The Causes of Canine Myocarditis and Myocardial Fibrosis Are Elusive by Targeted Molecular Testing: Retrospective Analysis and Literature Review

Abstract

Myocarditis can cause death or permanent heart damage. As epidemiologic and etiopathologic data for canine myocarditis are lacking, we performed a retrospective study using nucleic acid extracted from archived (2007 to 2015) tissues from myocarditis cases and control dogs without myocardial lesions. Heart tissue from pediatric/juvenile and adult dogs was tested with a comprehensive panel of conventional and real-time polymerase chain reaction (PCR) assays targeting recognized agents of canine myocarditis based on a literature review and informed by the comparative epidemiology of human myocarditis. The PCR screen, which included canine parvovirus 2 (CPV-2), canine distemper virus, canine herpesvirus, Borrelia spp, West Nile virus, adenovirus, parainfluenza virus, pneumovirus, respiratory coronavirus, influenza virus, Bartonella spp, Rickettsia spp, Mycoplasma spp, and Neospora caninum, did not detect agents in 35 of 66 cases (53%; 95% confidence interval [CI], 41%–65%) and was frequently negative in adults (21/26); by comparison, agents were not detected in 27 of 57 controls (47%; 95% CI, 35%–60%). Canine distemper virus, herpesvirus, adenovirus, coronavirus, parainfluenza virus, Mycoplasma haemocanis, and N. caninum were occasionally detected in both cases and controls; thus, PCR detection was not considered to indicate causation. We previously reported that CPV-2 continues to be associated with myocarditis in young dogs despite widespread vaccination; in adults, CPV-2 was detected in 2 of 26 cases and 4 of 22 controls. As several agents were similarly detected in cases and controls, it is unclear if these are cardiopathogenic, incidental, or latent. West Nile virus was detected at the analytic limit in 1 adult case. We did not detect Borrelia spp, Bartonella spp, Rickettsia spp, or influenza A virus in the myocarditis cases. These data demonstrate the limitations of current targeted diagnostic tests and the need for additional research to identify unknown agents and develop testing strategies for canine myocarditis.

Although myocarditis is a devastating disease resulting in permanent cardiac damage or death, the cause is not known for most veterinary cases. Systemic viral, bacterial, or protozoal infections are implicated, but in practice, the underlying cause is often not identified as there are few confirmatory clinical tests for canine myocarditis. Diagnosis is most frequently made at postmortem, and published investigations are largely limited to necropsy case reports and case series. Screening for all of the reported causes of canine myocarditis in individual clinical cases is cost prohibitive and often unrewarding, and thus targeted testing strategies are needed.

Myocarditis is characterized by nonspecific clinical signs, including acute weakness, syncope, exercise intolerance, respiratory difficulties, and/or abdominal distension; arrhythmias or left- or right-sided congestive heart failure are potentially fatal complications.⁹¹ A presumptive diagnosis is based on a combination of patient history, clinical signs, disease course, physical examination findings, echocardiographic and electrocardiographic abnormalities, and elevated circulating cardiac troponin I (cTnI) levels, which is unfortunately not specific to myocarditis.^20,43,49 Treatment is largely supportive.

The substantial deficiencies in clinical diagnosis and treatment of canine myocarditis are related to a paucity of epidemiologic data. In humans, viral infections are associated with the majority (69%) of myocarditis cases.⁴⁰ In contrast, the prevalence of viral myocarditis in the dog is undetermined. Autopsy series have been invaluable in understanding the epidemiology of human myocarditis and have informed specific molecular testing recommendations that have improved myocarditis diagnosis.^{13,22,23,40,67,72,97,108}

In the dog, there have been limited molecular investigations of myocarditis. Maxson et al⁸¹ tested for a small number of viral agents (canine parvovirus 2 [CPV-2], canine adenovirus [CAV], canine herpesvirus [CHV]) not detected in their small cohort of canine myocarditis cases (n = 9). Our previous retrospective study demonstrated a significant association of CPV-2 with myocarditis/myocardial fibrosis in young dogs 2 years of age or less; however, the cause of myocarditis remained unknown in the majority of cases, and the role of CPV-2 in adult myocarditis was not investigated.⁴⁴ As there have been few advances in our ability to diagnose and treat myocarditis in companion animals, more information is needed to elucidate the underlying causes.

The goal of this study was to direct testing strategies, improve disease diagnosis, and contribute to the understanding of the etiopathogenesis of this devastating disease. We first performed a literature review for primary reports of natural causes of canine myocarditis. Based on this review and the human literature, we hypothesized that viral infections are frequent causes of canine myocarditis, with different agents affecting pediatric/juvenile and adult dogs. We then performed a retrospective case-control study to test this hypothesis on archived tissue from a cohort of dogs representing different age classes. Myocardial samples were tested with a comprehensive panel of validated polymerase chain reaction (PCR) assays to screen for possible causes of canine myocarditis in our region (northeastern United States), encompassing reported causes of canine (CPV-2, canine distemper virus [CDV], CHV, Borrelia burgdorferi, West Nile virus [WNV], Bartonella spp, Rickettsia spp, and Neospora caninum) and human myocarditis (adenovirus, herpesviruses, and influenza virus), as well other agents common in the canine population.⁴⁶

Materials and Methods

Literature Review

A literature review was performed using the Medline Database from 1970 to April 2017 searching for primary reports of natural causes of canine myocarditis using the search terms canine and dog, in conjunction with myocarditis and myocardial inflammation. For causes of myocarditis cited in textbooks, we attempted to identify a primary resource.^24,80 Reports were excluded if they described experimental induction of disease, described an unconfirmed/unknown agent (ie, not identified by culture, molecular methods, other ancillary testing, or specific histologic characteristics), lacked description of pathologic features, did not identify the causal agent in the heart, or were learning exercises (ie, “Pathology in Practice” manuscripts). Reports describing endocarditis, pericarditis, or epicarditis without extension into the myocardium were excluded. The search was focused on causes, demographic information, and pathologic features.

Retrospective Case Selection

Formalin-fixed, paraffin-embedded (FFPE) tissue from the Cornell Animal Health Diagnostic Center (AHDC) pathology archives was used to evaluate dogs with myocarditis and/or myocardial fibrosis for infectious agents. A search of archived cases from June 2007 to November 2015 was performed to identify adult dogs 2 or more years of age of any breed or sex using similar diagnostic terms consisting of myocardial fibrosis, myocarditis, cardiac fibrosis, chronic myocarditis, fibrosing cardiomyopathy, cardiac failure, restrictive cardiomyopathy, or end-stage heart. Controls were identified from a similar age distribution and time period using the aforementioned keywords as exclusion criteria to identify adult dogs with ostensibly healthy hearts. Our previously described cohort of pediatric/juvenile cases and controls that were 2 years of age or less was also tested.⁴⁴

Histopathology was evaluated by a veterinary pathologist (K.K.) blinded to disease status. Cardiac tissue was scored histologically as previously described to assess the severity of necrosis, inflammation, and fibrosis on a scale of 0 to 3 (0 = absent, 1 = mild [up to 10%], 2 = moderate [between 10% and 25%], 3 = severe [over 25%]).⁴⁴ Myocardial fibrosis was evaluated by Masson trichrome stain. Gram stain, Gomori-methenamine silver, periodic–acid Schiff, and/or Warthin-Starry histochemical stains were performed on a subset of cases to assist in detection of organisms. Histochemical stains were performed by the AHDC Histopathology Laboratory.

