Introduction

Diets composed of plant-based (PB) foods and devoid of animal products are receiving growing attention across the food chain [1,2]. Interest in extruded PB diets for pets is suspected to be due, in part, to the body of evidence in people demonstrating several health-promoting effects of PB diets. Specifically, dietary patterns characterized by an abundance of healthful plant ingredients have been associated in people with decreased risk of cardiovascular disease, type 2 diabetes, and some cancers [3–5]. PB dietary patterns have also been associated with reduced circulating levels of inflammatory biomarkers in people [6,7]. Other factors contributing to shifts toward plant-based diets globally include concerns surrounding the welfare of animals used for food production and the impact of animal agriculture on the environment [1,8].

PB ingredients have been used in commercially prepared pet diets since the extrusion process was first introduced into the pet food industry in the 1950s. Certain PB foods (e.g., rice, corn) serve as functional ingredients in the extrusion process in that they influence shape, texture, and expansion of kibble. The nutritive content and digestibility of PB ingredients commonly used in commercial pet foods have been studied extensively [9–13], and it is well-established that PB ingredients can serve as viable sources of energy, dietary fiber, digestible carbohydrates, and a variety of other essential and non-essential nutrients for dogs [2,14]. In recent years, consumer demand for alternative protein sources has prompted researchers to investigate amino acid (AA) digestibility and protein quality of several PB ingredients, such as pulses (e.g., beans, lentils), protein concentrates (e.g., pea protein, potato protein), and soy [15–18]. Using precision-fed rooster assays, Reilly and colleagues demonstrated high digestibility of all the indispensable AAs, except methionine, in selected pulse ingredients and high digestibility of all the indispensable AAs in PB protein concentrates [15,16]. Protein quality, assessed by digestible indispensable amino acid scores (DIAAS), indicated methionine was the first-limiting AA in all the pulse ingredients and in all but one of the protein concentrates tested [15,16]. These findings suggest some PB ingredients are viable protein sources for canine and feline diets and emphasize the importance of inclusion of complementary proteins when formulating diets rich in certain PB ingredients.

Despite recommendations and regulations for nutrient content in the global pet food industry, AA deficiencies are still detected in some dogs consuming commercially prepared diets [19–22]. A well-documented cause of myocardial dysfunction in dogs and cats is a deficiency of the sulfur-containing AA, taurine. Dogs with taurine-deficient cardiomyopathy or other nutritionally-related myocardial dysfunction often begin to show improvement in their echocardiographic parameters 12 weeks after treatment or resolution of the underlying cause, but less is known about the time required for nutritionally-related cardiomyopathy to appear echocardiographically in dogs [21–23]. Unlike cats, dogs can synthesize taurine from dietary methionine and cystine; therefore, taurine supplementation has been thought to be unnecessary in dogs who consume a complete and balanced diet. However, taurine deficiency has been documented in dogs fed commercial diets that have been formulated to meet the requirements of Association of American Feed Control Officials (AAFCO) Dog Food Nutrient Profiles [19–22]. Potential explanations include breed-related differences in minimum methionine requirements [24], which are not accounted for by AAFCO, and non-nutrient factors, such as feeding protocols of pet owners, which directly impact caloric intake and nutritional status of pets. More recently, concerns surrounding a possible association between certain commercial diets and dilated cardiomyopathy (DCM)-like disease in dogs have been accumulating [22,23,25–27], and many of the implicated diets have a nutritional adequacy claim stating the diet has been formulated to meet the requirements of AAFCO Dog Food Nutrient Profiles. The discrepancies between nutritional adequacy claims on pet food labels and the health status of dogs consuming the diets underscore the importance of prospective in vivo evaluation of commercially available pet foods.

To date, there are no studies evaluating the effects of a PB diet on any naturally-occurring chronic disease in dogs. Moreover, clinical studies assessing for potential negative effects of commercial extruded PB diets on general and/or cardiac health of dogs are lacking. The objectives of this prospective interventional study were to assess the short-term AA, clinicopathologic, and echocardiographic findings in healthy adult dogs fed a commercial extruded PB diet (PBD) in which pea protein was the primary protein source and to compare the post-PBD values to baseline values and to control dogs fed a commercial extruded traditional diet (TD). We hypothesized circulating AA levels, clinicopathologic, and echocardiographic findings of dogs fed a PBD would be similar to baseline and to healthy controls.

Materials and methods

This study was conducted under a Ross University School of Veterinary Medicine (RUSVM) Institutional Animal Care and Use Committee (IACUC) approved protocol (18.05.19) with all dog owners providing written consent prior to inclusion of their pet(s) in the study.

