Lactoferrin: what the evidence actually shows

Lactoferrin has one of the most impressive property lists in the supplement aisle: it binds iron with enormous affinity, starves bacteria of that iron, punches holes in bacterial membranes, calms inflammatory signalling, and shows up in colostrum at the exact moment a newborn’s gut is being colonised. Every one of those statements is true in a test tube. The question this page is about is different and harder: lactoferrin is a large, folded, glycosylated protein, and you are proposing to heat it, spray-dry it, seal it in a capsule and drop it into stomach acid. What arrives at the other end, and what has that actually been shown to do in a person? The answer varies enormously by indication — from a defensible second-line option to a claim that the best-powered trial ever run flatly contradicted.

What lactoferrin actually is

Lactoferrin is an ~80 kDa glycoprotein of the transferrin family, folded into two lobes that each clamp a single ferric iron ion. It is secreted by glandular epithelium into milk, colostrum, tears, saliva and other mucosal fluids, and it is also stored in the secondary granules of neutrophils and released at sites of infection.^[1] Its biological job is not nutrition. It is iron custody — holding iron so tightly, and across a far wider pH range than transferrin, that microbes competing for the same iron cannot get at it.

The commercial reality is narrower than the biology. Practically every lactoferrin product you can buy is bovine lactoferrin (bLf), ion-exchange purified from cheese whey or skim milk. Human and bovine lactoferrin are closely homologous but not identical, and the field runs on bLf because it is available by the tonne and has been through the US GRAS process: cow’s-milk-derived lactoferrin has received “no questions” letters from FDA, and in 2025 and 2026 so did two notices for recombinant bovine lactoferrin made by precision fermentation in the yeast Komagataella phaffii.^[23]

Recombinant human lactoferrin is a different product and should not be conflated with it. The best-studied version, talactoferrin alfa, was developed as a drug and taken into phase II/III trials in severe sepsis and in preterm infants.^[22][18] Rice-expressed human lactoferrin exists too. But every GRAS notice for a human lactoferrin — from rice or from bovine milk — was withdrawn before FDA finished evaluating it.^[23] If a supplement label says “lactoferrin” with no species, it is bovine.

The four mechanisms, and how many of them survive a capsule

The mechanism story is genuinely interesting, and it is worth separating the parts that depend on the intact protein from the parts that do not.

• Iron sequestration. Apo-lactoferrin scavenges free iron from its surroundings, which is bacteriostatic rather than bactericidal — it slows growth by starvation. This is the one mechanism that requires the protein to arrive iron-free, and it is the mechanism a heavily iron-saturated product simply cannot perform.^[4]
• Lactoferricin. In 1992 Bellamy and colleagues isolated an antimicrobial domain near the N-terminus, in a region distinct from the iron-binding sites, by cleaving human and bovine lactoferrin with gastric pepsin. The bovine peptide — an 18-residue disulfide loop — was the more potent of the two, killing Gram-negative and Gram-positive strains at 0.3–3.0 µM.^[2] That is the elegant part of the story: the stomach does not only destroy lactoferrin, it also manufactures the weapon. It is also where the honesty has to come in, because a systematic review of lactoferrin’s oral fate reports that lactoferricin is undetectable in gastrointestinal digesta — generating it in a beaker with purified pepsin is not the same as recovering it from a gut.^[5]
• Immunomodulation. Pooled across human trials, lactoferrin supplementation lowered interleukin-6 by a mean of about 25 pg/mL, without moving C-reactive protein or NK-cell cytotoxicity.^[20] This is the mechanism doing most of the work in the iron literature, for reasons we come to below.
• Receptor-mediated effects on intact protein. Some lactoferrin really does cross intact. The cleanest demonstration is old and still striking: structurally intact 78 kDa maternal lactoferrin was purified from the urine of preterm infants fed human milk, with the authors concluding that the degradation of absorbed lactoferrin in vivo was “remarkably limited.”^[6] Note the population. A preterm infant’s stomach is a gentler environment than an adult’s — less acid, less pepsin — and that caveat runs through everything below.

