Effect of type of diet on blood and plasma taurine concentrations, cardiac biomarkers, and echocardiograms in 4 dog breeds

Abstract Background Associations of diet with dilated cardiomyopathy are under investigation. Objectives That cardiac assessment would show abnormalities in healthy dogs eating grain‐free (GF) diets or diets with Food and Drug Administration (FDA)‐listed ingredients of concern (peas, lentils, or potatoes) as top 10 ingredients (FDA‐PLP), but not in dogs eating grain‐inclusive (GI) diets or diets without FDA‐listed ingredients of concern (PLP) in the top 10 ingredients (NoFDA‐PLP). Animals One hundred eighty‐eight healthy Doberman Pinschers, Golden Retrievers, Miniature Schnauzers, and Whippets. Methods This study was an observational cross‐sectional study. Echocardiograms, cardiac biomarkers, and blood and plasma taurine concentrations were compared between dogs eating GF (n = 26) and GI (n = 162) diets, and between FDA‐PLP (n = 39) and NoFDA‐PLP (n = 149) diets, controlling for age and breed. Demographic characteristics, murmurs, genetic status, and ventricular premature complexes (VPCs) during examination were compared between dogs eating different diet types. Results No differences in echocardiographic variables, N‐terminal pro‐B‐type natriuretic peptide or whole blood taurine were noted between dogs eating different diet types. Dogs eating GF diets had higher median high‐sensitivity cardiac troponin I (hs‐cTnI) (GF 0.076 ng/mL [Interquartile range (IQR), 0.028‐0.156] vs. GI 0.048 [IQR, 0.0026‐0.080]; P < .001) and higher median plasma taurine (GF 125 nmol/mL [IQR, 101‐148] vs GI 104 [IQR, 86‐123]; P = .02) than dogs eating GI diets. Dogs eating FDA‐PLP diets had higher median hs‐cTnI (0.059 ng/mL [IQR, 0.028‐0.122]) than dogs eating NoFDA‐PLP diets (0.048 [IQR, 0.025‐0.085]; P = .006). A greater proportion of dogs eating FDA‐PLP diets (10%) had VPCs than dogs eating NoFDA‐PLP diets (2%; P = .04). Conclusions and Clinical Importance Higher hs‐cTnI in healthy dogs eating GF and FDA‐PLP diets might indicate low‐level cardiomyocyte injury.

Conclusions and Clinical Importance: Higher hs-cTnI in healthy dogs eating GF and FDA-PLP diets might indicate low-level cardiomyocyte injury. Theoretical reasons for dog foods that are grain free (GF) or high in peas, lentils, or potatoes to be associated with nDCM include nutritional deficiencies, poor nutrient bioavailability, impaired taurine metabolism or precursor availability, nutrient-to-nutrient interactions, or presence of toxic substances. 9 Genetic makeup or other modifying factors could also influence disease expression in some dogs and explain why not all dogs eating these diets are affected. Alternatively, it remains possible that this apparent association is confounded by other factors. 10 Although some affected dogs are taurine deficient (especially Golden Retrievers 7 ), thereby strengthening the association with food, the majority of dogs with suspected nDCM have normal blood taurine concentrations. 6 The role of taurine is uncertain at this time. [11][12][13][14] Nutritional dilated cardiomyopathy is a complex disorder with uncertain etiologies, undetermined contribution of dietary factors, variable phenotype, and unclear extradietary modifiers of disease expression. Many diseases have a spectrum of severity ranging from mild to overt disease, and so it is possible that some healthy dogs have subclinical cardiac abnormalities related to diet type. One study reported echocardiograms from clinically healthy Golden Retrievers, 8 but additional studies evaluating for subclinical disease related to diet are needed to facilitate efforts at early intervention and to enhance understanding of the natural history and pathophysiology of nDCM.
The purpose of this observational study was to evaluate the effect of diet type on blood and plasma taurine concentrations, cardiac biomarkers, and echocardiograms in apparently healthy dogs of 4 breeds.
We hypothesized that echocardiographic chamber dimensions and cardiac biomarkers would be higher in dogs eating diets that are GF or high in peas, lentils, or potatoes, compared with dogs eating diets without these characteristics, and that taurine concentrations would not differ between dogs eating diets with and without these characteristics.

