Validation of in vitro digestion using simulated small intestinal fluid with

specific digestive activity to predict the metabolizable energy of feed

ingredients for duck

L. Zhang,∗ F. Zhao,∗,1 H. Zhang,∗ G. Z. Bian,† Y. M. Wang,∗ X. Yang,∗ and H. Li†

∗
The State Key Laboratory of Animal Nutrition, Institute of Animal Sciences, Chinese Academy of Agricultural
Sciences, Beijing 100193, China; and † Guangdong Haid Group Co., Ltd, Guangzhou 511400, China

ABSTRACT Two experiments were conducted to val-
idate a method to prepare simulated small intestinal
fluid (SSIF) for in vitro digestion in ducks. Experi-
ment 1 compared the in vitro digestible energy (IVDE)
of SSIF to endogenous small intestinal fluid (ESIF) on
four feeds. The ESIF 1 or 2 obtained from two groups of
jejunal cannulated ducks offered diet 1 (3,050 kcal/kg
of ME and 19.95% of CP) or 2 (2,801 kcal/kg of ME
and 14.90% of CP) was purified into raw enzyme power
(REP) 1 or 2. SSIF 1 to 3 or 4 to 6 were prepared to
mimic ESIF 1 or 2, respectively. The enzyme sources
were REP 1 for SSIF 1 and 4, REP 2 for SSIF 2 and
5 or reagent enzymes for SSIF 3 and 6, respectively.
The IVDE of each feed was determined with SSIF or
ESIF. Experiment 2 was to validate whether REP 1 was
more effective than only reagent enzymes to prepare
SSIF. Ten feeds were determined with pepsin following
Key words: simulated small intestinal fluid, duck, in vitro digestion, metabolizable energy
SSIF 1 or 3 for IVDE 1 or 2, respectively. The accu-
racy of prediction model of true metabolizable energy
(TME) from IVDE 1 or 2 was evaluated to validate
the efficacy of SSIF. In experiment 1, higher activities
of amylase, trypsin and chymotrypsin were observed in
ESIF 1 than ESIF 2 (P < 0.05). The IVDE determined
with SSIF 1 and 2 or 3 and 4 were more comparable
to that of ESIF 1 or 2 than determinations with SSIF
3 or 6. In experiment 2, the mean IVDE 1 or 2 was
97.22% or 96.23% relative to TME, respectively, and
both were highly correlated with TME (P < 0.01; R2
≥ 0.98). However, the residual SD of TME prediction
model with IVDE 1 was less than that generated with
IVDE 2 (55 vs. 71 kcal/kg). In conclusion, the IVDE
determined with in vitro digestion of pepsin following
SSIF prepared with REP can predict accurately TME
of feed for ducks.
2019 Poultry Science 98:1280–1287
http://dx.doi.org/10.3382/ps/pey450

INTRODUCTION of porcine (Valdes and Leeson, 1992) or bovine (Bryan

Accurate energetic values of feed ingredients are crit-
ical to formulating diets for poultry. However, in vivo
metabolizable energy (ME) assays are time-consuming,
expensive, and laborious. Animal nutritionists have de-
veloped several in vitro digestion procedures to rapidly
predict effective energy of feeds (Boisen and Eggum,
1991; Valdes and Leeson, 1992; Boisen and Fernádez,
1997; Regmi et al., 2008; Losada et al., 2009, 2010).
In vitro digestion procedures have established dynamic
data of feed ingredients for pigs (INRA-AFZ-INAPG,
2004), but in vitro digestible energy (IVDE) can-
not accurately predict ME for poultry (Clunies and
Leeson, 1984; Clunies et al., 1984; Valdes and Lee-
son, 1992; Losada et al., 2009, 2010). Generally, sim-
ulated small intestinal fluid (SSIF) of poultry is made

C 2018 Poultry Science Association Inc.
Received April 19, 2018.
Accepted September 2, 2018.
1
Corresponding author: zsummit@hotmail.com
et al., 2018) pancreatin or porcine small intestinal
fluid (Sakamoto et al., 1980; Clunies and Leeson, 1984;
Clunies et al., 1984). The homogeneity between SSIF
and in vivo intestinal fluid is challenging for several
reasons. Firstly, the digestibility of SSIF prepared with
the wt/vol or vol/vol concentration of pancreatin or
endogenous small intestinal fluid (ESIF) is not re-
peatable in different laboratories because the activi-
ties of digestive enzymes vary across sources (Furuya
et al., 1979; Sakamoto et al., 1980; Clunies et al., 1984;
Valdes and Leeson, 1992). Secondly, SSIF of rooster is
sometimes composed of enzymes derived from livestock
(Valdes and Leeson, 1992), the efficacy of which can dif-
fer from that of the endogenous digestive enzymes se-
creted by poultry (Crevieu-Gabriel et al., 1999). Lastly,
the activities of digestive enzymes in SSIF do not
match in vivo values of host animals because little data
were reported on the composition of in vivo intestinal
fluid in poultry.
We hypothesized that determinations of IVDE would
differ when using SSIF prepared with raw enzymes

