Iron deficiency anemia diagnostic criteria in primary care

Updated American Society of Hematology guidance places the principal cutoff at serum ferritin ≤30 ng/mL in adults without inflammation and supports thresholds up to 50 ng/mL in selected higher-risk populations.

That change is not a cosmetic adjustment to a laboratory reference range. It alters case identification in primary care, where iron deficiency is often evaluated before the cause is established and where a normal mean corpuscular volume may create false reassurance. The appropriate clinical diagnostic criteria for adult iron deficiency anemia therefore require more than a haemoglobin value and a ferritin result interpreted in isolation. Haemoglobin establishes anaemia; ferritin estimates iron stores; transferrin saturation helps determine whether iron is actually available for erythropoiesis.

The ferritin cutoff is a diagnostic decision, not a biological cliff

Serum ferritin is useful because it generally reflects stored iron. It is also an acute-phase reactant, which is precisely why the apparently simple question—whether ferritin is normal—becomes methodologically unstable in outpatient medicine. A patient with infection, cancer, chronic inflammation or another inflammatory condition may have ferritin concentrations elevated by the underlying process, while tissue iron availability remains inadequate.

For adults without evidence of inflammation, the updated ASH threshold is ferritin ≤30 ng/mL. This is more clinically defensible than treating values in the 15–30 ng/mL range as automatically reassuring. The older, narrower threshold may retain specificity for depleted stores, but its sensitivity for detecting iron deficiency in the primary-care population is inadequate if used as the only gatekeeper.

A large primary-care cohort study involving 255,351 adults illustrates the practical consequence. When different ferritin thresholds were applied, the recorded rates of iron deficiency were:

These figures do not prove that every additional diagnosis generated by a higher cutoff represents clinically significant deficiency. They do demonstrate that the selected threshold determines the apparent burden of disease. A laboratory interval is not an oracle; it is a classification rule with consequences for sensitivity, specificity and subsequent investigation.

A ferritin value above 15 ng/mL is not evidence that iron deficiency has been excluded. It may simply indicate that the diagnostic threshold was set too low.

The distinction matters particularly in early deficiency, in patients with ongoing blood loss and in individuals whose haemoglobin has not yet fallen substantially. Iron depletion can precede overt anaemia. If clinicians require both a low haemoglobin and a markedly depleted ferritin before acting, the diagnostic process identifies the endpoint of the disease rather than its clinically useful beginning.

Why haemoglobin alone is an incomplete endpoint

Iron deficiency anaemia is, by definition, associated with anaemia, but the diagnostic process should not be reduced to haemoglobin classification. A patient can have iron deficiency before the haemoglobin crosses the laboratory’s lower reference boundary. Conversely, anaemia may be caused by chronic kidney disease, inflammation, haemolysis, nutritional deficiency, marrow disease or occult blood loss unrelated to iron depletion.

The minimum outpatient interpretation should connect three domains:

• Haemoglobin, to establish whether anaemia is present and to assess its severity in the context of the laboratory and clinical setting.
• Serum ferritin, to estimate iron stores, with the threshold adjusted for inflammatory and high-risk states.
• Transferrin saturation, or TSAT, to assess the proportion of circulating transferrin carrying iron and to identify restricted iron transport.

Mean corpuscular volume remains useful for pattern recognition, but it is not a sufficiently sensitive exclusion test. Microcytosis supports iron-restricted erythropoiesis, yet iron deficiency may be present before red-cell indices become clearly abnormal. A normal MCV should not overrule a compatible ferritin and TSAT pattern, particularly when there is chronic blood loss or an inflammatory comorbidity.

The same caution applies to red-cell distribution width and other indices sometimes used to distinguish iron deficiency from thalassaemic traits. They can contribute to a differential diagnosis, but none substitutes for the clinical question of whether iron stores and iron delivery are adequate. A microcytic anaemia diagnostic algorithm that begins and ends with MCV is convenient, not comprehensive.

Higher thresholds in heavy menstrual bleeding and pregnancy

The diagnostic threshold should be more conservative in populations where the pre-test probability of iron deficiency is already high. ASH guidance suggests considering ferritin thresholds up to 50 ng/mL or lower in higher-risk groups, including people with heavy menstrual bleeding and pregnant individuals with anaemia.

This does not mean that every ferritin result below 50 ng/mL proves iron deficiency in every adult. It means that the cost of delayed recognition is different in a patient with continuing substantial menstrual blood loss or pregnancy-associated haematological demands. The same numerical result can carry different diagnostic weight when the prior probability changes.

Heavy menstrual bleeding should therefore be treated as a clinical variable, not as background documentation. The relevant history includes persistence, impact on daily functioning, the presence of clots or flooding, intermenstrual bleeding, postpartum status and known uterine or endocrine disease. It also determines whether oral replacement alone is likely to address the problem. Replenishing iron while leaving ongoing blood loss unexplained is not disease management; it is inventory correction.

