Chronic Fatigue: Finding the Cause

Healthcare · §611

You have been tired for a year. Maybe three. Long enough that "I'm just tired" has stopped being a sentence and become a personality — the gym lapsed, dinners shrank, the people who used to invite you stopped. Here is the thing almost nobody tells you: tiredness that will not quit almost always has a findable, treatable cause hiding behind it — iron, B12, thyroid, sleep apnea, depression, the recovery from an infection your body never finished. A blood panel and an honest sleep study answer the question, and the version of you that gets the answer is the one with afternoons again.

Test · As-needed Evidence Moderate Chapter Healthcare

The main lift here is daily energy — when the testing lands on a real driver, the recovery curve is steep. Focus and mood follow, because brain fog and the bad days are downstream of most of what the labs find. Sleep does too if the answer is apnea, which is one of the most under-caught conditions in adults past forty. The price is a lab panel and an afternoon at the clinic; every step of the pathway is solid, guideline-anchored medicine. The honest catch: sometimes what the testing rules out matters more than what it finds — and that is its own kind of useful.

Tiredness is your brain telling you that supply is below demand. Supply is oxygen reaching your cells, the hormone that sets your metabolic floor, the neurotransmitters that drive motivation, the slow-wave sleep that resets you overnight, and the immune system finally standing down after the last virus cleared. Demand is the load of your day. Because the signal is downstream of so many separate systems, the cause is rarely guessable from the outside — but the list of usual suspects is short, and they are nearly all testable.

Five families handle the bulk of cases. Sleep — either not enough of it, or hours that look fine on the bedside clock but are getting chewed up by an airway that keeps collapsing. Thyroid — the hormone, called thyroxine, that sets your basal metabolic rate; too little of it and the whole engine runs cold. Iron and B₁₂ — the substrate your blood and your brain run on, depleted slowly by menstruation, by certain common medications, by age, by a vegetarian diet without a B₁₂ supplement. Mood — depression compresses energy more than most readers will admit to, and shows up as physical tiredness as often as it does as sadness. And the post-infection state — your immune system left in the "on" position months after the bug itself is gone, which the COVID years made every primary care doctor familiar with.

What the labs actually catch

Each of the five families has serious trial backing. The one that surprises clinicians most often: iron deficiency without anemia. The standard blood test looks at hemoglobin and calls it normal — but ferritin, the storage iron, can be empty long before your hemoglobin drops, and an empty ferritin tank starves your mitochondria and the enzymes your brain needs to make neurotransmitters. This is most of why a fifth to a third of menstruating women feel tired all the time.

B₁₂ deficiency is the second under-detected finding. Serum B₁₂ on its own is a noisy proxy — a chunk of the population walks around with results in the so-called normal band while their cells cannot actually use the vitamin. A backup test called methylmalonic acid catches that functional deficiency the basic number misses Stabler 2013. Two of the most-prescribed drugs in the world quietly deplete B₁₂ over years: metformin (one in fourteen long-term users becomes deficient in the RCT data) de Jager 2010 and the acid-suppressing PPIs like omeprazole (the odds of deficiency rise 65% after two years on them, in a study of 26,000 patients) Lam 2013.

Sleep apnea — your airway collapsing dozens of times an hour without you remembering it — is present in roughly one in ten men 30 to 49 and closer to one in six in older bands Peppard 2013. Women are systematically under-diagnosed because they tend to show up with the fatigue and the insomnia rather than the loud snoring everybody pattern-matches to. Overt low thyroid reliably resolves once it is replaced Garber 2012. Depression carries fatigue as one of its formal diagnostic symptoms, and as one of the most persistent residual ones when treatment only partially works Targum and Fava 2011. Post-viral fatigue runs about 11% at six months across a number of different infections — mono, Q fever, and Ross River virus all showed the same rate in an Australian cohort that followed people prospectively Hickie 2006 — and chronic fatigue is the single most-reported persistent symptom in the largest UK study of hospitalised COVID survivors Evans 2021.

What it costs to keep ignoring it

The version of you carrying this for another year gradually stops doing the things you used to be the person who did. The hill to the bus is longer than it was. You decline the dinner. The book on the nightstand stays where it is. You start lying about why you can't make it. Over a year that becomes a habit; over five years a habit becomes a personality, and the social signal flips — the people who used to invite you stop, and you stop noticing they stopped.