1280
 NOVEL SIMULATED DIGESTIVE FLUID TO PREDICT ME 1281
power (REP) purified from ESIF compared to SSIF Table 1. Composition and nutrient contents of diet for ducks
made with reagent sources of amylase, trypsin and (as-fed basis).
chymotrypsin. Using ducks as an animal model, the ob- Ingredient Basal diet Diet 1 Diet 2
jectives of this study were to (1) compare the IVDE of
feed determined with ESIF or SSIF, which prepared Corn, % 70.34 66.24 66.24
Soybean meal, % 23.95 25.54 17.69
with REP or only reagent enzymes and (2) validate Soybean oil, % 1.42 3.79 3.00
whether the REP was more effective than only reagent DL-Methionine, % 0.07 0.08 0.06
enzymes to prepare SSIF for in vitro digestion. L-lysine, % 0.04 0.01 0.07
Dicalcium phosphate, % 1.68 1.81 1.81
Limestone, % 1.20 1.23 1.08
Sodium chloride, % 0.30 0.30 0.30
MATERIALS AND METHODS Vitamin-mineral premix1 , % 1.00 1.00 1.00
Rice hull, % – – 8.75
Experimental procedures were approved by the ani- Total, % 100 100 100
mal care and welfare committee of the Institute of An- Nutrient content2
DM, % 89.11 88.73 89.53
imal Science, Chinese Academy of Agriculture Sciences CP, % 17.72 19.95 14.90
(Beijing, China). Ether extract, % 5.00 8.05 6.37
Ash, % 5.66 5.6 6.16
Crude fiber, % 4.42 4.24 7.35
ME/(kcal/kg) 2 2,949 3,050 2,801
Experimental Design
1
Supplied per kilogram of diet: vitamin A, 2,500 IU, vitamin D3, 400
Experiment 1 The objective of experiment 1 was IU, vitamin E, 10 IU, vitamin K3, 0.5 mg, thiamine 1.8 mg, riboflavin,
to compare the in vitro digestibility of SSIF to ESIF 4.0 mg, vitamin B6 3.0 mg, vitamin B12 7 μ g, pantothenic acid 11.0 mg,
nicotinic acid 55.0 mg, folic acid 0.5 mg, biotin 120 μ g; choline chloride
on corn, wheat, soybean meal, and cottonseed meal 750 mg, Cu (as copper sulfate) 8 mg, Fe (as ferrous sulfate) 80 mg, Mn
digested in a computer-controlled simulated digestion (as manganese sulfate) 60 mg, Zn (as zinc sulfate) 40 mg, I (as potassium
system (CCSDS) to simulate the process of small in- iodide) 0.35 mg, Se (as sodium selenite) 0.15 mg.
2
ME values were calculated according to the AME values of feedstuffs
testinal digestion. Forty-eight 15-wk-old white Peking for chickens (China feed database, Ministry of Agriculture, China, 2013)
ducks (3.25 ± 0.50 kg) were divided into two treat- and others were determined values on DM basis.
ments of 24 birds each. Ducks were surgically fitted
with T-shaped cannulas at Meckel’s diverticulum ac- trypsin, and chymotrypsin of ESIF in ducks (Zhao et
cording to the method described by Zhao et al. (2016) al., 2007). Ten ingredient samples including three grains
and placed in individual cages (0.45 m × 0.38 m × (corn, sorghum, and wheat), four plant protein meals
0.51 m) in a temperature-controlled room (25◦ C) un- (soybean meal, cottonseed meal, rapeseed meal, and
der 12 h of light per day. Each cannulated duck was peanut meal), and three grain byproducts (two samples
fed 20 ml of 10% (wt/vol) dextrose solution three times of distiller’s dried grains with solubles [DDGS; sam-
daily for 3 d after cannulation followed by ad libitum ples A and B] and rice bran) obtained from Guangdong
access to water and a basal diet (Table 1) designed to Haid Group Co., Ltd (Guangzhou, China; Table 3). The
satisfy NRC recommendations (1994). The ESIF 1 or 2 IVDE of 10 ingredients was determined in a CCSDS
was obtained from cannulated ducks randomly offered with pepsin following SSIF 1 (IVDE 1) or with pepsin
diet 1 (3,050 kcal/kg of ME and 19.95% of CP) or diet following SSIF 3 (IVDE 2). The IVDE of each sample
2 (2,801 kcal/kg of ME and 14.90% of CP) according to was determined for five replicates. The true metaboliz-
the method reported by Zhao et al. (2007). REP 1 or 2 able energy (TME) of each feed was determined for six
was purified from ESIF 1 or 2, respectively, according to replicates with four ducks in each. Good correlation be-
the method described by Zhang (2014). The specific ac- tween in vitro and in vivo determinations is necessary to
tivities of REP 1 and 2 were 28,020.9 vs. 25,634.3 U/g compare novel in vitro digestion techniques (Boisen and
for amylase, 5,382.1 vs. 4,865.7 U/g for trypsin, and Eggum, 1991). Thus, higher correlation between IVDE
2,708.0 vs. 2,281.7 U/g for chymotrypsin, respectively. and TME indicates SSIF is more comparable to ESIF.
The SSIF 1 to 3 or 4 to 6 were prepared to mimic ESIF The R2 and accuracy of the prediction model of TME
1 or 2, respectively. The main sources of enzymes were from IVDE 1 or 2 were used to validate the efficacy of
REP 1 (SSIF 1 and 4), REP 2 (SSIF 2 and 5), or reagent SSIF.
enzymes (SSIF 3 and 6). All SSIF matched the activities
of amylase, trypsin and chymotrypsin in ESIF by sup- IVDE Determination
plementation of reagent grade amylase (Sigma A3306,
Sigma-Aldrich Co., St. Louis, MO), trypsin (Amer- In Vitro Digestion of Small Intestine In vitro in-
sco 0785, Amersco Inc., Solon, OH), and chymotrypsin testinal digestion was conducted in a CCSDS. The
(Amersco 0164, Amersco Inc., Solon, OH; Table 2). The mix frequency, composition of small intestinal buffer
IVDE of each feed was determined with the in vitro di- solution, digestion time, digestion temperature, and
gestion of SSIF or ESIF in five replicates. wash procedure for digested product at the stage of
Experiment 2 Experiment 2 was to validate whether small intestinal digestion were described by Zhao et al.
REP 1 was more effective than only reagent enzymes (2014b). In brief, 2 g grains or 1 g non-grain sample
to prepare SSIF matching the activities of amylase, and 20 ml SSIF or ESIF were added into the dialysis
1282 ZHANG ET AL.
Table 2. Composition and enzyme activities of small intestinal fluid.

