Journal of Zoo and Wildlife Medicine 27(4): 468-476. 1996Copyright 1996 by American Association of Zoo Veterinarians

MAMMARY CANCER IN CAPTIVE WILD FELIDS AND RISKFACTORS FOR ITS DEVELOPMENT: A RETROSPECTIVESTUDY OF THE CLINICAL BEHAVIOR OF 31 CASES

Lisa A. Harrenstien, D.V.M., Linda Munson, D.V.M., Ph.D., Ulysses S. Seal, Ph.D.,and The American Zoo and Aquarium Association Mammary Cancer Study Group

Abstract: This retrospective study was initiated because of the perceived high incidence of mam-mary gland cancer in zoo felids in which progestin contraception has been used. Our purpose wasto describe the clinical behavior of these tumors and identify risk factors for their development.Clinical and historic records were reviewed from 31 captive wild felids with palpable and histo-logically confirmed mammary gland neoplasia. All mammary masses were classified histopatho-logically as carcinomas, with cribiform, solid, cyst-papillary, and adenocarcinoma patterns. Thoracicradiography was a sensitive antemortem procedure to detect metastases. Although other abnormal-ities (anorexia/lethargy, ventral mass or ulcer visible, anemia, neutrophilia, hypercalcemia, or azo-temia) were noted, concurrent geriatric diseases often confounded clinical interpretation. Therefore,these findings were not specific for a diagnosis of cancer. Metastases were most common in lymphnode, lung, and liver. Mammary cancer was most common in animals that had been implanted withmelengestrol acetate (MGA)-impregnated silastic devices (n = 29, 94%). This association wasstatistically significant (P < 0.002). There was no significant difference in age at diagnosis ofmammary cancer in MGA-treated felids (x ± SD = 13.87 ± 2.60 yr) versus non treated felids(16.33 ± 0.94 yr). Mammary cancer behaves as aggressively in wild felids as in domestic cats, andlong-term exogenous progestin exposure is a risk factor in its development. Zoos using MGA infelids should be cognizant of the strong association between progestin exposure and mammarycancer development and should employ alternative contraceptive methods whenever possible.

Key words: Felids, mammary cancer, melengestrol acetate, contraceptives.

INTRODUCTION

Mammary gland neoplasia in captivewild felids has been reported with increasedfrequency and concern in recent years, andmost. of these cases have been malig-nant. 1.3.4.8.18 Disease surveillance by the Con-traceptive Advisory Group of the American

From the Department of Pathology, University ofTennessee College of Veterinary Medicine, P.O. Box1071, Knoxville, Tennessee 37901, USA (Harrenstien,Munson); and the International Union for the Conser-vation of Nature/Species Survival Commission Con-servation Breeding Specialist Group, 12101 JohnnyCake Ridge Road, Apple Valley, Minnesota 55124,USA (Seal). The AZA Mammary Cancer Study Groupincludes Akron Zoological Park (Gary Riggs, D.V.M.),Baltimore Zoo (Michael R. Cranfield, D.V.M.), Bran-dywine Zoo (Lin Klein, V.M.D.), Buffalo ZoologicalGardens (Allan W. Prowten, D.V.M.), Caldwell Zoo(Doyle D. Starnes, Jr., DYM.), Calgary Zoo (VirginiaHoneyman, D.V.M.), Chaffee Zoological Gardens ofFresno (Ronald P. Gentzler, D.V.M.), Cheyenne Moun-tain Zoological Park (Paul P. Calle. V.M.D.), DallasZoo (Bonnie L. Raphael, D.V.M.),

Zoo and Aquarium Association (AZA) hasshown that exposure to exogenous proges-tins is a strong risk factor for developmentof severe endometrial hyperplasia and uter-ine carcinoma in captive wild felids'>" andis a suspected risk factor for developmentof mammary cancer in domestic'v'? andcaptive wild'>" felids.

