Managing pig flow through breeding management

John Carr

Murdoch University and Portec Australia

www.portec.com.au

Introduction

Achieving an even pig flow is associated with setting and meeting a suitable breeding target associated with a known farrowing rate. The farrowing rate is defined as “the number of sows which farrow to a given number of services” [(number of females bred)/(number of females farrowed)]/100%¹.

This paper addresses some of the causes of the variability in farrowing rate and proposes that variability can be reduced if producers factor consider not all females in the breeding pool have identical expected breeding outcome.

The farm health team – owner, manager, stockperson and veterinarian – should determine a breeding target based on the required pig flow². Once the weight of finishing/sale pigs is set, the number of females required to be farrowed to achieve the required pig flow output is easily calculated. The aim is to maximize the legal output of the farm³. The number of weaners per batch should not vary over the year; otherwise the finishing meat output will vary. The number of females bred is then determined for each batch to ensure that the farrowing, nursery and finishing spaces are filled.

Females cannot be transferred to different batches to achieve an “on-average” acceptable output. The modern farm has to stop using averages and embrace absolutes to maintain health.

This paper examines five variables that need to be considered to make a better judgment of the number of breeding females required to fulfill the pig flow targets:- parity; the previous lactation length; the weaning to service interval, repeat breeders and the effect of time of the year when mated.

Materials and methods

Farm records from the Pig Care database (PigCHAMP Inc) from the United States of America (US) were obtained for the years 2002-2004 and analysed for the farrowing rate as it varied over each of the variables:- parity, lactation length, wean to breeding period, seasonal variation and effect of returns to breeding. In addition, data from 15 farms in Western Australia (WA), using PigCHAMP as a recording programme, were combined into one unit and analysed.

The farms in WA were situated at latitude 30 to 32S. The farms in US were situated between latitudes 30 and 49N.

The data was extracted from the Pig Care and PigCHAMP records, inputted into Excel spreadsheets and results displayed graphically. Similarities or differences between different variables were analysed using the Student’s t test. The results were then analysed to produce a predicated value for each of the variables.

Results

· The effect of age (parity) of the gilt or sow on her subsequent farrowing rate

Parity 0 is defined as gilt until she has been mated¹. The effect of parity on the subsequent farrowing rate is shown in table 1.

Table 1

The effect of the female’s parity on her farrowing rate in US and WA

Parity	0	1	2	3-6	7+	Totals
US
FR %		76	80	81	76
Breedings	116577	92126	82657	173857	24508	489,724
% population	24	19	17	36	5
WA
FR %		75	79	81	78
Breedings	12362	9380	8156	18523	2799	51,220
% population	24	18	16	36	5

There is a 6% increase in the farrowing rate as the female ages from a gilt to a parity 2 sow. Sows of parity 7+ demonstrated a 3-5% reduction in farrowing rate as compared with 3 to 6 parity sows. The US data indicated a higher farrowing rate in the younger females (1^st to 2^nd parity). The production was identical in 3 to 6 parities in US and WA industries. The older sows preformed better in WA than in the US.

Figure 1

Effect of parity of the female pig on her farrowing rate

· The length of lactation of a sow on her subsequent farrowing rate

Within the US the farrowing rate gradually increases until a lactation of 15 days after which it stabilizes until a lactation length of 24 days. Lactations longer than 24 days demonstrated a progressive reduction in the subsequent farrowing rate.

In WA, the farrowing rate in lactations less than 15 days was extremely variable. There was a general trend for the farrowing rate to increase until a lactation of 18 days. After 24 days lactation the farrowing rate continued to improve.

Statistical analysis revealed no significant differences between days in the US or WA data. However, a statistical difference of p<0.01 was revealed when groups were examined 0-14 vs. 15-24 days and 15-24 vs. 25-35 days of lactation and subsequent farrowing rate.

Table 2

The farrowing rate with the length of the previous lactation

Days	12	13	14	15	16	17	18	19	20	21	22	23	24
US	71	75	77	78	78	79	79	79	79	79	79	79	77	%
WA	67	79	75	71	76	79	81	81	82	80	79	82	81	%

Figure 2

Farrowing rate in relationship to lactation length

The lactation length was examined with regard to the parity of the sow. The US and WA data indicated a similar pattern between different parities with increases to 15 or 18 day lactation respectively. Afterwards there is a plateau effect. The number of sows prior to day 10 was very small. The WA data is more variable which is associated with small numbers of individual records on single lactation length days.

Figure 3

The farrowing rate in relationship to lactation length by parity of the US sow

Figure 4

The farrowing rate in relationship to lactation length by parity of the WA sow

Parity

Figure 5

The wean to 1^st service interval with lactation length in WA

The 1^st parity sow wean to service interval was extended (above 8 days) for lactations shorter than 17 days in the WA data.