Nucleic Acid Extraction

Nucleic acid was extracted from two 10-mm sections of archived FFPE cardiac tissue sectioned into sterile Eppendorf tubes. The microtome blade was changed between samples and the microtome face was cleaned with 70% ethanol to minimize cross-contamination. Samples were incubated overnight in lysis buffer with proteinase K at 60°C. After incubation, the tubes were briefly spun down and allowed to cool to room temperature. Lysates (200 μL) were transferred to a 96-well plate for extraction using the Biomek 4000 automated workstation (Beckman Coulter, Indianapolis, IN) Total Nucleic Acid preset program with the Agencourt Formapure Kit (Beckman Coulter) according to manufacturer instructions. Nucleic acid was quantified with a Nanodrop ND-1000 spectrophotometer (Thermo-Fisher, Waltham, MA); the average nucleic acid concentration was 336 ng/μl.

PCR

All PCR experiments were run with appropriate controls, including a no-template reaction as well as a confirmed positive for each pathogen. The 95% confidence intervals (CIs) of total proportions with PCR detection were calculated using the modified Wald method and GraphPad QuickCalcs (GraphPad Software, La Jolla, CA).

Conventional PCR

Conventional PCR was performed using housekeeping gene RPS19 primers and CPV-2 VP1-VP2 primers as previously described.^15,44,84 Conventional PCR for Neospora canis was performed as previously described with the following alterations to PCR conditions: denature (94°C; 1 minute), annealing (50°C; 1 minute), and primer extension (74°C; 3.5 minutes) repeated for 40 cycles.¹³⁰ PCR reaction mixtures were separated on a 1.5% agarose gel (Thermo Scientific, Waltham, MA) and run at 125 V for 45 minutes. Amplicon sequencing was performed by the Cornell University Biotechnology Resource Center and analyzed using a BLAST search of GenBank.

Real-Time PCR

Samples were diluted to 50 ng of DNA/ml. All real-time PCR (rtPCR) testing was performed semiquantitatively with platform-specific cycle threshold (Ct) values interpreted according to interquartile ranges based on serial dilutions of positive controls across the dynamic range of the assay. The Ct value is the number of cycles when amplification is detected and is inversely proportional to the amount of target nucleic acid in the sample.

rtPCR testing for WNV was performed at the microliter scale on the StepOne platform (Applied Biosystems, Foster City, CA).³⁸ To avoid false negatives, WNV PCR was not performed on cases/controls submitted during low-risk periods for WNV exposure (ie, 3 weeks after the first hard frost until an ambient temperature supporting mosquito activity). rtPCR was performed using primers and probe to amplify the CPV-2 VP2 gene as previously described.⁴⁴ rtPCR for CHV-1 was performed on the 7500-FAST platform (Applied Biosystems) using published primer/probes with the following PCR conditions: 10 minutes at 48°C, then 3 minutes at 95°C, followed by 40 cycles of 15 seconds at 95°C, then 60 seconds at 60°C.⁷⁰ Reactions were carried out in a 25-μl volume (4 μl of nucleic acid per reaction) using PathID multiplex 1-step real-time PCR master mix (Applied Biosystems). rtPCR for Mycoplasma haemocanis was performed as previously described.^8,96

Nanoliter-Scale rtPCR

Reverse transcription and preamplification were performed together using a pool of 18 primer sets (for either respiratory or tick-borne pathogens) and the TaqMan FastVirus 1-step master mix (Applied Biosystems). Preamplified products were loaded onto custom nanoscale PCR plates (OpenArray; Applied Biosystems) printed in a 18 × 3 format.⁴⁶ The pathogens targeted included canine respiratory coronavirus (CRCoV, betacoronavirus), CDV, CAV, canine parainfluenza virus (CPiV), influenza A, Mycoplasma cynos, and Streptococcus zooepidemicus. The CAV assay is optimized for detection of CAV-2. The tick-borne pathogens targeted included Bartonella spp, B. burgdorferi, Borrelia mayonii, Borrelia miyamotoi, and Rickettsia spp. To avoid false negatives, tick-borne pathogen testing was not performed on puppies less than 6 weeks of age. Nanoscale panels were run at the AHDC.

Immunohistochemistry and In Situ Hybridization

Immunohistochemistry (IHC) for CDV (mouse monoclonal gift from James A. Baker Institute of Cornell University, Ithaca, NY), CPV-1 (rabbit polyclonal gift from James A. Baker Institute Carmichael Laboratory Cornell University), adenovirus (MAB8052; Millipore, Temecula, CA), WNV (clone 7H2; VRL, Gaithersburg, MD), and coronavirus (FIPV3-70; Custom Monoclonals International Sacramento, CA), was performed on a subset of cases by the AHDC Histopathology Laboratory. Tissue with confirmed infection and sections incubated without primary antibody served as positive and negative controls. In situ hybridization for CHV-1 and CPV-2 was performed as previously described with appropriate controls.^44,56

Results

Literature Review

According to search criteria, the causes and characteristics of natural, nonexperimental myocarditis in dogs described in 45 reports are summarized in Table 1 (and detailed in Suppl. Table S1). There were 13 reports of viral myocarditis, including WNV,^18,25,74 CPV-2,^{1,11,27,44,50,52,71,88,102,111} and CPV-1/minute virus.⁵¹ There were 15 reports of protozoal myocarditis, including Leishmania infantum,^103,121 N. caninum,^{6,58,68,82,85,90,98} Trypanosoma cruzi (Chagas disease),^{31,64,127,128} Hepatozoon americanum,⁹ and Toxoplasma gondii.⁵³ There were 9 reports of bacterial myocarditis, including Bartonella spp,^14,35,41,109 B. burgdorferi,^33,57,73 Citrobacter,²⁸ Clostridium piliforme,¹³¹ and Staphylococcus spp.⁵⁷ There were 6 reports of fungal myocarditis, including Inonotus tropicalis,¹⁰¹ Coccidioides immitis,^16,100 Aspergillus spp,^21,61 and Blastomyces dermatitidis,¹⁰⁷ and a report of algal myocarditis caused by Prototheca spp.¹¹⁵ Case reports and case series represented the vast majority of reports (43/45). Ten reports described features of myocarditis in dogs of various ages; 30 reports included pediatric/juvenile cases, and 26 reports described features of adult cases. Ten reports described CPV-2 myocarditis in pediatric/juvenile dogs and not in adults. Fungal and algal cases were mostly described in adults.

Table 1.

Summary of Features of Natural Myocarditis Previously Reported in the Dog.

