Introduction. Variation in live weight in growing pigs causes a number of problems for producers and results in significant economic loss. However, it is important to bear in mind that biological variation exists for a reason. It is the foundation of natural and artificial selection and as such is the basis for genetic improvement in performance traits. In addition, even if we could minimize all of the genetic and environmental causes of such variation there would still be a wide range of growth rates and, therefore, live weights in our commercial pig populations. This paper sets out to describe typical variation observed in live weight under commercial conditions, to detail the major causes of such variation, and to discuss approaches to minimizing and(or) managing variation programs. Much of the information presented is from an extensive program of research that has been carried out on commercial facilities in Illinois in collaboration with The Maschhoffs, a large-scale swine production company.
Measurement of Variation in Live Weight. Measuring variation in live weight in a population of pigs is conceptually very simple but problematic because it requires that large numbers of pigs are weighed periodically. Performance traits, such as growth rate and live weight, are generally normally distributed. Knowledge of the mean and standard deviation (SD) of live weight effectively defines any population. We can estimate the weight of all pigs in the population based on these parameters given that for a normally-distributed population we expect that 68.3%, 95.5%, and 99.7% of the values for the population to be within ±1, ±2, or ±3 SD of the mean, respectively. A useful statistic to use when describing variation is the Coefficient of Variation (CV) which is the SD expressed as a percentage of the mean. Changes in the variation in weight within a population of pigs appear to be relatively predictable (Shull, 2013). In general, we would expect the CV of live weight to decrease by ~0.7 percentage units for each 10 kg increase in weight.
Problems Resulting from Weight Variation. Variation in live weight is evident at birth and problems due to this variation start at this stage. Light birth weight piglets are disadvantaged compared to their heavier contemporaries and are particularly at risk in the early post-partum period. They are more susceptible to chilling and less able to compete for access to teats with consequent reduction in colostrum and nutrient intake. These factors increase the risk of preweaning mortality from crushing, starvation, and/or disease. Problems resulting from low birth weight pigs are likely to have increased over recent years because of the considerable genetic improvement of litter size and the consequent reduction in average piglet birth weights. Quiniou et al. (2002) showed that increases in numbers born from ≤11 to ≥16 piglets were associated with a reduction in average piglet birth weight from 1.59 to 1.36 kg (mean decrease of 35 g for each additional piglet born). This corresponded to an increase in the percentage of piglets weighing <1 kg from 7 to 23 %; less than 50 % of piglets weighing <1 kg survived to weaning compared to more than 90 % for those weighing >1 kg.
Variation in weight observed in commercial populations of pigs at the end of the finishing period causes problems for producers in meeting the weight targets of the slaughter plants.
Reducing and/or Managing Variation in Live Weight. It is important to distinguish between approaches that actually will reduce the variation in live weight within a population compared to management approaches that attempt to minimize the impact of whatever level of variation exists. In theory, reducing variation is the preferred approach; however, even if we do everything possible to minimize variation, the range in weights within a population at the end of finishing will still be substantial and we will need to rely on approaches to manage that variation. Minimizing variation in live weight requires that the variation in growth rate within the population to be reduced. An important consideration is which pigs in the population to target. Increasing the growth rate of all of the pigs in a population will not reduce variation but it will result in more of the lighter weight pigs reaching the minimum weight required by the market before the building needs to be emptied. It makes no sense to reduce the growth rate of the fastest growing pigs as this will only increase both the time to slaughter for the population and also production costs. Targeting increasing the growth rate of the lightest pigs in the group would increase the average growth rate of the population as well as reduce variation in weight. A key question is what proportion of the population should be targeted? Schinckel et al. (2004) suggested that the lightest 20% of the population at weaning were responsible for the majority of the subsequent variation in weight. Peterson (2008) compared the growth performance from birth to slaughter of pigs from the light, moderate, and heavy thirds of the population based on birth weight and showed that the lightest group grew slower on average than the other 2 weight groups that had similar growth rate.
