Starch sources and their influence on extrusion parameters, kibble characteristics and palatability of dog diets

Abstract

This study aimed to evaluate the effects of different starch sources on the extrusion process, kibble characteristics and diet palatability for dogs. Seven diets with corn, brown rice, sorghum, potato starch, sweet potato flour, chickpea and pea as the only starch source were evaluated. The dietary inclusion of the starch sources varied between 330.13 and 643.84 g/kg to provide similar starch concentrations in all diets (around 300 g/kg). The conditioning temperature and the volume of water added to the preconditioner during the processing of the diets were measured, and the extruder variables evaluated were knife speed, feed rate, screw speed, amperage and productivity. For the physical characteristics of kibbles, density, size, expansion index and hardness were measured. Porosity variables evaluated were total pore area, average pore area and number of pores. Diet palatability was evaluated in 16 dogs. Six paired palatability tests were performed, in which all diets were compared to the corn diet, with two consecutive days per test, totalling 32 repetitions (16 dogs × 2 days/test). Diets with tubers presented lower kibble density and higher size, expansion index and hardness than the diets with cereals and pulses (p < 0.001). Diets with tubers had more pores than pulses (p < 0.001). The inclusion of pea results in kibbles that were less expanded and with fewer pores (p < 0.001). Dogs preferred all diets compared with the corn-based diet (p < 0.001), except for the sweet potato diet. The dietary inclusion of tubers as starch sources results in higher kibble expansion, size, hardness, porosity and lower density. The inclusion of pea results in kibbles that are less expanded and with a lower number of pores. The diet’s moisture content and texture affect diet palatability in dogs.

HIGHLIGHTS

• Tubers, such as potato, as starch sources result in higher kibble expansion, size, hardness and porosity.

• Starch sources with a high fibre concentration, such as pea and chickpea, result in diets with kibbles less expanded and with less porosity.

• The moisture content and physical characteristics of kibbles affect the diet palatability in dogs.

Keywords:

• Amylose
• amylopectin
• cereals
• pulses
• tubers

Introduction

The extrusion process aims to cook, shape and texturise a homogeneous mass of ingredients by combining moisture, pressure, temperature and mechanical friction in a short time (Riaz 2007). It is the main process used worldwide to produce commercial pet diets (Spears and Fahey 2004). Among the ingredients included in the dough, starch sources are the main responsible for the rheological properties that characterise most of the extruded products. In general, the starch fraction works as a thermoplastic polymer during extrusion. When leaving the extruder, the pressure and temperature differential induce the water vaporisation that deforms the gelatinised starch. When water, energy and time are sufficient in the process, starch granules lose their crystallinity, swell and disrupt, forming an amorphous mass that binds all food components forming a continuous structure (Ding et al. 2005). This results in a cellular structure in the kibbles, responsible for their expansion and crispness formation, which are important for the appearance, palatability and digestibility of the diets for dogs (Baller et al. 2018; Ottoboni et al. 2019).

Starch is found in several plant species as a reserve digestible carbohydrate, being abundant in cereal grains (580–830 g/kg on dry matter basis, DM; Carciofi et al. 2008; Bazolli et al. 2015), pulses (270–570 g/kg DM; Carciofi et al. 2008; He and Wei 2017) and tubers (950–970 g/kg DM; Carciofi et al. 2008; Domingues et al. 2019). Although corn and broken rice are widely used in extruded dog foods, the interest in different starch sources has increased, due to commercial trends or specific formulations (e.g. diets for diabetic and obese dogs), exploring the use of whole grains (brown rice and sorghum), pulses (lentil, chickpea and pea) and tubers (potato and cassava) (Englyst et al. 2003).

