Low reproductive performance and high sow mortality in a pig breeding herd: a case study

Sow performance is a key component of the productivity of commercial pig farms. Reproductive failure in the sow is common in pig production. For every 100 sows served, 89 should farrow. In absence of specific diseases such as porcine parvovirus, pseudo-rabies, swine fever, leptospirosis and brucellosis, management failures are the most important causes of loss. A syndrome associated with reproductive inefficiency, and post-service vaginal discharge and high sow mortality in a commercial pig farm is described. Pregnancy failures exceeded 20% and sow mortality exceeded 12% for two consecutive years. The abnormal post-service vaginal discharge rate was 1.7% during the period of investigation. An investigation involving an analysis of farm records, a review of breeding management practices, clinical examinations, laboratory analysis and examination of urogenital organs was conducted. The main contributing factors found were a sub-optimal gilt breeding management, an inadequate culling policy in combination with a sub-optimal culling rate and the presence of cystitis in more than 1% of the urogenital organs examined. The high sow mortality rate was related to an aged breeding herd. A control programme was recommended based on management changes involving oestrus detection, movement of gilts post-service, hygiene in the service area, boar exposure post-service and urinary acidification. This programme failed to increase the farrowing rate due to incomplete implementation of the recommendations made. The farrowing rate increased to 86.5% subsequent to a farm manager change in January 2005, which resulted in complete implementation of the control programme.

: Vaccination programme (1999)(2000)(2001)(2002)(2003)(2004)(2005)(2006) Parvovirus / Erysipelas 1 E. coli 2 Aujeszky's 3 Gilts 2 ml at selection 2 ml, 3 weeks later 2 ml at 6 and 3 weeks prefarrowing 2 ml at selection Sows 2 ml, 7 days postfarrowing 2 ml at weaning 2 ml every six months Boars 2 ml every 6 months -2 ml every six months PEER REVIEwED conducted in the presence of a boar once a day from Monday vto Wednesday. Sows were served in the service area in the presence of a boar. Ninty-eight per cent of services were by artificial insemination (AI). The timing of services in relation to first detection of onset of oestrus is outlined in Table 2. Semen doses were delivered three days a week and some repeat sows were served by boars. Sows were moved immediately after second service to dry sow accommodation. They were pregnancy tested at 28 days after service and transferred to the farrowing rooms on day 110 of pregnancy. The lactating feed curve is outlined in Table 3. Additional water was available from nipple drinkers. The average lactation length was 26 days. Replacement gilts were home bred. During 2003, an average of five gilts were served per week. Before September 2003, gilts were selected at six months of age and raised up to 115 kg bodyweight in finishing pens, then moved to a gilt pool shed accommodating 60 gilts in three pens, and a boar in an adjoining pen. The stocking density was 2.7 m 2 per gilt. Gilts were offered a gilt rearing ration ad libitum until served. In September 2003, due to a continuous fertility problem in gilts, the farm manager established the replacement gilts in finishing pens, in three areas: 1. A holding pool with selected gilts (95-120 kg), offering a finisher diet ad libitum; 2. A service pool (120-150 kg), offering a lactating sow diet ad libitum. It accommodated 49 gilts in seven pens. The stocking density was 2.7 m 2 per gilt; and, 3. A service pen (Figure 1). Oestrus detection in the gilt service pool was conducted once a day. Gilts were artificially inseminated in a service pen in the presence of a boar and transferred three to five days after second service to sow stalls in the dry sow house until day 110 of pregnancy. They were offered dry sow ration following the same feeding regime as pregnant sows (Table 4). The farrowing rate was low for two consecutive years and sow mortality rate was high. The manager contacted the author´s practice requesting a veterinary investigation

aNalysis of farM reCords
The farm records were examined in January 2004.

