Animal, page 1 of 7 © The Animal Consortium 2019 animal

doi:10.1017/S1751731119001587

Modeling nutritional and performance factors that influence

the efficiency of weight gain in relation to excreted nitrogen
in weaning piglets
K. M. Cardinal1† , M. S. Vieira1, M. B. Warpechowski2, P. K. Zielgelmann3, L. Montagne4,
I. Andretta1 and A. M. L. Ribeiro1
1
Department of Animal Science, Federal University of Rio Grande do Sul, Porto Alegre, RS 91540-000, Brazil; 2Department of Animal Science, Federal University of Paraná,
Curitiba, PR 80035-050, Brazil; 3Department of Statistic, Federal University of Rio Grande do Sul, Porto Alegre, RS 91509-000, Brazil; 4INRA/AgrocampusOuest, UMR
1079, Systèmes d’Elevage, Nutrition Et Humaine, F. 35000 Rennes, França

(Received 6 August 2018; Accepted 18 June 2019)

One of the most debated topics in pig production is the need to study, understand and change the production system in order to
improve nutrient efficiency, becoming more environmentally friendly. The nitrogen excretion has highly deleterious effects on the
environment, and it is necessary to develop tools that help to reduce the excretion of this compound without compromising
productivity. Therefore, two models were generated to estimate the efficiency of weight gain in relation to excreted nitrogen in
post-weaning piglets. Data for testing these models were obtained from previous master and PhD studies carried out at the
Federal University of Rio Grande do Sul, Animal Science Laboratory using piglets in the post-weaning phase with results for
performance and digestibility. The database that was constructed was composed of raw data from 10 studies carried out
between 2000 and 2016, on a total of 726 piglets weaned at ages between 17 and 28 days, and to which 62 different
treatments were applied. An exploratory analysis of the data was done by evaluating scatter plots and histograms, and variables
representing different treatments were used in a stepwise multiple linear regression analysis, with the F-test used as the selection
criterion. Two models were generated that either considered the nitrogen retained or not, to estimate the ratio between weight
gain and excreted nitrogen using generalized linear model procedure. The authors analyzed the behavior of each variable to
evaluate whether the equation generated was biologically coherent. Weight gain, dry matter intake, nitrogen intake,
metabolizable energy intake, retained nitrogen and urinary nitrogen were all significant ( P<0.001) variables in model I, and in
model II the variable fecal nitrogen was also included. The models had high coefficients of determination ( R2 of model I and II
were 0.9013 and 0.8271, respectively), and the nitrogen ingested variable was the one that most strongly influenced growth
efficiency. When the retained nitrogen variable was removed from the model, there was a reduction in the fit of the equations. It
was possible to conclude that both of the two models generated could be applied and the amount of nitrogen ingested had the
greatest influence on growth efficiency related to nitrogen excretion.

Keywords: meta-analysis, nutrition, protein, pollutant, swine

Implications
The use of easy-to-apply equations to predict the efficiency of
weight gain to excreted nitrogen in weaning piglets has great
relevance within the context of modern, sustainable pig farms.
The equations generated in this study will serve as a supple-
mentary tool for digestibility assays in research institutions. It
will indirectly benefit pig producers and the environment since
†
Present address: Faculdade de Agronomia, Avenida Bento Gonçalves, 7712 –
Agronomia, RS 91540-000, Brazil. E-mail: katia.zootecnia@hotmail.com
the modeling carried out in this study will help to reduce nitro-
gen excretion associated with maintenance of weight gain.
Introduction
Maximizing the growth performance of pigs has traditionally
been the goal of pig producers and nutritionists. Diets are
generally formulated to achieve this goal, and the diets
using excessive nutrients result in a surplus of nutrients
excreted in feces and urine (NRC, 2012; Pomar et al.,