In pregnancy, the interpretation must also be integrated with gestational context, haemoglobin, symptoms, bleeding history and the broader obstetric assessment. The threshold up to 50 ng/mL is intended to reduce missed deficiency in a higher-risk setting, not to convert ferritin into a stand-alone treatment trigger without diagnostic review.

A practical approach is to separate two questions that are often collapsed:

1. Is iron deficiency plausible or likely at this ferritin level in this patient?

2. What is causing the deficiency, and is the cause still active?

The first question determines whether iron studies support deficiency. The second determines whether outpatient management is safe and sufficient.

Chronic inflammation changes the meaning of ferritin

Inflammation is where the single-marker approach fails most conspicuously. Ferritin can rise in response to inflammatory signalling and may appear acceptable even when available iron is inadequate. In this setting, the statement that ferritin above 30 ng/mL excludes iron deficiency is not clinically defensible.

For adults with chronic inflammation, cancer or infection, ASH guidance supports interpreting ferritin together with TSAT. A ferritin level below 100 ng/mL alone has low sensitivity in this population: 0.38. When ferritin below 100 ng/mL is combined with TSAT below 20%, sensitivity increases to 0.91. The difference is not an argument for treating every low TSAT as proof of absolute iron depletion. It is an argument against using ferritin as though inflammation had not altered its operating characteristics.

TSAT below 20% is particularly relevant because it indicates restricted iron transport, whether the underlying mechanism is depleted stores, impaired mobilisation or both. The distinction between absolute and functional iron deficiency remains clinically important. Absolute deficiency describes inadequate total iron stores. Functional deficiency describes insufficient iron availability for erythropoiesis despite the presence of stored iron that cannot be adequately mobilised.

That distinction should prevent two opposite errors. The first is dismissing a low TSAT because ferritin is not very low. The second is labelling every inflammatory anaemia as iron deficiency and proceeding without investigating the inflammatory or systemic disease. Laboratory interpretation should narrow the differential; it should not abolish it.

In inflammatory disease, ferritin is a noisy proxy for iron stores. TSAT is not a decorative add-on to the panel; it changes the sensitivity of the diagnostic strategy.

A useful outpatient panel in this setting therefore includes ferritin and TSAT alongside the complete blood count and the clinical assessment of inflammation. If ferritin and TSAT are discordant, the discordance is information. It may indicate inflammatory sequestration, mixed anaemia, ongoing blood loss, chronic kidney disease or an analytical and biological limitation that requires reassessment rather than reflexive certainty.

Gastrointestinal workups require a more sensitive threshold

Anaemia in an adult patient undergoing evaluation for a gastrointestinal source presents a different diagnostic problem. Here the objective is not merely to identify severe iron-store depletion. It is to avoid missing iron deficiency that may signal chronic occult blood loss.

For anaemic patients being evaluated for gastrointestinal causes, a ferritin cutoff below 45 ng/mL improves diagnostic sensitivity compared with traditional cutoffs below 15 ng/mL. That threshold should not be interpreted as proof that the gastrointestinal tract is the source. It indicates that iron deficiency is sufficiently plausible to justify appropriate evaluation in the clinical context.

The distinction is essential. A low ferritin supports iron deficiency; it does not identify the lesion, establish malignancy, or determine whether the bleeding is gastrointestinal. The workup still depends on age, sex, symptoms, medication exposure, family history, bowel habit changes and the severity and persistence of anaemia. The laboratory result sets the diagnostic direction, not the final anatomical diagnosis.

In primary care, the common failure mode is sequential fragmentation: anaemia is documented, iron is prescribed, haemoglobin is rechecked, and investigation stops when the count improves. Haemoglobin response can confirm that iron was relevant to the erythropoietic deficit, but it does not explain why iron was lost. A corrected endpoint does not retrospectively make the cause benign.

The second failure mode is the opposite: a ferritin slightly above the traditional threshold is used to defer evaluation. In a patient with anaemia and possible gastrointestinal blood loss, the <45 ng/mL threshold is more appropriately regarded as a sensitivity-oriented trigger. It reduces the risk that a low-grade or partially masked deficiency is dismissed before the cause has been considered.

A primary-care diagnostic algorithm that reflects the evidence

A practical algorithm should be explicit enough to prevent predictable errors, but not so rigid that it ignores the patient’s context.

1. Confirm the haematological abnormality.

Review haemoglobin, MCV and the complete blood count rather than treating an isolated haemoglobin value as the diagnosis. Note whether the pattern is microcytic, normocytic or mixed. Normocytic indices do not exclude early iron deficiency.

