Evaluation of two raw diets vs a commercial cooked diet on feline growth

Abstract

Objectives

The objective of this study was to determine if two raw feline diets were nutritionally adequate for kittens.

Methods

Twenty-four 9-week-old kittens underwent an Association of American Feed Control Officials’ (AAFCO) 10 week growth feeding trial with two raw diet groups and one cooked diet group (eight kittens in each). Morphometric measurements (weight, height and length), complete blood counts, serum chemistry, whole blood taurine and fecal cultures were evaluated.

Results

Overall, the growth parameters were similar for all diet groups, indicating the two raw diets used in this study supported feline growth, within the limitations of an AAFCO growth feeding trial. Kittens fed the raw diets had lower albumin (P = 0.010) and higher globulin (P = 0.04) levels than the kittens fed the cooked diet. These lower albumin levels were not clinically significant, as all groups were still within normal age reference intervals. A red cell microcytosis (P = 0.001) was noted in the combination raw diet group. Increases in fecal Clostridium perfringens were noted in all groups, along with positive fecal Salmonella serovar Heidelberg and Clostridium difficile toxin in the combination raw diet group.

Conclusions and relevance

The majority of the parameters for feline growth were similar among all groups, indicating the two raw diets studied passed an AAFCO growth trial. In theory, it is possible to pass an AAFCO growth trial but still have nutrient deficiencies in the long term due to liver and fat storage depots. Some of the raw feeders had elevated globulin and microcytosis, likely associated with known enteropathogenic exposure. Disease risks to both pets and owners are obvious. Additional research in this area is needed to investigate the impact of raw diets on the health of domestic cats.

Introduction

Feeding of raw food diets to both cats and dogs has been an area of ongoing controversy and debate among veterinarians, breeders and owners. Proponents argue that raw meat diets are the ancestral food of dogs and cats, and these diets represent the optimal nutritional profile for health and longevity. Proponents describe numerous benefits, including increased digestibility, improved skin and coat quality, improved stool characteristics, enhanced immune function and decreased incidence of many diseases (ie, obesity, diabetes mellitus, allergies, feline lower urinary tract disease, arthritis and cancer).^1,2 All of these claims are based on anecdotal evidence compared to scientific studies. In contrast, opponents cite several peer-reviewed studies addressing food safety and public health concerns,^3–6 along with questions on whether these diets are nutritionally balanced. ⁷ Cats, as obligate carnivores, obtain all their necessary nutrients from animal tissues.

Few studies have been carried out examining growth and blood parameters in cooked vs raw fed domestic felids. One study evaluated feeding whole ground raw rabbit vs a cooked diet to kittens over a 13 month period (AG Glasgow, NJ Cave, SL Marks, et al, unpublished data [http://www.vetmed.ucdavis.edu/ccah/local-assets/pdfs/Role_of_diet_feline%20health_Glasgow.pdf]). The purpose of that study was to define a ‘gold standard’ diet for cats that could subsequently be used for comparison with commercial diets. The growth curve of the kittens on the raw and cooked diets was identical, indicating the raw rabbit diet supported normal growth but was not significantly better or worse than the cooked diet. Positive aspects of the raw diet included improved stool quality, coat quality and positive palatability. Although the raw rabbit diet outwardly appeared to be nutritionally adequate for the kittens, one of the kittens died suddenly and was later diagnosed with dilated cardiomyopathy secondary to taurine deficiency. All of the other kittens on the raw rabbit diet were subsequently evaluated and 70% were found to have heart muscle changes compatible with taurine deficiency. For the remaining months of the study, the raw rabbit diet was supplemented with taurine and taurine levels in the kittens returned to normal.

Nutritional adequacy claims for commercial pet foods are based on procedures and protocols established by the Association of American Feed Control Officials (AAFCO). ⁸ One method of establishing an adequacy claim in the USA is to conduct a feeding trial following AAFCO-established protocols. Adequacy claims can also be made by comparing the calculated nutrient content of diet formulas with a nutrient data bank, or actual chemical analysis of the diets, with minimum AAFCO guidelines. The pet food industry uses ingredients that can have a wide variation in actual nutrient content and bioavailability. Therefore, feeding protocols are considered the ‘gold standard’ in evaluating pet food for nutritional adequacy and performance. The protocol for following an AAFCO kitten growth trial requires a minimum of eight 9-week-old kittens in both the test diet group and control group. The control group must have previously passed an AAFCO growth trial. The kittens are fed the diets exclusively for 10 weeks with weight, hemoglobin, hematocrit, whole blood taurine and serum albumin measured before and at the end of the 10 week trial. The test diet passes if the average gain is similar to the control diet group and week 10 average laboratory values are equal to or greater than hemoglobin 10 g/dl, packed cell volume (PCV) or hematocrit 29%, albumin 2.7 g/dl and taurine 300 nmol/ml.