Intestinal fluid
Items ESIF1 1 SSIF2 1 SSIF 2 SSIF 3 ESIF 2 SSIF 4 SSIF 5 SSIF 6

Enzyme sources
ESIF from duck fed diet 1 +
ESIF from duck fed diet 2 +
REP3 purified from ESIF
REP 1 + +
REP 2 + +
Reagent grade enzyme4
Amylase + + + + + +
Trypsin + + + + + +
Chymotrypsin + + + + + +
Digestive enzyme activities, U/mL
Amylase 433.7 433.7 433.7 433.7 361.2 361.2 361.2 361.2
Trypsin 114.8 114.8 114.8 114.8 89.8 89.8 89.8 89.8
Chymotrypsin 39.8 39.8 39.8 39.8 28.8 28.8 28.8 28.8
1
ESIF = endogenous small intestinal fluid. The intestinal digesta was collected for 1 h every 4 h
beginning at 9:30 on alternate days from day 38 to 98 after intestinal cannulation. After the collection
of digesta from each time, ESIF was obtained by centrifuging digesta for 10 min at 1,250 g and 4◦ C
according to the procedure described by Zhao et al. (2007). In each diet treatment, ESIF collected from
24 ducks were pooled for purification of digestive enzymes and in vitro digestion.
2
SSIF = simulated small intestinal fluid.
3
REP = raw enzyme power.
4
Reagent grade enzymes include: amylase (Sigma A3306, Sigma-Aldrich Co., St. Louis, MO), trypsin
(Amersco 0785, Amersco Inc., Solon, OH), and chymotrypsin (Amersco 0164, Amersco Inc., Solon, OH).
“+” indicates the enzyme was added into the simulated small intestinal fluid.

Table 3. The nutrient contents of feed ingredients (on DM basis).

Category Sample CP, Crude Ether Ash, GE1 ,

1
% fiber, % extract, % % kcal/kg

Grain Corn 9.26 1.21 4.81 1.27 4,425

Sorghum 10.52 0.95 4.56 1.51 4,529
Wheat 14.52 1.23 2.98 1.41 4,416
Plant protein meal Soybean meal 49.29 6.60 3.45 6.60 4,700
Cottonseed meal 51.98 9.51 3.13 6.93 4,663
Rapeseed meal 42.28 10.36 3.90 7.09 4,683
Peanut meal 56.56 3.91 5.09 5.07 4,803
Grain byproduct Corn DDGS2 A 32.53 4.33 14.62 4.53 5,331
Corn DDGS B 30.08 4.38 12.28 4.98 5,214
Rice bran 11.76 15.2 15.46 9.51 4,879
1
GE = Gross energy.
2
DDGS = Distiller’s dried grains with solubles.