Heat and digestion: the fact most product pages get wrong

This is where a lactoferrin page earns its place, because it is the variable that determines whether anything else on the label matters. Lactoferrin is sensitive to thermal denaturation and to other physicochemical stresses, and the comprehensive review of the field states the practical consequence plainly: extraction, powder formation and the processing parameters of finished products have to be optimised specifically to minimise undesired denaturation.^[1] That is not a marketing talking point; it is a manufacturing constraint.

The most useful single experiment on this is thirty years old. Using differential scanning calorimetry on bovine lactoferrin, Paulsson and colleagues found two thermal transitions whose enthalpies depended on the protein’s iron status. Pasteurisation affected only the lower-temperature transition, and pasteurised lactoferrin bound bacteria and inhibited E. coli just as well as unheated lactoferrin. UHT treatment completely denatured both forms, reduced their ability to bind bacteria, and abolished the antibacterial activity entirely.^[4] So the correct statement is not “heat destroys lactoferrin” and it is not “processing doesn’t matter” — it is that there is a real threshold, pasteurisation sits below it, and UHT sits above it.

Digestion is the second filter, and it is the harsher one. Lactoferrin is sensitive to the physicochemical stresses of the gastrointestinal tract, which is precisely why an entire research literature exists on encapsulation and stabilisation technologies intended to get it past the stomach.^[5] The one controlled human demonstration that processing changes digestion comes from the neonatal intensive care unit: in a randomised trial in preterm infants, pasteurising human milk measurably altered the kinetics of gastric proteolysis compared with raw milk.^[7] Real effect, real population — and again, not an adult stomach.

Apo versus holo: the product variable nobody prints on the label

Bovine lactoferrin is sold as a single ingredient name, but it is not a single material. A dedicated paper on commercial bLf preparations found that their physico-chemical heterogeneity directly influences their effectiveness, and that iron saturation in particular governs both thermal stability and resistance to proteolysis. The authors are blunt that the efficacy of commercial bLf “can be erratic,” depending on structural integrity, degree of iron and other metal saturation, N-glycosylation, desialylated forms and contaminants.^[3] Five N-glycosylation sites are themselves part of what makes the protein resist proteolysis.^[3]

And the two properties you might want pull in opposite directions:

• Iron-saturated (holo) lactoferrin is the tougher molecule — more resistant to heat-induced change, with a single thermal transition that survived pasteurisation.^[4] It is also the form that, in the same experiments, did not inhibit bacterial growth at all, pasteurised or not.^[4] Full iron loading buys stability by giving up the mechanism.
• Iron-free (apo) lactoferrin is the active one — and the fragile one. It is also the form with the clearest evidence for helping iron absorption. In a stable-isotope crossover study in 25 Kenyan infants, adding 1.41 g of apo-lactoferrin (0.56% iron saturation) to a ferrous-sulfate test meal raised fractional iron absorption by 56% — 9.8% versus 6.3% — while intrinsically labelled holo-lactoferrin (47.26% saturation) delivered absorption no different from ferrous sulfate alone.^[8]

That study also settles a question buyers rarely think to ask. Its 1.41 g of holo-lactoferrin at 47% saturation carried 1.5 mg of iron; the apo form carried 0.08 mg.^[8] Scale that to a 250 mg capsule and you get, at most, a couple of tenths of a milligram — against 65 mg of elemental iron in a standard 325 mg ferrous sulfate tablet. Whatever lactoferrin does for iron status, it is not doing it by supplying iron. Keep that in mind for the next section, because it reframes the whole anemia literature.

The evidence, graded by indication

Iron status: the best case, and the trial that complicates it

Start with the trial that built the reputation. In a prospective, randomised, double-blind study in 100 pregnant women with iron-deficiency anemia, 100 mg of bovine lactoferrin twice daily was compared with 520 mg of ferrous sulfate once daily for 30 days. Haemoglobin, ferritin and serum iron rose significantly in both groups, with no significant difference between them, and the median scores for abdominal pain and constipation were significantly worse on ferrous sulfate.^[9] That is a clean, well-designed result and it is the honest heart of the lactoferrin case: comparable, and easier to take.