Exogenous progestins have been used in

Erie Zoological Gardens (Kenneth J. Felix, D.V.M.),Franklin Park Zoo (John L Curtin, D.V.M.), JacksonvilleZoological Gardens (c. Douglas Page, DYM.), Knox-ville Zoological Gardens (Don Gillespie, DYM., PatrickJ. Morris, DYM., Edward C. Ramsay, D.V.M.), Los An-geles Zoo (Cynthia E. Stringfield, D.V.M.), Memphis Zoo(E. Michael Douglass, DYM.), NYZSlThe Wildlife Con-servation Society (Paul P. Calle, DYM.), Reid Park Zoo(Thomas O. Miller, D.V.M.), Riverbanks Zoological Park(Barbara T Baker. D.V.M.), Sacramento Zoo (NadineLamberski, D.V.M.), S1. Louis Zoological Park (RandallE. Junge, DYM.), Sunset Zoo (James W. Carpenter,M.S., DYM.), and Toledo Zoo (Timothy Reichard,D.V.M.). Present address (Harrenstien): Veterinary Med-ical Teaching Hospital, School of Veterinary Medicine,University of California. Davis, California 95616, USA

468

HARRENSTIEN ET AL.-MAMMARY CA CER IN CAPTIVE WILD FEUDS 469

captive felids for contraception, dermato-logic treatment, and behavior modification.Melengestrol acetate (MGA) is the mostcommonly used progestin contraceptive inthese felids, although megestrol acetate(MA, Ovaban=, Schering-Plough AnimalHealth, Union, New Jersey 07033, USA)and medroxyprogesterone acetate (MPA,Provera'", Pharmacia and Upjohn, Kalama-zoo, Michigan 49001, USA) have also beenused. Silicone implants, impregnated withMGA and placed s.c. or i.m., have been themost common devices used for contracep-tion in captive felids over the past 20 yrbecause of MGA's high potency and effi-cacy, ease of administration (implant place-ment once every 2-3 yr), reversible contra-ceptive effect (upon removal or depletion ofthe implant), minimal diabetogenic or ad-renosuppressive effects, and minimal be-havioral effects.v-"

The purposes of this study are to describethe clinical behavior of mammary cancer incaptive wild felids, to determine the relativevalues of diagnostic procedures and clinicalfindings in the confirmation of malignancy,and to determine whether MGA exposure,species, or nulliparity are significantly as-sociated with occurrence of mammary can-cer in captive wild felids.

MATERIALS AND METHODS

All U.S. facilities housing wild felidswere solicited for historical information re-garding occurrence of mammary neoplasiain their felid collections. Responses werecollated with cases identified through theAZA Contraceptive Advisory Group's dis-ease surveillance database. Fifty-three fe-lids with mammary neoplasia were identi-fied. Full clinical records were available for31 of the animals, and these were the basisof subsequent analysis (Table 1).

Histopathologic examination of mam-mary tissue from all 31 animals was per-formed by one of us (L.M.) for tumor clas-sification (adenoma vs. carcinoma). All ofthe mammary masses were carcinomas, andspecific histologic patterns were seen in

Table 1. Captive wild felids affected by mammarycancer (n = 31).

Species

Dum-tion

Age at Average oftime of dosage MGAdiag- of expo-nosis MGN sure(yr) (mglkg) (mo)

10.0 5.58 11512.5 6.19 8516.2 8.33 10515.8 4.46 9312.3 5.39 7213.0 6.09 4814.9 5.17 8010.8 6.30 788.2 5.82 75

20.0 4.12 6015.0 5.03 5612.0 4.83 3611.7 5.23 11814.5 7.83 1215.4 unk" 79I\.6 6.54 5012.7 5.94 7015.716.4 12.49 11511.9 12.37 10818.1 6.03 3613.4 10.47 10617.0 24.00 3615.7 12.63 2615.0 4.56 9314.6 6.22 14915.4 9.42 8017.014.3 11.56 12810.0 17.92 5114.2 30.32 26

Tiger (Panthera tigris)

Jaguar tPanthera onca)

Lion (Panthera lea)

Cougar (Felis concolor)

Leopard (Panthera pardus)Jungle cat (Felis chaus)

, MGA = melengestrol acetate.bunk = unknown.

combination in these tumors. Further sub-classification and correlation of pattern withclinical behavior was therefore not attempt-ed.