| Inclusion criteria
Dogs were included if they were considered to be clinically healthy, had been eating a commercial dog food for at least 6 months, and were a purebred Whippet, Golden Retriever, Doberman Pinscher, or Miniature Schnauzer. We specifically selected 2 breeds with no known genetic predisposition to dilated cardiomyopathy (Miniature Schnauzer and Whippet), 1 with known heritable dilated cardiomyopathy (Doberman Pinscher), and 1 that appears to be at increased risk for nDCM (Golden Retriever), although appropriate epidemiological data for this condition are lacking. Four breeds with notably different characteristics were chosen for this study to test the hypotheses in diverse breeds without introducing extensive breed variability. The study was not designed to evaluate each breed individually.

| Exclusion criteria
Dogs were not included if they were fed more than 1 type of food within the 6 months before presentation, were receiving taurine supplementation, were receiving cardiac medications, or had known cardiac disease. Dogs with clinically important extracardiac diseases were excluded, as were female dogs that were pregnant, lactating, or in estrous.

| Examination and diet history
Body weight, body condition score (1-9 scale), 15 muscle condition score (normal, mild, moderate, or severe muscle loss), 16 and results of a physical examination were recorded for all dogs. Owners filled out a diet history form detailing specific foods, treats, supplements, and the duration of feeding. Biotin containing supplements were specifically noted because of the potential for mega-doses of this vitamin to interfere with high-sensitivity cardiac troponin I (hs-cTnI) measurement. 17,18

| Categorization of diets
The ingredient list of the diets fed to enrolled dogs were reviewed for categorization based on the presence of grains and separate categorization based on whether peas, lentils, or potatoes were listed in the top 10 ingredients based on the FDA's advisory (https://www.fda. gov/animal-veterinary/outbreaks-and-advisories/fda-investigationpotential-link-between-certain-diets-and-canine-dilatedcardiomyopathy). Diets were categorized as GF if there were no grain-containing or grain-derived ingredients listed and as grain inclusive (GI) if grain-containing or grain-derived ingredients were listed.
Grain-containing or grain-derived ingredients included whole-grain products or refined-grain products made from any part of wheat, rice, oats, corn, barley, or another cereal grain. 19 Oils were not considered a grain product. Diets were separately categorized as having FDAlisted ingredients of concern (peas, lentils, or potatoes) in the top 10 ingredients (FDA-PLP) or as without FDA-listed ingredients of concern (peas, lentils, or potatoes) in the top 10 ingredients (NoFDA-PLP). Inclusion of taurine and methionine as dietary ingredients was noted because these amino acids could affect blood taurine concentrations. 20 21 Left ventricular volumes in diastole and systole were indexed to body surface area (LVVd/m 2 and LVVs/m 2 , respectively). 22 Ejection fraction (EF) was calculated as the percent change in LV volume between diastole and systole. The right-sided short-axis view at the level of the chordae tendineae was used for M-mode measurements of LV size including LV internal dimension in diastole and systole normalized for body length using allometric scaling according to 2 formulas (normalized LV internal diameter in diastole [LVIDdN] and normalized LV internal diameter in systole [LVIDsN], respectively). 23,24 Fractional shortening (FS) was calculated as the percent change in LV diameter between diastole and systole using M-mode measurements. Left atrium to aortic diameter ratio (LA/Ao) was determined in the right-sided short-axis parasternal view in mid-diastole using 2D measurements. 25 Rhythm during the echocardiogram was noted.