Approaches to reducing variation have not been widely researched and there is a dearth of controlled studies on which to base recommendations for minimizing variation. In addition, many of the studies that have attempted to reduce variation have significant limitations. Some have not been large enough to accurately estimate treatment differences in weight variation and others have measured variation in the early growth period only and have not followed the pigs through to the end of finishing. The approaches to reducing variation in weight discussed below require validation, ideally in large-scale commercial studies, before they could be recommended. In addition, several of these potential approaches may be difficult to apply and/or not cost effective
Genetic Factors: Associated with the substantial genetic improvements in litter size that has occurred over recent years has been a correlated reduction in average piglet birth weight but also an increase in the within-litter variation in birth weight (reviewed by Campos et al., 2012). Genetic improvement in litter size is continuing and variation in birth weight could actually increase. Published heritability estimates for growth rate are generally within the range of 0.2 to 0.3 which suggests that a proportion of the observed variation in weight within a population is due to the sires and dams that are used to produce the population. Minimizing the number of sires used should reduce genetic variation in growth rates within a population. However, the impact of such approaches on CV at slaughter weight is unlikely to be great (Tokach, 2004). Interestingly, Archer et al. (2003) reported that cloned and non-cloned pigs had similar variation in live weight.
Sex Effects: Differences between barrows and gilts for growth performance have been well documented. Rearing the sexes in separate facilities, an approach that has often been used to allow sex-specific nutritional programs to be used, should reduce the variation in weight within each sex compared to growing them in the same facility.
Birth Weight: There is no doubt that light birth weight pigs will grow slower on average than heavy birth weight litter mates. However, the relationship between birth weight and subsequent growth rate to harvest weight is not linear. Peterson (2008) showed that increases in birth weights up to 1.5 kg were accompanied by increasing growth rates; above that weight there was no effect of birth weight on growth rates. Variation in birth weight is responsible for a significant proportion of the variation observed in subsequent growth, however, there is little that producers can do at the present time to reduce this variation.
Weaning Age: Variation in weaning age can be a major contributor to variation in live weight at the end of finishing. Main et al. (2004) compared weaning ages of 12, 15, 18, and 21 days of age; the earlier weaning ages were associated with lower weights and increased CV for weight at all times from weaning to the end of finishing. Weaning at 21 days of age and minimizing the variation in age within a population of pigs could potentially have a large impact on variation in weight in a population at the end of finishing.
Parity of Dam: Quesnel et al. (2008) found that the CV of birth weight increased in higher parity sows, being around 20 % for parity 1 and 2 females and between 22 and 25 % for parity 3 to >7 sows. Some production systems segregate gilts and their piglets in separate facilities than used for higher parity sows; however, further segregation of piglets on the basis of parity of dam in an attempt to minimize variation in weight at birth and subsequently could be practically difficult and is likely to increase production costs.
Sorting by Weight: Sorting pigs by weight into more uniform weight groups early in the growth period has been advocated in the belief that the reduced variation will be maintained through to slaughter weight. However, a number of studies (e.g., Wolter et al., 2002) have shown that this approach reduces variation at the time of sorting but not at the end of finishing. There is no benefit in sorting pigs by weight to reduce variation unless some additional treatment is going to be applied to the sorted groups. There is evidence that pens that have pigs with a wide weight range can actually be emptied earlier than those with a narrower weight range. For example, in studies cited by Gonyou (1998), pigs were sorted in groups with either low or high (normal) variation in weight at ~25 kg live weight. Overall population weight gain was not affected by weight variation within the pen; however, the pens with high weight variation were emptied 3 to 5 days before those with low weight variation. Maintaining weight variation in pens may be an approach to managing weight variation at the end of finishing.
Split-suckling: Donovan and Dritz (2000) showed that split-suckling of litters, the practice of allowing the smallest pigs in the litter to suckle the sow for a short period of time soon after birth without competition from the bigger piglets, can reduce within-litter variation in weaning weight, but only for relatively small litters (≤9 piglets born alive). However, the impact on variation in weight at the end of finishing needs to be established.
Health Status: Several studies have suggested that while the presence of disease can increase variation in live weight within a population, the use of strategies to control disease, such as multiple-site production and All In/All Out management, can reduce variation (Schinckel et al., 2002). In addition, prompt action to treat any disease condition will minimize its impact on the health and growth of the pigs and is likely to minimize any impact of the disease on variation.
