Three peer-reviewed trials report that oxygenated drinking water improves feed conversion in broilers. The largest and best-powered trial reports that it does not.

Both statements are accurate. The tension between them is the most commercially useful information in this literature, because it indicates when a growth benefit should be expected — and when a producer is buying something else instead. Most supplier material cites the three positive trials and omits the fourth. This page works through all four, why they diverge, what survived the null result, and what a producer should conclude.

This is the poultry-specific companion to our broader review of oxygenated drinking water for poultry and livestock, which also covers pigs, cattle and the underlying mechanisms.

1. The state of the evidence

Four Broiler Trials, One Apparent Contradiction

Broilers have the strongest evidence base of any species in this field. Shin et al. (2016), a three-arm randomised trial in Poultry Science, found feed conversion improved by up to 11.4%, body weight gain up, abdominal fat down, and IgG/IgM concentrations elevated, all at P<0.05. Two independent Egyptian groups then replicated the direction: Abdel-Baky et al. (2023) and El-Deeb et al. (2025) both found body weight and feed conversion improvements alongside enhanced antioxidant capacity, with El-Deeb additionally reporting significantly improved antibody titres against Newcastle disease and avian influenza.

Three controlled trials, three separate groups, consistent direction. In most contexts that is a reasonable evidence base. Then the largest study in the literature reported a null result on exactly the endpoint the other three had moved.

840
Ross 308 broilers in Khattak et al. (2025) — the largest trial in this literature
12
Replicate pens per treatment, under commercially representative housing at Scotland's Rural College
19–22%
By which both groups already exceeded Ross 308 breed performance standards
No
Significant difference in feed conversion ratio or final body weight between oxygenated and tap water

That is a real null result on the primary growth endpoints, not a discrepancy to be explained away. It deserves to be reported as prominently as the three positive trials — which, in supplier material, it generally is not.

2. Reconciling them

Why the Trials Disagree

Three explanations are compatible with the published data. They are not mutually exclusive, and they carry very different commercial implications, so it is worth separating them rather than settling on whichever one suits the pitch.

01

Baseline headroom

The reading we find most defensible. If a flock is already performing 19–22% above breed standard, there is very little growth left for any intervention to capture, whatever the underlying physiology does. Under this explanation the growth benefit is real but conditional on the baseline being oxygen-limited in the first place.

02

Statistical power and publication

Smaller trials overstate effect sizes when they reach significance at all, and null results from small trials are less likely to be written up at all. The three positive results could therefore be inflated relative to the true effect. Note that this explanation predicts a real but smaller benefit than 11.4% — not the absence of one.

03

Context and delivered dose

The trials were run in different production environments and did not necessarily administer the same thing. Heat stress, stocking density, ventilation and water temperature all bear on how oxygen-limited a bird is; and where dissolved oxygen is not verified at the drinker line rather than at the point of treatment, apparent disagreement between trials may partly be disagreement about the dose actually delivered.

3. What survived

The Ceiling Effect — and What Held Up Under It

A null result on growth is not a null result on everything

In the same study, under the same already-optimal conditions that left no growth headroom, welfare and carcass-quality endpoints separated clearly between treatments. That combination is what makes the ceiling-effect reading more defensible than simply concluding the intervention does nothing: a treatment that did nothing would not be expected to move hock burn, feather condition, fat deposition and muscle myopathy prevalence at the same time.

It also inverts the usual sales argument. The endpoints that held up under the hardest test are not the ones a producer is typically sold on.

Endpoint in Khattak et al. (2025) Result Commercial character
Feed conversion ratioNo significant differenceGrowth
Final body weightNo significant differenceGrowth
Hock burn37.3% of birds lesion-free vs. 12.3% in controlsAudited welfare
Feather condition scoreSignificantly improvedAudited welfare
Breast cleanlinessSignificantly improvedAudited welfare
Abdominal fat−12%Carcass
Thigh yield+2.6%Yield
White striping prevalenceReducedMeat quality
Gut microbiome compositionDifferential signatures on metagenomic sequencingMechanistic

Hock burn and feather condition are audited welfare indicators under most retailer and assurance schemes, which gives them direct commercial consequence independent of any growth effect.

A disclosure that belongs with this study

One co-author on Khattak et al. (2025), R. Pearson, is affiliated with Oxcel, a commercial nanobubble water company whose unpublished trial data is referenced elsewhere in this field. The study was peer-reviewed and published in Poultry Science, and a null result on the primary growth endpoints is not the outcome a commercial sponsor would choose — which, if anything, strengthens the finding rather than weakening it. But the affiliation is rarely mentioned in summaries of the paper, and a reader is entitled to know it and check the methods themselves.

4. For a buyer

What This Means If You Are Evaluating a System

The question worth asking is not whether oxygenated water works. It is which of the two benefit classes you are actually buying, and whether your operation has the headroom for the one you are being sold.

Where a growth benefit is more plausible

  • Flocks performing at or below breed standard. This is the direct implication of the Khattak baseline, and the single most informative number you can put on the table before a supplier quotes you a figure
  • Known oxygen-limiting conditions — heat stress, high stocking density, constrained ventilation, warm source water. Biologically coherent, but not yet tested head-to-head against a cooler or lower-density context
  • Operations with a documented disease or gut-health challenge, where the immune and microbiome findings have somewhere to act

Where a growth benefit should not be assumed

  • Flocks already exceeding breed standard on optimised husbandry. This is precisely the condition under which the largest trial found nothing on growth
  • Any species other than broilers. No published growth data exists for laying hens, turkeys, quail, ducks or geese
  • Any quoted figure not accompanied by the baseline it was measured against. An 11.4% feed conversion improvement means nothing without knowing what the control group was achieving

The welfare and carcass benefits are the more consistent of the two classes on current evidence, because they held under the conditions designed to be hardest on the intervention. Whether they translate into money depends on your assurance scheme, your carcass grading and your current hock burn and feather scores — which is a calculation on your own numbers, not a claim a supplier can make for you.

