The energy credit: how much can a probiotic really give back?

Because those are not the same thing, and the difference is a matrix.

An enzyme comes with one. A phytase carries a credit tied to the phytate it destroys; a xylanase carries one tied to the arabinoxylan it breaks down. There is a substrate, a dose response and a number the supplier will stand behind, so the formulator takes the credit, cuts the specification, and banks the saving. Feed is 60 to 70% of the cost of producing a broiler, so this is where the money is.

Probiotics have largely stayed outside that arrangement.

It is logical enough that they should not. A probiotic is, by definition, a live organism, and a live organism is not a fixed quantity in the way an enzyme is. A probiotic has to survive pelleting, germinate, colonise, and then compete with whatever is already living there. Every one of those steps is contingent on the strain, the dose, the diet and the challenge in the house. The response is variable by nature.

Which leaves an obvious question, largely unasked: if a credit is taken anyway, how many kilocalories does a probiotic actually give back?

The mechanism is real, and worth following, because it explains almost everything that follows. The claim on a technical sheet is usually some version of one idea: that the additive helps the bird express more energy from the same diet, putting a few otherwise-wasted kilocalories back on the credit side of the formula.

The caecum is the principal site of microbial fermentation of dietary fibre in the chicken. Fibre that the bird's own enzymes cannot touch is broken down there by a specialised community of bacteria: Bacteroides and Prevotella act as generalists, able to attack a wide range of plant polysaccharides, while Ruminococcus, Fibrobacter, Clostridium and Roseburia are the cellulolytic specialists. Ruminococcus degrades fibre to simple sugars using cellulase and the cellulosome, a surface-bound enzyme complex.⁽²'³⁾

The end products are short-chain fatty acids, principally acetate, propionate and butyrate. Acetate is absorbed and used by the liver to synthesise fat, propionate supplies gluconeogenesis (the manufacture of glucose), and butyrate is the main fuel of the gut lining itself.⁽²⁾ Through this the bird valorises part of the fibre fraction and recovers energy its own enzymes could not reach.

The amount it recovers is what a matrix credit is meant to capture.

Fermentation can only ever add so much energy, because a chicken is poorly built for it. A pig ferments in a large hindgut where feed lingers for hours, and the short-chain fatty acids produced there may supply 30 to 76% of its basal energy needs.⁽⁶⁾ A chicken is different. Feed clears its gut in three or four hours, against around twenty in a human, and fermentation is limited to the caeca, two small blind pouches that draw in only the finest, most fluid material while the bulk of the feed passes them by.⁽⁴'⁵⁾

The upshot is that fermentation meets only an estimated 5 to 30% of a bird's energy needs,⁽⁴⁾ and where a flock falls in that range depends largely on the diet: towards the top on a fibrous ration, towards the bottom on a modern corn-soy diet that is low in fibre and high in energy. The energy a probiotic can valorise is, in a broiler, modest from the outset.

The wider performance picture is modestly positive. Pooled across the broiler literature, Bacillus supplementation improves feed conversion efficiency by a standardised mean difference of around −0.33 and daily gain by about 0.37 g per bird per day, though with heterogeneity and some evidence of publication bias.⁽⁹⁾

But none of that tells a formulator what to put in the matrix. To price a matrix you need energy retention itself, measured as AMEn, the energy the bird keeps once losses in the droppings are subtracted. Yet few trials report AMEn for every diet they test, which is precisely what is needed to value a credit.

A study by Nusairat and Wang, published in Frontiers in Veterinary Science, did.⁽¹⁾

The design is set out in figure 1. Six corn-soy diets were compared in 2,496 Ross 708 broilers grown to 42 days: a standard positive control, a negative control formulated 130 kcal/kg below it, and four supplemented arms built on that negative control.

The probiotic was included at 100 g per tonne, delivering 10⁵ CFU per gram of feed, which is 10⁸ CFU per kilogram. That is a commercially realistic dose, not a research-scale one, and both the xylanase activity and the Bacillus counts were confirmed by analysing the finished feed.

One detail of the design becomes critical later. The 130 kcal energy reduction was achieved by partially replacing corn with soybean hulls, which raised the fibre content of the reduced-energy diets.

The probiotic alone returned 21 kcal/kg, and 22 kcal/kg in the same measurement at 21 days. Both were statistically significant, and remarkably consistent.

