What 90% of swine professionals do not understand about gestating sow feeding

Today I want to go over what looks like 90% of swine professionals do not understand about gestating sow feeding (and even the NRC got it wrong, and I respect the NRC for most things).

If you are in a hurry, here is the summary:

Most swine professionals do not understand that the sow changes it's priorities throughout the gestation period. In early gestation, the priority is herself; thus, she will lose her pregnancy and return to estrus if anything undesirable happens. However, in late gestation, things change a lot:

1) Now that the fetuses are well developed, the gilt or sow very altruistically prioritize the fetuses instead of themselves.

2) If you increase the feed allowance in general (a.k.a. bump-feeding) or if you increase the nutrient density specifically (50% more amino acids or energy, for instance), little to nothing will happen. Well, way more negatives than positives will happen - more on that later.

3) The only exception is if she is thin, but if she is thin on day 90, she has an underlying disease or major body condition scoring issues and training to be corrected by the staff. This prioritization in late gestation is very much like in lactation, where the priority is the litter. She loses body condition at the expense of producing milk for her litter.

Now, let's deep dive into the above concept...

For the last several decades, the number of total born piglets increased significantly. As birth weight went down, people tried to fix that through diet, which clearly made our herds fat and didn't fix the situation. 

As fetal growth increases exponentially in late gestation, many folks think the amount of feed should increase. That's not the case for several reasons. To begin with, the sow prioritizes the litter in late gestation, and we can make an analogy to the lactating sow: she also prioritizes the litter. The sows are going to lose body condition to focus on their litters. That's exactly the same thing in late gestation for the sow, which is different from early gestation where if the sow is in a negative energy or protein balancing, she will likely return to estrus. 

There's a quote by Sherlock Holmes that says:

"It is a capital mistake to theorize before one has data. Insensibly one begins to twist facts to suit theories, instead of theories to suit facts​."

It applies to the situation where we tried to look for the obvious, let's get more feed, where when you look at the data, it's straightforward that's not the case.

Buitrago et al. (1974) fed 700 grams per day of an energy equivalent to a corn-soybean meal diet throughout the whole gestation, compared with 1.5 kg a day and 2.4 kg. The 2.4 and 1.5 kg were enough to keep the birth weight of the pigs. Seven hundred grams a day, as you could guess, was not enough. So we know that 700 grams per day are low.

Ezekwe and Opoku (1988) did a study where they fasted the sows for up to two weeks before farrowing to see these effects of prioritization. They had a control diet of 1.8 kg a day, seven-day fasting, and 14-days fasting. They gave vitamins and minerals during this fasting, but no proteins and energy. There was no statistical difference in birth weight across the treatments. That is just one more example of the prioritization of the sows. Let's emphasize that for obvious welfare reasons, we do not recommend anyone trying.

Then there are many studies over the years focusing on increasing the amount of feed during the early, middle, or late gestation, based on the development of the muscle fibers. Dwyer et al. (1994), Nissen et al. (2003), Bee (2004), Heyer (2004), and Lawlor et al. (2007) didn't find effects on increasing the amount of feed.

In late gestation, there are Cromwell et al. (1989), Miller et al. (2000), Nick Shelton et al. (2009), Jose Soto et al. (2011). There was my large-scale study as well in 2015. More recently, Mallmann et al. has done studies in this area as well. Some of these studies improved birth weight by about 28 grams. However, that's only 2% of the birth weight, and it was not consistent.

Then, the other argument for increasing the feeding in late gestation is: there's the mammary complex development, we need to give more feed. It turns out that several studies didn't find this was helpful. The more feed you give, the more fat cells are in the mammary complex, which reduces the amount of colostrum and reduces the amount of feed intake in lactation.

Amdi et al. (2014) looked at thin gilts vs. fat gilts. Thin gilts were defined as an average of 12 millimeters of back fat. Fat gilts were defined as an average of 19 millimeters of back fat. They had a diet of 3.1 Mcal of Metabolizable Energy per kg. So basically, they fed 1.8, 2.5, and 3.5 kg a day. The conclusion was if the gilt is thin, which is 12 millimeters, you need to provide more feed. But if the gilt is fat, and 19 millimeters is likely fat for a gilt, any of those levels are fine.