In our study, an AAFCO growth protocol was used to compare a high protein (>40% dry matter [DM]) cooked canned kitten diet with two raw meat diets for nutritional adequacy in kittens. Morphometric measurements of weight, length and height were measured in addition to blood parameters (complete blood count [CBC], complete serum chemistry and whole blood taurine). The primary goal was to determine if two raw food diets, a commercial complete raw diet and a combination raw diet in which a commercial supplement is mixed with raw meat obtained by the owner and intended to be complete and balanced, were adequate for passing an AAFCO growth claim. It is well known that raw meat may be contaminated with a variety of pathogenic microbes. ⁶ While this study was not primarily a food safety study, periodic monitoring of feces for pathogenic bacteria was felt to be necessary owing to the potential effects of pathogens on growth and blood parameters in the kittens. Our hypotheses were that the raw food diets would pass an AAFCO growth feeding trial, and there would be no differences in morphometric measurements or blood parameters but probable that fecal enteropathogens would be detected.

Materials and methods

All kittens were born and raised at The University of Tennessee Veterinary Medical Center and research facility, and their care was in compliance with the Guide for the Use and Care of Laboratory Animals. The experimental protocol was approved by the Institutional Animal Care and Use Committee.

Animals

Twenty-four domestic shorthair kittens born over a 3 year period were used for the growth feeding trial. All kittens were born from the same queen and tom. The queen was bred each year to have 1–2 litters per year. Subsequently, five litters of kittens were run through a feeding trial sequentially (total of five feeding trials completed) over a 3 year period. Kittens were weaned at 6–7 weeks of age and were eating solid food prior to beginning the feeding trial at 9 weeks of age. All kittens were weaned on both a canned (Science Diet Kitten Healthy Development Liver and Chicken entrée minced; Hill’s Pet Nutrition) and dry extruded growth diet (Science Diet Kitten Healthy Development Original dry cat food; Hill’s Pet Nutrition). The kittens were given physical examinations, were deemed healthy and were negative for intestinal and external parasites 1–3 days prior to starting the feeding trial. All kittens were housed in individual metabolism cages during the feeding trial and allowed daily group play. At the end of the study, the kittens were transferred to a permanent feline colony that is used for dietary and other non-invasive research, or adopted out to private homes.

Diet

The kittens were randomized to three different dietary groups using a random numbers table. Each diet group consisted of eight kittens. Diet group A (five males, three females) was fed a cooked food (Evo Turkey and Chicken canned cat and kitten cat food; Natura Pet Products). This diet was a canned food and was chosen to match closely the commercial raw diet moisture and protein content. Treatment group B (five males, three females) was fed a commercial frozen raw diet (Chicken and Clam Frozen Raw Diet; Wild Kitty Raw All Natural Cat Food). Both diet A and B had the following AAFCO statement attached: ‘formulated to meet nutritional levels established by AAFCO for all cat life stages’. Treatment group C (six males, two females) was fed a combination raw food diet. This diet was made with raw chicken (boneless, skinless chicken breast; Tyson Foods) obtained from a local grocery store and mixed, according to manufacturer’s instructions, with a commercial supplement (TC Feline Plus Cat Food Premix with beef liver; TC Feline) designed to balance a raw meat diet. Although no nutrient analysis of this diet when mixed with raw meat was available, it was chosen because of its popularity and the limited number of these types of combination raw diets available for feline owners. This diet had no AAFCO statement of nutritional adequacy. A composite sample of each diet was submitted to Eurofin Laboratory (Des Moines, IA, USA) for proximate analysis of moisture, crude protein, crude fat, ash, crude fiber, taurine and calories. Table 1 is the proximate analysis of each diet on an energy basis (g nutrient/1000 kcal) along with the ingredient lists. The combination raw diet was prepared every 2 weeks and immediately frozen. All raw foods were kept frozen until 1 day before feeding, when they were transferred to a refrigerator in preparation for feeding the next day. Three times daily, kittens were offered their respective diet at three times the resting energy requirement (RER) for the first 4 weeks of the feeding trial (kittens 9–13 weeks of age) and 2.5 times the RER from weeks 5 to 10 of the feeding trial (kittens 13–19 weeks of age), to ensure adequate food intake. In accordance with AAFCO protocols, the length of the feeding trial was 10 weeks. Water was available at all times.

Table 1.

Calculated dietary nutrient composition (g/1000 kcal) and metabolizable energy (ME)

Item	Cooked*	Commercial raw †	Combination raw ‡
ME (kcal/g)	1.46	1.56	0.95
Protein	89.5	93.3	176.4
Fat	72.5	64.2	27.5
Digestible carbohydrate	7.5	13.7	–
Taurine	0.69	0.67	2.01
Moisture (%)	71.8	70.7	78.6

^*Ingredients: turkey, chicken, turkey broth, chicken broth, chicken meal, herring, carrots, whole egg, salmon meal, natural flavor, carrageenan, tomato flakes, cottage cheese, L-ascorbyl-2-polyphosphate, apples, guar gum, vitamin E supplement, vitamin A supplement, vitamin D3 supplement, vitamin B12 supplement, thiamine mononitrate, niacin supplement, d-calcium pantothenate, pyridoxine hydrochloride, riboflavin supplement, folic acid, biotin, zinc amino acid chelate, cobalt amino acid chelate, copper amino acid chelate, manganese amino acid chelate, potassium iodide, inulin, herring oil, choline chloride, potassium chloride, salt, sunflower oil, taurine, sodium phosphate, beta carotene

