49

a 23% increased risk of death (22.0 per10,000 exposures). This analysis of medicalrecords for more than 7 million office visitsand over 2 million dogs demonstrates thefeasibility of using large electronic databas-es to test hypotheses generated by sponta-neous adverse event reports to the UnitedStates Food and Drug AdministrationCenter for Veterinary Medicine. In addition,information can be generated on baselineoccurrences of certain conditions in a largepopulation of dogs presented to veterinaryhospitals across the United States.

INTRODUCTIONProHeart 6 (moxidectin) was launched byFort Dodge Animal Health (Overland Park,KS) in June 2001 with an indication to pre-vent canine heartworm disease caused byDirofilaria immitis for 6 months and to treatexisting larval and adult stages of the caninehookworm Ancylostoma caninum. SinceProHeart 6 was introduced to the market,the United States Food and Drug

Safety Profile of Moxidectin(ProHeart 6) and Two OralHeartworm Preventives in Dogs Larry T. Glickman, VMD, DrPH*

Nita W. Glickman, MPH, PhD*

George E. Moore, MS, DVM*

Rami Cobb, BVSc†

Stephen A. Connell, DVM†

Mitch Morrison‡

Hugh B. Lewis, BVMS§

KEY WORDS: Adverse events, dogs,heartworm, moxidectin (ProHeart 6), phar-macovigilance, vaccines

ABSTRACTMedical records of a nationwide veterinarypractice (Banfield, the Pet Hospital) wereevaluated to determine the incidence ofadverse events and particular health prob-lems following administration of the sus-tained-release injectable heartwormpreventive moxidectin (ProHeart 6), 2 oralmonthly heartworm preventives, and/or vac-cines in dogs. Similar information wasreviewed for dogs receiving neither heart-worm preventives nor vaccines. The safetyprofile of these products was comparable.However, ProHeart 6 was associated with a27% increased risk of mast cell tumor (2.1per 10,000 exposures), while one of the oralheartworm preventives was associated with

Intern J Appl Res Vet Med • Vol. 3, No. 2, 2005

*Department of Veterinary Pathobiology ‡Section of Veterinary Information Services School of Veterinary MedicinePurdue UniversityWest Lafayette, IN

†Fort Dodge Animal HealthOverland Park, KS

§Banfield, the Pet HospitalPortland, OR

Presented in part to the Veterinary Medicine AdvisoryCommittee of the Food and Drug Administration, Center forVeterinary Medicine, Rockville, MD, at a public hearing onProHeart 6, January 31, 2005.

Intern J Appl Res Vet Med • Vol. 3, No. 2, 200550

Administration’s (FDA) Center forVeterinary Medicine (CVM) has reportedthat they have received nearly 5,500 reportsof serious adverse drug reactions attributedto ProHeart 6.1 Of these adverse eventreports, at least 1,900 were thought by CVMto be unrelated to the concurrent administra-tion of other drugs or vaccines. Many ofthese events were judged by CVM to besevere, including more than 600 reports ofdeath. Following discussions with CVM,Fort Dodge Animal Health announced onSeptember 3, 2004, that it was voluntarilyceasing production and recalling ProHeart 6from the market pending a review by anindependent scientific panel.

Prelicensing studies of ProHeart 6 byFort Dodge Animal Health did not reveal anyserious adverse events. Healthy dogs treatedwith ProHeart 6 either 1, 3, or 5 times at therecommended dose of 0.17 mg/kg did notdemonstrate any clinical signs, laboratoryfindings, or necropsy lesions associated withsystemic or target organ toxicity.2 In addition,when ProHeart 6 was administered to healthy10-week old puppies at 3 or 5 times the rec-ommended dosage or to genetic lines ofCollie dogs sensitive to administration ofivermectin, no clinical or laboratory abnor-malities were observed. Field studies ofProHeart 6 in 374 client-owned dogs treatedtwice at 6-month intervals at the recommend-ed dose by veterinarians in 4 different statesresulted in the following observed adverseevents: vomiting (3 dogs), diarrhea (2 dogs),weight loss (2 dogs), listlessness (1 dog),injection site pruritus (1 dog), and elevatedbody temperature (1 dog). However, preli-censing safety and efficacy studies such asthese typically involve fewer than 500 dogsand lack sufficient statistical power to identi-fy relatively rare, though possibly serious,adverse events. As a result, rare drug-associ-ated adverse events are not well characterizeduntil after widespread marketing.3