The second-order damage is the part that actually shortens your life. The exercise you can't do becomes the weight that worsens the apnea that was already fragmenting your sleep, in a feedback loop you cannot break without naming. The thyroid running cold for a decade quietly worsens your cholesterol. The depression nobody named carries cardiovascular mortality of its own. Most readers carrying chronic fatigue have been carrying it for two or three years before they take it seriously enough to test. The cost of that delay, compounded across all the things you didn't do, is the years themselves.

The lab panel that answers the question

This is what to ask for at the appointment, or to order through a direct-to-consumer service in places that allow it. The first pass is one blood draw and reads cheap. Resist the urge to skip the items that "feel fine" — the whole point is to catch the ones that are quietly off.

First-pass panel:

Complete blood count with differential — for anemia and the rest of the red cell picture.
Ferritin — and read it against a threshold of 50 µg/L, not the bottom of the lab's "normal" range. Under 30 is clearly low; under 50 is the band where supplementation has been shown to lift fatigue in trials Vaucher 2012.
TSH — the thyroid screen. If it is abnormal, the lab should reflexively add a free T4 on the same sample Garber 2012.
Vitamin B₁₂. If the result comes back in the low-normal band of 200 to 400, ask for a methylmalonic acid as the follow-up — that is the test that confirms a functional deficiency the basic number can miss Stabler 2013.
HbA_1c and a fasting glucose — for diabetes and pre-diabetes, which often present as fatigue before they present as anything else.
25(OH) vitamin D — common, treatable, cheap.
A comprehensive metabolic panel for the liver, kidney, and electrolyte basics.

If the first pass comes back clean and the fatigue is still there in a month, the second pass is where most of the hidden cases live.

Second-pass when the first pass is unrevealing:

A home sleep test, or a full in-lab study if your insurance covers it. Order this even if you do not snore, even if you are not overweight, and especially if you are a woman — those are the cases that get missed for years Peppard 2013.
A mental-health screen. The PHQ-9 for depression and the GAD-7 for anxiety take five minutes each and are validated; they are not the same as "asking how you are."
A real medication review. Sedating antihistamines like diphenhydramine, beta-blockers, gabapentin, opioids, benzodiazepines, sleep-disrupting amounts of alcohol, daily cannabis — any one of these can be the answer on its own, and combinations stack.
An honest post-infection history. If the fatigue dates from within six months of a viral illness, the path forward looks different — pacing rather than push-through, with the ME/CFS criteria becoming applicable at the six-month mark NICE 2021.

What to skip. Salivary cortisol panels marketed as "adrenal fatigue" testing — the diagnosis they screen for is not a real medical condition Cadegiani and Kater 2016. Empirical thyroid replacement for a borderline TSH just to see if it helps — the largest trial of that approach showed no symptomatic benefit, and the over-treatment carries real costs Stott 2017. B₁₂ shots without a test confirming you need them. And any IV-drip "energy infusion" sold without a workup behind it.

When tiredness is the wrong thing to test for

The workup in the previous section assumes the fatigue is the main thing. A small set of warning signs put a different question first.

Get a clinical evaluation soon — not a lab order form — if your fatigue comes with any of these:

Unintentional weight loss of more than five percent of your body weight in six months.
Night sweats that soak the sheets.
Fever you cannot explain.
Lumps under the skin, especially in the neck, armpit, or groin, that you can feel.
Blood in your urine or stool, or unexplained bruising.
A new severe headache, vision changes, or any focal neurological symptom — weakness on one side, slurred speech, sudden numbness.
Chest pain on exertion, or new shortness of breath walking up a flight of stairs you used to handle.

These shift the differential toward things that need a clinician's hands now — cancers, occult infections, cardiac disease — rather than an outpatient panel and a wait-and-see.

If you find iron deficiency on the workup, do not start supplementing without your clinician's sign-off in two specific situations. Hemochromatosis — a genetic condition where you absorb iron too well — is more common than people realise, and adding iron to it is harmful. And in anemia tied to chronic disease or chronic inflammation, the iron is locked up rather than absent; oral iron will not fix that picture and may make it worse Weiss and Goodnough 2005. The full iron panel — ferritin alongside iron saturation and transferrin — distinguishes the two.

The myths that keep people tired

The biggest single myth in this space is adrenal fatigue — the idea that chronic stress has worn out your adrenal glands and you need supplements, salivary cortisol panels, and a special diet to "support" them. It is not a real diagnosis. A systematic review of 58 studies found no consistent biomarker basis and no validated link between the salivary cortisol curves the wellness industry sells and the symptom complex they promise to treat Cadegiani and Kater 2016. Genuine adrenal insufficiency — Addison's disease — is rare, dangerous, and diagnosed by a morning blood cortisol and a stimulation test in a hospital lab, never by a mail-in saliva kit Bornstein 2016. The "adrenal fatigue" diagnosis is what gets sold to readers when their real cause is something else on this list and a clinic wants to keep them on the books.