1
Cardinal, Vieira, Warpechowski, Zielgelmann, Montagne, Andretta and Ribeiro
2017). Consequently, there is an excessive environmental
impact caused by the imbalance between the supply of
nutrients and the nutritional need of the pigs (Monteiro
et al., 2016; Andretta et al., 2017).
One of the main nutrients that have negative impacts on
the environment is nitrogen (N). During the storage,
treatment, and release of waste in the soil, nitrogen can form
polluting compounds, such as nitrate, ammonia and nitrous
oxide (McAuliffe et al., 2016). Laws and regulations to limit
the production and use of manure as fertilizer have thus
become increasingly restrictive, in an attempt to mitigate
these environmental impacts of N.
In post-weaning piglets in particular, low feed intake
associated with high potential protein deposition results in
the use of diets with high CP levels. There have been numer-
ous studies on topics that may result in reduced nutrient
excretion, such as on dietary protein reduction using
synthetic amino acid supplementation (Berrocoso et al.,
2013; Gloaguen et al., 2013 and 2014), use of exogenous
enzymes (Guggenbuhl et al., 2012; Metzler-Zebeli and
Zebeli, 2013; Yáñez et al., 2013) and improvements in man-
agement strategies (Hutchings et al., 2013; Montes et al.,
2013; Garcia-Launay et al., 2014). However, high safety mar-
gins are still widely used in conventional feeding programs.
The genetic advances obtained in pig farms were aimed at
obtaining animals with higher lean tissue gain and reduced
maintenance requirements (Boddicker et al., 2011), lower
heat production (Barea et al., 2010) and better use of
nutrients. Since growth rate and feed intake are related to
the productive efficiency, more efficient animals will conse-
quently be less polluting. The analysis of six genetic lines of
pigs showed that from 1969 to 2004 there was a 19%
improvement in N retention, resulting in 20% less excretion
(Knap and Bishop, 2000). Similarly, Shirali et al. (2012) and
Morel and Wood (2005) showed that N excretion was lower
in lean pigs or in pigs with a greater tendency for meat
deposition.
Meta-analysis is a useful tool for grouping data and
quantifying results from studies where the focus of the
research on a particular topic was only secondary, or even
nonexistent (Sauvant et al., 2008). Therefore, it is possible
to make new inferences on a distinct topic based on a data-
base of information from previous studies. Thus, the aim of
the present study was to generate predictive equations to
estimate the efficiency of weight gain v. excreted nitrogen
in post-weaning piglets and investigate the influence of
different variables on this relation.
Material and methods
A database was constructed in an electronic spreadsheet
using data from previous studies selected on the basis of
the following criteria. To be included, studies must have:
(i) been carried out at the Federal University of Rio Grande
do Sul (UFRGS) Animal Science Laboratory, basis of
published theses and dissertations; (ii) used piglets in
2
the post-weaning phase (up to 30 days after weaning);
(iii) presented performance results; and (iv) presented
digestibility results. The first criterion was established to
allow a database of raw data collected under comparable
conditions. After the selection step, detailed information
about the studies was included in the database, including
bibliographic aspects (trial and year), experimental char-
acteristics (ingredients and nutritional composition of
diets, sex, age of animals and duration of the trial) and
results (body weight, weight gain, feed intake, feed con-
version, energy, fat, protein and dry matter digestibility
coefficients, CP, lysine, energy, calcium and phosphorus
intake, urinary nitrogen, fecal nitrogen (FN) and retained
nitrogen (RN)). In each column a descriptive characteristic
or result was inserted, and in each line the characteristics
and results of an experimental unit (cage with one pig in
seven studies or mean of two pigs in three studies) were
described, totaling 144 columns and 491 rows. Some var-
iables were calculated from tabulated data to increase the
total number of variables, such as lysine intake, feed con-
version and the relationship between weight gain and
excreted nitrogen, which was considered as the main
response variable in this study. The methodologies
described in the literature were used for the elaboration
of the database and for the coding of the data (Sauvant
et al., 2005; Lovatto et al., 2007). Codes were used as
moderating variables in the analysis to account for vari-
ability among the compiled trials (e.g., effect of study).
The database was built from the responses observed in 10
studies carried out between 2000 and 2016. The database
was composed of data for a total of 726 animals (218 females
and 508 barrows), and the sum of the treatments of the indi-
vidual studies results in 62. All piglets came from commercial
genetic lines and were weaned between 17 and 28 days of
age. The piglets’ average ± SD body weight at the beginning
of the pre-starter phases (1 to 14 days post-weaning) was
5.05 ± 2.90 kg and 13.38 ± 4.50 kg at the end of the starter
phase (15 to 28 days post-weaning). They were individually
housed in metabolic cages and kept in thermal comfort
during the experimental periods.
The main ingredients used in the diets, their average com-
positions and other descriptive statistics related to the feeds
used are presented in Table 1. The diets were formulated with
nutritional levels recommended by Rostagno et al. (2011) or
NRC (2012) and provided ad libitum. Corn and soybean meal
were the main sources of energy and protein used in 60%
and 98% of the diets, respectively.
An exploratory analysis of the data was done by evaluat-
ing scatter plots and histograms. Correlation analyses were
performed to pre-select the covariate to be used (Table 2;
r<0.700), and to avoid those that would result in multicolli-
nearity within the same model. The pre-selected variables
were then used in a stepwise multiple linear regression analy-
sis, with the F-test used as the selection criterion. Significant
variables were selected to remain in the regression equa-
tions, while non-significant variables were sequentially
removed; inflation variance factors were also recorded. A
 Modeling weight gain/excreted nitrogen in piglets

