PDF Download
PubReader
Export Citation
Email the Author
Share Facebook
Share Twitter
Cited by 0 Article
Metrics (View:181)
INTRODUCTION
As the prices of corn, wheat, and soybean meal have risen with fluctuations [1,2], alternative sources have gained increasing attention as a practical solution to reduce feed costs and reliance on traditional feed ingredients [1]. Because nutrient concentrations, including protein and fiber, vary among the alternative feed ingredients, an accurate evaluation of digestibility of nutrients in alternatives is necessary to replace the conventional sources in swine diets [3].
The determination of nutrient digestibility requires animal experiments which are expensive, laborious, and time-intensive. In contrast, in vitro procedures mimicking the digestive and absorptive systems in the gastrointestinal tract of pigs to estimate nutrient digestibility in feed ingredients are cost-effective and time-saving [4–7]. In vitro procedures have been used for evaluating feed ingredients in previous studies [8,9]. However, in vivo nutrient digestibility of feed ingredients may not be accurately estimated solely by in vitro disappearance due to the differences in nutrient composition affecting digestion under in vivo and in vitro conditions [10]. The addition of nutrient concentrations as independent variables may increase the prediction accuracy for nutrient digestibility in pigs. However, to our knowledge, most prediction equations estimating in vivo nutrient digestibility were based on in vitro disappearance as the sole independent variable [5,11]. For that reason, the objectives of this study were to assess nutrient digestibility of various feed ingredients using in vitro procedures and to develop prediction equations for estimating standardized ileal digestibility (SID) of crude protein (CP) and apparent total tract digestibility (ATTD) of gross energy (GE) using in vitro nutrient disappearance and nutrient composition in conventional and alternative feed ingredients fed to pigs.
MATERIALS AND METHODS
Ten feed ingredients, including rice, corn, soybean hulls, wheat, wheat bran, palm kernel expellers, copra meal, cashew nuts, rapeseed meal, and soybean meal, were employed to develop prediction equations for estimation of in vivo nutrient digestibility using in vitro nutrient disappearance and nutrient concentrations (Table 1). The 10 feed ingredients were chosen to provide a variety of CP and fiber concentrations.
aNDF, amylase-treated neutral detergent fiber; ADF, acid detergent fiber.
The values for SID of CP in rice, wheat, and cashew nuts and the values for ATTD of GE in rice, corn, soybean hulls, wheat, wheat bran, palm kernel expellers, copra meal, cashew nuts, and rapeseed meal were from in-house data (Table 2). A total of 44 data for SID of CP in corn, soybean hulls, wheat bran, palm kernel expellers, copra meal, rapeseed meal, and soybean meal were collected from 21 published papers [12–32]. A total of 19 data for ATTD of GE in soybean meal were collected from 9 published papers [8,11,12,15,16,23,30,31,33]. In the search conducted on PubMed and Google Scholar, the keywords included SID of CP, ATTD of GE, feed ingredients, and pigs. The research articles were manually selected according to the title and nutrient composition of the test ingredients, including CP, fat, fiber, and ash. During the process of screening, data of sows and weaned piglets were removed. The averaged values were calculated and used for each ingredient when multiple data were available for an ingredient.
1) A total of 44 data for SID of CP in corn, soybean hulls, wheat bran, palm kernel expellers, copra meal, rapeseed meal, and soybean meal fed to growing pigs were obtained from 21 research papers and a total of 19 data for ATTD of GE in soybean meal fed to growing pigs were obtained from 9 research papers [8,11–33]. The values for SID of CP in rice, wheat, and cashew nuts, as well as the values for ATTD of GE in rice, corn, soybean hulls, wheat, wheat bran, palm kernel expellers, copra meal, cashew nuts, and rapeseed meal were determined by our research group as marked with hyphens.
In vitro ileal disappearance (IVID) of dry matter (DM) and CP in 10 feed ingredients was evaluated by mimicking the digestive system of pigs based on the procedure developed in the literature [6,34]. The particle size of test ingredients was reduced to less than 1.0 mm using a grinder before the in vitro assays. In step 1, one gram of each ingredient was placed into a flask with a capacity of 200 mL, followed by supplementation with 25 mL of 0.1 M sodium phosphate buffer solution (pH 6.0) and 10 mL of 0.2 M HCl (pH 0.7). To mimic the pigs’ stomach digestion, HCl (1 M) and NaOH (1 M) were poured into the flask to make pH 2.0. Then, 1.0 mL of 10-mg/mL pepsin solution (P7000, Sigma-Aldrich) was poured into the flask. To prevent fermentation by microorganism, 0.5 mL of chloramphenicol solution (5.0 g/L ethanol; C0378, Sigma-Aldrich) was added. Then, the flasks were agitated using a shaking incubator (LSI-3016R, Daihan Labtech, Namyangju, Korea) at 39°C for 6 hours.