The second myth is "my labs came back normal, so there's nothing wrong." Lab reference ranges are population statistics, not clinical decision thresholds. Ferritin of 22 µg/L is "in range" on most US lab cards and is the band the trial above showed responds to iron Vaucher 2012. A TSH of 5.0 sits at the high tail of the normal distribution; whether to treat is genuinely debated and the largest trial said it does not help symptoms Stott 2017. A serum B₁₂ of 280 pg/mL is "normal" and yet a measurable fraction of people in that band have functional deficiency that methylmalonic acid will expose Stabler 2013. "Normal" means inside a range that includes both healthy people and people whose results are quietly costing them.

The third is the caffeine myth — that more coffee, taken earlier, fixes this. Caffeine blocks the receptor your brain uses to signal tiredness; it does not pay off any of the underlying debt. If you need it to function, the debt is still there and your sleep architecture is being eroded on top of whatever else is wrong.

The fourth is the assumption that "just sleep more" is the universal answer. It works if the problem is duration. It does almost nothing if the problem is fragmented sleep from an airway that keeps collapsing — and the difference between eight hours in bed and eight hours of restorative sleep is most of what apnea steals.

The pattern depends on who you are

Different populations carry different suspects. The lab panel does not change much, but where the yield is highest does.

If you menstruate, iron is the first thing to check, and the most likely thing to find. A fifth to a third of menstruating women run ferritin levels that are quietly costing them — heavy periods are the main driver, and pregnancy and breastfeeding compound it. Ask for ferritin specifically; a "normal CBC" is not the same as a normal iron status, because hemoglobin holds steady until the deficiency is severe. The trial that established treatment for this band — women with normal blood counts but ferritin under 50 — is the one to point your clinician at if they push back Vaucher 2012. Sleep apnea is also under-diagnosed in women, because the textbook OSA patient is a snoring 50-year-old man and the female pattern looks more like fatigue plus insomnia; if the iron is fine and the tiredness persists, push for a sleep study even if no one has told you that you snore Peppard 2013.

After 50, the B₁₂ picture changes. Your stomach makes less of the protein that helps you absorb the vitamin from food, and ten to fifteen percent of older adults have a measurable deficiency. Serum B₁₂ alone is a noisy test in this band — ask for methylmalonic acid as the follow-up if your B₁₂ falls anywhere between 200 and 400 pg/mL Stabler 2013. Sleep apnea prevalence also rises through this band, and treating it has cardiovascular as well as energy payoff.

If you have been on metformin for diabetes for more than two years, run a B₁₂ and a methylmalonic acid annually. The RCT data put the number needed to harm at roughly one in fourteen over four years — meaning if you and thirteen other people are on metformin for that long, one of you is going to be deficient de Jager 2010. Same logic applies to chronic acid-suppressing PPIs like omeprazole or esomeprazole: the odds of B₁₂ deficiency rise by about two-thirds after two years on them, and most chronic PPI prescriptions in primary care are stronger than the indication actually warrants Lam 2013. The fix is either supplementation or de-prescribing the drug; talk to whoever wrote the prescription.

If your fatigue started within six months of a viral illness — mononucleosis, COVID, the bad flu that knocked you flat — the path forward is different. The post-viral fatigue rate runs around 11% at six months across a number of acute infections Hickie 2006, and chronic fatigue is the single most-reported persistent symptom in the largest UK study of post-hospitalised COVID survivors Evans 2021. The current evidence-aligned approach is pacing — staying just under the threshold that triggers a crash — rather than push-through exercise. If the fatigue passes six months, the formal ME/CFS criteria become applicable, and the right next step is a clinician who actually believes you and knows the NICE 2021 guideline NICE 2021.

And if you are vegetarian or vegan, B₁₂ is non-negotiable as a supplement — plant foods do not contain a usable form — and if you menstruate, iron deserves the same close eye as it does for omnivores.

Where this goes wrong

The most common failure is what doctors privately call lab-and-forget. You order the panel, the results come back, your ferritin reads 22, your clinician glances at the hemoglobin and says "your iron is fine," and the actual finding never becomes a prescription. Three years later you are still tired and the file still has the same number in it. The countermove: ask for the actual numbers, read them against the trial thresholds in the protocol above, and push back specifically on the ferritin if it is under 50.

