Observational study of factors associated with nursery pig performance

Abstract

The objective of this study was to determine the association between body weight of pigs at 7 wk of age, specific management factors, and previous body weight. Among 3736 pigs from 8 commercial farms, greater 7-wk weight was associated with greater birth weight (P = 0.001), greater weaning weight (P = 0.001), and earlier age at weaning (P = 0.001). Farms with the earliest average weaning age had the highest health status. Farms with later weaning ages tended to use nurse sows for light-weight pigs. Pigs from farms using all-in/all-out management in the nursery had greater 7-wk weights (P = 0.01). The total mortality and culling rate during the nursery phase (weaning to approximately 7 wk of age) was 6.8%. Higher losses were observed among pigs with low weaning weights (< 4.1 kg) compared with those with higher weaning weights. More of the variation due to 7-wk weight was due to individual sow rather than to farm of origin.

Introduction

One of the main limitations to satisfactory post-weaning performance is low weaning weight (1). Animal growth is regulated by many factors, including pathogen pressure, environment, genetics, and nutrition (2,3). Many factors have been associated with variation in pre-weaning piglet weight; up to 35% of this variation is due to specific management factors, such as all-in/all-out (AI/AO) practices (4). The desire to optimize pig through-put in swine facilities has highlighted the number of slow-growing, small pigs at close-out. It is difficult to achieve AI/AO pig flow when there is substantial variation in growth rate (4–10). Variability in post-weaning growth performance should be minimized in order to maximize space utilization, especially in AI/AO pig management systems (4–10).

Few studies have directly addressed the causes of between-pig variation in body weight, largely because of the substantial expense and effort required to weigh pigs individually (11,12). Only recently have researchers reported body-weight variation within groups of pigs and have related this variation to environmental factors (10,11). The objective of this study was to determine the association between the weight of pigs at 7 wk of age, specific management factors, and previous body weight.

Materials and methods

Herds

This study involved 3736 pigs from 8 commercial farms (Table I). The farms were selected because they represented a wide range of management styles and were within a 1.5-h drive of the University of Guelph; in addition, the producers were willing to participate. The herds ranged in size from 150 to 1200 sows and included farrow-to-finish pigs, farrow-to-feeder pigs, and off-site nursery units. The herds varied widely in disease status and were categorized as low, moderate, or high health status on the basis of the presence of several pathogens, including Mycoplasma hyopneumoniae, Porcine reproductive and respiratory syndrome virus (PRRSV), Actinobacillus pleuropneumoniae, Sarcoptes scabiei, and Brachyspira hyodysenteriae (Table II).

Table I.

Herd and production characteristics of 8 commercial Ontario swine farms

					Weight (kg)
Farm number	Number of sows	Number of pigsa	Average weaning age (d)	Weaning delayedb	At birth	At weaning	At 7 wkc	Creep feedd	Production type	Genetic source
1	750	687	22	Yes	2	6.5	14.4	Yes (9 d)	F–F	Purchased F1
2	200	158	23	No	2	6	13.2	Yes (14 d)	F–F	Purchased F1
3	200	271	24	No	1.5	5	11.6	Yes (16 d)	F–F	Purchased F1
4	150	262	22	No	2	5.5	11.7	No	F–F	Home-reared rotational cross
5	250	261	21	No	1	6	15.4	Yes (9 d)	F–F	Synthetic linese
6	700	715	20	No	2	6	13.4	No	F–f	Purchased F1
7	120	104	16	No	1.5	5	18.3	Yes (10 d)	MOS	Synthetic linese
8	500	335	27	Yes	1.5	5	13.5	No	F–F	Home-reared rotational cross

F–F — farrow-to-finish; F–f — farrow-to-feeder; MOS — moved to off-site nursery

^aNumber of pigs weighed into the study; total 3736

^bLight-weight pigs were kept in the farrowing room on a nurse sow after the expected weaning age

^cAdjusted because of variable age distribution

^dFed to the pigs while they are nursing

^eBreeding stock producer

Table II.