The data was examined to reveal the range in lactation lengths. The mean lactation length in the US was 18 days whereas it was 21 days in WA.

Figure 6

The population distribution at each lactation length

· The influence of the weaning to first service interval on the subsequent farrowing rate

In the US data the farrowing rate rose rapidly between days 2 to 4 post-weaning. Between days 3-6 the farrowing rate was 79-83%. Post-day 6 post-weaning the farrowing rate rapidly fell to day 10, reaching 68% after day 11. The farrowing rate rose to a secondary peak on day 16, after which time the farrowing rate again fell.

In WA, a similar pattern was seen, although the results from day 0-2 post-weaning were higher than US data at 76-78% farrowing rate. A dip was seen on day 7 which also bottomed out at days 10 to 11. After this the pattern of farrowing rate change per day post-weaning becomes extremely variable with no clear trend.

Table 3

The influence of the weaning to 1^st service interval on the subsequent farrowing rate

Days	0	1	2	3	4	5	6	7	8	9	10	11	12
US	67	63	71	79	83	83	79	74	72	68	68	71	73	%
WA	76	76	78	78	82	80	78	74	69	68	68	69	73	%

Statistical analysis of the US data demonstrated that there was no significant difference when each day was compared to the next day. However, when groups of days were examined, statistical differences were noted, table 4.

Table 4.

Analysis of the statistical significance of the wean to 1^st service interval on the subsequent farrowing rate

Comparison between group 1 and group 2		US	WA
Array 1	Array 2	P value	P value
0-2 days	3-6 days	0.008	0.015
0-3 days	4-6 days	0.009	0.009
3-6 days	7-16 days	0.009	0.016

Figure 7

The wean to 1^st service interval on subsequent farrowing rate

The relative proportions of sows with a specific wean to service interval is illustrated in figure 5. A secondary peak in numbers of sows cycling in both the US and WA data is seen around day 26 post-weaning.

Figure 8

Percentage of females bred on each wean to service period in days

The secondary peak was noticed around day 27 in the US data and examined in more detail.

Figure 9

Detail of the females bred between wean to 1^st service 14 to 33 days

The rise was noticed between days 25 to 29 in the US data. A similar rise was not particularly evident in the WA data.

· Effect of a repeat breeder

Sows occasionally fail to conceive or maintain their pregnancy and return to oestrus. The number of returns to oestrus since the last weaning event and subsequent fertility was analysed.

Table 5

The effect of repeat breeder on the farrowing rate

Breeding #	0 (post- weaning oestrus)	1^st return	2+ returns
US	80	63	46	FR %
WA	80	66	55	FR %

Figure 10

Farrowing rate with number of post-weaning returns to oestrus

· Effect of seasons

The variation in farrowing rate is predicable over the year. The farrowing rate varied from 81-73% in the US (8% variation) and 84-69% in WA (15% variation). The farrowing rate was not significantly different in the US for the months January to October. However, November and December were significantly different from the other months with a drop of 4% in the farrowing rate.

Table 6

The variation in the farrowing rate per month over this time frame

Days	JAN	FEB	MAR	APL	MAY	JUN	JUL	AUG	SEP	OCT	NOV	DEC
US
2002	75	78	79	79	80	79	80	80	81	80	78	77	%
2003	78	80	80	79	78	79	79	79	78	78	75	73	%
2004	74	78	78	78	79	79	80	80	79	79	76	75	%
WA
2002	80	77	74	72	69	73	78	79	79	80	81	79
2003	80	79	82	81	73	77	81	79	82	82	84	82
2004	82	83	82	75	75	77	77	77	79	79	80	79	%
Averages for the period 2002-2004
US	76	79	79	79	79	79	80	80	79	79	76	75	%
WA	81	80	79	76	72	76	77	79	80	81	81	80	%

Figure 11

Monthly farrowing rate in US centered on December

Figure 12

Monthly farrowing rate in WA centered on June

Figure 13

Average farrowing rate with month of farrowing

In Western Australia the monthly variation was pronounced with a large variation experienced throughout the year. The summer reduction in farrowing rate extended over a 3-4 month period starting in February with a peak bottom in May. In WA the average reduction in farrowing rate was around 8%.

There was no specific response detected by producers in either the US or WA to the inevitable seasonal variation. Both industries recorded their lowest services in February.

Figure 14

Percentage of females bred per month

Discussion

Failing to achieve the same number of farrowings per batch is a major cause of poor pig production resulting in variable output. Farms often consider their production by the number of sows living on the farm rather than a true limiting factor, such as the number of farrowing crates or the floor space available in finishing. Farmers de-stablise their pig flow by failing to set and then achieving adequate breeding targets².