The second is single-finding tunnel vision. The lab catches one thing — say, a TSH of 5.2 — and the assumption becomes "found it, done." You start levothyroxine, nothing much happens, and the actual answer was the sleep apnea you also have. The workup is a differential; finding one thing does not rule out a second. If treatment for the first finding does not move the needle in eight to twelve weeks, finish the panel.

The third is starting iron and quitting because of the GI side effects. Roughly two in five people on standard oral iron get nausea or constipation; many stop before the ferritin even moves. The fix is alternate-day dosing — taking the iron every other day instead of daily — which improves absorption and dramatically improves tolerance. Take it with vitamin C and away from coffee, tea, and dairy.

The fourth is staying on a treatment that is not working. Levothyroxine for a borderline thyroid number that does not improve symptoms after a few months should be reconsidered, not continued indefinitely. The largest randomised trial showed no symptomatic benefit from treating subclinical hypothyroidism, even when the lab number corrected Stott 2017. If your tiredness has not budged after a fair trial, the diagnosis was probably wrong.

The fifth is treating a fake diagnosis instead of the real one. Months of licorice root, adaptogens, "cortisol manager" capsules, and follow-up salivary tests, while the actual driver — the apnea, the iron, the depression — progresses untouched. The wellness industry is not stupid; it has built a parallel system because the real one is slow and brusque. But the supplements do not work for the thing they claim to treat Cadegiani and Kater 2016, and the months spent on them are months the real cause keeps eating.

The sixth is pushing through post-viral fatigue with forced exercise. For decades the standard recommendation was graded exercise therapy — adding small amounts of activity each week until you got back to baseline. The 2021 NICE guideline removed that recommendation entirely because the evidence showed it makes post-exertional malaise worse, not better, in people with ME/CFS NICE 2021. If exertion makes you crash for days afterward, the right strategy is pacing, not adding load.

The seventh is conflating "tired" with "sleepy." Sleepy — you fall asleep in the waiting room, you nod off at red lights — points toward sleep debt, sleep apnea, narcolepsy. Tired-but-can't-sleep, the kind where you lie awake at 2am exhausted, points toward depression, anxiety, hyperthyroidism, or a circadian phase problem. The blood panel does not distinguish them; the way you describe the symptom to your clinician does.

What happens when you find it

The recovery curve depends on which lab lit up. Iron repletion lifts fatigue in roughly four to twelve weeks once the ferritin clears the symptomatic threshold — the trial in non-anemic women saw a near-halving of fatigue scores by week twelve Vaucher 2012. B₁₂ replacement, taken as high-dose oral or as injections if absorption is the problem, lifts cognitive and energy symptoms inside a few weeks, though the neuropathy recovery is slower and not always complete if the deficiency was prolonged Stabler 2013. Levothyroxine for confirmed overt low thyroid is six to twelve weeks before you feel the difference, and the lab number tells you when the dose is right.

CPAP for moderate-to-severe apnea is the one with the dramatic curve. The first few nights are awkward — the mask is strange, you tug at it in your sleep, the air pressure takes adjusting to — and then on the third or fourth night you wake up before your alarm, sharp and rested, and you realize you have not been sleeping in years. Your partner notices the room is quiet for the first time in a decade. The morning headache you had stopped questioning is gone. By six months the daytime energy is at a level you had filed under "before."

Antidepressant or therapy response, when the right name was depression, runs four to eight weeks for the floor under your day to lift. The shift is not euphoria — it is the return of the bandwidth you had been spending on staying functional, redirected to the things you actually care about. The relationship that had been suffering under your irritability stops being a slow leak.

Post-viral fatigue is the hardest case. Most people recover over months to a year, but a minority do not, and the literature on the persistent group does not yet have a definitive treatment beyond pacing and patience NICE 2021. The honest payoff here is smaller: knowing the name of what you have, finding the patient community that has been there longer than you have, and not making yourself worse with the wrong strategy.

The compound is the actual point. A year of restored energy is a year of social investment, a year of training, a year of choices made from capacity rather than depletion. Five years of it compounds further. The reader at 58 who finally found and fixed their apnea at 47 has a different decade in front of them than the one who shrugged it off — and the difference is mostly invisible to either of them until they look back.