^†Ingredients: free range organic chicken, apples, Atlantic clams, beets, broccoli, carrots, chicken hearts, chicken liver, cod liver oil, dried kelp, dried yeast, flax, flax seed, lecithin, mushrooms, water sufficient for processing, oysters, peas, rice bran, spinach, wheat germ, wheat germ oil

^‡Ingredients: skinless chicken breast, supplement (freeze-dried bovine bone [New Zealand], egg yolk, whey protein isolate, beef liver, freeze-dried krill, taurine, cellulose, kelp, vitamin E, vitamin D3, vitamin A, vitamin B complex)

Food intake

To determine food intake, all food was weighed before and after each offering and the difference determined. The majority of food was eaten within 15–20 mins of presentation. Any food not ingested within 4 h of presentation was removed to prevent microbial contamination or proliferation. Previous studies using the same rooms showed evaporation rates of <5% in 24 h so any water loss was of minor consideration (C Kirk, 2015, personal communication). Feeding bowls were cleaned between feedings, using hot water and commercial dishwashing liquid.

Morphometric measurements

Body weight, body condition score (BCS), height (from bottom of the foot to top of scapula) and length (from tip of nose to base of tail) were recorded weekly by the same individual (BAH). Body condition scoring is a process used to asses an animal’s fat stores and, to a lesser extent, muscle mass. ⁹ Fat cover is evaluated over the ribs, down the topline, around the tailbase and ventrally along the abdomen. Body condition scoring was performed using the nine-point scale with 5/9 being ideal weight, 1/9 being very thin and 9/9 being grossly obese. ⁹ In accordance with AAFCO growth protocols, the average body weight gain was not to be less than the average for the concurrent control group, minus the allowance for normal variation over the 10 week period. ⁸

Nutritional adequacy

An AAFCO growth protocol for kittens requires hemoglobin, PCV, whole blood taurine and serum albumin be measured and recorded at the end of the feeding trial. The average final hemoglobin, PCV, whole blood taurine and serum albumin concentrations should not be less than as follows: hemoglobin, 10.0 g/dl (no individual <8.0 g/dl); PCV, 29% (no individual <26%); taurine, 300 nmol/ml (no individual <200 nmol/ml); albumin, 2.7 g/dl (no individual <2.4 g/dl); or the average for the concurrent control group minus the allowance for normal variation over the 10 week period. ⁸ In accordance with AAFCO protocols, these hematologic variables along with a CBC and plasma biochemical analysis were assessed to determine nutritional adequacy. Whole blood (2.0 ml) was collected by jugular venepuncture at weeks 0, 5 and 10 of the study. Of the 2.0 ml, 0.5 ml blood was collected in EDTA and 0.5 ml was collected in sodium heparin and submitted for a CBC and complete chemistry analysis at the University of Tennessee Veterinary Medical Center Clinical Pathology Laboratory. The remaining 1 ml was placed in a sodium heparin tube and submitted to the Amino Acid Analysis Laboratory at the University of California Davis for whole blood taurine analysis. The CBC and complete chemistry analyses were run on the same day as collection. The whole blood taurine analysis was completed within 48 h of submission to the Amino Acid Analysis Laboratory.

Fecal cultures and fecal scoring

Composite feces for each of the five feeding trial groups were cultured for Salmonella, Campylobacter and Clostridium species at weeks 0, 5 and 10 of the feeding trial, with the exception of the first two feeding trial groups when only week 0 and week 10 composite fecals were cultured. Fecal cultures were performed at the University of Tennessee Veterinary Medical Center Clinical Bacteriology and Mycology Laboratory (UTVMCCBML). These composite cultures were not separated by diet. If a composite culture was positive for Salmonella, Campylobacter or Clostridium species, all kittens in that group had individual fecal cultures submitted to determine the source. Feces were scored and recorded daily by the same individual (BAH) during the 10 week feeding trials. Scoring was carried out using a five-point scale as follows: 1 = watery, liquid that can be poured; 2 = soft, unformed stool; 3 = soft, moist, formed stool; 4 = dry, well-formed stool; 5 = hard, dry pellets. In addition, if individual kittens developed fecal scores of 2 or less for more than 2 days, individual fecal cultures were submitted for Salmonella, Campylobacter and Clostridium species testing at the UTVMCCBML.

Food cultures

If any subsequent group or individual fecal cultures were positive for Salmonella or Campylobacter species, samples of all three diets were cultured for Salmonella and Campylobacter species by the UTVMCCBML.