Postmarketing surveillance of veterinarydrug-associated adverse events currentlydepends on passive, spontaneous reportingby pet owners and veterinarians directly to

the pharmaceutical company or CVM. Whenthe number of spontaneous reports is suffi-cient to signal a potential safety problem, thesame regulatory agency (CVM) that licensedthe drug reviews the findings and mayimplement several decisions. These mayrange from no further action required,requesting changes to the approved labeling,requesting the drug be voluntarily withdrawnfrom the market, referring the matter to aCVM Advisory Committee for advice, orrequesting that further studies be conductedto better understand the postmarketingobservations. Spontaneous reporting ofadverse events alone, whether to the pharma-ceutical company or CVM, cannot by itselfbe used to calculate incidence rates of drug-associated adverse events, since the numberof adverse events that actually occur and thenumber of dogs that received a drug (thepopulation at risk) are both unknown.Following voluntary recall of ProHeart 6,Fort Dodge Animal Health asked one of theauthors (LTG) to conduct an epidemiologi-cal study to determine the incidence ofpotential adverse events associated withProHeart 6 and to compare these with thesafety profile of 2 oral, monthly heartwormpreventives and vaccines. This article reportsthe results of such a study utilizing the elec-tronic medical records of Banfield, the PetHospital, a national veterinary hospital.

METHODS

Data SourceBanfield, the Pet Hospital, was founded in1955 in Portland, OR. By 2004 Banfieldoperated a national network of 403 full-serv-ice primary care animal hospitals in over 40states with approximately 1.4 million activepatients and 60,000 patient-visits per week.Banfield hospitals are paperless and utilizeproprietary software (PetWare) to createelectronic medical records that are uploadedweekly to a central data warehouse, wherethey are stored in Oracle (Redwood Shores,C A ) format. Medical records for the periodof time from January 1, 2002, to August 31,2004, when ProHeart 6 was recalled, were

51

transferred to Purdue University as pipe-delimited ASCII files. The files includedinformation on 7,075,250 encounters (officevisits) for 2,047,809 dogs. The files wereconverted into data sets for analysis.

Data AnalysisPotential adverse events were categorized asliver related, neurological, ocular, immunemediated, allergic, anaphylaxis, cardiac,cancer, or death, based on a combination ofclinical signs or laboratory findings (Table1). The incidence of adverse events for eachexposure group was calculated using SASsoftware (version 9.1.3, SAS Institute, Cary,NC, USA) and expressed as either the num-ber of adverse events per 10,000 encountersor as the number of adverse events per10,000 days at risk (for example, 10 days atrisk could be 1 dog followed for 10 dayspostexposure or 10 dogs followed for oneday each postexposure). Formal statisticalanalysis to evaluate differences betweenexposure types was generally not done dueto the very large sample sizes. That is, thepower to detect very small differences inadverse event rates for common outcomesbetween exposure groups was extremelyhigh, even when such differences wereunlikely to have any clinical significance.However, the same was not necessarily truefor less common outcomes or when multi-variate analyses were performed.Multivariate logistic regression models weredeveloped using SAS version 9.1.3 softwarewith the PROC LOGISTIC procedure.Results were expressed in terms of oddsratios, 95% confidence interval of the oddsratio, and P values. A P value < 0.05 wasconsidered statistically significant.

Study AssumptionsIt was assumed that oral, monthly heart-worm preventives had been administered byowners to dogs on the same day of theencounter in which they appeared in themedical record. The potential impact of thisassumption would be to u n d e r e s t i m a t e t h eincidence rate of adverse events associatedwith the 2 oral, monthly heartworm preven-

tive drugs, since some doses of oral heart-worm drugs were either never given to theirdog by owners or were given, but not at thebeginning of the 30-day follow-up period.

RESULTSFrom January 1, 2002, to August 31, 2004,there were 6,800,061 encounters for1,983,162 individual dogs that met study eli-gibility criteria. These encounters or expo-sures were grouped as follows (number ofencounters): ProHeart 6 with or without avaccine (735,654), Heartworm preventive 1with or without a vaccine (411,082),Heartworm preventive 2 with or without avaccine (18,405), any vaccine without con-current administration of a heartworm preven-tive (1,489,032), or none of these (4,144,984)(Table 2). The proportion of encounters asso-ciated with vaccination was 62.9% forProHeart 6, 59.9% for Heartworm 1, and65.1% for Heartworm 2. Vaccines were onlyadministered during 26.4% of the encountersin which no heartworm preventive was given.This may be because many of these dogs pre-sented with a health problem for which heart-worm preventive drugs or vaccines were notindicated, and/or they may have already beenon heartworm prophylaxis at the time of thevisit. That is, these dogs were likely to be lesshealthy than dogs given a heartworm preven-tive or vaccine.

