1. Introduction

The pet food sector increasingly emphasizes ingredient quality, nutritional adequacy, palatability, safety, and environmental performance. Animal-derived ingredients remain important in canine diets because they can provide concentrated protein and indispensable amino acids, although their nutritional value depends on raw-material composition and processing conditions [1,2,3,4].

Edible organs and other slaughter by-products may be valuable pet food ingredients rather than low-value waste streams. Their composition differs markedly among tissues: liver is generally nutrient dense and rich in several trace elements, whereas lung is typically leaner and contains a relatively high proportion of connective tissue. Consequently, organ type, ash content, connective-tissue content, and thermal history can influence protein solubility, amino acid availability, and digestibility [3,4,5,6]. Utilization of locally available by-products may also improve resource efficiency, but environmental benefits should not be assumed without a defined allocation method and life-cycle assessment [7,8].

Drying lowers water activity, facilitates grinding and storage, and can concentrate nutrients on an as-is basis. Nevertheless, an apparent increase in protein, fat, or minerals after drying primarily reflects removal of water rather than creation of nutrients. Heat intensity must also be controlled because excessive processing can reduce reactive lysine and promote Maillard reaction products, with potential consequences for protein quality [9,10,11]. Therefore, the nutritional value of a dried ingredient cannot be inferred from crude protein concentration alone.

Palatability is commonly assessed with acceptance tests or two-bowl preference tests. First choice and relative intake provide different information, and the design should account for bowl position, repeated observations, and individual dog effects, and paired data structure [12]. Recent studies demonstrate that animal-protein source and by-product inclusion can alter preference and nutrient utilization, but results vary with formulation and processing [1,2,13,14,15].

Horse meat production is culturally and economically important in Mongolia, yet scientific data on dried horse liver and lung in dog foods are limited. The objective of this study was therefore to characterize the proximate and selected mineral composition of hot-air-dried horse liver and lung, develop prototype dry dog foods, assess reported microbiological quality, and evaluate feed acceptance and two-bowl preference in adult dogs. The study was designed as a preliminary product-development investigation rather than a demonstration of complete and balanced nutritional adequacy.

2. Materials and Methods

2.1. Raw materials and preparation

Horse liver and lung meeting the requirements of the Mongolian standard MNS 1023:2007 [16] were obtained for product development. Visible connective tissue and contaminants were removed, and the organs were cut into slices approximately 5mm thick. The sliced liver and lung samples were dried in a hot-air drying oven at either 50°C or 70°C for 24 h, with 5 kg of each organ loaded per batch. Drying was continued until constant weight was achieved. Each temperature treatment was conducted using three independently prepared batches (n = 3). After drying, the materials were ground through an approximately 1 mm screen and stored in airtight packaging until formulation.

2.2. Product formulation and selection

Three prototype formulations containing 80%, 85%, or 90% dried horse liver or lung were prepared. Corn and peas were included at 2.5% each in all formulations, while oil was added at 15%, 10%, or 5%, respectively, resulting in a total formulation of 100%. The complete ingredient composition of the experimental formulations is presented in Table 1. Prototype III, containing 90% dried horse liver or lung, 2.5% corn, 2.5% peas, and 5% oil, was selected for subsequent evaluation based on satisfactory product formation during preliminary processing. No formal sensory evaluation was conducted at this stage.

2.3. Chemical and microbiological analyses

Moisture, crude protein, fat, and ash were determined according to MNS 3746-84, MNS 3745-84, MNS 3748-84, and MNS ISO 936:2003, respectively, and were reported on an as-is basis. Product pH was measured according to MNS ISO 2917:2000. Iron and copper were quantified by atomic absorption spectroscopy (AAS) following microwave-assisted acid digestion and expressed as mg/kg product on an as-is basis. Total carbohydrate and energy were calculated, and color was measured using a CR-20 colorimeter. Total aerobic bacterial counts were determined using a serial-dilution plate-count method, while Escherichia coli and molds were assessed using standard laboratory procedures. Chemical and microbiological analyses were conducted on three independently prepared batches of each formulation, with each batch analyzed in triplicate. Technical replicates were averaged within each batch, and data are presented as mean ± SD across the three independent batches (n = 3).

2.4. Dogs, feeding assessment, and palatability test

All procedures involving dogs were reviewed and approved by the Institutional Animal Care and Use Ethics Committee of the Mongolian University of Life Sciences (Approval number: VSBMR-2026/11) on 13 March 2026). The approved study covered the use of dogs for experimental evaluation of the newly developed pet food. All procedures were conducted in accordance with the approved protocol, with particular attention to animal welfare and the minimization of unnecessary stress or discomfort.