The following factors were examined aspotential risk factors for the development ofmammary cancer: species (n = 31), parityhistory (n = 24), contraceptive history (n= 31), and duration of treatment and aver-age dose of MGA (n = 29). Informationwas also collected regarding clinical signs

470 JOURNAL OF ZOO AND WILDLIFE MEDICINE

Table 2. Species distribution of captive wild felids affected by mammary cancer.

Species

No. (%)affected felids-

in database

No. (%) MGA-treatedfelids>

in database

No. felids withconfirmed mammary

carcinoma(% by species)

Tiger (Panthera tigris)Lion (Panthera lea)Jaguar (Panthera onca)Leopard (Panthera pardus)Cougar (Felix concolor)Jungle cat (Felis chaus)

48 (28%)21 (12%)16 (9%)17 (10%)9 (5%)3 (2%)

32 (36%)18 (20%)'13 (15%)'9 (10%)4 (5%)2 (2%)

17 (35%)'2 (10%)7 (44%)'1 (6%)2 (22%)2 (67%)

, Age-matched felids from the American Zoo and Aquarium Association Contraceptive Advisory Group's disease surveillancedatabase. Percentage is of total felids in the database (/1 = 170).

b MGA = meiengestroi acetate. Percentage is of all contracepted felids in database (n = 89).'P < 0.05.

and their duration before diagnosis (n =

31), age of patient at time of diagnosis (n= 31), thoracic radiographic findings (n =

10), thoracic ultrasonographic findings (n =

1), abdominal radiographic findings (n =

3), abdominal ultrasonographic findings (n= 4), hematologic findings from blood col-lected at the time of initial mammary nod-ule palpation (n = 25), and gross necropsyand histopathologic findings (n = 12 com-prehensive necropsies, n = 11 partial nec-ropsies). Two felids were exposed to dosesof MGA that were approximately doublethose of the rest of the study population, sothese high values were not included in sta-tistical analysis of MGA dose. Results ofradiographic and ultrasonographic exami-nations were evaluated if they were per-formed ::0;14days before necropsy, becausenecropsy findings were not expected to cor-relate with earlier radiographic and ultra-sonographic exams (30-120 days beforenecropsy).

To determine if species was a risk factor,cancer-affected felids were compared witha cohort of age-matched unaffected felids(of all species, regardless of their contra-ceptive history) from the AZA Contracep-tive Advisory Group's disease surveillancedatabase (n = 139). To determine if param-eters of MGA exposure (MGA dose or du-ration of exposure) were risk factors, thedisease surveillance database was used toidentify a cohort of felids without cancer

that were matched for species and age, re-gardless of their contraceptive history (n =

76).Continuous numerical data were statisti-

cally compared using Student's unpairedtwo-tailed t-test at a significance level of P< 0.05. Frequency data were statisticallycompared using Fisher's exact test at a sig-nificance level of P < 0.05.

RESULTSComparisons between mammary cancergroup and age-matched cohort

The species distribution of felids affectedby mammary cancer (n = 31) is similar tothat of all nonaffected felids in the diseasesurveillance database (n = 139), with theexception of tigers and jaguars, which havea higher occurrence of mammary cancer (P< 0.001 and P < 0.01, respectively) (Table2). In this database, MGA implant usagehas tended to be more common in jaguars(P < 0.05) and lions (P < 0.005) than inall other felids for which contraceptive his-tory is known (n = 164).

Comparisons between mammary cancergroup and age/species-matched cohort

Mammary cancer occurred more fre-quently in animals that had been exposedto MGA than in those that had not beenexposed (P < 0.002). Twenty-nine (94%)of the 31 cancer-affected animals in thisstudy had been given MGA (Table 1),

HARRENSTIEN ET AL.-MAMMARY CANCER IN CAPTIVE WILD FEUDS 471

Table 3. Sensitivity and specificity of antemortem diagnostic procedures in 31 captive wild felids withmammary gland cancer. Diagnostic procedures were performed 0-14 days before necropsy. For this study, thestandard for detection of masses was defined as necropsy results for corresponding anatomical sites.

