Published September 8, 2015

The association between immunoglobulin G in sow colostrum and piglet plasma1

C. Kielland,*2 V. Rootwelt, O. Reksen,* and T. Framstad*

*Department of Companion Animal Clinical Sciences, Faculty of Veterinary Medicine and Biosciences, Norwegian
University of Life Sciences, N-0033 Oslo, Norway; and Department of Production Animal Clinical Sciences,
Faculty of Veterinary Medicine and Biosciences, Norwegian University of Life Sciences, N-0033 Oslo, Norway

ABSTRACT: Colostrum provides newborn piglets
with energy and passive immunity and is essential
for survival of the piglets. The plasma concentration
of immunoglobulin G (IgG) in piglets is dependent
on several factors, most importantly the concentration
of IgG in sow colostrum (colostrum IgG). The main
aims of this study were to investigate the variation in
concentration of colostrum IgG between herds and the
individual sows within herd and to investigate factors
associated with plasma IgG concentrations in piglets
(piglet IgG). From 4 herds (A to D), 876 piglets from
62 sows were included in the study. Colostrum was
sampled from sows immediately after expulsion of
the first piglet and before the first suckling (t1), mid-
way through farrowing (just after the sixth piglet was
born; t2), and after the last piglet was born (t3). At
d1, 0.5mL blood from piglets was collected in tubes
containing EDTA, and IgG concentrations were ana-
lyzed. Mean colostrum IgG concentration across all
herds was 53.9 g/L. Herd A had mean colostrum IgG
of 38.3g/L, whereas the other 3 herds (B,C, and D)
had mean colostrum IgG of 47.4, 60.4, and 67.8 g/L,
respectively. Colostrum IgG at t1, t2, and t3 across all
herds was 56.2, 53.7, and 42.5 g/L, respectively. Mean
concentration of piglet IgG across all samplings was
21.7 g/L. Multilevel linear regression analysis was
performed with piglet IgG (g/L) as outcome. In this
model, the herd effect accounted for 9% of the total
variance and 34% of the variance resided at sow level.
Piglet IgG was associated with herd, birth order (n),
body mass index (BMI) > 17 (kg/m2), and colostrum
IgG at t1 (g/L) with an overall P-value < 0.01. Herd
D had the highest predicted mean level of piglet IgG.
The main model predicted that piglet IgG decreased
linearly by 0.4 g/L with each piglet born (P < 0.01).
The model also predicted an increase by 0.1 g/L for
each gram per liter extra colostrum IgG in colostrum
(P = 0.03). Piglets with a BMI above 17 kg/m2 had a
greater piglet IgG (+4.5 g/L) than those with a BMI at
17 kg/m2 or below (P < 0.01). Concentrations of colos-
trum IgG varied largely between herds and between
sows. The largest variation of piglet IgG was mainly
on the piglet level, supporting the complex nature of
IgG production and uptake. However, the strong asso-
ciation between colostrum IgG and piglet IgG shows
that increased IgG level in colostrum will improve the
levels of IgG in piglets and potentially increase sur-
vival of the piglets.

Key words: body mass index, colostrum, immunoglobulin G, piglet, plasma, sow, swine

2015 American Society of Animal Science. All rights reserved. J. Anim. Sci. 2015.93:44534462
doi:10.2527/jas2014-8713
INTRODUCTION

Colostrum provides newborn piglets with energy

for growth and heat production (Le Dividich, and
Noblet, 1984; Herpin et al., 1994), provides passive
1We would like to extend our thanks to the farmers participating
immunity, and is essential for health and survival
in this work-intensive study and also for letting us take samples (Sangild, 2003). In swine, the epitheliocorial nature
from their sows and piglets. of the placenta prevents passive immunity transfer to
2Corresponding author: camilla.kielland@nmbu.no
the fetus. Immunoglobulin G (IgG) is the most clini-
Received November 17, 2014.
cally important globulin during the first weeks of life
Accepted July 12, 2015.
4453
4454
Table 1. Descriptive data from the 4 piglet producing herds
A B
Item LandraceLandrace
Farrowing pen/creep area 7.7 m2/0.9 m2
Feeding of sows in 9.29 MJ NE/kg,
farrowing pen 8.26 g lysine/kg,
and 14.4%/kg CP
In farrowing pen >7 d before
farrowing
No. of sows2 15
No. total born piglets 185
No. total born piglets per sow 12.3
1Optimized liquid feed with the aim of providing 9.86 MJ NE/kg, 8.26 g lysine/kg, and 14.7%/kg CP.
2All sows from herds A, B, and D were from 1 batch, and the 12 sows from herd C were from 6 batches.
and IgG from colostrum is absorbed over the gastro-
intestinal tract during the first 24 to 48 h postpartum
(Sjaastad et al., 2012). However, absorption of IgG in
piglets is dependent on the timing of gut closure (Rooke
and Bland, 2002). Genotype, parity, age, vaccination
status, and endocrine status of the sow; fat composi-
tion of feed; and herd management are reported to in-
fluence colostrum yield and composition (Devillers et
al., 2007; Farmer and Quesnel, 2008; Eastwood et al.,
2014). Colostrum IgG varies greatly between sows. The
concentration of IgG in sow colostrum is reported to
range from 48.0 to 95.6 g/L in single herds (Klobasa
and Butler, 1987; Tuchscherer et al., 2002; Svendsen
et al., 2005; Couret et al., 2009; Foisnet et al., 2010;
Bovey et al., 2014). In addition to animal factors, herd
management is an important factor influencing piglet
IgG. In a review by Kirkden et al. (2013), they suggest
management solutions that will improve both colostrum
IgG and piglet IgG concentrations. Providing extra at-
tentions to small piglets in nonhomogenized litters may
be necessary to ensure absorption of enough IgG in pig-
lets (Muns et al., 2014). Uptake of IgG can be indirectly
measured by recording piglets IgG in serum after birth
(Svendsen et al., 2005). The aims of this study were to
investigate the variation in concentration of colostrum
IgG between herds and the individual sows within herd
and to investigate factors associated with plasma IgG
concentrations in piglets (piglet IgG).
