| HOME | HELP | FEEDBACK | SUBSCRIPTIONS | ARCHIVE | SEARCH | TABLE OF CONTENTS |
|
|
|
|||||||
|
|||||||||
ANIMAL GROWTH, PHYSIOLOGY, AND REPRODUCTION |
* Animal Breeding and Genetics Group, Wageningen University, 6700 AH Wageningen, The Netherlands;
and
Institute for Pig Genetics, 6640 AA Beuningen, The Netherlands; and
and
Division Animal Resources, Animal Sciences Group, Wageningen University and Research Centre, 8200 AB Lelystad, The Netherlands
 |
Abstract |
|---|
 Top Abstract IntroductionMaterials and MethodsResultsDiscussionImplicationsLiterature Cited |
|---|
Key Words: Birth Weight • Piglet Vitality • Placental Efficiency • Placental Weight • Preweaning Mortality
|
Introduction |
|---|
|
TopAbstractIntroductionMaterials and MethodsResultsDiscussionImplicationsLiterature Cited |
|---|
). Important risk factors for preweaning mortality are birth weight (BWB) and withinlitter variation in BWB (English and Smith, 1975; Roehe and Kalm, 2000; Tuchscherer et al., 2000).
Birth weight is highly dependent on placental nutrient supply. The latter is determined, to a large extent, by placental size (mass, surface area) and blood flow. A measure for the ability of the placenta to sustain fetal growth is placental efficiency (PLEFF; i.e., the ratio between BW and placental weight (PLW; Molteni et al., 1978; Biensen et al., 1999a; Kurtz et al., 1999). Because fetal mortality in pigs in which the number of embryos is not limiting is due to placental insufficiency, selection for smaller and more efficient placentas has been suggested as a means to decrease fetal mortality, and thereby increase litter size (Wilson et al., 1998; Ford et al., 2002). This increase in litter size is economically only of interest if it increases the number of pigs weaned. Until now, the effect of PLEFF on subsequent neonatal pig vitality has not been studied thoroughly.
The primary objective of this study was to examine the relationship between PLEFF and neonatal pig vitality as measured by the probability of preweaning death of live-born piglets. Furthermore, the data collected were used to establish whether PLEFF is a better predictor of the risk of preweaning mortality than BWB and PLW as such. Until now, most studies concerning relationships between PLEFF, PLW, and BWB in pigs have been done with gilts. Because the present study was performed with gilts and sows, the data were first analyzed to establish whether these relationships are affected by parity. The outcome of these analyses is also reported.
|
Materials and Methods |
|---|
|
TopAbstractIntroductionMaterials and MethodsResultsDiscussionImplicationsLiterature Cited |
|---|
, with slight modifications as described by Van Rens et al. (2002). Briefly, each expelled piglet was caught, and the umbilical cord was ligated with surgical silk containing a specific code. The umbilical cord was then cut between the piglet and tag, allowing the placental end of the cord with its tag to retract into the vagina. Subsequently, the piglet was earmarked with a number corresponding to the specific code on the tag. When all placentas of a litter were expelled, the piglets and placentas were weighed. Subsequently, the piglets’ front leg knees were taped to prevent grazing during suckling. Where unavoidable, piglets were crossfostered, preferably within 36 h of birth. Piglet mortality until weaning at 4 wk after birth was registered. Placental efficiency was calculated as the ratio between BWB (g) and PLW (g). The 118 litters in this study were produced by 104 pigs. Thus, 14 pigs produced two (n = 12) or three (n = 2) litters. In total, there were 31 first-parity litters and 87 litters of a higher parity. For the statistical analyses, each litter received a unique identification number, referred to as LIT.
Statistics
All data were analyzed with SAS (SAS Inst., Inc., Cary, NC), using procedures CORR, GLM, MIXED, and GENMOD. For drawing three-dimensional graphs, procedure G3D was combined with the ANNOTATE facility.
The Effect of Parity on the Relationships Among Traits.