Statistical analysis

Twenty-four kittens were randomly assigned to the three diet treatments with eight kittens per treatment, with each evaluated at 9, 14 and 19 weeks of age (weeks 0, 5 and 10 of feeding trial). A completely randomized design was used to compare means for kitten weight, height, length, serum chemistry variables, whole blood taurine and CBC. A mixed-model ANOVA (SAS, Version 9.2) included diet, repeated measures over weeks and the interaction. Models for weight, height, length and BCS also included sex. Least squares means were compared using Fisher’s least squares difference. The significance level was defined as P ⩽0.05, while 0.10 was defined as a trend. Normality and equal variance of residuals were examined for all dependent variables. No variables failed to meet these assumptions.

Results

Food intake, BCS and morphometric measurements

Table 2 lists the average DM intake (g/day), DM protein intake (g/day) and caloric intake (kcal/day) per day, along with mean BCS by diet treatment and sex. Males on the commercial raw diet had a trend toward less DM intake (P = 0.085) and less kcal (P = 0.065) per day than the other two diet groups. Both males and females on the combination raw diet had significantly more protein (P <0.0001) per day than the other two diet treatments owing to the high level of protein in this diet. Overall, there were no differences in BCS between the three diet groups and between males and females in each diet group.

Table 2.

Caloric, dry matter (DM) and protein DM intake per day, and body condition score (BCS) by diet treatment and sex

Total	Cooked	Commercial raw	Combination raw
kcal/day
Male	352.13 ± 17.40 a	284.08 ± 17.40a,b	340.60 ± 15.88 a
Female	301.48 ± 22.46a,b	268.30 ± 22.46 b	268.71 ± 27.51 b
DM g/day
Male	56.37 ± 2.91a,b	45.41 ± 2.91 c	62.65 ± 2.65 a
Female	48.29 ± 3.75b,c	42.45 ± 3.75 c	49.42 ± 4.60b,c
Protein DM g/day
Male	26.32 ± 1.99 c	22.57 ± 1.99 c	48.99 ± 1.82 a
Female	22.55 ± 2.57 c	24.76 ± 2.57 c	38.65 ± 3.15 b
BCS (1–9 scale)
Male	5.47 ± 0.14a,b	5.15 ± 0.14a,b	5.54 ± 0.13 a
Female	5.06 ± 0.18 b	5.03 ± 0.18 b	5.55 ± 0.22a,b

Data are ± SEM. Different letter superscripts in both rows and columns indicate P values ⩽0.05

Figure 1 illustrates mean weight gains in males vs females in all the treatment groups over the 10 week feeding trial. A significant difference in weight was found between the sexes (P = 0.025), with males having higher weight gains than females.

Figure 1.

Mean weekly kitten weight by sex (mean ± SEM)

Figure 2 illustrates weight gain for males and females by diet treatment. As the feeding trial progressed (weeks 2–10), males on the combination raw diet consistently had between 100 and 200 g higher weights. Male kittens on the commercial raw diet had initial higher weight gains compared with the males of the cooked diet (between 25 g and 100 g), but these differences ceased after week 5. None of these weight differences between the three treatments were statistically significant.

Figure 2.

(a) Male and (b) female kitten mean weekly weight by diet treatment (mean ± SEM)

The female kittens fed the cooked diet showed the greatest weight gains during the second half of the feeding trial. None of these weight differences were statistically significant between the three diet treatments. Both of the raw diets passed AAFCO guidelines for a growth claim consistent with their weights being equal or not less than 10% of the recipients of the cooked diet.

No significant differences in height or length were seen between the different diet treatments. See supplementary material for height and length figures (Figures 1–3 in the supplementary material).

CBCs

Hematologic mean values for the three treatment groups at weeks 0, 5 and 10 are shown in Table 3. Hemoglobin at weeks 5 and 10 was significantly higher in both the cooked and commercial raw group compared with week 0 (P = 0.0013), while the combination raw diet group saw minimal hemoglobin change over the 10 week period. Hematocrit was also significantly higher at weeks 5 and 10 compared with week 0 in the cooked group, while neither raw diet group saw significant changes in hematocrit over the 10 week period (P = 0.0012). Red cell mean cell volume (MCV) in the combination raw diet was significantly lower at weeks 5 and 10 compared with week 0 (P = 0.001), and was significantly lower at week 10 compared with the other two diets (P = 0.0009). There were no statistically significant differences in white blood cell, that is, neutrophil, monocyte or eosinophil, numbers between the three treatment groups. As mean hemoglobin levels were 10 g/dl and mean hematocrit levels were ⩾29%, both raw diets fulfilled this AAFCO requirements for providing a growth claim.

Table 3.