No. true No. truepositives! negatives!

No. felids no. false no. falseDiagnostic procedure tested positives negatives Sensitivity Specificity Test accuracy

Thoracic radiography 10 6/0 3/1 86% 100% 90%Thoracic ultrasonography 1 110 0/0 100% 100%Abdominal radiography 3 0/0 0/3Abdominal ultrasonography 4 2/0 0/2 50% 50%

whereas 48 (63%) of the 76 felids in thenonaffected cohort had been given MGA.MGA exposure was not significantly longerin cancer-affected felids (x ± SD = 74.24± 33.85 mo, n = 29) than in those felidswithout mammary cancer (58.56 ± 37.2mo, n = 48) (P < 0.08). Mean dosages ofMGA for cancer-affected felids (7.53 ±3.34 mg/kg, n = 27) and nonaffected felids(8.39 ± 4.58 mg/kg, n = 38) also did notdiffer significantly (P = 0.41).

Comparisons within mammary cancer group

For those felids with mammary cancer,age at time of diagnosis did not differ sig-nificantly between MGA-treated and non-treated felids (Table 1). Overall, the meanage at time of diagnosis for nulliparous an-imals (13.05 ± 1.87 yr, n = 9) did not differsignificantly from that for parous animals(14.07 ± 2.29 yr, n = 15) (P = 0.27). Spe-cifically, the mean age at time of diagnosisdid not differ significantly between thosefelids that had experienced pregnancy onlybefore their first MGA exposure (14.38 ±2.36 yr, n = 8) and those felids that hadbeen pregnant only after MGA exposure(11.17 ± 0.59 yr, n = 2) (P = 0.10), al-though those felids that were pregnant be-fore MGA treatment tended to be olderwhen cancer was first detected.

Antemortem clinical presentation

Zookeepers and veterinarians noted non-specific clinical signs in the felids of thisstudy. Signs of disease included anorexia!lethargy, vomiting, weight loss, diarrhea,

and abdominal distension (58%, 23%, 19%,6%, and 6% of animals respectively). Themore specific signs of mammary skin ul-ceration, respiratory difficulty due to pul-monary metastasis, or rear leg swelling dueto lymphatic obstruction were only seen ina small proportion of the cases (29%, 12%,and 6% of animals, respectively).

Data regarding specific sites and sizes ofmammary masses were not available forthis study.

Procedures to diagnose and detect metas-tasis of mammary cancer were initiated ata mean (±SD) of 9.28 (± 14.26) days (n =

29) after onset of signs. Antemortem diag-nostic techniques included physical exami-nation, thoracic and abdominal radiographyand ultrasonography, abdominal magneticresonance imaging, and hematologic test-ing. Definitive diagnosis was accomplishedby biopsy or aspiration of a palpable mam-mary mass or a suspected site of metastasisor by necropsy of those animals that had nopremonitory signs.

Thoracic radiography was a sensitive toolfor detection of pulmonary metastasis, andall radiographic lesions that were sugges-tive of cancer were confirmed at necropsy(high specificity) (Table 3). Thoracic radi-ographs were obtained for 10 animals, andradiographic abnormalities (bronchiolar ornodular patterns) were noted in six of thesefelids. Thoracic ultrasonography was per-formed in one animal, and two metastaticnodules were identified in the pulmonaryparenchyma. Abdominal radiography wasan insensitive test for detection of visceral

472 JOURNAL OF ZOO AND WILDLIFE MEDICINE

metastasis; abdominal metastases in threeanimals were not detected by this method(Table 3). Abdominal ultrasonography wasmore sensitive than radiography for detect-ing metastasis to liver, kidney, or spleen.However, two other animals had metastasisto several abdominal organs (liver, adrenalgland, ovary, and lymph node) that was notdetected by ultrasonography but was con-firmed by necropsy.