The effects of parity (i.e., gilt or sow) on the relationships 1) BWB to PLW, 2) PLEFF to BWB, and 3) PLEFF to PLW were analyzed using a mixed model, which initially included the random variable LIT and the fixed variables PLW (Relationships 1 and 3) or BWB (relationship 2), TNB (total number of piglets born), parity ("gilt" or "sow"), interaction of parity and TNB, and interaction of parity and either PLW or BWB. The relationships were assumed to be polynomial. To determine the order of the polynomial model, a stepwise forward procedure was used. In every step, a subsequent order of PLW or BWB and its interaction with parity was included in the model, but only if the last included order or its interaction with parity was significant. The forward procedure was continued until the highest included order and its interaction were not significant, or until the fourth order was reached. At each next step, all lower orders and their interactions were included in the model. Once the highest order was reached, interactions that were not significant were eliminated stepwise, removing in each step the nonsignificant interaction of the highest order. If the interaction of TNB and parity was nonsignificant, it was eliminated from the model as well. To draw the figures, the intercepts of the equations describing the relationships were calculated by combined use of the population means per parity for the variables in the equation and the estimated regression coefficients.
The Effects of Birth Weight, Placental Weight, and Placental Efficiency on Piglet Vitality.
Piglet preweaning death was defined as a binary trait, with a score of 0 for piglets alive at weaning and a score of 1 for live born piglets that died before weaning. Because of the binomial distribution of preweaning death (PWD), analysis was done by logistic regression using the GENMOD procedure. In all models, LIT was included as a random variable (in the procedure defined as repeated subject, type = exch), and the number of piglets born alive (NBA) was included as a continuous, linear variable. Furthermore, one of the continuous linear variables, BWB, PLW, or PLEFF, or a combination of two of these variables, was included into the model. Thus, six different relationships were analyzed. To avoid having the resulting estimates of the regression coefficients become too small, BWB, as well as PLW, was divided by 100 before inclusion in the model. Odds ratios (OR) were calculated from the regression coefficient (by raising e to the power of the regression coefficient). Regression coefficients are referred to as "ßTRAIT" in the present paper, in which "TRAIT" stands for the name of the trait to which the regression coefficient refers.
To draw the graphs, the estimates of intercept and regression coefficients, the population average of NBA, and a range of values of two of the variables, BWB, PLW, and PLEFF, were used, unless mentioned differently. For example, the relationship of the probability of dying before weaning (PPWD) with PLW and BWB was drawn using the following formula: PPWD = ea/(1 + ea), in which a = intercept + ßNBA x NBA + ß PLW/100 x PLW/100 + ßBWB/100 x BWB/100 ("intercept" and regression coefficients "ß" are estimates from the model).
|
Results |
|---|
|
TopAbstractIntroductionMaterials and MethodsResultsDiscussionImplicationsLiterature Cited |
|---|
. For gilts, the litter minimum for PLEFF varied from 2.50 to 5.29, and the litter maximum for PLEFF varied from 5.17 to 8.90, whereas in sows, the litter minimum of PLEFF varied from 1.78 to 6.12, and the litter maximum of PLEFF varied from 4.66 to 10.25.
|
illustrates the ratio of BWB to PLW (i.e., PLEFF) for a range of biologically possible birth weights and placental weights. In the resulting plane, the observed PLEFF values are shown for the 1,036 live born piglets that were successfully labeled.
|
. The relationships among BWB, PLW, and PLEFF (raw data) are presented in Figure 2 (panels a through c).
|
|
), in which ßgilt = 8.50 and ßsow = 9.09 were different (P <0.01). The P-values of TNB, parity, PLW x parity, PLW, PLW2, and PLW3 were <0.001, 0.052, 0.004, <0.001, <0.001, and 0.005, respectively.
The relationship of PLEFF to BWB was best described by a second-order equation (i.e., PLEFF = intercept + ßparity x TNB + 0.002286 x BWB + 6.90 x 10–7 x BWB2; Figure 2e), in which there was no difference between gilts and sows. The P-values of TNB, TNB x parity, parity, BWB, and BWB2 were 0.60, 0.021, 0.030, <0.001, and <0.001, respectively. The significantly different relationship of PLEFF to TNB between parities resulted in regression coefficients ßgilt = 0.075 and ßsow = –0.049.
The relationship of PLEFF to PLW was best described by a linear equation (i.e., PLEFF = intercept –0.06345 x TNB + ßparity x PLW; Figure 2f), in which ßgilt = –0.00875 and ßsow = –0.00681 were different (P <0.05). The P-values of TNB, parity, PLW x parity, and PLW were <0.001, 0.070, 0.015, and <0.001, respectively.