Herd health status

							Pig flow in room
Farm number	Mycoplasmaa	PRRSV	APP	Mangeb	Dysenteryc	Health status	Farrowing	Nursery
1	+	+	−	−	−	Moderate	Variable	AI/AOd
2	+	+	+	+	−	Low	CF	CF
3	+	−	+	−	−	Moderate	CF	CF
4	+	+	+	+	+	Low	CF	AI/AOd
5	−	+	−	−	−	High	CF	AI/AOd
6	+	−	−	−	−	High	AI/AO	AI/AOd
7	−	+	−	−	−	High	AI/AO	AI/AOe
8	+	+	+	+	+	Low	AI/AO	CF

PRRSV — Porcine reproductive and respiratory syndrome virus; APP — Actinobacillus pleuropneumoniae; CF — continuous flow; AI/AO — all-in/all-out

^aMycoplasma hyopneumoniae

^bSarcoptes scabiei var. suis

^cBrachyspira hyodysenteriae

^dAI/AO by room

^eAI/AO by nursery site

Experimental design and observations

The study was performed between July 1998 and February 1999. All pigs born during a 3- to 5-wk period in each herd were included to obtain a minimum of 150 pigs per farm. The pigs were identified at birth with colored ear tags and were individually weighed at birth, at weaning, and at approximately 7 wk of age. Individual records were kept for each pig that included age, sex, and clinical disease, as well as mortality in the pen, number of pigs per pen, feeder spaces, water access, pen space per pig, and regrouping of pigs.

At birth and at weaning, the pigs were weighed on a postal scale that measured in 0.1-kg increments (Pelouze Model 4010; Sunbeam Products, Aurora, Illinois, USA). At 7 wk of age, the pigs were weighed with a walk-on, mobile, digital scale (Reliable Scale Model 5; Reliable Scale, Calgary, Alberta) that measured in 0.1-kg increments.

Calculations and statistical analysis

Data were entered into a spreadsheet (Quattro Pro 8.0; Borland, Scotts Valley, California, USA) and then exported to Statistix (Analytical Software, Tallahassee, Florida, USA) for descriptive analysis. Weaning age was described as “early” for pigs weaned before 18 d, “late” for pigs weaned at 28 d or more, and “moderate” for pigs weaned between 18 and 27 d. This weaning-age variable was assigned to individual pigs. The age distribution for 7-wk weight varied by approximately 3 wk (51.3 ± 6.2 d). Therefore, 7-wk weight was standardized to the median age (50 d) by calculating the average daily gain (ADG) in kilograms for each pig in the nursery barn, subtracting 50 d from the recorded age at which the 7-wk weight was obtained, and subtracting the product of these 2 variables from the recorded 7-wk weight for each pig.

Simple linear regression was used to assess the association between adjusted 7-wk weight and birth weight, parity of the sow, total number of piglets nursing the sow, sex, genetics, weaning weight, weaning age, AI/AO management, and occurrence of clinical disease during the study. Factors that showed significance at P < 0.10 were kept for the subsequent analysis. The dependent variable (adjusted body weight of individual pigs at 7 wk of age) was then regressed on the significant factors in a multivariable regression analysis after controlling for random farm and sow effects using PROC MIXED (13) (PC/SAS, SAS System for Windows, version 8; SAS Institute, Cary, North Carolina, USA). All interactions between main effects and quadratic effects were also tested. Covariance parameter estimates were used to calculate the total variance due to farm effects, sow effects, and the error term. Unless stated otherwise, values are presented as least square means ± standard deviation (s).

Results

The across-farm average body weights were 1.69 ± 0.4 kg at birth, 5.80 ± 1.5 kg at weaning, and 14.89 ± 3.82 kg (adjusted) at 7 wk. The ADG in the nursery barn was 0.30 ± 0.11 kg. The coefficient of variation (CV) in the adjusted 7-wk weight was 26.6%.