Seventeen (68%) of the 25 animals inwhich blood work was performed had oneor more hematologic abnormalities. Hyper-calcemia was noted in seven (26%) ani-mals, including two with concurrent renallesions, one with metastasis-related renal le-sions, and three without apparent renal le-sions (Table 4). Anemia was noted in seven(26%) animals (Table 5); four of these ane-mias were normocytic in nature, one wasmicrocytic, and one was macrocytic. Neu-trophilia was diagnosed in four animals thatdid not have evidence of inflammatory dis-ease other than mammary cancer; the neu-trophil count in these animals was 30,695(±4,575) neutrophils/ul, and the range ofimmature neutrophils was 0-4,788 bands/fLl.

Renal values (blood urea nitrogen or cre-atinine) were elevated in two animals be-cause of histologically confirmed renaldiseases common in aged felids (cortical at-rophy, glomerulonephritis, interstitial ne-phritis with ischemic necrosis, or pyelo-nephritis). In two animals, peritoneal car-cinomatosis was so severe as to cause ure-teral stenosis, hydroureter, hydronephrosis,and resultant azotemia. One animal with re-nal metastasis and no other histologicallyconfirmed renal lesions had a creatinineconcentration of 4.9 mg/dl.

Hepatic metastases were also associatedinfrequently with abnormal hepatic values(aspartate aminotransferase, alanine ami-notransferase, bilirubin, or bile acids). Nineanimals showed hepatic metastasis (withoutother hepatic lesions) at necropsy, but onlytwo of these animals (out of four fromwhich blood was collected) showed eleva-

vi"Vl

"..c:C1:!coc,.s"0soc:1:!coVl

"'u"c,Vl

..c:al"

.:Vl0.0c::.ac:

1.1:(.)

'5iJo]OJc,o(.)

:5"U]c

"OJCIl

Vlo "c: »

o 0c: c:

·i'':..<::Co

"Co"2"Eo

bO

<aco..,osEB

~oseo'"oU

~"'"~ ~O'\O\O\O\C'"i00r-:"<t'.....•....

~N ~ ~""';C"l..-OOI'~c-t~~~~~~~c~....-;....-;c:~~ CI)

~ N ~ 0 N c: c: 0 "~ ~ ~ ~ ~ c ~

vi"<Il

".,sc:

~0..

.5

~c:

~'"<Il

".~0..<Il

g"•...B<Il

OJ>B.5

"()c:~~"0::...;8t;()

"0c:'"On>.•..'"EE'"E

HARRENSTIEN ET AL.-MAMMARY CANCER IN CAPTIVE WILD FEUDS 473

~ ~•..... ~ 001J:ir-:.,.;'<!"'<!"trlI •..... IV) 100--. __

~t;~~oo~NOO~v)-"'"'-" "-'" '-"'-" '-""'-"q..>::'0O\,,"!'0 ~

~ §~oo:g'Dg]""" g g 3

]0.o.s

o•....•

tions of serum liver enzymes. Of eight an-imals with abnormal hepatic values, four(50%) had hepatic lesions or hepatic me-tastases confirmed by necropsy. Alkalinephosphatase was not increased in any felidwith mammary cancer.

Animals survived for a variable period oftime after diagnosis. Twenty-eight cases(90%) were diagnosed antemortem, and theremaining three (l 0%) were diagnosed atnecropsy. Of the 27 felids that were eu-thanized, 14 were euthanized because ofrapidly deteriorating condition at 35.62(::!::49.30)days after diagnosis. Twelve wereeuthanized for elective reasons at 61.58(::!:: 171.44) days after diagnosis. Four ani-mals died spontaneously, at 0, 1, 2, and 390days, respectively, after diagnosis.

Necropsy examination

All mammary gland tumors evaluated for~ this study were classified as carcinomas.~ Further histopathologic characterization of.:OJ these tumors included combinations of cri-c.~ biform, solid, cyst-papillary, and adenocar-~ cinoma patterns. Twenty-eight felids (90%)c" had evidence of metastasis at necropsy. The

.12 distribution of metastases in the 12 animals:;;~ with comprehensive necropsies is listed in·ue Table 6. The distribution of metastases in.~ animals with limited necropsies (n = 11)~E was similar to that of animals with compre-E hensive necropsies.vi:e

DISCUSSION..a0.