Table 1 Descriptive statistics of the main ingredients and average Table 2 Pre-selected* variables used in a stepwise multiple linear
composition of the piglets’ diets regression analysis to compose models I and II to predict the weight
gain relative to excreted nitrogen (BWG/EN) of piglets
n Mean SD Min Max
All variables Pre-selected
Ingredient (%)
Corn 337 35.14 16.69 0.00 53.30 Soybean meal
Boiled corn 212 38.10 21.94 0.00 62.61 Corn
Soybean meal 481 17.02 9.01 0.00 35.94 Crude protein
Desativated soy 53 8.21 11.61 0.00 30.00 Lysine
Micronizated soy 116 11.55 11.29 0.00 30.00 Methionine
Isolated soy protein 93 2.17 1.95 0.00 5.00 Tryptophan
Whole milk powder 155 6.42 4.94 0.00 15.87 Threonine
Lactose 92 7.27 2.52 0.00 10.00 Metabolizable energy
Glycerol 31 7.58 5.14 0.00 15.00 Metabolizable energy intake X
Corn gluten 399 3.81 1.66 2.00 7.00 Crude protein: Metabolizable energy
Sugar 491 2.99 0.05 2.55 3.00 Dry matter intake X
Swine blood plasma 303 3.19 1.33 1.00 5.00 Nitrogen intake X
Milk serum 336 12.97 4.99 3.63 20.00 Lysine intake X
Nuklospray E501 63 25.08 5.04 20.00 30.00 Methionine intake
Yeast extract 30 2.53 1.81 0.00 5.00 Inicial live weight X
Beta-glucano2 30 0.02 0.01 0.00 0.04 Final live weight
Promosoy 903 61 5.31 1.97 3.32 7.24 Weight gain X
Raw wheat 58 7.89 17.44 0.00 45.76 Feed consumption X
Boiled wheat 58 7.89 17.44 0.00 45.76 Feed conversion ratio
Raw rice 58 7.87 17.38 0.00 45.62 Fecal nitrogen X
Boiled rice 58 7.87 17.38 0.00 45.62 Urinary nitrogen X
Skimmed milk powder 58 12.00 0.00 12.00 12.00 Retained nitrogen Fixed1
Barley 39 10.00 0.00 10.00 10.00
*The variables were pre-selected after correlation analysis (r<0.700).
Oil 246 2.93 1.11 0.73 5.50 1
Fixed: Variable was fixed by the authors to compose the model I and not
Integral rice meal 63 6.97 4.96 0.00 15.00 compose the model II.
Fat 119 3.43 1.65 1.28 5.40
Phosphate 491 1.33 0.27 0.98 1.92
L-Lysine 491 0.43 0.08 0.29 0.57 each variable to evaluate whether the equation generated
DL-Methionine 491 0.22 0.06 0.10 0.35 was biologically coherent and sensible. Descriptive statistics
L-Threonine 430 0.15 0.03 0.06 0.23 (means, deviations and ranges) were calculated for each
L-Tryptophan 311 0.03 0.03 0.00 0.10 selected variable. After generating the first model, the pro-
Acidifying 491 0.27 0.11 0.00 0.40 cedure was performed again, but with the ‘nitrogen reten-
Zinc oxide 491 0.12 0.12 0.00 0.40 tion’ variable removed from the model. Each individual
Antibiotic 491 0.04 0.03 0.00 0.10
animal represented the experimental unit in all studies,
Calculated composition
ME4 (MJ/kg) 491 14.85 7.69 13.36 16.87
and the code for study effect was considered as a random
CP5 (%) 491 20.17 1.32 17.50 24.00 effect into the model. Graphs for predicted and observed val-
Total lysin (%) 491 1.44 0.11 1.13 1.71 ues were generated for each equation using the values of the
Ca6 (%) 491 0.73 0.07 0.60 0.93 trials selected to compose the database. All analyses were
P7 (%) 491 0.48 0.02 0.43 0.55 performed in the statistical software Minitab 18 (Minitab
Fat (%) 391 6.51 2.94 1.17 11.15 Inc., State College, PA, USA), and the variables were consid-
n=Number of rows of database with the ingredient; SD=standard deviation;
ered to have significant effects when P<0.05.
Min=minimum; Max=maximum.
1
Nuklospray E50®(Sloten do Brasil Ltda, Santos, São Paulo, Brasil).
2
Betaglucano (Betamune, Biorigin – São Paulo, Brasil).
3
Promosoy 90® (Central Soya Inc., Chicago, Ill). Results
4
Metabolizable energy.
5
Crude protein. The pre-selected variables are described in Table 2. They had
6
Calcium.
7
Phosphorus. values of correlation between 0.014 and 0.690, and the vari-
ables with the highest values were RN and nitrogen intake (NI)
(0.689) followed by NI and body weight gain (BWG) (0.683).
model was generated to estimate the ratio of weight gain to Descriptive statistics for the variables that were significant in
excreted nitrogen (BWG/EN) using a generalized linear the models are presented in Table 3, and the adjusted coefficient
model procedure. The model was evaluated through graphi- values of the explanatory variables are presented in Table 4.
cal inspection and performing a Ryan–Joiner test for normal- Model II did not include the RN variable, but included
ity on the residuals. The authors analyzed the behavior of FN g/day), which was not included in model I. Figure 1 shows