The Effects of Birth Weight, Placental Weight or Placental Efficiency on Piglet Vitality
Results related to piglet vitality are presented in Table 3 and Figure 3. For all models used, NBA had a significant positive effect on the chance of dying before weaning. In other words, even after correction for, and thus independent of, their BWB, PLW, or PLEFF, piglets that were born alive in a larger litter had a higher chance of dying before weaning than piglets that were born alive in a smaller litter. The odds of a live-born piglet dying before weaning increased with a factor 1.15 to 1.20 for each additional live-born piglet (Table 3, Models 1 through 6).
|
|
, Model 1; Figure 3a). This increase in the chance of dying before weaning with a decrease in birth weight tended (P = 0.05) to become steeper with an increase in placental weight (Table 3, Model 4; Figure 3b), but was not significantly affected by placental efficiency (Table 3, Model 5).
Within a litter, independent on NBA, piglets with a light placenta had a greater (P = 0.0128) chance of dying than did litter mates with a heavier placenta (Table 3, Model 2). This relationship changed, however, after including BWB into the model. Piglets with a light placenta tended (P = 0.05) to have a lower chance of dying than did litter mates with a similar birth weight but with a heavier placenta. The steepness of this relationship was highly dependent on the BWB of the piglets (Table 3, Model 4; Figure 3b). Including PLEFF instead of BWB in the model resulted again in a significant negative association of preweaning death with PLW. Piglets with a light placenta had a greater (P <0.001) chance of dying than litter mates with a similar PLEFF but with a heavier placenta. This relationship, which was steepest for piglets with a low PLEFF, became less (P <0.001) steep as PLEFF increased (Table 3, Model 6; Figure 3c).
The significant negative association of preweaning death to PLEFF (Table 3, Model 6) became a tendency when PLW was excluded from the model. Hence, piglets with a low PLEFF tended (P = 0.08) to have a higher chance of dying than littermates with a high PLEFF (Table 3, Model 3).
|
Discussion |
|---|
|
TopAbstractIntroductionMaterials and MethodsResultsDiscussionImplicationsLiterature Cited |
|---|
The relationship of BWB to PLW confirms the relationship described in gilts at d 35 (Van Rens and Van der Lende, 2000), d 105 (Vallet, 2000; second-order relationship, but with a decrease in fetal weight for the largest placentas), and d 110 to 112 of gestation (Leenhouwers et al., 2002), and at term (Van Rens and Van der Lende, 2002; Mesa et al., 2003). Apparently, the dependency of fetal weight and birth weight on placental size decreases as placental size increases. The restriction of fetal growth when the placenta is relatively large can be due to a limited fetal growth capacity or to a restriction of placental functioning in large placentas (Vallet, 2000; Van Rens and Van der Lende, 2000, 2002; Vallet et al., 2002). A comparable nonlinear relationship between birth weight and placental weight as described here for the pig was found for the offspring of a mouse interspecific back cross (Kurtz et al., 1999). In contrast, however, in offspring of a mouse intraspecific cross, fetal weight was linearly related to placental weight (Kurtz et al., 1999). The difference in relationship appeared to be due to an additional population of fetuses amongst the hybrid fetuses that had a much larger placenta (leading to more variation in placental and fetal weights) compared with placental size of the purebred fetuses.
Because of the nonlinearity of the relationship of BWB to PLW, PLEFF (i.e., the ratio between these two traits) is a complicated trait. As a consequence, placental efficiency at relatively high placental weights does not give the same information as placental efficiency at relatively low placental weights.
In contrast to the significant curvilinear relationship of PLEFF to BWB (after correction for TNB and parity) as described in the present dataset, Vonnahme et al. (2002) found no significant relationship of placental efficiency with weight of conceptuses at d 25, 36, and 44. The symmetry of the curvilinear relationship in the present study caused a nonsignificant linear correlation between PLEFF and BWB, which agrees with the lack of linear correlation previously described for d-110 fetuses (Biensen et al., 1999b) and for piglets at term (Biensen et al., 1999a).
For the curvilinear relationship of PLEFF to BWB (shown in Figure 2e), estimated PLEFF for the range of observed BWB varied between 4.59 and 5.54 for gilts and between 4.23 and 5.39 for sows, respectively. However, for the linear relationship of PLEFF to PLW (shown in Figure 2f), estimated PLEFF for the observed range of PLW varied between 2.93 and 7.04 for gilts and between 2.24 and 6.80 for sows, respectively. Therefore, a change in PLW appeared to have a much larger effect on the value of PLEFF than a change in BWB, which is due to the larger variation in PLW than in BWB.