However, it is known that the composition of these raw materials, mainly the differences in the crystalline form of their starch granules, amylose:amylopectin ratio and fibrous concentration can influence starch gelatinisation, viscosity and shear inside the extruder barrel (Eliasson 2004). For example, pulses, such as pea and chickpea, present higher amylose (240–490 g/kg DM) and fibre concentration (around 170 g/kg DM) than cereals (around 200–300 g/kg amylose and 80 g/kg fibre on DM basis in corn) (Hoover et al. 2010; Pérez and Bertoft 2010), which may impact the rheological properties of starch during the extrusion process, resulting in kibbles with different characteristics (Domingues et al. 2019; Pezzali and Aldrich 2019). The differences in kibble texture due to different behaviours of starch sources during the extrusion process may affect diet palatability and digestibility (Koppel et al. 2015; Domingues et al. 2019) and post-prandial response in dogs (Vastolo et al. 2023).

Modifications caused by extrusion on nutritional quality and microbiological safety of the diets are already described in the literature (Riaz 2007; Tran et al. 2008; Gibson and Alavi 2013). However, few published studies evaluated the effects of starch sources on kibble characteristics and diet palatability of pet food (Bazolli et al. 2015; Baller et al. 2018; Domingues et al. 2019; Pezzali and Aldrich 2019). As a differential, this study provides an evaluation of alternative starch sources to corn in extruded dog food. Thus, the objective of this study was to evaluate the influence of different starch sources on extrusion parameters, physical characteristics of kibbles and diet palatability in dogs.

Materials and methods

All animal care and experimental procedures were approved by the Animal Use Ethics Committee of the Agricultural Sciences Sector of the Federal University of Paraná, Curitiba, PR, Brazil, under protocol n. 001/2020.

Experimental diets

Seven diets containing different starch sources corn, brown rice, sorghum, potato starch, sweet potato flour, chickpea and pea were evaluated. The diets were formulated to have similar starch (300 g/kg starch), protein and fat concentrations and to satisfy the nutritional requirements of adult dogs according to the European Pet Food Industry Federation (FEDIAF 2019). The ingredients of the evaluated diets are presented in Table 1.

Table 1. Ingredients of the experimental diets.

Ingredients (g/kg)	Corn	Brown rice	Sorghum	Potato	Sweet potato	Chickpea	Pea
Corn	433.90	–	–	–	–	–	–
Brown rice	–	447.45	–	–	–	–	–
Sorghum	–	–	450.13	–	–	–	–
Potato	–	–	–	330.13	–	–	–
Sweet potato	–	–	–	–	366.28	–	–
Chickpea	–	–	–	–	–	643.84	–
Pea	–	–	–	–	–	–	565.63
Poultry offal meal	267.88	264.36	255.61	293.16	292.12	243.82	227.20
Isolated swine protein	147.43	138.73	146.97	170.68	171.84	45.92	81.26
Cellulose	55.00	41.43	45.30	66.81	46.20	20.00	26.05
Poultry fat	49.71	61.91	58.94	93.41	74.75	9.11	53.33
Poultry hydrolyzate	20.00	20.00	20.00	20.00	20.00	20.00	20.00
Potassium chloride	7.05	7.09	5.00	6.78	8.78	2.00	7.50
Sodium chloride	5.00	5.00	4.02	5.00	5.00	2.00	5.00
Mineral-vitamin supplement^a	3.00	3.00	3.00	3.00	3.00	3.00	3.00
Adsorvent	2.00	2.00	2.00	2.00	3.00	2.00	2.00
Choline chloride	2.00	2.00	2.00	2.00	2.00	2.00	2.00
Calcium propionate	2.00	2.00	2.00	2.00	2.00	2.00	2.00
Taurine	1.00	1.00	1.00	1.00	1.00	1.00	1.00
DL-methionine	0.80	0.80	0.80	0.80	0.80	0.08	0.80
BHT	0.15	0.15	0.15	0.15	0.15	0.15	0.15
BHA	0.08	0.08	0.08	0.08	0.08	0.08	0.08
Citric acid	3.00	3.00	3.00	3.00	3.00	3.00	3.00

^aEnrichment per kg of product^–1: vitamin A (retinol) = 20.000 IU; vitamin D3 = 2.000 IU; vitamin E (alpha-tocopherol α) = 48 mg; vitamin K3 = 48 mg; vitamin B1 = 4 mg; vitamin B2 = 32 mg; pantothenic acid = 16 mg; niacin = 56 mg; choline = 800 mg; Zn as zinc oxide = 150 mg; Fe as ferrous sulphate = 100 mg; Cu as copper sulphate = 15 mg; I as potassium iodide = 1.5 mg; Mn as manganese oxide = 30 mg; Se as sodium selenite = 0.2 mg; antioxidant = 240 mg.

