Sow factors allowed a high feeding level or ad libitum access to

J .J. Eissen et al. Livestock Production Science 64 2000 147 –165 151

3. Sow factors allowed a high feeding level or ad libitum access to

feed during gestation, they consume more feed than The interrelationships between milk production, needed to meet their energy requirement during changes in body weight and composition, and vol- gestation. Sows fed a high gestation feeding level, untary feed intake during lactation Fig. 1 are however, have a lower voluntary feed intake during complex and the controls of partitioning of nutrients lactation than sows fed according to their require- between milk secretion and deposition or mobilisa- ments during gestation, and mobilise more reserves tion of body reserves are poorly understood. The during lactation e.g. Mullan and Williams, 1989; physiological drive of lactating sows to produce milk Yang et al., 1989; Dourmad, 1991; Weldon et al., at the expense of other body functions is, however, a 1994a; Xue et al., 1997; Revell et al., 1998a; Fig. 2. key component of the metabolic state of lactating Le Cozler et al. 1998a showed that a higher feeding sows and is controlled by factors like litter size, level during rearing ad libitum vs. 80 of ad parity and genotype Pettigrew et al., 1993. It will libitum also led to a lower voluntary feed intake be discussed below how the sow factors body weight during lactation Fig. 2. Points are connected per and body composition at farrowing, parity, litter size study in Fig. 2 to indicate that each study shows an and genotype may affect voluntary feed intake of effect of a similar magnitude. lactating sows. A higher feeding level during rearing and or gestation generally results in a higher body weight 3.1. Body weight and body composition at and body fatness of sows at farrowing. The higher farrowing body fatness at farrowing seems to be associated with the lower lactation feed intake Dourmad, 1991; Because feed intake during early lactation is Williams and Smits, 1991; Revell et al., 1998a, generally too low to meet the energy requirements, whereas the effect of body weight at farrowing high producing sows mobilise body reserves to seems to be small O’Grady et al., 1985; Williams supply energy and nutrients for milk production and and Smits, 1991; Weldon et al., 1994a. The effect of hence to maintain and stimulate the growth rate of body fatness was illustrated by regression analyses of piglets Mullan and Williams, 1989. When sows are daily feed intake during lactation on backfat thick- Fig. 2. Relationship between daily feed intake during rearing ♦ or gestation other symbols and voluntary feed intake during lactation of a sow: j Revell et al., 1998a; d Mullan and Williams 1989; m Dourmad 1991; ♦ Le Cozler et al., 1998a; h Xue et al., 1997; s Weldon et al., 1994a. 152 J .J. Eissen et al. Livestock Production Science 64 2000 147 –165 ness at farrowing: Yang et al. 1989 estimated a directly into the blood circulation and very little 21 21 slope of 2 18 and 2 129 g day mm for glycerol is reesterified into triacylglycerides in adip- primiparous and multiparous sows, respectively, ose cells. Nonesterified fatty acids, on the other 21 Dourmad 1991 estimated a slope of 2 63 g day hand, may be reesterified into triacylglycerides to a 21 mm for primiparous sows and Koketsu et al. larger extent than glycerol before leaving adipocytes 21 21 1996a estimated a slope of 2 19 g day mm Revell et al., 1998a. Levels of NEFA are therefore across all parities. better indicators for fat mobilisation than for body Body weight and fat depots influence feed intake fatness per se. presumably by modulation of long-term regulation Concentrations in blood of NEFA during late mechanisms. The various control mechanisms are gestation were not significantly affected by rearing or either independently unique in action or synergistic gestation feeding level Weldon et al., 1994a; Revell and may vary according to the phase of the lactation et al., 1998a; Le Cozler et al., 1998b. Levels of period. How these may interact remains to be glycerol, however, were significantly higher in fat elucidated. Several studies focused on one or two of sows, which supports the mechanism described the mechanisms. Most of these studies used different, above Revell et al., 1998a. During the first weeks usually two, feeding levels during rearing or gesta- of lactation, levels of NEFA and glycerol were tion resulting in relatively fat and lean sows at always higher in fat than lean sows Weldon et al., farrowing. During lactation all sows were fed ad 1994a,b; Revell et al., 1998a; Le Cozler et al., libitum. Mechanisms possibly explaining the effect 1998b. However, it is not clear whether these higher of body composition on lactation feed intake of sows levels found post partum in fat sows cause, or are a are turnover of body fat tissue, insulin and leptin consequence of, the lower feed intake of fat sows, levels in blood and cerebrospinal fluid, presence of involving another mechanism. insulin resistance and glucose intolerance, and levels of milk production and body protein reserves. These 3.1.2. Insulin and leptin five mechanisms are described below, followed by As animals fatten, there is a gradual increase in paragraphs about meal eating behaviour of fat and basal blood insulin Woods et al., 1985, 1998 and lean sows, the effect of