Ten clinically healthy adult dogs (five males and five females), 2–4 years of age, with a mean body weight of 17.8 ± 1.3kg and a body condition score of 4–5 on a 9-point scale, participated in the feeding assessment. Dogs underwent veterinary examination, routine vaccination, and deworming before the study. Feed was offered twice daily, and fresh water was available ad libitum. The feeding assessment lasted 28 days, comprising 21 days of adaptation followed by 7 days of observation. Daily amounts of food offered and refused were recorded. The individual dog was considered the experimental unit (n = 10). Repeated measurements obtained from the same dog were treated as within-dog observations rather than independent biological replicates. Feeding-performance variables were summarized as mean ± SD.

A two-bowl preference assessment was conducted using the same 10 adult dogs to compare the selected test food with a commercial control. Each dog was simultaneously offered 200g of the test food and 200g of the commercial control in separate bowls for 30min. The assessment was conducted once daily for 5 consecutive days, providing five observations per dog and 50 first-choice observations in total. Bowl positions were alternated between the left and right sides on successive test days to minimize positional bias. First choice was defined as the food first approached and consumed by the dog.

First-choice preference (%) was calculated as (number of first choices for the test food / total valid first-choice observations) × 100. The preference ratio was calculated as test-food intake / control-food intake, while the preference index was calculated as [test-food intake / (test-food intake + control-food intake)] × 100. Repeated observations from the same dog were considered within-dog measurements rather than independent biological replicates, and the individual dog was treated as the experimental unit (n = 10).

2.5. Statistical analysis

Data are presented as mean ± standard deviation. The source file reports one-way ANOVA with Tukey’s test, a chi-square test for first choice, and SPSS version 29.0. For a two-food comparison measured in the same dogs, paired analysis or a mixed model with dog as a random effect is generally more appropriate than one-way ANOVA. The exact analysis should be rerun from the raw dataset. Statistical significance was set at P < 0.05.

3. Results

3.1. Effect of drying on proximate composition

Drying reduced the moisture content of horse liver from 71.2% to 7.5–8.2% and that of lung from 78.2% to 7.9–8.1% (Table 2). On an as-is basis, crude protein increased from 22.09% to 41.3–41.62% in liver and from 17.9% to 41.26–41.68% in lung. These changes mainly reflect water removal. The dried organ values produced internally coherent proximate totals of approximately 99–100%. The relatively high crude protein concentrations indicate that dried horse liver and lung are concentrated protein-containing ingredients; however, these values should not be interpreted as evidence of superior protein quality because amino acid composition and protein digestibility were not evaluated in the present study (Table 2).

3.2. Iron and copper concentrations

Reported iron and copper concentrations were higher in dried than fresh organs on an as-is basis (Table 3). Iron increased from 213.2 mg/kg in fresh liver to 502–506 mg/kg after drying and from 112.3 mg/kg in fresh lung to 473–556 mg/kg. Copper increased from 8.1 to 10.9 mg/kg in liver and from 3.15 to 6.5–7.5 mg/kg in lung. Following hot-air drying, crude protein, fat, iron, and copper concentrations were higher when expressed per 100 g of product on an as-fed basis. However, these differences coincided with a substantial reduction in moisture content and therefore primarily reflect concentration of dry-matter components resulting from water removal. To account for this effect, nutrient concentrations were additionally expressed on a dry-matter basis. The dry-matter-based values provide a more appropriate comparison of the relative nutrient composition before and after drying and were therefore used for interpretation of the effects of the drying process (Table 3). Iron and copper concentrations were reported on an as-is basis. Therefore, the higher concentrations observed after drying should be interpreted primarily as a concentration effect associated with moisture removal rather than as evidence of increased mineral content or mineral retention.

3.3. Selected product composition and microbiological findings

The chemical composition, calculated metabolizable energy, and microbiological quality of the liver- and lung-based dry pet food products are presented in Table 4. Moisture contents were 7.10% and 7.50% in the liver- and lung-based products, respectively, and were below the maximum value specified in MNS 5688:2006 [17]. Both experimental formulations were characterized by relatively high crude protein concentrations. The liver-based product contained 42.01% crude protein, while the lung-based product contained 41.60%. Crude fat contents were 15.00% and 14.20%, respectively. Total carbohydrate concentrations were comparatively low, at 3.60% in the liver-based formulation and 3.80% in the lung-based formulation. Ash concentrations were similar between the two products, with values of 6.75% and 6.72%, respectively. Calculated metabolizable energy was 287.14 kcal/100 g in the liver-based product and 279.60 kcal/100 g in the lung-based product. The slightly greater calculated energy density of the liver-based formulation corresponded with its marginally higher crude protein and crude fat concentrations. Microbiological analysis showed total bacterial counts of 3.0 × 10² CFU/g in the liver-based product and 2.0 × 10² CFU/g in the lung-based product. Escherichia coli and molds/fungi were not detected in either formulation under the applied analytical conditions. Overall, both experimental formulations met the evaluated chemical and microbiological criteria specified in MNS 5688:2006 [17]. However, these results should not be interpreted as a comprehensive food-safety assessment because water activity, Salmonella, other relevant pathogens, and storage stability were not evaluated (Table 4).