MATERIAL AND METHODS
The experimental protocol for this study did not
require approval by the Norwegian Animal Research
Authority due to an exception for such procedures in the
Norwegian regulations for animal testing (FOR 1996-
01-15 no. 23, Regulation of animal testing, 2: Scope).
Kielland et al.
Herd
C D
Sow
LandraceYorkshire
Boar
LandraceDuroc LandraceDuroc LandraceDuroc
7.5 m2/0.9 m2 7.3 m2/0.8 m2 7.0 m2/1.3 m2
Liquid feed1 9.86 MJ NE/kg, 8.26 g lysine/kg,
and 14.7%/kg CP
45 d before 3 wk before farrowing
farrowing
17 12 18
208 222 261
12.2 18.5 14.5
Animals
The study material comprises LandraceYorkshire
sows and piglets from 4 piglet-producing herds (A to
D); details are shown in Table 1. Herds A and B were
conventional breeding herds with lactation and transi-
tion phase, and the other 2 herds were farrow to 30 kg
satellite herds from the same sow pool system (C and
D). In a sow pool system, sows are artificially insemi-
nated at the central unit. Then, sows are transported
to a satellite herd 3 wk before the expected farrowing
date and they stay there until weaning of piglets. After
weaning, the sows are returned to the central unit,
where they are again inseminated or culled.
In the present study, 62 sows were included. All
sows from herds A, B, and D were from 1 batch, and
the 12 sows from herd C were from 6 batches. The
number of animals included in the experiment was
set to ensure reliable data recording and proper ani-
mal monitoring during the experiment. Data for from
herds C and D have been used in 3 previous studies
already published (Rootwelt et al., 2012a,b, 2013). In
these papers, however, IgG values were displayed at
group level and not used on individual piglet level as
in the present study. Data from herds A and B have not
previously been shown. All sows were inseminated
with heterospermic semen from LandraceLandrace,
LandraceDuroc, or DurocDuroc boars. The study
period was from June 2009 to November 2011. All
herds were located in the east and southeast of Norway.
From 62 sows, 876 piglets were initially included in
the study, and any animal with missing data at the sow
or piglet level was excluded from the statistical analy-
ses and only piglets with complete data on all vari-
ables were used in the main regression analysis.
 Immunoglobulin G in sows and piglets
Management
Detailed housing and management characteris-
tics are presented in Table 1. All herds used the same
vaccination regime, for example, vaccination against
Escherichia coli, porcine parvovirus, and Erysipelothrix
rhusiopathiae. In each herd, gilts and sows were indi-
vidually kept in loose-housed standard farrowing pens
with a piglet creep area. The size of pens varied, but
all pens were of equal size within each herd (details
in Table 1). The sows were fed a commercial lactation
diet, including small amounts of hay before farrow-
ing (Table1). All sows had ad libitum access to water.
Farrowing was attended and allowed to start naturally.
The automatic temperature regulation of the gestation
rooms were set in all herds at 18 to 19C. Manual inter-
vention during farrowing was performed by the farmer
if the birth interval between 2 piglets exceeded 90 min.
In herd D, the first 6 to 8 firstborn piglets were dried
and placed under a heat lamp until the second colos-
trum sample (midway through farrowing [just after the
sixth piglet was born; t2]) was taken; see more details
in the paper by Rootwelt et al. (2012b). Thereafter, the
remaining piglets were placed at the udder. All herds
had a mean weaning age of 33.0 d (1.7).
Recorded Variables
Variables recorded at the herd level were introduc-
tion to farrowing pen, size of farrowing pen, and type
of breed used when inseminating. Sow level variables
recorded at each farrowing were parity, farrowing length,
birth interval, litter size, and colostrum IgG.
Colostrum was sampled from sows immediately
after expulsion of the first piglet and before the first
suckling (t1), at t2, and after the last piglet was born
(t3). The definition of last piglet born was when the
contractions had stopped and after the expulsion of
placenta. Colostrum was collected from 3 random
teats located in the anterior, middle, and posterior
part of the udder. Colostrum from these 3 teats was
pooled to 5 to 10 mL at each time point (t1, t2, and t3).
The samples were frozen at 40C and later analyzed
for concentration of colostrum IgG with an in-house
single radial immunodiffusion test kit (VMRD Inc.,
Pullman, WA). All of the laboratory analyses included
2 duplicates from the same sample. The mean of the
duplicates is the final value used and compared with
the other values obtained from the rest of the samples.
Time elapse from expulsion of the first piglet until
the birth of each subsequent piglet was recorded (birth
time) as well as sex. All live piglets were weighed at
birth (birth weight [BiW]; n = 805). In addition, we also
obtained the weight in 566 piglets the next day (weight
on d 1 [WD1]). The time from birth to second weighing
4455
was registered. Body length was measured within 24 h
postpartum from the os occipitale to the root of the tail.
Body mass index (BMI) was calculated using BW (kg)
and the square of body length (m2) at birth: BMI = BiW
(kg)/the square of body length (m2).
At d 1, 0.5 mL whole blood was evacuated from
the vena jugularis externa/interna/communis of the
piglets and collected in tubes containing EDTA as
an anticoagulant (n = 692). Time from birth to blood
sampling was recorded. The blood samples were sub-
sequently stored at room temperature until they were
centrifuged at 3,000 g for 10 min at 20C the next
working day. Plasma was then frozen at 40C until
analysis of IgG with an in-house single radial immu-
nodiffusion test kit (VMRD Inc.). Values below the
lower detection limit of 3.8 g/L were defined as 0.