body fatness on voluntary leptin concentrations Woods et al., 1998. Con- lactation feed intake in relation to stage of lactation, centrations in the blood are the difference between and optimum body composition of sows at farrow- release in the blood and breakdown or uptake. Basal ing. insulin is defined as the amount of insulin measur- able in the blood in the absence of exogenous 3.1.1. Turnover of body fat tissue influences such as feed. It is generally estimated after Firstly, fat is stored in the body with a continuous a fast of 12–24 h. Insulin secretion is stimulated turnover which involves the release of fatty acids and acutely in response to the intake of a meal, whereas glycerol Forbes, 1988. The release into the blood- leptin secretion is not. The mechanisms governing stream is greater when the amount of body fat is leptin secretion remain to be elucidated, but insulin greater. The concentrations of mentioned substrates appears to play a key role Woods et al., 1998. or the extent of oxidation may act as signals that Insulin is secreted from pancreatic beta cells, where- could be read by the liver and sent to the brain via as leptin is a product of the obese gene which is vagal nerves Williams, 1998. Therefore, the sow expressed only in fat tissue. It is known that insulin may use the rate of fat metabolism to regulate and and leptin from the blood can penetrate into the monitor its energy status and hence voluntary feed cerebrospinal fluid at a slow rate, and that levels intake Williams, 1998. Glycerol may be a better within the cerebrospinal fluid can be considered as indicator of body fatness than nonesterified fatty an indicator over time of the levels in the blood acids NEFA because the only source of plasma Woods et al., 1985, 1998. Cerebrospinal fluid glycerol is from the breakdown of triacylglycerides levels change relatively slowly and thus are a more Revell et al., 1998a. Moreover, almost all glycerol stable parameter than blood levels. Higher levels of from the breakdown of triacylglycerides is released blood insulin and leptin as a result of a high feeding J .J. Eissen et al. Livestock Production Science 64 2000 147 –165 153 level during gestation would, therefore, lead to intolerance presumably results in a lower clearance higher levels of both in cerebrospinal fluid at farrow- rate of glucose from the blood after a meal. As a ing, which may inhibit feed intake Williams, 1998; consequence, use of peripheral glucose is likely Woods et al., 1998. Insulin and leptin then would decreased and voluntary feed intake may be reduced act as a homeostatic mechanism for body fat at the to maintain blood glucose concentrations. Further- brain level. more, lower blood insulin levels as a result of No references were found in which concentrations glucose intolerance may enhance the mobilisation of leptin and insulin in the cerebrospinal fluid or and oxidation of stored adipose tissue as oxidation of concentrations of leptin in blood of fat and lean sows NEFA is depressed to a lesser degree. The latter also were measured. Xue et al. 1997, Revell et al. may reduce voluntary feed intake. 1998a and Le Cozler et al. 1998b collected blood In a number of studies, sows were fed according samples during late gestation after fasting and found to a normal or high feeding level during gestation no difference in basal insulin levels between fat and resulting in relatively lean and fat sows, and blood lean sows. Weldon et al. 1994b reported no effect glucose and or insulin levels in late gestation and or of gestation feeding level on basal insulin level on lactation were determined, usually after an overnight day 1 of lactation. Basal insulin level at day 15 of of fast and after infusion of glucose. During late lactation was even lower for the high gestation gestation, Xue et al. 1997 found fat sows to be feeding level group Xue et al., 1997. more glucose intolerant, illustrated by higher glucose These results suggest that the contrast between and lower insulin concentrations in fat sows after lean and fat sows was not large enough to find glucose infusion compared with lean sows. Revell et differences in basal insulin levels or that differences al. 1998a did not find differences between fat and are more likely to be measured in cerebrospinal fluid lean sows in late gestation after glucose infusion. At Revell et al., 1998a. day 1 of lactation, Weldon et al. 1994b found symptoms of insulin resistance as insulin peak 3.1.3. Insulin resistance and glucose intolerance secretion after glucose infusion was not affected by Another possible mechanism that also involves gestation feeding level, but the rate at which glucose insulin is the development of insulin resistance and was cleared from the blood was much slower for the or glucose intolerance. Insulin usually regulates both fat sows. Xue et al. 1997 infused sows with blood glucose levels and fat mobilisation, with the glucose at day 15 of lactation and again found fat result that oxidation of NEFA is depressed and sows to be more glucose intolerant, the impaired oxidation of blood glucose is stimulated Kronfield, glucose tolerance being more severe during lactation 1971; Revell and Williams, 1993. Excessive feed than during late gestation. Weldon et al. 1994a and intake during gestation may cause the sow to become Revell et al. 1998a studied blood glucose and insensitive to