The batch breeding target is based on the number of females to farrow/the farrowing rate. The farrowing rate is defined as the (number of females bred/number of sows farrowed)*100%¹. Producers utilize averages to compute needs whereas the pig works in whole numbers and absolutes. This paper looks at five variables which, if included in calculations, would allow producers to set more realistic breeding targets to eliminate or reduce output variations.

There are three types of females that are available to a farm to meet the batch breeding target.

Figure 15

The available breeding females and major factors associated with each that may affect the farrowing rate

Parity of the breeding female

A farrowing rate of 76% (US) or 75% (WA) was recorded for gilts. The reduction of 6% is sufficient to require adjustments to be made in breeding targets on farms where there were more than 12 farrowing crates per batch or where a larger than normal number of gilts were included in a batch. There were very similar parity profiles used by the US and WA industries.

Weaned sows lactation length

The expected farrowing rate was stable at 78-79% (US) for a lactation length of 15 to 24 days and at 79-82% (WA) for lactation lengths of 18-24 days. The US data indicates a lower farrowing rate for the longer lactation length (over 24 days) which is contra to previous reports where farrowing rate increases with an increased lactation length⁴. The differences in 15-18 day lactations and the subsequent farrowing rate between the two countries is possibly a difference in nutritional routines. The use of maize corn vs. small grains for example. There was little difference in farrowing rate and lactation length with parity in the US or WA data. Other reports have indicated⁵ that gilts with a short than 19 day lactation had an increased wean to service interval which can be expected to reduce farrowing rates. This trend was recognised in the WA but not in the US data. Parity 1 sows did not follow any particularly different pattern to parities 2 to 6 sows. Parity 7+ sows followed a more chaotic pattern in both the US and WA data sets. The two countries results diverged markedly when the lactation was in excess of 24 days. In the US the subsequent farrowing rate reduced, whereas in WA the farrowing rate increased with lactation length. The number of sows involved in the data extreme was low and the effect may be a sampling error, nutritional effect or a feature of farming, in particular animal selection for future breeding retention in the herd. WA generally weans 2 days later (18 days US vs. 21 WA). Two percent of the WA sample had a lactation over 28 days, whereas, these represent less than 0.5% of the US data set.

The inter-relationship of these different events is illustrated by analysis of the wean to service data by lactation length. In WA 1^st parity sows with a lactation length shorter than 17 days had an extended wean to service interval of 10 days. This has been previously described in US data with a 9 day wean to service interval of 1^st parity sows with 18 day lactations.

Effect of wean to 1^st service interval

The impact of wean to 1^st service interval followed reported patterns^4,6. The data describes a rapid rise in farrowing rate from 3 to 6 days post-weaning, followed by an equally rapid decrease from 7-11 days post-weaning. After 11 days post-weaning the farrowing rate again progressively increases but never reaches the farrowing rate expected for sows with a 4-5 day wean to 1^st service interval.

A noticeable secondary rise in sow cycling at 26/27 days post-weaning is seen in the US data. This is associated with sows that cycled normally 4-7 days post-weaning but were missed and recycled normally at 22-31 days (normal oestrus 18-24 days). The WA data does not specifically demonstrate this peak at 28 days.

There was an extremely long ‘tail’ in the wean to service data with individual sows recording over 200 days since weaning before being bred. Farmers should remember that for every 10 days the sow eats 25 kg of feed with no significant increase in her value. If the sow does not visibly cycle by day 30 post-weaning she should be culled as uneconomic.

Again there was a numerical difference in the farrowing rate between the US and WA herds. In the US data, extended weaning to 1^st service interval results in a progressively poor prognosis for subsequent farrowing rate. In the WA data, the trend-line was not clear. The peak of 3-6 days was significant (p<0.001) when compared to the surrounding days, in both the US and WA data.

Repeat breeders

The data demonstrated that a failure to conceive to a previous breeding event reduces the expected farrowing rate at the next mating event. A similar trend was seen in both the US and WA data, although the US herd had a lower farrowing rate on their 2+ breeding than similar sows from WA.

Seasonal effect

In the US sows farrowing in November and December experienced an average 4% reduction in the farrowing rate. In the WA data, the seasonal effects were more pronounced with a 9% average drop. The seasonal reduction in WA was deeper and longer than experienced in the US. This may be associated with the wider latitude range in the US data or that the temperature range in WA is greater than regularly experienced in the US.

Production of a predicated value adjustment for the batch breeding target.