BHA: Butylated hydroxyanisole; BHT: Butylated hydroxytoluene.

Extrusion process and physical characteristics of kibbles

The ingredients were ground in 0.8 mm sieves and extruded in a single-screw extruder (Ferraz, E-130; Ribeirão Preto, SP, Brazil). The conditioning temperature and the volume of water added to the preconditioner during the processing of the diets were measured. The temperature was measured with a digital infra-red thermometer (LaserGrip Model GM400, São Paulo, SP, Brazil) every 10 min. The following extruder variables were measured every 10 min: knife speed (Hz), feed rate (Hz), screw speed (Hz) and amperage (A). The productivity was kept constant at approximately 1070 kg/h, and the extrusion variables and diets used in the study were considered only after extrusion stabilisation. After the extrusion process, the diets were dried in a horizontal dryer for 25 min at a temperature from 52 (end of the dryer) to 110 °C (beginning of the dryer) and coated with poultry fat and liquid palatant.

One batch per diet was produced and 20 aliquots per diet were collected for the analysis of physical characteristics of kibbles. The density, size, expansion index (EI) and hardness of the experimental diets were measured in 20 samples per treatment. Density was expressed as the ratio of diet weight (grams) to volume (liters). Kibble size was determined using a digital calliper (MTX-316119). The EI was calculated as the ratio between the kibble’s radial size and the extruder die’s diameter.

For hardness analysis, 20 intact kibbles from each diet were randomly selected. These samples were analysed using a durometer (Ethik Technology; 298 DGP II hardness tester, Brazil), which measures the diametrically applied force required to break a kibble. The force was measured in Newton (N) and converted and expressed in kgf/cm².

Diets were examined under scanning electron microscopy (SEM) at 10x magnification to determine kibble porosity (three replicates per treatment). Each kibble was cut longitudinally to allow better visualisation of the pores. Kibble pore area (mm²) was measured using the ImageJ® software (Wayne Rasband, National Institutes of Health, Bethesda, MD, USA). The number of pores was manually counted in the image generated at SEM. All pores presented in the image were counted after applying the threshold tool of ImageJ® to highlight the pores. The variables evaluated were: total pore area, average pore area and number of pores, in accordance with Domingues et al. (2019).

Chemical analysis

Diets and starch sources were analysed for dry matter at 105 °C (DM105), ether extract in acid hydrolysis (EEAH, method 954.02), ash (method 942.05), calcium (method 927.02), phosphorus (method 984.27), nitrogen (N, method 954.01) and then crude protein (CP) was calculated as N × 6.25, according to the Association of Official Analytical Chemists (AOAC, 1995). Starch was analysed according to Hendrix (1993). The total dietary fibre, insoluble fibre and soluble fibre were analysed according to Prosky et al. (1988). Gross energy (GE) was determined by an isoperibol calorimeter (Parr Instrument Co., Model 1261, Moline, IL, USA). All analyses were made in triplicate and repeated if they varied more than 5%.

Palatability assay

For the palatability test, 16 adult Beagle dogs (eight males and eight females), with 4 years of age, mean body weight of 10.4 ± 1.76 kg, and body condition score of 5.6 ± 0.6 (Laflamme 1997) were used. The dogs maintained their usual routine during the experimental period. They had free access to an outdoor area (1138 m²) under supervision for 4 h/day for voluntary exercise and socialisation with other dogs and people. Outside this period, they were housed in pairs in covered masonry kennels (5 m long x 2 m wide) with bed and access to water ad libitum. The kennels had bars on the side walls allowing limited visual interaction with neighbouring dogs. The room temperature ranged from 17 °C to 26 °C with a 12-h light-dark cycle. Only during the palatability test (for about 30 min.), dogs were housed individually in the same kennels.

