Pork producers and breeding herd managers typically record a single reason for culling a sow. Often, they simply check a box in their recordkeeping program without reporting other health-related conditions that may be occurring simultaneously.

Classifying cull sows into various categories is done by lay personnel and serves as an evaluation of health status, which is frequently used to make management decisions.

Many recordkeeping software systems do not allow producers to record multiple reasons for culling a sow. Typically, the reason for culling is based on external signs and does not incorporate evidence of internal lesions or diagnostic testing results.

Evaluating sow carcasses at harvest plants presents an opportunity to confirm the reasons recorded on the farm and characterize the factors that may cause sows to be culled. Studies that investigate the presence of gross lesions in harvest plants are few.

Postmortem examinations of sow reproductive organs can potentially identify the reason for reproductive failure. A comparison of the on-farm culling records and the observations made at the plant can provide a means of evaluating the accuracy of recorded culling codes.

Standardizing Culling Codes

A retrospective cohort study compared reported reasons for culling with production data and lesions observed at harvest. Records on 923 uniquely identified sows from eight conventional, farrow-to-wean farms with common standard operating procedures were collected at two harvest plants in the Midwest. Individual production records and the reason for culling were obtained.

Criteria were established to identify sows that did not match the culling codes of old age, did not conceive, anestrous, poor body condition, farrowing productivity and cesarean section.

Old age was considered an appropriate culling code if a sow was greater than Parity 5. Sows typically reach maximum body weight at the fifth parity; therefore, sows culled at or before their mature size is reached, arguably are not old.

The “did not conceive” code was considered an accurate culling reason if sows had not conceived within 45 days after weaning. The 45-day limit was established by adding three days for an early return to estrus after weaning and two normal, 21-day estrous cycles. This was consistent with procedures reported by farms participating in the project.

Anestrous was considered an acceptable culling code if sows were culled eight or more days after weaning. The eight-day limit is considered at the upper range of a normal wean-to-estrus interval and anestrous could not be definitively determined prior to this time.

Body condition was considered an acceptable reason to cull a sow if body condition score was less than 3. Body condition score was evaluated prior to slaughter using a scale of 1 (thin) to 5 (fat) [See Body Condition Scoring Guidelines in “Managing Sows in Gestation” Blueprint, National Hog Farmer, April 15, 2006].

Farrowing productivity was established as an acceptable culling code if sows had poor number born alive, poor number weaned, or poor milking ability and were culled at Parity 2 or greater.

Culling Parity 1 sows for poor litter performance was considered unacceptable because those traits are lowly heritable, indicating there is a large environmental influence.

Culling codes that could not be evaluated at the harvest facility, such as lameness, prolapse and other illnesses and management reasons, were accepted as correct without evaluating additional criteria.

Although lameness could not be determined at the harvest facilities, front and rear foot lesions were evaluated and recorded at the harvest facility. Foot lesions included cracked hooves, pad lesions, abscesses on any surface of the foot, overgrown toes and missing dewclaws.

The culling codes of “not found, unknown” or “sudden death” were not acceptable because the sows being evaluated made it to the harvest facility; therefore, those sows were in fact found, and were actively culled for some reason.

The Real Reasons Sows Failed

Of the 923 sows evaluated, 209 (23%) appeared to have had an inaccurate reason recorded for why they were culled. (See Table 1.) The relatively high frequency of culling code errors could result in health, management and financial decisions being based on erroneous information.

Taking a closer look at the individual reasons for culling, old age was improperly recorded in 19% of sows. Of the 322 sows recorded as being culled for old age, 40 (12%) were culled more than 21 days after weaning, 10 (3%) were Parity 5 or younger, and 12 (4%) were culled more than 21 days after weaning and were Parity 5 or younger.

In the study, sows from all parities (1 to 13) were identified as being culled for “old age.” Farms that cull for old age at a specific parity automatically forces the parity average down. This management practice changes herd parity distribution and may make it increasingly difficult to reach the point at which the average sow pays for herself — approximately three parities.

The code “did not conceive” was inappropriately recorded in 48 of 172 (28%) sows culled for this condition. Of the 48 sows culled less than 45 days after weaning, 43 (88%) were culled 29 to 44 days after weaning, indicating they may have been given just one opportunity to conceive.

Anestrous was inaccurately recorded in a relatively small proportion — 7 of 123 sows (6%). Within 30 days of weaning, 59 (48%) were culled for anestrous.

Body condition was improperly recorded in 31 of 90 sows (34%), meaning those sows had a body condition score of 3 or greater at slaughter. Of the 90 sows culled for body condition, 63 (70%) had a body condition score of 1 or 2. And, of the 73 sows with a body condition score of 1, only 27 (37%) had a culling code of poor body condition recorded at the farm.

Farrowing productivity was not appropriately recorded in 23 of 73 sows (32%). Of the 73 sows, 16 (22%) were culled at greater than 21 days after weaning, five (7%) were Parity 1 sows, and two (3%) were culled at greater than 21 days after weaning and were Parity 1 sows.

In addition to the 23 sows with an improper culling code, 16 (22%) averaged more pigs born alive/litter/parity than the study average (10.93). However, since the farrowing productivity code used by several farms represented poor number born alive, number weaned or milking ability, determining the underlying productivity reason for removal using this culling code was not possible.

Cesarean section was unacceptably recorded in 14 of 15 (93%) sows. Of those, 13 (87%) were pregnant at the harvest facility. The Cesarean sectionculling code appeared improperly applied because sows often die or are euthanized following the procedure.

Lameness was recorded as the culling code for 83 sows; as previously noted, lameness was not evaluated at the harvest facility. What we were able to observe and count were the number of lesions (cracked hooves, foot pad abrasions, abscesses, missing dew claws, etc.,) on each foot. The number of lesions per foot and per animal were then totaled.

Total number of front foot lesions in the lameness culling code tended to be higher (P 0.05) between the lameness culling code and the other culling codes combined (1.19 vs. 1.16). Although the presence or absence of foot lesions was recorded in the current study, the relationship between the severity of foot lesions and lameness was unknown.

Prolapse, other illness and management were recorded as the culling codes for 11, 8 and 2 sows, respectively. Prolapses were confirmed in 4 of 11 sows from harvest facility research.

Not found, unknown and sudden death were recorded as the culling codes for 18, 5 and 1 sow, respectively.

Based on the study's findings, herd investigations and interventions that rely heavily on farm reported assessments of clinical condition may have significant limitations.

Differences within and between farms, despite common management and standard operating procedures, suggest that developing health interventions in integrated systems will still require assessment of individual farm characteristics and management. Training farm personnel in the correct diagnosis and reporting of clinical conditions of sows could be very beneficial.

Table 1. Frequency of Sow Culling Codes and Improper Culling Codes From Eight Farms Within an Integrated U.S. Production System
Cull code Frequency % Frequency of improper culling codeb %
Old agea 322 35 62 19
Did not conceive 172 19 48 28
Anestrous 123 13 7 6
Body condition 90 10 31 34
Lameness 83 9 0 0
Farrowing productivity 73 8 23 32
Not found 18 2 18 100
Cesarean section 15 2 14 93
Prolapse 11 1 0 0
Other illness 8 1 0 0
Unknown 5 1 5 100
Management 2 0 0 0
Sudden death 1 0 1 100
Total 923 100 209 23
aAverage parity at removal of culled sows = 5.44.
bCulling codes not meeting defined criteria.