Reproductive performance
The farrowing rate fell from 86% in 2001 to 77% in 2002-2003. The reproductive losses were mainly due to repeat breeders and sows found to be non-pregnant (Table 5).
Farrowing rates were low for all parities in 2003. Gilts had an exceptionally low rate of 66%. Sows of parity over six had a farrowing rate of 83% ( Table 6). The reproductive losses in sows and gilts for 2003 were compared (Table 7). Repeats, found not-in-pig and sold in-pig rates were high in sows and especially in gilts.

Parity profile
The herd parity profile was high. In Januar y 2004, the average parity, excluding unser ved gilts, was five. Aherne (2002) suggested an optimal average herd parity of between 2.5-3.0. Although the replacement rate was adequate at 40.5%, the low farrowing rate in gilts resulted in insufficient gilt farrowings (Table 8). Consequently, the culling rate was reduced and old sows (parity over six) were retained. The parity distribution of the herd was examined ( Table  9) and compared with the ideal parity distribution suggested by Finn (2002). The percentage of young animals (less than parity three) was satisfactor y in the period October 01 to Februar y 02, but became too low from July 2002.

Sow feed consumption
The average daily feed intake (ADFI) per sow during lactation was 5.5 kg in 2003. This was lower than the target of 6.2 kg suggested by Carroll (2005).

past laboratory reports
Blood samples were collected from 20 sows in March 2003. Seram samples were analysed for evidence of porcine reproductive and respiratory syndrome (PRRS), and infection caused by Leptospira australis, Leptospira bratislava and Leptospira muenchen. The results were negative in all cases with the exception of low titres (1/100) in three sows to Leptospira muenche, which were not considered significant.

CliNiCal eXaMiNatioNs
A farm visit was carried out on January 20, 2004. The following areas were assessed:

Service routine
Sows were served in the presence of a boar, allowing nose-to-nose contact between the sows and the boar during service. Service technique was found to besatisfactory.

Gilt services management
Oestrus detection was conducted using the back-pressure test in the absence of a boar. Gilts in oestrus were served immediately in a boar pen and again 24 hours later. They were brought back after service to the pen where detection occurred and moved to sow stalls three to five days after second service.

Vaginal discharges
One sow was found with mucopurulent vaginal discharge 11 days post-service. Discharges are considered to be abnormal after seven days post-service (Muirhead 1986). Abnormal vaginal discharge, mastitis or agalactia was not observed in lactating sows.

Sow stalls
Six stalls in the main dry sow house were damaged ( Figure  2), allowing some sows to escape and injure other sows by mounting them.

Service area hygiene
Washing and disinfection was not being carried out. A build-up of faeces was evident behind some sows ( Figure 2).

Dry sow house lighting
Fluorescent lights were switched on for six hours a day. Light intensity at sow level was 200 lux in different areas of the house.

Boar exposure
Only one boar was present in the gilt pool area. In the dry sow houses boar contact was not maintained after service.

Sow body condition
Body condition was scored at weaning, using a visual condition scoring system based on a 10-point scale described by Whittemore (1998). Condition score two is thin and boney, while eight is evident fat. A minimum condition score of around fouris the target at weaning and sows should farrow with a minimum condition score of about six. Scores above seven and below three are unacceptable at any time. Sow condition was variable.
Between 20-25% of the weaned sows had a condition score four or less, and 75-80% had a score of five.

Lactating sow feeding and water supply
Some feed troughs appeared very clean suggesting underfeeding. Water was supplied via a computerised liquid feed system in the morning and evening time; water nipples were not present in the troughs.

Sow mortality
No dead sows were available for necropsy during the investigation period. A number of culled sows were condemned due to emaciation. On-farm sow euthanasia level was high due to injuries, lameness or poor general health.

saMple ColleCtioN
The following samples were collected as part of the laboratory investigation:

Water samples
Drinking water samples were collected and submitted for analysis. Coliform organisms were detected, but numbers were lower than acceptable limits for livestock (less than 5,000 total coliforms /100 ml and less than 1,000 faecal coliforms / 100 ml) suggested by the US Environmental Protection Agency (1973).