tubing of digestion chamber. The upper and lower small
intestinal buffers each continuously circulated for 7.5 h.
Undigested residue was dried and assessed for gross en-
ergy (GE) to calculate the IVDE as described below.
In Vitro Digestion of Gizzard-intestine In vitro di-
gestion of gizzard-intestine occurred in a CCSDS. The
pepsin activity in simulated gastric fluid, the activities
of amylase, trypsin, chymotrypsin in SSIF, and diges-
tion procedure were described by Zhao et al. (2014b).
After the simulated digestion, the undigested residues
were dried and analyzed for GE to calculate the IVDE
as described below.
Duck ME Assay
The TME values of 10 feed ingredients in experiment
2 were determined using 120 18-wk-old Peking drakes
of similar weight (3.8 to 4.0 kg) provided by the An-
imal Husbandry and Aquaculture Research Center of
Guangdong Haid Group (Guangzhou, China). Ducks
were randomly divided into five groups, each with six
replicates of four ducks raised in individual cages. In
the first ME trial, 1 of the 5 groups was used for the
determination of endogenous energy losses (EEL) and
each of the four remaining groups was used to deter-
mine the ME content of corn, sorghum, wheat, and
basal diet of corn starch. After the ME trial, there
was a 14-d rest period when ducks were provided with
ad libitum access to a commercial diet and allowed to
swim in a pool. After 14-d, the same 120 ducks were
randomly reassigned into five groups of 24 ducks (six
replicates of four ducks) to determine the EEL and
ME content of soybean meal, cottonseed meal, rape-
seed meal, and peanut meal. Following a second 14-d
rest period, 96 of the same 120 ducks were randomly
selected and assigned into four groups of 24 ducks
(six replicates of four ducks) to determine the EEL
 NOVEL SIMULATED DIGESTIVE FLUID TO PREDICT ME
and ME content of corn DDGS A, DDGS B, and rice
bran.
The TME bioassay method was in accordance with
Adeola et al. (1997) with minor modifications as fol-
lows: After 5 d acclimatization and feed withdrawal
for 36 h, ducks were force-fed with 60 g of sample to
determine AME. Excreta was collected for 36 h, dried
for 48 h at 65◦ C, and re-equilibrated with air for 24 h
before analysis. Corn, sorghum, and wheat were fed as
the only dietary ingredients, but soybean meal, cotton-
seed meal, rapeseed meal, peanut meal, corn DDGS,
and rice bran were tested as part of a complete diet
(60% corn starch + 40% test ingredient). Each ingre-
dient or mixture was ground through a 2-mm screen
before pelleting. Pellets measured 4 mm in diameter
and 6 mm long and were air-dried to 14% moisture be-
fore force-feeding. Force-feeding was conducted with a
stainless steel funnel with a narrow stem (40 cm long
and 1.0 cm inner diameter). Before determining EEL,
ducks were deprived of feed for 36 h. Data were used to
correct AME to TME.
Chemical Analysis
Before chemical analysis, samples were ground finely
in a laboratory mill fitted with a 0.3 mm mesh screen.
The dry matter (DM) content (method 934.01; AOAC,
1990) was determined by oven drying at 105◦ C for
5 h. Feed ingredients, excreta, and residue samples
were analyzed for GE by a Parr 6400 automatic adi-
abatic calorimeter (Parr Instrument Co., Moline, IL)
with benzoic acid as the calibration standard. Sam-
ples of ingredients and diets were analyzed for CP
(Kjeldahl N; method 954.01; AOAC, 1990), crude
fiber (method 962.09; AOAC, 1990), and ether extract
(method 920.39; AOAC, 1990).
Pepsin activity was determined as described by
Sturkie (1976) with hemoglobin as substrate. Amy-
lase activity was determined as described by Dahlqvist
(1962) with soluble starch as substrate. Trypsin activ-
ity was determined as described by Wirnt (1974a) us-
ing Nα-p- toluolsulfonyl-L- arginine methyl ester hy-
drochloride (T4626, Sigma-Aldrich CO., St. Louis, MO)
as a substrate. Activity of chymotrypsin was deter-
mined as described by Wirnt (1974b) using N-benzoyl-
l-tyrosine ethyl ester (B6125, Sigma-Aldrich CO., St.
Louis, MO) as the substrate. Total protein of ESIF was
measured using a protein quantification kit (A045-2,
Nanjing Jiancheng Bioegineering Institute, Nanjing,
China).
Calculation and Statistical Analysis
The TME of corn, sorghum, wheat, and diet were
calculated as follows: TME (kcal/kg) = (energy intake
– energy output + EEL)/feed intake. The TME of rice
bran, soybean meal, cottonseed meal, rapeseed meal,
peanut meal, and DDGS were calculated according to
1283
the different approach as follows: TME = (TMEt –
TMEb × b)/(1 – b). Where: TMEt were the TME of the
complete diet (basal diet plus the ingredient); TMEb
were the TME of the basal diet; b was the proportion
of the basal diet in the complete diet. The following
formula: IVDE = [(sample DM weight × sample GE)
− (defatted residue DM weight × defatted residue GE)
+ GE of dry residue of digestive enzymes]/sample DM
weight was used to calculate IVDE.
The TTEST procedure (SAS Institute, 1990) as-
sessed whether the activities of digestive enzymes in
ESIF were affected by diet. The general linear model
procedure (SAS Institute, 1990) assessed whether IVDE
differed when determined with SSIF compared to ESIF.
Tukey’s multiple comparison test was conducted at a
significance level of α = 0.05. The REG procedure of
SAS (SAS Institute, 1990) developed linear regression
models to predict the TME from IVDE. The coefficient
of determination (R2 ) and residual SD indicated quality
of the prediction models (Kaps and Lamberson, 2004).
RESULTS AND DISCUSSION
IVDE of Feeds Determined with ESIF or SSIF
In experiment 1, the activities of amylase, trypsin,
and chymotrypsin as well as total protein concentration
were greater in the ESIF of ducks fed diet 1 than diet
2 (Table 4; P < 0.05). The greater dietary ME and
CP obviously induced activities of digestive enzymes
in ESIF of ducks. This phenomenon agreed with the
results of our previous studies (Zhao et al., 2007; Ren
et al., 2012; Zhao et al., 2012). Therefore, two types
of ESIF (1 or 2) with different activities of digestive
enzymes can be collected from intestinally cannulated
ducks fed diets 1 or 2.
The small intestine is the main position for the diges-
tion and absorption of dietary nutrients. Historically,
researchers assumed the ability of SSIF and ESIF to
digest feeds in vitro was similar. However, few have
compared simulated and endogenous digestive fluid in
previous experiments (Boisen and Fernández, 1997;
Wilfart et al., 2008; Swiech, 2017). The IVDE of ingre-
dients digested with SSIF or ESIF are shown in Table
5. The IVDE was similar for each of wheat and soy-
bean meal determined with SSIF 1 or ESIF 1, but the
IVDE for corn or cottonseed meal was less for SSIF 1
compared to ESIF 1 (P < 0.05). The IVDE determined
with SSIF 1 was 97.81 to 100.34% that determined with
ESIF 1. The IVDE for corn, wheat, and cottonseed meal
was less for SSIF 2 compared to ESIF 1, whereas the
opposite was true for soybean meal (P < 0.05). The
IVDE determined with SSIF 2 were 95.84% to 101.68%
that determined with ESIF 1. The IVDE of each feed
ingredient was less for determinations with SSIF 3 com-
pared to SSIF 1, 2, or ESIF 1 (P < 0.05). The IVDE
determined with SSIF 3 was 82.41% to 98.14% that de-
termined with ESIF 1. In the second group, the IVDE
was similar for ingredients determined with SSIF 4 or
1284 ZHANG ET AL.