Results

Amino acids

Twenty-nine AAs were assayed. AAs were assessed for normality and the majority were judged to not be normally distributed. Those that were normally distributed were: alanine, arginine, histidine, and tyrosine. AA results for both groups of dogs are presented (Table 1). In the control group, one essential AA, histidine, was higher at 4 weeks compared to baseline (P = .017). In the PB diet group, there was statistical evidence of change in the 10 essential AAs and in 12 of the non-essential AAs. All essential AAs, except methionine, were higher after 4 weeks on the PBD. Seven of the non-essential AA, including taurine (plasma and whole blood), were higher after 4 weeks on the PB diet while 5 non-essential AAs were lower after 4 weeks on the PB diet. There were no differences in AA concentrations between the 2 groups of dogs at baseline. At 4 weeks, there were meaningful differences in 5 of the AAs between the 2 groups; cystathionine and whole blood taurine were higher in the PBD group and leucine, methylhistidine-3, and serine were lower in the PBD group. Established reference intervals were available for 25 of the AA assays [30]. Post-PBD AA values, including those that decreased, were within or above the reference intervals, except glutamine and cystine.

Download:

• PPT
  PowerPoint slide
• PNG
  larger image
• TIFF
  original image

Table 1. Plasma amino acid and whole blood taurine concentrations in dogs fed a traditional diet or a PB diet for 4 weeks.

https://doi.org/10.1371/journal.pone.0258044.t001

Body weight, BCS, hematology, serum biochemistry, and urinalysis

Baseline and post-diet body weight and select hematologic, serum biochemical, and UA data from PBD dogs and from comparison dogs are summarized (Table 2). At the 12-week visit, 2 of the PB diet dogs did not have urine samples available; therefore, data from 28 dogs were included in the analysis. There were no substantial changes in body weight or BCS over the 12-week study period in either group of dogs, and there were no meaningful differences in body weight or BCS between the PB diet dogs and the control dogs at baseline or at 12 weeks. Further, no evidence of nutritional deficiency (e.g., poor haircoat, muscle loss) was detected on physical examination of the dogs at any of the timepoints (baseline, 1 week, 4 weeks, 12 weeks). Median hematologic and biochemical values for both groups of dogs were within normal limits at baseline and at 12 weeks. In the control group, hemoglobin was lower at 12 weeks compared to baseline (P = .003). In the PB diet group, ALP (P = .047) and urine pH (P = .005) were higher at 12 weeks compared to baseline. There were no meaningful hematologic or biochemical differences between the 2 groups at baseline. At 12 weeks, the urine pH of the PB diet group was higher than the urine pH comparison group (P = .022). At 12 weeks, two dogs (7%) on the PB diet had an alkaline phosphatase (ALP) concentration that was above the reference interval and 1 dog (3%) had an alanine aminotransferase (ALT) concentration above the reference interval.

Download:

• PPT
  PowerPoint slide
• PNG
  larger image
• TIFF
  original image

Table 2. Body weight and select hematologic and serum biochemical data in dogs at baseline and after being fed a traditional diet or a plant-based diet for 12 weeks.

https://doi.org/10.1371/journal.pone.0258044.t002

Echocardiographic findings

Echocardiographic data for all dogs is summarized (Table 3). There were no substantial changes over the 12-week study period in the control group. In the PBD group, LVIDd was higher 12 weeks post-diet compared to pre-diet (P < .001). Normalized LVIDd was also higher post-diet (P = .031). There were no meaningful changes in LVIDs (P = .138), normalized LVIDs (P = .317), or FS (P = .411) 12 weeks post-PB diet. There was no statistical evidence of difference between the 2 groups of dogs for any of the echocardiographic parameters at baseline or at 12 weeks.

Download:

• PPT
  PowerPoint slide
• PNG
  larger image
• TIFF
  original image

Table 3. Left ventricular echocardiographic findings in dogs receiving a traditional diet or a plant-based diet for 12 weeks.

https://doi.org/10.1371/journal.pone.0258044.t003

Discussion

In this prospective diet study, several changes were observed in the AA profiles of 30 healthy adult dogs after consuming a commercial extruded PBD for 4 weeks. On the basis of the PBD containing a viable primary protein source (pea protein), we hypothesized AA concentrations in the PBD group would be similar to baseline and the control group, which was not the case in this cohort of client-owned dogs. In addition, after 12 weeks on a PBD, clinicopathologic and echocardiographic testing in the dogs demonstrated changes in some of the variables.