Group	Etiology	Age Group	Region	Gross Cardiac Lesions	Cardiac Histopathology	Ancillary Test(s)	Other Lesions	References
Virus	West Nile virus	P/J,74 A18,25,74	US-Midwest18,74 West25	Fibrinous atrial epicarditis74	Lymphocytes,18,25,74 macrophages,74 neutrophils25 with degenerated or necrotic CM,18,25,74 vasculitis25	IHC;18,74 PCR m18,25 OT;74 VI OT;25 serology18	Polioencephalomyelitis;18,25,74 interstitial nephritis; tubular necrosis18	18, 25, 74
	Canine parvovirus 2	P/J1,11,27,44,50,52,71,88,102,111	Africa–South Africa,11 US111-Northeast,27,44 Canada,52 Australia,102 South Korea,88 Europe-lreland,71 England50	Pale heart/mottling;11,27,52,102 ventricular dilation;11,52,102 cardiac enlargement;27,102 LAT dilation;27 pericardial effusion71,102	CM necrosis, necrotizing1,11,27,44,50,102,111 to lymphoplasmacytic myocarditis;1,11,27,44,50,102,111 reactive fibrosis;11,27,52,102 endocardial fibrosis;11,27 CM basophilic-amphophilic INIB;11,27,44,50,52,71,102,111 mineralization of degenerate CM;1 macrophages/multinucleated giant cells1,27,44	EM: 20-nm IN viral Particles m;27,50,52,71,102 OT;50 IFA m52 OT;50 homogenate inoculum produced PV enteritis;52 IHC;1,42,111 ISH;44,88 SN;50 PCR44	Pulmonary edema/congestion;11,27,52,102 PV enteritis;27,88,50 interstitial pneumonia;27,50,102 hepatic congestion, periacinar vacuolization, necrosis;52,102 leukoencephalomalacia;1 lymphoid depletion50,88	1, 11, 27, 44, 50, 52, 71, 88, 102, 111
	Canine parvovirus 1	P/j	US-South	Tan to pale gray streaks to irregular areas	Mineralized areas of necrosis with giant cells	VI, IFA OT	Enteritis with large INIB in enterocytes; lymphoid depletion	51
Protozoa	Trypansoma cruzi	P/J,127 P/J-A31,64,128	US-South,127 Southwest,64,128 Mexico31	Enlarged RAT127 or heart;64 ventricular pallor;127 yellow, gray, or white spots-zones128	Lymphoplasmacytic, histiocytic/granulomatous pancarditis;127,128 necrotizing myocarditis/CM degeneration/necrosis;31,127 protozoal pseudocysts with amastigotes with parallel kinetoplasts in CM64,127,128	PCR m;31,127 serology;64 EM;128 flagellates via blood culture; parasitemia in inoculated mice128	Pulmonary edema;127,128 acute centrilobular hepatocyte necrosis;127 splenic lymphoid hyperplasia; amastigotes in macrophages;64,127 granulomatous Gl leiomyositis, myositis128; gliosis with amastigotes128	31, 64, 127, 128
	Leishmania infantum	A,121 P/J-A103	Europe-Spain,121 Brazil103	Epicardial hemorrhage, linear pallor, cardiomegaly121	Mononuclear to necrotizing121 or fibrinoid vasculitis;103 nonsuppurative/lymphoplasmacytic to granulomatous myocarditis;103,121 contraction band necrosis;103 fibrosis;103,121 protozoa in macrophages103	Kinetoplast DNA PCR;121 IHC OT121 m;103 serology121	Granulomatous uveitis with intracellular protozoa;121 renal, hepatic vasculitis121	103, 121
	Neospora caninum	A,6,85 P/J6,58,68,82,90,98	US-Northeast,85 Australia,82,98 Europe,6 Canada90 Africa–South Africa,58 Turkey68	Raised, mottled red/white, infarcts;85 irregular, small, firm, pale, raised lesions;85 pericardial fluid;85 pallor/discoloration82,90	Infarcts;85 CM parasitophorous vacuoles with tachyzoites;6,58,82,85,90 focal endocarditis, subendocardial osseous-chondroid metaplasia;85 lymphoplasmacytic myocarditis6,82,85,90 with neutrophils;85 necrosis to necrotizing myocarditis with protozoa;68,90,98 fibrosis, calcification58	Multiplex PCR85; IHC m58,68,82,85,90,98 OT;6,90,98 maternal serology;82,98 rt PCR m98 OT;68,98 EM m90 OT;6 nested PCR/S98	Hepatic chronic passive congestion;85 lymphoplasmacytic inflammation in kidney, skeletal muscle, cerebral white matter85	6, 58, 68, 82, 85, 90, 98
	Hepatozoon canis (americanum)	P/J-A	US- Southwest (TX)	NR	Pyogranulomatous inflammation, various developmental stages; vasculitis with mineralization	Histo; intraleukocytic organism on blood smear	Skeletal muscle cysts, pyogranulomas with merozoites; interstitial pneumonia with vasculitis; pyogranulomatous glomerulonephritis, pancreatitis; periosteal exotoses, pyogranulomas	9
	Toxoplasma gondii	P/j	Brazil	Focal white streak	Necrotizing myocarditis with dystrophic mineralization, intralesional apicomplexan protozoa	Apicomplexan 18S ribosomal DNA PCR/RFLP	Necrotizing hepatitis with INIB (CAdV-1); necrotizing bronchitis/bronchiolitis, interstitial pneumonia with epithelial ICIB (CAdV-2, CDV): GI villous loss with cryptal necrosis (CPV-2c); demyelination with astrocyte INIB (CDV)	53
Bacteria	Bartonella vinsonii	A41,109	Europe-Germany,109 US-West41	NR	Lymphoplasmacytic, histiocytic, neutrophilic myocarditis extending from vegetative endocarditis41,109	ITS PCR/S,41,109 Pap31 PCR/S valve/m41	Aortic valve neutrophilic, histiocytic endocarditis with gram-negative coccobacilli;109 chronic renal disease, neoplasia, diabetes mellitus41	41, 109
	Bartonella vinsonii and α-Proteobacterium	P/J-A	US-South	Valvular vegetative endocarditis	Coronary arteries disrupted by fibrinoid necrosis, purulent exudate; random neutrophils, fibrin, macrophages, debris	ITS PCR/S valve/blood/tissue; serology	NR	14
	Bartonella henselae	A	US-Northeast	Cardiomegaly; pinpoint, coalescing, tan to white nodular foci33	Coalescing, transmural, pyogranulomatous, lymphoplasmacytic pancarditis; regional CM degeneration/necrosis	ITS PCR/S (fresh frozen)	Valvular endocarditis; pyogranulomatous tubulointerstitial nephritis, pneumonia, meningoencephalitis, uveoscleritis, lymphadenitis, tracheitis, cystitis, splenitis, periarteritis	35
	Borrelia burgdorferi	A,57,73 P/J33	Europe- Poland,57 US-Northeast,73 Midwest33	Chamber dilation;57,73 LV hypertrophy, MV degeneration;57 vegetations;73 aortic embolism;57 LV infarct;55 hydropericardium;33,57 coalescing white to tan areas33	Lymphoplasmacytic infiltrate with CM degeneration;57,73 atrial fibrosis;57 conduction system lymphocytic infiltrates;57 granulocytic aoritis with mural thomboembolism, myocarditis;57 infarction;57 pyogranulomatous myocarditis with CM degeneration, necrosis, mineralization, fibroplasia73	IHC33,77 (abnormal cystic forms);57 serology;57 silver + organisms73	Lymphocytic inflammation in the brain;57 lymphoplasmacytic nephritis;33 ileocecal intussusception;33 pulmonary edema;33 centrilobular hepatocyte vacuolation or degeneration/necrosis;33 ascites, hydrothorax33,57	33, 57, 73
	Citrobacter koseri	P/J	Europe-lreland	Cardiomegaly; few raised, firm, ovoid ventricular lesions	Transmural suppurative myocarditis with fibroplasia, mineralization of hyalinized CM	Gram-negative bacilli in and macrophages extracellular; m culture	Thymic depletion; periacinar hepatocyte necrosis; small intestine villous atrophy with loss of enterocytes, crypt abscesses (speculated CPV-2)	28
	Clostridium piliforme	P/J	US-West	Transmural, irregular, coalescing, raised, pale gray areas	Macrophages, neutrophils with lymphocytes; loss/degeneration of CM	Filamentous bacilli with silver stain m OT	Necrotizing hepatitis, enterocolitis, GI leiomyositis; adrenal cortical adenitis; lymphoid depletion	131
Fungus	Inonotus tropicalis	A	US-South	Transmural firm, coalescing, white, ill-defined nodules; thickened myocardium, pericardium; pericardial effusion	Transmural coalescing granulomatous inflammation with epithelioid macrophages, multinucleate giant cells, lymphocytes, plasma cells; 5- to 7-μm-wide septate hyphae with parallel to bulging walls	Histo; RNA gene ITS, D1/D2 regions PCR/S OT	Thoracic, abdominal effusion; pyogranulomatous inflammation, hyphae in lymph node, lung, gastric wall, liver, spleen, bone marrow, brain	101
	Coccidioides immitis	P/J100 A16,100	US-South16,100	Pale yellow elliptical lesions;100 fibrinopurulent pericarditis, epicarditis;100 granulomatous pericarditis16	Fibrinopurulent pericarditis/epicarditis with CM atrophy/destruction;100 granulomas with spherules, granulomatous pericarditis16	Histo;16,100 intradermal Coccidiodin test;100 serum CF100	Lung granulomas, fibrosis;16,100 granulomas/purulent exudate bronchial LN, liver, bones, pericardium, subcutis;16,100 anterior uveokeratitis with spherules;100 catarrhal enteritis100	16, 100
	Aspergillus alabamensis	A	US-Midwest	Coalescing raised, tan foci	Pyogranulomas with moniliform hyphae and short, septate, 4.5- to 5-μm-wide hyphae with parallel walls, laterally branching globose accessory conidia (aleuriospores)	β-Tubulin gene PCR/S	Disseminated pyogranulomas, fungi; segmental necrotizing vasculitis in the kidney, pancreas, stomach	21
	Aspergillus terreus	A	Australia	Ill-defined, pale, circular lesions	Degenerate neutrophils, macrophages with necrotic calcified CM; hyphae with laterally branching aleuriospores	Culture; histo	Granulomas, infarcts in kidney, spleen, skeletal system	61
	Blastomyces dermatitidis	A	US-various	Granulomas	Granulomas with epicardial/pericardial ascending aorta adventitial involvement, stenosis of main pulmonary artery; lymphoplasmacytic myocarditis	Histo (biopsy, aspirates); serology	Granulomas in lung, spleen, kidney, adrenal glands, brain, tongue, eye	107
Algae	Prototheca spp	A	Australia	NR	Prototheca granulomas	Histo; culture	Granulomas in colon, kidney, eyes, liver, brain, pancreas, lymph nodes, muscle	115