Hematology values by diet treatment and week

Analyte/week	Cooked	Commercial raw	Combination raw	Reference (mean ± 1 SEM) 10
Hemoglobin (g/dl)
0	10.28 ± 0.39 d	10.86 ± 0.39c,d	10.81 ± 0.39c,d	9.8 ± 0.2
5	11.52 ± 0.39a,b,c	12.18 ± 0.39 a	11.35 ± 0.39a,b,c,d	10.1 ± 0.3
10	11.94 ± 0.39a,b	11.99 ± 0.39a,b	10.86 ± 0.39b,c,d	11.0 ± 0.4
Hematocrit (%)
0	31.30 ± 1.13 d	33.09 ± 1.13b,c,d	32.10 ± 1.13c,d	33.3 ± 0.7
5	35.41 ± 1.13a,b	36.00 ± 1.13a,b	34.61 ± 1.13a,b,c	33.1 ± 1.6
10	36.46 ± 1.13 a	35.51 ± 1.13a,b	33.51 ± 1.13a,b,c,d	34.9 ± 1.1
MCV (fl)
0	43.62 ± 0.56a,b,c	44.89 ± 0.56 a	44.91 ± 0.56 a	47.8 ± 0.9
5	42.99 ± 0.56b,c	43.81 ± 0.56a,b,c	42.47 ± 0.56 c	44.5 ± 1.8
10	44.15 ± 0.56a,b	44.65 ± 0.56 a	39.54 ± 0.56 d	43.1 ± 1.5
MCHC (g/dl)
0	33.07 ± 0.36 a	32.84 ± 0.36 a	33.60 ± 0.36 a	29.5 ± 0.4
5	32.57 ± 0.36 a	33.81 ± 0.36 a	32.84 ± 0.36 a	31.3 ± 0.9
10	32.71 ± 0.36 a	33.75 ± 0.36 a	32.36 ± 0.36 a	31.6 ± 0.8
Total WBC‡
0	11.40 ± 1.19 a	10.44 ± 1.19 a	10.89 ± 1.19 a	ND
5	11.19 ± 1.19 a	12.01 ± 1.19 a	10.72 ± 1.19 a	ND
10	11.575 ± 1.187 a	10.89 ± 1.19 a	13.650 ± 1.19 a	ND
Absolute neutrophils (× 103/mm3)
0	5.64 ± 0.64 a	4.90 ± 0.64 a	5.44 ± 0.64 a	ND
5	4.64 ± 0.64 a	6.07 ± 0.64 a	5.15 ± 0.64 a	ND
10	4.09 ± 0.64 a	4.45 ± 0.64 a	4.92 ± 0.64 a	ND
Absolute monocytes (× 103/mm3)
0	0.315 ± 0.073 a	0.304 ± 0.073 a	0.367 ± 0.073 a	ND
5	0.242 ± 0.073 a	0.246 ± 0.073 a	0.378 ± 0.073 a	ND
10	0.200 ± 0.073 a	0.278 ± 0.073 a	0.396 ± 0.073 a	ND
Absolute eosinophils (× 103/mm3)
0	0.551 ± 0.090 a	0.544 ± 0.090 a	0.574 ± 0.090 a	ND
5	0.560 ± 0.090 a	0.565 ± 0.090 a	0.491 ± 0.090 a	ND
10	0.563 ± 0.090 a	0.561 ± 0.090 a	0.652 ± 0.090 a	ND

Data are mean ± SD. Different letter superscripts in both rows and columns indicate P values ⩽0.05

MCV = mean cell volume; MCHC = mean cell hemoglobin concentration; ND = no data available; WBC = white blood cells

Plasma biochemical values and whole blood taurine

Plasma mean biochemical results from all three groups at the three time points (weeks 0, 5 and 10) are shown in Table 4. Albumin was significantly lower in both the raw diets at week 5 and 10 compared with the cooked diet, but it was still within normal reference intervals (P = 0.010). The cooked group had albumin levels above normal reference intervals. Globulin was significantly higher in both raw diets at week 5 compared with the cooked diet (P = 0.04), and also in the combination raw diet at week 10 compared with the cooked diet (P = 0.04). There were no significant differences in total protein between the diets at each time point, but over the 10 week period within each diet group, plasma values for total protein were significantly higher in the cooked and combination raw groups at weeks 5 and 10 compared with week 0 (P <0.0001). The cooked group had significant increases in albumin over the 10 weeks (P = 0.003), while the combination raw group had significant increases in globulin over the 10 week period (P <0.007). At the end of the feeding trial, total protein and taurine levels were highest in the kittens fed the combination raw diet, with taurine being significantly higher in this group at week 10 compared with the other two dietary treatments (P = 0.04). The combination raw diet also had three times the taurine content compared with the other two diets. Taurine concentrations in the cooked group were similar at all three time points, but in both of the raw diets differences over time were noted. The commercial raw diet taurine levels significantly decreased from week 0 to week 10 (P = 0.04), while the combination raw diet taurine levels significantly increased from week 0 to week 5 and week 0 to week 10 (P = 0.04). It is unknown if the taurine levels in the commercial raw diet would have continued to drop and become inadequate or stabilized without extending the feeding trial beyond 10 weeks. Both the combination raw diet and cooked diet had additional taurine listed on the ingredient list. Mean serum taurine levels for both the raw diet groups were >300 nmol/ml, and mean serum albumin levels were >2.7 g/dl, thus fulfilling this requirement for an AAFCO growth claim.

Table 4.