Table 4. Digestive enzyme activity of ESIF1 from ducks fed diets 1 or 2.

Items ESIF 12 ESIF 23 P value

Digestive enzyme activity, U/ml

Amylase 463.1 ± 40.4 395.1 ± 45.5 0.0210
Trypsin 115.4 ± 8.7 94.6 ± 5.3 0.0005
Chymotrypsin 47.4 ± 4.1 31.4 ± 3.0 < .0001
Total protein concentration, mg/mL 3.5 ± 0.3 3.0 ± 0.2 0.0067
1
ESIF = endogenous small intestinal fluid. The intestinal digesta was collected for 1 h every 4 h
beginning at 9:30 on day 33, 35, and 37 after intestinal cannulation. After the collection of digesta
from each time, ESIF was made by centrifuging digesta for 10 min at 1,250 g and 4◦ C according to the
procedure described by Zhao et al. (2007). In each replicate, an equal volume of ESIF from four ducks
in each collection time was pooled to determine enzyme activity.
2
ESIF 1 was the endogenous small intestinal fluid from jejunal cannulated ducks with six replicates
of four ducks fed diet 1.
3
ESIF 2 was the endogenous small intestinal fluid from jejunal cannulated ducks with six replicates
of four ducks fed diet 2.

Table 5. The IVDE1 of feed ingredients determined with ESIF2 and SSIF.3

Item Corn Wheat Soybean meal Cottonseed meal

IVDE determined with small intestinal fluid, kcal/kg

ESIF 1 4,092 ± 15a 3,918 ± 13a 3,564 ± 24b 2,692 ± 11a
SSIF 1 4,030 ± 13c 3,929 ± 10a 3,576 ± 13b 2,633 ± 19b
SSIF 2 4,059 ± 14b 3,886 ± 19b 3,624 ± 10a 2,580 ± 23c
SSIF 3 3,926 ± 5d 3,845 ± 14c 2,937 ± 9c 2,535 ± 6d
P value < 0.0001 < 0.0001 < 0.0001 < 0.0001
Ratio of IVDE determined with SSIF to value determined with ESIF 1, %
SSIF 1/ESIF 1 98.48 100.28 100.34 97.81
SSIF 2/ESIF 1 99.19 99.18 101.68 95.84
SSIF 3/ESIF 1 95.94 98.14 82.41 94.17
IVDE2 determined with small intestinal fluid, kcal/kg
ESIF 2 4,076 ± 7a 3,881 ± 19a 3,555 ± 19a 2,601 ± 31a
SSIF 4 4,009 ± 8b 3,880 ± 8a 3,532 ± 16a 2,538 ± 27b
SSIF 5 4,009 ± 11b 3,896 ± 11a 3,541 ± 12a 2,540 ± 14b
SSIF 6 3,889 ± 12c 3,836 ± 12b 2,873 ± 20b 2,468 ± 7c
P-value < 0.0001 < 0.0001 < 0.0001 < 0.0001
Ratio of IVDE determined with SSIF to value determined with ESIF 2, %
SSIF 5/ESIF 2 98.36 100.39 99.61 97.65
SSIF 4/ESIF 2 98.36 99.97 99.35 97.58
SSIF 6/ESIF 2 95.41 98.84 80.82 94.89
1
IVDE = in vitro digestible energy (DM basis). The values represent mean of five replicates per sample and are expressed as mean ± SD.
2
ESIF = endogenous small intestinal fluid. The intestinal digesta was collected for 1 h every 4 h beginning at 9:30 on alternate days from day 38
to 98 after intestinal cannulation. After the collection of digesta from each time, ESIF was made by centrifuging digesta for 10 min at 1,250 g and
◦
4 C according to the procedure described by Zhao et al. (2007). In each diet treatment, the ESIF collected from 24 ducks was pooled for in vitro
digestion.
SSIF = simulated small intestinal fluid.

5 for each of four feed ingredients. The IVDE of corn
and cottonseed meal was less for SSIF 4 or 5 compared
to ESIF 2 (P < 0.05), but there were no significant
differences for wheat and soybean meal. The IVDE de-
termined with SSIF 4 or 5 was 97.58% to 100.39% that
determined with ESIF 2. The IVDE in each of feed in-
gredients was least when determined with SSIF 6 com-
pared to SSIF 4, 5, or ESIF 2. The IVDE determined
with SSIF 6 was 80.82% to 98.84% that determined
with ESIF 2. In the small intestine of poultry, more than
10 types of digestive enzyme contribute to digestion
(Whittow, 2000). Previous studies focused on the ac-
tivities of amylase, trypsin, chymotrypsin, and lipase in
digesta or intestinal fluid (Furuya et al., 1979; Sakamoto
et al., 1980; Fan, 2003; Zhao et al., 2007; Ren et al.,
2012) because these enzymes influence extent of diges-
tion of dietary nutrients. The current study focused
on activities of amylase, trypsin, and chymotrypsin be-
cause they have a greater influence on the GE digestibil-
ity in each of four feeds compared to other enzymes in
ESIF. Previous studies reported that amylase, trypsin,
and chymotrypsin can account for more than 86% of the
activity of ESIF (Xie, 2011; Yan et al., 2012). The dif-
ferences of digestible ability between SSIF 1 or 2 and
ESIF 1, and between SSIF 4 or 5 and ESIF 2 indi-
cate SSIF made of REP 1 (supplemented with a small
quantity of reagent enzymes) was closest to the ESIF
relative to REP 2 (supplemented with relatively more
reagent enzymes). In general, the digestibility of en-
zymes is dependent on their enzymatic properties. Pre-
vious studies have documented differences in pH activ-
ity profile and hydrolysis kinetics of enzymes from pigs
and chickens (Crevieu-Gabriel et al., 1999). This study
utilized REP purified from the ESIF of ducks. There-
fore, the digestive enzymes of REP had the same enzy-
matic properties as that of in vivo digestive enzymes.
Table 6. The determined TME1 and predicted values from IVDE2 of 10 samples (on DM basis).