"§ Exposure of captive wild felids to MGA~0. greatly increases their risk of developing

mammary cancer. Regular (continuous)treatment of domestic cats with MPA orMA has been associated with increasedprevalence of mammary carcinoma, but ir-regular (infrequent) administration hasnot.'? In our study, no relationship could befound between duration of MGA exposureand incidence of cancer. However, given thelong (2::3 yr) duration of contraception af-forded by MGA in its currently availableform, presumably all MGA treated captivewild felids have been subject to progestin

('t') OI)~ -: ~ l.f"!"';,,'<!"oot-t-N'<!"C:N '<!" '<!"

<Il <Il

" ">.>.

r-::&t'!t; r--: r-: r.n 0\trl C:'<!"t-O\\OON 0" 0;-ON V) \0 C:>' ('f)

<Il

o "c: >.

,-;.

§oc

474 JOURNAL OF ZOO AND WILDLIFE MEDICINE

Table 6. Sites of metastasis noted in comprehen-sive necropsies of 12 captive wild felids with mam-mary cancer.

Site No. cases Frequency

Lymph node(s) 9 75%Lung/pleura 8 67%Liver 7 58%Adrenal gland 4 33%Spleen 3 25%Peritoneum 2 17%Heart 2 17%Uterus 2 17%Ovary 2 17%Kidney 1 8%Bone 1 8%Intestine 1 8%Vagina 1 8%Eyelid I 8%No metastasis 1 8%

exposure for a period that should be con-sidered continuous and long term, and thusour findings would be consistent with thoseof domestic cat studies.

No association was found in this studybetween MGA dosage and occurrence ofmammary cancer, possibly because of thebroad dosage range of MGA in both cancer-affected and unaffected captive wild felids.Studies have shown that concurrent proges-tin and estrogen exposure increases the riskof mammary cancer in women, IO endoge-nous estrogens enhance mammary tumori-genesis in dogs and cats," and suppressionof ovarian activity confers a protective ef-fect in dogs. II In captive wild felids, follic-ular activity is often not suppressed by cur-rently used dosages of MGA (unpubl. data),thus a concurrent estrogen/progestin hor-monal environment is common in MGA-exposed felids and could be an importantmechanism for mammary carcinogenesis inthese animals.

The increased occurrence of mammarycancer in tigers and jaguars may be due pre-dominantly to the increased exposure ofthese species to MGA. Jaguars also appearto have a high prevalence of ovarian!" anduterine" cancer. It is unclear why lions(with their significantly higher prevalence

of MGA exposure) do not have a high oc-currence of mammary cancer.

Neither pre-MGA nor post-MGA preg-nancy(ies) were associated with a delay inonset of mammary cancer. This findingcontrasts with those in human studies, inwhich pregnancy early in life has beenfound to confer a protective benefit againstmammary cancer because of early differ-entiation of stem cells that would otherwisebe susceptible to neoplastic transforma-tion.> Only two cancer-affected felids hadbeen pregnant after MGA exposure; thislow number is believed to reflect generalreproductive management practices in zoosand should not be interpreted as supportingtheories regarding possible protective ben-efits of pregnancy after MGA exposure.

The clinical signs of disease in this studywere not specific for cancer. All cancer-af-fected felids had palpable mammary nod-ules, but it is unknown how many felidswith nodules from mammary hyperplasia orother nonneoplastic causes were excludedfrom this study. The most accurate ante-mortem procedures for diagnosis of mam-mary cancer were thoracic radiography andevaluation of fine-needle aspirates or biopsysamples. Other procedures that were helpfulin increasing the index of suspicion for can-cer, although they were less commonly as-sociated with positive findings, were tho-racic ultrasonography, abdominal ultraso-nography, serum chemistry analysis (to de-tect hypercalcemia), and hematologicanalysis (to detect anemia or neutrophilia).