Serum biochemistry values by diet treatment and week

Analyte/week	Cooked	Commercial raw	Combination raw	RI
Total protein (g/dl)
0	5.91 ± 0.14d	6.09 ± 0.14b,c,d	6.04 ± 0.14c,d	4.8–6.5 11
5	6.24 ± 0.14a,b,c	6.30 ± 0.14a,b,c,d	6.39 ± 0.14a,b	5.4–6.8 8
10	6.37 ± 0.14a,b,c	6.21 ± 0.14a,b,c,d	6.52 ± 0.14a	5.4–6.8 8
Albumin (g/dl)
0	3.64 ± 0.07b	3.64 ± 0.07b	3.57 ± 0.07b	2.4–3.0 11
5	3.99 ± 0.07a	3.67 ± 0.07b	3.67 ± 0.07b	2.5–3.6 8
10	4.01 ± 0.07a	3.65 ± 0.07b	3.64 ± 0.07b	2.5–3.6 8
Globulin (g/dl)
0	2.27 ± 0.12d	2.45 ± 0.12b,c.d	2.46 ± 0.12c,d	NA
5	2.25 ± 0.12d	2.625 ± 0.12a,b,c	2.71 ± 0.12a,b	NA
10	2.36 ± 0.12cd	2.56 ± 0.12a,b,c,d	2.89 ± 0.12a	NA
Whole blood taurine (nmol/ml)
0	503.0 ± 47.4b,c	605.9 ± 47.4a,b	524.0 ± 47.4b,c	300–600 12
5	575.6 ± 47.4a,b,c	553.6 ± 47.4a,b,c	652.7 ± 47.4a	300–600 12
10	502.1 ± 47.4b,c	470.5 ± 47.4c	637.6 ± 47.4a	300–600 12
Blood urea nitrogen
0	19.12 ± 1.83c	21.37 ± 1.83b,c	22.50 ± 1.83b,c	16–33 11
5	17.88 ± 1.83c	22.75 ± 1.83b,c	28.62 ± 1.83a,b	19–34 12
10	18.62 ± 1.83c	19.37 ± 1.83c	25.12 ± 1.83a,b	19–34 12
Creatinine (mg/dl)
0	0.587 ± 0.047e	0.637 ± 0.047d,e	0.625 ± 0.047e	0.6–1.2 11
5	0.937 ± 0.047b	0.787 ± 0.047c	0.737 ± 0.047c,d	0.4–0.9 8
10	1.075 ± 0.047a	1.012 ± 0.047a,b	1.025 ± 0.047a,b	0.4–0.9 8
Potassium (mEq/l)
0	4.84 ± 0.19a,b	4.82 ± 0.19a,b,c	5.07 ± 0.19a	5.0–6.2 8
5	4.30 ± 0.19c,d	4.40 ± 0.19b,c,d	4.70 ± 0.19a,b,c	4.1–5.9 8
10	4.04 ± 0.19d	4.39 ± 0.19b,c,d	4.67 ± 0.19a,b,c,d	4.1–5.9 8
Bicarbonate (mmol/l)
0	19.0 ± 1.05a	18.0 ± 1.25a	16.8 ± 1.25a	ND
5	19.3 ± 0.98a	18.9 ± 0.98a	18.5 ± 0.98a	ND
10	16.5 ± 0.98a	18.4 ± 0.98a	18.7 ± 0.98a	ND
Anion gap
0	21.44 ± 1.51a	21.98 ± 1.78a	21.68 ± 1.78a	ND
5	23.66 ± 1.41a	21.15 ± 1.41a	24.16 ± 1.41a	ND
10	24.84 ± 1.41a	21.56 ± 1.41a	22.21 ± 1.41a	ND

Data are mean ± SD

RI = reference interval; NA = not available; ND = no data available

Within each time period, there were no significant differences in acid–base balance (bicarbonate and anion gap) between the three diets.

Potassium concentrations decreased significantly in the cooked group from week 0 to week 5 (P = 0.0009). Both raw diet groups had higher levels of potassium than the cooked diet group at weeks 5 and 10, with a trend toward significantly higher levels in the combination raw group compared with the cooked group at week 10 (P = 0.078).

There were no significant differences in creatinine concentrations between the three treatment groups, but creatinine levels increased significantly in all three groups over time (P <0.0001). Consistent with the protein content of the combination raw diet, blood urea nitrogen (BUN) was significantly higher in the combination raw group at week 5 compared with the cooked diet group (P = 0.011), and at week 10 compared with both the cooked and commercial raw diets.

Fecal and food cultures

Table 5 lists the results of composite fecal cultures submitted for each group of kittens by week. Four of the five groups had increases in Clostridium perfringens over the 10 week period. Two composite samples were positive for Salmonella enterica subspecies: group 2, week 10; and group 3, week 5. In the second group, Salmonella serotype Enteritidis was cultured at week 10, but all subsequent individual cultures, repeated three times at weekly intervals, were negative. In the third group, Salmonella serotype Heidelberg was cultured and found to be from a kitten on the combination raw diet. This kitten had no clinical signs and fecal scores of 4 throughout the feeding trial. Both Salmonella serovars isolated are commonly reported in raw poultry. ¹³ All three diets fed during these periods were tested and found negative for Salmonella species. Eight individual kittens with diarrhea were cultured throughout the feeding trials. In four of the kittens (three cooked feeders and one combination raw diet feeder), diarrhea was potentially associated with >1000 colonies of C perfringens. In three of the kittens with diarrhea (three cooked diet feeders), no Salmonella, Campylobacter or Clostridium species were found. One kitten with diarrhea on the combination raw diet was found to have C difficile toxin. All eight kittens were treated with 7 days of metronidazole with resolution of loose stool. The two kittens on the combination raw diet with known fecal bacteria (Salmonella serovar Heidelberg and C difficile toxin) had elevated lymphocytes (kitten with Salmonella species, week 5: 8540; week 10: 11,800; and kitten with C difficile, week 10: 8200).