Univariate AnalysesThe number of doses of ProHeart 6 admin-istered monthly by Banfield veterinariansincreased over time, with more being givenduring the peak of mosquito activity fromMarch to September (Figure 1). The rate ofany adverse event per 10,000 encounterswas higher for dogs that received any of theheartworm preventives concurrent with vac-cination compared with dogs that receivedany of the heartworm preventives withoutvaccination (Table 2). However, for dogsthat did not receive any heartworm preven-tive, vaccination was associated with alower rate of adverse events compared withdogs that were not vaccinated.

Intern J Appl Res Vet Med • Vol. 3, No. 2, 2005

Intern J Appl Res Vet Med • Vol. 3, No. 2, 200552

Table 1. Categorization of Potential Drug- and Vaccine-Associated Adverse Events UsingMedical Records of Dogs Presented to Banfield, the Pet Hospital*

Disease Category Adverse Event CriteriaLiver disease Liver: any diagnosis Diagnosis: hepatopathy, hepatitis, hepatitis

encephalopathy, hepatitis acute, hepaticdisease

Liver: elevated ALP ALP ≥ 393 IU/LLiver: elevated ALT ALT ≥ 236 IU/LLiver: elevated GGT GGT ≥ 24 IU/LLiver: elevated total bilirubin Total bilirubin ≥ 1.0 mg/dLLiver: any elevated enzyme Any elevated enzyme (ALP, ALT, GGT, or

total bilirubin)Liver: any diagnosis plus any Any liver disease diagnosis plus any elevated enzyme elevated enzymeLiver: any diagnosis or any Any liver disease diagnosis or any elevated enzyme elevated enzyme

Neurological disease Neurological: any diagnosis Diagnosis: encephalopathy,or exam finding meningitis, epilepsy, behavioral disorders of

unknown origin, seizure-acquired, shock(cardiovascular); exam finding: paresis,paralysis, ataxia

Ocular disease Ocular: any diagnosis Diagnosis: optic neuritis, retinal or exam finding degeneration, anisocoria; exam finding:

visual deficit, abnormal visual acuity

Immune-mediated Thrombocytopenia Diagnosis: thrombocytopenia;disease thrombocytopenia, immune mediated

Immune-mediated disease Diagnosis: Immune-mediated disease or AHAImmune-mediated disease plus Diagnosis: Immune-mediated diseaseabnormal laboratory value or AHA with an abnormal reticulocyte countImmune-mediated disease: any Any immune-mediated disease adverse event

Allergic reaction Allergic reaction Diagnosis: allergic reaction, drug reaction,drug induced disease, acute allergic reaction,vaccine reaction, urticaria, drug eruption

Cardiac disease Cardiac murmur Diagnosis: cardiac murmurCardiac arrhythmia Diagnosis: cardiac arrest, atrial fibrillation,

atrial premature contractions, atrialtachycardia, bundle branch block, heart block1st degree, heart block 2nd degree, heartblock 3rd degree, cardiac arrhythmia,ventricular premature contractions,ventricular tachycardia

Cancer Mast cell tumor Diagnosis: mast cell tumorLymphosarcoma Diagnosis: lymphosarcoma Histiocytoma Diagnosis: histiocytomaCancer: any Any cancer diagnosis

(mast cell, lymphosarcoma, or histiocytoma)

*ALT indicates alanine aminotransferase; ALP, alkaline phosphatase; GGT, gamma glutamyltransferase; AHA, autoim-mune hemolytic anemia.