| Statistical analysis
Commercial software was used for data analysis (SAS 9.4, Cary, North Carolina). Data are presented as median and interquartile range (IQR).
Differences in the baseline characteristics of dogs in each diet type group were compared using Fisher's exact test or chi-square test and Mann-Whitney test for categorical and continuous variables, respectively. Differences in selected echocardiographic variables (LVIDdN and LVIDsN using both allometric formulas, LA/Ao, FS, LVVD/m 2 , LVVS/m 2 , EF), whole blood and plasma taurine concentrations, hs-cTnI concentrations, and NT-proBNP concentrations were compared between dogs eating GF and GI diets and between dogs eating FDA-PLP and NoFDA-PLP diets using analysis of variance tests.
Crude, unadjusted analyses were performed to examine the association between the different diet categorizations and our principal study outcomes. We then performed multivariable regression analyses examining these associations while controlling for the dogs' age and breed, yielding multivariable adjusted P values.
The prevalence of murmurs, occurrence of arrhythmias noted on monitoring electrocardiogram during echocardiography, and genetic positivity for Doberman Pinschers (heterozygous or homozygous for either mutation) were compared between dogs eating GF and GI diets, and between dogs eating FDA-PLP and NoFDA-PLP diets using Fisher's exact test. Significance was set at P < .05. There were no significant differences in terms of age, weight, sex, or time eating the diet before enrollment according to the type of diet consumed (Table 1). There were no differences in genetic status between Doberman Pinschers fed GF and GI diets or between Doberman Pinschers fed FDA-PLP and NoFDA-PLP diets. No differences in murmur prevalence were present between diet types. All murmurs were low grade (1/6 or 2/6). The proportion of dogs with ventricular premature complexes (VPCs) noted on the monitoring electrocardiogram during echocardiography was significantly higher for dogs eating FDA-PLP diets compared with NoFDA-PLP diets (P = .04).

| Study population
The majority (92%) of dogs in the study were assessed as has having normal muscle condition. Muscle condition scores were not significantly different between diet categorizations after controlling for age and breed (median and IQR scores were normal for all groups;

| Diets and supplements
The 188 dogs in this study were fed 72 diets, which were represented by 34 different brands. All diets were dry kibble. Table 2 shows the number of dogs from each breed that were eating each diet type.
Twenty-six dogs were eating GF diets and 162 dogs were eating GI diets. The proportion of dogs eating GF diets that included taurine or methionine as ingredients (21/26; 81%) was greater than the proportion of dogs eating GI diets that included taurine or methionine as ingredients (63/162; 39%, P < .001).  .04 Note: Median and interquartile range are shown for the first 3 variables. Number and percentage of the total are shown for the remaining variables.
Genetic status was considered positive if Doberman Pinschers were heterozygous or homozygous for either the DCM1 or DCM2 mutations. P < .05 was considered significant (bolded). Abbreviations: DP, Doberman Pinschers; FDA-PLP, diets with Food and Drug Administration ingredients of concern (peas, lentils, or potatoes) in the top 10 ingredients; GF, grain-free diets; GI, grain-inclusive diets; GR, Golden Retrievers; NoFDA-PLP, diets without Food and Drug Administration ingredients of concern (peas, lentils, or potatoes) in the top 10 ingredients; VPCs, ventricular premature complexes.
3.2.2 | Categorization based on FDA ingredients of concern in top 10 ingredients

| Echocardiography, cardiac biomarkers, and taurine concentrations
After multivariable regression, no differences were found between diet categorizations (GF vs GI and FDA-PLP vs NoFDA-PLP) for any echocardiographic variables, NT-proBNP, or whole blood taurine concentrations (Table 3). Median hs-cTnI (P < .001) and median plasma taurine concentrations (P = .02) were both significantly higher in dogs eating GF diets compared to GI diets and median hs-cTnI concentration (P = .006) was significantly higher in dogs eating FDA-PLP diets compared to NoFDA-PLP diets (Table 3).

| SNAP 4Dx
Twenty-eight dogs enrolled at the University of Florida showed elevated serum hs-cTnI concentrations and so underwent in-house SNAP 4Dx testing. All dogs were negative for Dirofilaria immitis antigen, Anaplasma phagocytophilum antibody, Anaplasma platys antibody, Ehrilichia canis antibody, and Ehrlichia ewingii antibody. One dog was determined to be positive for Borrelia burgdoferi C6 antibody. This dog was eating a diet categorized as GI and NoFDA-PLP.

