Different carbohydrate sources in dog foods supported overall health and cardiac function: an 18-mo prospective study in healthy adult dogs

Abstract

A link between dilated cardiomyopathy (DCM) and dog foods marketed as grain-free has been suggested. In this randomized, parallel-group, double-blind study, the impact of 4 foods with different ingredient profiles on echocardiographic parameters and cardiac biomarkers was assessed in 60 dogs over 18 mo. Foods included a grain-free diet with potatoes and peas (GF + PPo), a grain-inclusive diet with peas and pea fiber (G + PPF), a grain-inclusive diet without peas or potatoes (G), and a grain-free diet with potatoes (GF + Po). Echocardiographic parameters, blood and urinary taurine, and serum cardiac troponin-I and NT-BNP were assessed at 6, 12, and 18 mo. No clinically significant changes or between-group differences were observed in cardiac troponin-I or NT-BNP. Whole blood and plasma taurine levels remained within the normal range and were unaffected by diet. Despite diet-by-time interactions in wLVIDd, wLVIDs, fractional shortening, and ejection fraction (P < 0.05), all dogs were considered clinically normal regarding DCM. Twenty-four dogs were diagnosed with mild endocarditis by study end, which may explain the observed echocardiographic changes. These data demonstrate that cardiac function was supported in healthy adult dogs fed foods formulated to provide similar nutrition through different ingredient profiles. These results demonstrate the importance and effectiveness of balanced, high-quality nutritional formulations.

Both grain-free and grain-inclusive diets providing complete and balanced nutrition supported normal cardiac function in healthy adult dogs in an 18-mo study.

Graphical Abstract

Graphical Abstract.

Introduction

Dilated cardiomyopathy (DCM) is a disease of the myocardium that is characterized by impaired systolic dysfunction, dilation, and impaired contraction of the left ventricle, and thinning of the ventricle wall that can lead to arrhythmias and sudden death (Dukes-McEwan et al., 2003; McCauley et al., 2020). In dogs, the progression of DCM can be slow and its clinical signs and symptoms, which include congestive heart failure, breathlessness, lethargy, anorexia, exercise intolerance, and sudden fainting, may not manifest until the final stages of the disease (Dukes-McEwan et al., 2003; Mansilla et al., 2019).

DCM is the second most common heart disease in dogs and has been reported in approximately 0.5% of dogs evaluated in U.S. veterinary referral hospitals, with a higher prevalence in large and giant breeds and in purebred dogs (0.65%) versus mixed breeds (0.16%; Sisson et al. (2000); Dukes-McEwan et al., 2003). A prior retrospective analysis focusing solely on cardiology patients between 2011 and 2019 documented a 3.9% incidence rate (Quest et al., 2022). However, authors found no statistically significant alteration in the general population’s DCM prevalence between 2000 and 2019. The prevalence of DCM is particularly high (>3.0%) in Deerhounds, Dobermanns, Irish Wolfhounds, Great Danes, and Boxers. An increased prevalence of DCM has also been reported in Golden Retrievers, Saint Bernards, and Bulldogs (Meurs, 2003; Bélanger et al., 2005; McCauley et al., 2020).

Although diagnosis of DCM requires documentation of left ventricular dilation, depressed systolic function, and increased sphericity of the left ventricle (Dukes-McEwan et al., 2003), evaluating cardiac biomarkers can provide a more comprehensive assessment of preclinical DCM (McCauley et al., 2020). Measurement of serum concentrations of the N-terminal end of the brain natriuretic peptide (NT-proBNP) provides a measure of chronic strain on the myocardium, while assessment of cardiac troponin-I can indicate acute myocardial damage (Langhorn and Willesen, 2016; McCauley et al., 2020).

Traditionally, DCM has been considered a primarily inherited disease, although other recognized causes of DCM include endocrine disorders, myocarditis, chronic tachycardia, and specific nutrient deficiencies (Phillips and Harkin, 2003; Sanderson, 2006; Vollmar et al., 2019; McCauley et al., 2020; Freid et al., 2021). In recent years, concern regarding the association of specific foods with the development of DCM has been raised (US FDA-CVM, 2018; Mansilla et al., 2019), although no definitive link has so far been established.

In 2018, the U.S. Food and Drug Administration (FDA) questioned the association between DCM in dogs with the consumption of foods marketed as grain-free and those containing potatoes and pulse ingredients, such as peas and lentils (US FDA-CVM, 2018). Note that there are no regulations that define foods that can be considered grain-free, although product claims must be truthful and not misleading; thus, these foods cannot include traditional sources of grains, such as rice, corn, oats, barley, and wheat (US Food and Drug Administration, 2018a). However, foods marketed as grain-free can include carbohydrates sourced from plants, such as white potatoes, sweet potatoes, green and yellow peas, chickpeas, and lentils, which can provide similar functionality in processing compared with grains. While the FDA has issued updates on this report, no causal link between grain-free foods and DCM has been established, and in December 2022, the FDA announced it would not provide additional updates until future research allows more definitive conclusions (US FDA-CVM, 2019a, b, 2022).

Nonetheless, several studies have shown that DCM and its subclinical echocardiographic precursors can occur in dogs eating pulse-based, grain-free, or other nontraditional or boutique foods (e.g., raw foods that are grain-free or foods that are manufactured by a small pet food company with < $1billion in global sales), with peas being the most common ingredient in the foods of the reported cases (Kaplan et al., 2018; Ontiveros et al., 2020; Quilliam et al., 2023; Coppinger et al., 2024). Other studies have failed to find echocardiographic changes in dogs eating grain-free foods (Adin et al., 2021; Leach et al., 2023), although increases in blood markers of cardiac disease have been noted (Adin et al., 2021). Moreover, switching dogs to foods with grains or that exclude pulses has been reported to improve cardiac function and clinical outcomes in dogs diagnosed with DCM (Freid et al., 2021; Haimovitz et al., 2022; Walker et al., 2022). However, additional changes were made in many of these studies, such as changes in protein source, levels of micronutrients, format of foods, and added medications, which likely confounded these findings.

One explanation proposed for the association between diet and DCM involves the relative deficiencies in taurine, a sulfur-containing amino acid. Taurine is found exclusively in animal protein and is absent from plant-sourced proteins (McCauley et al., 2020). Dogs can synthesize taurine from methionine and cysteine but require dietary methionine to meet daily requirements (Mansilla et al., 2019). Still, data suggest that rates of taurine synthesis vary among dog breeds, with larger dogs exhibiting lower rates of taurine synthesis than smaller dogs (Ko et al., 2007). Other factors that may affect taurine synthesis include the preferential consumption of sulfur-based amino acids, including the precursors of methionine, by the gut microbiome (Backus et al., 1994).