3
Cardinal, Vieira, Warpechowski, Zielgelmann, Montagne, Andretta and Ribeiro

Table 3 Descriptive statistics of the significant explanatory variables selected to compose models I and II to predict the weight gain relative to
excreted nitrogen (BWG/EN) of piglets
Variable Mean SD Min Max Median

Dependent
Body weight gain/excreted nitrogen 90.99 31.01 37.76 185.52 86.98
Covariate
Urinary nitrogen (g/day) 2.75 1.74 0.37 14.32 2.50
Fecal nitrogen (g/day) 2.33 1.63 0.41 9.89 1.98
Retained nitrogen (g/day) 15.10 6.32 2.49 32.75 14.10
Body weight gain (g/day) 406 147 135 918 383
Dry matter intake (g/day) 590 213 191 1155 569
Nitrogen intake (g/day) 20.20 8.31 7.19 45.59 19.17
Metabolizable energy intake (MJ/kg) 9.836 3.787 0.669 20.18 9.30
SD = standard deviation; Min = minimum;Max = maximum.

Table 4 Model parameters of the explanatory variables adjusted for the weight gain relative to excreted nitrogen (BWG/EN) of piglets
Body weight gain/excreted nitrogen
Model I Model II1
Variables Coefficient CSE P-value Coefficient CSE P-value

Intercept −156.56 9.14 <0.001 83.83 2.25 <0.001

Urinary nitrogen (g/day) −1.774 0.67 0.009 −15.813 0.74 <0.001
Fecal nitrogen (g/day) – – – −11.143 0.71 <0.001
Retained nitrogen (g/day) 3.267 0.00 <0.001 – – –
Body weight gain (g/day) 0.178 0.00 <0.001 0.179 0.00 <0.001
Dry matter intake (g/day) 0.025 0.11 0.047 0.020 2.86 0.047
Nitrogen intake (g/day) −4.844 0.24 <0.001 −1.857 0.32 <0.001
Metabolizable energy intake (MJ/kg) 0.007 0.00 <0.001 0.012 0.00 <0.001
R2 90.13 82.71
CSE=Coefficient standard error.
1
The analysis of model II did not include nitrogen retained (RN).

the observed BWG/EN values compared to those predicted by demonstrating greater accuracy for low and medium values
models I and II. of BWG/EN in both equations.
The equation generated for model I was:

BWG=EN ¼ 156 56  1:77 UN þ 3:26 RN þ 0:17 BWG Discussion

þ0:02 DMI  4:84 NI þ 0:007 MEI R2 ¼ 0:9012
(1)
While that for model II was:
BWG=EN ¼ 83:83  15:81 UN  11:14 FN þ 0:179 BWG