The next step was performed to mimic the small intestine digestive system of pigs. Initially, 10 mL of 0.2 M phosphate buffer solution (pH 6.8) and 5 mL of 0.6 M NaOH solution were added to each flask. Next, HCl (1.0 M) and NaOH (1.0 M) were poured into the flasks to adjust the pH to 6.8, followed by the addition of 1.0 mL of 50-mg/mL pancreatin solution (P1750, Sigma-Aldrich). The flasks were agitated using the shaking incubator at 39°C for 18 hours. Then, 5 mL of sulfosalicylic acid solution (20%) was added to each flask, and the flasks were kept at room temperature for 30 minutes to precipitate undigested protein. After precipitation for 30 minutes, the undigested samples were filtered using glass filter crucibles (CFE Por. 2; ROBU® Glasfilter-Geraete GmbH, Hattert, Germany), which had been pre-weighed and had contained 0.5 g of celite (Celite 545, Daejung Chemicals & Metals, Siheung, Korea). Then, 10 mL of 95% ethanol was poured twice to wash glass filter crucibles containing undigested samples, followed by 99.5% acetone washing twice. After that, undigested samples in the crucibles were dried at 80°C for 24 hours. After being cooled in a desiccator for 1 hour, IVID of DM in the test ingredients was calculated by weighing the crucibles. The undigested residues in the filter crucibles were collected for CP analysis to calculate the IVID of CP. The blank values were obtained to correct the contents of DM and CP in the undigested residues not derived from feed ingredient samples, based on the prediction equations using amounts of pepsin and pancreatin added [4]. Three replicates were conducted for each ingredient.
In vitro total tract disappearance (IVTTD) of DM and organic matter (OM) in 10 test feed ingredients was evaluated by mimicking the digestive system of pigs based on the procedure developed in the literature [5,34]. In steps 1 and 2, only sample weight, enzymes concentration, and incubation period were altered from 2-step in vitro assays. For evaluating the IVTTD of DM and OM in feed ingredients, 0.5 g of each sample was digested with pepsin solution (25 mg/mL) in step 1 and pancreatic solution (100 mg/mL) in step 2 with the incubation periods of 2 and 4 hours in step 1 and 2, respectively. In step 3 for mimicking the condition of large intestine fermentation, 10 mL of 0.2 M EDTA solution was added to each flask. Then, 30% acetic acid or 1 M NaOH was added to adjust the pH to 4.8. The samples in the flasks were supplemented with 0.5 mL of a cellulolytic enzyme mixture (Viscozyme® L, Sigma-Aldrich) and agitated in the incubator at 39°C for 18 hours. The samples were filtered using glass filter crucibles as suggested in IVID assays, and the residues in the crucibles were dried at 130°C for 6 hours. Additionally, the ash content in the residues was measured for calculating the IVTTD of OM in test ingredients. A blank was included during the IVTTD procedure to correct the contents of residual DM and OM, which were considered to derive from exogenous enzymes. Three replicates were conducted for each ingredient.
Prior to chemical analyses, the particle sizes of test ingredient samples were reduced by grinding to pass through a 1.0-mm screen. DM (method 930.15), CP (method 990.03), OM (method 942.05), ether extract (EE; method 920.39), ash (method 942.05), amylase-treated neutral detergent fiber (aNDF; method 2002.04), and acid detergent fiber (method 973.18) in feed ingredients were determined according to the AOAC [35]. Gross energy contents in test ingredients were also determined using a bomb calorimeter (6400 Automatic Isoperibol Calorimeter, Parr Instrument Company, Moline, IL, USA).
The calculations of in vitro DM disappearance followed the equations below [7]:
where DMing (g) is the quantity of DM in feed ingredients, DMresidues (g) is the quantity of DM residues after the IVID or IVTTD procedure, and DMblank (g) is the quantity of DM residues in the blank estimated using an estimation model for the 2-step in vitro digestion procedure [4]. The quantity of DM residues in the blank after the IVTTD procedure was measured. The calculation of IVID of CP followed the equation below [7]:
where CPing, CPresidues, and CPblank are the CP contents in a test ingredient, the undigested residues, and the blank value estimated using the exogenous enzyme doses according to the literature [4], respectively. The calculation of IVTTD of OM followed the equation below [7]:
where OMing (g) is the quantity of OM in feed ingredients, OMresidues (g) is the quantity of OM in the undigested residues after the IVTTD procedure, and OMblank (g) is the quantity of OM in the blank after the IVTTD procedure.