While incompletely understood, taurine is found in high concentrations in the heart, where it is involved in the utilization of cellular energy and the maintenance of membrane integrity (Ripps and Shen, 2012). Taurine is also involved in the formation of bile salts through its conjugation with cholesterol-derived bile acids (Ridlon et al., 2016). While most bile acids are reabsorbed and recycled, some are lost in the feces and some are consumed by the gastrointestinal microbiome (Ellegård and Andersson, 2007; Pezzali et al., 2020; Clark et al., 2023), resulting in a loss of taurine and an alteration of the fecal bile acid pool.

The higher concentration of soluble fiber provided by peas and other pulse ingredients compared to grains may also increase the likelihood that the fibers bind with bile acids such as taurocholic acid (TCA), supporting a larger microbial population and increasing their likelihood of being utilized by intestinal microbes or excreted in the stool (Hoving et al., 2018; Donadelli et al., 2020; Pezzali et al., 2021). Increased microbial activity associated with higher concentrations of soluble fiber may also increase the likelihood of bile acid catabolism, reducing the likelihood of bile acid resorption. Therefore, it has been suggested that grain-free foods with high concentrations of legumes may reduce taurine bioavailability and lead to changes in cardiac parameters similar to those observed in DCM (Bokshowan et al., 2023).

To evaluate the potential link between diet and DCM, this study randomized healthy adult dogs to 1 of 4 complete and balanced foods with different ingredient profiles to assess cardiac function and overall health over the course of 18 mo. We hypothesized that the test foods would not negatively impact overall health or heart function indices.

Materials and Methods

Study overview

This study was a prospective, double-blind, randomized, parallel-group trial conducted between December 2, 2019, and August 25, 2021 (Figure 1). All procedures were directed by a protocol approved (CP901) by the Institutional Animal Care and Use Committee of Hill’s Pet Nutrition, Inc., in Topeka, KS, USA. All methods were carried out in accordance with relevant institutional and national guidelines and regulations.

Figure 1.

Study design.

Population

Purebred and mixed-breed adult dogs of any size were considered for inclusion if they were deemed healthy and had body condition scores (BCS) of 2 to 4 on a 5-point scale. Dogs were excluded if they had systemic diseases of the heart, including DCM, heart failure, or cardiovascular disease, kidney disease, or endocrine disease. All animals were colony housed and treated in accordance with Hill’s Global Animal Welfare Policies. They were allowed normal socialization and enrichment activities, including interactions with other dogs, people, and toys. Fresh water was provided ad libitum. The study did not interfere with the normal routine of the animals.

Study design and dietary interventions

The study consisted of a 15-d screening period in which all dogs were fed the same basal food. Dogs that met the inclusion criteria were stratified by breed, age, sex, reproductive status, and weight and were randomly assigned to one of the 4 test foods using a complete block design (Figure 1).

Test foods included (1) a grain-free food with potatoes and peas (GF + PPo); (2) a grain-inclusive food with peas and pea fiber (G + PPF); (3) a grain-inclusive food with no peas or potatoes (G); and (4) a grain-free food with potatoes (GF + Po). Ingredient and nutrient profiles of the 4 foods are shown in Tables 1 and 2. All foods were formulated using Concept5 (CFC Tech Services, Inc., Pierz, Minnesota, USA), and nutritional analysis of finished foods was conducted by a commercial laboratory (Eurofins Scientific, Inc., Des Moines, IA, USA) using official methods published by AOAC International (2019). All test foods were designed to provide complete and balanced nutrition for adult dogs and met the nutritional concentration requirements set forth by the Association of American Feed Control Officials for canine adult maintenance foods (AAFCO, 2019). The test foods were nutritionally similar, included similar quantities of sulfur-containing amino acids, added taurine and B vitamins, and had similar digestibility profiles (Morris et al., 2023), although the G + PPF food included increased fiber.

Table 1.

Ingredients in test foods as a percentage of the total recipe

	Test food3
Ingredient, %	GF + PPo	G + PPF	G	GF + Po
Chicken	13.3	16.0	15.0	10.4
Potato starch	16.4	–	–	26.4
Potatoes	16.4	–	–	26.4
Pea protein	16.8	16.6	–	–
Yellow peas	16.4	15.0	–	–
Chicken fat	7.1	6.4	4.1	5.1
Whole grain corn	–	5.0	10.0	–
Cracked pearled barley	–	5.0	10.0	–
Whole grain wheat	–	–	13.5	–
Soybean oil	3.4	3.5	2.5	3.8
Potato protein	–	–	–	12.6
Brewers rice	–	8.3	3.8	–
Chicken liver flavor	3.0	3.0	3.0	3.0
Whole grain sorghum	–	–	10.0	–
Corn protein meal	–	–	10.0	–
Pea fiber	–	9.7	–	–
Brown rice	–	5.0	–	–
Lactic acid	1.2	1.2	1.2	1.2
Pork liver flavor	1.2	1.2	1.2	1.2
Soybean meal	–	–	4.6	–
Flaxseed	1.1	1.0	1.1	0.9
Dried beet pulp	–	–	1.5	2.1
Powdered cellulose	–	–	1.0	2.5
Dicalcium phosphate	0.8	0.5	0.5	1.5
Calcium carbonate	1.0	0.9	0.7	0.6
Chicken meal	–	–	3.0	–
Iodized salt	0.4	0.4	0.4	0.4
Choline chloride	0.4	0.3	0.4	0.5
DL-methionine	0.4	0.3	0.2	0.3
Potassium chloride	–	–	0.6	0.6
Vitamin Premixes1	0.3	0.3	0.3	0.3
Taurine	0.1	0.1	0.1	0.1
L-Lysine	–	0.01	0.97	0.01
Mineral premixes2	0.06	0.06	0.06	0.04
Mixed tocopherols	0.05	0.05	0.05	0.05
Cysteine	0.04	0.04	0.04	0.04
L-Tryptophan	0.04	0.01	0.06	0.03
Magnesium oxide	–	–	–	0.06

¹Vitamin premixes included rice hulls as a carrier for this blend. The individual vitamin compounds included vitamin E, L-ascorbyl-2-polyphosphate, niacin, thiamin, Vitamin A, calcium pantothenate, riboflavin, biotin, vitamin B12, pyridoxine, folic acid, vitamin D, choline chloride, and beta-carotene.

²Mineral premixes included calcium carbonate as a carrier for this blend. The individual mineral compounds included ferrous sulfate, zinc oxide, copper sulfate, manganese oxide, dicalcium phosphate, iodized salt, potassium chloride, magnesium oxide, calcium iodate, and sodium selenite.

³GF + PPo, grain-free + peas and potatoes; G + PPF, grain-inclusive + peas and pea fiber; G, grain-inclusive without peas or potatoes; GF + Po, grain-free + potatoes.

Table 2.