Abbreviations: A, adult; CDV, canine distemper virus; CF, complement fixation; CM, cardiomyocytes; CPV-2, canine parvovirus 2; EM, electron microscopy; GI, gastrointestinal; histo, histopathology; IB, viral inclusion body; IC, intracytoplasmic; IFA, immunofluorescence; IHC, immunohistochemistry; ISH, in situ hybridization; ITS, internal transcribed spacer DNA; IN, intranuclear; LAT, left atrium; LN, lymph node; LV, left ventricle; m, myocardium; MV, mitral valve; NR, not reported; OT, other tissue; PCR, polymerase chain reaction; PCR/S, polymerase chain reaction and sequencing; P/J, pediatric/juvenile; PV, parvoviral; RFLP, restriction fragment length polymorphism; RAT, right atrium; US, United States; VI, virus isolation.

Case Demographics

A search of the AHDC archives (June 2007 to November 2015) identified 1911 canine cases from the search period. Twenty-seven adult cases and 22 controls were identified using search criteria. PCR was performed on 26 cases with available tissue (Suppl. Table S2). The demographics of pediatric/juvenile cases and controls was previously reported.⁴⁴ Cases and controls were primarily from the northeastern United States.

The median age (minimum, maximum) of adult myocarditis/myocardial fibrosis cases was 6 years (2, 14). There were 5 castrated males, 4 intact males, 12 spayed females, and 5 intact females. Case breeds included mixed breeds (2), Maltese (2), Staffordshire Bull Terrier (2), Labrador Retriever (6), Beagle Hound (2), and 1 each of Greyhound, Shetland Sheepdog, French Spaniel, Golden Retriever, Rottweiler, Bichon Frise, Scottish Terrier, Pomeranian, Siberian Husky, Welsh Corgi, Weimaraner, and Puli. The median age of control cases was 9 years (2, 13). The controls included 6 castrated males, 3 intact males, 8 castrated females, and 5 intact females. Control breeds included Labrador Retriever (1), German Shepard (2), Staffordshire Bull Terrier (2), French Bulldog (2), Rat Terrier (2), Mixed Breed (2), Cocker Spaniel (2), and 1 each of Chihuahua, Spaniel, Coton De Tulear, Bulldog, Boston Terrier, Australian Shepard, Beagle Hound, Wheaton Terrier, and Cane Corso. The distribution of myocarditis cases by submission month is shown in Figure 1. Adult myocarditis cases occurred at a low frequency throughout the calendar year while pediatric/juvenile cases were most numerous in the summer (June/July) and fall (October/November).

Figure 1.

Canine myocarditis submissions for necropsy from June 2007 to November 2015 grouped by month and age.

The summary results of PCR testing are shown in Table 2. Not all PCR tests were performed for all cases/controls due to the aforementioned exclusion criteria. Nucleic acid from 12 different pathogens was detected by PCR in 14 pediatric/juvenile cases and 19 pediatric/juvenile controls, as well as in 5 adult cases and 11 adult controls. More than 1 agent was detected in the myocardium of 4 pediatric/juvenile cases and 4 pediatric/juvenile controls, as well as in 2 adult cases and 4 adult controls. The PCR testing did not detect any agents in 35 of 66 cases (53%; 95% CI, 41%–65%), including 14 of 40 pediatric/juvenile and 21 of 26 adult cases. No agents were detected in 27 of 57 controls (47%; 95% CI, 35%–60%).

Table 2.

Polymerase Chain Reaction Detection of Infectious Agents in Canine Myocarditis Cases and Controls by Age Class.

	All Cases, No. Positive/Total No. (%; 95% CI)	Cases <2 y, No. Positive/Total No. (%; 95% CI)	Controls <2 y, No. Positive/Total No. (%; 95% CI)	Cases >2 y, No. Positive/Total No. (%; 95% CI)	Controls >2 y, No. Positive/Total No. (%; 95% CI)
Bartonella spp	0/44	0/18	0/0	0/26	0/0
Borrelia spp	0/44	0/18	0/0	0/26	0/0
Mycoplasma haemocanis	3/54 (6; 1–16)	2/28 (7; 0–24)	1/35 (3; 0–16)	1/26 (4; 0–21)	0/22
Canine adenovirus	3/52 (6; 1–16)	2/28 (7; 0–24)	1/35 (3; 0–16)	1/24 (4; 0–22)	0/22
Canine respiratory coronavirus	1/54 (2; 0–11)	1/28 (4; 0–19)	1/35 (3; 0–16)	0/26	1/22 (5; 0–24)
Canine distemper	5/54 (9; 4–20)	5/28 (18; 7–36)	2/35 (6; 0–20)	0/26	5/22 (23; 10–44)
Canine parainfluenzavirus	1/54 (2; 0–11)	1/28 (2; 0–19)	0/35	0/26	0/22
Influenza A	0/54	0/28	0/35	0/26	1/22 (5; 0–24)
Canine herpesvirus 1	3/50 (6; 1–17)	2/27 (7; 0–25)	8/35 (23; 12–39)	1/23 (4; 0–23)	0/0
West Nile virus	1/32 (3; 0–17)	0/18	0/0	l/l4 (7; 0–31)	0/0
Canine parvovirus 2	16/66 (24; 15–36)	l4/40a (35; 22–50)	2/41a (5; 0–17)	2/26 (8; 0–25)	4/22 (18; 7–39)
Neospora caninum	7/49 (14; 7–27)	6/23 (26; 12–47)	9/35 (26; 14–42)	1/26 (4; 0–21)	3/22 (14; 4–34)

Abbreviation: CI, confidence interval.

^aCases/controls <2 years from Ford et al.⁴⁴

M. haemocanis was detected by rtPCR in pediatric/juvenile myocarditis cases, pediatric/juvenile controls, and adult myocarditis cases but not in adult controls. In pediatric/juveniles, M. haemocanis was detected with M. cynos. The rtPCR-positive cases and controls were in the low-positive range of the assay (Ct values 21.2–30.0), at or below the analytic detection limit, and attempted amplicon sequencing was unsuccessful.