bility studies and, consequently, the number of variables is
þ0:02 DMI  1:85 NI þ 0:01 MEI R2 ¼ 0:8271
(2)
Where BWG/EN is the ratio between body weight gain and
excreted nitrogen; UN is urinary nitrogen; FN is fecal nitro-
gen; RN is nitrogen retained; BWG is weight gain; DMI is
dry matter intake; NI is nitrogen intake; and MEI is metabo-
lizable energy intake.
The graphs for observed and predicted values are shown
in Figure 1. The values presented linear behavior (P<0.001),
The equations generated in this study make it possible to
extend the exploration of data obtained in digestibility stud-
ies. Besides analyzing the efficiency in the use of a nutrient or
ingredient, it is also possible to analyze the results in order to
increase environmental sustainability.
All predictive variables were obtained only from digesti-
limited in the database. For that reason, some variables were
calculated and all variables with r<0.07 were considered as
predictors. This is a feature that allows the models to be
applied in the future, since the equations are simple to inter-
pret and include variables that are often used in digestibility
studies, such as ‘Fecal Nitrogen’, ‘Urinary Nitrogen’ or
‘Retained Nitrogen’. Another limitation of database was that
some variables, as ingredients, were not equally distributed
among experiments, which reduced their potential as
predictors.

4

200
R 2:90.13
Observed BWG/EN
150
100
50
0 0
0 50 100 150
Predicted BWG/EN
Equation I
Figure 1 (colour online) Comparison between observed values of the weight gain/excreted nitrogen (BWG/EN) of piglets and those predicted using models I
and II.
Previous research demonstrated that nutrition and feed-
ing practices have a direct influence on the environmental
impact generated by pig production chain (Cherubini
et al., 2015; McAuliffe et al., 2016). The effect of increased
N intake on performance variables and nitrogen excretion is
well documented in the literature. However, the relationships
among these variables have not previously been investigated
to understand the main factors that affect this relationship.
From an economic and environmental point of view, it is
desirable that BWG is maximized while N excretion is
minimized (Pomar et al., 2017).
In general, the means of feed intake and weight gain
observed in the database constructed in this study were
within the expected range for the evaluated range of piglet
weights. The good performance observed was likely related
to the high nutrient digestibility coefficients of the feed used,
which reached mean values >80%. This was due to the qual-
ity of the diets, which were formulated with ingredients of
high digestibility and associated with good animal health
status.
As expected, model I had a higher coefficient of determi-
nation than model II (R2=0.9013 v. 0.8271), due to the better
adjustment of predicted values by information on N reten-
tion. This result is coherent with the importance of the
self-explanatory RN variable within model I. Although model
II presents a lower R2, this model may also be an alternative
to be used in studies that measure urinary N and fecal N. The
relation between predicted and observed BWG/EN is not
totally linear (Figure 1). The models fit better for the BWG/
EN ranges where there are more observations in the database
for the models to be calibrated, and therefore it is possible to
observe that higher values of BWG/EN make the equations
less precise. It may be associated with different factors, such
the ingredients’ digestibility, that decrease along the feeding
phases, or the age of the pigs, because of the decrease in
nitrogen retention efficiency. Also, equation 2 has high vari-
ability than equation 1 with higher BWG/EN values. Since the
main difference between the equations 1 and 2 is FN and
nitrogen retained, the greater variability indicates that the
nitrogen retained can be more accurate that the first one.
Modeling weight gain/excreted nitrogen in piglets
200
R 2:82.71
Observed BWG/EN
150
100
50
200 0 50 100 150 200
Predicted BWG/EN
Equation II
Future studies should be focused to better understand these
patterns to improve the capacity of the models to predict high
values of BWG/EN.
The ability of weaned piglets to preferentially deposit
protein over fat demonstrates the efficiency of growth perfor-
mance and N excretion in these animals. The average nitro-
gen retention rate (retention/intake) observed in the studies
that composed our database was 75%, showing that the effi-
ciency of the use of this nutrient was high, and it is in accor-
dance with values of nitrogen retention found by Straarup et
al. (2006). In contrast, when pigs were evaluated in the
grower-finishing phase, when they were fed on cereal-based
diets, only 35% of the ingested nitrogen was retained
(Dourmad et al., 1999 and 2009). Shirali et al. (2012)
revealed that nitrogen excretion is changing throughout
the entire growing period. Greatest nitrogen retention rate
(32%; nitrogen retention/NI), least nitrogen excretion and
greatest retention were found during early growth from 60
to 90 kg. Nitrogen efficiency gradually decreased with
increasing growth stages.
Energy and protein are the major nutritional factors
involved in protein gain efficiency in young piglets. Van
Milgen and Noblet (2003) reported that the response of pro-
tein gain to energy intake is curvilinear, and protein enhance-
ment in growing animals is protein- and energy-dependent.
The relationship between energy and BWG/EN in our results
showed that higher metabolic energy intake (MEI) resulted in
increased growth efficiency. According to Möhn and De
Lange (1998), in a stress-free environment with adequate
supplies and intake of energy and nutrients, the amount
of protein deposition is determined by energy intake or
the genetic potential for maximum deposition. Thus, the
main nutritional limitation of protein deposition in young
pigs is energy consumption (Dunshea et al., 1998).
The strategy of increasing dietary energy concentration has
been used to minimize the impacts of the lower performance of
piglets in the initial phase (Beaulieu et al., 2006; Oresanya et al.,
2008). Ribeiro et al. (2016) and Arnaiz et al. (2009) observed
higher feed intake and N retention when dietary energy content