Analyzed¹ nutrient profile of test foods on a dry matter basis

		Test food5
Nutrient	AAFCO Min4	GF + PPo	G + PPF	G	GF + Po
DM, % as fed		90.6	91.2	90.7	90.8
ME2 (calculated), kcal/kg		4,108	4,022	3,970	3,933
Ash, %		6.1	5.9	6.1	6.0
Crude protein, %	18.0	27.1	30.4	27.3	26.1
Crude fat, %	5.5	18.8	18.9	15.7	15.3
NFE3 (calculated), %		44.6	38.6	48.0	49.1
Crude fiber, %		3.4	6.2	2.8	3.5
Total dietary fiber, %		8.6	16.5	11.0	8.2
Insoluble fiber, %		7.1	13.3	9.1	7.5
Soluble fiber, %		1.6	3.2	1.9	0.7
Arginine, %	0.51	1.92	2.15	1.26	1.46
Histidine, %	0.19	0.66	0.76	0.62	0.63
Isoleucine, %	0.38	1.13	1.24	1.02	1.30
Leucine, %	0.68	1.97	2.26	2.87	2.37
Lysine, %	0.63	2.12	2.37	2.30	2.16
Methionine, %	0.33	0.88	0.87	0.75	0.83
Methionine & cysteine, %	0.65	1.22	1.27	1.17	1.18
Phenylalanine, %	0.45	1.22	1.36	1.32	1.40
Phenylalanine & tyrosine, %	0.74	1.99	2.23	2.15	2.37
Threonine, %	0.48	1.06	1.17	0.95	1.35
Tryptophan, %	0.16	0.35	0.36	0.35	0.39
Valine, %	0.49	1.32	1.47	1.25	1.59
Taurine, ppm		1126	1083	1332	1452
Calcium, %	0.50	1.09	1.10	1.00	1.07
Phosphorus, %	0.40	0.77	0.78	0.73	0.78
Potassium, %	0.60	0.98	0.88	0.86	0.84
Sodium, %	0.08	0.35	0.35	0.31	0.34
Chloride, %	0.12	0.56	0.57	1.20	0.86
Magnesium, %	0.06	0.12	0.16	0.12	0.10

¹Nutritional analysis was conducted by Eurofins Scientific, Inc. (Des Moines, IA, USA) using official methods of analysis published by AOAC International (2019).

²ME, Metabolizable energy. Calculated ME based on modified Atwater values.

³NFE, Nitrogen-free extract. Calculated NFE as follows: % NFE = 100 % – (% ether extract + % crude protein + % ash + % crude fiber).

⁴AAFCO, Association of American Feed Control Officials. Minimum values based on the AAFCO Official Publication, 2019 (AAFCO 2019).

⁵GF + PPo, grain-free + peas and potatoes; G + PPF, grain-inclusive + peas and pea fiber; G, grain-inclusive without peas or potatoes; GF + Po, grain-free + potatoes.

Once enrolled, dogs were fed once daily based on their metabolic body weight (BW) for 18 mo. Food intake was recorded daily. To ensure that all dogs were given sufficient food to maintain their ideal BW, the desired kcal/day for each dog was calculated using the maintenance energy requirement for laboratory dogs recommended by the National Research Council (130 kcal/d × BW in kg; National Research Council, 2006). Dogs were weighed monthly, and food intake was adjusted accordingly.

Criteria for removal from the study included an average weight loss greater than 10% and the use of medications, treatments, or vaccinations administered for emergency intervention that would affect study results. Any adverse events were recorded, reported to the attending veterinarian, study coordinator, and principal investigator. Dogs that developed health conditions that require medical interventions such as surgery, medication, or other therapies could have been removed from the study at the discretion of the principal investigator or attending veterinarian. Routine vaccinations and heartworm, flea, and tick preventives were permitted and recorded.

Study procedures

At baseline and after 6, 12, and 18 mo, each dog underwent a physical exam, echocardiogram, blood collection, fecal analysis, and urinalysis (Figure 1). BW and BCS were evaluated monthly, where BCS were determined by a single, trained individual blinded to dietary treatment using the 1 to 5 scale (1 = very thin and 5 = obese). Whole blood was collected from fasted dogs for 12 h for complete blood count (CBC) and taurine analysis.

Serum was collected for analysis of serum chemistry, cardiac troponin-I, and NT-BNP analysis. CBC, serum chemistry, plasma NT-NBP, serum troponin I, whole blood and plasma taurine and amino acids, urinary taurine excretion, and fecal bile acid excretion were analyzed in accordance with established methods. CBCs in whole blood were conducted using Sysmex XT-2000iV (Sysmex Corp, Kobe, Hyogo, Japan), while serum chemistry was evaluated using Roche Cobas c501 (Roche Diagnostics Corp., Indianapolis, Indiana, USA). Plasma NT-NBP and serum troponin I were analyzed by IDEXX Laboratories, Inc. (Westbrook, Maine, USA). Whole blood, plasma, and urine taurine were analyzed by the University of California Davis School of Veterinary Medicine Amino Acid Laboratory (Davis, California, USA) using previously reported methods (Kim et al., 1995; Delaney et al., 2003). Fecal bile acids were analyzed by the Texas A&M Veterinary Medical Diagnostic Laboratory (College Station, Texas, USA) using established methods (Guard et al., 2019).

Echocardiography was performed on dogs without sedation by a veterinary radiologist blinded to the dietary treatments using a commercial ultrasound machine (MyLab 30, Esaote SpA, Genoa, Italy) with specific measurements carried out using the 3.5 MHz dedicated cardiac probe. If a dog would not sit for the echocardiogram, light sedation was given and noted on their report. The following measurements were taken from at least 3 cardiac cycles and the mean was recorded: the left atrial to aortic diameter (LA/Ao) ratio obtained from the right parasternal short-axis 2D; the left ventricular internal diameter at end diastole (LVIDd), left ventricular internal diameter at end systole (LVIDs), interventricular septal thickness at end diastole (IVSd), interventricular septal thickness at end systole (IVSs), left ventricular posterior wall dimensions at end diastole (PWd), left ventricular posterior wall dimensions at end systole (PWs), aortic diameter (AOD), and left atrial diameter (LA) measured on the M-mode echocardiogram obtained from the right parasternal short-axis view. M-mode values were used to calculate the fractional shortening (FS) and ejection fraction (EF) percentage. Doppler flow interrogations for detecting mitral and tricuspid regurgitation were also conducted. All the M-mode echocardiographic values were normalized and calculated according to the allometric scaling of Hall and colleagues (Hall et al., 2008): adjusted variable (w) = variable (cm) / (0.795 × BW^⅓).