Diets containing sorghum, brown rice, potato, sweet potato, chickpea and pea were compared one by one to the diet containing corn, totalling six comparisons. Each comparison was performed for two consecutive days. The two foods were offered simultaneously to the dogs once a day (08:00 h). The amount provided was 30% higher than the NRC (2006) recommendations for the maintenance of adult dogs to assure the presence of leftovers of one of the diets. Once one of the diets was completely consumed, both bowls were removed, and the remaining amount was quantified. The relative position of the feeders was alternated on the second day of the experiment so that the animal was not conditioned to the feeding site.

Palatability was determined considering the intake ratio and the first choice between the diets offered to the dogs. The first choice was considered by the first bowl that the animal approached. To determine the intake ratio, the food offered and the leftovers were quantified for later calculation.

Calculations and statistical analysis

The energy consumption (kW/h) and the specific mechanical energy (SME) transferred to the mass (kW/h/ton) were calculated according to Riaz (2007), with the following equations: $$\surd \text{Electric \hspace{0.17em}phase\hspace{0.17em} of \hspace{0.17em}the\hspace{0.17em} system}\times \text{feeding\hspace{0.17em}rate}\times \text{amperage}\times {\text{motor\hspace{0.17em}cos}}_{\mathrm{\gamma }}/1000.$$

In which, motor cos_γ = 0.86 $$\text{SME}\left(\text{kW}/\mathrm{h}/\text{ton}\right)=\left(\text{energy\hspace{0.17em}consumption}\right)\times 1000/\text{extruder \hspace{0.17em}output\hspace{0.17em}}\left(\text{kg}/\mathrm{h}\right)$$

Extrusion variables were not subjected to statistical analysis due to only two observations per treatment.

The density, size, EI and porosity data of kibbles were analysed for normality by the Shapiro–Wilk test (p < 0.05) and submitted to analysis of variance (ANOVA), totalling 20 repetitions per treatment (except for porosity data that resulted in three repetitions per treatment) with a significance level of p < 0.05. When the F test of ANOVA indicated statistical difference, means were compared by Tukey’s test (p < 0.05).

For the palatability test, the intake ratio of the diets was calculated according to the equation: $$\text{Intake\hspace{0.17em}ratio}=\mathrm{g}\text{\hspace{0.17em}consumed\hspace{0.17em} of \hspace{0.17em}diet}\mathrm{A}\text{or}\mathrm{B}/\text{total\hspace{0.17em}}\mathrm{g}\text{\hspace{0.17em}consumed\hspace{0.17em}}\left(\mathrm{A}+\mathrm{B}\right)$$

Palatability data were analysed considering a completely randomised design. The results of the intake ratio were compared by paired Student t-test and the first choice by the chi-square test, both at 5% significance level, totalling 32 repetitions per test (16 dogs × 2 days of evaluation). Data were analysed using the SAS statistical package (version 8, SAS Institute Inc., Cary, NC, USA).

Results

Chemical composition of the diets

Although the diets were formulated to have similar chemical composition, there were some differences in starch concentrations among diets (Table 2), varying from 269.4 (pea diet) to 373.7 g/kg DM (brown rice diet). There were also differences (Table 2) in CP concentrations (320.8 g/kg in the brown rice diet to 354.9 g/kg DM in the chickpea diet) and TDF concentrations (76 g/kg DM in the potato diet to 170.6 g/kg DM in the pea diet).

Table 2. Analysed chemical composition (g/kg, dry matter basis) of starch sources and experimental diets.