Table 5.

Composite fecal cultures (colonies)

Group/week	Fecal cultures
Group 1
0	No Salmonella, Campylobacter or Clostridium species
10	500 C perfringens
Group 2
0	100 C perfringens
10	1000 C perfringens, Salmonella serotype Enteritidis; all subsequent individual cultures negative
Group 3
0	500 C perfringens
5	Salmonella serovar Heidelberg – combination raw fed kitten
10	1000 C perfringens
Group 4
0	500 C perfringens
5	1000 C perfringens
10	1000 C perfringens
Group 5
0	100 C perfringens
5	1000 C perfringens
10	1000 C perfringens

Discussion

Optimal growth rates in young animals are dependent on supplying adequate amounts of essential amino acids along with dietary nitrogen for synthesis of non-essential amino acids. Rogers et al found that kittens showed maximal weight gains at a wide dietary ratio of essential amino acid nitrogen to total amino acid nitrogen (E:T). ¹⁴ Compared with previously reported data on weight gain in kittens, ¹⁵ all kittens in our study, regardless of dietary treatment, had weight gains that were at the high end or above previously reported ranges. Males on the combination raw diet had the highest weight gains, but neither of the raw diets resulted in significantly higher weight, height or length gains. It is likely that the essential amino acid and total nitrogen required for optimal tissue accrual was maximized at levels of 90 g protein/1000 kcal. Thus, the higher protein level of 176 g protein/1000 kcal in the combination raw diet did not result in improved growth rates. Other potential physiological parameters, such as production of immune cells, biogenic amines and antioxidant levels that were not directly measured may have been influenced by the higher protein level in the combination raw diet.

As previously stated, the pet food industry uses ingredients that can have a wide variation in actual nutrient content and bioavailability. While AAFCO feeding trials are considered the ‘gold standard’ in evaluating pet food for nutritional adequacy, even these procedures have been previously cited as flawed. ¹⁶ Amino acid requirements and adequacy are generally based on weight gain or nitrogen balance for growth and maintenance, respectively. Levels of nutrients for other optimal functions such as immune function or antioxidant activity are generally not accounted for in AAFCO profiles. Additionally, the time frame for various feeding protocols may be too short to determine some nutritional deficiencies. The time frame for a kitten growth trial is only 10 weeks. ⁸ In theory, it is possible to pass an AAFCO growth feeding trial but still have inappropriate levels of some nutrients with respect to long-term feeding. Certain nutrients have body depot stores that will not be depleted or detected within this time frame. These nutrients include vitamins A, D and B₁₂, and, potentially, the minerals copper and iron.

Protein energy nutrition is known to have a direct effect on albumin production. Both of the raw diet groups had lower albumin levels at weeks 5 and 10 than the cooked diet group. As all the raw diet feeders’ albumin values were within previously published normal reference intervals, ¹⁰ the clinical significance of the difference between diet groups is likely negligible. In contrast with albumin, significantly higher globulin concentrations at week 10 were found in the combination raw diet group. Albumin synthesis is influenced by oncotic pressure, but inflammation with associated acute phase proteins and cytokine production, along with metabolic acidosis can inhibit albumin synthesis. No acid–base differences were noted between the groups. Inflammation cannot be ruled out as a source of higher globulin in the combination raw diet groups as elevated lymphocytes were noted at the same time as exposure to enteropathic fecal bacteria.

High serum creatinine and BUN concentrations have been previously reported in non-domestic felines fed high-protein diets. ¹⁷ All the groups had creatinine levels higher than previously published reference intervals. ¹⁰ BUN, a more consistent indicator of dietary nitrogen intake, was significantly higher in the kittens fed the combination raw diet secondary to its very high protein level compared with the cooked and commercial raw diets.

