Why a “Normal” Ferritin May Still Signal Iron Deficiency

The lab's lower limit may start with the wrong comparison group

A ferritin result of 18 µg/L can come back flagged as normal, yet still fall in a range that deserves clinical attention. The problem may begin before the blood draw: many laboratory reference intervals were built from people assumed to be healthy, without first confirming that they had adequate iron stores.

That approach mixes iron-replete people with people whose iron stores are already low. The resulting distribution can push the lower boundary of "normal" downward. A 2025 study by Troike et al. tested what happened when the reference group instead consisted of iron-replete patients. Its results support the concern that conventional ferritin lower limits can miss iron deficiency, especially in women.

This is an argument for better interpretation, not a license to diagnose yourself from one number. Ferritin is useful, but inflammation can raise it independently of iron sufficiency. Hemoglobin, transferrin saturation, symptoms, medical history, and the reason for testing still matter.

Reference interval and diagnostic cutoff are different numbers

Laboratories often report a reference interval: the range observed in a selected reference population. A diagnostic cutoff answers a different question: which value best separates people with and without a condition for a particular clinical purpose. Confusing those concepts turns a technical issue into bad medical advice.

Troike and colleagues used receiver-operating-characteristic analysis in a patient cohort stratified by iron status, transferrin saturation, and hemoglobin. Rather than using an unscreened general population as the benchmark, they used iron-replete patients as the reference comparator. The study reported recalculated lower reference limits of 16.9 µg/L for females and 30 µg/L for males, versus conventional laboratory lower limits commonly near 10–12 µg/L.

The same analysis produced higher Youden Index–optimal diagnostic cutoffs: 45 µg/L for females and 70 µg/L for males. The Youden Index selects the point that balances sensitivity and specificity in that dataset; it does not establish a universal treatment threshold. At 45 µg/L, sensitivity for iron deficiency in females increased by 44% relative to conventional laboratory cutoffs. The publication details provided here do not report confidence intervals or the cohort size, so the precision of that estimate cannot be judged from the available report.

A ferritin of 18 µg/L may be "in range" on a lab report while remaining below the 45 µg/L diagnostic cutoff identified for women in this single ROC-analysis study.

Why women's ferritin values need separate interpretation

The female and male cutoffs in the Troike analysis differed substantially. Menstrual iron losses, average body iron stores, and differences in the reference population could all contribute. The study cannot separate those explanations completely.

Population data offer a plausible biological reason to avoid treating all adult women as one uniform reference group. In two European cohorts, CoLaus and PREVEND, Kastrati et al. found that iron-biomarker profiles differed significantly by menopausal status. That study did not create ferritin reference intervals or validate a 45 µg/L cutoff. Still, it supports the narrower point that a reference group containing many premenopausal women may have systematically lower ferritin values than one composed of postmenopausal women.

Lower average values do not automatically mean deficiency. They do mean that a population-derived lower limit can answer a different question than a clinician asking whether a person has enough available iron for their circumstances.

Ferritin alone cannot settle the diagnosis

Ferritin measures stored iron, but it also rises during inflammation. In heart failure, for example, Alak et al. reported that ferritin and transferrin saturation can mislead because systemic inflammation confounds traditional iron markers. Their retrospective cohort study of 322 patients developed a ferritin/CRP-and-albumin composite score for mortality risk stratification. It does not validate higher ferritin cutoffs in otherwise healthy adults, but it illustrates why a ferritin value should not travel alone.

A randomized study in Cambodian women makes the same practical point from another angle. Pei et al. compared baseline hemoglobin, hepcidin, ferritin, and total body iron stores as predictors of a hemoglobin response after 12 weeks of daily iron supplementation. The four measures were equally strong diagnostic predictors in that population. The result does not show that ferritin thresholds are irrelevant; it shows that fine-tuning one marker may not outperform a broader assessment.

A useful clinical discussion often includes ferritin, hemoglobin, serum iron, and transferrin saturation. Hepcidin may add information in selected settings, though availability and interpretation vary. The right panel depends on the person: heavy menstrual bleeding, pregnancy, inflammatory disease, kidney disease, recent infection, diet, gastrointestinal symptoms, and prior supplementation can all change the meaning of the result.

Higher thresholds may find more cases, but they can also create new errors

Raising a diagnostic cutoff increases sensitivity in the Troike cohort, which means fewer iron-deficient women were missed by the study's comparison standard. It can also reduce specificity, meaning more people without iron deficiency may fall below the cutoff. The reported 44% sensitivity gain does not tell us the absolute number of additional diagnoses, the tradeoff in false positives, or whether people in the 15–30 µg/L range have symptoms or improve with treatment in primary care.

Those unanswered questions matter. A biomarker cutoff is clinically useful only if it improves decisions and outcomes, not merely classification. Trials in more specialized settings use much higher definitions for specific reasons. For example, a randomized preoperative trial by Kaserer et al. enrolled surgical patients with iron deficiency defined as ferritin below 100 ng/mL or transferrin saturation below 20%, with or without anemia. That disease-setting threshold should not be copied into routine testing for healthy people.

Pregnancy is another setting with its own risks and screening questions. Auerbach et al. studied iron deficiency in first-trimester non-anemic pregnant women, while the USPSTF has found insufficient evidence for routine screening in asymptomatic populations. That institutional caution helps explain why a promising diagnostic proposal does not become standard practice overnight.

What to do with a low-normal result

If ferritin is below 45 µg/L in a woman or below 70 µg/L in a man, the Troike study suggests that a clinician should not dismiss the value as definitively normal solely because it clears a laboratory lower limit. That is a prompt for context, not an automatic prescription for iron.

• Review symptoms and likely sources of iron loss, including menstrual or gastrointestinal blood loss.
• Check hemoglobin and transferrin saturation alongside ferritin; consider inflammatory context when ferritin seems inconsistent with the rest of the picture.
• Discuss repeat testing, further evaluation, or treatment with a clinician before taking high-dose iron. Excess iron has risks, and unexplained deficiency deserves an explanation.

The central finding is worth taking seriously: conventional ferritin reference intervals may have been calibrated with people who already had low iron stores. But the proposed 45 µg/L and 70 µg/L diagnostic cutoffs come from one institution and one study cohort, without independent replication in the evidence reviewed here. A low-normal ferritin should invite a better workup, not a new universal rule.