INTRODUCTION
Nutrient digestibility is a key factor in assessing the nutritional quality of companion animal diets, as it determines the proportion of ingested nutrients available for metabolic use [1]. Apparent total tract digestibility (ATTD) significantly impacts feed efficiency, fecal characteristics, and overall nutritional status. Meanwhile, undigested residues that reach the hindgut undergo microbial fermentation, producing metabolites that influence gut health and fecal quality [2,3]. Highly digestible diets are therefore particularly important for companion dogs, including the heterogeneous mixed-breed populations that account for a substantial fraction of the global pet-dog population [4].
Body size and breed are well-known modulators of canine digestive physiology: small-breed dogs exhibit higher mass-specific basal metabolic rates and shorter gastrointestinal transit times than large-breed dogs, while large-breed dogs have more developed cecal and colonic compartments that prolong transit time and increase fermentative activity [4–6]. These anatomical and functional differences suggest that digestive efficiency may vary across breeds; however, systematic in vivo digestibility data for the heterogeneous mixed-breed populations that dominate the global pet-dog population remain scarce.
Among breeds, copy-number variation in the amylase gene has been associated with breed-level differences in carbohydrate digestive capacity [7], implying that breed-specific adaptations may extend beyond carbohydrate metabolism to overall nutrient utilization. Validation of in vitro digestibility methods against in vivo data in genetically diverse populations is therefore needed if the resulting equations are to be applied outside of single-breed reference colonies.
In vitro digestibility methodologies have become accepted alternatives to in vivo trials due to their standardized conditions, rapid throughput, and reduced ethical concerns [8]. Previous studies have demonstrated a strong correlation between in vitro and in vivo digestion rates, supporting the validity of in vitro methods as a reliable approach for evaluating comparative digestibility [9,10].
Most existing canine digestibility research has used Beagles as a standardized model or studied populations without explicit breed specification [1,11], leaving the in vivo digestibility profile of mixed-breed dogs—and its agreement with in vitro estimates—largely uncharacterized across life stages. We hypothesized that the agreement between two-step in vitro digestion and in vivo ATTD would vary with both the nutrient and the life stage in mixed-breed dogs. Therefore, the present study was conducted to evaluate, across three life stages (puppies, adults, and seniors), the level of agreement between a two-step enzymatic in vitro digestion method and in vivo apparent total tract digestibility for dry matter, organic matter, crude protein, gross energy, crude fiber, and ether extract in mixed-breed dogs, and to derive life-stage–specific regression equations linking the two approaches.
MATERIALS AND METHODS
The experimental diet, comprising oat, turkey, and chicken breast meal as primary ingredients, was formulated in our laboratory specifically for this study. The diet was formulated to meet the AAFCO [12] nutrient profile for all life stages including growth, ensuring that nutrient supply was adequate for puppies, adult, and senior dogs without re-formulation (Table 1). Manufacturing was carried out under a contract arrangement at a pilot-scale extrusion facility, following a standard sequence of dry-ingredient grinding through a 1.0-mm screen, mixing in a horizontal ribbon mixer, steam conditioning, single-screw extrusion, drying to a shelf-stable moisture, perilla-oil coating, and ambient cooling.
Vitamin and mineral premix supplied per kg of diets: 3,500 IU vitamin A; 250 IU vitamin D3; 25 mg vitamin E; 0.052 mg vitamin K; 2.8 mg vitamin B1 (thiamine); 2.6 mg vitamin B2 (riboflavin); 2 mg vitamin B6 (pyridoxine); 0.014 mg vitamin B12; 6 mg Cal-d-pantothenate; 30 mg niacin; 0.4 mg folic acid; 0.036 mg biotin; 1,000 mg taurine; 44 mg FeSO4; 3.8 mg MnSO4; 50 mg ZnSO4; 7.5 mg CuSO4; 0.18 mg Na2SeO3; 0.9 mg Ca(IO3)2.
A two-step in vitro digestion procedure adapted from Hervera et al. [13] was performed in six independent replicates of the experimental diet. Each replicate consisted of a separately weighed subsample of the same bulk experimental diet, processed through the full two-step procedure described below.
Preparation phase: Prior to digestion, all samples were oven-dried at 65°C to a constant weight and subsequently ground into a fine powder with a particle size of less than 1.0 mm.