insulin, probably by affecting insulin insulin concentrations during lactation without infus- receptor number and or affinity. The sow will then ing glucose. Weldon et al. 1994a found lower exhibit a smaller response in glucose clearance to the concentrations of insulin for fat sows, whereas same amount of insulin Weldon et al., 1994b. A concentrations of glucose were not influenced, which high feeding level during gestation may also cause also points towards glucose intolerance. Differences the sows to become glucose intolerant, possibly by in insulin levels were especially clear during early decreasing the number of glucose receptors and lactation. Revell et al. 1998a did not find an effect reducing sensitivity of beta-cells in the pancreas to of gestation feeding level on glucose or insulin glucose Murray et al., 1990; cited by Xue et al., concentrations during mid and late lactation. 1997. When a sow becomes glucose intolerant, she Le Cozler et al. 1998b varied the feeding level will exhibit a smaller response in blood insulin to the during rearing, while all sows received the same same amount of glucose Xue et al., 1997. Both amount of feed during gestation. The difference insulin resistance and glucose intolerance may lead between fat and lean sows at farrowing was therefore to higher glucose concentrations after a meal. mainly a carry over effect from the rearing period. Development of insulin resistance and or glucose Infusion of sows with glucose in late gestation did 154 J .J. Eissen et al. Livestock Production Science 64 2000 147 –165 not affect rate of glucose clearance; however, a similarly affected during gestation, which is sup- higher peak insulin secretion was observed for fat ported by the fact that Head et al. 1991, Head and sows, showing that fat sows may be more resistant to Williams 1991, 1995 and Revell et al. 1998b insulin. During mid lactation, fat sows showed a used suboptimal diets during gestation in order to poorer glucose tolerance and appeared to be more support fat deposition at the expense of lean deposi- resistant to insulin than lean sows after glucose tion. These diets were used to create fat sows at infusion. It seems, therefore, that impaired glucose farrowing that had a similar net weight gain during clearance is more likely caused by body composition gestation compared with lean sows at farrowing, at farrowing per se, than by a high feeding level which received a more optimal diet during gestation. during the preceding gestation. In conclusion, it seems that differences in glucose 3.1.5. Body protein reserves intolerance and insulin resistance between fat and Another reason for fat sows to have a lower lean sows may, at least partly, explain the lower voluntary feed intake might be the lower supply of voluntary lactation feed intake of fat sows, although endogenous substrates for milk production Wil- results are not fully unambiguous. Results of the liams, 1998, which could also be linked to the studies in which glucose was infused may differ due reduced milk output of fat sows in the studies of to the dose used, as Weldon et al. 1994b and Xue et Head et al. 1991, and Head and Williams 1991, al. 1997 infused 1 g glucose per kg body weight, 1995. Fat sows have less protein reserves to supply whereas Revell et al. 1998a and Le Cozler 1998b substrates for milk production compared with lean infused 0.06 g and 0.5 g per kg body weight, animals at a similar weight Revell et al., 1998a. In respectively. studies that changed feeding level during gestation and not diet composition, fat sows were heavier at 3.1.4. Milk production farrowing Weldon et al., 1994a,b; Xue et al., 1997; Head et al. 1991 and Head and Williams 1991 Le Cozler et al., 1998a,b and may have had a reported that fat sows, in comparison with lean sows, similar or even higher amount of body protein had a lower capacity to secrete energy in milk reserves than lean sows. If milk output is limited by because they had fewer milk secretory cells. This the supply of endogenous amino acids, then the may have caused the significantly reduced litter capacity of the animal to produce milk is reduced. growth of the fat sows reported by Head and Limited body protein reserves may therefore reduce Williams 1995. Revell et al. 1998b reported that milk production and hence the voluntary feed intake milk yield was about 15 higher in lean than fat of sows Williams, 1998. sows, which was also reflected in litter growth. A Mahan and Mangan 1975 found that voluntary lower milk production may diminish the drive to eat feed intake during lactation was reduced when sows and reduce voluntary feed intake of sows Fig. 1. In were fed a diet low in protein during gestation and most other studies, however, litter growth was not lactation. However, voluntary feed intake during affected by a high gestation feeding level and fatness lactation was not reduced when the diet during at farrowing, indicating that the effect of fewer milk gestation was high in protein, indicating that endog- secretory cells of fat sows was probably limited e.g. enous protein reserves may limit voluntary feed Dourmad, 1991; Weldon et al., 1994a,b; Xue et al., intake under certain circumstances. Mahan 1998 1997. found that multiparous sows consumed more feed In dairy cattle, rapid rates of growth in the during the first week of lactation and primiparous prepubertal period are