A significant linear decrease of PLEFF with an increase of PLW (after correction for TNB and parity) was found. A comparable, but curvilinear instead of linear decrease of PLEFF with an increase of PLW was reported for d-25, -36, and -44 fetuses (Vonnahme et al., 2002), d-105 fetuses of unilaterally hysterectomized-ovariectomized gilts (Vallet, 2000), and for human babies at term (Molteni et al., 1978). Furthermore, comparable negative linear correlations between PLEFF and PLW were described for d-110 fetuses (Biensen et al., 1999b) and piglets at term (Biensen et al., 1999b; Mesa et al., 2003).
The role of PLEFF on piglet vitality was, until now, hardly studied. In contrast to the findings of Biensen et al. (1999a), both underlying components of PLEFF (i.e., BWB and PLW) were significantly negatively associated to the chance of dying before weaning (Models 1 and 2). The PLEFF itself merely tended to be negatively associated to the chance of dying before weaning (Model 3). From the results of Models 5 and 6, it can be concluded that BWB plays a greater role in the effect of PLEFF on the chance of preweaning death than PLW. Piglets with the same BWB but with different PLEFF (which thus is caused by a difference in PLW) had a comparable chance of dying before weaning (Model 5). In contrast, among the piglets with the same PLW but with different PLEFF (and thus different BWB), the piglets with a higher PLEFF (and thus higher BWB) had a significantly lower chance of dying than piglets with the lower PLEFF (and thus lower BWB; Model 6). Following a similar line of reasoning, the same conclusion (i.e., BWB has a greater role in the effect of PLEFF on preweaning death than PLW) can be drawn from the results of Models 3 and 4.
Among piglets with the same PLEFF, those with a high BWB or PLW had a significantly lower chance of dying before weaning than piglets with a low BWB (Model 5) or a low PLW (Model 6).
In conclusion, the effect of PLEFF on preweaning mortality was highly dependent on the value of its underlying traits (i.e., BWB and PLW). Of these traits, BWB played a much greater role in the effect of PLEFF on the chance of preweaning death than PLW. Because a difference in PLEFF can be a result of a difference in BWB, PLW, or both, and because animals with similar PLEFF do not have to have similar BWB and PLW, it is very difficult to use PLEFF as a reliable predictor for the chance of dying before weaning. The extent to which BWB depends on PLW is dependent on the value of PLW (see discussion above). Placental functioning and fetal growth capacity, however, also have effects on the value of BWB. Therefore, of the three traits (BWB, PLW, and PLEFF), BWB remains the best predictor for the chance of preweaning mortality, and it also is the easiest to measure.
|
Implications |
|---|
|
TopAbstractIntroductionMaterials and MethodsResultsDiscussionImplicationsLiterature Cited |
|---|
|
Footnotes |
|---|
2 Correspondence: P.O. Box 338, 6700 AH Wageningen (phone: +31-317-482335; fax: +31-317-483929; e-mail: birgitte.vanrens{at}wur.nl).
Received for publication July 1, 2004. Accepted for publication September 22, 2004.
|
Literature Cited |
|---|
|
TopAbstractIntroductionMaterials and MethodsResultsDiscussionImplicationsLiterature Cited |
|---|
This article has been cited by other articles:
|
|
 |
|
  G. Wu, F. W. Bazer, J. M. Wallace, and T. E. Spencer BOARD-INVITED REVIEW: Intrauterine growth retardation: Implications for the animal sciences J Anim Sci, September 1, 2006; 84(9): 2316 - 2337. [Abstract] [Full Text] [PDF]
|
|
|
|
|
|
H. Mesa, T. J. Safranski, K. M. Cammack, R. L. Weaber, and W. R. Lamberson Genetic and phenotypic relationships of farrowing and weaning survival to birth and placental weights in pigs J Anim Sci, January 1, 2006; 84(1): 32 - 40. [Abstract] [Full Text] [PDF]
|
|
| ||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||
| HOME | HELP | FEEDBACK | SUBSCRIPTIONS | ARCHIVE | SEARCH | TABLE OF CONTENTS |
in Panels a to c; – – – in Panels d to f). Raw data (Panels a, b, and c) and regression curves calculated from the model as described in the text (Panels d, e, and f). The relationship of BWB to PLW (Panel d) was best described by a third-order equation, in which there was a parity effect (P <0.001). The relationship of PLEFF to BWB (Panel e) was best described by a second-order equation. The parity effect was not significant. The relationship of PLEFF to PLW (Panel f) was best described by a linear equation in which there was a parity effect (P <0.05). All equations are given in the results section of the text.