Pooling four such trials in 600 women, a 2017 meta-analysis found haemoglobin at four weeks favoured lactoferrin by 0.77 g/dL (95% CI 0.04–1.55, P = 0.04), with significantly fewer gastrointestinal side effects.^[10] Read the subgroups before you read the conclusion, though. The overall confidence interval nearly touches zero; there was no significant difference in mild anemia across three trials and 372 women; and the entire moderate-anemia signal came from a single trial of 228 women.^[10] The authors nonetheless concluded that lactoferrin “should be the iron replacement agent of choice” in pregnancy — a recommendation that travels considerably further than the data underneath it.

Two structural problems sit under this literature, and both matter more than any individual effect size.

First, the trials come from a small number of centres. The pregnancy work is dominated by a handful of groups — principally in Rome and Naples, and in Egyptian university hospitals — often using the same protocol, the same 100 mg twice-daily dose, and in some cases the same commercial material.^[9][15] Replication across independent centres is what turns a promising result into a reliable one, and this field has less of it than its citation count suggests.

Second, there is a live retraction in the middle of it. One of the most-cited head-to-head trials — a single-centre Egyptian randomised study of 200 pregnant women comparing lactoferrin with ferrous sulfate — was retracted in 2023.^[11] Its numbers have not gone anywhere. A 2022 meta-analysis in Nutrients, which concluded that lactoferrin is a “superior supplement to ferrous sulfate” for serum iron, ferritin and haemoglobin, draws on that trial among its sources.^[12] If you are evaluating a lactoferrin claim that cites a meta-analysis, it is worth checking whether the inputs are still standing.

That same 2022 analysis contains the most interesting mechanistic finding in the whole area, and it cuts against the intuitive story. Compared with ferrous sulfate, lactoferrin produced better serum iron, ferritin and haemoglobin — while producing lower fractional iron absorption (−2.08%) and a large drop in IL-6 (−45.6 pg/mL).^[12] In other words, the proposed mechanism is not that lactoferrin delivers iron better. It is that it lowers inflammation, which lowers hepcidin, which releases the body’s own sequestered iron — the mechanism proposed by the group that has studied anemia of inflammation in pregnant and non-pregnant women with oral lactoferrin.^[15] If that is right, lactoferrin should work best in inflammatory anemia and worst in plain dietary iron deficiency.

Which is roughly what the largest and most rigorous trial found. In 2026, a double-blind, randomised, parallel-group non-inferiority trial in Dhaka randomised 555 non-pregnant Bangladeshi women with iron-deficiency anemia to 200 mg bLf, 400 mg bLf, or 60 mg elemental iron as ferrous sulfate, daily for 12 weeks, with a pre-specified non-inferiority margin of −0.25 g/dL. Mean haemoglobin change was −0.2 g/dL on 200 mg, 0.0 g/dL on 400 mg, and +1.1 g/dL on ferrous sulfate — mean differences of −1.2 and −1.1 g/dL against the active control, P < 0.0001. Lactoferrin did not fail to be non-inferior; it was inferior, and by a margin roughly four times the one the trial was designed to rule out. Adverse events were comparable across all three groups.^[13] The authors’ conclusion is unambiguous: these findings do not support bovine lactoferrin as a substitute for ferrous sulfate.

For context on where lactoferrin sits among iron treatments generally: a network meta-analysis of 53 randomised trials in 9,145 pregnant women found that only intravenous iron sucrose and ferric carboxymaltose beat oral ferrous sulfate on haemoglobin, and that the evidence for every other intervention compared with ferrous sulfate was insufficient.^[14] Lactoferrin is not the exception to that verdict; it is inside it.