Weaning age ranged from 14 to 29 d, with a mean of 19.7 ± 3.5 d, and was dependent on farm-management policies. Pigs weaned at an early, moderate, or late time were weaned at ages 17.16 ± 2.9 d, 21.2 ± 1.5 d, and 27.4 ± 1.1 d, respectively. Early-weaned pigs weighed 5.2 ± 1.3 kg at weaning and 16.5 ± 3.8 kg at 7 wk (CV 23.4%), with an ADG of 0.334 ± 0.1 kg. Pigs weaned at a moderate age weighed 6.07 ± 3.7 kg at weaning and 13.6 ± 3.23 kg at 7 wk (CV 23.6%), with an ADG of 0.266 ± 0.09 kg. Late-weaned pigs weighed 6.9 ± 1.9 kg at weaning and 12.8 ± 3.04 kg at 7 wk (CV 23%), with an ADG of 0.25 ± 0.12 kg.

The total mortality and culling rate during the nursery phase (weaning to approximately 7 wk of age) was 6.8%. Of the pigs that died or were culled during this phase, 49.6% were of low weaning weight (< 4.1 kg), and 21.2% were pigs that weighed > 5.8 kg at weaning. The nursery mortality rate was farm-dependent and ranged from 1.2% to 10.2%. Weight variation at the nursery barn was also farm-dependent. The within-farm CV of weaning weight was 17.4% to 40.7% across the farms.

The fixed-effect model explained 20.3% of the variation in 7-wk weight (Table III). Greater adjusted 7-wk weights were associated with greater birth weights, greater weaning weights, earlier age at weaning, and AI/AO production. As birth weight increased by 0.5 kg, the adjusted 7-wk weight increased by 0.79 kg: pigs that were 1.0 kg heavier at weaning had an adjusted 7-wk weight that was 0.80 kg greater. Early-weaned piglets were 2.6 kg heavier than late-weaned pigs and 0.86 kg heavier than pigs weaned at a standard age. AI/AO production was associated with an increase in 7-wk weight of 4 kg compared with continuous-flow (CF) production.

Table III.

Factors associated with weight of nursery pigs at 7 wk of age

Variablea	Coefficientb	Standard error	Partial r2c (%)	P value
Intercept	7.08	0.70	—	< 0.001
Birth weight	1.58	0.16	4.7	< 0.001
Weaning weight	0.80	0.04	12.7	< 0.001
Standard weaning age	−0.86	0.28	0.4	< 0.001
Late weaning age	−2.60	0.38	2.2	< 0.001
AI/AO pig flow	5.42	0.93	0.3	< 0.01

^aFarm and sow nested within farm were included as random variables

^bIndicates the predicted change in the 7-wk weight as the variable is changed by 1 unit; for example, if the weaning weight increased by 1 kg, the 7-wk weight increased by 0.8 kg

^cIndicates the proportion of the variation in the 7-wk weight explained by the variable; for example, changes in weaning weight accounted for 12.7% of the total variation in 7-wk weight

The estimated variance in the farm-to-farm adjusted 7-wk weight was 7.93%, the sow-to-sow variance within the farms was 28.03%, and the error-term variance, which included the pig-to-pig variance, was 64.03% of the total variance.

Discussion

Birth weight and weaning weight both had a significant influence on nursery pig performance (Table III), consistent with previous findings (4,14,15). Although the total number of pigs born, the number of pigs nursing the sow, and the parity of the sow were associated with weaning weight, after controlling for weaning weight these factors were not associated with nursery pig growth (16). Sex was not associated with weaning weight or 7-wk weight. Some of the farms were below the achievable targets for today’s genetic pig lines of 15 kg at 7 wk of age and an ADG of 0.429 kg (4,12). Of the weight variation at 7 wk of age, 12.7% was explained by the weaning weight (1).

As in previous studies (17,18), greater ADG and greater 7-wk weight was reached by pigs weaned before 18 d of age. Dritz et al (19) had the same results and concluded that they were due to the time during which the piglet is in contact with a higher concentration of pathogens (i.e., in the farrowing room and with the sow). Although the weaning-age variable was assigned to individual pigs, the farms with the lowest average weaning age also had the highest health status (Table II). Similarly, the farms with the highest weaning ages had the lowest health status. The farms with higher weaning ages tended to keep light-weight pigs in the farrowing rooms for an additional week after the rest of the litter was weaned. The disease status of the farms with older weaned pigs and the use of nurse sows may partially explain the relationship between weaning age and nursery-pig weight gain.