The GLM procedure of SAS (SAS Inst., Cary, NC, USA) was used for data analyses. The fixed effect was test ingredient and least-square means were calculated for IVID and IVTTD of nutrients. The comparison of mean values was made with Tukey’s adjustment using the PDIFF option. The experimental unit was a flask. The CORR procedure of SAS was used for correlation analyses among nutrient compositions, in vitro disappearance, SID of CP, and energy digestibility. The REG procedure of SAS was used to develop novel prediction equations estimating SID of CP or ATTD of GE based on IVID of CP, IVTTD of DM, nutrient compositions, or both as independent variables. The statistical significance and tendency levels were declared at p < 0.05 and p < 0.10.
RESULTS
In the 10 feed ingredients, the concentrations of CP ranged from 8.2% to 48.5%, EE ranged from 0.3% to 47.8%, aNDF ranged from 1.0% to 64.7%, ADF ranged from 0% to 47.4%, and ash ranged from 0.8% to 8.4% on an as-is basis (Table 1). The range of GE contents in test ingredients was from 3,610 to 6,524 kcal/kg. The average SID of CP and ATTD of GE in the 10 test ingredients ranged from 62.3% to 96.0% and from 58.2% to 96.4%, respectively (Table 2).
The IVID of DM in wheat was the greatest (p < 0.05), followed by rice, corn, and cashew nuts (Table 3). The IVID of CP in cashew nuts and soybean meal was the greatest (p < 0.05), followed by rice, wheat, and rapeseed meal. Soybean hulls showed the lowest IVID of DM and CP (p < 0.05). The IVTTD of DM in rice was the greatest (p < 0.05), followed by cashew nuts, soybean meal, wheat, and corn. The IVTTD of OM in rice, cashew nuts, and soybean meal were the greatest (p < 0.05), followed by corn, wheat, and rapeseed meal. Palm kernel expellers showed the lowest IVTTD of DM and OM (p < 0.05).
Each least squares mean represents 3 observations.
The EE was positively correlated with GE (r = 0.97; p < 0.001; Table 4). The SID of CP showed a negative correlation with aNDF (r = −0.81; p < 0.01). The SID of CP was positively correlated with IVID of CP (r = 0.82; p < 0.01). The ATTD of GE was negatively correlated with aNDF (r = −0.92; p < 0.001). The ATTD of GE was positively correlated with IVTTD of DM (r = 0.89; p < 0.001).
The most suitable model for SID of CP was: SID of CP (%) = 16.55 + 0.89 × IVID of CP – 2.00 × ash with R2 = 0.89, and p < 0.001 (Table 5). The most suitable model for ATTD of GE was: ATTD of GE (%) = 42.68 + 0.57 × IVTTD of DM – 2.27 × ash with R2 = 0.94, and p < 0.001 (Table 6).
| Item | Regression coefficient parameter | Statistical parameter | |||||
|---|---|---|---|---|---|---|---|
| Intercept | IVID of CP | Ash | aNDF | RMSE | R2 | p-value | |
| Equation 1 | 11.59 | 0.85 | - | - | 6.28 | 0.67 | 0.004 |
| Standard error | 16.85 | 0.21 | - | - | - | - | - |
| p-value | 0.511 | 0.004 | - | - | - | - | - |
| Equation 2 | 42.95 | 0.53 | - | −0.19 | 5.19 | 0.81 | 0.003 |
| Standard error | 20.02 | 0.23 | - | 0.09 | - | - | - |
| p-value | 0.069 | 0.054 | - | 0.066 | - | - | - |
| Equation 3 | 16.55 | 0.89 | −2.00 | - | 3.88 | 0.89 | < 0.001 |
| Standard error | 10.49 | 0.13 | 0.53 | - | - | - | - |
| p-value | 0.159 | < 0.001 | 0.007 | - | - | - | - |
aNDF, amylase-treated neutral detergent fiber; RMSE, root mean square of error.
| Item | Regression coefficient parameter | Statistical parameter | |||||
|---|---|---|---|---|---|---|---|
| Intercept | IVTTD of DM | Ash | aNDF | RMSE | R2 | p-value | |
| Equation 1 | 28.29 | 0.64 | - | - | 6.74 | 0.78 | 0.001 |
| Standard error | 9.30 | 0.12 | - | - | - | - | - |
| p-value | 0.016 | 0.001 | - | - | - | - | - |
| Equation 2 | 76.69 | 0.15 | - | −0.39 | 5.87 | 0.86 | 0.001 |
| Standard error | 26.98 | 0.28 | - | 0.21 | - | - | - |
| p-value | 0.025 | 0.608 | - | 0.102 | - | - | - |
| Equation 3 | 42.68 | 0.57 | −2.27 | - | 3.89 | 0.94 | < 0.001 |
| Standard error | 6.41 | 0.07 | 0.55 | - | - | - | - |
| p-value | < 0.001 | < 0.001 | 0.005 | - | - | - | - |
aNDF, amylase-treated neutral detergent fiber; RMSE, root mean square of error.
