The research trail
The papers that set every threshold and pattern.
Organized by lab marker, then by life stage. Every threshold traces to a primary paper or a consensus body. Preliminary evidence is tagged.
Version 0.1.1 (April 2026) · 700+ appraised papers across 16 patterns.
Evidence by marker
Per-marker thresholds and sources
Ferritin
Ferritin measures the protein your body stores iron in. When iron supply is healthy, ferritin stays above a certain level. When supply drops, ferritin drops first — before red blood cells change, before hemoglobin falls, and often before you feel different. Ferritin is the single most useful iron lab. It is also distorted by inflammation, so we read it in context.
| Life stage | Thresholds (μg/L) | Badge | Primary sources |
|---|---|---|---|
| Non-pregnant premenopausal | Replete ≥ 50 · Incipient 30–50 · Depletion 25–30 · IDE < 25 | High quality | Mei 2021 · Addo 2025 · Galetti 2021 · Tarancon-Diez 2022 |
| Pregnancy T1 | IDE threshold: < 25.8 · Adequate reserves ≥ 60 | High quality | Mei 2025 · McCarthy 2024 · Bothwell 2000 |
| Pregnancy T2 | IDE threshold: < 18.3 | High quality | Mei 2025 |
| Pregnancy T3 | IDE threshold: < 19.0 | High quality | Mei 2025 |
| Postpartum (within 6 months) | IDE threshold: < 30 | High quality | Milman 2011 · Butwick 2021 · Akesson 2002 |
| Perimenopausal | IDE threshold: < 25 | High quality | Kim 2025 · Parker 2020 · Means 2024 · Merlo 2023 |
| Postmenopausal | IDE threshold: 22–38 (age-dependent) · Hematologic response: 25 | High quality | Means 2024 · Foy 2019 · Kim 2025 · Warne 2017 · Wouters 2019 |
| Overload evaluation trigger (any life stage) | Ferritin > 1000 warrants evaluation regardless of TSAT | High quality | Waalen 2008 · Adams 2005 |
Hemoglobin
Hemoglobin is the oxygen-carrying protein inside your red blood cells. Once iron supply to the bone marrow has been constrained long enough, hemoglobin falls. When it falls below a life-stage reference threshold, the research calls the pattern anemia.
| Life stage | Anemia threshold (g/L) | Badge | Primary sources |
|---|---|---|---|
| Non-pregnant premenopausal | < 120 | Consensus | Braat 2024 · Addo 2019 · WHO |
| Pregnancy T1 | < 110 | Consensus | Braat 2024 · Young 2023 · WHO |
| Pregnancy T2 | < 105 | Consensus | Braat 2024 · Young 2023 · Shinar 2018 |
| Pregnancy T3 | < 100 | Consensus | Young 2023 |
| Postpartum (1-week reference) | < 110 | High quality | Milman 2011 |
| Postpartum (8-week reference) | < 120 | High quality | Milman 2011 |
| Postmenopausal | < 120 (WHO); some evidence suggests 130 | Preliminary | Wouters 2019 · Netz 2024 |
| Severe anemia (any life stage) | < 70 triggers urgent evaluation | High quality | Mansukhani 2023 · Butwick 2016 |
| Severe postpartum (within 6 months) | < 80 triggers urgent evaluation | High quality | Mansukhani 2023 · Butwick 2016 |
Transferrin saturation (TSAT)
TSAT measures how much of your iron-transport protein is actually carrying iron right now. It is more variable than ferritin — it shifts by time of day, by how recently you ate, and during inflammation. We read it in combination with ferritin, not alone.
| Threshold | Value | Badge | Primary sources |
|---|---|---|---|
| Low — reduced iron delivery to cells | < 20% | High quality | EASL 2022 · Cacoub 2019 · Lim 2014 |
| Overload evaluation trigger (premenopausal) | > 45% paired with ferritin > 200 | Consensus | EASL 2022 · Adams 2005 · Allen 2008 |
| Overload evaluation trigger (postmenopausal) | > 50% paired with ferritin > 300 | Consensus | EASL 2022 · Adams 2005 |
| Sensitive screening trigger (female) | > 32% | High quality | Lim 2014 · Lou 2024 |
Red cell indices (MCV, RDW)
MCV measures the average size of your red blood cells. RDW measures how much cell size varies within your population of red cells. Together they give clues about what is driving an abnormal hemoglobin — iron deficiency tends to produce small cells with wider size variation; hemoglobinopathy traits tend to produce small cells with tighter size variation; B12 or folate deficiency tends to produce larger cells.
| Threshold | Value | Badge | Primary sources |
|---|---|---|---|
| Microcytic (small red cells) | MCV < 80 fL | High quality | WHO reference ranges · multiple cohort validations |
| Macrocytic (large red cells) | MCV > 100 fL | Consensus | Points away from iron — most commonly B12 or folate deficiency |
| Elevated RDW (supports iron-deficiency over hemoglobinopathy) | ≥ 14.5% | Preliminary | Van Zeben 1990 · multiple cohorts (not yet meta-analyzed) |
C-reactive protein (CRP)
CRP is a general marker of inflammation. Ferritin is an acute-phase reactant — it rises during infection, injury, or chronic inflammation regardless of actual iron stores. When CRP is above 5 mg/L, we flag ferritin as potentially inflated and lower the confidence of iron-stage classifications. Above 10 mg/L, we treat ferritin as unreliable without clinician-interpreted adjustment.