Effective Environment: Ensuring that all pigs experience the same or a very similar environment will reduce variation in growth rates and weight for age within populations. Standardizing the number of pigs within a facility as well as providing a consistent rearing environment are important considerations. In addition, restricting access to feed or water can reduce the growth rate of pigs and is also likely to increase variation within the population. Ensuring that there are sufficient drinkers and feeder spaces for all of the pigs in the pen will help minimize variation.
Floor Space: On commercial units, pigs are kept at floor spaces that reduce average growth rates but increase the total live weight produced from the facility. Surprisingly, studies have shown that crowding pigs has no impact on variation in weight within a populations.
Managing Variation. The major opportunity for managing variation is in the late finishing phase of production. Increasing the growth rates of all or part of the population of pigs at this stage will increase the number of animals that reach the desired weight window for the slaughter plant before the barn needs to be emptied. Removing and shipping the heaviest pigs from pens maximizes the proportion of the population in the target weight window and also increases the growth rate of the remaining pigs by on average 10% (DeDecker, 2006). Feeding 5 to 10 ppm ractopamine for the final 3 to 4 weeks prior to slaughter has been shown to increase growth rates by around 10% (Apple et al., 2007). The effects of pig removal and ractopamine on increasing growth rates are likely to be additive and used jointly these two approaches could increase growth rates in the last weeks of finishing by 20% or more.
Potential Nutritional Approaches to Reducing Variation. In theory, nutritional approaches to reducing variation in live weight at the end of finishing should target reducing variation in birth weight and/or increasing pre- and post-weaning growth rates, either of the entire population or of the lightest pigs.
Reducing Variation in Birth Weight: Variation in live weight starts in utero and a number of nutritional approaches to manipulating intra-uterine growth have been evaluated. A detailed discussion of this complex area is beyond the scope of this paper. A list of approaches for which published information is available suggesting positive impacts on the mean and variation in birth weight is presented below.
Maintaining Sow Body Condition in Lactation: Sows in poor condition at weaning and those with high body weight and fat loss during lactation have been shown to have greater embryo mortality and to produce smaller litters with greater variation in birth weight in the subsequent parity (Quesnel et al., 2008). However, ensuring that sows are weaned in good body condition is a target of practical sow nutrition programs and it is unlikely that poor body condition is a major cause of variation in well managed sow units.
Arginine: Several studies have shown that feeding relatively high levels of arginine in late pregnancy increases numbers born and birth weights (Wu et al., 2007). However, other studies have not been able to repeat this finding (Garbossa et al., 2015)
Feeding Level: There are claims that increasing the feeding level for sows during gestation, particularly during the later stages, will increase birth weights. However, there is limited published information in the scientific literature to support such claims. Lawlor et al. (2007) showed that substantial increases in feeding level (up to 50 or 100%) at various times in gestation had little effect on birth weight, variation in birth weight, or subsequent performance of the progeny.
Increasing Pre- and Post-weaning Growth Rates: Nutritional approaches to increasing postnatal growth have included cross-fostering smaller piglets onto sows with high milk production, spilt-suckling, providing the piglets with access to liquid milk replacer in the farrowing pen and the nursery, feeding complex rather than simple post-weaning starter diets, and using liquid feeding, and feeding fermented diets immediately post-weaning. All of these approaches have been shown to increase growth rate and live weight during the time that they are applied, however, these improvements have not generally been sustained to slaughter weights.
High-fat Diets: Feeding diets with added fat (up to 8%) to growing-finishing pigs under commercial conditions increases growth rates (De la Llata et al., 2001; Spencer et al., 2005). It is not clear if this will have any effect on variation. However, this is an approach that could be used to manage variation by increasing the number of pigs in a population that reach the minimum weight for the plant before the barn needs to be emptied.