53Intern J Appl Res Vet Med • Vol. 3, No. 2, 2005

Intern J Appl Res Vet Med • Vol. 3, No. 2, 200554

The incidence of any liver-relatedadverse event per 10,000 encounters washigher for dogs receiving ProHeart 6 thanfor dogs receiving either of the 2 oral,monthly heartworm preventives, regardlessof whether a vaccine was administered ornot (Table 3). However, the incidence ofliver-related adverse events was not as highas for dogs in the group that received noheartworm preventive drug or vaccine.When the incidence of liver-associatedadverse events was measured per 10,000days at risk, the rates were comparable fordogs that received Heartworm 1 only (1.15),

ProHeart 6 only (1.17), or vac-cine only (1.27). The mean num-ber of days at risk for dogsreceiving ProHeart 6 alone was29.2 compared with 27.2 and27.4 for dogs receivingHeartworm 1 or Heartworm 2,respectively. This same patternof risk persisted when liver-relat-ed adverse events were evaluat-ed separately for any laboratoryabnormality or any clinical diag-nosis consistent with liver dys-function (data not shown).

The incidence of allergicreactions was similar for dogsthat received ProHeart 6, any ofthe oral, monthly heartwormpreventives, or vaccine alone.However, the incidence of aller-gic reactions was consistentlyhigher for vaccinated comparedwith unvaccinated dogs, regard-less of the heartworm preventivethey received. In addition, theincidence of allergic reactionsper 10,000 days at risk was 0.52for dogs that receivedHeartworm 1 only, 0.68 forthose receiving Heartworm 2only, and 0.62 for the ProHeart6 only dogs, versus 1.65 fordogs that received a vaccineonly. More than 95% of allallergic reactions occurred in thefirst 3 days following an expo-

sure, while liver-associated adverse eventswere more evenly distributed over the 30-day follow-up period. The rate of anaphy-laxis was lower than for any of the otheradverse events studied. The highest rate ofanaphylaxis was observed in the group ofdogs that received Heartworm 2 plus vac-cine, but this rate was based on only 2observed anaphylactic events.

The incidence of neurological, ocular, orimmune-mediated events were all relativelylow and there was no obvious associationwith respect to exposure group, except for

Figure 1. Number of doses of ProHeart 6 administered todogs by veterinarians at Banfield, the Pet Hospital, fromJanuary 1, 2002, to August 31, 2004.

Figure 2. Risk of potential liver-related adverse events indogs as a function of age following administration ofProHeart 6 by veterinarians at Banfield, the Pet Hospital. Riskwas based on results of multivariate logistic regression.**Using the odds ratios (OR) for age and for the interaction of age xProHeart 6 administration, respectively, from the multivariate logisticmodel, the following equations compute the risk of potential liver-related adverse events:Risk (for a 1-year old) = 0.854 x 1.043 = 0.89Risk (for a 7-year old) = (0.853 x 1.043)7 = 1.14

55

the dogs that received no heartworm pre-ventive or vaccine; these dogs had the high-est risk. There was no apparent associationbetween the incidence of cardiovascular dis-ease and exposure group except for dogsthat received no heartworm preventive orvaccine, which had the highest risk.

The death rate per 10,000 encounterswas higher for Heartworm 1 alone than itwas for Heartworm 2 or ProHeart 6,whether administered alone or with a vac-cine, or for dogs in the vaccine-only group(Table 4). A similar pattern was observedwhen the death rate was measured per10,000 days at risk (data not shown). Thehigher death rate for dogs in the group thatneither received a heartworm preventive norwere vaccinated is indicative of the fact thatthese dogs probably presented for a medicalcondition rather than for preventive medi-cine or a wellness program (Table 4).

The incidence of cancer was higher fordogs that received ProHeart 6 alone or withvaccine than it was for dogs in the otherexposure groups, with the exception of dogsthat did not receive either a heartworm pre-ventive or vaccine. The incidence of mastcell tumor per 10,000 encounters was higherfor dogs that received ProHeart 6 either withor without a vaccine, compared with dogsthat received either of the 2 oral, monthlyheartworm preventives, or vaccine alone(Table 5). In contrast, no such associationwith ProHeart 6 was observed for lym-phosarcoma or histiocytoma. The incidenceof mast cell tumor per 10,000 days at riskwas 0.024 for dogs that received Heartworm1 only, 0.0 for Heartworm 2 only, 0.072 forProHeart 6 only, and 0.043 for dogs thatreceived a vaccine only. The incidence ofmast cell tumor per 10,000 encounters fordogs that had received 1, 2, 3, 4, or 5 ormore doses of ProHeart 6 was 1.91, 2.01,1.39, 3.31, and 3.35, respectively, showingno statistically significant dose-responserelationship with subsequent treatments. Theclinical significance of a diagnosis of mastcell tumor within 30 days of treatment is notknown. In addition, it was not known if

these dogs had ever been given ProHeart 6by a non-Banfield veterinarian.