2. Order ferritin with TSAT when the clinical probability is meaningful.

Ferritin alone may be adequate as an initial test in adults without inflammatory disease, but adding TSAT becomes important when inflammation, infection, cancer, chronic kidney disease or a discordant blood picture is present.

3. Apply the threshold to the clinical context.

Use ferritin ≤30 ng/mL as the principal adult threshold in the absence of inflammation. Consider a threshold up to 50 ng/mL or lower in higher-risk groups such as patients with heavy menstrual bleeding or pregnant individuals with anaemia. Do not use 30 ng/mL as an exclusion boundary in inflammatory states.

4. Interpret ferritin below 100 ng/mL with TSAT in inflammatory states.

A TSAT below 20% strengthens the diagnosis of iron-restricted erythropoiesis. The combination is substantially more sensitive than ferritin below 100 ng/mL alone, with sensitivity reported as 0.91 versus 0.38.

5. Escalate the causal assessment when anaemia is unexplained or persistent.

Menstrual blood loss may be sufficient to explain deficiency in some patients, but it should not automatically be used to explain every case. In men, postmenopausal patients and others with unexplained anaemia, gastrointestinal causes require particular attention.

6. Separate replacement from investigation.

Starting iron therapy does not complete the diagnostic process. It addresses the deficit while the clinician determines whether blood loss, malabsorption, dietary insufficiency, inflammatory disease or another condition is maintaining it.

7. Reassess discordant results rather than forcing a label.

Ferritin that appears acceptable alongside TSAT below 20%, persistent anaemia or active inflammation is not a clean negative result. It is a reason to revisit the model of the case.

This approach is intentionally less convenient than a single ferritin cutoff. Clinical medicine rarely improves when convenience is mistaken for validity.

Treatment decisions should follow, not replace, diagnosis

Outpatient anaemia management protocols often become treatment-centred because iron is inexpensive, familiar and easy to prescribe. That does not make the diagnostic question optional. Iron supplementation may be appropriate, but the indication, route, duration and follow-up depend on the severity of anaemia, tolerance, suspected cause, ongoing blood loss and comorbidity.

The diagnostic threshold should not be inflated merely to justify treatment in a patient with nonspecific fatigue, nor should treatment be withheld because ferritin is not below an obsolete value. Both positions misuse the laboratory test. The purpose of a higher threshold is to improve sensitivity where the consequences of missed deficiency are clinically relevant, while preserving the requirement to interpret the result against haemoglobin, TSAT and the underlying disease context.

A further point deserves emphasis: symptomatic improvement, if it occurs, is not a validated diagnostic endpoint by itself. Symptoms such as fatigue, reduced exercise tolerance and impaired concentration are nonspecific. A response to iron may support the diagnosis, but it cannot establish the cause of deficiency or rule out coexisting disease.

The same evidentiary discipline applies to follow-up. A rising haemoglobin is reassuring but incomplete. The clinician still needs to know whether iron stores are being restored, whether blood loss continues and whether the initial diagnosis adequately explains the full laboratory pattern. When anaemia persists despite plausible replacement, the differential should expand rather than simply prompting repeated prescriptions.

What the updated thresholds change in practice

The revision is best understood as a change in diagnostic sensitivity, not as permission to diagnose iron deficiency from any ferritin value that is merely below the population reference interval. Its practical implications are specific:

• Ferritin ≤30 ng/mL supports iron deficiency in adults without inflammation.
• Ferritin up to 50 ng/mL may be clinically relevant in higher-risk settings, including heavy menstrual bleeding and pregnancy-associated anaemia.
• Ferritin alone is insufficiently sensitive in chronic inflammation, cancer or infection.
• Ferritin below 100 ng/mL combined with TSAT below 20% is substantially more sensitive than ferritin below 100 ng/mL alone in inflammatory states.
• Ferritin below 45 ng/mL is a more sensitive threshold when anaemic patients are being assessed for gastrointestinal causes.
• Traditional values of 12–15 ng/mL should not be treated as universal exclusion thresholds.

The clinical applicability is therefore straightforward, although not simplistic: primary-care clinicians should stop using a profoundly low ferritin as the sole definition of iron deficiency. The more defensible strategy is context-adjusted interpretation of ferritin, systematic use of TSAT when inflammation may distort ferritin, and a separate causal workup when anaemia or blood loss remains unexplained.

The updated thresholds do not create a miracle diagnostic shortcut. They correct a familiar methodological error—confusing a highly specific late marker with a sufficiently sensitive screening criterion. For independent practitioners, the recommendation is strict: use the higher thresholds where the guideline context supports them, pair ferritin with TSAT in inflammatory states, and do not allow iron replacement to substitute for finding the reason the patient became iron deficient.