Study endpoints

The primary endpoints of the study included changes in echocardiographic parameters (wLVIDd, WLVIDs, FS, and EF), whole blood and plasma taurine, cardiac troponin-I, and NT-BNP. Secondary endpoints include changes in BW, body condition score, fecal bile acids, and the following echocardiographic parameters: IVSd, IVSs, PWs, AOD, and LA. Changes in the dogs’ CBC results, chemistry panels, and physical exam results were also monitored, as were reported adverse events.

Statistical analysis

For the analysis of all endpoints, the normality of data was tested using the UNIVARIATE procedure on the residuals of the data. When the data did not meet assumptions for normality, statistical analysis was performed on the natural logarithmic transformation of the data, and the data were then back-transformed for reporting purposes.

All blood, urine, fecal, and echocardiogram variables were analyzed using a linear mixed model (MIXED procedure) with dietary treatment (diet), month of study (time), and the diet by time interaction as fixed effects. Time was included as a random effect in the model with animal as the subject to account for subject-to-subject variation among the dogs. Because time was a classification variable, an appropriate covariance structure was selected using the corrected Akaike information criterion fit statistic to account for correlations between repeated measurements over time. An unstructured covariance structure was used to account for the correlation between the random intercepts and slopes. The Kenward-Roger adjustment was used to approximate the degrees of freedom and reduce bias of variance estimates. Least squares means were separated using Tukey’s adjustment, and effects were considered significant at P ≤ 0.05. All statistical analyses were conducted using SAS, version 9.4 (SAS Institute, Cary, NC, USA).

Results

Disposition and baseline characteristics

Sixty dogs were enrolled (n = 15 dogs per group), and 54 dogs completed the study (Figure 2). Six dogs were removed from the study due to health concerns deemed unrelated to test foods by the attending veterinarian, including a cancer diagnosis (n = 4) and decreased health (n = 2).

Figure 2.

Study disposition.

Demographic and baseline characteristics of participating dogs are summarized in Table 3. There were no significant differences among groups related to the sex, weight, or age of the dogs. A wide range of dog breeds participated in the study, including Australian Shepherds, Beagles, Fox Terriers, Golden Retrievers, Labrador Retrievers, and a variety of mixed breeds. At baseline, the mean age of the dogs was 9.9 ± 1.9 yrs, the mean weight was 15.8 ± 7.7 kg, and 31 (51.6%) were male, one of which was intact. All other dogs were spayed or neutered.

Table 3.

Demographics and baseline characteristics. Data are presented as absolute counts or as means ± standard deviation

Signalment	Overall	Test food1
		GF + PPo (n = 15)	G + PPF (n = 15)	G (n = 15)	GF + Po (n = 15)
Animals, n	60	15	15	15	15
Male, n	31	8	8	7	8
Female, n	29	7	7	8	7
Body weight, kg	15.8 ± 7.7	15.6 ± 8.6	14.6 ± 5.9	15.7 ± 8.2	15.9 ± 9.5
Average age, yr	9.5 ± 1.9	10.6 ± 1.9	9.7 ± 1.8	10.2 ± 1.9	9.2 ± 1.7

¹GF + PPo, grain-free + peas and potatoes; G + PPF, grain-inclusive + peas and pea fiber; G, grain-inclusive without peas or potatoes; GF + Po, grain-free + potatoes.

BW and dietary intake

BW, body composition score, dietary intake, and caloric intake were similar among groups during the 18-mo study (Table 4; P > 0.050). Per kg of metabolic BW (BW^0.75), dogs consuming GF + PPo ate less than dogs consuming G (P = 0.014). While a main diet effect was observed for caloric intake per kg BW^0.75 (P = 0.049), no differences among dietary treatments were observed with Tukey’s adjustment.

Table 4.

Effect of diet, day of study (time), and the diet × time interaction on body weight, body condition, and food intake over the 18-mo study

	Test food1					P-values
Variable	GF + PPo (n = 15)	G + PPF (n = 15)	G (n = 15)	GF + Po (n = 15)	SEM2	Diet	Time	Diet × Time
Body weight, kg	23.2	22.3	22.1	24.4	3.53	0.870	<0.001	0.111
Metabolic body weight, kg0.75	10.3	10.1	10.0	10.8	1.24	0.879	<0.001	0.058
Body condition score	3.2	3.4	3.2	3.3	0.13	0.260	<0.001	0.659
Intake, g/d	223	230	255	247	23.2	0.710	<0.001	0.365
Intake, kcal/d	839	833	920	875	84.5	0.859	<0.001	0.400
Intake, g/kg BW0.75/d	27.9a	31.3a,b	32.1b	29.4a,b	1.03	0.014	<0.001	0.414
Intake, kcal/kg BW0.75/d	105	113	116	104	3.7	0.049	<0.001	0.384

¹GF + PPo, grain-free + peas and potatoes; G + PPF, grain-inclusive + peas and pea fiber; G, grain-inclusive without peas or potatoes; GF + Po, grain-free + potatoes.

²SEM, pooled standard error of the mean.

^a,b,cWithin row, diets with different superscripts differ at P ≤ 0.050.

Echocardiographic endpoints

None of the dogs showed any clinical evidence of DCM on ultrasound. The veterinary radiologist who conducted the echocardiograms, who was blinded to the dietary treatments, deemed all results clinically normal with respect to DCM. However, indicators of chronic degenerative valve disease (CDVD) were noted in several dogs. While no dogs were diagnosed with CDVD at baseline, 10 dogs in the G + PFF group were diagnosed with mild valvular endocardiosis by the end of the study, compared to 5 dogs each in the GF + PPo and G groups and 4 dogs in the GF + Po group.

Variables measured via echocardiograms are presented in Table 5. A diet effect was observed for both FS and EF (P = 0.014 and 0.005, respectively), with both being higher in G and GF + Po compared to G + PPF. A diet effect was also observed for weight-adjusted left ventricular internal dimensions (wLVIDs; P = 0.003), in which both G and GF + Po were lower than G + PPF.

Table 5.