associated with substantial sows during the whole lactation when offered a diet reductions in milk production in all subsequent with a higher protein content during gestation. Revell lactations Little and Kay, 1979; Sejrsen et al., et al. 1998a used a low and high protein diet during 1982. Recent reports suggest that the deleterious lactation. The dietary supply of protein increased effect is associated with feeding heifers a diet with voluntary feed intake during weeks 3 and 4 of an inadequate protein:energy balance resulting in an lactation, possibly by increasing milk production and excessive fat deposition during the prepubertal hence the drive to consume feed. period Mackenzie and Revell, 1998. Gilts may be In summary, low body protein reserves may limit J .J. Eissen et al. Livestock Production Science 64 2000 147 –165 155 21 21 milk production and, therefore, voluntary feed intake estimated slope was 2 63 g day mm for aver- of sows during lactation. However, protein reserves age feed intake during the whole lactation, whereas 21 21 only seem a limiting factor for feed intake when the the slope was 2 95 g day mm when only the protein supply of the lactation diet is not optimal in average feed intake during week 1 of lactation was relation to the body composition of the sow. considered. These results are not surprising as differences in 3.1.6. Meal eating behaviour fatness between fat and lean sows get smaller during Dourmad 1993 and Weldon et al. 1994a the course of lactation due to the greater losses of studied meal eating behaviour during lactation of tissue of fat sows during the early phase of lactation sows that were fed close to ad libitum or restricted Le Cozler et al., 1999. during gestation, respectively. Weldon et al. 1994a found that ad libitum fed, and thus fatter sows, had a 3.1.8. Optimum body composition at farrowing lower daily feed intake during lactation by eating It is clear that the sow’s feed intake during fewer meals rather than smaller meals. In contrast, lactation is controlled in some way and that the Dourmad 1993 found that fat sows had smaller ingested nutrients are integrated with body reserves. meals, which were shorter in duration rather than The amount of body reserves at farrowing has an fewer meals, especially during the first 2 weeks of important influence on subsequent reproductive per- lactation. Both used a meal criterion of , 10 min to formance because it determines the extent that summarise information of feeding bouts into meals. reserves can be mobilized during lactation without Results of Weldon et al. 1994a suggest that gas- affecting the interval between weaning and sub- trointestinal signals mainly affect meal ending, sequent mating Mullan and Williams, 1989. High whereas results of Dourmad 1993 suggest that a body reserves at farrowing lead to a reduced vol- metabolic control was also partly involved. Both untary feed intake during lactation, as shown above, studies agree that fat sows use more time to absorb and excessive weight loss. Excessive weight loss has and utilise ingested nutrients before starting the next been associated with several common reproductive meal. This is reflected by the larger meal to meal problems as already mentioned in Section 1. Also, interval in case of a similar meal size Weldon et al., overfeeding during gestation is not recommended 1994a and a similar meal to meal interval in case of because of the increased frequency of farrowing smaller meals Dourmad, 1993. These results may problems in fat sows Dourmad et al., 1994. Low point towards a higher level of insulin resistance and feed allowances during gestation lead to an increased glucose intolerance of fat sows compared with lean voluntary feed intake and consequently less weight sows. loss of sows during lactation. The increase in vol- untary feed intake during lactation, however, gener- 3.1.7. Effect of body fatness at farrowing on ally does not compensate for the lower intake during voluntary lactation feed intake in relation to stage gestation. Therefore, feed intake over the whole of lactation cycle of gestation and lactation is reduced when a Revell et al. 1998a concluded that, during the normal lactation length of about 4 weeks is consid- first 2 weeks of lactation, voluntary feed intake ered e.g. Dourmad, 1991; Xue et al., 1997; Revell et mainly depends on body fatness whereas, during the al., 1998a. As a result, a low gestation feeding level latter phase of lactation, also other effects like decreases backfat thickness and body weight at protein content of the diet may effect voluntary feed weaning and tends to delay the return to oestrus after intake. Dourmad 1991 and Le Cozler et al. 1999 weaning, especially in high producing sows Dour- found that voluntary feed intake during the first part mad, 1991. of lactation was significantly affected by body Long-term performance of sows is best served by fatness at farrowing, whereas overall lactation feed minimising fluctuations in body weight and fat intake was not. Dourmad 1991 presented regres- reserves, so avoiding extremes of body condition and sion coefficients of average daily feed intake during subsequent poor performance Cole, 1982; Aherne different periods of the lactation on backfat thickness and Kirkwood, 1985. Chemical body composition at at farrowing, which illustrated the same effect. The farrowing should therefore be