Preterm infants: a textbook case of a large trial overturning small ones

If you want to understand why evidence hierarchies exist, this is the case study. Colostrum is rich in lactoferrin, preterm infants get little of it, and late-onset sepsis and necrotising enterocolitis are among the things that kill them. A series of small trials supported supplementation, and the 2020 Cochrane review pooled twelve randomised trials in 5,425 infants to a typical risk ratio of 0.82 for late-onset sepsis (number needed to treat 25), with no effect on NEC stage II or III (RR 1.10) and none on all-cause mortality.^[16] The reviewers graded that evidence low certainty and said why in unusually direct language: “the presence of publication bias and small studies of poor methodology that may inflate the effect size make recommendations for clinical practice difficult.”^[16]

Then somebody ran the trial properly. ELFIN randomised 2,203 infants born before 32 weeks’ gestation across 37 UK hospitals to enteral bovine lactoferrin at 150 mg/kg per day (maximum 300 mg/day) or sucrose placebo, until 34 weeks’ postmenstrual age, with parents, caregivers and outcome assessors blinded and 99% primary-outcome capture. Late-onset infection occurred in 316 of 1,093 infants on lactoferrin (29%) versus 334 of 1,089 on placebo (31%)— adjusted risk ratio 0.95 (95% CI 0.86–1.04), P = 0.233. The conclusion: enteral bovine lactoferrin does not reduce late-onset infection, and the data do not support its routine use.^[17]

The contrast is the lesson. A 120-infant randomised trial of recombinant human lactoferrin (talactoferrin) in the same population reported 50% fewer hospital-acquired infections — a striking result, and exactly the size and kind of trial that ELFIN was built to adjudicate.^[18] When a dozen small positive trials meet one large, well-conducted, adequately powered trial, the large trial is the one that tells you what happens.

None of this is directly about a capsule in an adult. The dose is weight-scaled and enormous, the population is neonatal, and infection prevention in a NICU is not a consumer question. But ELFIN is the single best-powered test of the general proposition that oral lactoferrin prevents infection — and it came back null.

Gut, immunity and the rest of the list

Helicobacter pylori. The most defensible non-iron use is as an adjunct to eradication therapy. A meta-analysis of nine randomised trials in 1,343 patients found intention-to-treat eradication of 86.6% with lactoferrin versus 74.4% without (OR 2.26, 95% CI 1.70–3.00), with fewer total side effects (9.1% vs 16.3%) and a marked reduction in nausea (OR 0.15).^[19] Two honest caveats: these were small trials, mostly published before 2009, and eradication regimens have changed substantially since — the comparator that lactoferrin was added to is not the comparator your gastroenterologist would use today. This is adjunct-level evidence for a prescribed regimen, not a reason to self-treat an infection that requires diagnosis and antibiotics. For the broader question of what actually has evidence at the gut barrier, see larazotide and KPV.

Immune and respiratory claims. This is where consumer marketing is loudest and the data are least supportive — specifically for the people buying it. A systematic review and meta-analysis of 25 studies found respiratory tract infection incidence reduced in infants and children (OR 0.78, 95% CI 0.61–0.98) but not in adults (OR 1.00, 95% CI 0.76–1.32) — a confidence interval centred precisely on no effect. Lactoferrin did lower IL-6 in adults, but did not move C-reactive protein or NK-cell cytotoxicity, and the reviewers noted that clinical studies remain limited.^[20] An anti-inflammatory biomarker effect with no infection benefit is exactly the kind of finding that gets laundered into a “clinically proven immune support” claim. It is not one. Lactoferricin belongs to the same broad family of cationic host-defence peptides as LL-37, and the same gap between in-vitro potency and human outcomes applies to both.

COVID-19. Grade this low and move on. Lactoferrin generated an enormous wave of pandemic-era claims on the strength of in-vitro antiviral activity. The best test of them, the LAC randomised double-blind placebo-controlled trial, gave 800 mg/day of bovine lactoferrin or placebo to 218 hospitalised patients with moderate-to-severe COVID-19 on top of standard care and found no difference in either primary outcome — death or ICU admission (RR 1.06, 95% CI 0.63–1.79) and discharge or NEWS2 ≤ 2 within 14 days (RR 0.85, 95% CI 0.70–1.04). It was, the authors noted, extremely well tolerated.^[21] Safe and ineffective is a perfectly respectable trial result. It is not a selling point.