Blood samples
Blood samples were collected from 20 sows and subjected to serological assays for PRRS, Aujeszky's disease (AD), Swine Influenza (SI) and Leptospira spp were carried out. PRRS and AD results were negative. Swine influenza H1N1 and H3N2 titres were negative or less than or equal to 1/80 in some sows, which was not considered significant.
No titres for Leptospira australis titres were detected. Leptospira bratislava and Leptospira muenchen titres were negative except for a titre of 1/100 which was not considered to be significant.

eXaMiNatioN of urogeNital orgaNs
Ten sows were culled on January 29, 2004 as a result of poor reproductive performance and their urogenital tracts were examined. The findings of the examination are outlined in Table 10.    (Table 11). Seventyeight per cent of bladders examined had thickened walls and hyperaemia of the mucosa. All kidneys were normal and E. coli was isolated from the bladder of sow 1596.

difereNtial diagNosis
1. PRRS was excluded on the basis of negative serological results; 2. Aujeszky's disease was excluded on the basis of negative serological results; 3. Swine influenza H1N1 and H3N2 was excluded on the basis of absence of clinical signs and insignificant serological titres (lower or equal to 1/80); 4. Porcine parvovirus was excluded on the basis of an effective vaccination programme and absence of an increase in mummifications which remained below 0.5% of pigs born. Muirhead (1986) specified less than 0.5% as a target level for mummifications; 5. Leptospira infections were excluded on the basis of negative serology results and negative FAT results on fallopian tube samples; 6. Cystitis; 7. Poor gilt management routine, including inadequate oestrus detection, overfeeding regime post-service and stress associated with possible mounting of served gilts still on heat by pen mates; and, 8. Inadequate culling policy.

diagNosis
Following examination of farm records, clinical, urogenital examinations and laboratory investigations, the main contributory factors to the problem were: Infectious: • Cystitis, associated with E. coli infection.
Non-infectious: • Sub-optimal oestrus detection frequency; Sub-optimal gilt oestrus detection technique; Overfeeding during the immediate post service period; and, Inadequate culling policy leading to an aged breeding herd.

reCoMMeNded CoNtrol prograMMe: reCordiNg of aborNMal post-serViCes disCharges
The farm manager was instructed to record the following details for each sow: Sow identification and parity; • Number of days post-service; and, • Subsequent reproductive performance. The farmer • was also instructed to collect vaginal swabs in a sterile manner. A demonstration of the correct technique to avoid contamination was provided.

Recording of presumptive cause of sow death
This was introduced in January 2004 and was determined by the clinical assessment of the carcase, within the limitations, by the farm manager or submission when possible to the laboratory.

Oestrus detection -service programme
It was recommended that oestrus detection be carried out twice a day for gilts and sows in the presence of a mature boar. Services were to be carried out when standing oestrus was first detected and repeated 24 hours later. This recommendation wasnot implemented.

Management of gilt services
Gilts were to be moved from the gilt pool accommodation to gilt pens in the sow service area at the first recorded oestrus and then served in stalls on the second recorded oestrus. This recommendation was not implemented.

Target gilt services per week
In January 2004, it was advised to increase gilt services to 10 services per week, to compensate for low farrowing rates and to allow for an increased culling rate of old sows.

Hygiene programme in the service area
Removal of faeces from behind the sows and power washing and disinfection between batches was recommended for the prevention of ascending urogenital infections. The recommendations were not implemented.

Modification of sow stalls
The rear of the stalls in the dry sow house was modified in April 2004 to ensure that all sows were confined in stalls Figure 3).

Boar exposure
It was recommended that boar contact in the dry sow houses and gilt pool area be increased. This measure was disregarded.

Pregnancy testing frequency
Pregnancy testing was conducted twice, at 28 days postservice and repeated at 42-50 days.

Gilt nutrition
A revised feeding programme was recommended (Table 12).

Nutrition during lactation
Feed levels were adjusted daily according to individual sow condition and appetite.

Light intensity in service area
Light intensity was increased to 250 lux at sow level. A timer was installed to ensure exposure to 14 hours light per day.