Multivariate Analyses Separate multivariate logistic regressionmodels were developed for the risk ofadverse events including liver disease, aller-gic reactions, cancer, and death, in order tocontrol for potential confounding effectsand to identify interactions between inde-pendent variables. The independent vari-ables included in each model were exposuregroup (ProHeart 6, Heartworm 1,Heartworm 2, and vaccine), ProHeart 6 dosenumber, age, weight, use of nonsteroidalanti-inflammatory drugs (NSAIDs), orsteroid drugs. All possible 2-way interac-tions of these independent variables werealso evaluated.

In the liver disease model (Table 6),steroid use was associated with a 25%increased risk, while ProHeart 6 was associ-ated with a 15% reduction in risk. Each addi-tional dose of ProHeart 6 resulted in an 8%reduction in the risk of liver-associatedadverse events. However, there was evidenceof a significant interaction (effect modifica-tion) between age and ProHeart 6. Using thebest-fit equation from the logistic regressionmodel, the relationship between the risk ofliver-associated adverse events and ProHeart6 administration was graphed as a function ofage (Figure 2). The risk of a liver-associatedadverse event in dogs receiving ProHeart 6increased with increasing age. ProHeart 6was associated with a decreased risk of liverdisease in dogs less than 4 years of age andan increased risk of liver disease in dogsgreater than 4 years of age.

In the allergic reaction model (Table 6),ProHeart 6, Heartworm 1, vaccines,NSAIDs, and steroid use were all associatedwith increased risk, with vaccines havingthe strongest effect. However, each addi-tional dose of ProHeart 6 was associatedwith a significant 7% reduction in the riskof allergic events. In the model for death(Table 6), Heartworm 1 was associated witha significant 23% increased risk, whereas

Intern J Appl Res Vet Med • Vol. 3, No. 2, 2005

Intern J Appl Res Vet Med • Vol. 3, No. 2, 200556

57

ProHeart 6 was associated with a 71%decreased risk. However, age significantlymodified the effect of ProHeart 6 such thatthe protective effect of ProHeart 6 for deathapplies primarily to low-weight, low-agedogs without concurrent vaccine. Each addi-tional dose of ProHeart 6 further reducedthe risk of death by 9%.

In the mast cell tumor model (Table 6),ProHeart 6 was associated with a 26%increased risk, whereas the 2 oral, monthlyheartworm preventives and vaccines wereunassociated with mast cell tumor risk.NSAIDs also were associated with a 422%increased risk of mast cell tumor. Steroidadministration was associated with a 182%increased risk of lymphosarcoma, which canbe explained by the fact that Banfield vet-erinarians administer long-term steroids aspart of the lymphosarcoma treatment proto-col. None of the exposure groups was asso-ciated with an increased risk ofhistiocytoma, but vaccines were protective.

DISCUSSION Mosquito-transmitted canine heartworminfection has been diagnosed in dogs in manyparts of the world and is endemic in the 48contiguous states in the United States.4 W h i l ethe prevalence of heartworm infection in dogsand the length of the mosquito season variesfrom state to state, the peak heartworm trans-mission generally occurs in July and Augustand may last for 6 months above the 37th par-allel. Despite widespread availability ofmonthly heartworm preventives, the infectionrate increased in the 1990s, while the use ofheartworm preventives declined.5 S u r v e y shave shown that compliance (reliable monthlytreatment by owners) is problematic and is alimiting factor in the control of heartworm inthe dog population.6 In June 2001 moxidectinin the form of ProHeart 6 was approved andlaunched in the United States by Fort DodgeAnimal Health to prevent canine heartworminfection for 6 months and to treat existinglarval and adult stages of the canine hook-worm. In addition to its duration of action, themajor advantage of this product is the fact

that it does not depend on dog owners toadminister it on a monthly basis. This productprovides an avenue for continuous protectionagainst heartworm infection while decreasingdependence on owner compliance comparedwith monthly preventives. Since experimentalstudies have shown that ProHeart 6 has simi-lar efficacy to the commonly used oral,monthly heartworm preventives, and sincecompliance with an injectable drug likeProHeart 6 is greater than for heartworm pre-ventives that depend on administration by dogowners, ProHeart 6 is likely to be more effec-tive for pet dogs than oral, monthly heart-worm preventives. The results of this studyindicate that the incidence of adverse eventsfollowing administration of ProHeart 6 isindeed comparable to 2 oral, monthly heart-worm preventives, despite the fact that nearly5,500 reports of adverse events following theuse of ProHeart 6 have been submitted to theCVM at the time of this study.