| DISCUSSION
This study of 4 dog breeds found significantly higher hs-cTnI concentrations in healthy dogs eating GF diets compared to dogs eating GI diets, and in healthy dogs eating FDA-PLP diets compared to dogs eating NoFDA-PLP diets, and these differences persisted after controlling for the dogs' age and breed. Serum concentrations of cardiac troponin I are low in healthy dogs. 27 Myocardial cell injury is associated with release of troponin I into the circulation, allowing for its detection by analyzers. 27 Abbreviations: FDA-PLP, diets with Food and Drug Administration ingredients of concern (peas, lentils, or potatoes) in the top 10 ingredients; GF, grain-free diets; GI, grain-inclusive diets; NoFDA-PLP, diets without Food and Drug Administration ingredients of concern (peas, lentils, or potatoes) in the top 10 ingredients.
compared to the GI group. Oral supplementation with taurine and dietary inclusion of taurine and methionine increase plasma taurine concentrations independent of whole blood taurine concentrations 34,35 but 1 study showed both higher whole blood and plasma taurine concentrations in dogs eating diets fortified with taurine and methionine. 20 Alternatively, the higher plasma taurine concentrations in dogs eating GF diets might be caused by cardiomyocyte injury, which could be supported by the finding that hs-cTnI was also higher in dogs eating GF diets compared to GI diets. Circulating taurine concentrations have historically been used as a biomarker of myocardial injury in people, although whole blood concentrations seem more closely associated with injury than plasma concentrations in some studies. [36][37][38] Elevated plasma taurine concentrations in conjunction with low intramyocardial taurine concentrations in a rat model of isoproterenol-induced myocardial injury resulted from defective myocardial taurine transport and increased sarcolemmal taurine release associated with cardiomyocyte damage. 39 Taurine supplementation to rats increases expression of the myocardial taurine transporter, restoring taurine movement into the myocardial cells. 39 Our results contrast with a study of healthy Golden Retrievers, which found no difference in plasma taurine concentrations, but lower whole blood taurine concentrations in dogs eating non-traditional diets compared to traditional diets. 8 The reasons for these contrasting findings are unclear, but might be influenced by the other, diverse breeds included in our study, different diets that were fed to dogs in each study, and different diet characterizations used in each study. While the reasons that dogs consuming GF diets had higher plasma taurine concentrations in our study are unknown, this finding could be related to dietary inclusion of taurine or methionine as an ingredient, sample handling artifacts, or myocardial disease and injury.
Although the number of dogs with VPCs noted on monitoring electrocardiography during the echocardiogram was relatively low, a greater proportion of dogs eating FDA-PLP diets had VPCs than those eating NoFDA-PLP diets. No echocardiographic differences were found between healthy dogs in this study eating GF and GI diets or between healthy dogs in this study eating FDA-PLP and NoFDA-PLP diets. In addition, no diet type differences were detected for NT-proBNP. The lack of differences in echocardiographic measurements is in contrast to a recent study in Golden Retrievers in which FS was lower and LV diameters were larger in those eating nontraditional diets compared to traditional diets. 8 Similar to postulated reasons for contrasting findings regarding taurine concentrations in that publication, our study enrolled 3 breeds in addition to Golden Retrievers and utilized different diet categorizations. The absence of echocardiographic differences between dogs eating different diet types in the current study could be because these dogs were clinically healthy with a low disease prevalence, breed differences and variability, low statistical power, or to true absence of differences. We intentionally chose to study 2 breeds that are not predisposed to dilated cardiomyopathy (Miniature Schnauzer and Whippet), 1 breed that is predisposed to genetically-based dilated cardiomyopathy (Doberman Pinscher), and 1 breed that appears predisposed to nDCM and taurine Note: Median and interquartile range are shown with crude, unadjusted P values obtained from the univariable analyses and adjusted P values after multivariable regression analysis controlling for age and breed. P < .05 after multivariable adjustment was considered to be statistically significant (bolded). Abbreviations: EF, ejection fraction; FDA-PLP, diets with Food and Drug Administration ingredients of concern (peas, lentils, or potatoes) in the top 10 ingredients; GF, grain-free diets; GI, grain-inclusive diets; hs-cTnI, high-sensitivity cardiac troponin I;  The conflicts of interest reported by the authors did not influence data collection, data interpretation, or manuscript preparation.

OFF-LABEL ANTIMICROBIAL DECLARATION
No off-label antimicrobial use.