Effect of diet, day of study (time), and the diet × time interaction on echocardiographic variables over the 18-mo study

	Normal reference1	DCM reference2	Test food					P-values
Variable			GF + PPo (n = 15)	G + PPF (n = 15)	G (n = 15)	GF + Po (n = 15)	SEM5	Diet	Time	Diet × time
LVIDd, cm		6.2 ± 1.1	3.07	3.25	3.06	3.12	0.145	0.782	0.444	0.021
LVIDs, cm		5.2 ± 1.0	1.59	1.78	1.51	1.57	0.094	0.194	0.680	0.001
IVSd, cm		1.0 ± 0.2	1.15	1.18	1.19	1.21	0.045	0.831	0.046	0.960
IVSs, cm		1.2 ± 0.3	1.86	1.91	1.92	1.95	0.074	0.829	0.077	0.971
PWd, cm			1.32	1.27	1.35	1.35	0.042	0.476	0.255	0.153
PWs, cm			2.03	1.96	2.09	2.08	0.069	0.528	0.252	0.427
AOD, cm			1.70	1.67	1.68	1.67	0.056	0.983	0.001	0.980
LA, cm			2.81	2.82	2.78	2.76	0.104	0.969	0.106	0.880
LA/AO			1.66	1.70	1.65	1.65	0.021	0.258	< 0.001	0.601
wLVIDd	1.73 ± 0.11	2.11 ± 0.32	1.57	1.67	1.54	1.54	0.038	0.051	0.081	0.010
wLVIDs	1.15 ± 0.13	1.65 ± 0.29	0.81a,b	0.92a	0.76b	0.77b	0.032	0.003	0.732	< 0.001
FS, %	34.0 ± 5.1	16.0 ± 5.0	48.6a,b	45.3a	50.9b	50.0b	1.26	0.014	0.416	0.010
EF, %	61.0	<40.0	80.6a,b	77.0a	82.8b	82.0b	1.19	0.005	0.346	0.004
wIVSd			0.59	0.61	0.60	0.60	0.012	0.765	0.038	0.798
wIVSs			0.95	0.99	0.97	0.97	0.022	0.664	0.175	0.960
wPWd			0.68	0.66	0.68	0.67	0.014	0.745	0.062	0.152
wPWs			1.04	1.02	1.06	1.03	0.020	0.535	0.177	0.340
wAOD			0.87	0.87	0.85	0.83	0.018	0.393	< 0.001	0.931
wLA			1.44	1.46	1.40	1.37	0.027	0.078	0.497	0.691

¹LVIDd, left ventricular internal diameter at end diastole; LVIDs, left ventricular internal diameter at end systole; IVSd, interventricular septal thickness at end diastole; IVSs, Interventricular septal thickness at end systole; PWd, left ventricular posterior wall dimensions at end diastole; PWs, Left ventricular posterior wall dimensions at end systole; AOD, Aortic diameter; LA, left atrial diameter; LA/AO, left atrial to aortic diameter ratio; Adjusted variables (w) = Variable (cm) / (0.795 × BW^⅓) according to Hall et al., 2008; FS, fractional shortening; EF, ejection fraction.

²Adapted from Hall et al., 2008, and Visser et al., 2019.

³Adapted from Freid et al., 2021, and Dukes-McEwan et al., 2003.

⁴GF + PPo, grain-free + peas and potatoes; G + PPF, grain-inclusive + peas and pea fiber; G, grain-inclusive without peas or potatoes; GF + Po, grain-free + potatoes.

⁵SEM, pooled standard error of the mean.

^a,b,cWithin row, values with different superscripts differ at P ≤ 0.050.

Diet-by-time interactions were observed for LVIDd, wLVID diastole (wLVIDd), LVIDs, wLVID systole (wLVIDs), FS, and EF (Figure 3). The wLVIDd of dogs consuming GF + Po decreased from baseline to the end of the study (P = 0.025), whereas the wLVIDd of dogs consuming the other foods were similar over time. Dogs consuming G + PPF had greater wLVIDs at months 12 and 18 compared to both G (P = 0.022 and 0.002, respectively) and GF + Po (P = 0.043 and 0.002, respectively). Both FS and EF were lower in G + PPF at month 18 compared to G (P = 0.004 and 0.003, respectively) and GF + Po (P = 0.040 and 0.020, respectively), as expected given the observed changes in left ventricular diameter over time. Additionally, EF in dogs consuming G + PPF was lower at the end of the study (month 18) compared to baseline (P = 0.046). All other echocardiographic variables were unaffected by the diet-by-time interaction.

Figure 3.

Changes in (A) weight-adjusted left ventricular internal diameter in diastole (wLVIDd); (B) weight-adjusted left ventricular internal diameter in systole (wLVIDs); (C) fractional shortening (%); and (D) ejection fraction (%) over the 18-mo study for each dietary treatment: grain-free plus peas and potatoes (GF + PPo), grain-inclusive plus peas and pea fiber (G + PPF), grain-inclusive without peas or potatoes (G), and grain-free plus potatoes (GF + Po). The gray shaded area in A and B and the black line in C and D represent the normal reference range reported in Hall et al. (2008) and Visser et al. (2019). The red line in A and B and the red shaded area of C and D represent the reference for dogs with dilated cardiomyopathy reported in Freid et al., 2021, and Dukes-McEwan et al., 2003. Within each month of study, bars with different superscripts indicate differences at P < 0.050. Within treatment, bars with asterisks (*) indicate a difference at P < 0.050.

CBC, blood chemistry, and blood, fecal, and urinary markers

Results of all CBC and chemistry panels are summarized in Tables 6 and 7. These data were subject to review by the attending veterinarian, who determined that the laboratory values were not indicative of clinically significant findings. This assessment was based on the observation that all values were either within established normal reference ranges or, in instances where values deviated from the normal range, did not present any discernible indicators of clinical illness. Cardiac troponin-I levels were below the limit of detection (<0.2 ng/mL) and could not be statistically analyzed.

Table 6.

Effect of diet, day of study (time), and the diet × time interaction on complete blood count parameters of dogs consuming 1 of the 4 dietary treatments over 18 mo

	Normal range1	Test Food2					P-values
Variable4		GF + PPo (n = 15)	G + PPF (n = 15)	G (n = 15)	GF + Po (n = 15)	SEM3	Diet	Time	Diet × time
Basophils, %		0.107	0.093	0.107	0.100	0.020	0.953	0.383	0.138
Basophils, k/uL	0.0 to 0.1	0.009	0.007	0.008	0.007	0.0017	0.839	0.124	0.084
Eosinophils, %		5.7	4.5	4.9	4.2	0.57	0.264	0.226	0.299
Eosinophils, k/uL	0 to 1.4	0.49a	0.33a,b	0.35a,b	0.32b	0.047	0.034	0.041	0.072
Hematocrit, %	33.0 to 58.7	45.9	48.2	48.9	48.7	0.99	0.110	0.017	0.278
Hemoglobin, g/dL	10.5 to 20.1	15.3	16.1	16.5	16.4	0.36	0.091	<0.001	0.248
Lymphocytes, %		22.4	24.9	24.3	25.8	1.82	0.580	0.413	0.888
Lymphocytes, k/uL	0.3 to 3.9	2.0	1.9	1.8	2.2	0.24	0.686	0.503	0.451
MCH, pg	21 to 27	23.3	23.2	23.4	23.5	0.17	0.595	0.230	0.610
MCHC, g/dL	30.1 to 41.9	33.4	33.4	33.8	33.5	0.16	0.357	<0.001	0.296
MCV, fL	63.0 to 78.3	69.8	69.4	69.3	70.3	0.60	0.588	0.001	0.477
Monocytes, %		6.7a	4.9b	6.2a,b	5.7a,b	0.44	0.027	0.385	0.280
Monocytes, k/uL	0.0 to 1.4	0.60a	0.36b	0.43b	0.41b	0.040	<0.001	0.219	0.278
Neutrophils, %		65.0	65.6	65.4	64.0	1.84	0.920	0.639	0.938
Neutrophils, k/uL		6.0	4.9	4.8	4.8	0.42	0.133	0.788	0.952
Platelets, k/uL	140 to 540	235	219	220	261	13.8	0.081	0.032	0.678
RBC, M/uL	4.48 to 8.53	6.57	6.93	7.11	6.95	0.158	0.104	<0.001	0.590
RDW, fL	33.0 to 39.4	35.7	35.3	35.8	36.1	0.29	0.259	0.035	0.225
Reticulocytes, %	0.31 to 1.58	0.90	0.77	0.68	0.79	0.070	0.152	0.373	0.069
Reticulocytes, M/uL		0.060	0.054	0.049	0.055	0.0051	0.454	0.567	0.097
WBC, k/uL	4.0 to 18.2	9.0	7.6	7.3	7.5	0.55	0.096	0.623	0.863