By the third quarter of 2004, ProHeart 6was the number-two product sold in theUnited States for heartworm prevention,with a 24% market share. ProHeart productshave also been registered in a number ofinternational markets since 2001, includingAustralia, Canada, the European Union(France, Greece, Italy, Portugal, and Spain),Korea, and Japan. The length of activityclaims and active ingredient concentrationvary depending on the market. For example,ProHeart 12 has gained a 47% market sharein Australia, where it offers 12 months ofprotection and contains 3 times the amountof moxidectin as ProHeart 6. In Italy, thesame product (trade name GUARDIAN SR)is expected to achieve a 35% market shareby the end of 2004. Shortly after the USlaunch of ProHeart 6, CVM expressed con-cern about a number of reports of allergic-type reactions after administration, rangingfrom mild and self-limiting to severe ana-phylactoid reactions.

When Fort Dodge Animal Healthannounced that it was voluntarily ceasingproduction and recalling ProHeart 6 fromthe US market until resolution of CVM

Intern J Appl Res Vet Med • Vol. 3, No. 2, 2005

Intern J Appl Res Vet Med • Vol. 3, No. 2, 200558

Table 6. Stepwise Multivariate Logistic Regression Risk Analysis for Potential Adverse Events FollowingAdministration of ProHeart 6, 2 Oral Heartworm Preventive Drugs, Steroids, NSAIDs, or Vaccines.*

Adverse Event Odds 95% StatisticalRatio Confidence Limits Significance

Liver

Age 1.16 1.16 1.17 <.0001Weight 0.99 0.99 0.99 <.0001ProHeart 6 0.85 0.79 0.92 <.0001Steroid 1.24 1.04 1.49 0.015ProHeart 6 each additional dose 0.92 0.88 0.95 <.0001Age x ProHeart 6 1.04 1.03 1.05 <.0001Weight x steroid 1.00 1.00 1.00 0.024

Allergic reactionHeartworm preventive 1 1.11 1.03 1.20 0.005ProHeart 6 1.48 1.37 1.60 <.0001ProHeart 6 each additional dose 0.93 0.90 0.97 0.002Age 0.84 0.81 0.88 <.0001Vaccine 2.90 2.51 3.36 <.0001NSAID 1.64 1.23 2.20 0.001Steroid 2.11 1.71 2.61 <.0001Heartworm preventive 1 x age 1.04 1.01 1.07 0.001Heartworm preventive 1 x weight 0.99 0.99 1.00 0.029Vaccine x NSAID 0.41 0.30 0.58 <.0001Age x vaccine 1.11 1.06 1.15 <.0001Weight x vaccine 0.99 0.99 0.99 0.005Age x steroid 0.89 0.84 0.94 <.0001Weight x steroid 0.99 0.98 0.99 0.011

Mast cell tumorAge 1.27 1.23 1.30 <.0001Weight 1.01 1.01 1.01 <.0001ProHeart 6 1.26 1.01 1.59 0.044NSAID 5.22 2.80 9.73 <.0001Age x NSAID 0.92 0.85 0.99 0.038

LymphosarcomaAge 1.36 1.28 1.44 <.0001Weight 1.01 1.00 1.02 0.001Vaccine 0.36 0.21 0.63 0.001Steroid 2.82 1.12 7.06 0.027

HistiocytomaAge 0.72 0.69 0.75 <.0001Weight 1.01 1.01 1.01 <.0001Vaccine 0.62 0.50 0.77 <.0001

*Only factors significant in each model at P < 0.05 are shown.continued on page 59

59

safety concerns based on adverse eventreports it had received, regulatory agenciesin other countries also reviewed the safetyrecord of ProHeart in dogs. To date, no reg-ulatory agency outside of the United Stateshas found ProHeart to present a clear andpresent danger to dogs, and with the excep-tion of Korea, have not taken any action torecall the product. For example, based on areview of suspected adverse reactions, theCanadian Veterinary Drugs Directoratedetermined that “an immediate recall is notwarranted for ProHeart 6 in Canada.”7