Item	Starch sources
	Corn	Brown rice	Sorghum	Potato	Sweet potato	Chickpea	Pea
Dry matter	887.8	887.6	877.0	857.9	915.1	912.6	881.5
Ash	25.1	14.5	23.0	3.0	17.9	29.5	28.9
Crude protein	74.5	137.7	107.3	3.7	27.8	240.3	255.4
EEAH^a	38.1	41.1	20.6	5.3	11.0	71.3	23.9
Gross energy (kcal/kg)	4829.2	4429.9	4321.8	4217.4	4126.4	4860.7	4545.6
Total starch	700.6	767.5	753.7	960.4	786.9	452.7	476.2
Total dietary fibre	52.3	68.9	53.3	29.7	112.3	257.0	237.0
Insoluble fibre (IF)	46.8	61.2	46.9	27.6	68.3	226.9	208.6
Soluble fibre (SF)	5.5	7.7	6.4	2.1	44.0	30.1	28.4
IF:SF	8.5	7.9	7.3	13.1	1.5	7.5	7.3
Ca	0.3	0.3	0.4	0.3	1.1	1.5	0.5
P	2.4	2.9	3.9	0.7	0.6	3.8	3.3
	Diets
	Corn	Brown rice	Sorghum	Potato	Sweet potato	Chickpea	Pea
Dry matter	963.3	927.6	933.5	951.1	970.3	953.6	941.1
Ash	79.7	87.1	81.3	77.0	81.7	89.8	84.7
Crude protein	328.4	320.8	338.3	330.5	350.1	354.9	345.0
EEAH^a	128.7	123.1	126.3	124.8	123.8	138.5	123.0
Gross energy (kcal/kg)	4784.4	4861.6	4863.4	4959.1	4916.9	4987.1	4882.6
Total starch	304.1	373.7	353.3	350.3	288.3	291.5	269.4
Total dietary fibre	81.7	77.3	76.0	94.3	96.2	169.1	170.6
Insoluble fibre (IF)	76.4	70.8	69.1	89.1	75.1	146.4	146.9
Soluble fibre (SF)	5.3	6.5	6.9	5.2	21.1	22.7	23.7
IF:SF	14.4	10.9	10.0	17.0	3.56	6.44	6.19
Calcium	15.0	14.5	13.2	13.5	15.0	15.0	12.3
Phosphorus	7.9	7.9	7.8	6.9	7.7	7.7	8.3

^aEEAH: ether extract in acid hydrolysis.

Extrusion variables and physical characteristics of kibbles

The results of variables evaluated during the extrusion process are presented in Table 3. In general, the sorghum, potato and chickpea diets presented higher temperature in the preconditioner. In addition, the potato and chickpea diets demanded higher water addition in the preconditioner, and the potato and sweet potato diets resulted in higher SME in the extruder barrel.

Table 3. Means of extrusion variables in diets containing different starch sources for dogs.^a

Item	Corn	Brown rice	Sorghum	Potato	Sweet potato	Chickpea	Pea
Preconditioner
Temperature (ºC)	65.7	55.9	76.4	80.8	68.6	85.6	63.5
Water addition (L/h)	257.5	260.0	227.5	300.0	210.0	295.0	247.5
Extruder
Screw speed (Hz)	39.0	39.0	37.0	41.0	42.5	42.5	37.2
Feeding rate (Hz)	15.0	15.0	15.0	15.0	15.0	15.0	15.0
Amperage (A)	56.4	60.9	54.4	67.2	67.2	48.2	63.1
SME (kW/ton/h)^b	17.27	18.65	16.68	20.60	20.60	14.76	19.33
Die
Number of holes	25	25	25	25	25	25	25
Hole diameter (mm)	4.5	4.5	4.5	4.5	4.5	4.5	4.5

^aN = 2 observations.

^bSME: Specific mechanical energy transferred to the mass.

