Fecal lactoferrin

Lactoferrin sits in the secondary granules of neutrophils and is released where those cells degranulate. When neutrophils cross into the intestinal wall, lactoferrin reaches the gut lumen and leaves the body in stool. Measuring it there gives a number that tracks neutrophil traffic into the bowel, which is the reason the assay exists.

The protein is the same one described on the definitional entry for lactoferrin, and its granule storage is covered under biology. This section treats it purely as a laboratory measurement: what it detects, how accurate it is, and against what reference standard.

These pages describe the science of the test and never interpret an individual result. A result belongs to the clinician who ordered it, who holds the history the number has to be read against.

What the assay measures

Commercial assays are immunoassays, most of them enzyme-linked and reported in micrograms of lactoferrin per gram of stool. The 2003 study that established the marker used a polyclonal antibody ELISA on fresh samples from 271 people, and the separation it found between groups is wide enough to explain why the test caught on.

What the assay cannot do is tell whose lactoferrin it has found. The entry on what fecal lactoferrin is covers the analyte's origin in the neutrophil specific granule, and the one population where that origin does not hold: in a 1997 Lima cohort every fully breast-fed control child without diarrhea tested positive, because milk lactoferrin reaches an infant's stool intact.

Mean concentrations in that cohort were 440 µg/g in 104 patients with Crohn's disease and 1,125 µg/g in 80 with ulcerative colitis, against 1.27 µg/g in 31 patients with irritable bowel syndrome and 1.45 µg/g in 56 healthy controls. Those are standard errors of 128 and 498 on the two disease means, so the spread within the disease groups is large. Even so, the gap between inflamed and uninflamed bowel runs to two or three orders of magnitude, and a marker with that much room between its groups tolerates an imprecise cutoff.

The study reported specificity of 90% for identifying inflammation in active inflammatory bowel disease, and 100% specificity for ruling out irritable bowel syndrome when lactoferrin was raised. Its title claims the marker is both sensitive and specific. The abstract gives specificity figures and no sensitivity figure, so the first half of that claim rested on the concentration separation rather than on a stated number.

Pooled accuracy, and why the figures differ

Four pooled analyses have been published. They report different numbers because they asked different questions, and the difference between those questions is the most useful thing in this literature.

Sensitivity is stable across all of this, at roughly 0.78 to 0.82. Specificity is not. Asked to separate inflammatory bowel disease from irritable bowel syndrome, the assay reaches 0.94 and 0.95. Asked to detect endoscopic activity in people who already carry an IBD diagnosis, it falls to 0.79, and the confidence interval on that figure runs from 0.62 to 0.89.

That drop is not a contradiction between the studies. Distinguishing an inflamed bowel from a functional disorder is an easier discrimination than grading how inflamed a known-inflamed bowel currently is. Mosli and colleagues used endoscopy as the reference standard in every included study, which is the stricter comparison, and their conclusion was that the marker's value "must be considered in specific clinical contexts" rather than read off a single accuracy figure.

Two weaknesses in the pooled figures

The Wang analysis reports the highest specificity of the four and is the least solid of them. Its subgroup results give specificity of 100% for both Crohn's disease and ulcerative colitis while the overall pooled specificity is 0.95, which cannot both hold in the way stated, and its published conclusion describes the marker as having "a high specificity and a modest specificity". It was included here because it is one of only two meta-analyses addressing diagnosis directly, and its pooled sensitivity agrees with the others.

The Menees analysis is the negative reading in this set and deserves more weight than its brevity suggests. Screening 1,252 citations, the authors found just two lactoferrin studies meeting prospective diagnostic-cohort criteria, against eight for calprotectin. They concluded that C-reactive protein at or below 0.5 mg/dL, or calprotectin at or below 40 µg/g, essentially excludes inflammatory bowel disease in someone with IBS symptoms, and that lactoferrin analyzed on its own had little clinical utility. That reflects the size of the evidence base available to them rather than a demonstrated failure of the assay.

Against fecal calprotectin

Calprotectin occupies almost the same clinical role and has been studied more heavily. Two datasets compare them directly.

Mosli and colleagues pooled all three markers against the same endoscopic reference standard, which makes their figures the cleanest comparison available.

Calprotectin is the more sensitive of the two stool markers and lactoferrin the more specific, and the intervals overlap on both counts. Both beat C-reactive protein on sensitivity by a wide margin, which is the finding least in dispute anywhere in this literature: a normal CRP misses about half of endoscopically active disease.

The 2008 German study measured all four markers in the same 139 patients, every one of whom had an ileocolonoscopy. At adjusted cutoffs, overall diagnostic accuracy in inflammatory bowel disease was 80.0% for lactoferrin, 80.0% for calprotectin, 74.1% for neutrophil elastase and 64.0% for C-reactive protein. Calprotectin performed best in Crohn's disease at 81.4% and lactoferrin best in ulcerative colitis at 83.3%. The authors stated that none of the three stool markers is consistently superior, and that all three outperform CRP.

Combining markers helped more than choosing between them. A composite of the stool markers, CRP and a disease-specific clinical index reached 95.3% accuracy in ulcerative colitis, against 79.0% for the clinical scores alone.

What the number cannot do

Raised fecal lactoferrin means neutrophils are entering the bowel. It does not say why. Infection, drug-induced injury and malignancy all recruit neutrophils, so the assay is a test for inflammation rather than a test for any particular disease. This is why every pooled figure above is tied to a stated comparison group, and why a specificity of 0.95 against irritable bowel syndrome carries no promise of the same performance against an undifferentiated population.

Cutoffs are also not standardized across platforms. Langhorst and colleagues improved every marker's accuracy by adjusting cutoffs to their own cohort, which is a warning about transferring a threshold between laboratories rather than a result in its own right. A figure quoted without its assay and its cutoff is close to meaningless, and comparison of results between platforms needs the platform named.

How far that goes is set out in the lactoferrin test entry, which tabulates the assay, population and threshold used by six validation studies. Kane reported micrograms per gram of stool and Walker micrograms per milliliter, with no conversion stated in either paper, and the only ROC-derived threshold published alongside its full performance figures comes from a cohort of 52 children.

Feeding the protein is a separate question from measuring it, and the trials that have tried it report fecal markers going in more than one direction. In one randomized trial of formula fortification in 96 anemic infants, diarrhea became less frequent at the higher fortification level while fecal calprotectin in the same infants came out higher. Those readings and the pooled trial evidence behind lactoferrin and gut health sit on the supplementation side of the subject rather than the diagnostic one.