CAV was detected by rtPCR in pediatric/juvenile cases and pediatric/juvenile controls, as well as in an adult case but not in adult controls. Although CAV rtPCR-positive cases and controls were in the moderate Ct range for the nanoscale platform (11.86–13.19), adenoviral inclusions and antigen were not detected in the myocardium of cases or controls by histopathology and IHC, respectively.

CRCoV was detected by rtPCR in a single pediatric/juvenile case, pediatric/juvenile control, and adult control but not in adult cases. Although CRCoV PCR-positive cases and controls had moderate Ct values (range, 15.05–19.37), CV antigen was not detected in the myocardium of cases by IHC.

CDV was detected by rtPCR in pediatric/juvenile cases, pediatric/juvenile control, and adult controls but not in adult cases. The median Ct value for pediatric/juvenile cases was 13.74 (range, 7.61–15.86); the Ct values for pediatric/juvenile controls were 16.99 and 17.05. The 7.61 value was considered high positive, and all others were moderate. CDV antigen was not convincingly detected by IHC in the myocardium of PCR-positive cases or controls. In pediatric/juvenile cases and controls, CDV nucleic acid was codetected with nucleic acid of other agents, including CAV (2/5 cases, 1/2 controls), M. cynos (1/5 cases), CPiV (1/5 cases), CPV-2 (2/5 cases), and Neospora (2/5 cases). The median CDV Ct value for adult controls was 18.13 (range, 13.30–19.30) and in the moderate range similar to pediatric/juvenile controls.

CPiV was detected by PCR in 1 pediatric/juvenile case (moderate Ct value 17.99) but not in pediatric/juvenile controls, adult cases, or adult controls.

CHV-1 was detected by PCR in pediatric/juvenile cases, pediatric/juvenile controls, and 1 adult case. The Ct range for pediatric/juvenile cases was 31.23 to 38.44 (median 37.94 Ct) or approximately 1 viral genome copy per microliter (range, 25.64–39.91). CHV-1 was detected in the myocardium of an 8-day-old PCRþ case using in situ hybridization (ISH) with signal localized to endothelial cells (case 40, Fig. 2) along with endothelial inclusions, cardiomyocyte degeneration, mineralization, hemorrhage, and mild neutrophilic inflammation considered secondary to myocardial damage. Herpesviral inclusions were not observed in other cases or controls.

Figures 2–5.

Heart, dog. Figure 2. Systemic canine herpesvirus 1 infection, case No. 40. There are glassy amphophilic endothelial inclusions (arrow) along with degeneration and necrosis of individual cardiomyocytes. Hematoxylin and eosin. Inset: Herpesviral nucleic acid (brown staining) is localized to endothelial nuclei by in situ hybridization. Figure 3. Myocardial necrosis and inflammation, case No. 18. The cardiac microvasculature is lined by plump endothelial cells with prominent darkly staining nuclei (arrow) with marginating intravascular leukocytes. Viral agents were not detected in this case by polymerase chain reaction screen. Figure 4. Cardiomyocyte degeneration, mineralization, and multinucleate giant cells, case No. 22. In this case not immunoreactive for canine parvovirus 1, there are multifocal small areas where necrotic cardiomyocytes are mineralized with scattered multinucleated cells (arrow) at the periphery of the necrotic foci. Hematoxylin and eosin. Figure 5. Tyzzer’s disease, case No. 31. A pediatric case with lymphoplasmacytic myocarditis and rod-shaped bacteria in cardiomyocytes morphologically consistent with Clostridium piliforme (arrow). Modified Steiner.

WNV was detected at or below the analytic detection limit and was thus considered low positive (Ct value 37.7) in 1 adult case but not detected in pediatric/juvenile cases, pediatric/juvenile controls, or adult controls. In the low-positive case, WNV antigen was not detected in the myocardium via IHC.

Neospora nucleic acid was detected in pediatric/juvenile cases, pediatric/juvenile controls, an adult case, and adult controls. Amplicon sequences from 2 of 7 PCR-positive pediatric/juvenile cases matched N. caninum sequences in Genbank (top 3 matches KX683873.1, LN714476.1, and KF6498848.1) with high (95%–96%) identity. Sequencing of other cases failed due to signal noise. Neospora PCR detection occurred without histological evidence of characteristic protozoa.

CPV-2 was detected in adult cases and controls in the low-positive range. The Ct range for adult cases was 27.69 to 31.33, and the median Ct value for adult controls was 31.01 (Ct range, 29.61–32.60); with standard curve interpolation, the range of PCR-positive adult cases was 8.35 × 10¹ and 1.15 × 10³ copies and the median quantity (minimum, maximum) of controls was 1.05 × 10² copies (3.35 × 10¹, 2.88 × 10²). CPV-2 was not detected in the myocardium of adult cases using ISH.

The Lyme exposure status was unknown for the majority of cases except for 2 adult cases with medical histories of positive serology, but B. burgdorferi was not detected in cardiac tissue in any cases or controls by rtPCR. B. miyamotoi is also present in the northeastern US region but was not detected in any samples. We did not detect Bartonella spp, Rickettsia spp, or S. zooepidemicus in any of the cases or controls regardless of age.

Cardiac histologic features were not associated with detection of agents by PCR. Demographic, histologic, and tissue detection characteristics of cases with PCR detection are presented in Table 3. In the adult cohort of 26 cases, 18 had some evidence of cardiomyocyte necrosis, 20 had inflammation, and 9 had some degree of myocardial fibrosis; 11 cases had a combination of necrosis, inflammation, and fibrosis. Five cases had predominantly necrosis, 9 cases had a combination of necrosis and inflammation, 9 cases were predominately inflammatory, and 3 cases had predominantly fibroplasia or fibrosis. A review of case histories and pathology reports revealed systemic disease in 15 cases, including 10 cases with systemic inflammation consistent with sepsis. Significant lesions were limited to the myocardium in 7 cases, including 5 cases that were predominately inflammatory. There were also 4 cases of mild cardiac necrosis, inflammation, or fibrosis, which most likely represented incidental changes. One pediatric case had multifocal to regional cardiomyocyte necrosis with dark, inclusion-like material in endothelial nuclei (case 18, Fig. 3), and 1 pediatric case had multifocal cardiomyocyte necrosis with mineralization and multinucleate cells also not immunoreactive for CPV-1 (case 22, Fig. 4).⁴⁴ CPV-1 was not detected by IHC in the subset of pediatric/juvenile cases with suggestive myocardial necrosis, mineralization, and giant cells.

Table 3.

Demographic, Histologic, and Tissue Detection Characteristics of Cases With Agent(s) Detected by PCR Screen.