Culling policy
A guideline to the recommendations made is outlined below: • Weaned sows with shoulder sores, ulcerated hocks or traumatic injuries; • Sows with abnormal vulva discharge; and, • Old sows (parity over seven).

Water supplied to pregnant sows
An extra six litres of water per day, between the morning and the evening feeding times was provided to each pregnant sow using a computerised liquid feed system. The extra water supplied was generally consumed.

Dietary urinary acidifier
Ammonium chloride (97%) was included in the dry sow ration at a rate of 6 kg per tonne and in the lactating sow ration at a rate of 2.5 kg per tonne for eight consecutive weeks, commencing in late January 2004. showed that 23% of the sows in the herd were over parity six compared with 26% in 2003.   An average of seven gilts were served per week from January-October 2004. 5. The level of abnormal post-service vaginal discharges recorded from January-June 2004 was 1.7% (Table 15). The type of discharge observed was creamy without blood in sows between seven to 50 days after service. The duration of the discharge varied from one to five days. 6. Vaginal swabs were collected from four sows that had abnormal post-service discharge and sent to the laboratory for bacteriological examination. The results are summarised on Table 16.
7. Sow mortality reduced from 12.9% (October-December 2003) to 7.8% (July-September 2004). One sow was submitted to the laboratory for necropsy. A splenic torsion was reported. A list of presumptive causes of sow death recorded by the farm manager is outlined in Table 17. 8. Examination of urogenital organs from infertile gilts and sows (parity range 27) on October 20, 2004 revealed that incidence of cystitis had not fallen (Table 18). Samples for bacteriological analysis were collected from the uterus, bladder and kidney where gross changes were observed and a urine sample was collected. Culture techniques were conducted on the material collected. E. coli was isolated from uterus and bladder of sow 1795. Eight blood samples were collected and seriological examination for Leptospira spp. was negative. 9. Subsequent case outcomes also resulted in staff changes. Due to an incomplete implementation of recommendations, the unit manager, who started working in the unit in September 2001, was laid off and a new manager was appointed in January 2005.

subsequeNt Case outCoMe ii
In a meeting with the new unit manager in January 2005 the previous recommendations were reviewed. Emphasis was made on the following points: • Implementation of an active culling policy; • Introduction of twice-a-day oestrus detection policy in the presence of a boar; • Implementation of gilt services in stalls in the sow service area; • Upgrading service area hygiene; and, • Frequent exposure between served sows/gilts and boars. After the full implementation of recommendations significant improvements were observed: 1. The farrowing rate increased from 78% (January-March 2005) to 85% for sows and gilts, and to 84% for gilts only for the period April-October 2005; and, 2. Sow mortality rate reduced to 4.1% (April-October 2005).

disCussioN
The impaired fertility problem affected mainly gilts and first parity sows; however, older sows had a higher farrowing rate. Therefore the high parity profile may in the short term have been beneficial to the fertility of the herd. It is normally stated that oestrus period in the sow lasts 24-60 hours and ovulation occurs 38-42 hours after the onset of heat. Service should take place up to 12 hours before ovulation. A delay in mating could result in lowered embryo survival (Muirhead 1991). Oestrus detection occurred every 24 hours. Therefore, some gilts or sows may have been on the point of ovulation when first detected resulting in conception failure. White (2006) suggested checking for signs of oestrus at least every 12 hours.

