to dogs and people, because of their herbivorous diet, may have had a protective effect. In addition, fresh forage or hay and food supplements in the typical equine diet may provide enough antioxidant activity to decrease doxorubicin toxicity in horses with healthy hearts. Exercise also has been shown to counteract doxorubicin cardiotoxicity.[36] Good cardiovascular health and regular activity in horses in this study may explain better tolerance of the drug.

This study had inherent limitations that may have precluded accurate detection of cardiotoxicity. The ECHOs were performed by different cardiologists, cTnI assays were carried out using 3 different techniques and only 4 horses were available for necropsy after complete treatment. Another limitation of this study was sample size and short follow-up duration. Because cardiotoxicity occurs at a low frequency in people and dogs, the sample size in this study may not have been sufficient to allow conclusions to be drawn. In addition, conclusions regarding statistical analyses performed posthoc with low power should be interpreted cautiously.

Doxorubicin can be administered safely to horses at a dosage of 70 mg/m2 IV at 3-week intervals for up to 6 cycles. Additional studies are needed to confirm the therapeutic value of doxorubicin as either a single agent or in combination with other drugs in the treatment of horses with cancers.


Supported in part by the Cunningham and Doyle Charitable Trust Fund and the UC Davis Center for Equine Health. The authors thank Catherine Glines for her help with scheduling, sample acquisition, and client communication throughout the clinical trial.

Conflict of Interest: Authors disclose no conflict of interest.


  1. 1

    Adriamycin, Bedford Laboratories, Bedford, OH

  2. 2

    Angiocath, BD Medical, Sandy, UT

  3. 3

    Interlink System, 10 drops/mL, Baxter Healthcare Corporation, Deerfield, IL

  4. 4

    Extension set, Smith Medical ASD Inc, Dublin, OH

  5. 5

    T-port, Braun Medical Inc, Bethlehem, PA

  6. 6

    Stratus CS Troponin I assay, Dade Behring, Newark, DE (normal range: ≤0.06 ng/mL). Assay performed at the Heart Station, New Bolton Center, University of Pennsylvania

  7. 7

    Beckman Access 2 Troponin I assay, Beckman Coulter, Inc, Chaska, MN (normal range: 0.00–0.06 ng/mL). Assay performed at the University of California – Davis, Medical Center

  8. 8

    Immulite 2000 Troponin I assay, Siemens Medical Solutions, Los Angeles, CA (normal range: 0.01–0.07 ng/mL). Assay performed at the University of California – Davis, Veterinary Medical Teaching Hospital

  9. 9

    Hemocult, Beckman Coulter Inc, Chaska, MN

  10. 10

    SPSS Version 11.5, SPSS Inc, Chicago, IL

  11. 11

    Normosol-R, Hospira, Inc, LakeForest, IL

  12. 12

    Plasma-Lyte A, Abbott Laboratory, North Chicago, IL