Case No.	Age	Agents Detected by PCR	IHC/ISH	Myocardial Lesions	Other Lesions
Pediatric/juvenile casesa
2a	28 d	Neospora	ND	Degenerate and mineralized CM with inflammation	Suppurative pyelonephritis and acute neutrophilic bronchopneumonia
9a	14 d	Neospora	CPV-1 IHC−	Multifocal myocardial degeneration and necrosis	Enteritis, lymphoid depletion, and interstitial pneumonia
10a	3 m	CAV, CDV, Mycoplasma haemocanis, Mycoplasma cynos	AV IHC−; CDV IHC−	Chronic extensive lymphohistiocytic myocarditis	Mild interstitial pneumonia
18a	14 d	Neospora	ND	Myocardial necrosis; dark nuclei; aortic medial necrosis and mineralization	Infected tail dock site; systemic mineralization
19a	3 d	Neospora	ND	Vasculitis with myocardial mineralization	Urogenital infection
22a	10 d	Neospora	CPV-1 IHC−	Myocardial necrosis with mineralization and multinucleated cells	Interstitial pneumonia
23a	1 y	CDV, Neospora	CDV IHC±	Myocardial fibrosis	Multiorgan hemorrhage (suspect anticoagulant toxicity)
25a	2 y	CHV-1	ND	Severe coalescing necrotizing myocarditis with mineralization	NSL
28a	2 m	CAV, CDV, parainfluenza	AV IHC−; CDV+ (spleen only)	Focal myocardial necrosis and focal subacute pulmonary valvular endocarditis	Multifocal, acute peribronchial hemorrhage
30a	3 m	CDV, CPV-2	CDV IHC−	Multifocal acute hemorrhagic myocardial and vascular necrosis	Moderate hypophysitis, mild meningitis, multiorgan fibrinoid vascular necrosis, and hemorrhage fibrin thrombi
29a	4 m	CRCoV	CoV IHC−	Severe necrotizing myocarditis and ischemic necrosis with interstitial fibrosis	Hepatic centrilobular necrosis; suppurative splenitis; multisystemic fibrin thrombi; neutrophilic, lymphoplasmacytic, and histiocytic ileocolitis with crypt abscesses (CPV-2, CoV IHC−)
33a	5 m	M. haemocanis, M. cynos	ND	Chronic lymphohistiocytic myocarditis with fibrosis	NSL
38a	1 m	CDV, CPV-2	CPV-2 IHC lymph node +, heart -; CDV IHC−(heart, lung)	Severe neutrophilic myocarditis and endomyocarditis with necrosis	Bacterial septicemia, lymphoid depletion, interstitial pneumonia secondary to CPV-2 infection
40a	8 d	CHV-1	CHV-1 ISH+	Myocardial necrosis with mineralization and endothelial intranuclear inclusions	Systemic intranuclear inclusions; acute necrotizing multifocal interstitial pneumonia, splenitis, hepatitis, and enteritis; acute renal cortical necrosis
Adult cases
a3	5 y	Neospora	ND	Atrial epicarditis and myocarditis	Uterine rupture and sepsis
a9	10 y	WNV	WNV IHC−	Multifocal minimal-mild necrosis	Choroid plexus carcinoma
a25	5 y	CAV	AV IHC−	Multifocal myocardial necrosis	Glomerulonephritis (Lyme serology positive)

Abbreviations: AV, adenovirus; CAV, canine adenovirus; CM, cardiomyocytes; CDV, canine distemper virus; CoV, coronavirus; CPV-1, canine parvovirus 1; CPV-2, canine parvovirus 2; CRCoV, canine respiratory coronavirus; IHC, immunohistochemistry; ISH, in situ hybridization; ND, not done; NSL, no significant lesions; PCR, polymerase chain reaction; PV, parvovirus; WNV, West Nile virus.

^aFrom Ford et al.⁴⁴

Aerobic bacterial culture isolated Enterococcus spp (1/3) and Escherichia coli (2/3) from myocardial or other tissues of 3 pediatric/juvenile cases, but intralesional bacteria were not evident in examined myocardial sections. Bacterial culture was not performed on tissues from adult cases. Gram or modified Steiner stain was performed on 8 pediatric/juvenile cases with bacteria identified in 3. Bacteria morphologically consistent with C. piliforme were identified in 1 pediatric case (case 31, Fig. 5) with silver histochemical stain.

Discussion

Despite a comprehensive PCR screen, the cause of myocardial lesions remained undetermined for a large proportion of cases in our study. In particular, the cause of myocarditis was not known for the majority of adult cases, demonstrating the limitations of available molecular tests. Furthermore, agent detection by PCR did not imply causation: M. haemocanis, CAV, CRCoV, CPiV, CDV, and CHV were not associated with myocarditis as detection occurred with equivalent frequency in myocarditis cases and controls. Several agents of myocarditis (T. cruzi, L. infantum, H. americanum, C. immitis) were not identified in our cohort but would not occur in our region (northeastern United States) without travel history to the respective endemic areas.

Our research builds on previous research of canine myocarditis, which has been largely limited to case reports and case series reporting Bartonella spp,^14,35,41,109 Borrelia burdorferii,^33,57,73 Citrobacter koseri,²⁸ C. piliforme,¹³¹ Staphylococcus spp,⁵⁷ L. infantum,^103,121 I. tropicalis,¹⁰¹ N. caninum,^{6,58,82,85,90,98} T. cruzi (Chagas disease),^{31,64,127,128} WNV,^18,25,74 CPV-2,^{1,11,27,44,50,52,88,102} CPV-1/minute virus,⁵¹ H. americanum,⁹ T. gondii,⁵³ C. immitis,^16,100 Aspergillus spp,^17,21,61 B. dermatitidis,¹⁰⁷ and Prototheca spp¹¹⁵ as natural causes. There are reports of cardiac involvement with CHV⁹⁴ and CDV⁵⁵ experimental infection. The frequency and current prevalence of canine myocarditis agents are unknown except for areas with endemic Chagas disease or Leishmania infection.^64,103,127

Viral Agents

The viral agents tested for by the PCR panel were CPV-2, CRCoV, CDV, CAV, CPiV, influenza A, CHV-1, and WNV; most agents were occasionally detected in cases and controls.

The majority of reports of canine myocarditis describe CPV-2 myocarditis. Features were reviewed in our previous retrospective study, which demonstrated a significant association between CPV-2 and myocarditis/myocardial fibrosis in young dogs.⁴⁴ CPV-2 is cardiotropic and causes myocarditis in susceptible puppies but also contributes to canine myocarditis due to gastrointestinal (GI) barrier disruption, immunosuppression, and bacterial myocarditis.^28,53,123 In this study, CPV-2 was detected in several adult myocarditis cases; however, the viral quantity was low and not detected by ISH, and detection frequency was similar in adult controls. Considering the replication requirements of parvoviruses for mitotically active S-phase cells, these results suggest PCR detection of latent or subclinical infection.^92,93 Available data do not support CPV-2 as a primary cause of myocarditis in adult dogs.^81,87

The related CPV-1 (minute virus of canines) is a cause of neonatal enteritis with large intranuclear inclusions, lymphoid depletion, and less frequent focal to diffuse areas of myocardial necrosis with mineralization and giant cells.⁵¹ CPV-1 was not detected in selected cases in this cohort by IHC and appears to be an infrequent cause of canine myocarditis in neonates only.

CDV-associated necrosis and calcification of the myocardium of suckling pups is described in textbooks.⁶⁰ Nevertheless, primary descriptions of myocardial degeneration, necrosis, mineralization, and CDV inclusions are limited to experimental studies using gnotobiotic neonatal Beagles inoculated with virulent CDV at 5 to 7 days but not in pups inoculated at 10 to 21 days of age.⁵⁵ A literature search did not identify reports of naturally occurring canine myocarditis with CDV as a single agent, which suggests the role of CDV in myocarditis is primarily as an immunosuppressive agent.^53,125

WNV was detected at the analytic limit in 1 adult case in our study. Myocarditis, along with meningoencephalitis and multifocal hemorrhage, is a prominent lesion in native and exotic birds infected with vector-borne West Nile virus; comparatively, myocardial involvement is rare in nonavian species.¹¹⁴ Clinical WNV disease is rare in dogs, but necrotizing variably lymphocytic and neutrophilic myocarditis with vasculitis has been reported in dogs that also have WNV polioencephalomyelitis.^18,25,74

The test panel used here was informed in part by the comparative epidemiology of human myocarditis. In humans, the most frequent agents associated with myocarditis include adenovirus, herpesviruses (cytomegalovirus, Epstein-Barr virus, human herpesvirus 6, herpes simplex viruses 1 and 2), enterovirus, hepatitis C virus, influenza viruses A and B, parvovirus B19, and rhinovirus.^{12,40,67,72,97} Other causes (including parainfluenza, group A Streptococci, and Mycoplasma spp) have rarely been detected via serology or culture in patients with clinical features of myocarditis.⁶³

CAV has not been reported as a cause of canine myocarditis. In our cohort, CAV was infrequently detected by PCR in cases and controls; however, adenoviral antigen was not detected by IHC. CAV-1 (infectious canine hepatitis) infects hepatocytes and endothelial cells.³² Clinical disease is infrequent due to widespread vaccination and most often occurs in inappropriately vaccinated dogs under 1 year of age often in concert with stress, overcrowding, and/or coinfection.¹¹⁸ CAV-2, the agent of infectious tracheobronchitis, causes either inapparent infection or respiratory disease often as a coinfection with a variety of viruses and bacteria, where dogs are cohoused.¹¹⁷ Available data do not support a role for CAV in canine myocarditis.