Since ProHeart products are all manu-factured at one site by the same method,using the same ingredients, one may specu-late that dogs in the United States somehowreact differently to ProHeart 6 than dogs inother countries. A more likely explanationhowever, is that the CVM in the UnitedStates interpreted adverse event reports toProHeart 6 differently than regulatory agen-cies in other countries, or that US pet own-ers are more inclined to report adversereactions. An analysis of Banfield data inthis study showed that allergic reaction ratesfollowing ProHeart 6 administration werecomparable overall to those for 2 oral heart-

worm preventives and far less than the inci-dence when any of the heartworm preven-tives studied were administered with avaccine, or when vaccine alone was admin-istered. Also, the incidence of allergicevents following ProHeart 6 administrationby Banfield veterinarians decreased by26.4% from the first quarter of 2002 to thethird quarter of 2004, when ProHeart 6 wasrecalled (data not shown).

The major finding from this epidemio-logical postmarketing study involvingalmost 2 million dogs seen by Banfieldhospitals over a 2-year period was similari-ty in the safety profiles of ProHeart 6 andtwo oral heartworm preventives. WhileProHeart 6 was associated with anincreased incidence of liver-related adverseevents in univariate analysis, this sameincrease was not found in either the days-at-risk analysis or when controlling for con-founding factors such as age in multivariateanalyses. Another important finding wasthat ProHeart 6, the 2 monthly heartwormpreventive drugs, and vaccines were allassociated with a clinically significantincrease in the incidence of allergic reac-tions, especially during the first few days

Intern J Appl Res Vet Med • Vol. 3, No. 2, 2005

Table 6. c o n t i n u e dAdverse Event Odds 95% Statistical

Ratio Confidence Limits SignificanceDeathHeartworm preventive 1 1.23 1.05 1.42 0.008ProHeart 6 0.29 0.22 0.38 <.0001ProHeart 6 each additional dose 0.91 0.85 0.97 0.005Vaccine 0.55 0.44 0.68 <.0001Age 1.21 1.17 1.24 <.0001Weight 0.99 0.99 0.99 <.0001ProHeart 6 x vaccine 2.06 1.61 2.64 <.0001Age x Heartworm preventive 1 0.96 0.94 0.99 0.021Age x ProHeart 6 1.03 1.01 1.05 0.001Age x vaccine 0.96 0.93 0.98 0.002Age x steroid 1.05 1.01 1.09 0.019Age x NSAID 1.03 1.01 1.05 0.021Weight x Heartworm preventive 1 0.99 0.99 1.00 0.037Weight x ProHeart 6 1.01 1.00 1.01 0.001Weight x NSAID 1.01 1.01 1.01 <.0001

*Only factors significant in each model at P < 0.05 are shown.

Intern J Appl Res Vet Med • Vol. 3, No. 2, 200560

post-exposure. In addition, one of the oralheartworm preventive drugs (Heartworm 1)was associated with a clinically significantincreased risk of death within 30 days fol-lowing administration.

The only potential adverse event studiedthat was independently associated with anincreased incidence following administrationof ProHeart 6 was mast cell tumor.However, the absolute magnitude of theincreased risk of mast cell tumor associatedwith ProHeart 6 alone (2.1 per 10,000 dosesadministered) or ProHeart 6 plus vaccine(1.9 per 10,000 doses administered) wassmall. For example, it would require aprospective study of 600,000 dogs receivingProHeart 6 alone and 600,000 dogs receiv-ing ProHeart 6 plus a vaccine to detect thissmall difference in mast cell tumor incidence90% of the time with a Type I error of0.05%. Also no significant dose-responserelationship was observed between the num-ber of ProHeart 6 doses a dog had receivedand the risk of developing mast cell tumor.Two-year studies in mice8 and rats9 did notfind any evidence of moxidectin-related tar-get organ toxicity or tumorigenicity.Furthermore, a plausible mechanism has notbeen proposed whereby moxidectin or simi-lar chemical compounds can induce or pro-mote cancer formation, especially within 30days of administration.