A note on what is inference here. The Khattak baseline finding is published. The extension from it — that oxygen-limited flocks should show more benefit — is our reading of the evidence, and no trial has yet stratified by baseline performance to test it directly. We flag it as interpretation rather than presenting it as a result.

5. Due diligence

Six Questions Worth Asking a Supplier

01

What was the control group achieving?

Every percentage improvement is relative to a baseline. If the baseline is not disclosed, the percentage is not interpretable — and a large improvement over a poorly performing control tells you nothing about your own flock.

02

Is this peer-reviewed or commercial data?

Both can be informative, but they are not the same class of evidence. Commercial trial reports in this field frequently lack disclosed absolute figures, statistical analysis, or independent verification. Ask which category a number belongs to.

03

Was the null trial included in your evidence pack?

Khattak et al. (2025) is the largest study in the literature. A supplier who has not read it is not across the field; a supplier who has read it and omitted it has made a choice about what to show you.

04

What dissolved oxygen reaches the drinker?

Not what the unit achieves in a tank. Concentration measured at the line the birds actually drink from, under working flow, at the far end of the run — with the measurement method stated.

05

Which species is the data from?

Broiler data is broiler data. If the proposal is for layers, turkeys or quail, ask explicitly whether any species-specific trial exists — because at the time of writing, none does.

06

What happens if it does not replicate?

A supplier confident in conditional benefits should be willing to structure a pilot with pre-agreed endpoints and a defined decision point. Reluctance to define what failure looks like is itself informative.

6. The rest of poultry

Gaps, Not Null Results

The distinction matters and is routinely blurred. A null result means a trial was run and found nothing. A gap means no trial has been run at all. Everything below is a gap.

Laying hens

One published study, and it measured disinfection only — reduced bacterial counts in cloacal samples. No egg production, feed conversion, body weight, egg quality or welfare data exists. Given a 60–80 week production life, the feather-condition and abdominal-fat findings would matter considerably more here than across a five-week broiler cycle. This is arguably the highest-value unfilled gap in the field.

Turkeys

Same fast-growth, oxidative-stress profile as broilers, with breast myopathies a more severe commercial problem. That makes the white striping signal directly relevant if it replicates. No published study exists.

Japanese quail

Heart rates of 300–400 bpm, mass-specific oxygen consumption 1.4–1.8× that of broilers, and a 35–42 day cycle that makes controlled trials fast and inexpensive. The most practical species in poultry for a bridging trial, and no one has run one.

Ducks and geese differ meaningfully in gut microbiome and caecal fermentation architecture, so extrapolation from gallinaceous birds is weaker again. See the full livestock review for the extrapolation reasoning across species.

7. What would settle it

The Trial That Has Not Been Run

The disagreement in this literature is resolvable, and the design that would resolve it is not exotic. A trial that stratified by baseline flock performance — running the same intervention concurrently in a high-performing and an underperforming production context, with the same endpoints and the same delivered dose verified at the drinker — would directly test the ceiling-effect explanation against the alternatives. If the benefit is conditional on headroom, the effect size should scale inversely with baseline performance. No published trial has tested this.

Until it does, the honest position is that the growth benefit is unpredictable at the level of an individual operation, and the welfare and carcass benefits are the better-supported claim. We would rather say that plainly than present a range of outcomes as though the top of it were the expectation.

Common questions

Frequently Asked Questions

Does oxygenated water improve broiler feed conversion?

Sometimes. Three controlled trials say yes; the largest and best-powered says no. The most defensible reading is that the benefit is conditional on how oxygen-limited the baseline flock already is, rather than universal.

Why did the largest trial find no growth effect?

Both groups already exceeded breed performance standards by 19–22%. With that little headroom, no intervention could have shown a further growth improvement — which is a different finding from the intervention doing nothing.

What did that trial still find?

Significantly improved hock burn, feather condition and breast cleanliness, abdominal fat down 12%, thigh yield up 2.6%, reduced white striping, and differential gut microbiome signatures — all under the same optimal conditions.

Is there evidence for laying hens?

No production evidence. The single published study measured disinfection only. Egg production, feed conversion, egg quality and welfare in layers are entirely untested.

Is there evidence for turkeys or quail?

None published for either. Both are biologically plausible candidates, but plausibility is not evidence, and broiler figures presented as applicable to them are extrapolation.

How much oxygen does the treatment add?

Standard farm water typically carries 3–9 mg/L dissolved oxygen; nanobubble treatment can reach roughly 25–120 mg/L and hold it for days rather than minutes. What matters commercially is the concentration measured at the drinker line, not in the tank.

References

Scientific References

The mouse, pig and beef cattle studies referenced in the broader review are not repeated here. See the full reference list for all published evidence across species, and for the commercial trial reports excluded from it.

Evaluating a System for Your Flock?

Inpelor assesses whether the published evidence actually supports a proposal for your species, your baseline performance and your production context. We do not manufacture equipment, and our assessment is never tied to a purchase.

Request a Consultation →