Buying it: dose, form, price, and who should bother

Dose. The consumer-relevant range is narrower than the label range. The pregnancy anemia trials used 100 mg twice daily (200 mg/day);^[9] the 2026 non-inferiority trial tested 200 and 400 mg/day;^[13] the pooled inflammation signal appeared at around 200 mg/day;^[20] the COVID trial used 800 mg/day.^[21] Preterm dosing at 150 mg/kg/day^[17] is weight-scaled clinical dosing and tells you nothing about an adult capsule. So: 200–400 mg/day is where the human data live, and the common 250 mg once-daily product sits inside it. Nothing supports going higher. The label variables that actually separate one product from another — species, iron saturation, processing and verified content — are laid out in our criteria for judging a lactoferrin label.

What to actually check before you pay.

• Species, stated explicitly. It should say bovine. If it says only “lactoferrin,” assume bovine — and if it implies human lactoferrin at a supplement price, that is a claim FDA has never let through a completed GRAS notice.^[23]
• Iron saturation, in a number. Almost nobody prints it, and it is the variable a dedicated study of commercial preparations identified as governing both stability and function.^[3] Email the manufacturer. A company that controls its saturation will tell you; a company that cannot answer is not controlling it. Given that apo is the antibacterial and absorption-enhancing form,^[4][8] low saturation is the one you want — and it is also the more fragile one, which is the trade-off you are paying someone to manage.
• Protein content per capsule, third-party verified. This is a purified dairy fraction, and “milk protein concentrate” is not the same product. You want a certificate of analysis giving actual lactoferrin content, not just capsule weight — the same discipline we apply to any bulk protein supplement.
• Do not pay extra for “undenatured,” “raw” or enteric-coated. Both are mechanistically plausible.^[4][5] Neither has been tested against conventional material on a human outcome, and the clinical evidence base was built on conventional material. A premium for an untested variable is a premium for a story.

Price, in the only unit that matters. Work in dollars per gram of protein, not dollars per bottle. A 250 mg × 60 capsule bottle contains 15 grams of lactoferrin; at a typical US retail price of roughly $30–45 that is about $2–3 per gram, or $0.50–0.75 a day at 250 mg — call it $15–25 a month. Bulk bLf powder runs cheaper per gram but costs you dose accuracy and, usually, the certificate of analysis. Hold that against the comparator: a month of ferrous sulfate costs a few dollars and, in the largest head-to-head trial, raised haemoglobin by 1.1 g/dL where lactoferrin raised it by nothing.^[13]

Who this plausibly makes sense for.

• Someone with iron deficiency who genuinely cannot tolerate oral iron. The tolerability advantage is the single most consistent finding in this entire literature — less abdominal pain and constipation than ferrous sulfate in the head-to-head trials,^[9] fewer GI side effects in the pooled analysis,^[10] and no adverse-event penalty in the 2026 trial.^[13] An iron supplement you abandon at week two treats nothing. But this is a second-line trade made with a clinician, with ferritin and haemoglobin rechecked on a schedule, and with clear eyes about the fact that the largest trial found it substantially less effective. Modern alternatives — alternate-day dosing, lower elemental doses, different salts, or IV iron — should be on the table in the same conversation.^[14]
• Someone whose anemia is inflammatory rather than purely dietary. This is where the proposed hepcidin/IL-6 mechanism actually predicts benefit,^[12][15] and it is the most scientifically coherent case for the compound. It is also a diagnosis, not a self-assessment.
• Someone on a prescribed H. pylori eradication regimen whose clinician is open to an adjunct.^[19] Adjunct is the operative word.

Who is wasting money.