This epidemiologic study also demon-strates the advantage of using electronicmedical records from a large primary careveterinary practice rather than relying on thecurrent pharmacovigilance system of pas-sive or spontaneous reporting of drug-asso-ciated adverse events in animals to CVMand drug companies. The latter depends onveterinarians or dog owners deciding whatconstitutes an adverse event and requiresthem to report such events either by tele-phone or in writing. It is well known thatpassive systems are plagued by a varyingdegree of underreporting.1 0 In addition, newsreleases or Internet postings about adverseevents associated with a product may biaspeople to report adverse events, whereas

Internet reports are less likely to influencethe occurrence of potential adverse eventsgleaned directly from medical records.Other advantages to the use of medicalrecords over spontaneous reporting are thatit facilitates calculation of drug-associatedadverse event rates (absolute risk) based onthe population at risk or number of dosesadministered and allows adjustment of theserates for potential confounding factors (age,breed, weight) and for concurrent adminis-tration of other drugs or vaccines.

Epidemiological studies are not intendedto replace passive surveillance followingmarketing of new veterinary products.Rather, they are most useful for testinghypotheses or alarms raised through passivepostmarketing surveillance. Epidemiologicalstudies of this type will be facilitated in thefuture by the trend toward computerizationof veterinary medical records and the growthof large corporate and private veterinarypractices. The results of the epidemiologicalstudy described in this report do not indicatemajor safety concerns regarding ProHeart 6,which affords 6 months of continuous pro-tection against heartworm prevention andreduces dependence on owner compliancefor its effectiveness. Since no drug is com-pletely safe, veterinarians can use the resultsof this study to select the most appropriateheartworm preventive strategy for individualdogs. The practice of evidence-based veteri-nary medicine ultimately depends on theavailability of unbiased findings from con-trolled population-based studies.

ACKNOWLEDGMENTSThis study was funded in part by grant RO1CI 000093 from the US Centers for DiseaseControl and Prevention, and by a contractwith Fort Dodge Animal Health.

We appreciate the help of the Banfieldveterinarians who generated the medicalrecords reviewed in this study and Dr. ScottCampbell, Chairman and Chief ExecutiveOfficer of Banfield, the Pet Hospital, formaking this valuable resource available toPurdue University at no cost.

61

REFERENCES 1 . US Food and Drug Administration Center for

Veterinary Medicine. Dear Doctor letter con-cerning ProHeart 6 (Moxidectin) from FortDodge Animal Health. July 22, 2002. Availableat h t t p : / / w w w . f d a . g o v / c v m. Accessed March2 0 0 5 .

2. Fort Dodge Animal Health. 3 Month TargetAnimal Safety Study (Toxicity) with MoxidectinCanine SR (Sustained Release) InjectableFormulation. Overland, KS:0899-C-US-4-98,GASD 06–15.00, 1998.

3 . US General Accounting Office. Food and DrugA d m i n i s t r a t i o n : Effect of User Fees on DrugApproval Times, Withdrawals, and Other AgencyA c t i v i t i e s . Publication GAO-02-958. Washington,DC: General Accounting Office; 2002.

4. McCall JW, Guerrero J, Genchi C, Kramer L.Recent advances in heartworm disease. VetParasitol. 2004;125:105–130.

5. American Heartworm Society. AmericanHeartworm Society reminds dog owners toremain vigilant against heartworm disease. [pressrelease] March 22, 2003. Available at

http://www.heartwormsociety.org/MediaRelease.htm. Accessed March 2005.

6 . American Animal Hospital Association. The Pathto High-quality Care: Practical Tips forImproving Compliance. Lakewood, CO: AmericanAnimal Hospital Association; 2003:3–8.

7. Veterinary Drugs Directorate, Health CanadaHealth Products and Food Branch. HealthCanada’s Update On the Safety Evaluation ofProHeart 6 Sustained Release Injection.Available at http://www.bcvma.org/cfm.Accessed March 2005.

8. Goldenthal EI. Chronic Dietary Toxicity andOncogenicity with AC 301,423 in Mice.Mattawan, MI: American Cyanamid;1992. Study141–031. Protocol 971–89–155.

9. Zoetis T. Chronic Dietary Toxicity andOncogenicity Study with AC 301,423 in Rats.Vienna, VA: American Cyanamid; 1992.Hazelton–Washington Study 362–202, protocolnumber 971–89–156.

10. Strom BL. Potential for conflict of interest in theevaluation of suspected adverse drug reactions.JAMA. 2004; 292:2643–2646.

Intern J Appl Res Vet Med • Vol. 3, No. 2, 2005