The changes in essential AAs, with 9 of 10 being higher at 4 weeks compared to baseline, may be explained in part, by differences in the nutrient content of the baseline (traditional) diet and the PBD (S1 Appendix). Methionine, for example, which was lower in the PBD group compared to baseline, was lower in the nutrient analysis of the PBD compared to the TD. Taurine also was lower in the PBD nutrient analysis; however, plasma and whole blood taurine were significantly higher in the PBD group compared to baseline. Whole blood taurine was also higher in the PBD group compared to controls. While there are no studies that report taurine or other AA concentrations in dogs fed a PBD to compare our results to, there are studies that have investigated the impact of certain plant ingredients on taurine and AA concentrations in dogs and cats. In a study of healthy dogs fed a purified low protein diet, beet pulp, a natural source of dietary fiber, was shown to decrease protein digestibility and increase bile acid excretion thereby reducing circulating taurine levels [31]. In a different study, dogs fed commercially prepared traditional diets containing whole grain brown rice as the first plant ingredient had significantly lower whole blood taurine concentrations than dogs fed a diet with multiple plant protein sources [30]. Healthy cats fed a purified diet containing a high concentration of rice bran developed critically low plasma and whole blood taurine levels compared to controls [32]. Recently, pulses (e.g., peas, lentils), which are limiting in methionine and abundant in fermentable fiber, have been speculated to be associated with some cases of taurine-deficient or other nutritionally-related cardiomyopathy in dogs [22,27,33]. The PBD used in this study did not contain beet pulp or rice bran, but it did contain peas, pea protein, and brown rice as the first 3 ingredients. In addition, the PBD was supplemented with taurine as was the TD; therefore, it is possible the bioavailability of sulfur-containing AAs in either diet was suboptimal and was offset by taurine supplementation. Nevertheless, our finding of higher whole blood taurine levels in dogs fed a PBD for 4 weeks compared to baseline and to controls suggests diets containing only plant ingredients can be formulated and processed to achieve taurine and other nutrient targets in dogs.

Although clinical evidence of amino acid deficiency was not observed in the dogs in this study, two dispensable AAs, glutamine and cystine, were below the established reference intervals in both groups of dogs at 4 weeks. At baseline, however, glutamine concentrations in both groups were above the reference interval. Being that glutamine is a major source of nitrogen for purine synthesis, glutamine levels can fluctuate widely depending on the nitrogen status of the animal [34]. Thus, the plasma glutamine concentrations reported here may simply reflect normal interday variation. Blood collection, sample preparation, and laboratory analysis were standardized; therefore, errors related to sampling and/or analysis are an unlikely cause of the low glutamine levels but cannot be definitively excluded. Cystine was low at baseline and at 4 weeks in both groups of dogs. Protein-bound cystine is lost when plasma proteins are removed before amino acid analysis unless the sample is deproteinized with 6% sulfosalicylic acid [35]. Our samples were not deproteinized prior to shipment to the AA lab; therefore, the low cystine levels in this study were likely due to storage loss during or after shipment.

On the basis that this was an exploratory study, and each test of AA was considered independent of the other, multiplicity adjustments were not made. Being that AA metabolism is influenced by other AAs, statistical significance of differences in the AAs observed in this study must be interpreted with caution. Further, while changes were seen in several of the AA values, interpreting clinical relevance, particularly in regard to the reference intervals, is challenging. The published reference intervals for AA concentrations in dogs are from a single time point [30], and it is unknown whether meaningful day-to-day variations exist in dogs under normal conditions. Gray and colleagues evaluated the effect of 48-hour fasting on AA status in dogs and found that plasma methionine changed 43.7% from baseline after just 12 hours of fasting yet remained within the established reference interval [36]. Intraday variation in AA levels, in association with exercise and diet, has also been observed in dogs [37]. It is possible the fluctuations observed in the AA concentrations of our subjects represent normal interday variation as seen in other species [38,39]. Long-term, randomized, controlled diet trials with AA concentrations from multiple time points would provide valuable information regarding potential normal AA variations in dogs and assist in interpreting the impact of diet.

Vomiting and diarrhea are common clinical signs of primary gastrointestinal (GI) disease in dogs and often occur secondary to non-GI disease (e.g., renal, hepatic) as well which makes them practical and useful signs for monitoring by owners and caretakers [40,41]. The percentage of dogs on the PB diet who had a vomiting or diarrhea event during the 12-week study period (47%) was similar to the comparison group (50%). While 29% of the events in the PB diet group were accompanied by a logical explanation (e.g., known dietary indiscretion), many of the events did not have an explanation. Because the dogs in this study were client-owned and had no environmental or social restrictions, it is possible that 1 or more of the events had an infectious, allergic, or toxic etiology. Nevertheless, the events were transient and did not require medical intervention. We cannot exclude a PBD-related etiology; however, this is considered less likely since a greater percentage of dogs not on the PBD had vomiting or diarrhea at some point during the study period.