Feeding to Requirements for Amino Acids: Feeding to the pigs’ requirements for amino acids will optimize growth performance. In practice, feeding programs are designed for populations (pens, rooms, or entire buildings) and normally diets are formulated for the “average” pig in the population. This results in large numbers of pigs in the population being either under or over supplied with amino acids which is, in part, responsible for variation in growth performance. It has been suggested that dividing the population into weight sub-populations and feeding each sub-population of these according to requirements would reduce population variation. However, the amino acid requirements of the weight sub-populations have not been established. Amino acid requirements change over time as the animals increase in weight and, practically, frequent changes in the diet fed are made to maintain alignment between the pigs requirement for and dietary supply of amino acids (i.e., phase feeding). In theory, increasing the number of phases should allow closer alignment between the animals’ requirement and dietary supply of amino acids and should reduce variation in growth rates within a population. However, several studies have actually found that using fewer dietary phases had no negative impact on overall growth performance. In fact, Moore et al. (2012) showed that feeding one diet from 22 to 106 kg live weight gave the same overall growth performance and carcass quality as using 3 diets or changing the diet every week. The major reason for this apparent anomaly is that pigs are able to compensate for periods of restricted growth when the amino acid supply is limiting by increasing growth performance later when supply becomes adequate.
Conclusions. In summary, variation in growth rates and live weights within pig populations is inherent and difficult to change. Practically, the factors that are likely to have the biggest impact on population variation (and which producers should focus on to minimize variation) are health status, weaning age, variation in age within the population, and the uniformity of the environment within the facility. Approaches to managing variation should focus on increasing the growth rate of either the entire population or of the lightest pigs. Feeding to requirements for amino acids, use of high fat diets, feeding ractopamine at the end of finishing, and removing the heaviest pigs from the pen for shipping to the plant are likely to be the most practical and cost effective approaches to managing variation.
Presented at CLANA 2016 in Cancun, Mexico.
Apple, J. K., P. J. Rincker, F. K. McKeith, S. N. Carr, T. A. Armstrong, and P. D. Matzat. 2007. Prof. Anim. Sci. 23:179-196.
Archer, G. S., S. Dindot, T. H. Friend, S. Walker, G. Zaunbrecher, B. Lawhorn, and J. A. Piedrahita. 2003. Biol. Repro. 69:430-436.
Campos, P. H. R. F., B. A. N. Silva, J. L. Donzele, R. F. M. Olivera, and E. F. Knol. 2012. Anim. 6:797-806.
DeDecker, J. M. 2006. PhD Dissertation, University of Illinois. De la Llata, M., S. S. Dritz, M. D. Tokach, R. D. Goodband, J. L. Nelssen, and T. M. Loughin. 2001. J. Anim. Sci. 79:2643-2650.
Donovan, T. S., and S. S. Dritz. 2000. J. Am. Vet. Med. Assoc. 217:79-81.
Garbossa, C. A. P., F. M. Carvalho Junior, H. Silveira, P. B. Faria, A. P. Schinckel, M. L. T. Abreu, and V. S. Cantarelli. 2015. J. Anim. Sci. 93:2872-2884.
Gonyou, H. W. 1998. Allen D. Leman Swine Conference pp 126-128.
Main, R. G., S. S. Dritz, M. D. Tokach, R. D. Goodband, and J. L. Nelssen. 2004. J. Anim. Sci. 82:1499-1507.
Moore, K. L., B. P. Mullan, and J. C. Kim. 2012. Anim. Prod. Sci. 53:52-56. Peterson, B. A. 2008. PhD Dissertation, University of Illinois. Quesnel, H., L. Brossard, A. Valancogne, and N. Quiniou. 2008. Anim. 2:1842-1849.
Quiniou, N., J. Dagorn, and Gaudre. 2002. Livest. Prod. Sci. 78:63-70.
Schinckel, A. P., and B. A. Craig. 2002. Prof. Anim. Sci. 18:219-226.
Schinckel, A. P., J Ferrel, M. E. Einstein, S. A. Pearce, and R. D. Boyd. 2004. Prof Anim. Sci. 20:79-86.
Shull, C. M. 2013. PhD Dissertation, University of Illinois. Spencer, J. D., A. M. Gaines, E. P. Berg, and G. L. Allee. 2005. J. Anim. Sci. 83:243-254.
Tokach, M. D. 2004. Adv. Pork Prod. 15:281-290. Wolter, B. F., M. Ellis, S. E. Curtis, E. N. Parr, and D. M. Webel. 2002. J. Anim. Sci. 80:2242- 2246.
Wu. G., F. W. Bazer, T. A. Davis, L. A. Jaeger, G. A. Johnson, S. W. Kim, D. A. Knabe, C. J. Meininger, T. E. Spencer, and Y-L Yin. 2007. Livest. Sci. 112:8-22.