3.4. Feed intake

Daily intake, intake normalized to body weight, eating time, and food residue did not differ significantly between the commercial control and the test food. These results support short-term acceptance but do not establish digestibility, nutritional adequacy, or health benefits (Table 5).

3.5. Two-bowl preference assessment

The test food accounted for 68% of reported first-choice observations. Mean consumption was 186.4 ± 15.2 g for the test food and 131.3 ± 17.8 g for the control food, with a reported P-value of 0.013. The reported preference ratio was 1.42 ± 0.11 and preference index was 58.7 ± 3.8%. These outcomes suggest preference for the test food, although the raw repeated-measures dataset and statistical model are needed to confirm the inference (Table 6). The denominator and number of test days must be reported. First-choice results should not be described as “68% of dogs” unless the sample size supports that percentage.

4. Discussion

The principal finding of this preliminary study was that hot-air drying produced low-moisture horse liver and lung powders with high as-is crude protein concentrations. This result is technologically expected because removing water concentrates dry matter. The increase should therefore be described as concentration rather than generation or improvement of nutrients. Similar caution is needed when interpreting iron and copper because as-is concentrations rise as water is removed; mineral retention must be calculated from dry-matter values and material yield.

The crude protein concentration of the dried organs was high, but crude protein alone does not establish protein quality. Animal by-product meals vary in connective tissue, ash, amino acid composition, and digestibility [3,4,5,6]. Studies of dry pet foods have shown that raw-material type and prior processing can affect soluble protein, amino acid availability, and in vitro or in vivo digestibility [1,2,3,4]. Consequently, amino acid profiling, reactive lysine measurement, and a validated digestibility assay are necessary before the horse-organ ingredients can be characterized as nutritionally superior.

Thermal processing supports preservation and can improve desirable aroma and color, but the original statement that 85°C for 120 min was based on a “Billion reaction” was scientifically incorrect. The relevant process is the Maillard reaction. It can improve aroma and browning while simultaneously decreasing reactive lysine and forming Maillard reaction products [9,10,11]. Because the present study did not measure lysine availability or Maillard products, neither a benefit nor a nutritional penalty can be concluded from browning alone.

Short-term intake of the test food was comparable to the commercial control, and the reported two-bowl results suggest a preference for the horse-organ product. Palatability tests are useful for product development, but first choice, intake ratio, and absolute consumption are not interchangeable outcomes [12]. Recent studies demonstrate that animal-derived ingredients and selected by-products can increase preference without necessarily producing equivalent changes in digestibility or gastrointestinal outcomes [2,13,14,15]. The present palatability finding should therefore be limited to the tested conditions and population.

The microbiological results are encouraging but incomplete for a safety claim. Commercial dog food quality assessment requires validated sampling, defined detection limits, appropriate indicator and pathogenic organisms, and consideration of storage stability [2,18]. The current file does not report water activity, Salmonella testing, packaging permeability, replicate batches, or time-series microbiological measurements. A three-month refrigerated storage statement cannot be generalized to ambient shelf life without a controlled shelf-life study.

A major limitation is the internal inconsistency of the finished-product proximate results. Moisture, protein, fat, carbohydrate, and reported ash/minerals sum to more than 100%, indicating a transcription, unit, or analytical error. This inconsistency also undermines the calculated energy value and any claim of compliance with MNS 5688:2006 [17]. In addition, the reported body-weight-normalized intake is incompatible with the stated daily intake and average body weight. These values must be recalculated from raw laboratory and animal records before submission.

Finally, this study did not measure apparent total tract digestibility, amino acid balance, vitamin concentrations, calcium-to-phosphorus ratio, long-term clinical outcomes, or complete-diet adequacy. Therefore, claims that the food strengthens joints and bones, improves coat quality, supports digestive health, or is suitable for both dogs and cats are not supported by the presented data and were removed. The appropriate interpretation is that dried Mongolian horse liver and lung are promising high-protein ingredients that warrant further formulation and validation studies.

5. Conclusions

Hot-air-dried horse liver and lung showed potential as protein-rich ingredients for experimental dry dog food formulation. The developed foods showed acceptable short-term intake, and the two-bowl assessment indicated preference for the test food under the conditions evaluated. However, the present study did not assess amino acid balance, nutrient digestibility, vitamin concentrations, calcium-to-phosphorus ratio, long-term health outcomes, or complete-diet adequacy. Therefore, the findings should be limited to ingredient composition, short-term acceptance, and palatability, and should not be interpreted as evidence of complete nutritional adequacy or specific health benefits.