¹IDEXX reference range; GF + PPo, grain-free + peas and potatoes.

²G + PPF, grain-inclusive + peas and pea fiber; G, grain-inclusive without peas or potatoes; GF + Po, grain-free + potatoes.

³SEM, pooled standard error of the mean.

⁴MCH, mean cell hemoglobin; MCHC, mean cell hemoglobin concentration; MCV, mean cell volume; RBC, red blood cells; RDW, red cell distribution width; WBC, white blood cells.

^a,b,cWithin row, diets with different superscripts differ at P ≤ 0.050.

Table 7.

Effect of diet, day of study (time), and the diet × time interaction on serum chemistry parameters of dogs consuming 1 of the 4 dietary treatments over 18 mo

	Normal range1	Test food2					P-values
Variable		GF + PPo (n = 15)	G + PPF (n = 15)	G (n = 15)	GF + Po (n = 15)	SEM3	Diet	Time	Diet × time
Albumin, g/dL	2.9 to 4.0	3.53a	3.77b	3.69a,b	3.75a,b	0.062	0.032	0.268	0.297
Albumin/ Globulin	1.1 to 2.0	1.42	1.58	1.57	1.56	0.054	0.140	0.006	0.156
ALP, U/L	30 to 439	207	169	105	119	57.2	0.567	< 0.001	0.390
ALT, U/L	22 to 268	49	52	55	56	6.7	0.829	0.018	0.995
Bilirubin, mg/dL	0.0 to 35.1	0.052a	0.072a,b	0.084b	0.078a,b	0.0087	0.046	< 0.001	0.579
BUN, mg/dL	6.4 to 20.5	15.63	15.61	13.85	13.91	1.020	0.389	0.127	0.091
BUN/Creatinine	10.3 to 33.6	23.9	26.1	21.5	20.8	1.55	0.071	0.124	0.017
Calcium, mg/dL	0.8 to 20.1	10.39	10.52	10.42	10.54	0.094	0.587	0.232	0.148
Chloride, mmol/L	105 to 115	110.49	111.95	110.93	111.43	0.471	0.144	< 0.001	0.510
Cholesterol, mg/dL	133 to 401	197	202	209	208	10.8	0.836	0.043	0.966
Creatinine, mg/dL	0.41 to 0.82	0.68	0.64	0.67	0.70	0.046	0.841	0.613	0.400
Globulin, mg/dL	1.9 to 2.9	2.56	2.45	2.39	2.45	0.071	0.405	0.010	0.144
Glucose, mg/dL	58 to 96	82.4	84.8	84.4	80.7	1.71	0.250	< 0.001	0.606
Hemolysis		7.7	12.7	13.9	15.6	2.45	0.104	0.264	0.450
IRF, %		14.6	13.8	14.4	15.6	1.11	0.659	0.307	0.564
Lipemic, mg/dL		7.7a,b	6.6a	9.6a,b	17.7b	2.96	0.027	0.379	0.483
Magnesium, mg/dL	1.7 to 2.3	1.96	2.02	1.96	1.92	0.033	0.202	0.002	0.021
Na:K	28 to 37	32.3	33.1	32.7	32.8	0.50	0.703	0.027	0.641
Phosphorus, mg/dL	2.9 to 6.1	4.33a	4.04a,b	3.88a,b	3.84b	0.126	0.026	0.010	0.031
Potassium, mmol/L	4.0 to 5.3	4.58	4.53	4.55	4.55	0.063	0.954	0.040	0.677
Sodium, mmol/L	144 to 150	147.8a	149.0b	148.1a,b	148.4a,b	0.29	0.028	0.003	0.099
Total Protein, g/dL	5.1 to 6.5	6.09	6.22	6.08	6.20	0.085	0.514	0.619	0.222
Triglycerides, mg/dL	34 to 429	81a	84a,b	108a,b	134b	14.6	0.034	0.446	0.252

¹IDEXX reference range; GF + PPo, grain-free + peas and potatoes.

²G + PPF, grain-inclusive + peas and pea fiber; G, grain-inclusive without peas or potatoes; GF + Po, grain-free + potatoes.

³SEM, pooled standard error of the mean.

^a,b,cWithin row, diets with different superscripts differ at P ≤ 0.050.

Whole blood, plasma, and urinary markers of heart health are presented in Table 8. Day of study (time) affected all variables (P < 0.001) except urinary creatinine, which was evaluated to normalize urinary taurine. Whole blood taurine, plasma taurine, urinary taurine, urinary creatinine, and NT-BNP were unaffected by diet and showed no diet-by-time interactions. The urine taurine:creatinine ratio was reduced in test food GF + PPF (P = 0.042) but showed no diet-by-time interaction (P = 0.206).

Table 8.

Effect of diet, day of study (time), and the diet × time interaction on blood and urine markers of taurine status and heart health over the 18-mo study

	Normal reference range1	Test food2					P-values
Variable		GF + PPo (n = 15)	GI + PPF (n = 15)	G (n = 15)	GF + Po (n = 15)	SEM3	Diet	Time	Diet× Time
Whole blood taurine, ng/mL	>200	288.6	287.7	261.8	271.7	11.01	0.231	<0.001	0.421
Plasma taurine, ng/mL	>60	125.9	132.9	124.4	118.1	6.11	0.388	<0.001	0.783
NT-BNP1, pmol/L	<800	534	584	646	662	52.5	0.248	0.037	0.561
Urine taurine:creatinine, nmol:nmol		0.54	0.70	0.77	0.66	0.064	0.063	0.082	0.786

¹Reference range from Delaney et al., 2003, and de Lima and Ferreira, 2017.

²GF + PPo, grain-free + peas and potatoes; G + PPF, grain-inclusive + peas and pea fiber; G, grain-inclusive without peas or potatoes; GF + Po, grain-free + potatoes.