Colostrum yield was estimated by recording and
adding the weight gain for each piglet from birth to d 1
(Couret et al., 2009). Estimated measures of colostrum
intake per piglet were calculated using a formula de-
signed by Theil et al. (2014): colostrum intake (g) = 106
+ (2.26 WG1) + (200 BiW)+ (0.111 TS) (1,414
WG1)/TS + (0.0182 WG1)/BiW. Time suckling (TS)
was calculated from the time of birth until time of sec-
ond weighing. Weight gain (WG1) was the difference
between BiW and WD1.
Finally, regarding the regression analysis with pig-
let IgG concentration as the outcome, only 644 piglets
of the 692 originally blood sampled animals were used.
Four sows were excluded from the regression analysis
due to missing values of colostrum IgG at t1, and their
piglets were, therefore, also excluded. In litters with
few piglets (<8), colostrum sampling was performed
only at t1 and t2, and this led to several missing values
at t3. Additionally, in herd D, no samples were collected
at t3 due to the protocol in this herd. We also had miss-
ing WD1 values; however, this variable was not used in
the regression analysis and did not influence the number
of animals in the regression analysis.
Statistical Analyses
Descriptive statistics were performed using JMP
(SAS Inst. Inc., Cary, NC) and STATA (Stata SE/10
for Windows; Stata Corp., College Station, TX). This
was performed on the herd, sow, and piglet level. Box
plots were used to explore and describe the develop-
ment of IgG concentrations in colostrum throughout
the farrowing in relation to factors at the sow level.
The outcome variable chosen for the regression
analysis was piglet plasma IgG (g/L) on d 1. Predictor
variables used in the analysis at sow level were par-
ity, farrowing length, litter size, average birth intervals,
and colostrum IgG at t1. Predictor variables at piglet
4456 Kielland et al.
Figure 1. Box plot of colostrum immunoglobulin G (g/L) concentra-
tions from 62 sows in 4 herds at the 3 time points at farrowing: immediately
after expulsion of the first piglet and before the first suckling (t1), midway
through farrowing (just after the sixth piglet was born; t2), and after the last
piglet was born (t3). Value at t2 and t3 include only herds A, B, and C. Box
indicates median and upper (75th) and lower (25th) percentile. Vertical lines
are upper and lower adjacent value. Colostrum was sampled at t1, t2, and t3.
level were sex, birth order, birth time, and BMI at birth.
Linearity between the continuous and dichotomous
variables was investigated with graphs using a logit
function in STATA, creating a scatterplot smoothing
(lowess) line between the 2 variables. Plots, best linear
fit, and R2 were used to explore how different predictors
explained the variation in colostrum IgG and piglet IgG.
As one of the aims was to explore the level of varia-
tion on each hierarchical level, the multilevel mixed ef-
fects linear regression model (xtmixed in STATA) was
built by including herds and sows as random effects.
However, as the 4 herds were individually selected and
not randomly included in the study, the final model in-
cluded herd as fixed effect (1) and sow as random ef-
fect. Therefore, none of the predictor variables at herd
level could be included in the statistical model due to
collinearity. This final model was used to estimate pre-
dictors associated with 644 piglet IgG samples.
In the model-building process, each variable was sep-
arately tested without any random effects included, and
variables with a P-value < 0.20 within this univariable
analysis were considered for inclusion in the final model.
When building the final model, a forward stepwise
technique was used, starting with the variables with
the lowest P-values from the separate variable analy-
sis (Kielland et al., 2010). Therefore, any distortion
and confounding could be observed as each variable
was included. Confounding variables were tested by
running the model with and without that variable.
Running the model with and without extreme values
of predictors was also performed. When exploring
influencing values, no values were deleted from the
analysis due to high influence. The variables giving a
model with the best fit were chosen. If variables were
highly correlated with each other (|| > 0.8; Dohoo et
al., 2009), only 1 of these variables was included.
In all analyses, statistical significance was consid-
ered with a P-value < 0.05. Goodness of fit for each
model was evaluated and the best model was chosen.
Normality plots of standardized residuals were evalu-
ated and potential outliers and observations with large
influence were explored. The model was tested with
and without possible influencing points.
RESULTS
Sow and Litter Characteristics
Total mean parity was 2.7, ranging from 1 to 8, of
which 28% were primiparous, 25% were parity 2, 24%
were parity 3, and 23% were parity 4 to 8. Average litter
size was 14.1 piglets, and the average farrowing length
was 4 h, ranging from 1 to 12 h. Average birth interval
between piglets was 18.3 min, ranging from 0 min to
8h. One piglet had an interval of 8 h after the previous
piglet, and this piglet was born dead and as the last one
in that litter. After excluding that 1 piglet from the de-
scriptive analysis, the range in birth interval was 0 min
to 3 h. Number of piglets that were dead at birth was,
on average, 1.27 per litter, ranging from 0 to 7 piglets.
Average colostrum yield per sow at d 1, estimated from
piglets individual weight gain, was 1.4 kg, ranging
from 0.33 to 2.69 kg. Estimated colostrum yield was
not significantly associated with litter size (P = 0.69).
Mean colostrum IgG across all samples was
53.9g/L. Herds A and B had a mean colostrum IgG of
38.3 and 47.4 g/L, respectively, whereas herds C and D
had a mean colostrum IgG of 60.4 and 67.8 g/L. Mean
colostrum IgG at t1, t2, and t3 for all herds decreased
throughout farrowing: 56.2, 46.8, and 42.5 g/L, respec-
tively (Fig. 1; Table 2). Mean values at t2 and t3 do not
include herd D. No significant difference in colostrum
IgG concentrations across parity was found (P > 0.05).
Piglet Characteristics
The average BiW was 1.46 kg, ranging from 0.48
to 2.53 kg. At d 1, the weight was, on average, 1.59 kg,
ranging from 0.48 to 2.92 kg. Average weight differ-
ence on d 1 was 0.12 kg, ranging from 0.43 to 1.06 kg.