DISCUSSION
In the present study, the analyzed GE, CP, EE, fiber, and ash contents in the test ingredients were consistent with the values found in the previous studies [8,11–13,15,36]. Among the evaluated ingredients, cashew nuts contained the highest GE content because of the greater EE concentration compared to the other test ingredients [37].
The IVID of CP values in soybean hulls, wheat bran, copra meal, and soybean meal determined in the present study were within the range of data reported in the previous research [1,34,38]. The greater IVID of CP in cashew nuts, soybean meal, and rice compared with the other test ingredients is likely because of the relatively low fiber concentration in cashew nuts, soybean meal, and rice [39,40]. Fibers are more difficult to digest compared to starch, protein, and lipids because pigs excrete fewer endogenous fiber-degrading enzymes [41]. Additionally, fibers in feed ingredients negatively affect the digestibility of other nutrients surrounded by or bonded to fiber fractions, hindering nutrient breakdown and absorption [39,40]. In the in vitro condition, exogenous enzymes may be hindered by fiber fractions, resulting in a decrease in in vitro disappearance [1]. By the same token, rice showed the greatest IVTTD of DM and OM, and palm kernel expellers showed the lowest IVTTD of DM and OM among the test ingredients mainly because of the fiber concentrations. The values for IVTTD of DM and OM in corn, soybean hulls, wheat, wheat bran, palm kernel expellers, copra meal, rapeseed meal, and soybean meal determined in the present study were within the range of data reported in the previous studies [1,2,11,34,42–44].
The positive relation of GE and EE observed in this work may be a result of the high energy content of lipids [37]. This study found a negative relation between aNDF and SID of CP or ATTD of GE, supporting that fiber is not digested by pigs and disrupts the digestion of other nutrients in feed ingredients, which is confirmed by previous studies [1,45,46]. The positive correlations between IVID of CP and SID of CP, as well as IVTTD of DM and ATTD of GE observed in the present study are in agreement with previous studies that reported relations between in vitro disappearance and in vivo digestibility, indicating that in vitro assays can be adopted to predict SID of CP or ATTD of GE in pigs [5,6,8].
Incorporating aNDF as an additional independent variable in addition to IVID of CP in prediction equations for SID of CP improved the prediction accuracy as the SID of CP was highly correlated with aNDF concentration in feed ingredients. With the same token, the inclusion of aNDF in the prediction model for ATTD of GE increased the determination coefficient as ATTD of GE was highly correlated with aNDF in feed ingredients. However, considering the strong negative correlation between aNDF and SID of CP (r = −0.81, p < 0.01) or ATTD of GE (r = −0.92, p < 0.001), the increased determination coefficients by the inclusion of aNDF in the models for SID of CP and ATTD of GE were not that large (ΔR2 = 0.14 and 0.08, respectively). The reason for the lack of large influence of aNDF on the accuracy of prediction models is likely due to the high correlation between aNDF and in vitro disappearance, indicating that the effects of aNDF on in vivo digestibility may already have been reflected when in vitro assays were conducted.
Incorporating ash as an independent variable in addition to in vitro nutrient disappearance for estimating SID of CP or ATTD of GE enhanced the prediction accuracy represented by determination coefficients (ΔR2 = 0.22 and 0.16, respectively) compared with the prediction equations with in vitro nutrient disappearance as a sole independent variable. Interestingly, ash content was not correlated with SID of CP and ATTD of GE. The reason for the improved prediction accuracy by the inclusion of ash as an additional independent variable remains unclear. However, previous studies reported that ash contents are negatively correlated with in vivo CP digestibility, the amount of digestible CP, and energy utilization in feed ingredients fed to pigs, likely because ash may interfere with nutrient digestion [3,8,47]. Similarly, ash contents in the ingredients tended to be negatively correlated with ATTD of GE (r = −0.59, p = 0.074) in the present study. However, the correlation between ash contents and SID of CP (r = −0.40, p = 0.254) was not significant which may be due to the lack of sufficient number of ingredients or sufficiently large range of ash contents or both in the present study. In addition, the effects of ash on in vivo digestibility of energy and nutrients may not be fully reflected under in vitro conditions. Further research is warranted to investigate the role and mechanisms of ash in the prediction models for energy and nutrient digestibility.
In conclusion, SID of CP and ATTD of energy in feed ingredients fed to growing pigs can be estimated using prediction models developed in the present study based on in vitro nutrient disappearance and ash concentrations as independent variables.