Physical characteristics of kibbles (Table 4) from diets containing potato starch had greater size and EI and lower density when compared to kibbles from diets based on cereals and pulses (p < 0.001), except for the density and size of kibbles of sorghum and the density of kibbles of pea, that did not differ (p > 0.05). These variables also did not differ between potato and sweet potato kibbles (p < 0.05). On the other hand, kibbles from diets containing pea had EI similar to the diets with corn and brown rice and lower than the other diets (p < 0.001).

Table 4. Means of physical characteristics of kibbles in diets containing different starch sources for dogs.

Item	Corn	Brown rice	Sorghum	Potato	Sweet potato	Chickpea	Pea	SEM^1	p-value
Density (g/L^3)	409.50^b	459.60^a	364.70^c	356.10^cd	343.70^d	447.20^a	371.70^c	3.893	<0.001
Size (mm^3)	8.81^cd	8.97^cd	9.89^bc	11.35^a	10.49^ab	8.33^d	10.20^b	0.116	<0.001
EI² ^3	1.96^cd	1.99^cd	2.15^bc	2.52^a	2.33^ab	2.26^b	1.85^d	0.026	<0.001
Hardness (kgf/m²)^3	9.65^b	8.87^b	8.96^b	13.92^a	10.60^b	4.74^c	10.14^b	0.276	<0.001
Total area of pores (mm²)^4	35.04^ab	28.24^b	57.81^ab	65.48^a	63.49^a	33.04^ab	24.22^b	3.805	0.003
Average pore area (mm^2)^4	1.17^bc	0.94^c	1.92^a	2.18^a	1.74^ab	1.10^c	0.81^c	0.058	<0.001
Number of pores^4	36^cd	41^bc	42^bc	46^ab	55^a	30^d	30^d	0.5	<0.001

¹SEM = Standard error of mean; ²EI = Expansion index; ^a,b,c,d Distinct letters indicate difference by Tukey’s test (p < 0.05). ³N = 20 samples per treatment; ⁴N = 3 replicates per treatment.

Kibbles from the chickpea-based diet showed a density similar to brown rice and higher than the other starch sources (p < 0.001). Furthermore, kibbles from the chickpea diet had a similar size to diets containing corn and brown rice and were lower when compared to the other diets (p < 0.001).

Hardness was higher in kibbles containing potato in relation to the other diets and lower for chickpea-based kibbles (p < 0.001). The other diets showed similar hardness (p > 0.05).

Total pore area was greater in tuber-based diet kibbles compared to pea-based diet kibbles (p = 0.003). A greater average pore area was found in kibbles of the diet containing potato in relation to corn, brown rice and pulses (p < 0.001). The pulses-based and brown rice-based diet presented a lower average pore area compared to sorghum and tubers (p < 0.001). Diets containing tubers had a higher number of pores compared with diets containing cereals and pulses (p < 0.001), while diets containing pulses had a lower number of pores compared with all diets, except corn (p > 0.05, Figure 1).

Figure 1. Scanning electron microscopy (10x magnification) of the longitudinal section of kibbles of diets with different starch sources. (a) Corn, (b) brown rice, (c) sorghum, (d) potato, (e) sweet potato, (f) chickpea, and (g) pea.

Palatability

There was a difference in the first choice only between corn vs. chickpea diets, with dogs preferring the chickpea diet (p < 0.05). All comparisons resulted in a lower intake ratio to the corn diet in comparison to the other starch sources (p < 0.001), except for the sweet potato diet, which did not differ from corn diet (p > 0.05, Table 5).

Table 5. Number of first choices to diet A – corn (n) and intake ratio (%) of experimental diets.

Diet (A × B)	n¹	Intake ratio²	p-value
		A
Corn × Brown rice	14	37.80	0.039*
Corn × Sorghum	13	6.93	<0.001*
Corn × Potato	12	16.20	<0.001*
Corn × Sweet potato	11	40.49	0.102
Corn × Chickpea	7*	9.12	<0.001*
Corn × Pea	11	26.07	<0.001*

^aNumber of first choices to diet B is obtained by 32-n.

^bIntake ratio: [g ingested from diet A or B/total g ingested (A + B)] × 100.