CHV-1 infection is associated with fatal systemic disease in neonatal puppies (recently reviewed in Jager et al⁵⁶). In neonates, characteristic necrosis and hemorrhage can be observed in most organs; myocardial degeneration with or without necrosis can be a feature of systemic infection.^32,94 Latent infection is established in ganglia, which may remain subclinical or manifest as mild, recurrent genital, respiratory, or ocular disease.^32,56 Available data do not support a role for CHV-1 in myocarditis in adult dogs.

Acute myocarditis can be a complication of influenza virus infections in humans. Canine influenza virus infection can result in respiratory disease, most frequently in group-housed dogs.⁹⁹ CRCoV and CPiV both frequently contribute to canine respiratory disease with asymptomatic infection or mild respiratory illness; both viruses were detected in a single pediatric/juvenile myocarditis case. CPiV replicates in epithelial cells of the upper respiratory tract but is occasionally associated with viremia.³⁹ Strains of canine enteric CoV (distinct from CRCoV) have been associated with severe systemic disease, and highly pathogenic enteric CoV is an emerging cause of systemic vasculitis in dogs.^19,132 Available data do not support a role for influenza A, CPiV, or CRCoV in canine myocarditis.

Bacterial Agents

In our cohort, bacteria were identified in myocardial sections in 3 cases; unfortunately, bacterial culture was not routinely performed in cases. Bacteria, particularly Streptococcus, Staphylococcus, and Bartonella spp, frequently cause mural and valvular infective endocarditis, which has been reviewed by others.^{69,78,119,120} Reported causes of bacterial myocarditis include Bartonella spp,^{14,35,41,78,109} B. burdorferii,^33,57,73 C. koseri,²⁸ C. piliforme,¹³¹ Rickettsia rickettsii,¹⁰⁴ Staphylococcus spp,⁵⁷ and anaerobic infection⁴⁵.

In the dog and other species, C. piliforme can cause necrotizing enterocolitis, hepatic necrosis, and occasionally necrotizing myocarditis.¹³¹ As C. piliforme is difficult to culture, diagnosis is based on observation of characteristic microscopic lesions in young animals with intracellular bacilli (demonstrated by silver histochemical stains) in cells at the periphery of areas of necrosis; serologic tests and PCR are also available.

The northeastern United States is a Lyme disease endemic area with high rates of B. burgdorferi infection in dogs. While infection is frequent, clinical disease associated with canine B. burgdorferi infection is infrequent and generally nonspecific, including fever, arthritis, anorexia, lymphadenopathy, glomerulonephritis, neurologic signs, and rarely myocarditis. Lyme carditis is rare in humans and usually presents as a self-limiting atrioventricular block during acute infection, although a broad range of cardiac involvement has been reported.^42,67 Lyme carditis may be a direct inflammatory reaction to spirochete infection or may be immune mediated, similar to Lyme nephritis.^20,36,42 There are intriguing studies associating B. burgdorferi with human dilated cardiomyopathy, but epidemiologic data are lacking.^66,95,113 The European Concerted Action on Lyme Borreliosis (EUCALB) has proposed a clinical case definition for diagnosis of Lyme carditis in humans, which includes acute-onset atrioventricular conduction or other rhythm disturbance, myocarditis or pancarditis, and the detection of specific serum antibodies, supported by the detection of Borrelia spp (culture and/or PCR) in endomyocardial biopsy.¹¹² Cardiologic manifestations are not well documented in dogs, and there are few reports associating Borrelia spp with canine myocarditis, usually via IHC, and no specific criterion for the diagnosis of cardiac borreliosis in dogs.^33,57,73,76 In dogs, descriptions vary from pyogranulomatous³³ to chronic lymphoplasmacytic and histiocytic myocarditis with fibrosis,^57,73 the latter similar to human Lyme carditis.⁴² We did not detect Borrelia spp by PCR in our cohort of cases from a Lyme endemic area, but tissue detection in experimentally infected dogs can be sporadic.^2,116 These data indicate Borrelia spp are not frequently associated with lesions of canine myocarditis; however, the lack of detection by PCR does not rule out Lyme carditis in humans. Furthermore, the host-bacterium interactions responsible for manifestations of Lyme disease are not well defined.

In dogs, pyogranulomatous lymphoplasmacytic endocarditis and myocarditis are associated with infection with several species of the Bartonella genus.^{14,35,41,78,105,109} In our cohort, we did not detect Bartonella spp by PCR in the hearts of cases or controls regardless of age; however, PCR was performed on nucleic acid from FFPE tissue, which has been demonstrated to affect the sensitivity of Bartonella detection.³⁵

M. haemocanis, the agent of canine hemotropic mycoplasmosis, was detected by rtPCR in occasional cases and controls. Hemotropic mycoplasma is transmitted primarily by tick bite, transfer of infected blood during aggressive contact, and possibly perinatally.^26,65,86 Clinical disease (ie, hemolytic anemia) is rare in nonsplenectomized, immunocompetent dogs. We attribute the detection of M. haemocanis secondary to immune changes related to systemic disease.

Protozoal Agents

Protozoal agents were not observed histologically in our cohort, but we detected N. caninum by PCR in both cases and controls. N. caninum infection is associated with progressive hind limb paresis, ataxia, and ascending paralysis most severe in congenitally infected puppies. Encephalitis, myositis, myocarditis, pneumonia, and dermatitis have also been reported.^6,58 Infection is often disseminated with Neospora zoites in multiple tissues, but lesions and organisms are occasionally confined to the heart.⁹⁰ The prevalence of N. caninum infection and frequency of subclinical infection in the canine population are unclear; seroprevalence has been reported in up to 37.8%, with higher prevalence in farm dogs in rural environments compared to urban settings (reviewed elsewhere).^10,36,75 Subclinically infected dams can transmit the organism transplacentally, and successive litters may be born infected.³⁶ Variable vertical transmission and subclinical infection have been suggested by serosurvey of breeding dogs: N. caninum seroreactivity was reported in up to 52% (32/61) of pups born to seropositive dams while 22 of 88 (25%) pups demonstrated clinical signs of neosporosis.⁷ In our study, the clinical relevance of detection of N. caninum by PCR in the absence of Neospora zoites was difficult to determine; thus, subclinical infection and contamination must be considered.¹²⁶ In dogs, T. gondii myocarditis has been reported with viral immunosuppression.⁵³

T. cruzi, the hemoflagellate protozoan agent of Chagas disease (American trypanosomiasis), causes chronic pancarditis in dogs and humans. Chagas disease is limited geographically due to the range of the triatomine insect vector (reduviid or “kissing” bugs), which is present in the southern United States, Mexico, Central America, and South America. Seroprevalence, up to 23.6%, varies according to location, with higher seroprevalence in sporting and working breeds.^31,47,64 Intracellular amastigotes within pseudocysts are present primarily in cardiac muscle. Acute infection is associated with granulomatous and necrotizing myocarditis and numerous T. cruzi pseudocysts. Organisms are less frequent (or absent) with chronicity and are accompanied by lymphoplasmacytic infiltrate and fibrosis.^106,127 During acute infection, low serum antibodies may result in false-negative results via serum immunofluorescence testing.¹²⁷