Serum potassium concentration was found to be significantly lower in the kittens fed the cooked diet compared with the raw diets at weeks 5 and 10. High-protein diets can increase metabolic acid load secondarily to the metabolism of methionine and cysteine to hydrogen sulfite and sulfate. If metabolic acidosis is present, potassium moves from the intracellular to extracellular environment in exchange for hydrogen ions. Because there were no differences in bicarbonate or anion gap between the three diets, acid–base balance does not seem likely as a cause of increased serum potassium seen in the raw diets. The potassium requirement in young felids is dependent on the level of protein in the diet. ¹⁸ The uptake of cation amino acids into the cell is exchanged for intracellular potassium. This exchange could account for the different levels found between the cooked diet group (lower protein) and combination raw diet group (higher protein). Differences in the raw vs cooked diet group’s serum potassium concentrations may also have been due to differing levels of potassium in each diet, but dietary potassium levels would need to be analyzed to confirm this

Replacement of large fetal red blood cells by smaller postnatal red cells occurs in the neonate within the first few weeks of life. ¹⁹ These changes are associated with reductions in hemoglobin, hematocrit, mean cell volume and mean cell hemoglobin content. Red cell levels reach their nadir at 4 weeks of age. Maturation of red cell indices similar to adult levels occurs by 3–4 months of age. ²⁰ Hemoglobin levels may not reach adult levels until 4–6 months of age, depending on the availability of iron in the diet. ²⁰ As common with other mammals, hemoglobin levels are reduced during the nursing period because of reduced concentrations of iron in milk along with lagging hemoglobin production and a rapidly proliferating red cell volume. ²¹ In our kittens, both hemoglobin and hematocrit continued to rise in the cooked diet group, as would be expected with further maturation of the hematopoietic system. In contrast, both raw diets saw modest increases in hematocrit and hemoglobin, which tended to plateau as the feeding trial progressed. Red cell microcytosis, as determined by previously published pediatric reference intervals, ²² was also noted in the kittens fed the combination raw diet as the feeding trial progressed.

Nutritional factors associated with reduced erythroid cell production and accompanying microcytosis include inadequate levels of pyridoxine, iron or copper. The combination raw diet had additional pyridoxine with inclusion of a B-vitamin supplement, but deficiencies in copper and/or iron may have caused the microcytosis as the diets were not tested for mineral levels and no information was available from the manufacturer. Seven of the eight kittens on the combination raw diet had MCV levels <43 fl at week 10, with four of these also having hematocrits <33%. Subclinical infection with associated anemia of inflammation could also have caused reduced erythroid production in the kittens fed the combination raw diet. Inflammation leads to iron retention within cells of the reticuloendothelial system resulting in limited iron availability for erythroid production. Acute phase proteins and cytokines negatively affect proliferation and differentiation of erythroid progenitor cells, along with inhibiting iron uptake.^1–3 In addition, the direct toxic effects of cytokine induced free radicals, including nitric oxide and superoxide anion, negatively affect erythroid cell proliferation. ⁴

Owing to composite rather than treatment group-specific fecal cultures, specific generalizations cannot be made. In 4/5 feeding trial groups, increases in C perfringens were noted over the 10 week period. C perfringens has been found to be a normal intestinal inhabitant of cats on canned commercial diets. ²³ C perfringens lacks many of the enzymes necessary for amino acid biosynthesis, thus thriving in high-protein environments. ²⁴ Previous studies in birds and mammals have shown increases in C perfringens with increased dietary protein.^25–27 It seems likely the diarrhea was associated with increased levels of C perfringens, but whether it was due to increased numbers or expression of enterotoxin is unknown. Toxigenic C difficile was also cultured from a kitten on the combination raw diet. While the diarrhea resolved with administration of metronidazole, gastroenteritis and potential shedding to other caregivers would be present.

While no cases of diarrhea were associated with Salmonella species, it was cultured twice during the feeding trial periods and in one case was traced back to a kitten fed the combination raw diet. Susceptibility and severity of infection depends on multiple factors, including virulence of the pathogen strain, infectious dose and host resistance. Host resistance factors include age, immunocompetence, stress, steroid administration and presence of chronic disease. While this kitten’s case was subclinical, elevated lymphocytes were noted. Risks of illness to susceptible animals and caregivers would be a significant concern.

Limitations of this study include the large difference in protein and fat content between the combination raw diet and the other two diets, and differences in ingredients used in all diets. Ideally, a single diet ration would have been used, with commercial heat processing applied to one portion and not to the other. This would have eliminated any potential differences in macronutrient content, protein quality and variance in ingredients. Owing to cost restrictions, complete nutrient analysis of each of the diets, including vitamin and mineral levels, was not carried out. Having a complete dietary analysis would have allowed for detection and analysis of other nutrient considerations important for growth. Further limitations include the use of composite fecal cultures rather than treatment group-specific cultures. Diagnostic specific techniques for C difficile (combination toxin testing by ELISA and organism detection by culture or PCR), C perfringens (ELISA for enterotoxin along with PCR for enterotoxic strains) and use of food industry standards for Salmonella species and Campylobacter jejuni food monitoring would have provided a more accurate assessment of pathogen exposure. ²⁸ Along with more specific pathogen testing, examination of serum acute phase proteins and cytokines would have provided evidence for the presence or absence of an inflammatory response.

Conclusions

Overall, the growth parameters were similar for all diet groups, indicating the two raw diets in this study supported feline growth within the limitations of an AAFCO growth feeding trial. Salmonella serovar Heidelberg and C difficile enterotoxin were detected in the feces of some of the raw feeders, substantiating pathogen exposure and risks to both animal and owner. Further research in this area is needed to investigate the impact of raw diets on the health of domestic cats.