*Number of first choices to diet A differs by the chi-square test and intake ratio by the paired t-test (p < 0.05).

Discussion

As expected, the physical characteristics of kibbles, mainly density, expansion, hardness and porosity are highly influenced by the starch source used in the diet. Besides, the other ingredients from the formulation and processing conditions, such as the mechanical and thermal energy transferred to the dough, residence time in the preconditioner and extruder barrel, and moisture can also influence kibble characteristics (Riaz 2007).

Starch composition significantly influences starch behaviour during food extrusion. Tuber starches, such as potato and sweet potato, contain a higher proportion of amylopectin (700–800 g/kg) relative to amylose (200–300 g/kg; Pérez and Bertoft 2010), facilitating starch gelatinisation during extrusion. Furthermore, tubers contain high levels of phosphate monoesters, which are covalently bound to the amylose and amylopectin fraction (Schirmer et al. 2013). These phosphate groups contribute to higher viscosity, water binding capacity and swelling power, contributing to a low gelatinisation temperature (Hoover 2001; Singh et al. 2003), directly influencing the gelatinisation properties and retrogradation rate of starch (Karim et al. 2000; Schirmer et al. 2013).

Increasing starch gelatinisation improves kibble expansion (Bhattacharya and Choudhury 1994). A greater EI was observed in this study in kibbles of potato starch and sweet potato-based diets (2.52 and 2.33, respectively). The same was described in studies evaluating the inclusion of potato starch compared to corn (Domingues et al. 2019) and different types of processing on physical characteristics of kibbles of diets containing sweet potato flour (Borba et al. 2005). A higher expansion also results in kibbles with lower density (Bhattacharya and Choudhury 1994; Domingues et al. 2019).

Pulses present higher amylose (290–650 g/kg) concentration than tubers and cereals (200–350 g/kg) (Joshi et al. 2013; Thakur et al. 2019). Besides, pulses also present higher TDF concentrations, which may influence the kibble macrostructure. A high dietary fibre concentration may hinder the formation of an adequate cellular structure, as it conducts water vapour without the formation of cells. This reduces the radial expansion, increases specific and apparent density, and promotes the formation of harder kibbles with small pores (Robin et al. 2012; Monti et al. 2016; Souza et al. 2022). These characteristics may explain the lower kibble expansion of the diet with the inclusion of pea (TDF = 170.6 g/kg DM) and the lower radial size, and higher density of the chickpea-based diet (TDF = 169.1 g/kg DM).

High-density diets usually are associated with low porosity (Camire et al. 1990; Tiwari and Jha 2017). Size and number of pores in a kibble are related to density, and the formation of the porous structure is also dependent on the starch content and type of the formula (Gill 2002; Ah-Hen et al. 2014). This association was observed in kibbles of diets containing brown rice and chickpea, which presented higher density, lower total pore area, average pore per area and number of pores (only chickpea-based diet) compared to other diets. In turn, it was observed in tubers-based diets kibbles with lower density and higher total pore area, average pore per area and number of pores.

Although it was not the objective of this study, it is important to highlight that other studies have observed that these differences in amylose:amylopectin and fibre concentrations together with other physicochemical characteristics of starch sources influence diet digestibility and the post-prandial glycemic response in dogs. Dogs fed diets containing pulses and sorghum presented lower DM and energy digestibility and lower glycemic response compared to dogs fed potato, corn and broken rice diets (Carciofi et al. 2008; Domingues et al. 2019; Quilliam et al. 2021). Considering that, it is important to consider to petfood formulations the overall effect of these starch sources on the extrusion process, kibble characteristics and physiological responses in dogs.