As we look at the overall literature, I would say that based on the recent data, we don't need to bump feed the gilts either, even though one or two studies in the past saw a little bit of improvement in the gilts. Today, we don't have that strong evidence anymore. Gladly, genetic companies started updating their recommendations.

Then a swine nutritionist would say…

"Hey, it would be awesome if we could feed the sows every day a different amino acid and energy, right?"

Well...

The University of Guelph did a few studies on this area, where they fed either:

1) A control diet which was one level throughout the whole gestation period, and that is 2.2 kg per day of a diet with 2.5 Mcal of net energy per kg, and 0.56% digestible lysine, or

2) What they called the "precision feeding" program, which changed the amount of amino acid energy every day as the NRC 2012 suggests.

It turns out there were no meaningful differences in the results. For example, when you practice precision feeding, they gain more weight, which makes sense. As you give more amino acid energy, the sows will gain more weight, not the fetuses. However, I believe that precision feeding is valid and economical when you want to segregate the diets by parity groups.

Ren et al. (2015) documented, one more time, that the more weight a female gains during gestation, the more weight they lose during lactation. Then in 2015, as part of my Ph.D. study at K-State, we used 1,105 females. The total born, or litter size, was 14.5 on average, and we had about 16,000 baby pigs weighed at birth. Then we did a factorial design where we had gilts and sows, 11 grams of lysine a day vs. 20 grams of digestible lysine a day, and then 4.5 Mcal of net energy per day vs. 6.75. For a corn-soy diet, that was equivalent to 1.8 kg of feed a day vs. 2.7, just so you can translate that into feed amount.

When we did that, as we looked at different parameters, when we increased the amount of amino acids, the birth weight of the pigs didn't change. When we increased the amount of energy, it increased by 30 grams, which is questionable. I hypothesize that these 30 grams will not be translated into a lifetime performance because these 30 grams are likely glycogen reserves that would quickly burn right after birth. Hence, people need to be very careful when they start extrapolating this information. We also did an economic analysis at the time. Even if that slight improvement in birth weight had a positive impact on the pig all the way to the finishing phase, it wouldn't pay back because of the amount of feed you had to give to the sow. Then something that we saw that was obvious, and several other studies observed as well, is that when we gave more feed, there was an increased amount of stillborn. More specifically, it was coming from the energy. That was the reason... fat sows have more stillborn as you would expect.

Then more recently, Thomas et al. (2018) did a fascinating study where they fed 11 g of digestible lysine per day, 13.5 g, 16 g, or 18.5 g, throughout the whole gestation period. The authors also had gilts and sows, and overall, they gain weight as you'd expect whn you increase amino acid density. They had a considerable litter size, between 15 and 16 overall for the herd. The treatments didn't affect total born, but it's evident that this whole conversation holds true even with today's litter size. Interestingly, they observed that when you increase the amount of dietary amino acids in the sows, not in the gilts, there was a slight reduction in stillborn. In the first study, more energy was negative, and in this study, more amino acid was positive. These studies seem complementary, meaning, when we make our sows fatter, it is harmful. If we make them "stronger" (i.e., more muscular), it helps in this case, but now it comes back to the economics. Thomas et al. (2018) also looked at the number of pigs weaned by treatment, and there was no difference.

A review from 2010 by Rozeboom reported a study that fed 1.8, 2.3, or 2.7 kg of feed per day throughout the whole gestation...

Guess what happens in lactation when you feed this amount of feed during gestation?

The more feed you give, the less feed they eat in lactation.

If we look at the economics of bump feeding, we're talking about $5-10 per sow per year, and that's a substantial amount of money. We're not even adding the negative effects of bump feeding on reducing lactation feeding intake and increasing stillborn rate.

Many of you are familiar with the Caliper concept developed by Mark Knauer from North Carolina State University. I like it because it's a very objective and fast measurement of the body condition of the sows, and that's probably the take-home message...