Hypercalcemia has been associated withmammary cancer in a variety of species, butthis association has never been documentedin felids. Moderate hypercalcemia may besecondary to primary renal disease, but themarked hypercalcemia (without hyperphos-phatemia) seen in five felids of this studyis suspected to be due to other mechanismssuch as secretion of tumor-related humoralfactors (e.g., parathyroid hormone-relatedprotein [PTHrP]) or metastasis-inducedbone resorption." Bone metastasis was con-firmed in one felid of this study; the lack

HARRENSTIEN ET AL.-MAMMARY CANCER IN CAPTIVE WILD FEUDS 475

-'

of routine necropsy examination of boneprecludes discussion of bone involvementin the other cases. In immunohistochemicalstudies (now in progress), we are trying todemonstrate higher levels of PTHrP inmammary carcinoma tissues of hypercal-cemic felids of this study.

Anemia was noted in seven mammarycancer-affected captive wild felids. Anemiais the most common hematologic disorderof human cancer patients" and may haveseveral causes, including chronic inflam-matory disease (a normocytic anemia),blood loss (microcytic), hematopoietic dys-plasia (macrocytic), and (in cats) concurrentfeline leukemia virus infection (normocyticor macrocytic)." Any of these factors couldhave played a role in development of ane-mia in the felids of our study, althoughnone of our cases were known to be leu-kemia test positive. One leopard in thisstudy had undergone chemotherapy (adria-mycin and cyclophosphamide) and was pre-sumed to be in renal failure, thus the che-motherapy or decreased renal production oferythropoietin could have contributed to heranemia.

Neutrophilia has been associated withseveral types of widely metastatic cancer,especially those with pulmonary metasta-ses.' This association may be due to tumor-induced release of cytokines (e.g., colonystimulating factors) or to a nonspecific hostresponse to tumor growth. All four of theneutrophilic felids of this study had wide-spread metastases (including to the lung)evident at necropsy.

It is difficult to draw specific prognosticconclusions from the diagnosis-to-deathtime intervals of this study. Zoo staff mem-bers have become aware of the aggressivenature of mammary cancer in domestic catsand captive wild felids (which has beenconfirmed by this study) and therefore oftenchoose euthanasia soon after a diagnosis ofmammary cancer is made to minimize du-ration of animal suffering. The cancer casesin this study demonstrated widespread me-tastasis at necropsy. Once clinical signs of

disease were noted, mammary cancer hadlikely already metastasized and would havebeen difficult or impossible to eradicate.>Histopathologic examination of mammarynodules is necessary for confirmation ofmalignancy (vs. hyperplasia, which is com-monly reported, or adenoma development,which is rare in felids) so that appropriatemedical management decisions can bemade.

CONCLUSIONS

All 31 mammary neoplasms evaluated inthis study were malignant, 94% of thesecancer-affected felids had been contracept-ed with MGA. In the felids of this study,thoracic radiography was the most usefulantemortem procedure to detect metastasis.Mammary cancer should be considered asa differential diagnosis in hypercalcemic oranemic captive wild felids. Short durationof clinical signs and widespread metastasisof mammary cancer in these animals areindicative of the aggressive nature of mam-mary carcinoma in zoo felids.

Acknowledgments: This project was support-ed by funds from the New York Zoological So-ciety Nixon Griffis Fund for Zoological Re-search and from the American Zoo and Aquar-ium Association Conservation EndowmentFund/Ralston-Purina Big Cat Survival Fund. Wethank Dr. Joyce Merryman for her help in manu-script preparation and histotechnologists BrendaCollins and Ron Moulton of the University ofTennessee Histopathology Laboratory for theirtechnical assistance.

LITERA TURE CITED

1. Chandra, S., and D. C. Laughlin. 1975. Virus-like particles in cystic mammary adenoma of a snowleopard. Cancer Res. 35: 3069-3074.

2. Chinn, D. R., R. K. Myers, and J. A. Matthews.1985. Neutrophilic leukocytosis associated with met-astatic fibrosarcoma in a dog. J. Am. Vet. Med. Assoc.186: 806-809.

3. Frazier, K. S., M. E. Hines, C. Ruiz, A. J. Herron,and N. H. Altman. 1994. Immunohistochemical dif-ferentiation of multiple metastatic neoplasia in a jaguar(Panthera onca). J. Zoo Wild!. Med. 25: 286-293.