considered as an 156 J .J. Eissen et al. Livestock Production Science 64 2000 147 –165 optimum trait, taking the expected reproductive As litter size increased from 3 to 13 piglets, average daily feed intake of sows during lactation increased performance and feed intake during the following gradually by 0.6 kg, from 4.4 to 5.0 kg. Weaned litter lactation of sows into account Dourmad, 1991. For sizes of 14 and 15 piglets were not associated with a example, Yang et al. 1989 advised a target backfat higher lactational feed intake relative to litters having thickness P2 at first parturition of 20 mm. 7 to 13 piglets. O’Grady et al. 1985 estimated in a multiple regression analysis a linear 0.22 and a 3.2. Litter size 2 21 21 quadratic response 20.01 kg pig day for In response to greater suckling intensity, sows litter size, indicating a maximum feed intake at a nursing more piglets produce more milk Auldist and litter size of 14 piglets. According to the formulae of King, 1995; Toner et al., 1996; Auldist et al., 1998; O’Grady et al. 1985, daily feed intake increases by Revell et al., 1998b. Auldist et al. 1998 estimated 0.96 kg when litter size increases from 3 to 13 a significant linear relationship between milk yield piglets. The intercept A in the study of O’Grady et Y; kg day and litter size LS: 8, 10, 12 or 14 al. 1985 was arbitrarily taken as 5.2 to avoid an piglets: Y 5 5.98 1 0.689 3 LS and Y 5 8.20 1 overlap of data points of this study with other studies 0.324 3 LS for early day 10 to 14 and late day 24 in Figs. 3 and 4. Auldist et al. 1998 studied feed to 28 lactation, respectively. Toner et al. 1996 intake of sows nursing 6, 8, 10, 12 or 14 piglets and studied milk yield of sows nursing 6, 7, 8 or 10 did not find a linear relationship between feed intake piglets and also estimated a significant linear rela- and litter size. However, feed intake of sows nursing tionship between milk yield and litter size. Sows six piglets was lowest and feed intake of sows with a greater litter size and milk production have a nursing eight piglets was also lower than feed intake greater need to use energy and may, therefore, have a of sows nursing larger litters. It should be noted that larger voluntary feed intake Fig. 1. Auldist et al. 1998 limited daily feed intake to Fig. 3 shows relationships between lactation feed maximally 5 kg. intake and litter size. Litter size at weaning ranged Yang et al. 1989 used different gestation feeding from 3 to 15 piglets in a data set containing levels over four parities, resulting in lean and fat information of 19,393 litters Koketsu et al., 1996a. sows at farrowing. They studied ad libitum feed Fig. 3. Comparison of the relationships between lactation feed intake and litter size: j Koketsu et al., 1996a; d Auldist et al., 1998; m 2 O’Grady et al., 1985: lactation feed intake 5 A 1 0.224 3 LS 2 0.008 3 LS where A corresponds to the intercept and effects of other factors A arbitrarily taken as 5.2 and LS stands for litter size. J .J. Eissen et al. Livestock Production Science 64 2000 147 –165 157 Fig. 4. Comparison of the relationships between lactation feed intake and parity: j Koketsu et al., 1996a; d Mahan 1998; m O’Grady et 2 al., 1985; ♦, h Neil et al., 1996. Within the study of O’Grady et al., 1985: lactation feed intake 5 A 1 0.297 3 P 2 0.022 3 P where A corresponds to the intercept and effects of other factors A arbitrarily taken as 5.2 and P stands for parity. Within the study of Neil et al., 1996: ♦, sows received a simplified diet during the gestation period; h, sows received a conventional diet during the gestation period. intake maximally 7 kg day of sows nursing 6 or 10 compensate completely for the increased energy piglets but did not find a significant effect of litter demand. Dourmad 1991, using equations proposed size on voluntary feed intake. This was especially by Noblet and Etienne 1989, calculated that an clear in the lean sows. Fat sows at farrowing nursing increase in milk energy output of 1 MJ day induces 10 piglets, however, had a higher feed intake than fat a 45 g day increase in feed intake, which represents sows nursing six piglets at parities 1 to 4. This only about 40 of supplemental energy required for indicates that body condition at farrowing might milk production. Koketsu et al. 1996a used a affect a sow’s response in feed intake to increasing general rule of thumb ‘1.8 kg day plus 0.45 kg 21 21 litter sizes. piglet day ’, which has been recommended by In summary, voluntary feed intake of lactating Tokach and Dial 1992; cited by Koketsu et al., sows nursing relatively small litters increases with 1996a as a guideline for producers to use in feeding increasing litter size. Apparently, factors that limit lactating sows. Sows nursing small litters ,7 feed intake of sows nursing small litters can be piglets consumed more than suggested by the guide- overridden or terminated when litter size and there- line, but most sows consumed decidedly less. This fore milk production is increased. The increase in illustrates that high producing sows consume an feed intake, however, seems to be following a insufficient amount of feed to meet the energy diminishing increment-type pattern, indicating that needed to adequately support lactation Koketsu et limiting factors can only be overridden to a certain al., 1996a. extent and or other factors become limiting for sows nursing larger litters. 