Light-weight pigs (< 4.1 kg) entering the nursery had a higher rate of death or culling (49.6%) and were lighter at 7 wk of age than pigs that had been heavier entering the nursery (> 5.8 kg). These larger pigs had a nursery death and culling rate of 21.2%. Factors that may contribute to the higher mortality rate in light-weight pigs include the possibilities that light-weight pigs are carriers of disease (20), are poor converters of feed (21), and may consume less feed than heavier pigs.

The post-weaning ADG was higher on farms with AI/AO nursery management than on farms with CF nurseries (2,14,22). Organisms become established in CF nurseries, and pathogens spread to younger animals from older pigs (2). Unfortunately, it was not possible to separate the confounding effects of disease status from AI/AO management. As can be seen from Table II, the 3 farms with the least disease used AI/AO management, whereas the CF nursery facilities had a moderate or low health status. Therefore, the impact of the AI/AO variable is likely a combination of management style and disease status. After weaning, depressed feed intake, poor growth rate, increased disease susceptibility, and behaviour problems may be caused by a combination of factors, including space restriction, large group size, restricted feeding, type of feeder, and environmental stresses such as excessive heat or cold, or drafts, in addition to the influences of moving, mixing, and separation from the dam and litter mates (23–25). Optimal production requires control of multiple factors. The producer must maximize the genetic potential of the particular type of pig by controlling room environment to reduce temperature fluctuation, controlling disease, and providing an appropriate diet at various ages (11,14,19).

Sow-to-sow differences explained more of the variation in 7-wk weight than farm-to-farm differences. Sows likely differ in their mothering ability, milking ability, and production of antibodies against infectious-disease organisms. The farm per se describes little of the variation in 7-wk weights. Although pigmanship and other farm variables likely impact 7-wk weight, these are overwhelmed by the large within-farm variation. Appropriate stocking density and feeder design, subsorting, and euthanasia of small or sick pigs have all been suggested as methods for reducing variation in growth rates (5,10).

Other researchers have suggested that variation in pig weight increases as pigs grow (4,6). From our study, one may conclude that the CV of incoming weight does not increase, because it was 25.2%, 26.3%, and 25.6% at birth, weaning, and 7 wk of age, respectively. However, absolute variation in weight does increase as the pig ages, because 25.2% of a 1.5-kg pig is 0.378 kg, whereas 25.6% of a 35-kg pig is 8.96 kg.

There was wide variation in the within-farm CV of weaning weight: from 17.4% to 40.7%. This illustrates the importance of conducting on-farm research in commercial units. Producers have an achievable goal of 17% but perhaps will set realistic goals just below their current level. The nursery phase is critical for the pig’s performance and perhaps the most challenging and dynamic phase of swine production. The pig’s growth in the growing-finishing barn will be improved if the pig is heavier and healthier when leaving the nursery barn (4,25,26).

The weight of a pig at a certain stage of life is strongly affected by its weight at an earlier stage. In this study, birth weight and weaning weight explained 4.7% and 12.7% of the variation in 7-wk weight, respectively. Pigs that were 1 kg heavier at weaning were predicted to be 0.8 kg heavier at 7 wk. Perhaps some of the heaviest pigs do not maintain the 1-kg advantage during the nursery phase because they do not adapt to solid feed as well as some of the average-weight pigs. As weight on entry into the nursery has a significant influence on growth, producers should try to maximize the weaning weight to improve nursery performance. Thus, management strategies to manage light-weight nursery pigs better than, and differently from, medium and large pigs, are becoming increasingly important.

Although the use of standardized 50-d weights does introduce some bias, this was assumed to be the most appropriate outcome measurement. As a pig gets older and larger, the ADG increases. Therefore, use of the standardized weight would result in a higher ADG for pigs weighed when they were older than 50 d and a lower ADG for pigs weighed earlier than 50 d. This standardized-weight calculation could conceivably have reduced the variation in ADG in this observational study. This would effectively have resulted in a bias towards the null hypothesis when factors associated with the outcome were modelled. It can therefore be concluded that the associations found in this study were valid.

In conclusion, the wide variation in 7-wk weight was associated with birth weight, age at weaning, weaning weight, and AI/AO production.