4. Gillette, D. M., H. M. Acland, and L. Klein.

476 JOURNAL OF ZOO AND WILDLIFE MEDICINE

1978. Ductu1ar mammary carcinoma in a lioness. J.Am. Vet. Med. Assoc. 173: 1099-1102.

5. Hahn, K. A, L. Bravo, and J. S. Avenell. 1994.Feline breast carcinoma as a pathologic and therapeu-tic model for human breast cancer. In Vivo 8: 825-828.

6. Hayes, H. M., Jr., K. L. Milne, and C. P. Mandell.1981. Epidemiological features of feline mammarycarcinomas. Vet. Rec. 108: 476-479.

7. Hernandez, J. F., B. B. Fernandez, M. Chertach,and P. A. Gage. 1975. Feline mammary carcinomasand progestogens. Feline Pract. 5: 45-48.

8. Hruban, Z., W E. Carter, T Meehan, P. Wolff,W. A Franklin, and S. Glagov. 1988. Complex mam-mary carcinoma in a tiger (Panthera tigris). J. ZooAnim. Med. 19: 226-230.

9. Madewell, B. R, and B. F. Feldman. 1980.Characterization of anemias associated with neoplasiain small animals. J. Am. Vet. Med. Assoc. 176: 419-425.

10. Miller, D. R, L. Rosenberg, D. W. Kaufmann,P. Stolley, M. E. Washauer, and S. Shapiro. 1989.Breast cancer before age 45 and oral contraceptives.Am. J. Epidemiol. 129: 269-280.

11. Misdorp, W. 1991. Progestagens and mammarytumors in dogs and cats. Acta Endocrinol. 125: 27-3).

12. Misdorp, W, A. Romijn, and A. A. M. Hart.1991. Feline mammary tumors: a case-control studyof hormonal factors. Anticancer Res. 11: 1793-1798.

13. Munson, L. 1993. Adverse effects of contra-ceptives in carnivores, primates, and ungulates. Proc.Am. Assoc. Zoo. Vet. 1993: 284-289. (Abstr.)

14. Munson, L. 1994. A high prevalence of ovarianpapillary cystadenocarcinomas in jaguars. Vet. Pathol.31: 604. (Abstr.)

15. Munson, L., and R. J. Mason. 1991. Patholog-ical findings in the uteri of progestogen-implanted ex-otic felids. Proc. Am. Assoc. Zoo. Vet. 1991: 311-312.(Abstr.)

16. Munson, L., J. E. Stokes, and L. A. Harrenstien.1995. Uterine cancer in zoo felids on progestin con-traceptives. Vet. Pathol. 32: 578. (Abstr.)

17. Randel, R. D., C. J. Callahan, R E. Erb, H. A.Garverick, and B. L. Brown. 1972. Effect of melen-gestrol acetate on plasma progesterone, luteinizinghormone and total corticoids in dairy heifers. Am. Soc.Anim. Sci. 35: 389-397.

18. Raphael, B. L., S. L. Huntress, and T G. Curro.1990. Reproductive disorders associated with proges-togen implants in a group of exotic felids. Proc. Am.Assoc. Zoo. Vet. 1990: 282-283. (Abstr.)

19. Rosol, T J., and C. C. Capen. 1992. Mecha-nisms of cancer-induced hypercalcemia. Lab. Invest.67: 680-702.

20. Rutteman, G. R., and W. Misdorp. 1993. Hor-monal background of canine and feline mammary tum-ours. J. Reprod. Fertil. Suppl. 47: 483-487.

21. Seal, U. S., R Barton, L. Mather, K. Olberding,E. D. Plotka, and C. W. Gray. 1976. Hormonal con-traception in captive female lions. J. Zoo Anim. Med.7: 12-20.

22. Thomas, D. B., and E. A. Noonan. 1991.Breast cancer and depot-medroxyprogesterone acetate:a multi-national study (The WHO Collaborative Studyof Neoplasia and Steroid Contraceptives). Lancet 338:833-838.

23. Weller, R E. 1982. Paraneoplastic disorders incompanion animals. Compend. Cont. Ed. Pract. Vet. 4:423-428.

Received for publication 3 July 1995.