3.3. Parity Auldist et al. 1998 studied the effect of litter size on a sow’s losses of body weight and backfat Body weight increases with parity. Higher parity thickness during lactation and estimated significant heavier sows, therefore, have higher maintenance positive linear relationships. The increase in vol- requirements during lactation and may be expected untary feed intake of sows nursing six to 14 piglets to consume more feed Fig. 1. On the other hand, was for any increase in litter size inadequate to lower parity sows are still not fully grown and have 158 J .J. Eissen et al. Livestock Production Science 64 2000 147 –165 larger energy and protein requirements for body shows the relationships between lactation feed intake growth than higher parity sows. Parity number may, and parity in four studies. Again, the intercept A in therefore, affect the partitioning of energy and or the study of O’Grady et al. 1985 was taken as 5.2, protein between maternal tissue and milk during to avoid overlap of data points of this study with the lactation. This is supported by Pluske et al. 1998 others. O’Grady et al. 1985 estimated in a multiple 21 who made primiparous sows anabolic during lacta- regression analysis a linear 0.30 kg parity number 21 tion by super-alimentation provide sows with nutri- day and a quadratic response 20.02 kg parity 2 21 21 tion via stomach cannulae to achieve a feed intake of number day for parity, with a maximum feed about 125 of ad libitum. They concluded that intake between parity 6 and 7. According to this primiparous sows seem to partition extra energy into formula, daily feed intake of sows increases by 0.73 body growth rather than milk production, whereas kg from parity 1 to 7. Koketsu et al. 1996a found a multiparous sows show an increase in milk yield significant lower feed intake for primiparous sows Matzat et al., 1990; cited by Pluske et al., 1998. than for older sows in their study about factors Results presented by Cole 1990 suggest that affecting lactation feed intake on commercial herds. primiparous sows need a higher dietary energy intake Feed intake increased by 0.81 kg, from 4.51 kg to avoid maternal losses in live weight and condition parity 1 to 5.32 kg parity 9. Mahan 1998 found during lactation than second parity sows. This would a significant quadratic increase in weekly and total indicate that primiparous sows have higher mainte- lactation feed intake by parity parities 1 to 5; nance requirements. In particular, primiparous sows however, litter size during lactation was linearly are of relatively low absolute size in relation to their affected by parity in this study. Mahan 1998 also productivity and nutrient demands Whittemore, found an effect of parity that was related to protein 1996, which may cause gastrointestinal limitations content of the diet. A higher protein content in the to be more severe for primiparous than for higher diet increased voluntary feed intake of primiparous parity sows. For example, Sinclair et al. 1996 sows during the full lactation, whereas this was not recorded feed intake of primiparous and parity three the case for multiparous sows. This may be a sows. Sows of both parities had a similar feed intake reflection of higher maternal protein requirements of during the first 2 weeks of lactation; however, primiparous sows; however, lower protein body primiparous sows had a lower feed intake during reserves of primiparous sows may also play a role. weeks 3 to 5. Feed intake increased significantly with increasing Differences in milk yield have been found be- parity in the study of Neil et al. 1996. They did not tween parities. However, experiments in which clear present information about levels of litter size; how- conclusions are drawn are scarce, since estimates ever, they mentioned that litter size increased with must be made during successive cycles. Summa- parity and that ad libitum feed intake of sows was rising the results presented by Etienne et al. 1998, not affected by litter size. milk production increases from the first to second O’Grady et al. 1985 and Koketsu et al. 1996a parity, reaches a maximum and is similar from the concluded that the gradual increase in feed intake second to fourth lactation and slowly decreases that occurs with advancing parity seems to be afterwards. Differences in milk production between consistent with the increase in maintenance energy parities, however, also partly reflect differences in requirements associated with age-related increases in litter size. According to Ferreira et al. 1988 and body weight. Three other studies also had body Vanschoubroek and Van Spaendonck 1966; cited by weight and backfat thickness recordings of sows. Etienne et al., 1998, second parity sows produced Primiparous sows lost significantly Mahan, 1998 or about 11 and 26 more milk, respectively, than numerically Neil et al., 1996; Sinclair et al., 1996 primiparous sows. more body weight and numerically more backfat NRC 1987 summarised voluntary feed intake thickness Neil et al., 1996; Sinclair et al., 1996 records from many sources and reported an average during lactation compared with multiparous sows. daily feed intake during lactation of 5.17 kg per day The latter results would indicate that the energy per sow, with gilts consuming 15 less. Fig. 4 intake increases more than energy requirements for J .J. Eissen