Gastric phase: Each container received 25 mL of phosphate buffer (0.1 M, pH 6.0) together with 10 mL of HCl solution (0.2 M, pH 0.7). The pH was brought to 2.0 by titrating with 1 M HCl and 1 M NaOH. Gastric digestion was initiated by the addition of 1 mL of pepsin solution (10 mg/mL; ≥ 250 units/mg solid, P7000, pepsin from porcine gastric mucosa; Sigma-Aldrich). To inhibit microbial contamination, 1 mL of chloramphenicol solution (C0378, chloramphenicol; Sigma-Aldrich; 5 g/L in ethanol) was also introduced. Flasks were then covered with Parafilm M® film and incubated at 39°C for 2 h in a shaking incubator (SWB-35; Hanyang Science Lab).
Small intestinal phase: Once cooled to room temperature, each flask received 5 mL of 0.6 M NaOH and 10 mL of phosphate buffer (0.2 M, pH 6.8), followed by pH adjustment to 6.8 with 1 M HCl and NaOH. Intestinal digestion was then simulated by adding 1 mL of pancreatin solution (100 mg/mL; 4 × USP, P1750, pancreatin from the porcine pancreas; Sigma-Aldrich). The flasks were sealed with Parafilm M® film and incubated at 39°C for 4 h with continuous agitation.
Sample collection and filtration phase: Following digestion, undigested residues were recovered by vacuum filtration through pre-weighed, pre-dried glass filter crucibles (Gooch Type Filter Crucibles, PYREX®). Each flask was rinsed three times with distilled water during the filtration step. Residues in the crucibles were subsequently washed with two sequential additions of 10 mL of 95% ethanol and 10 mL of 99.5% acetone.
To determine dry matter (DM) content following in vitro digestion, filter crucibles containing undigested residues were oven-dried at 70°C for 24 h. Subsequent ashing at 550°C for 4 h was conducted to obtain organic matter (OM) values. Crucibles were cooled to room temperature before weighing. All proximate analyses followed AOAC methods [14]: DM (method 930.15), OM (method 942.05), crude fiber (CF; method 978.10), and ether extract (EE; method 920.39). Crude protein (CP) was quantified by the Dumas combustion method (Rapid MAX N-Exceed, Elementar), and gross energy (GE) was measured using a bomb calorimeter (Parr 6400 Bomb Calorimeter, Parr Instrument).
In vitro DM digestibility was calculated as follows:
In vitro digestibility of OM, CP, GE, CF, and EE was derived using the equation below:
where Nr = nutrient concentration in residues (DM %), Nd = nutrient concentration in diet (DM %), and IDDM = in vitro DM digestibility (%).
This experiment received approval (approval # 202412-CNU-409) from the Institutional Animal Care and Use Committee of Chungnam National University, Daejeon, Korea. Dogs were housed and handled in accordance with the approved procedures throughout the study.
Eighteen mixed-breed dogs of comparable medium body size (mature weight approximately 8–11 kg) were used in this study. The dogs were derived from medium-sized parent stock, were of unidentified pedigree, and showed no dominant morphological traits of any single recognized breed. Animals were allocated to three life-stage groups: six puppies (< 1 year old; initial body weight, BW: 10.68 ± 1.55 kg), six adults (2–7 years old; BW: 8.34 ± 0.38 kg), and six seniors (> 8 years old; BW: 8.87 ± 0.98 kg). Within each life stage, sex was balanced (3 males and 3 females). Each dog was individually housed in a kennel maintained at 23°C. The total study duration was 17 days, with the first 7 days designated as an adaptation period. Maintenance energy requirements (MER) were estimated on the basis of metabolic BW (mBW) using the following equations:
Calculating the MER used the following formula:
Daily feed requirements of each dog were determined using MER, and the dogs were fed twice a day at 9:00 and 17:00.
The ATTD of DM, OM, CP, GE, CF, and EE were determined using 0.5% Cr2O3 incorporated into the diet as an indigestible external marker. Fresh fecal samples were collected on days 3 through 6 of the collection period. Both freshly collected fecal samples and the corresponding diet samples were immediately frozen at −20°C for subsequent analysis. Upon completion of the experiment, fecal samples were oven-dried at 70°C for 72 h and ground through a 1 mm screen. Nutrient digestibility of DM, OM, CP, GE, CF, and EE was analyzed using the dried fecal samples. Analytical methods for DM (method 930.15), OM (method 942.05), CF (method 978.10), and EE (method 920.39) followed AOAC methods [15]. CP and GE were determined by Dumas combustion (Rapid MAX N-Exceed, Elementar) and bomb calorimetry (Parr 6400 Bomb Calorimeter, Parr Instrument), respectively.
