Oxygen carrying, blood sugar & the boring causes

What actually decides this outcome, in order of size

This is the one fatigue pathway where the effect sizes are known, the tests are cheap and the fix is specific. Ranked by how often each one turns out to be the answer:

1. Iron stores, well before anemia, at a ferritin the laboratory prints as normal. The trial that settles this randomized 198 menstruating women with a ferritin under the study threshold and a normal hemoglobin to 80 mg/day of elemental iron or placebo for 12 weeks: fatigue fell 47.7% against 28.8%, a difference of 18.9% (95% CI 3.2 to 34.5, P=0.02) Vaucher 2012. Non-anemic iron deficiency is a real and treatable cause of tiredness, and a normal Complete Blood Count (CBC) with Differential does not exclude it because hemoglobin is the last thing to fall, roughly 120 days behind the stores.
2. Where the ferritin threshold actually sits, which is not at 15 ng/mL. Physiologically derived thresholds in blood donors put the value for iron-deficient erythropoiesis far above the bottom of most laboratory reference ranges Addo 2022, which is why a result flagged as normal by the lab is routinely the finding. A reference interval is the central 95% of a population, not a statement about whether your marrow has iron.
3. B12 status, which one number cannot report, because only part of it reaches a receptor. Serum cobalamin is mostly bound to haptocorrin and unavailable to cells, so the assay measures a pool that is not the pool that matters; guidelines are explicit that a clinical picture with a borderline value should be resolved with a metabolite Devalia 2014, and the biomarker literature says the same Harrington 2024.
4. The post-meal glucose curve, not the fasting number, which can sit at 88 mg/dL through all of it. Fasting glucose can sit at 88 mg/dL while a mixed meal drives an excursion above 180 mg/dL and a reactive fall two hours later, and the fall is what gets described as an afternoon crash. Fasting insulin catches the compensating half of that at the beta cell, years before the glucose itself moves off 88 mg/dL.
5. Thyroid, which acts through a nuclear receptor while the test reports a pituitary signal. A frankly high TSH with a low free T4 is a finding, and the pituitary takes 6 to 8 weeks to report a change. Subclinical hypothyroidism is not the same thing: in older adults randomized to levothyroxine, symptom and tiredness scores did not improve against placebo Stott 2017. Being in the upper 2.5% of a reference range is not, by itself, a reason to be tired.

The order to run these in, and what has to be true first

Draw before you buy anything, because every item on this pathway is cheap enough that people take all of them at once and then cannot tell which one worked. The order is set by which result invalidates the others, and hs-CRP invalidates ferritin.

1. One tube answers most of it, and hs-CRP has to be on it. Ferritin with hs-CRP (High-Sensitivity C-Reactive Protein) on the same draw, Complete Blood Count (CBC) with Differential for the mean corpuscular volume, Iron Panel (Iron, TIBC, Transferrin Saturation) with Transferrin for saturation, Vitamin B12 with Methylmalonic Acid (MMA), Folate, Serum, HbA1c (Hemoglobin A1c) with Fasting Insulin, and TSH (Thyroid-Stimulating Hormone) with Free T4 (Thyroxine). Ferritin is an acute-phase reactant, so a value drawn while CRP is elevated overstates the stores behind it and a normal-looking number can be masking depletion.
2. Correct iron as a single, at 60 to 120 mg of elemental iron rather than the 18 mg in a blend. Iron or Iron Complex where ferritin is low and transferrin saturation is under 20%; Desiccated Beef Liver supplies heme iron with the food matrix but not at a therapeutic dose. Alternate-day dosing exists as a strategy because a dose of iron raises hepcidin for roughly 24 hours and hepcidin degrades ferroportin on the enterocyte, so a second dose the same day is absorbed worse than the first.
3. Correct B12 as the right form, and only after MMA has been read, because a dose destroys the serum assay within 24 hours. Methylcobalamin (B12) or B12, with Methylfolate (5-MTHF) only if folate is the deficient one, because folate repletion can improve the hematology of B12 deficiency while the neurological damage continues. B-Complex is a hedge for the rest of the water-soluble half and is not a treatment for a deficiency any single in it could correct, since renal clearance removes most of it within hours.
4. Then the glycemic half, which is a food problem with a pharmacological assist acting on AMPK. Berberine activates AMPK and lowers post-meal excursions; Dihydroberberine exists because berberine's oral bioavailability is under 1% and is limited by P-glycoprotein efflux. Myo-Inositol acts as a second messenger downstream of the insulin receptor, Chromium is only interesting where intake is genuinely low, and Magnesium is a cofactor for the hexokinase and phosphofructokinase steps that spend ATP before any is made.
5. Electrolytes is the one item here that is not a nutritional status. Sodium and potassium are replaced against acute losses, they change how somebody feels within 30 to 60 minutes, and they will not change any number on the panel above. Vitamin D sits at the end of this list because 25(OH)D has a half-life of 2 to 3 weeks and is not usually the answer to this complaint.