et al. Livestock Production Science 64 2000 147 –165 159 maintenance and milk production of sows with intake, despite a leaner body composition, seems to increasing parity. Differences in glucose tolerance be contradictory to the previous conclusion that, between lower and higher parity sows, however, may within genotype, lean sows have a higher feed also play a role. Kemp et al. 1996 orally adminis- intake. An explanation for this apparent paradox tered glucose after an overnight fast at day 105 of could be that feed intake during lactation is depend- gestation to sows of different parities. This resulted ing on the difference between the actual body fatness in significant higher blood peak levels and area under of a sow at farrowing, which mainly depends on the the curve up to 75 min after administration of feeding strategy during gestation, and the potential glucose for lower parity sows compared with higher body fatness at farrowing, which depends on the parity sows, indicating that lower parity sows may be genotype. The smaller the difference between actual more glucose intolerant or insulin resistant. and potential body fatness, the lower the feed intake of a sow during lactation would be. This would mean 3.4. Genotype that the sows selected for a low daily feed intake were indeed absolutely leaner, but, compared to their 3.4.1. Selection genetic potential for body fatness, relatively fatter In several species, feed intake data of lactating than the sows selected for a high daily feed intake. animals were recorded to estimate genetic parameters The results of Kerr and Cameron 1996b are in for voluntary feed intake. In a small data set, Van Erp agreement with the estimated genetic correlation et al. 1998 estimated a heritability of 0.19 for between voluntary daily feed intake during the voluntary feed intake of lactating sows. In dairy growth phase and during lactation for pigs r 5 g cattle, Van Arendonk et al. 1991 estimated a 0.92 60.50; Van Erp et al., 1998. Archer et al. heritability of 0.46 for dry matter intake, and Koenen 1998 estimated a genetic correlation of 0.51 be- and Veerkamp 1998 and Van Elzakker and Van tween post weaning voluntary feed intake and feed Arendonk 1993 estimated heritabilities varying intake at maturity in mice. from 0.18 to 0.37 and 0.18 to 0.42, respectively, Kerr and Cameron 1996a estimated genetic depending on the stage of lactation. These results relationships between performance traits, measured show that voluntary feed intake of lactating animals during the growth phase from 30 to 85 kg of body is a heritable trait which can be changed by selec- weight, and reproduction traits. Genetic correlations tion. of average daily feed intake and daily gain during the Kerr and Cameron 1996b reported substantial growth phase with litter weight at weaning were 0.42 variation in feed intake during lactation between and 0.52, respectively. This suggests that gilts select- primiparous sows of Large White lines after seven ed for a high daily feed intake or daily gain during generations of divergent selection for lean growth the growth phase exhibit an increase in milk pro- rate, lean feed conversion or daily feed intake during duction during lactation, which may affect the vol- the growth phase from 30–85 kg of body weight untary feed intake of these gilts during lactation. under ad libitum feed intake Cameron and Curran, Kerr and Cameron 1995, 1996b reported a reduced 1994. Lactation feed intake was lowest for sows of daily gain as a correlated response of piglets the low lean growth rate line and highest for sows of suckling primiparous sows selected for low daily the high daily feed intake and the low lean feed feed intake during the growth phase after five and conversion line. Sows selected for low daily feed seven generations of selection. Piglets suckling intake during the growth phase consumed signifi- primiparous sows selected for high daily feed intake cantly less feed during lactation than sows selected did not grow any faster than piglets of unselected for high daily feed intake Kerr and Cameron, control sows Kerr and Cameron, 1995 or grew only 1996b. Body weight prior to farrowing and litter faster during the second part of lactation Kerr and size during lactation were not different, whereas Cameron, 1996b. These results suggest that a low backfat thickness prior to farrowing was lower for feed intake of sows can inhibit milk yield rather than the low daily feed intake line. The lower feed intake that a high feed intake can enhance milk yield during lactation of sows selected for low daily feed Mackenzie and Revell, 1998. 