The ATTD of nutrients was calculated using the following formula:
where Nf = concentration of nutrient in feces, Nd = concentration of nutrient in the diet, Cd = concentration of Cr2O3 in the diet, and Cf = concentration of Cr2O3 in the fecal.
For the in vivo experiment, the individual dog was considered the experimental unit; for the in vitro experiment, each independent digestion replicate was treated as the experimental unit. Differences in digestibility among life-stage groups were tested by one-way analysis of variance (ANOVA) followed by Tukey’s honestly significant difference (HSD) test for pairwise comparisons. The assumptions of normality of residuals (Shapiro–Wilk test) and homogeneity of variance (Levene’s test) were verified prior to ANOVA; where assumptions were violated, the non-parametric Kruskal–Wallis test was applied as a sensitivity check, and the conclusions were unchanged. Orthogonal contrasts were used to compare the in vitro method against the in vivo measurements within each life stage. The relationship between the two methods was examined by linear regression analysis within a general linear model (GLM) framework. All statistical computations were carried out using JMP® Pro version 16.0.0 (SAS Institute). Differences were regarded as statistically significant at p < 0.05.
RESULTS
The in vitro and in vivo digestibility of DM, OM, CP, GE, CF and EE in puppies, adult dogs, and senior dogs are presented in Table 2. The in vitro digestibility of DM, OM, CP, CF, and EE was significantly higher (p < 0.05) than in vivo digestibility in all life stages. The in vitro digestibility of GE was also significantly higher (p < 0.05) than in vivo digestibility in all life stages except puppies.
Among in vivo measurements, a significant life-stage effect was detected only for gross energy digestibility (p < 0.05), with senior dogs showing a lower value than adult dogs (p < 0.05). For the remaining nutrients (DM, OM, CP, CF, and EE), no significant life-stage difference was observed (p > 0.05; Table 3).
The statistical relationships between in vitro and in vivo digestibility as linear regression equations are shown in Table 4. In puppies, strong correlations were observed for GE, CF, and EE (r² = 0.93, 0.87, and 0.91, respectively), whereas DM, OM, and CP showed moderate correlations (r² = 0.60, 0.40, and 0.55, respectively). In adult dogs, OM and GE exhibited the strongest correlations (r² = 0.96 and 0.98, respectively), followed by CF (r² = 0.85), while DM, CP, and EE showed moderate correlations (r² = 0.67, 0.56, and 0.51, respectively). In senior dogs, strong correlations were observed for most nutrients, with EE, DM, and GE showing the highest values (r² = 0.93, 0.91, and 0.87, respectively), whereas CP exhibited a moderate correlation (r² = 0.66).
DISCUSSION
The present study assessed and compared nutrient digestibility in mixed-breed dogs at three life stages: puppies, adults, and seniors, both in vitro and in vivo. Consistently, the in vitro digestibility values were higher than the corresponding in vivo values across all measured nutrients, aligning with findings from previous studies [13,15,16]. This systematic overestimation in in vitro methods can be attributed to the lack of consideration for endogenous nitrogen and fat losses that occur in vivo. In living animals, fecal output includes not only undigested dietary residues but also epithelial cell debris, mucus secretions, and bacterial biomass [17,18]. Furthermore, the two-step enzymatic procedure does not account for hindgut microbial activity, where resident bacteria modify undigested substrates through fermentation [19,20]. Additionally, the static nature of in vitro systems fails to replicate the dynamic variations in enzyme secretion, gastric emptying rates, and intestinal motility that occur in response to meal composition and individual physiological status [21,22].
Contrast analysis showed a progressive decline in in vivo digestibility values from puppies to seniors for most nutrients. This age-related decrease in digestive efficiency can be attributed to several physiological changes associated with aging, including reduced pancreatic enzyme secretion, decreased brush border enzyme activity, and a diminished intestinal absorptive surface area due to villous atrophy [23,24]. Prior studies have also shown that senior dogs experience alterations in gut microbiota composition, characterized by reduced bacterial diversity and lower populations of beneficial fermentative bacteria [3,25], which may further hinder nutrient utilization efficiency. The smallest difference between in vitro and in vivo values observed in puppies suggests that their heightened digestive capacity during growth, marked by increased enzyme secretion and enhanced absorptive function, more closely resembles the optimal conditions simulated in vitro [26]. The rapid tissue accretion and organ development during the growth phase necessitate maximal nutrient extraction efficiency, and the developing gastrointestinal system appears optimized for high digestive performance [27,28].