This average weight gain was influenced by a reduction
in BW from birth to d 1 in 4% of the piglets. When these
piglets were excluded from the descriptive analysis, the
average weight gain was 138 g. Male piglets (n = 456)
weighed significantly more (60 g) at birth than females
(n = 366, P = 0.03), with 1.48 and 1.42 kg, respectively.
 Immunoglobulin G in sows and piglets 4457

Table 2. Mean concentrations of colostrum immunoglobulin G (IgG) in the 4 herds at 3 time points (t1, t2, and
t31) during farrowing and concentrations of IgG in piglet plasma (piglet IgG) at d 12
Herd
A B C D
Item g/L SD g/L SD g/L SD g/L SD
Colostrum IgG at t1 40.5 12.2 53.5 13.9 69.6**2 30.3 64.2** 26.5
Colostrum IgG at t2 37.3 9.3 46.7 5.6 58.5** 18.7 69.93*** 28.9
Colostrum IgG at t3 34.8 7.7 41.1 12.4 54.6*** 14.0 4
Piglet IgG at d 1 18.2 6.4 19.7 5.8 22.2 7.7 26.3 11.4
1Colostrum was sampled from sows immediately after expulsion of the first piglet and before the first suckling (t1), midway through farrowing (just after
the sixth piglet was born; t2), and after the last piglet was born (t3).
2Values marked with * differ from values in herd A (baseline): *P < 0.05; **P < 0.01; ***P < 0.005.
3No piglets suckled between t1 and t2 in herd D.
4In herd D, no samples at t3 were collected due to the protocol in this herd.

Mean concentration of piglet IgG at d 1 was 21.7
g/L, ranging from 0 to 58 g/L. Estimated measure of
colostrum intake, using the formula by Theil et al.
(2014), was 353 g per piglet, ranging from 123 to 595
g, with only 3% of the piglets below 200 g. Estimated
colostrum intake for each piglet was significantly as-
sociated with litter size (P < 0.01), with an estimate
of 11 g lower intake with each additional piglet born.
Mean BMI was 20.1 kg/m2, ranging from 12.7 to 36.5
kg/m2.
When exploring the association between the out-
come (piglet IgG) and the predictor variables, the rela-
tionship between piglet IgG and BMI was significant but
not a straight line (Fig. 2). Therefore, during the model-
building process, we found a cut off at 17 kg/m2. When
the BMI was above 17.0 kg/m2, it was no longer sig-
nificantly associated with piglet IgG. With that as back-
ground, the BMI variable was categorized into 2 groups.
One group was from 12.7 to 17 kg/m2 (n = 71; baseline),
and another was from 17.1 to 36.5 kg/m2 (n = 805). This
categorization significantly improved the model.
Associations between Colostrum
Immunoglobulin G and Plasma Immunoglobulin G
Explorative graphs of the total data set indicate
a clear correlation between colostrum IgG at t1 and
piglet IgG at d 1 (Fig. 3) and a distinct herd difference
(Fig. 4). The lowess curve in Fig. 3 indicates that the
piglet IgG level in plasma flattens out when colostrum
IgG is above 60 g/L.
Multilevel linear regression analysis was per-
formed with piglet IgG (g/L) as outcome and details
are shown in Table 3. Because of strong herd effects,
herd was included as fixed effect. Piglet IgG was as-
sociated with herd, birth order (n), BMI > 17 (kg/m2),
and colostrum IgG at t1 (g/L). Herd D had the highest

Figure 3. Scatter plot between piglet immunoglobulin G (IgG) and

Figure 2. Scatter plot of body mass index by piglet immunoglobu-
lin G concentration in plasma on d 1, with a locally weighted smoothing
(lowess) line (solid) between the 2 variables and a fitted regression line
(broken line).
colostrum IgG concentrations. Colostrum was sampled immediately after
expulsion of the first piglet and before the first suckling (t1) and before first
suckling. Piglet plasma IgG concentrations were sampled at d 1. Solid line
illustrates a locally weighted smoothing (lowess) line and the broken line
is a fitted linear line with 95% confidence interval (CI) bands.
4458
Figure 4. Scatter plot between piglet immunoglobulin G (IgG) and
colostrum IgG concentrations across 4 herds. Colostrum was sampled im-
mediately after expulsion of the first piglet and before the first suckling (t1)
and before first suckling. Piglet plasma IgG concentrations were sampled
at d 1. Solid line illustrates a lowess line and broken line is a fitted linear
line with 95% confidence interval (CI) bands.
predicted mean level of piglet IgG. Piglet plasma IgG
concentrations decreased linearly by 0.4 g/L with each
piglet born and increased by 0.1 g/L with each gram
per liter that the colostrum IgG increased. Piglets with
a BMI above 17 kg/m2 had a 4.5 g/L greater piglet IgG
than those with a BMI at 17 kg/m2 or below.
From this model, one could compare predicted
piglet IgG of 2 average piglets in 2 different herds.
Predicted piglet IgG concentrations of a piglet born as
number 7, with a BMI above 17 kg/m2, which suckled
a sow with colostrum IgG of 56 g/L at t1, and was
born in herd D would have an estimated piglet IgG
of 25.0 g/L. A piglet with identical characteristics in
herd A would have an estimated IgG of 17.3 g/L. The
difference between these 2 piglets is mainly due to the
herd level difference (Table 3).
When including sow as a random effect in the mod-
el, 36% of the unexplained variance of plasma IgG be-
tween piglets resided at sow the level and 64% of the
unexplained variation remains at the piglet level. This
indicates that there is a large similarity between each pig-
let within each litter. When running the model with both
herd and sow as random effects, the model calculates the
level of unexplained variance on each of the 3 hierarchi-
cal levels. The variance at the sow level stays constant
at 34%. Nine percent of the variance at the piglet level
resides at the herd level and 57% at the piglet level.