The domestic dog is the intermediate host for 2 species of Hepatozoon, an arthropod-borne protozoal infection. Old World hepatozoonosis (Hepatozoon canis) occurs worldwide and has been reported in Asia, southern Europe, Africa, South America, and the southern United States. H. canis meronts, which have a unique wheel-spoke arrangement of micromerozoites, are reported to be very plentiful in the myocardium and other tissues of wild African carnivores. Conversely, they do not frequently involve the myocardium of South African domestic dogs and are not associated with myocardial inflammation. Dogs with severe infections can have multiorgan involvement with splenomegaly, hepatomegaly, and lymphadenomegaly.^4,5,48,83 New World hepatozoonosis is primarily found throughout the southeastern United States and Gulf coast of Texas. H. americanum parasites are frequently found in skeletal and cardiac muscle, which can cause pyogranulomatous myositis. Inflammation is scant around “onion skin” cysts or developing schizonts but robust in response to released merozoites with pyogranulomatous vasculitis and vascular mineralization.^3,9,48

Protozoa of the genus Leishmania cause disease in dogs in the Mediterranean region and parts of Africa, Asia, and the Americas. Lesions are most commonly found in the skin and hemolymphatic organs, but granulomatous and lymphoplasmacytic myocarditis with cardiomyocyte loss, vasculitis, fibrosis, and amastigotes within myocardial macrophages has also been reported in dogs.^103,121

Fungal Agents

None of the cases in our cohort had evidence of fungal myocarditis.

In endemic areas of the southwestern United States, C. immitis causes restrictive pericarditis, epicarditis, and myocarditis as part of disseminated coccidioidomycosis.^{16,59,79,100,110} Most dogs present with pulmonary coccidioidomycosis following inhalation of infective arthrospores. Young, large breed, hunting, working, or sporting dogs are most at risk. Disseminated infection frequently involves bone, joint, and lymph node, but pericardial involvement can lead to heart failure, arrhythmia, syncope, and sudden death.⁹⁵ Pericardial infection can be diagnosed based on intralesional spherules in biopsy specimens from pericardectomy samples.⁵⁴ Intradermal coccidioidin test and serum coccidioidomycosis complement fixation tests are prone to false negatives.¹⁰⁰

Granulomatous myocarditis can be observed as a component of systemic mycotic infection with identification based on fungal histomorphology, culture, serology, and/or PCR.^17,21,61,101 Myocardial and great vessel involvement have been reported in cases of canine systemic blastomyocosis.¹⁰⁷

Algal Agents

Myocardial involvement is frequent in dogs with disseminated protothecosis. Prototheca zopfii and Prototheca wickerhamii have been associated with often abundant myocardial organisms.¹¹⁵ None of the cases in our cohort had evidence of protothecosis.

Agents With Secondary Myocardial Effects

Several infections are associated with myocardial lesions, but available literature does not indicate direct cardiac involvement. Myocardial involvement is observed in dogs with systemic vasculitis due to R. rickettsii, possibly related to disseminated endothelial infection or secondary immune-mediated mononuclear inflammatory reaction. Myocardial vasculitis and hemorrhage can be accompanied by small vascular thrombi, focal ischemic necrosis, and left atrial and mitral valve endocarditis, but infection of cardiomyocytes has not been demonstrated.¹⁰⁴

Dogs with babesiosis and clinical anemia are reported to have cardiac arrhythmia and nonspecific cardiac lesions, including fibrin microthrombi, lymphoplasmacytic and variably neutrophilic infiltrates, and necrotic foci.³⁷ Furthermore, cTnI is elevated in dogs with immune-mediated hemolytic anemia or complicated infection, although infection of cardiomyocytes has not been reported.⁷⁷ Cardiac changes may the result of cardiac infection or secondary to clinicopathologic abnormalities caused by Babesia (eg, anemia, disseminated intravascular coagulation, electrolyte alterations, renal insufficiency). Similarly, cTnI elevations and cardiac arrhythmia reported with acute natural Ehrlichia canis infection were correlated with the severity of anemia and signs of systemic illness (ie, tachycardia, fever, leukocytosis); infection of cardiomyocytes has not been reported.³⁴ African horse sickness occasionally causes disease in dogs, usually associated with consumption of uncooked, infected horse meat.^89,122,124 Necropsy findings in dogs include respiratory (“dunkop”-like) lesions with immunoreactivity for African horse sickness primarily labeling pulmonary microvascular endothelial cells similar to infected horses. In the myocardium of infected horses, immunoreactivity is localized to endothelial cells, circulating leukocytes, and infrequently red blood cells but not cardiomyocytes; subendocardial myocardial hyaline degeneration and necrosis with possible myocardial IHC labeling have been reported in 1 dog.^30,124

Conclusion and Testing Recommendations

Our study demonstrates the challenges of determining the causes of canine myocarditis. A comprehensive PCR screen did not determine the cause of myocarditis in most canine cases in our study; furthermore, agents detected by PCR were often not considered to be causative. Regardless, our study and the data available to date inform some testing recommendations.

Although CPV-2 is well recognized as a cause of myocarditis in susceptible puppies, PCR detection was similar in adult cases and controls in our study and CPV-2 myocarditis has not been reported in adult dogs; thus, CPV-2 is not considered associated with myocarditis in adult dogs.

Many of the PCR-detected pathogens (CDV, CHV, and CAV) are associated with subclinical, latent, or otherwise inapparent infection, particularly in adult animals. As immunization with modified live vaccine (CDV, CPV-2, CAV-2, and CPiV) is recommended for dogs beginning as early as 6 weeks of age, PCR may be detecting vaccine virus and not field strains. The codetection of multiple agents infrequently associated with disease (due to widespread vaccination) suggests detection of vaccine virus. PCR detection at least represents evidence of exposure to an infectious agent; thus, ultimately, it is unclear if PCR-detected agents are cardiopathogenic, incidental, or latent as clinical recognition of myocarditis usually occurs as viral titers are waning.

Evidence indicating age restriction of some agents supports initial development of age-specific testing recommendations, primarily based on excluding tests for agents that have not been documented as a cause of myocarditis in an age group, although screening of larger cohorts of myocarditis cases is needed. Based on available data and pathogenesis of individual agents, we suggest testing to include CPV-2 and CHV in neonatal and pediatric/juvenile patients and CPV-1 and CDV only in cases of neonatal myocarditis. Data available to date do not support testing for CDV, CHV, or CPV in cases of adult myocarditis.⁸¹ WNV and B. burgdorferi appear to be infrequent causes of myocarditis, and the decision to test should be made on a case-by-case basis. Despite the prevalence in human myocarditis, it does not appear that CAV or influenza is associated with canine myocarditis. Bacterial culture (augmented by special histochemical stains) and virus isolation are recommended for all myocarditis cases as unrecognized/novel pathogens may be identified. Fungal, protozoal, and algal pathogens can often be detected via histologic examination, although the number of organisms can vary according to chronicity and previous treatment. The occurrence and incidence of immune-mediated myocarditis, a diagnosis made by exclusion in humans based on IHC criteria in the absence of PCR pathogen detection, are unknown in the dog.

In humans, endomyocardial biopsy paired with molecular testing (PCR for viral genomes and leukocyte IHC) is the gold standard for diagnosis. Unfortunately, ancillary testing to identify microbial agents of canine myocarditis is usually not elected because screening for all of the reported causes of myocarditis is cost prohibitive; furthermore, when testing is performed, it is after death and does not always reveal an underlying cause.^29,62,129 Current diagnostic modalities are insufficient to identify the causative agents in most cases. There is a need to identify agents and establish criteria for the diagnosis of immune-mediated myocarditis, particularly in adult dogs. New approaches are needed to identify emerging, unrecognized, or novel pathogens. This study establishes groundwork to stimulate the development of novel testing strategies for the diagnosis and thus investigation of the etiopathogenesis of canine myocarditis.