was increased. In our model equations, increased MEI resulted
5
Cardinal, Vieira, Warpechowski, Zielgelmann, Montagne, Andretta and Ribeiro
in improved growth v. N excretion efficiency. Therefore, the
strategy of increasing the concentration of energy in feed could
also be used to reduce the environmental impact of raising pigs,
since the model shows an increment of the efficiency increasing
energy intake.
The result of increasing NI, in both equations, was a reduc-
tion in growth efficiency. However, nutritional plans with
high levels of CP are frequently used for weaned piglets
under practical conditions. For most nutrients ingested, maxi-
mum retention is achieved on a decreasing efficiency curve,
and this is particularly true for proteins and amino acids; as
ingestion increases, there is an increase in rates of degrada-
tion and excretion by the animal (Rezaei et al., 2013). High
levels of dietary protein increase amino acid deamination,
with a consequent increase in heat production (Roth et al.,
1999; Van Milgen and Noblet, 2003) and reduction in the
energy available for protein deposition (Le Bellego et al.,
2001), making the growth process less efficient in terms of
both performance and excretion. Webb et al. (2014) reviewed
studies that evaluated the relationship between reduced CP
in the diet and N excretion and concluded that for each 1 unit
reduction in CP, nitrogen excretion can be reduced by up to
10%. Therefore, considering the high efficiency of weaned
piglets in depositing protein and the high safety margins used
in diet formulations, the best strategy to improve efficiency
and reduce the amount of N excreted would be to reduce
their intake of this element, and perhaps tolerate lower
growth rates (Yin et al., 2010; Gloaguen et al., 2014).
The estimated BWG/EN ratio used in this study can assist
in improving the productivity and sustainability of current
swine production methods, and is also easy to interpret. It
should, however, be remembered that a static model is
not the best predictor of efficiency in all practical situations,
where other factors also interfere in the dynamics evaluated
herein, such as animal health status, thermal environment,
age and genetic capacity for lean tissue gain. Genetic
advances have produced animals with greater potential gains
of lean tissue, less requirement for maintenance (Boddicker
et al., 2011), lower heat production (Barea et al., 2010) and
better efficiency of nutrient usage. Since growth rate is linked
to production efficiency, this implies that if the animal per-
forms better, associated with better feed conversion ratio,
it will pollute less.
The vast majority of research on the impact of N excretion
on swine production systems has focused on the final phase
of the productive cycle, due to the higher volume of waste
produced and the lower efficiency in nutrient utilization com-
pared to those of piglets in the initial phase. However, when
considering intensive production systems, the volume of
waste produced by animals in the initial phase is also poten-
tially impactful and should be considered.
Conclusions
The models generated to estimate the efficiency of weight gain
v. excreted nitrogen had high coefficients of determination,
6
and the amount of nitrogen ingested was the variable that
had the greatest influence on efficiency in both models.
When the ‘nitrogen retained’ variable was removed from
the model, there was a reduction in the fit of the model
equation.
Acknowledgements
We thank the Animal Research Laboratory (LEZO) of the
Universidade Federal do Rio Grande do Sul – Brazil for
providing the material and data of the digestibility assays.
K. M. Cardinal, 0000-0002-8419-509X
Declaration of interest
We declare no conflict of interest, including financial, personal
or other relationships with other people or organizations that
could inappropriately influence this work.
Ethics statement
This study was performed with data from published theses and
dissertations; consequently there was no use of animals. In this
situation, it is not necessary for the study to be approved by the
Animal Ethics Committee of the Universidade Federal do Rio
Grande do Sul.
Software and data repository resources
None of the data were deposited in an official repository.
References
Andretta I, Haushild L, Kipper M, Pires PG and Pomar C 2017. Environmental
impacts of precision feeding programs applied in pig production. Animal 12,
1990–1998.
Arnaiz V, Ribeiro AM, Kessler AM, Raber M and Kuana S 2009. Efecto del peso al
destete, temperatura ambiental y energía metabolizable del pienso en lechones
recién destetados. Revista Brasileira de Ciências Agrárias 4, 472–478.
Barea R, Dubois S, Gilbert H, Sellier P, Van Milgen J and Noblet J 2010. Energy
utilization in pigs selected for high and low residual feed intake. Journal of
Animal Science 88, 2062–2072.
Beaulieu A, Levesque C and Patience J 2006. The effects of dietary energy con-
centration and weaning site on weaning pig performance 1. Journal of Animal
Science 84, 1159–1168.
Berrocoso J, Saldaña B, Serrano M, Cámara L, Ibáñez M and Mateos G 2013.
Influence of crude protein content, ingredient complexity, feed form, and dura-
tion of feeding of the phase I diets on productive performance and nutrient
digestibility of Iberian pigs. Journal of Animal Science 91, 1237–1246.
Boddicker N, Gabler N, Spurlock M, Nettleton D and Dekkers J 2011. Effects of ad
libitum and restricted feeding on early production performance and body com-
position of Yorkshire pigs selected for reduced residual feed intake. Animal 5,
1344–1353.
Cherubini E, Zanghelini GM and Tavares JMR 2015. The finishing stage in swine
production: influences of feed composition on carbon footprint. Environment,
Development and Sustainability 17, 1313–1326.
Dourmad J, Seve B, Latimier P, Boisen S, Fernandez J, Van der Peet-Schwering C
and Jongbloed A 1999. Nitrogen consumption, utilisation and losses in pig
production in France, The Netherlands and Denmark. Livestock Production
Science 58, 261–264.
Dourmad J-Y, Hassouna M, Robin P, Guingand N, Meunier-Salaün M and Lebret B
2009. Influence of pig rearing system on animal performance and manure
composition. Animal 3, 606–616.

