Lactoferrin, graded by the evidence.

Lactoferrin is a single-chain glycoprotein of roughly 80 kilodaltons, built from two similar lobes. Each lobe binds one ferric iron ion with a tenacity few biological molecules match, and holds it across a range of acidity that would strip most carriers. The protein occurs in milk, most abundantly in colostrum, and also in tears, saliva, nasal and intestinal secretions, and in the granules of neutrophils, where it is released at sites of inflammation. That last property is what turns the protein into a laboratory measurement: fecal lactoferrin tracks neutrophil traffic into the bowel, and pooled analyses put its specificity for inflammatory bowel disease against irritable bowel syndrome at 0.94 to 0.95. The published thresholds for that measurement are another matter: six validation studies of the lactoferrin test share no common unit, and only one of them reports a cutoff with its sensitivity and specificity attached. A fuller account of what lactoferrin is covers the crystal structure, the iron site and the second origin in the neutrophil.

Its best-characterized action follows directly from the iron binding. By sequestering free iron it denies an essential nutrient to bacteria that need it, which is where the antimicrobial reputation began. Subsequent work has described effects that do not depend on iron at all, including direct interaction with bacterial membranes and with host cell receptors. Those mechanisms, and the receptor and transport work behind them, are set out under the biology of lactoferrin.

How much of the protein is present depends on the species and on when the milk was taken. A cross-sectional study of 824 lactating women measured 3.16 g/L in colostrum, 1.73 in transitional milk and 0.90 in mature milk. A series through bovine lactation found 0.83 g/L at the first milking, falling to 0.09 by the third week. Compared stage for stage, human colostrum carries about four times the lactoferrin of bovine colostrum, and mature human milk about ten times that of mature cow's milk, so any single human-against-bovine ratio depends on which stage each sample came from. Every figure sits with the assay that produced it in lactoferrin concentration by source.

What that means for a person taking a supplement is a separate question, and a much less settled one. The ledger below is this site's answer to it.

What the graded claims add up to

Nine claims carry a grade. Two reach Supported, three are graded Not supported, and none reaches Established, which on this scale needs more than one adequately powered human trial pointing the same way. No claim about lactoferrin currently clears that bar.

The two supported claims have a design in common. Both tested the protein alongside or against an existing treatment rather than on its own. Pooled across nine randomized trials and 1,343 participants, adding bovine lactoferrin to antibiotic eradication therapy raised Helicobacter pylori clearance from 74.44 to 86.57 percent. Four trials in 600 pregnant women with iron deficiency anemia found oral lactoferrin comparable to ferrous sulfate on hemoglobin response and easier to tolerate, and the adverse events those trials counted are set beside three other populations under lactoferrin side effects. Lactoferrin benefits carries each claim with the design and the participant count behind it.

The three claims graded Not supported failed in different ways. ELFIN is the largest trial of lactoferrin ever run: 2,203 infants born before 32 weeks, randomized across 37 UK hospitals to bovine lactoferrin or to sucrose, with late-onset infection in 316 of 1,093 against 334 of 1,089 and an adjusted risk ratio of 0.95 (95% CI 0.86 to 1.04). The LAC trial gave 800 mg a day to 218 hospitalized adults with COVID-19, passed the enrollment its own power calculation called for, and returned 1.06 (0.63 to 1.79) for death or admission to intensive care. OASIS never got that far: designed for 1,280 adults with severe sepsis, it was stopped at 305 on its monitoring board's advice, with in-hospital mortality higher on treatment at 28.1 against 17.8 percent.

The same sequence appears twice in that list, and it is why the scale keeps a separate grade for a claim that was tested and failed. Small randomized trials in preterm infants reported that lactoferrin prevented infection; ELFIN looked for that effect and did not find it. A phase 2 trial of 194 adults with severe sepsis reported 28-day mortality of 14.4 percent against 26.9 on placebo; the larger trial of the same drug at the same dose by the same route then reversed the direction. Both early reports stay in circulation and are still cited, which is the gap a grade of Not supported exists to close.

How this site works

Every claim gets graded on the best available evidence, not the most recent study and not the most favorable one. A grade of Laboratory only is a description rather than a criticism: a great deal of genuinely interesting lactoferrin research has been done in cells, and reporting it as though it were a clinical finding is the error, not the research itself.

Where the honest grade is Insufficient, the page still gets written. Recording that nobody has run an adequate trial is useful, and it is information that is otherwise hard to find.

Figures carry their units and their source. The concentrations above span more than an order of magnitude, and quoting one of them without saying which milk and which stage of lactation it describes is the most common error in writing on this subject. What pasteurization does to those figures is a separate measurement question, and the section on lactoferrin sources pairs each concentration with the method that measured it.

Units get the same treatment. Two of the validation studies behind the stool assay disagree on them: one reports micrograms per gram of stool, the other micrograms per milliliter of extract, and neither states a conversion, so their numbers cannot be placed on one axis. Lactoferrin reference data collects tables of this kind, every row carrying the assay behind it. That section is thin at present, and saying so is better than filling it with figures whose method nobody recorded.

The published record has gaps of its own, and for this subject they have been counted. A search of the ClinicalTrials.gov intervention field on 11 November 2026 returned 158 interventional lactoferrin trials. Eighty-five are marked completed and nine have posted a summary result. Those nine enrolled 747 people between them, while the 76 completed records with nothing posted enrolled 13,969. A further 36 records carry the registry's own status of Unknown, 8,099 planned participants between them, and fifteen of the 36 were last updated before 2020.

An unposted record is not an unreported trial. A registry entry and a journal article are separate obligations, and lactoferrin clinical trials works through a case where the result reached print and the record was never updated. What the counts do establish is that the visible studies were not drawn at random from the studies that ran, which is why a claim here is graded on the best single trial rather than on a tally of positive ones. A 2013 analysis in the BMJ found the same pattern in general: of 585 large randomized trials completed before 2009, 171 were never published, and 133 of those had no results in the registry either. Lactoferrin research records trials with their status, including the ones that recruited and then reported nothing.

More on sourcing and method on the about page.