³SEM, pooled standard error of the mean.

Fecal bile acid data are presented in Table 9. All fecal bile acids were unaffected by the diet by time interaction. Ursodeoxycholic acid was affected only by time (P < 0.001), whereas all other fecal bile acids were affected by both diet (P < 0.050) and time (P < 0.050). Regardless of diet, ursodeoxycholic acid was lower at baseline compared to all other timepoints (P < 0.001). Cholic acid was increased at month 18 compared to baseline (P = 0.006), regardless of diet, and was greater with GF + PPo than both G and GF + Po (P = 0.036 and 0.001, respectively). Fecal chenodeoxycholic acid excretion was greater in GF + PPo compared to both G and GF + Po (P = 0.045 and 0.001, respectively) and, regardless of dietary treatment, was lower at baseline compared to all other timepoints (P < 0.001). Lithocholic acid excretion was reduced in G + PPF compared to all other diets (P = 0.025, 0.001, and 0.050, respectively) and, regardless of diet, was lower at baseline compared to month 12 and 18 (P = 0.026 and 0.138, respectively). Fecal deoxycholic acid excretion was greater with GF + Po compared to both G + PPF and G (P = 0.041 and 0.018, respectively) and, regardless of diet, was lower at baseline compared to all other timepoints (P < 0.050). GF + PPo increased total primary bile acid excretion compared to G and GF + Po (P = 0.035 and 0.001, respectively); total primary bile acid excretion was reduced at all timepoints compared to baseline (P < 0.050). GF + Po increased total secondary bile acid excretion compared to G + PPF and G (P = 0.025 and 0.014, respectively), and total secondary bile acids were lower at baseline compared to all other timepoints, regardless of diet (P < 0.050). Total bile acid excretion was higher at baseline compared to all other timepoints (P < 0.001), but with Tukey’s adjustment, no differences among dietary treatments were observed.

Table 9.

Effect of diet, day of study (time), and the diet × time interaction on fecal bile acids (ng/mg) over the 18-mo study

	Test food1					P-values
Bile acid, ng/mg	GF + PPo (n = 15)	G + PPF (n = 15)	G (n = 15)	GF + Po (n = 15)	SEM2	Diet	Time	Diet × time
Colic acid	999a	715a,b	491b,c	355c	92.8	0.001	0.006	0.332
Chenodeoxycholic acid	139a	90a	68b	44b	13.7	0.001	<0.001	0.053
Lithocholic acid	497a,c	319b	403a,b	594c	31.4	<0.001	0.010	0.287
Deoxycholic acid	4,056a,b	3,503a	3,387a	5,323b	283.5	0.015	<0.001	0.553
Ursodeoxychoic acid	59.2	52.2	50.7	53.7	1.68	0.240	<0.001	0.889
Total primary bile acids	1150a	811a,b	566b,c	404c	106.3	<0.001	0.002	0.283
Total secondary bile acids	4,755a,b	3,944a	3,883a	6,042b	307.0	0.010	<0.001	0.552
Total bile acids	6,875	5,278	4,826	6,858	351.2	0.027	<0.001	0.837

¹GF + PPo, grain-free + peas and potatoes; G + PPF, grain-inclusive + peas and pea fiber; G, grain-inclusive without peas or potatoes; GF + Po, grain-free + potatoes.

²SEM, pooled standard error of the mean.

^a,b,cWithin row, values with different superscripts differ at P ≤ 0.050.

Discussion

The current study, one of the largest studies to date assessing the impact of different foods on surrogate markers for DCM in healthy adult dogs, did not identify changes in cardiac or echocardiographic biomarkers indicative of clinical or subclinical DCM in any study group with any of the 4 test foods, each of which was formulated to provide a similar nutritional profile for adult dogs. Aside from increased fiber in G + PPF, all test foods were nutritionally similar, including comparable quantities of added sulfur-containing amino acids and B vitamins. Additionally, all test foods had similar macronutrient and energy digestibility profiles (Morris et al., 2023) and were formulated with the same quantity of added taurine, removing the confounding factor of the cardiologic impact of taurine deficiency. Dogs that completed the study maintained their BW, had normal CBC and serum chemistry profiles, and were considered clinically normal by the attending veterinarian.

There were also no notable changes in cardiac biomarkers of DCM during the 18-mo study. Cardiac troponin-I, which is known to increase proportionally to the degree of myocardial damage, was below the limit of detection (<0.2 ng/mL) in the majority of samples, preventing statistical analysis. NT-BNP levels in this study were also considered normal and unaffected by diets.

Although we observed significant diet-by-time interactions for wLVIDd, wLVIDs, FS, and EF, all dogs were considered clinically normal (with regard to DCM) by the blinded veterinary radiologist who performed the echocardiograms at all timepoints during this study. Additionally, the mean echocardiographic measurements from each food in this study are similar to those reported in over 1,100 healthy dogs across a wide range of BWs (3.0 to 68.5 kg), summarized in a meta-analysis by Hall and colleagues (Hall et al., 2008). As this comparison includes wLVIDs, FS, and EF, it seems reasonable to conclude that, from a clinical perspective, despite the statistical differences in these variables between treatments, the feeding of these study foods did not negatively impact the heart health and functionality of these dogs.

These results were consistent with those observed in other studies of healthy dogs fed grain-free foods (Cavanaugh et al., 2021; Leach et al., 2023). For example, one study of 32 purebred beagles and 33 mixed-breed hounds did not detect the development of cardiac dysfunction as assessed by cardiac biomarkers, echocardiographic parameters, or endomyocardial biopsies with 4 different foods (a high animal protein, grain-free food, a high animal protein grain-inclusive food, a low-animal protein grain-free food, and a low-animal protein grain-inclusive food; Leach et al., 2023). A treatment-by-day-by-breed effect was observed, suggesting that beagles who were fed a low-animal-protein grain-free food had a larger normalized left ventricular internal diameter at end-systole than those fed a grain-inclusive food; however, all variables were within normal reference ranges. FS for Beagles fed the grain-free food was also lower than those fed foods including grains.

Another study that evaluated the impact of healthy dogs fed either a commercial extruded pea-protein-based food (PBD, n = 34) and a traditional extruded food found no meaningful differences in LVIDs, normalized LVIDs, or FS, and no statistical evidence of between-group differences in any echocardiographic parameters (Cavanaugh et al., 2021). No clinically relevant hematologic, biochemical, or echocardiographic alterations were noted, although plasma amino acid and taurine levels were significantly higher in the PBD group. However, in the dogs receiving the PBD, LVID diameter and normalized LVID diameter were higher at the end of the study.