PEER REVIEwED
Gilts that have been moved into a boar pen for ser vice and returned to a group pen can experience increased stress levels sufficient to cause pregnancy failure (White 1996). In the present case study gilts were returned immediately after ser vice to group pens for a period of three to five days. This may have triggered adrenaline secretion due to excessive mating behaviour of gilts in oestrus by pen mates during the immediate post-ser vice period. White (1996) stated that sperm transport to the uterotubal junction and then to the fallopian tube is influenced by oxytocin release and that adrenaline is an antagonist to oxytocin. Boar exposure to ser ved gilts and sows in the unit was restricted. In this context, it is notable that White (1996) reported that the presence of a boar has a beneficial effect on the maintenance of pregnancy, and secondly that boar contact further helps to stimulate oestrus and its detection 18 to 23 days after ser vice if conception has failed to occur. Suboptimal light intensity (200 lux) was obser ved in the main dr y sow house. Lighting periods were insufficient (six hours per day). Alexander and Muirhead (1997) recommended a minimum of 220 lux during the dr y period for 14 hours a day to maintain a viable pregnancy. Young (2003) stated that high level feeding during the first few days of pregnancy is likely to increase hepatic blood flow and clearance of progesterone resulting in a depressed concentration of progesterone in the blood that may affect embr yo sur vival during early pregnancy. The feed level offered to the gilts and sows in the immediate post-ser vice period was high and may have caused conception failure. Almond (1995) stated that attempts to breed sows with signs of abnormal vulva discharge are usually futile. This is in agreement with this case study in which seven out of 15 discharging sows repeated. The recommended water requirement for pregnant sows is between 8-12 litres per day (Henr y and Apley 1999). Although adequate volumes of water were supplied to pregnant sows, (approximately 13-19 litres per sow per day), the extra water meant to encourage water consumption, thus reducing periods of recumbency and the inter vals between urination episodes. This may have a positive flushing effect on the bacteria in the bladder to increase as coliform bacteria replicate ever y 40 minutes (Carr et al. 1995) and cystitis develops when the bladder is not emptied frequently and bacteria concentrate in the urine increasing the risk of infection of the bladder. Deen and Xue (1999) found that where the gilt pool was too small, sow deaths increased due to sows being retained too long in the absence of a suitable replacement. In this case study, the number of gilts available for ser vice was sub-optimal and, as a result, old sows were retained increasing sow deaths and pregnancy losses due to on-farm euthanasia and casualty slaughter. Carroll (1999) stated that a sound culling policy is an integral part of herd management as old sows retained in herds lead to decreased productivity.
A significant reduction in sow mortality subsequent to staff change was obser ved. This was due to a high culling rate (68%) during the first quarter of 2005. Jones (1968) reported that urinar y tract disease was one of the main causes of death in sows kept under intensive conditions. In the present study, the sow death rate was over 9% throughout the period of investigation. Carr, Walton and Done (1995) stated that if annual sow deaths were in excess of 5%, then cystitis and pyelonephritis must be considered possible causes. The presence of cystitis in a high proportion of old sows may have predisposed to increased sow mortality. Thornton et al. (1998) studied two herds where subclinical urogenital infection (UGI) was present and where discharge rates remained below 2%. They reported that older sows have higher rates of UGI and that subclinical UGI during gestation possibly contributes to a reduction of farrowing rate because of early embryonic deaths, abortions and sow deaths. Similar findings were observed in the present study. The post-service vaginal discharge rate recorded (1.7%) involved mainly parity five to eight sows and was associated with reproductive failure. The use of ammonium chloride was considered to be of potential benefit as a urinary acidifier to aid in the control of cystitis. Carr et al. (1995) and Wendt (1998) stated that ammonium chloride can be useful in the prevention of urinary tract infections. The former author stated that ammonium chloride has little effect in the absence of antibiotics. However, Spillane (1998) reported a significant improvement in conception rate from the inclusion of ammonium chloride in sow feed in the absence of antibiotics. In the present case, a response to the urinary acidifier, in the absence of antibiotics was not observed. The improved per formance in the unit subsequent to a staff change was an interesting feature of this case. Almond (1995) stated that detailed investigations of sow infertility problems on large farms demonstrated that the human factor continues to be a primar y cause of unacceptable sow per formance.

CoNClusioN
This study highlights the importance of monitoring and evaluating breeding management practices for the achievement of optimum reproductive performance in pig herds.

aCKNowledgeMeNts
The author wishes to thank Michael Burke and Paul Spillane for their assistance and advice provided in writing this study.