Dunshea F, Eason P, Campbell R and King R 1998. Interrelationships between
dietary ractopamine, energy intake, and sex in pigs. Australian Journal of
Agricultural Research 49, 565–574.
Garcia-Launay F, Van der Werf H, Nguyen TTH, Le Tutour L and Dourmad J-Y
2014. Evaluation of the environmental implications of the incorporation of
feed-use amino acids in pig production using Life Cycle Assessment.
Livestock Science 161, 158–175.
Gloaguen M, Le Floc’h N, Corrent E, Primot Y and Van Milgen J 2014. The use of
free amino acids allows formulating very low crude protein diets for piglets.
Journal of Animal Science 92, 637–644.
Gloaguen M, Le Floc’h N, Primot Y, Corrent E and Van Milgen J 2013. Response
of piglets to the standardized ileal digestible isoleucine, histidine and leucine
supply in cereal–soybean meal-based diets. Animal 7, 901–908.
Guggenbuhl P, Waché Y and Wilson J 2012. Effects of dietary supplementation
with a protease on the apparent ileal digestibility of the weaned piglet. Journal
of Animal Science 90, 152–154.
Hutchings NJ, ten Hoeve M, Jensen R, Bruun S and Søtoft LF 2013. Modelling the
potential of slurry management technologies to reduce the constraints of
environmental legislation on pig production. Journal of environmental
management 130, 447–456.
Knap PW and Bishop SC 2000. Relationships between genetic change and infec-
tious disease in domestic livestock. BSAP Occasional Publication 27, 65–80.
Le Bellego L, Van Milgen J, Dubois S and Noblet J 2001. Energy utilization of low-
protein diets in growing pigs. Journal of Animal Science 79, 1259–1271.
Lovatto P, Lehnen C, Andretta I, Carvalho A and Hauschild L 2007. Meta-análise
em pesquisas científicas-enfoque em metodologias. Revista Brasileira de
Zootecnia 36, 285–294.
McAuliffe GA, Chapman DV and Sage CL 2016. A thematic review of life cycle
assessment (LCA) applied to pig production. Environmental Impact Assessment
Review 56, 12–22.
Metzler-Zebeli B and Zebeli Q 2013. Cereal β-glucan alters nutrient digestibility
and microbial activity in the intestinal tract of pigs, and lower manure ammonia
emission: a meta-analysis. Journal of Animal Science 91, 3188–3199.
Möhn S and De Lange C 1998. The effect of body weight on the upper limit to
protein deposition in a defined population of growing gilts. Journal of Animal
Science 76, 124–133.
Monteiro ANTR, Garcia-Launay F, Brossard L, Wilfart A and Dourmad JY 2016.
Effect of feeding strategy on environmental impacts of pig fattening in different
contexts of production: evaluation through life cycle assessment. Journal of
Animal Science 94, 4832–4847.
Montes F, Meinen R, Dell C, Rotz A, Hristov AN, Oh J and Dijkstra J 2013. Special
topics – Mitigation of methane and nitrous oxide emissions from animal oper-
ations: II. A review of manure management mitigation options. Journal of
Animal Science 91, 5070–5094.
Morel PCH and Wood GR 2005. Optimisation of nutrient use to maximise
profitability and minimise nitrogen excretion in pig meat production systems.
In III International Symposium on Applications of Modelling as an Innovative
Technology in the Agri-Food Chain, May 2005, Leuven, Belgium, pp. 269–275.