The birds agreed. Bodyweight gain to 42 days was 2,753 g on the probiotic against 2,730 g on the unsupplemented reduced-energy diet, a difference that was not significant. Only the antibiotic and the xylanase-probiotic combination produced birds heavier than the negative control, improving feed conversion by around 3 points.

Notice also that the enzyme and the probiotic together returned 48 kcal, not the 62 their separate results would predict.

Commenting on their findings, the researchers concluded that a blend of xylanase and Bacillus spp. "can be used as an alternative to antibiotic growth promoters in broiler diets even when formulated at ~4.0% less ME than a standard diet".

Some caution is warranted before building anything on a single trial. It was funded by an additive manufacturer, with one of the authors employed by the company; the conflict is properly declared, but it is worth knowing. And its diets were formulated to NRC (1994) requirements, a standard that the tenth revised edition has now superseded.

Nevertheless, the measurement at the heart of it is a straightforward one, it was taken at two separate ages, and it agreed with itself both times. So it is worth taking seriously.

Start with the arithmetic, because it needs no assumptions at all.

Vietnam is a useful test case for the high dependence of its feed raw materials⁽¹²⁾. At July 2026 prices, compound broiler feed runs at $381 per tonne, a live white-feather broiler fetches $1.03 per kg, and a probiotic costs about $2.00 per tonne of feed.

On that feed, 1 kcal/kg of energy credit is worth 12.5 cents per tonne. Which gives the number that ought to be on every matrix sheet and never is:

The whole business case therefore lives in a window between 16 and 21 kcal/kg. Below it, the additive does not cover its own cost. Above it, the difference is not coming from the bug. It is coming from the bird.

The trial priced exactly one case, and it is worth taking on its own terms rather than extrapolating from.

Consider what happened to the birds that were given the probiotic and had the energy taken away anyway. Over 42 days, those on the full-specification diet gained 2,829 g. Those on the 130 kcal-reduced diet, with the probiotic supplying what it could, gained 2,753 g. That is a shortfall of 76 g, or close to 3% of the finished bird. And because feed consumption did not differ across any of the six treatments, these birds ate precisely as much as their better-fed counterparts. They simply turned it into less meat.

The same fact shows up in the figure a mill actually watches. Feed conversion over 42 days was 1.677 on the full-specification diet and 1.722 on the reduced-energy diet with the probiotic, a deterioration of four and a half points.

Scale it to the tonne. A tonne of the full-specification feed produced 596 kg of bodyweight gain. A tonne of the reduced-energy feed, probiotic included, produced 581 kg. The credit cost 15 kg of liveweight for every tonne of feed made.

Take only what the probiotic gives you, and the additive pays for itself with 62 cents a tonne to spare, and the bird never notices. Draw 130 kcal against it and you are $1.79 a tonne worse off, because kilo for kilo the bird is worth 2.7 times the feed. Kilos of feed are cheap. Kilos of bird are not.

What a credit of 1.6%, or 49 kcal/kg, would cost, this trial cannot say, and drawing a straight line between the two columns would be guesswork. What it does say is that such a credit is more than double what the probiotic returns, and the difference has to come from somewhere. The pressure to take it tends to peak when raw material prices do, which is exactly when a lighter bird hurts most.

Take it in a fibrous diet, not a fatty one. The trial removed energy by adding soybean hulls, so it raised the very fibre the probiotic ferments. A commercial matrix is spent the other way, by pulling out fat. So 21 kcal/kg is probably a ceiling, and a high-fat, low-fibre formula will return less.

Do not stack it on an enzyme's credit. The xylanase and the probiotic returned 48 kcal together, not the 62 their separate results predict. Adding a probiotic credit on top of an NSPase matrix books the same energy twice.

Check where your specification starts. Ross 308 puts the grower at 3,100 kcal/kg.⁽¹¹⁾ A grower already running 3,028 has 72 kcal less room before it takes anything at all.

You are buying AMEn, not net energy. That figure ignores the heat a bird spends on each nutrient, and fat is the efficient source while fermentation is among the least. Swap one for the other and the number holds while the bird has less to spend.⁽⁸⁾

On this evidence a probiotic will hand you back around twenty kilocalories, in a diet that gives it something to ferment. Add an unpredictable challenge or simply take more than that, and you are not banking a saving. You are borrowing it from the bird.