Ratio of IVDE
IVDE, kcal/kg TME, kcal/kg to TME, % Difference
Category Ingredient IVDE 13 IVDE 24 Determined Predicted 15 Predicted 26 IVDE 1 IVDE 2 D17 D28

Grain Corn 3,996 ± 18 3,836 ± 4 4,006 ± 29 3,984 3,928 99.75 95.76 22 78

Sorghum 3,919 ± 13 3,849 ± 6 3,893 ± 18 3,921 3,940 100.67 98.87 –28 –47
Wheat 3,839 ± 14 3,693 ± 8 3,907 ± 31 3,855 3,798 98.26 94.52 52 109
Plant protein meal Soybean meal 3,556 ± 19 3,430 ± 11 3,625 ± 31 3,623 3,558 98.10 94.62 2 67
Cottonseed meal 2,568 ± 17 2,572 ± 19 2,777 ± 19 2,813 2,774 92.47 92.62 –36 3
Rapeseed meal 2,795 ± 19 2,838 ± 15 3,034 ± 24 2,999 3,017 92.12 93.54 35 17
Peanut meal 4,015 ± 16 4,037 ± 20 4,061 ± 43 3,999 4,112 98.87 99.41 62 –51
Grain byproduct DDGS9 A 3,777 ± 9 3,660 ± 18 3,694 ± 35 3,804 3,768 102.25 99.08 –110 –74
DDGS B 3,459 ± 16 3,475 ± 20 3,521 ± 28 3,543 3,599 98.24 98.69 –22 –78
Rice bran 2,719 ± 28 2,829 ± 5 2,972 ± 12 2,937 3,009 91.49 95.19 35 –37
Statistics
Mean 3,464 3,422 3,549 3,548 3,550 97.22 96.23 0 0
Minimum 2,568 2,572 2,777 2,813 2,774 91.49 92.62 –110 –78
Maxmium 4,015 4,037 4,061 3,999 4,112 102.25 99.41 62 109
TME equation from IVDE
Slope 0.820 0.913
Intercept 707 426
R2 0.99 0.98
Residual SD, kcal/kg 55 71
P-value < 0.001 < 0.001
1
TME values were determined for six replicates with four ducks in each and expressed as mean ± SD.
2
IVDE = in vitro digestible energy. The values represent mean of 5 replicates per sample and are expressed as mean ± SD.
3
IVDE 1 was determined with pepsin following simulated small intestinal fluid (prepared from REP 1 supplemented with small amounts of reagent enzymes) matched the activities of digestive
enzymes in duck fed normal diets (as described by Zhao et al. (2007)) in a computer-controlled simulated digestion system. REP1 was purified from endogenous small intestinal fluid (ESIF 1) of
ducks fed diet 1 (3,050 kcal/kg of ME and 19.95% of CP).
4
IVDE 2 was determined with pepsin following simulated small intestinal fluid (prepared from reagent enzymes) matched the activities of digestive enzymes in duck fed normal diets (as
NOVEL SIMULATED DIGESTIVE FLUID TO PREDICT ME

described by Zhao et al. (2007) in a computer-controlled simulated digestion system.