We can argue about all these things on sow feeding, but the question is: 

Are the sows fat? If they are fat, they need less feed.

The other thing that Mark Knauer shows is that if the sow is thin, where you can see the backbones, it's not good for performance. We know that.

Even my mom knows that. 

Now, if the sow is fat, that's where many people miss the boat sometimes. It's terrible for performance, pre-weaning mortality, lactation intake, to list a few.

A very interesting meta-analysis conducted by Kim et al. (2015) looked at back fat from many studies and multiple parameters. Looking at litter weight gain, it was evident that if the back fat is between 12 and 22 millimeters, that is good, meaning there is a wide range to work on these sows. However, if you go above 22 and below 12 mm, that's not good.

In another interesting study conducted by Mallmann et al. (2019), the authors fed 1.8, 2.3, 2.8, or 3.3 kg per day of a corn-soy diet. The more feed provided, the more weight the sows put as body weight, and then colostrum production, they were able to measure, and with increased feed allowance, colostrum production went down. The more feed they fed the sows, the less feed the sows consumed in lactation - I think you have seen that before haven't you?

Then, what was interesting was that Mallmann et al. (2019) followed them through four parities. They conducted the study on the first cycle of these gilts, close to a thousand gilts and 15,000 piglets weighed at birth. The higher the feed allowance, the lower the retention rate of these sows. These effects would very likely be magnified if he had fed this treatment throughout all four production cycles. Also, a lot of people say, "most of these studies are only one cycle." Even though several are one to two cycles, this study went all the way there and made it clear and closed this chapter in this discussion.

Then I think people often forget that "half" of the genetic makeup of the finishing pig comes from the "mom". The sow is being selected for feed efficiency and growth. That's why many farm managers recently have said, "I need to give less feed than in the past; otherwise, they get fat very quickly." So it's super interesting.

Then also, recently, a few genetic companies started to select for birth weight, which is great because the genetic companies will "fix" the birth weight problem since nutrition never did. At the same time, the nutritionist, production, and sow farm managers will bring those sows that are fat back to the ideal condition.

Then people ask, "Is there a major difference across genetic lines?"

The way I look at it is: first, there are only one or two studies with more than one genetic line in the same experiment - the Mallmann study was one of them. There were no differences between the two genetic lines tested.

And also, when you look at genetic lines, because it's super hard to have several genetic lines in the same study let alone conduct sow research at the correct sample size. We can talk about all these studies, and then someone shows up with a research paper with 10, 20, or 30 sows per treatment. We cannot look at that paper because we need 200, 300 sows per treatment to have a meaningful conclusion to production relevant parameters. Since a 4 x 4 factorial study between genetic lines and dietary treatments will likely not happen in the short term, we have to get what is available and run with it.

But if you look at a finishing pigs' difference between genetic lines...

Let's suppose that you see a 5 to 10% difference in finishing feed efficiency across the genetic lines out there. So with that number in mind, you can only assume that the difference from the sow will be around that number. So if one genetic line needs, let's say, 1.8 to 2 kg a day of a corn-soy diet to maintain that body weight or perform well, then that means a 10% difference it's 2.2 kg: from 2 kg to 2.2 kg per day during gestation is the range across genetic lines. But we see the genetic supplier's recommendations range from 2 to 3.5 kg per day during gestation, which is crazy. It's completely crazy.

One thing that we should take into consideration is how much importance is carried by a factorial model vs. an empiric model. Factorial is when you break down the protein pools and bring nutrition to its core components, which is fantastic. However, a lot of times, it may not replicate reality. So we need to stay open. In my view, we need to base most of our decisions on empirical data, meaning randomized large-scale commercial studies. Then, at a smaller level, look at the factorial approach to see how things compare, hypothesize, and validate with empiric research.

That is all I had. Thoughts?

Dr. Marcio Gonçalves

p.s. a few times a year, we open a few seats to the Elite Swine Nutritionist Program. To join the waitlist go to www.EliteSwineNutritionist.com