DISCUSSION
Concentrations of plasma IgG in piglets is said to
be dependent on the amount of available colostrum
(Devillers et al., 2011), amount of colostrum ingestion,
IgG absorption in the piglets gut (Jensen et al., 2001),
Kielland et al.
and gut closure (Rooke and Bland, 2002). According
to Jensen et al. (2001), 22.6% of colostrum IgG is ab-
sorbed by a piglets gut and the average piglet IgG has
been found to be 35 g/L (Svendsen et al., 2005). Present
findings on piglet IgG were lower than those found by
others (Svendsen et al., 2005). However, Svendsen et
al. (2005) used only 6 sows from 1 herd in a split study
design and these sows had a high average mean concen-
tration of colostrum IgG at almost 80 g/L. Our study in-
dicates a strong association between the concentration
of colostrum IgG and piglet IgG and that there are large
variations between both sow and herd. The high level
of colostrum IgG in the study by Svendsen et al. (2005)
may, therefore, explain why they found such large mean
piglet IgG concentration 27 h after birth.
The concentration of piglet IgG decreased with
each piglet born. As litter size was not significantly
associated with piglet IgG, this indicates that the time
of birth relative to farrowing start is essential. Hence,
being born late has a negative effect on the level of
piglet IgG at d 1. This is in agreement with another
study that concluded that the duration of farrowing
is likely to affect the level of passive immunity ac-
quired by piglets born late in the birth order (Bourne,
1969). Our results indicate that although being born
late is not beneficial for acquiring passive immunity,
the level of colostrum IgG counteracts that negative
effect. Other studies have found that increased litter
size is associated with increased neonatal death rate
(Baxter et al., 2013; Rutherford et al., 2013). This is
not only because of the increased litter size itself but
rather because of prolonged farrowing. As the IgG in
colostrum declines rapidly the first 24 h postpartum
(Rooke and Bland, 2002), being born late means less
available colostrum IgG, which may be one of the
causes of increased preweaning mortality.
The positive association between colostrum IgG
and piglet IgG supports the importance of high IgG
levels in colostrum. In a recent review by Theil et al.
(2014) on how to improve the quality and increase the
concentration of colostrum IgG, one of the important
nutrient factors is stated to be supplementation of fatty
acids during late gestation, and another study points out
that supplementation with long-chain n-6 and n-3 PUFA
is especially important (Mitre et al., 2005). Others have
found that vitamin supplementations given to sows
could improve the IgG absorption in piglets (Rooke and
Bland, 2002).
Body mass index was significantly associated with
piglet IgG at d 1, and the degree of association flattened
after the piglet reached a BMI of 17 kg/m2. This indi-
cates that piglets with BMI of 17 kg/m2 and below may
be too weak to acquire sufficient amounts of piglet IgG,
as in the group above 17 kg/m2. Interestingly, this is the
 Immunoglobulin G in sows and piglets 4459

Table 3. Results from a mixed effects linear regression model with the concentration of piglet plasma immuno-
globulin G (IgG) at d 1 as the outcome and multiple predictor variables
95% CI1
Predictor variables No. of piglets SE Lower Upper P-value
Herd
D 184 Baseline
A 176 7.7 2.3 12.1 3.2 0.01
B 170 5.9 1.9 9.6 2.3 0.01
C 114 4.7 2.2 8.9 0.4 0.03
Parity2
1 186 Baseline
2 173 1.1 1.9 2.6 4.7 0.57
3 139 2.8 2.4 1.9 7.4 0.22
48 146 3.6 2.3 0.9 8.2 0.12
Birth order 644 0.4 0.1 0.5 0.3 <0.01
Colostrum IgG at t1,3 g/L 0.1 0.0 0.0 0.1 0.03
BMI4 17 kg/m2 51 Baseline
BMI > 17 kg/m2 593 4.6 1.0 2.7 6.5 <0.01
Intercept5 644 19.1 2.7 13.8 24.4 <0.01
1CI = confidence interval.
258 sows were used in the analysis.
3t1 = immediately after expulsion of the first piglet and before the first suckling.
4BMI = body mass index.
5Random effect on sow level, accounting for dependency between each piglet in the same litter.

case regardless of all the other factors accounted for in
the model, such as herd, parity, birth order, and colos-
trum IgG. One explanation for why a BMI below 17
kg/m2 is negatively associated with piglet IgG may be
that piglets with a BMI below 17 kg/m2 might have ex-
perienced some level of intrauterine growth restriction
(IUGR; Baxter et al., 2013; Hales et al., 2013). Body
mass index may, therefore, give an indication of both
the level of body lipid content and the level of adipose
tissue present in newborns (Baxter et al; 2013). The re-
sults of Hales et al. (2013) indicate that body conforma-
tion is associated with postnatal survival risks and not
BiW alone. The same study reports differences in mor-
phology between piglets that had experienced IUGR
and those that had not. Intrauterine growth restriction
piglets have previously been found to have a lower BMI
than those categorized as normal (Attig et al., 2008).
Intrauterine growth restriction status was not identified
in the present study; however, IUGR is a practical mea-
sure to identify low viability piglets.
Because of the decreased relative body surface
area as BMI increases, piglets with a high BMI have
reduced loss of body heat and body fluids compared
with piglets with lower BMI, both of which are es-
sential during the critical newborn period. In litters
of unequal body sizes in particular, attending to less
vital piglets may be necessary to ensure absorption of
sufficient concentrations of IgG (Muns et al., 2014).
Identification of these piglets at risk, therefore, is criti-
cal to ensure oral supplementation with the aim of in-
creasing neonatal survival (Amdi et al., 2013). Oral
supplementation has been found to deliver compara-
ble IgG in the neonatal piglet (Campbell et al., 2012)
The association between colostrum IgG and piglet
IgG concentration differed among herds. One difference
was that in herd D, each piglet was dried and placed un-
der a heating lamp after birth. This management charac-
teristic in herd D may explain why it had both the high-
est descriptive and the highest predicted mean levels of
piglet IgG. Human support of weak piglets is reported
to increase their ability to acquire immunity (Kirkden
et al., 2013), and when piglets were either placed under
the heat lamp or both dried and placed under the heat
lamp, the mortality at weaning was reduced (Andersen
et al., 2009). Another distinct herd difference was the
time of sow introduction to the farrowing pen. A long
adjustment period is assumed to be the best practice and
may reduce stress. Stress can reduce the level of IgG
in colostrum, as stress has an impact on the immune
system (Couret et al., 2009). Our descriptive statistics
indicate a difference between herds A and B and herds
C and D regarding the colostrum IgG concentrations.