160 J .J. Eissen et al. Livestock Production Science 64 2000 147 –165 In conclusion, voluntary feed intake of sows general, pigs of the high gain lines have a higher during lactation can directly be changed by selection. daily feed intake than pigs of the low gain or control In practice, feed intake during lactation may indirect- lines. Clutter et al. 1998 found that the high weight ly be changed by selection for production traits, gain line had a significantly lower concentration of which may also affect litter performance. the putative satiety hormone cholecystokinin per unit of feed consumed compared with the low line. This 3.4.2. Background of genetic differences in supports the hypothesis that cholecystokinin may lactation feed intake play a role in genetic differences between lines for In general, genetic differences in lactation feed feed intake. Norton et al. 1989 reported higher intake of sows will, to some extent, reflect differ- blood glucose and insulin concentrations for a high ences in body weight and composition at farrowing, weight gain line whereas blood growth hormone and and in litter size and milk production during lacta- NEFA concentrations were higher for a low line. In tion. Meishan synthetic sows consumed significantly contrast, Arbona et al. 1988 showed greater basal more feed than Large White and Landrace sows blood growth hormone concentrations for pigs select- Sinclair et al., 1998. In that study, Meishan sows ed for high weight gain than for control pigs. Clutter were significantly lighter, had more backfat at day 1 et al. 1995 reported greater concentrations of of lactation and had a larger litter during lactation. circulating IGF-I for high weight gain line pigs Grandhi 1997 did two experiments with Hamshire compared with low line pigs throughout periods of and Yorkshire sows. During one experiment, Ham- feed deprivation and refeeding. Altogether, these shire sows ate significantly more and, in the second results indicate that selection for post weaning experiment, Hamshire sows tended to eat more. weight gain resulted in concomitant changes in These breed differences could at least partly be due endocrine and metabolic status of growing pigs. to the higher body weight of the Hamshire sows It is generally assumed that feed intake regulatory Grandhi, 1997. Landrace sows produced milk mechanisms are similar for mammalian species and significantly higher in protein content than Duroc chickens, although there are also some differences sows, whereas fat content was not different Shurson Barbato, 1994. Chickens selected for a high gain and Irvin, 1992, which may have contributed to the voluntarily consumed a volume of feed approaching higher lactation feed intake of Landrace sows in their the full capacity of their gastrointestinal tract, where- study. as chickens selected for a low daily gain consumed a Differences in glucose tolerance and insulin resist- small percentage of total capacity Barbato et al., ance between lines of sows may also affect voluntary 1984. Selection for body weight gain increased feed intake. Sows of a dam line had significantly villus surface area by 20-fold due to increases in higher peak levels of glucose and a higher area under crypt size and enterocyte migration rates Smith et the curve after oral administration of glucose in late al., 1990; cited by Barbato, 1994. The latter results, gestation compared with sows of a sire line Kemp et however, could also be due to the positive relation- al., 1996. This may point towards a higher level of ship between feed intake and villus surface area glucose intolerance and or insulin resistance of the Goodlad et al., 1987; cited by Barbato, 1994. dam line sows. Breed may also affect the partitioning O’Sullivan et al. 1992 found that high gain line of nutrients between maternal growth and lactation chickens had significantly higher levels of the en- Sinclair et al., 1996, 1998 and some breeds may be zyme trypsin than low gain line chickens of similar able to withstand heat stress more effectively than body weight or age. Furthermore, data reported by others Forbes, 1995. Barbato 1994 suggest a genetic basis for CNS Feed intake regulatory mechanisms that affect neurotransmitter levels, which seems to be related to voluntary feed intake during early life may, at least selection for weight gain. partly, also affect voluntary feed intake of mature In summary, results presented above show that animals. A number of studies presented results of differences between breeds or lines of breeds in blood parameters of growing pigs selected for high voluntary feed intake of lactating sows may involve or low daily gain during the growth phase. In a large number of factors in addition to body weight J .J. Eissen et al. Livestock Production Science 64 2000 147 –165 161 and body condition of sows and litter size milk at farrowing and genotype are the two sow factors production during lactation. Though there is a dearth that best can be used to increase feed intake capacity of information regarding the genetics of central and during lactation. Of these, genotype seems the most peripheral feed intake control mechanisms, it seems appropriate as body composition at farrowing should that mechanisms may act at the pre- and post- be considered as an optimum trait and, from this absorptive level, which shows that each factor men- point of view, voluntary feed intake during lactation tioned in Fig. 1 could be involved. An increase in should not be maximised per se, for example, by voluntary feed intake by selection is likely accom- reducing feeding level during gestation. These re- plished by cancelling or reducing effects of the most sults, therefore, suggest that voluntary feed intake limiting factors, whereas a decrease is likely ac- during lactation should be included in breeding complished by introducing new or intensifying ef- programmes. A higher feed intake during lactation fects of existing limiting factors. may be accomplished by direct selection for lactation feed intake or indirect selection, for example, for daily gain or daily feed intake during the growth

4. Conclusions and implications phase.