Modeling weight gain/excreted nitrogen in piglets
National Research Council (NRC) 2012. Nutrient requirements of swine, 11th
revised edition. National Academy Press, Washington, DC, USA.
Oresanya T, Beaulieu A and Patience J 2008. Investigations of energy metabo-
lism in weaning barrows: the interaction of dietary energy concentration and
daily feed (energy) intake. Journal of Animal Science 86, 348–363.
Pomar C, Andretta I and Haushild L 2017. Meeting individual nutrient require-
ments in growing pigs to optimize nutrient efficiency. In Achieving sustainable
production of pig meat (ed. J Wiseman), pp.1–14. Burleigh Dodds Science
Publishing, London, UK.
Rezaei R, Wang W, Wu Z, Dai Z, Wang J and Wu G 2013. Biochemical and physio-
logical bases for utilization of dietary amino acids by young pigs. Journal of
Animal Science and Biotechnology 4, 7.
Ribeiro AML, Farina G, Vieira MDS, Perales VA and Kessler ADM 2016. Energy
utilization of light and heavy weaned piglets subjected to different dietary
energy levels. Revista Brasileira de Zootecnia 45, 532–539.
Rostagno HS, Teixeira Albino LF, Donzele JL, Gomes PC, de Oliveira RF, Lopes DC
and Euclides RF 2011. Tabelas Brasileiras para aves e suinos: Composição de
Alimentos e Exigências Nutricionais, 3ª Edição. Universidade Federal de
Viçosa–Departamento de Zootecnia, Viçosa, Minas Gerais, Brazil.
Roth F, Gotterbarm G, Windisch W and Kirchgessner M 1999. Influence of dietary
level of dispensable amino acids on nitrogen balance and whole-body protein
turnover in growing pigs. Journal of Animal Physiology and Animal Nutrition
81, 232–238.
Sauvant D, Schmidely P and Daudin J 2005. Les méta-analyses des données
expérimentales: applications en nutrition animale. INRA Productions
Animales 18, 63–73.
Sauvant D, Schmidely P, Daudin J-J and St-Pierre NR 2008. Meta-analyses of
experimental data in animal nutrition. Animal 2, 1203–1214.
Shirali M, Doeschl-Wilson A, Knap PW, Duthie C, Kanis E, Van Arendonk JAM
and Roehe R 2012. Nitrogen excretion at different stages of growth and its asso-
ciation with production traits in growing pigs. Journal of Animal Science 90,
1756–1765.
Straarup EM, Danielsen V, Høy CE and Jakobsen K 2006. Dietary structured lipids
for post-weaning piglets: fat digestibility, nitrogen retention and fatty acid
profiles of tissues. Journal of Animal Physiology and Animal Nutrition 90,
124–135.
Van Milgen J and Noblet J 2003. Partitioning of energy intake to heat, protein,
and fat in growing pigs. Journal of Animal Science 81, 86–93.
Webb J, Broomfield M, Jones S and Donovan B 2014. Ammonia and odour
emissions from UK pig farms and nitrogen leaching from outdoor pig production.
A review. Science of the Total Environment 470, 865–875.
Yáñez J, Landero J, Owusu-Asiedu A, Cervantes M and Zijlstra R 2013. Growth
performance, diet nutrient digestibility, and bone mineralization in weaned pigs
fed pelleted diets containing thermostable phytase. Journal of Animal Science
91, 745–754.
Yin Y, Yao K, Liu Z, Gong M, Ruan Z, Deng D, Tan B, Liu Z and Wu G 2010.
Supplementing L-leucine to a low-protein diet increases tissue protein synthesis
in weaning pigs. Amino Acids 39, 1477–1486.