Immunoglobulin G acquisition in piglets is a pas-
sive absorption through the basolateral membrane of the
enterocyte. The main impact of keeping piglets warm is
a reduction of heat loss and improved thermoregulation
ability; therefore, more body energy is available for suck-
ling and competing for a teat. Therefore, keeping piglets
warm might improve absorption of IgG in the intestine,
but its main contribution is by enhancing vitality and
4460 Kielland et al.
capacity to suckle. The statistical model shows that the
majority of the variance resides at the piglet level, mean-
ing that when born, in whichever state, management
and herd factors can account for only a certain propor-
tion of the variation in IgG level in piglets. Therefore,
it is important to explore factors of improvement at the
piglet level, such as breeding for better vitality and an
even piglet size in the litter. It is tempting to speculate
whether prenatal maturity, as approximated by BMI, also
influences the development of the intestine. One study
has found that the intestinal mucosa is very sensitive to
dietary stimuli in the time just after birth (Jensen et al.,
2001). In addition, the motility of the intestines decreases
with low body temperatures (Mallet, 2002). As a conse-
quence, one may assume that keeping piglets warm gives
an overall positive effect on absorption of IgG.
Mean concentration of IgG in sow colostrum is
reported to be 61.8 mg/mL (Halliwell and Gorman,
1989). As in accordance with other studies (Quesnel,
2011; Rolinec et al., 2012), we found that colostrum
IgG decreased during farrowing. This was expected as
most colostrum is produced before farrowing (Theil
et al., 2014). In a review by Rooke and Bland (2002),
they state that the level of colostrum IgG also changed
greatly during the first 24 h after farrowing. Location of
udder sampling (front vs. back teats) is also reported to
effect the concentration of colostrum IgG of the sample
(Inoue et al., 1980). We sampled from 3 locations and
pooled the samples, thereby avoiding this difference.
Parity has been found to have an effect on colos-
trum quality (Devillers et al., 2007). Multiparous sows
usually have increased colostrum yield; primiparous
and old sows usually have a reduced colostrum yield.
Colostrum is found to be of better quality in multip-
arous sows, providing better passive immunity than
that in primiparous sows (Carney-Hinkle et al., 2013).
The parity effect on colostrum IgG was explored in our
study, but no significant difference was found. Studies
have found that the vaccination status of the sow and
immune difference between sows influence the level
of colostrum IgG (Bourne et al., 1975). All 4 study
herds used the same vaccination regime; nevertheless,
we found a large variation in colostrum IgG levels be-
tween sows. One explanation may be the normal vari-
ation of the immunoreaction of each sow (Gudding,
2010). In the sow, maternal glucocorticoid levels in-
fluence lactogenesis (Farmer and Quesnel, 2008). In
this study, no physiological or behavioral measures of
stress in the farms were recorded. We therefore can-
not conclude that farms A and B were operating un-
der greater stress levels than farms C and D. One may,
however, speculate whether factors affecting stress of
the sow explain the differences in colostrum quality
between the farms found in our study. One stress factor
could be the short introduction period to the farrow-
ing pen as previously mentioned. Another stress factor
may be group housing during gestation. However, an-
other study reported no difference between colostrum
IgG in sows housed in different systems (Zhao et al.,
2013).
The study went over a 2.5 y period, including sea-
sonal changes and some variability of stockpersons.
The feeding formulas remained the same throughout. In
a field study involving different sow operations, there
will always be some factors that must be adjusted for
by accounting for the unexplained variance at the herd
level. We feel confident that the aforementioned factors
did not vary between herds to an extent that would alter
our results and conclusions. However, the study must
be interpreted within the limits of a larger field study.
Colostrum yield per sow varies between 0.85 and
5.6 L/per day (Foisnet et al., 2010; Quesnel, 2011). In the
present study, colostrum yield was measured indirectly by
recording and adding the weight gain for each piglet from
birth to d 1 (Couret et al., 2009). One can argue that this
estimate depends on the litter size and the piglets capa-
bility to suckle. Colostrum intake is influenced by piglet
vitality, birth order, and total number of piglets born alive
(Quesnel et al. (2012). However, a significant association
between estimated colostrum intake by each individual
piglet and litter size was found, as piglets in large litters
received less colostrum. This indicates that litter size does
affect intake of colostrum (Devillers et al., 2007) but not
the concentration of piglet IgG in plasma, although this
seems contradictory. Therefore, litter size was included in
the model, but no signs of confounding effect were dis-
covered. One explanation of this contradictory finding
could be that the level of piglet IgG in plasma flattens out
above a certain level of colostrum IgG. Our descriptive
statistics indicate that such a level exists.
Several previous studies on colostrum IgG have
been performed in single herds. In single herd studies,
there are limitations in terms of low external validity.
External validity is largest in epidemiological stud-
ies, when statistical analysis are used, and where the
model accounts for both several predictors and for the
unexplained variation on all hierarchical levels. With
a mixed model, the distribution of the unexplained
variation across the levels can be calculated. The pres-
ent study shows how results regarding external valid-
ity are not necessarily robust, and extrapolation from
the results should be done with care.
Interestingly, even when including herd as a fixed
effect and indirectly accounting for various herd and
management practices, 9% of the variation was on the
herd level. This could mean that there are factors in
a hierarchical level above herd, climate, breed, etc.
that influence the piglet IgG concentrations. This un-
 Immunoglobulin G in sows and piglets
explained variation may be due to genetic difference
of either sow or boar, which is in accordance with the
review written by Farmer and Quesnel (2008). They
found literature referring to breeding line differences
in both colostrum composition and colostrum yield
(Farmer and Quesnel, 2008).