Utilising the figures of this paper, a simple calculator can be constructed which allows producers to counteract these predicable effects:

Link to an example Excel spreadsheet

Breeding number calculator example

Table 7

Summary of five major affecters of farrowing rate

Predictor	Variable	FR %	Variable	FR %	Variable	FR %
US
Parity	Gilt	76	Sow	80
Lactation length	<15 days	75	>15 days	80
Wean to 1^st breeding interval	0-2 days	65	3-6 days	82	>7 days	72
Return	1	64	2+	43
Season	Farrow Nov/Dec	75	Farrow other times	80
WA
Parity	Gilt	75	Sow	80
Lactation length	<16 days	75	>16 days	80
Wean to 1^st breeding interval	0-2 days	76	3-6 days	82	>7 days	70
Return	1	66	2+	55
Season	Farrow Apr-June	72	Farrow other times	81

The bold figures need to be taken into particular account.

The importance of making these adjustments to the breeding target can be realized using a couple of scenarios.

The example farm is a US 30 sow batch farm.

Question: Given different ratios of breeding females what would be the farrowing rate and number of farrowing sows?

In each example 5 weaned sows are replaced with an increase in another group

Table 8

Example influence of managed pig flow using a 30 sow batch farm in the US

Female bred class:	Normal batch	Example 1	Example 2	Example 3	Example 4
		wean to breeding interval	gilts	returns	Summer
# wean-bred 3-6 days	30	25	25	25	30
# wean-bred 7+ days	0	5	0	0	0
# 1^st return sows	4	4	4	9	4
# gilts	4	4	9	4	4
Total # bred	38	38	38	38	38
FR %	79	76	76	76	68
# farrowed	30	29	29	29	26
Acceptable	Yes	No	No	No	No

Note the ‘collapse’ in farrowing rate is not associated with any pathogen for example leptospirosis.

If each empty farrowing crate raises fixed costs by reducing meat output by approximately 700 kg – if this simple management mistake only occurred once a month on a weekly batching sow, costs are increased by a meat reduction of nearly 8.5 tonnes a year.

Such simple adjustments to the breeding target must be made farm specific. The paper demonstrates that global differences occur between different pig industries; this difference is also seen on individual pig farms. Likewise, additional features such as a change in genetics, new stockperson or a new AI technique may warrant additional overall reduction in the farrowing rate.

Additional observation

The use of two countries with similar farming styles and genetics, revealed marked differences in the expectations of their sows. Papers written on reproductive issues, which predict outcomes, must be read with caution and applied only when local issues such as local climate, nutritional and genetic base are fully appreciated and taken into account.

Conclusion

The breeding pool should not be considered as one homogeneous female but should be broken down into its discrete groups.

A simple spreadsheet, easily applied to a PDA, will allow a more stringent breeding target to be set. Having more control over the batch breeding target allows for a more even pig flow and subsequent reduction in variability in finishing output.

The results of this investigation reveal an average farrowing rate of 80% whereas many producers think that the normal farrowing rate should be 87+%⁵.

Different countries produce different production results and analysis must be used with caution when moving outside the range of the data set.

For every empty farrowing crate 9.5 pigs will not be sold – which at 72 kg dead weight is 684 kg of pig meat not paid for. This loss can significantly raise the cost per kg and ultimately make the farm economically unsustainable. A failure to breed sufficient animals each batch is the single most expensive mistake a farm can make.

Definitions¹

Gilt	A female that has arrived in the breeding herd, but has not yet been mated
Parity	The number of times a female has farrowed
Weaning to service interval	The interval between the date of weaning (0) and the date of first mating (service)
Repeat breeder	A sow that returns to oestrus before the anticipated farrowing date but more than 5 days after mating
Service	One or more completed matings within the service oestrus period, the maximum length of oestrus period being 5 days

References

1	Pig Health Recording, Production and Finance. A producer’s guide. The Pig Veterinary Society. ISSN 0141-3074
2	Carr, J. (1999). Development of Pig Flow Models. Pig Veterinary Journal 43: 38-53.
3	Minimum standards for the protection of pigs - Council Directive 2001/88/EC
4	Young, M. and Aherne, F. (2006). Productivity and economic impact of age at weaning. Sow productivity and reproduction. AASV seminar proceedings 37.
5	Pizarro, G. and Spronk, G.D. (2006). 90% farrowing rate – the methods and materials used to achieve normal reproduction. AASV seminar proceedings 37.
6	Leman, A.D. (1992). Optimizing farrowing rate and litter size and minimizing nonproductive sow days. Veterinary Clinics of North America: Food Animal Practice. 8: 609-621

Thanks

A special thanks to Ms Susan Olsen PigCare, who kindly compiled the US data from PigCare US Records.