As a limitation of this study, extrusion parameters were not subjected to statistical analysis, due to the lack of enough repetition, making it difficult to correlate these variables with their effects on the physical characteristics of kibbles. However, it seems that the rheological properties of potato and sweet potato starch increase the SME in the extruded barrel, probability due to the higher gelatinisation rate, resulting in greater stickiness during the extrusion process. The higher SME during the extrusion process of diets with potato starch in comparison to corn-based diets was also observed by Domingues et al. (2019). Another limitation of this study is the differences observed in the chemical composition of the diets, especially in the starch and protein concentrations, which were expected to be similar. These differences may be due to small analytical and/or dosage errors and were observed especially in the starch content of brown rice (373.7 g/kg DM) and pea (269.4 g/kg DM) diets.

Diet palatability can be influenced by several factors, such as the ingredients of the diet and its processing, the macrostructural characteristics of the kibbles, chemical composition, inclusion of different palatants and how they all relate to sensory properties, such as flavour, texture, shape and taste (Aldrich and Koppel 2015), demonstrating the complexity of the evaluation of palatability in dogs. Besides, food intake may also be influenced by diet digestibility. Considering that, one of the hypotheses for the difference in the intake ratio between corn compared to the other diets was the moisture content (moisture of the diets: corn = 36.7 g/kg; brown rice = 72.4 g/kg; sorghum = 66.5 g/kg; potato = 48.9 g/kg; chickpea = 46.4 g/kg, and pea = 59.0 g/kg). Moisture content is crucial for diet palatability and dogs can distinguish and choose diets with only two points more moisture when comparing diets with 80 and 100 g/kg moisture (Brito et al. 2010). On another hand, in studies evaluating diets with similar moisture content, dogs also preferred potato starch, pea and chickpea, instead of the corn-based diet (Domingues et al. 2019; Pacheco et al. 2021). These results demonstrate that not only the moisture content of the diets, but also other physical and chemical characteristics influence diet palatability for dogs.

It is possible that in diets with similar moisture content, the main effects of starch sources on diet palatability are their influence on kibble texture and consequently also on coating. Kibbles with lower density may increase crispness (strength and number of chewing movements) (Félix et al. 2010). In addition, density and highly porous kibbles are also associated with high oil absorption capacity (Gill 2002; Ah-Hen et al. 2014), affecting coating homogeneity. Fatty clusters tend to stick to the surface of the pet food, affecting food appearance and texture (Samant et al. 2021). Fat distribution on kibbles is important for good food palatability, and a worse coating efficiency could reduce food acceptance by dogs (Monti et al. 2016).

Conclusion

The inclusion of potato and sweet potato as starch sources results in higher kibble expansion, size, hardness, porosity and lower density. It also seems that potato and sweet potato starches increase the SME during the extrusion process. The inclusion of starch sources with high fibre concentration, such as pea, results in diets with kibbles less expanded and with less porosity. In addition, moisture content and kibble texture affect diet palatability in dogs.

Ethical approval

The experiment was approved by the Animal Use Ethics Committee of the Agricultural Sciences Sector of the Federal University of Paraná, Curitiba, PR, Brazil, under protocol n. 001/2020.

Credit authorship contribution statement

Gislaine Cristina Bill Kaelle: Investigation, Data Curation, Writing – Original Draft, Writing – Review and Editing. Taís Silvino Bastos: Investigation, Review. Renata Bacila Morais dos Santos de Souza: Investigation. Eduarda Lorena Fernandes: Investigation. Lorenna Nicole Araújo Santos: Investigation. Simone Gisele de Oliveira: Supervision, Writing – Review and Editing. Ananda Portella Félix: Conceptualisation, Data Curation, Project administration, Writing – Review and Editing.

Acknowledgements

The authors would like to thank H2Nutri and Oriente for providing the sweet potato and chickpea. Premier Institute for financial support. Center for Electronic Microscopy (UFPR) and Professor Itamar Francisco Andreazza (UFPR) for the partnership in the physical characteristics analysis of the kibbles. Coordenação de Aperfeiçoamento de Pessoal de Nível Superior (CAPES) for the scholarship of the first author.

Disclosure statement

The authors declare no conflict of interest.

Data availability statement

The data that support the findings of this study are available from the corresponding author ([email protected]), upon reasonable request.