The strong in vitro–in vivo agreement observed in senior dogs (DM, GE, EE r² = 0.87–0.93), despite their lower mean digestibility, can be reconciled by noting that the in vitro method captures relative differences in substrate susceptibility to enzymatic hydrolysis, whereas the absolute reduction in in vivo digestibility in senior dogs reflects age-related changes in pancreatic enzyme secretion, mucosal surface area, and microbial activity that act as a roughly uniform downward shift across samples. Because such a uniform shift preserves the rank order of samples, the slope of the regression line is reduced but its r² remains high. The relatively narrow within-group variability among senior dogs further contributed to the high coefficients of determination.
The linear regression analysis revealed that gross energy and organic matter digestibility had the strongest correlations between in vitro and in vivo methods. This suggests that the two-step enzymatic procedure most effectively simulates the digestion of energy-yielding macronutrients. The sequential hydrolysis using pepsin and pancreatin accurately mimics the gastric and small intestinal phases, where most lipid and carbohydrate digestion occurs. The chemical bonds cleaved during in vitro incubation closely resemble the enzymatic reactions occurring in vivo [8,13]. CP digestibility showed moderate correlations, likely due to the complexity of protein digestion, which involves the sequential action of multiple proteolytic enzymes with specific substrate affinities. Additionally, the resistance of different protein sources to enzymatic hydrolysis varies based on their amino acid composition, tertiary structure, and thermal processing history [11,29]. Prior studies have indicated that protein digestibility is particularly sensitive to ingredient quality and processing conditions. For instance, Maillard reaction products formed during extrusion can reduce protein bioavailability in ways not fully captured by in vitro methods [17,30]. Adult dogs exhibited the most consistent correlations across all nutrient parameters, reflecting the physiological stability characteristic of the maintenance life stage. At this stage, digestive enzyme profiles, gut microbiota composition, and intestinal morphology reach a mature equilibrium [31,32].
Because all observations originated from a single extruded oat–poultry-based diet, the regression coefficients reported here should be interpreted as life-stage–specific calibrations valid for diets of comparable composition, rather than as universal predictive equations. Within the present scope, the regression equations describe the distributional agreement between in vitro and in vivo digestibility values and may help reduce the need for repeated in vivo trials when evaluating closely related formulations [1]. The equations derived from adult dogs showed the highest agreement, making them the most informative reference among the three life-stage models for quality control of pet food formulations of similar matrix [22,30].
It should be noted that the sequential ethanol and acetone washes used in the filtration step of the present in vitro protocol (adapted from Hervera et al. [13]) may solubilize part of the residual lipid fraction, potentially shifting the absolute in vitro EE digestibility upward. Because all in vitro samples were processed identically, this offset is expected to be systematic across life stages and should not affect the relative comparison or the slope of the in vivo–in vitro regression. Consistent with this interpretation, the in vitro–in vivo EE agreement was among the strongest observed in the present dataset (r² = 0.91 in puppies, r² = 0.93 in senior dogs), indicating that the in vitro EE measurement tracks the corresponding in vivo digestibility with high fidelity despite the wash step. The absolute in vitro EE values reported here may nonetheless carry a small upward offset relative to a lipid-conserving wash protocol, and a comparative evaluation of the two wash schemes could further refine these absolute estimates in future studies.
Additionally, the in vitro approach supports systematic comparisons of novel ingredients and processing techniques, aiding in the development of optimized formulations tailored to specific physiological needs [19,33]. Moreover, the use of validated in vitro methods can reduce the number of animals needed for nutrient digestibility assessments while maintaining scientific rigor [34].
CONCLUSION
The two-step in vitro digestion method showed strong agreement with in vivo ATTD for energy and organic matter in adult dogs (r² = 0.96–0.98), and for ether extract, dry matter, and energy in senior dogs (r² = 0.87–0.93). Agreement was weaker for crude protein in all life stages (r² = 0.55–0.66) and for organic matter in puppies (r² = 0.40), indicating that the predictive value of the in vitro method is nutrient- and life-stage–dependent and should not be generalized uniformly. Within the scope of the single experimental diet and the mixed-breed population studied here, the regression equations derived from adult dogs may serve as a practical screening reference for diet evaluation, while broader application to chemically distinct diets will require additional multi-diet validation.