In contrast, a prospective cross-sectional study of healthy dogs eating a traditional food (including grains) or nontraditional food (including pulse ingredients as a primary ingredient) found that dogs eating nontraditional foods had lower indices of systolic function and larger left ventricular volumes compared with dogs eating traditional foods (Owens et al., 2023). It should be noted that dogs that were receiving taurine supplementation were excluded from the study. Similarly, a retrospective study of Irish Wolfhounds eating high-pulse or low-pulse foods found that high-pulse foods were associated with a higher prevalence of ventricular premature complexes, a potential early marker of cardiac abnormalities (Coppinger et al., 2024). Given these results, some authors have suggested that the screening of dogs eating high-pulse foods might be warranted to facilitate the early detection of DCM in this population (Owens et al., 2023).

Still, the onset of CDVD was noted in several dogs throughout the course of this study, which was expected given that the prevalence of cardiac disease increases with advancing age (Sisson, 2002). However, since none of the dogs in this study had diagnosed CDVD at baseline, dietary randomization could not account for differences in early markers of CDVD at baseline, which is a limitation of the current study. As a result, more dogs consuming G + PPF were diagnosed with CDVD than dogs consuming G or GF + Po, which could be one possible explanation for the changes in echocardiographic parameters observed over time in that treatment group. Other cardiac conditions are known to cause cardiac changes similar to those observed in DCM, such as chamber dilation and systolic dysfunction (Machen and Sleeper, 2015). For example, myxomatous mitral valve disease, the most common cardiovascular disease in dogs, has been shown to lead to gradual enlargement of the left side of the heart and accompanying reduction in both FS and EF (Boswood et al., 1765; Bonagura and Visser, 2022). Thus, it is possible that the observed increase in wLVIDs and associated decrease in FS and EF over time in dogs consuming G + PPF was a result of the progression of pre-existing subclinical CDVD.

In our study, blood taurine levels were within normal ranges in all treatment groups. Across all foods, whole blood taurine was 130% or greater the level considered for deficiency (<200 nmol/mL), and plasma taurine was 200% or greater the level considered for deficiency (<60 nmol/mL; Delaney et al., 2003). Urinary taurine excretion was also unaffected by diet. Other studies have shown variable effects of grain-free foods and legume-containing foods on taurine concentrations (Donadelli et al., 2020; Pezzali et al., 2020; Cavanaugh et al., 2021; Gizzarelli et al., 2021; Bakke et al., 2022; Clark et al., 2023).

The canine gut microbiota is known to metabolize the primary bile acids produced by the canine hosts into chemically distinct secondary bile acids (Rowe and Winston, 2024). Previous studies of dogs fed high-protein, high-fiber, and grain-free foods have reported that food appears to have a minimal impact on microbial-derived secondary bile acids in healthy dogs (Schmidt et al., 2018; Donadelli et al., 2020; Pezzali et al., 2020; Phungviwatnikul et al., 2021; Reis et al., 2021). In contrast, in our study, statistically significant among-group differences in fecal bile acids were observed throughout the study. Generally, primary fecal bile acid secretion was highest among dogs fed peas (GF + PPO and G + PPF) and lowest among dogs fed G and GF + PO, while total secondary bile acid secretion was highest among dogs fed GF + PO. However, the higher rates of bile acid secretion observed among dogs fed GF + PPO and G + PPF may relate more to the greater fat content in these foods than to the inclusion of peas in these foods. Additional research is needed to clarify how, and to what extent, changes in dietary composition affect the metabolism of primary bile acids and the production of secondary bile acids by the gut microbiome of healthy dogs.

Strengths of this study include its relatively long duration and its inclusion of multiple breeds and of dogs of a variety of sizes. However, this study had several limitations. An a priori power analysis was not conducted because of the practical limitations of animal availability. The dogs utilized in this research were uniquely accessible due to their transition from a preceding study, and these were the only group of subjects that fit the inclusion criteria that were available for this study. While this study was one of the largest studies of different foods on DCM biomarkers to date, a sample size of 60 may not be large enough to detect preclinical changes of DCM, which has a prevalence of approximately 0.5% in dogs of breeds who are not predisposed to DCM (Sisson D, 2000; Dukes-McEwan et al., 2003; Quest et al., 2022). Moreover, although all dogs had ad libitum access to water, we could not ensure that the dogs had a similar hydration status at each evaluation, and even mild fluid deficits can influence hemodynamic effects and affect echocardiographic results (Fine et al., 2010). In addition, our results may only be relevant to the specific nutritional formulation, processing method, and sources of ingredients evaluated in this study. Concentrations of antinutritional factors were similar across groups (Morris et al., 2023) and did not appear to affect outcomes, but other dog food formulations may contain different ingredients from a variety of sources and may be produced using different manufacturing processes or standards. Consequently, our results may not be representative of those observed with other dog foods that utilize alternative processing (e.g., freeze-dried or raw) or formulation standards.

Conclusion

After 18 mo, no significant changes in cardiac biomarkers were noted in dogs fed grain-free foods or grain-inclusive foods with or without peas, pea fibers, and potatoes. All evaluated foods appeared to have little to no impact on taurine concentrations or taurine excretion. No dogs showed any clinical evidence of DCM on echocardiograms, although the onset of CDVD was noted in several dogs, which was expected given its increasing prevalence in dogs with advancing age. Future research is needed to identify alternative variables and mechanisms that can influence the development and progression of DCM and to define the role of specific foods and the microbiome in altering the risk of DCM.

Glossary

Abbreviations

AE

adverse event

AOD

aortic diameter

CBC

complete blood count

CDVD

chronic degenerative valve disease

DCM

dilated cardiomyopathy

EF

ejection fraction

FDA

US Food and Drug Administration

FS

fractional shortening

G

grain-inclusive diet without peas or potatoes

G + PPF

grain-inclusive diet with peas and pea fiber

GF + Po

grain-free diet with potatoes

GF + PPo

grain-free diet with potatoes and peas

IVSd

interventricular septal thickness at end diastole

IVSs

interventricular septal thickness at end systole

LA

left atrial diameter

LA/Ao

left atrial to aortic diameter

LVIDd

left ventricular internal diameter at end diastole

LVIDs

left ventricular internal diameter at end systole

NT-proBNP

N-terminal end of the brain natriuretic peptide

PBD

pea-protein-based diet

PWd

left ventricular posterior wall dimensions at end diastole

PWs

left ventricular posterior wall dimensions at end systole

TCA

taurocholic acid

Conflicts of Interest Statement

All authors are current or past employees of Hill’s Pet Nutrition, Inc.

Author Contributions

All authors were involved in study design, data analysis and interpretation, and drafting and final approval of the manuscript. K.G. and C.S.: study design and execution. E.M., K.G., and L.H.: data analysis and interpretation. All: drafting & final approval of manuscript

Data Availability

The raw data supporting the conclusions of this article will be made available by the authors upon reasonable request.