Conclusion
Concentrations of colostrum IgG varied largely
between herds and between sows. The largest varia-
tion of piglet IgG was mainly on the piglet level, sup-
porting the complex nature of IgG production in sows
and uptake in piglets. However, the strong association
between colostrum IgG and piglet IgG shows that in-
creased IgG levels in colostrum will improve the lev-
els of IgG in piglets and potentially increase the sur-
vival of piglets as well.
LITERATURE CITED
Amdi, C., U. Krogh, C. Flummer, N. Oksbjerg, C. F. Hansen, and P.
K. Theil. 2013. Intrauterine growth restricted piglets defined
by their head shape ingest insufficient amounts of colostrum.
J. Anim. Sci. 91:56055613. doi:10.2527/jas.2013-6824.
Andersen, I. L., I. A. Haukvik, and K. E. Be. 2009. Drying and
warming immediately after birth may reduce piglet mortal-
ity in loose-housed sows. Animal 3:592597. doi:10.1017/
S1751731108003650.
Attig, L., J. Djiane, A. Gertler, O. Rampin, T. Larcher, S. Boukthir,
P. M. Anton, J. Y. Madec, I. Gourdou, and L. Abdennebi-Najar.
2008. Study of hypothalamic leptin receptor expression in
low-birth-weight piglets and effects of leptin supplementa-
tion on neonatal growth and development. Am. J. Physiol.
295:E1117E1125. doi:10.1152/ajpendo.90542.2008.
Baxter, E. M., K. M. D. Rutherford, R. B. DEath, G. Arnott, S. P.
Turner, P. Sande, V. A. Moustsen, F. Thorup, S. A. Edwards,
and A. B. Lawrence. 2013. The welfare implications of large
litter size in the domestic pig II: Management factors. Anim.
Welf. 22:219238. doi:10.7120/09627286.22.2.219.
Bourne, F. J. 1969. Antibody transfer and colostral whey proteins
in the pig. Vet. Rec. 84:607608. doi:10.1136/vr.84.24.607
Bourne, F. J., T. J. Newby, and J. W. Chidlow. 1975. The influence
of route of vaccination on the systemic and local immune
response in the pig. Res. Vet. Sci. 18:244248.
Bovey, K. E., T. M. Widowski, C. E. Dewey, N. Devillers, C.
Farmer, M. Lessard, and S. Torrey. 2014. The effect of birth
weight and age at tail docking and ear notching on the be-
havioral and physiological responses of piglets. J. Anim. Sci.
92:17181727. doi:10.2527/jas.2013-7063.
Campbell, J., S. Jacobi, Y. Liu, K. Hard Robertson, J. Drayton, I.
Medina, J. Polo, J. Crenshaw, and J. Odle. 2012. Evaluation
of immunoglobulin G absorption from colostrum supple-
ments gavaged to newborn piglets. J. Anim. Sci. 90:299301.
doi:10.2527/jas.51544.
Carney-Hinkle, E. E., H. Tran, J. W. Bundy, R. Moreno, P. S. Miller,
and T. E. Burkey. 2013. Effect of dam parity on litter performance,
transfer of passive immunity, and progeny microbial ecology. J.
Anim. Sci. 91:28852893. doi:10.2527/jas.2011-4874.
4461
Couret, D., A. Jamin, G. Kuntz-Simon, A. Prunier, and E. Merlot. 2009.
Maternal stress during late gestation has moderate but long-last-
ing effects on the immune system of the piglets. Vet. Immunol.
Immunopathol. 131:1724. doi:10.1016/j.vetimm.2009.03.003.
Devillers, N., C. Farmer, J. Le Dividich, and A. Prunier. 2007.
Variability of colostrum yield and colostrum intake in pigs.
Animal 1:10331041. doi:10.1017/S175173110700016X.
Devillers, N., J. Le Dividich, and A. Prunier. 2011. Influence of
colostrum intake on piglet survival and immunity. Animal
5:16051612. doi:10.1017/S175173111100067X.
Dohoo, I., W. S. H. Martin, and H. Stryn. 2009. Veterinary epidemi-
ologic research. 2nd ed. VER Inc., Charlottetown, PE, Canada.
Eastwood, L., P. Leterme, and A. D. Beaulieu. 2014. Changing the
omega-6 to omega-3 fatty acid ratio in sow diets alters se-
rum, colostrum, and milk fatty acid profiles, but has minimal
impact on reproductive performance. J. Anim. Sci. 92:5567
5582. doi:10.2527/jas.2014-7836.
Farmer, C., and H. Quesnel. 2008. Nutritional, hormonal, and
environmental effects on colostrum in sows. J. Anim. Sci.
87(Suppl. 13):5664. doi:10.2527/jas.2008-1203.
Foisnet, A., C. Farmer, C. David, and H. Quesnel. 2010.
Relationships between colostrum production by primiparous
sows and sow physiology around parturition. J. Anim. Sci.
88:16721683. doi:10.2527/jas.2009-2562.
Gudding, R. 2010. Vaksinasjon av dyr (Vaccination of animals).
Scandinavian Veterinary Press, Oslo, Norway.
Hales, J., V. A. Moustsen, M. B. F. Nielsen, and C. F. Hansen.
2013. Individual physical characteristics of neonatal piglets
affect preweaning survival of piglets born in a noncrated sys-
tem. J. Anim. Sci. 91:49915003. doi:10.2527/jas.2012-5740.
Halliwell, R. E. W., and N. T. Gorman. 1989. Veterinary clinical
immunology. Saunders, Philadelphia, PA.
Herpin, P., J. Le Dividich, D. Berthon, and J. C. Hulin. 1994.
Assessment of thermoregulatory and postprandial thermo-
genesis over the first 24 hours after birth in pigs. Exp. Physiol.