What gets bought for this that cannot move it

A multivitamin will not refill iron stores, and that is the commonest wasted year on this pathway. A blend carries roughly one reference intake, 18 mg, as a maintenance dose against daily losses. Refilling a depleted marrow is a therapeutic dose of a single for 8 to 12 weeks, and the reason people conclude that iron did not work for them is almost always that they never took a repleting dose of it.

Serum B12 stops meaning anything the moment you supplement, and keeps looking reassuring for months. An oral dose raises the serum value within a day while methylmalonyl-CoA mutase is still starved, which is exactly why the metabolite is the read-out and the single number is not Devalia 2014. Somebody who supplements first and tests second has destroyed the test and not treated the problem.

Glucose support is the wrong lever for a flat curve. If HbA1c (Hemoglobin A1c) is 5.1% and fasting insulin is 4 uIU/mL, an AMPK activator has no excursion to blunt, and the fatigue is coming from somewhere else. Berberine also inhibits CYP3A4 and P-glycoprotein, so it is the item on this pathway most likely to change the exposure of a prescription somebody is already taking.

And if every number here comes back clean, the answer is upstream of this page. A flattened cortisol curve produces the same description and needs a four-point measurement rather than a morning draw, which is Adrenal, cortisol rhythm & stress-driven fatigue. Unrefreshing sleep with normal iron and normal thyroid is What's keeping you awake — the upstream causes. Neither is on this shelf, and continuing to buy from it is how a year goes.

How you would know it was working, on a real read-out and a real timescale

Every claim here has a number under it, so this pathway is falsifiable in a way most of the site is not. The prediction: if iron or B12 was the cause, energy changes before the hematology does, and the marker moves first.

• Ferritin at 12 weeks with hs-CRP (High-Sensitivity C-Reactive Protein) beside it. Expect stores to rise well before Complete Blood Count (CBC) with Differential indices do, because a reticulocyte released today becomes an erythrocyte with a 120-day lifespan and the mean corpuscular volume is an average over that whole population. A ferritin that rose while CRP also rose has reported inflammation, not iron.
• Methylmalonic Acid (MMA) rather than Vitamin B12 to judge repletion. Methylmalonic acid falls only once methylmalonyl-CoA mutase has restarted, typically over weeks, and it is the number that can still say the treatment worked after the serum value has been rendered uninterpretable. Homocysteine moves with both B12 and folate and therefore cannot separate them.
• HbA1c (Hemoglobin A1c) at 12 weeks, with Fasting Insulin on the same draw. Insulin moves first because compensation precedes decompensation; if the glycemic story is right, insulin should fall while HbA1c barely moves, and if neither moves the excursion was not the problem.
• TSH (Thyroid-Stimulating Hormone) with Free T4 (Thyroxine) once, at the same time of day. TSH is pulsatile with a nocturnal peak and is roughly 30% lower in the afternoon, so a morning and an afternoon result are not comparable and a difference between them is the clock rather than the gland.

What will fool you. Iron taken with coffee or tea loses a large fraction of its non-heme absorption to polyphenol chelation, so a null result on a morning dose taken with breakfast is a null result about the timing. Any supplement containing biotin can distort immunoassay platforms that use streptavidin-biotin capture, which includes the common TSH, ferritin and B12 assays, so a high-dose B-complex belongs 48 hours away from the draw. And a week of good sleep will improve fatigue by more than any item on this page, which is why the retest matters more than the diary.

Sources read for these sections

• Vaucher P. Effect of iron supplementation on fatigue in nonanemic menstruating women with low ferritin: a randomized controlled trial. Canadian Medical Association Journal 2012;184(11):1247-54 · PMID 22777991
• Devalia V. Guidelines for the diagnosis and treatment of cobalamin and folate disorders. British Journal of Haematology 2014;166(4):496-513 · PMID 24942828
• Addo OY, et al. Physiologically based serum ferritin thresholds for iron deficiency in women of reproductive age who are blood donors. Blood Advances 2022 · PMID 35404995
• Harrington DJ. The application and interpretation of laboratory biomarkers for the evaluation of vitamin B12 status. Ann Clin Biochem 2024 · PMID 39367523
• Stott DJ, et al. Thyroid Hormone Therapy for Older Adults with Subclinical Hypothyroidism. New England Journal of Medicine 2017 · PMID 28402245