| Threshold | Value | Badge | Primary sources |
|---|---|---|---|
| Inflammation flag — ferritin potentially inflated | CRP ≥ 5 mg/L | High quality | Namaste 2017 · Thurnham 2010 · Williams 2019 · Ko 2024 |
| High inflammation — ferritin unreliable at face value without adjustment | CRP ≥ 10 mg/L | Consensus | Thurnham-BRINDA correction factors are population-level tools; applying them to a single draw is clinician-interpreted. |
Evidence by pattern
Per-pattern citations
| User-facing name | Research term | Citations |
|---|---|---|
| Your iron levels look healthy | Iron-replete (Stage 0) | Addo 2025 · Galetti 2021 |
| Early signs of iron running low | Incipient iron deficiency (Stage 1) | Galetti 2021 · Tarancon-Diez 2022 |
| Iron stores running low, not yet anemic | Iron depletion without anemia (Stage 2) | Addo 2025 · Mei 2021 |
| Early iron shortage showing in red blood cells | Iron-deficient erythropoiesis (Stage 3) | Primary: Mei 2021 · Addo 2025 · Brownlie 2004 Supporting: Mei 2025 · Choorapoikayil 2025 |
| Iron deficiency anemia | Iron deficiency anemia (IDA, Stage 4) | Mei 2021 · Addo 2025 · Skikne 2011 |
| Your ferritin cannot be read at face value right now | Inflammation inflates ferritin | Namaste 2017 · Thurnham 2010 · Williams 2019 · Ko 2024 |
| Iron in the body, but locked away | Functional iron deficiency | Primary: Kuragano 2020 · Rambod 2008 · Dignass 2018 · Skikne 2011 Supporting: Cacoub 2022 · Rohr 2023 · Svenson 2021 · Deniz 2022 |
| Anemia that comes from long-term inflammation | Anemia of chronic disease (ACD) | Primary: Skikne 2011 · Rohr 2023 · Dignass 2018 Supporting: Svenson 2021 · Murawska 2016 · Auerbach 2025 |
| Your iron markers are above the evaluation threshold (premenopausal) | Ferritin + TSAT above premenopausal overload evaluation threshold | Primary: EASL 2022 · Adams 2005 · Allen 2008 Supporting: Lim 2014 · Schmitt 2005 · Cullis 2018 · Adams 2015 |
| Your iron markers are above the evaluation threshold (postmenopausal) | Ferritin + TSAT above postmenopausal overload evaluation threshold | EASL 2022 · Adams 2005 · Warne 2017 |
| Very high ferritin — evaluation needed | Ferritin > 1000 μg/L | Primary: Waalen 2008 · Adams 2005 Supporting: Cullis 2018 · Goubran 2024 |
| Small red blood cells with adequate iron — thalassemia evaluation recommended | Microcytic anemia with normal or elevated iron studies | Primary: Cappellini 2020 · Van Zeben 1990 · Hans 2024 Supporting: Mondal 2021 · Thambiraj 2023 · Mettananda 2020 |
| Larger red blood cells — B12 or folate evaluation recommended | Macrocytic anemia signal | Out of this tool's iron scope. Routed to clinician for B12 / folate / thyroid evaluation. |
| Anemia present, but iron stores look adequate | Anemia with non-deficient ferritin, undifferentiated | Primary: Mei 2024 (ferritin staging boundary) · Skikne 2011 (anemia differential framework) Supporting: Addo 2025 · Cappellini 2020 (morphology-based subtype routing) Routes to MCV-based subtype differentiation when anemia is present but ferritin sits at or above the depletion threshold. |
Sourcing
The questions we investigated
We ran dedicated literature reviews for each of the questions below, prioritizing systematic reviews, meta-analyses, and large cohort or individual-participant-data studies. Where those weren't available, we appraised the best primary evidence we could find and tagged it accordingly.
- What ferritin levels signal iron deficiency at each life stage, and where are the overload thresholds?
- What hemoglobin levels define anemia across non-pregnant, pregnancy trimesters, postpartum, and menopause?
- Which secondary iron tests (TSAT, TIBC, serum iron, sTfR) add information beyond ferritin?
- How does inflammation distort ferritin readings, and when should ferritin not be read at face value?
- Which symptoms are research-linked to low iron, and at what stage do they typically appear?
- How do pregnancy, postpartum, perimenopause, and postmenopause shift iron physiology?
- How do dietary promoters and inhibitors of iron absorption work?
- What does the research say about iron repletion — oral vs. IV, dosing patterns, expected response?
- How do functional iron deficiency, anemia of chronic disease, and iron overload present?
- What questions are worth bringing to a clinician?
- Which populations are underrepresented in the literature, and where does that limit the tool's reach?
Limitations
Where the research is thinner
- Underrepresented populations. Thresholds derived predominantly from Western cohorts may not translate identically to every population.
- Pregnancy and postpartum. Threshold work is newer — strong, but built on fewer pooled studies than the non-pregnant literature.
- Perimenopausal transition. The menopausal-transition literature is smaller than the pre- and postmenopausal literature. Transitional thresholds carry more uncertainty.
- Gender-affirming hormone therapy (GAHT). Small but growing research base. We apply reference ranges per direction of hormone therapy, with a confidence caveat in the first 12 months.
- Symptom attribution in early-stage iron deficiency. Evidence at the earliest stages (Stages 1–2) is still developing. Effect sizes in randomized trials are modest and individual responses vary.
References
Every paper the tool relies on
Organized by what the paper contributes to the tool. DOI links go to the publisher record. Author-year citations in the tables above resolve to entries in this list. Click any citation chip in a row to jump to the reference.
Ready?
Read my labs
Start nowTakes about two minutes. Free. No account needed.