Baseline CBC, serum biochemistry, and UA were performed to confirm the health status of the dogs at the time of study inclusion and to provide data for comparison at the end of the 12-week study period. The authors recognize this level of pre- and post-diet testing may be cost prohibitive in other settings (e.g., clinical practice, pet food feeding trials); therefore, using AAFCO post-diet testing requirements [42] as a guide, an abbreviated panel of parameters was created (Table 2). The 2 dogs on the PBD with elevated ALP at the 12-week assessment had no clinical signs and were not taking any medications. The owner of the dog with elevated ALT at the 12-week assessment noted the dog recently had 2 episodes of urinary incontinence, and the dog had a new dermal mass on physical exam. Of the 2 dogs with elevated ALP, 1 had normal ALP at recheck 4 months later and the dog was still being fed the PBD at that time, which suggests the elevation in ALP was likely not diet-related, and the other dog with elevated ALP was lost to follow-up. The dog with elevated ALT had normal ALT on routine bloodwork 10 months later. The dog was no longer consuming the PBD at the time of recheck ALT; therefore, a diet-related cause of increased ALT cannot be excluded. The increase in urine pH in the PB diet group may be attributed to the lack of animal protein and/or abundance of vegetables in the diet [43]. PB diets in people have been shown to increase urine pH and reduce the risk of certain urinary diseases [44–46], and studies in dogs are needed.

Echocardiography was incorporated into the diagnostic plan of this diet study to help provide a comprehensive health assessment of dogs fed a diet for which there are no prior clinical studies. Evidence of overt left ventricular systolic dysfunction was not observed in dogs fed a PBD for 12 weeks. Although increases in LVIDd and LVIDs reached meaningfully different levels at 12 weeks, the observed changes ranged from 1–12% which is within established interday variation of left ventricular echocardiographic parameters in dogs [47]. To the authors’ knowledge, data on minimum or maximum amount of time for diet-induced myocardial dysfunction to develop in dogs do not exist. In a rat model of a vitamin D deficient diet, echocardiographic evidence of cardiac remodeling was detectable at 20 weeks; however, no earlier timepoints were included in that study [48]. The decision to reassess echocardiograms of dogs in this study at 12 weeks was largely based on knowledge that some dogs suffering from nutritionally-related cardiomyopathy begin to show improvement in their echocardiographic parameters as soon as 12 weeks after treatment or resolution of the underlying cause [21–23]. Early surveillance was also chosen in the interest of animal welfare.

A limitation of this study was lack of a randomized control group. The non-randomized group of 4 control dogs and the pre-diet testing provided valuable insight in some circumstances (e.g., cystine concentrations, clinical signs), but neither replaces the verification of observations provided by a randomized control group. Although most of the dogs were consuming the same brand of TD, only a third of dogs were consuming the same variety of food, and this may have impacted AA and other results. The nutrient content of the PB treats was not determined and treat intake was not standardized, which could have led to increased variability of nutrients and reduced our ability to detect differences among the 2 diet groups. Nutrient content of the diets was provided by manufacturers instead of analysis by the researchers. Dog owners and investigators who examined the dogs were not blinded, which could have resulted in reporting bias. A final limitation was the relatively short study period (12 weeks).

The study reported here demonstrated a cohort of 30 healthy adult and predominantly mixed breed dogs who consumed a commercial PBD for 4 weeks underwent changes to their AA profiles. Some of the changes in essential AAs were attributed to differences in AA content of the PBD compared to the AA content of a TD. Additionally, other dietary factors, such as amount, frequency, and global nutrient composition, may have impacted the AA concentrations of the dogs via effects on protein turnover and de novo AA synthesis. Dogs on the PBD did not develop any essential AA deficiencies or taurine deficiency, rather the PBD dogs had taurine concentrations higher than their pre-PBD values and higher than the whole blood taurine concentrations of the control dogs. Results of clinicopathologic and echocardiographic testing indicated the health of the dogs was maintained and did not deteriorate after 12 weeks on a PBD.

Our study evaluated a single commercial PBD; therefore, the results cannot be extrapolated to all PB diets since composition and nutrient availability vary based on formulation and processing. The study reported here supports the need for prospective in vivo studies of animal diets and suggests that client-owned pets may be an alternative to traditional study subjects. It should be emphasized that the findings of this study cannot be applied to other species (e.g., cats) nor should veterinary medical professionals or pet owners speculate about the long-term health effects of a PBD based on the results of this short-term study. Further research is needed to elucidate the potential long-term effects of PB diets on health and disease in dogs.

Supporting information

Acknowledgments

The authors thank Melody Montes for technical assistance, and Mariette Williams and Silvia Marchi for laboratory assistance. Portions of the results of this study were presented as an abstract at the 2019 American College of Veterinary Internal Medicine Forum.