5
Values are calculated based on TME = 0.820 × IVDE 1 + 707 (R2 = 0.99, P < 0.001).
6
Values are calculated based on TME = 0.913 × IVDE 2 + 426 (R2 = 0.98, P < 0.001).
7
Values are calculated as determined TME- predicted value from IVDE 1.
8
Values are calculated as determined TME- predicted value from IVDE 2.
9
DDGS = Distiller’s dried grains with solubles.
1285
1286 ZHANG ET AL.
However, the reagent grade amylase (Sigma A3306,
Sigma-Aldrich Co., St. Louis, MO) was from Bacil-
lus Licheniformis and reagent grade trypsin (Amer-
sco 0785, Amersco Inc., Solon, OH) and chymotrypsin
(Amersco 0164, Amersco Inc., Solon, OH) were puri-
fied from swine. Consequently, enzymatic properties of
reagent grade enzymes may differ from in vivo digestive
enzymes of ducks. Accordingly, SSIF 3 or 6 (made of
only reagent grade amylase, trypsin, and chymotrypsin)
greatly underestimated the IVDE compared to the
SSIF 1, 2, 4, and 5. Higher activities of digestive en-
zymes contained in REP led to less supplementation of
reagent enzyme. Greater activities of amylase, trypsin,
and chymotrypsin were present in REP 1 than REP
2. Therefore, the SSIF made of REP 1 (supplemented
with less reagent grade enzymes) more closely mimicked
the enzymatic properties of ESIF than SSIF made of
REP 2.
TME predicted from IVDE determined with
Pepsin-SSIF1 or Pepsin-SSIF3
A good correlation between in vitro and in vivo re-
sults is necessary to validate the efficacy of in vitro
digestion with pepsin-SSIF. In the experiment 2, the
mean IVDE 1, 2 and TME of 10 ingredients was 3,464,
3,422, and 3,549 kcal/kg of DM, with ranges from 2,568
to 4,015, 2,572 to 4,037 and 2,777 to 4,061 kcal/kg
of DM (Table 6), respectively. The mean IVDE was
97.22% (IVDE 1) or 96.23% (IVDE 2) of TME, which
indicates the present two-step in vitro digestion with
pepsin following SSIF 1 or 3 simulates the majority of
in vivo total tract digestion in major feed ingredients for
duck. We previously reported that IVDE 2 was 93.12%
of the TME for ducks across 30 samples of corn (range =
90.45% to 95.76%; Zhao et al., 2014b). This result sug-
gests that the activities of pepsin, amylase, trypsin, and
chymotrypsin are the key variables to modulate in vitro
digestion performed in the CCSDS. This interpretation
is consistent with our previous studies for ducks (Zhao
et al., 2014b) and roosters (Zhao et al., 2014a). In the
present experiment, the IVDE 1 of the three grains and
one soybean meal was greater than IVDE 2 and close to
TME. Thus, compared to the IVDE 2, the determina-
tions with IVDE 1 were more similar to TME (∼100%),
indicating the efficacy of SSIF 1 is better than SSIF
3 for grain and soybean meal, which further validates
the result of experiment 1. The IVDE 1 and IVDE 2
were similar for cottonseed meal, rapeseed meal, peanut
meal, and DDGS A and B, indicating the SSIF 1 and 3
had similar efficacy in these ingredients. However, the
IVDE 1 was less than IVDE 2 for rice bran. The enzy-
matic properties probably differ between microorgan-
ism and endogenous duck sources of amylase. Cellulase
may accompany the microorganism source of amylase
in SSIF 3, which makes an obvious contribution to in
vitro digestibility of ingredients high in fiber content.
Xie (2011) reported a similar phenomenon with wheat
bran for ducks.
Linear regression analysis revealed that strong corre-
lations with TME for IVDE 1 (P < 0.01; R2 = 0.99) and
IVDE 2 (P < 0.01; R2 = 0.98), respectively. However,
residual SD was lower for IVDE 1 relative to IVDE 2
(SD = 55 or 71 kcal/kg, respectively), indicating su-
perior accuracy of predicting TME from IVDE 1 com-
pared to IVDE 2. Similarly, the IVDE 1 was more accu-
rate than IVDE 2 to predict TME in three grains and
one soybean meal sample. Diets for ducks in China con-
tain more than 70% grains and soybean meal, so using
IVDE 1 may more accurately predict the ME of diets.
These results further imply the SSIF 1 is better than
SSIF 3 for in vitro digestion.
In conclusion, SSIF prepared with REP and supple-
mented with small amount of regent amylase, trypsin,
and chymotrypsin can reasonably assess the TME in
consideration of comparable digestibility and enzymatic
property with these of ESIF. The IVDE determined
with in vitro digestion of pepsin following the novel
SSIF may more accurately predict TME of feed ingredi-
ent for ducks compared to mixtures of reagent enzymes.
ACKNOWLEDGMENTS
This project was financially supported by the Young
Scientist Plan, the Innovation Team of the Chinese
Academy of Agricultural Sciences (Beijing, China) and
the scientific fund of Guangdong Haid Group Co., Ltd
(Guangzhou, China).
REFERENCES
Adeola, O., D. Ragland, and D. King. 1997. Feeding and excreta
collection techniques in metabolizable energy assays for ducks.
Poult. Sci. 76:728–732.
Association of Official Analytical Chemists (AOAC). 1990. Official
Methods of Analysis, 15th ed. Assoc. Off. Anal. Chem., Wash-
ington, DC.
Boisen, S., and B. O. Eggum. 1991. Critical evaluation of in-vitro
methods for estimating digestibility in simple-stomach animals.
NRR 4:141–162.
Boisen, S., and J. A. Fernandez. 1997. Prediction of the total tract
digestibility of energy in feedstuffs and pig diets by in vitro anal-
yses. Anim. Feed Sci. Technol. 68:277–286.
Bryan, D. D. S. L., D. A. Abbott, and H. L. Classen.
2018. Development of an in vitro protein digestibility as-
say mimicking the chicken digestive tract. Anim. Nutr. doi:
https://doi.org/10.1016/j.aninu.2018.04.007.