• Adults buying it to get fewer colds. The pooled adult respiratory data sit exactly on the null.^[20] The paediatric signal is real and is not transferable to you.
• Anyone taking it instead of iron for diagnosed iron-deficiency anemia. It delivers a fraction of a milligram of iron per capsule,^[8] and it lost decisively to ferrous sulfate in the only large rigorous comparison.^[13]
• Anyone paying a premium for “undenatured” without a saturation figure. You are being charged for the one variable that matters, measured on the one axis nobody has tested.
• Anyone buying it as an antimicrobial for an active infection. The in-vitro potency is real and the human infection-prevention record is the ELFIN result.^[17]

Safety, and the one place it is not benign

Oral bovine lactoferrin is well tolerated at the doses people use. The Cochrane reviewers reported no adverse effects across the included preterm trials;^[16] the 2026 trial found adverse events comparable across lactoferrin and ferrous sulfate arms;^[13] and the LAC investigators described an “excellent safety and tolerability profile” at 800 mg/day in hospitalised patients.^[21] The specific things worth knowing:

• Dairy protein allergy is a hard contraindication. This is a protein purified from bovine whey. Cow’s-milk protein allergy means avoid it. Lactose intolerance is a different condition and is not the same barrier — but a poorly purified product can carry residual milk components, which is one more argument for a certificate of analysis.^[3]
• Iron, in both directions. It supplies essentially no iron,^[8] so it cannot cause iron overload by delivery — but it is not inert around iron either. Apo-lactoferrin raised absorption of iron from a co-eaten meal by 56% in infants.^[8] If you have haemochromatosis or another iron-loading condition, that is a conversation with your clinician, not a shrug.
• Pregnancy and infants. Much of the human data was generated in pregnancy, which sometimes gets read as reassurance. It should not be: these are unapproved dietary supplements, and anemia in pregnancy is a condition that needs diagnosis, monitoring and a treatment chosen on effectiveness.^[14] Lactoferrin for infants is a formula-composition question for a paediatrician, not something to add at home.
• The one genuine harm signal. The phase II/III OASIS trial gave 1.5 g of recombinant human lactoferrin three times daily to adults with severe sepsis. It was stopped early for futility and safety after 305 patients: in-hospital mortality was 28.1% with talactoferrin versus 17.8% with placebo (P = 0.037) and three-month mortality 30.1% versus 20.4% (P = 0.036), with the authors concluding it “may even be harmful.”^[22] Different molecule, ten times the dose, and a critically ill population — this says nothing about 250 mg of bovine lactoferrin in a healthy adult. It says something important anyway: “it’s just a milk protein, it can’t hurt” is an assumption, and it has been falsified at least once.

The honest bottom line

Lactoferrin is a real molecule with real biology, and it is not a scam. It is also a protein being sold as though it were a small molecule — as though what is in the capsule is what reaches the target. Heat processing, iron saturation and gastric proteolysis each take a cut, none of them is disclosed on a label, and the one property that most of the marketing rests on (antibacterial activity) belongs to the iron-free form that is also the least stable.^[3][4]

On outcomes, the shape of the evidence is consistent and unflattering: the smaller the trial, the better lactoferrin looks. Twelve small preterm trials pooled to a real-looking benefit; ELFIN’s 2,203 infants found none.^[16][17] Four small pregnancy trials made it look at least as good as ferrous sulfate; 555 women in a properly powered non-inferiority trial found it clearly worse.^[10][13] A pooled adult immune effect on IL-6 comes with a respiratory-infection odds ratio of exactly 1.00.^[20]

What survives all of that is narrow and genuine: lactoferrin is gentler on the gut than ferrous sulfate, and it may help most in anemia driven by inflammation rather than by dietary shortfall.^[9][12] That makes it a legitimate second-line option for someone who cannot take iron, chosen with a clinician and checked with bloodwork. It does not make it a replacement for iron, an immune supplement for adults, or worth a premium for processing claims nobody has tested.

This article is research information, not medical advice. Bovine lactoferrin is sold as a dietary supplement and has no FDA approval for any indication — it is not approved to treat or prevent iron-deficiency anemia, infection, H. pylori, or any respiratory illness, and GRAS status for use in food is not a finding of clinical efficacy. It is purified from cow’s milk and must be avoided by anyone with cow’s-milk protein allergy. Iron-deficiency anemia requires diagnosis and monitoring; do not substitute a supplement for prescribed iron, and do not treat anemia in pregnancy, or give lactoferrin to an infant, without a clinician directing it. If you have haemochromatosis or another iron-loading disorder, or you are on treatment for H. pylori or any other infection, talk to a licensed clinician before starting it.