79:10111019. doi:10.1113/expphysiol.1994.sp003815.
Inoue, T., K. Kitano, and K. Inoue. 1980. Possible factors influenc-
ing the immunoglobulin G concentration in swine colostrum.
Am. J. Vet. Res. 41:11341136.
Jensen, A. R., J. Elnif, D. G. Burrin, and P. T. Sangild. 2001. Development
of intestinal immunoglobulin absorption and enzyme activities in
neonatal pigs is diet dependent. J. Nutr. 131:32593265.
Kielland, C., K. E. Be, A. J. Zanella, and O. sters. 2010. Risk
factors for skin lesions on the necks of Norwegian dairy cows.
J. Dairy Sci. 93:39793989. doi:10.3168/jds.2009-2909.
Kirkden, R. D., D. M. Broom, and I. L. Andersen. 2013. Piglet
mortality: Management solutions. J. Anim. Sci. 91:3361
3389. doi:10.2527/jas.2012-5637.
Klobasa, F., and J. E. Butler. 1987. Absolute and relative concen-
trations of immunoglobulins G, M, and A, and albumin in the
lacteal secretion of sows of different lactation numbers. Am.
J. Vet. Res. 48:176182.
Le Dividich, J., and J. Noblet. 1984. Effect of colostrum intake on
metabolic rate and plasma glucose in the neonatal pig in rela-
tion to environmental temperature. Biol. Neonate 46:98104.
doi:10.1159/000242039.
Mallet, M. L. 2002. Pathophysiology of accidental hypothermia.
Q. J. Med. 95:775785. doi:10.1093/qjmed/95.12.775
Mitre, R., M. Etienne, S. Martinais, H. Salmon, P. Allaume, P.
Legrand, and A. B. Legrand. 2005. Humoral defence improve-
ment and haematopoiesis stimulation in sows and offspring by
oral supply of shark-liver oil to mothers during gestation and
lactation. Br. J. Nutr. 94:753762. doi:10.1079/BJN20051569.
4462 Kielland et al.
Muns, R., C. Silva, X. Manteca, and J. Gasa. 2014. Effect of cross-
fostering and oral supplementation with colostrums on per-
formance of newborn piglets. J. Anim. Sci. 92:11931199.
doi:10.2527/jas.2013-6858.
Quesnel, H. 2011. Colostrum production by sows: Variability
of colostrum yield and immunoglobulin G concentrations.
Animal 5:15461553. doi:S175173111100070X.
Quesnel, H., C. Farmer, and N. Devillers. 2012. Colostrum intake:
Influence on piglet performance and factors of variation.
Livest. Sci. 146:105114. doi:10.1016/j.livsci.2012.03.010.
Rolinec, M., D. Bro, P. tastny, B. Glik, M. imko, and M.
Jurek. 2012. Immunoglobulins in colostrum of sows with
porcine reproductive and respiratory syndrome PRRS. J.
Cent. Eur. Agric. 13:303311.
Rooke, J. A., and I. M. Bland. 2002. The acquisition of passive
immunity in the new-born piglet. Livest. Prod. Sci. 78:1323.
doi:10.1016/S0301-6226(02)00182-3.
Rootwelt, V., O. Reksen, W. Farstad, and T. Framstad. 2012a.
Associations between intrapartum death and piglet, placental,
and umbilical characteristics. J. Anim. Sci. 90:42894296.
doi:10.2527/jas.2012-5238.
Rootwelt, V., O. Reksen, W. Farstad, and T. Framstad. 2012b.
Blood variables and body weight gain on the first day of life
in crossbred pigs and importance for survival. J. Anim. Sci.
90:11341141. doi:10.2527/jas.2011-4435.
Rootwelt, V., O. Reksen, W. Farstad, and T. Framstad. 2013.
Postpartum deaths: Piglet, placental, and umbilical character-
istics. J. Anim. Sci. 91:26472656. doi:10.2527/jas.2012-5531.
Rutherford, K. M. D., E. M. Baxter, R. B. DEath, S. P. Turner,
G. Arnott, R. Roehe, B. Ask, P. Sande, V. A. Moustsen,
F. Thorup, S. A. Edwards, P. Berg, and A. B. Lawrence.
2013. The welfare implications of large litter size in the do-
mestic pig I: Biological factors. Anim. Welf. 22:199218.
doi:10.7120/09627286.22.2.199.
Sangild, P. T. 2003. Uptake of colostral immunoglobulins by
the compromised newborn farm animal. Acta Vet. Scand.
44(Suppl. 1):S105S122. doi:10.1186/1751-0147-44-S1-S105.
Sjaastad, O. V., O. Sand, and K. Hove. 2012. Lactation. In:
Physiology of domestic animals. 2nd ed. Scandinavian
Veterinary Press, Oslo, Norway. p. 743757.
Svendsen, J., B. R. Westrm, and A. C. Olsson. 2005. Intestinal
macromolecular transmission in newborn pigs: Implications
for management of neonatal pig survival and health. Livest.
Prod. Sci. 97:183191. doi:10.1016/j.livprodsci.2005.04.001.
Theil, P. K., C. Lauridsen, and H. Quesnel. 2014. Neonatal pig-
let survival: Impact of sow nutrition around parturition on
fetal glycogen deposition and production and composi-
tion of colostrum and transient milk. Animal 8:10211030.
doi:10.1017/S1751731114000950.
Tuchscherer, M., E. Kanitz, W. Otten, and A. Tuchscherer. 2002.
Effects of prenatal stress on cellular and humoral immune
responses in neonatal pigs. Vet. Immunol. Immunopathol.
86:195203. doi:10.1016/S0165-2427(02)00035-1.
Zhao, Y., W. L. Flowers, A. Saravia, K.-J. Yeum, and S. W. Kim.
2013. Effects of social rank and gestation housing sys-
tems on oxidative stress status, reproductive performance,
and immune status of sows. J. Anim. Sci. 91:58485858.
doi:10.2527/jas.2013-6388.