Clunies, M., and S. Leeson. 1984. In vitro estimation of dry matter
and crude protein digestibility. Poult. Sci. 63:89–96.
Clunies, M., S. Leeson, and J. D. Summers. 1984. In vitro es-
timation of apparent metabolizable energy. Poult. Sci. 63:
1033–1039.
Crevieu-Gabriel, I., J. Gomez, J. P. Caffin, and B. Carre. 1999. Com-
parison of pig and chicken pepsins for protein hydrolysis. Reprod.
Nutr. Dev. 39:443–454.
Dahlqvist, A. 1962. A method for the determination of amylase in
intestinal content. Scand. J. Clin. Lab. Invest. 14:145–151.
Fan, H. P. 2003. Comparative study of the digestion of feed nutri-
ents between cockerel and drake. MS thesis. Chinese Academy of
Agricultural Sciences, Beijing.
 NOVEL SIMULATED DIGESTIVE FLUID TO PREDICT ME
Furuya, S., K. Sakamoto, and S. Takahashi. 1979. A new in vitro
method for the estimation of digestibility using the intestinal fluid
of the pig. Br. J. Nutr. 41:511–520.
INRA-AFZ-INAPG. 2004. Tables of Composition and Nutritional
Value of Feed Materials. D. Sauvant, J.-M. Perez, and G. Tran,
ed. Wageningen Acad. Publ., Wageningen, the Netherlands.
Kaps, M., and W. Lamberson. 2004. Biostatistics for animal science.
CABI Publishing, Wallingford, UK.
Losada, B., P. Garcı́a Rebollar, C. Álvarez, P. Cachaldora, M. A.
Ibánez, J. Méndez, and J. C. De Blas. 2010. The prediction of
apparent metabolisable energy content of oil seeds and oil seed
by-products for poultry from its chemical components, in vitro
analysis or near-infrared reflectance spectroscopy. Anim. Feed Sci.
Technol. 160:62–72.
Losada, B., P. Garcı́a Rebollar, P. Cachaldora, C. Álvarez, J.
Méndez, and J. C. De Blas. 2009. A comparison of the predic-
tion of apparent metabolisable energy content of starchy grains
and cereal by-products for poultry from its chemical components,
in vitro analysis or near-infrared reflectance spectroscopy. Span.
J. Agric. Res. 7:813–823.
National Research Council. 1994. Nutrient Requirements of Poultry.
9th rev. ed. Natl. Acad. Press, Washington, DC.
Regmi, P. R., W. C. Sauer, and R. T. Zijlstra. 2008. Prediction
of in vivo apparent total tract energy digestibility of barley in
grower pigs using an in vitro digestibility technique1. J. Anim.
Sci. 86:2619–2626.
Ren, L. Q., F. Zhao, H. Z. Tan, J. T. Zhao, J. Z. Zhang, and H. F.
Zhang. 2012. Effects of dietary protein source on the digestive en-
zyme activities and electrolyte composition in the small intestinal
fluid of chickens. Poult. Sci. 91:1641–1646.
Sakamoto, K., T. Asano, A. Furuya, and S. Takahashi. 1980. Esti-
mation of in vivo digestibility with the laying hen by an in vitro
method using the intestinal fluid of the pig. Br. J. Nutr. 43:389–
391.
SAS Institute. 1990. SAS/STAT User’s Guide: Statistics. SAS Inst.
Inc., Cary, NC.
Sturkie, P. D. 1976. Avian Physiology. 3rd ed. Springer-Verlag, New
York, NY.
Swiech, E. 2017. Alternative prediction methods of protein and en-
ergy evaluation of pig feeds. J Anim. Sci. Biotechnol. 8:39.
Valdes, E. V., and S. Leeson. 1992. Measurement of metabolizable
energy in poultry feeds by an in vitro system. Poult. Sci. 71:1493–
1503.
Whittow, G. C. 2000. Sturkie’s Avian Physiology. 5th ed. Acad.
Press, London, UK.
1287
Wilfart, A., Y. Jaguelin-Peyraud, H. Simmins, J. Noblet, J. van
Milgen, and L. Montagne. 2008. Kinetics of enzymatic digestion
of feeds as estimated by a stepwise in vitro method. Anim. Feed
Sci. Technol. 141:171–183.
Wirnt, R. 1974a. Chymotrypsin, measurements with n-benzoyll-
tyrosin ethyl ester as substrate. Pages 1009–1012 in Methods of
Enzymatic Analysis. 2nd ed. H. Ulrich Bergmeyer, ed. Verlag
Chemie, Weinheim, Germany.
Wirnt, R. 1974b. Trypsin, measurement with nα-p-toluenesulfonyl-
l-arginine methyl ester as substrate. Pages 1021–1024 in Methods
of Enzymatic Analysis. 3rd ed. H. Ulrich Bergmeyer ed. Verlag
Chemie, Weinheim, Germany.
Xie, M. L. 2011. Study on the preparation methods of simulat-
ing duck small intestine fluid. MS thesis. Yangzhou University,
Yangzhou.
Yan, F., F. Zhao, B. M. Mi, L. Zhang, Z. L. Qi, and H. F. Zhang.
2012. Effects of batch sources and activity contribution of di-
gestive enzymes in simulated intestinal fluid for ducks on diges-
tive ability of feed ingredients. Chin. J. Anim. Nutr. 11:2181–
2189.
Zhang, L. 2014. Study on the method for preparation of simulated
intestinal fluid based on enzyme powder purified from endogenous
digestive fluid in duck. MS thesis. Chinese Academy of Agricul-
tural Sciences, Beijing.
Zhao, F., S. S. Hou, H. F. Zhang, and Z. Y. Zhang. 2007. Effects
of dietary metabolizable energy and crude protein content on the
activities of digestive enzymes in jejunal fluid of Peking ducks.
Poult. Sci. 86:1690–1695.
Zhao, F., W. K. Zheng, G. Y. Hu, H. Li, and H. F.
Zhang. 2012. Study on the difference of intestinal fluid com-
position in duck fed a wheat-soybean meal and a corn-
soybean meal diets. Chin. J. Anim. Vet. Sci. 43:1401–
1407.
Zhao, F., L. Q. Ren, B. M. Mi, H. Z. Tan, J. T. Zhao, H. Li, H. F.
Zhang, and Z. Y. Zhang. 2014a. Developing a computer-controlled
simulated digestion system to predict the concentration of metab-
olizable energy of feedstuffs for rooster1. J. Anim. Sci. 92:1537–
1547.
Zhao, F., L. Zhang, B. M. Mi, H. F. Zhang, S. S. Hou, and Z. Y.
Zhang. 2014b. Using a computer-controlled simulated digestion
system to predict the energetic value of corn for ducks1. Poult.
Sci. 93:1410–1420.
Zhao, F., M. L. Xie, H. Li, and H. F. Zhang. 2016. Developing a
silicon t-shaped cannula to sample intestinal digesta in ducks.
Anim. Feed Sci. Technol. 219:175–180.