Related territory worth knowing

Several adjacent topics carry their own deeper treatments. Sleep apnea — what it does, how it gets treated — has its own entry. So does the family of upper-airway problems that fragment sleep without the textbook snoring pattern. Iron supplementation as a standalone protocol (alternate-day dosing, what form to take, how long for) deserves its own page once it lands in this catalogue. The same applies to B₁₂ supplementation strategy, particularly for the medication-induced cases.

If your fatigue passes the six-month mark and the workup has not surfaced a treatable driver, ME/CFS becomes the formal name the NICE guideline applies; the management is different from the symptomatic workup laid out here. Long COVID has overlapping features and its own emerging treatment landscape. Hashimoto's thyroiditis — the autoimmune form of low thyroid — has wrinkles the basic TSH + free T4 screen does not catch by itself, and the antibody test (anti-TPO) is the follow-up.

Worth knowing for general context: caffeine timing, alcohol's effect on sleep architecture, screen light at night, and morning sunlight all sit upstream of how restoring your baseline sleep architecture feels once the medical drivers are addressed.

Substance and claimed effects

Chronic fatigue, lower-case, is the symptom: persistent tiredness not relieved by rest, lasting weeks to months, often with day-after-day continuity rather than a clear bad-day / good-day rhythm. It is distinct from ME/CFS, the formal post-viral / multi-system illness NICE 2021 defines, although ME/CFS is one possible terminus of the chronic-fatigue workup. This entry treats chronic fatigue as a presenting complaint and lays out the differential diagnosis a reader (or their clinician) walks through to identify a treatable driver before either resigning to "it's just stress" or jumping to ME/CFS as a label.

The claimed effect of running the workup, when a treatable driver is found and addressed: substantial gains in energy (the dominant axis), focus (brain fog is downstream of most drivers), mood (B12 deficiency, hypothyroidism, OSA, and depression each carry mood cost), sleep (sleep-disordered breathing is one of the most under-diagnosed contributors), and short-term health. Longevity effects ride on the specific driver — untreated OSA, severe iron deficiency, untreated thyroid disease, and chronic untreated depression each carry independent mortality signal.

In scope: the recognised, clinically treatable causes of persistent fatigue — insufficient or fragmented sleep (including obstructive sleep apnea and upper-airway resistance), thyroid disease (overt and subclinical), iron status (anemia and low ferritin without anemia), B12 and folate deficiency, depression and anxiety, post-viral and post-COVID fatigue, glycemic dysregulation, vitamin D deficiency, and medication contributors (metformin, PPIs, sedating antihistamines, beta-blockers, benzodiazepines, alcohol, cannabis). Also in scope as a negative: the marketed diagnosis of adrenal fatigue, which the dossier debunks Cadegiani & Kater 2016. Out of scope: ME/CFS management itself (its own entry territory), Addison's disease and primary adrenal insufficiency as a standalone topic Bornstein 2016, and the detailed protocol layer of sleep apnea treatment.

Evidence by addressing question

Mechanism

Fatigue is the brain's signal that supply is below demand. Supply is oxygen-carrying capacity, mitochondrial substrate, thyroid hormone, neurotransmitter tone, restorative sleep architecture, and recovery from recent illness. Demand is cognitive, physical, and emotional load. Because the symptom is downstream of many independent systems, the workup is necessarily a differential rather than a single test.

Five upstream systems generate the bulk of presenting cases. First, sleep — either insufficient duration or fragmentation from sleep-disordered breathing, where airway collapse triggers micro-arousals through the night that the sleeper does not remember; AHI (apnea-hypopnea index) ≥15 marks moderate-to-severe OSA, present in roughly 10% of 30–49-year-old men and 17% in older age bands, with women under-counted because they present with fatigue and insomnia rather than snoring Peppard 2013. Second, the thyroid axis — thyroid hormone sets the metabolic floor for ATP synthesis, and overt hypothyroidism (elevated TSH + low free T4) reliably produces fatigue, cold intolerance, weight gain, and dry skin Garber 2012. Third, hematology — iron is required not only for hemoglobin but for mitochondrial cytochromes and the rate-limiting enzyme tyrosine hydroxylase in neurotransmitter synthesis, which is why ferritin can be too low for cellular use while hemoglobin is still in range Vaucher 2012. B12 maintains myelin and the methionine cycle; deficiency produces megaloblastic anemia and a demyelinating neuropathy that frequently presents as fatigue and cognitive slowing before the blood picture changes Stabler 2013. Fourth, mood — depression compresses energy, motivation, and reward signalling; fatigue is among the most common residual symptoms in partially-treated major depression Targum & Fava 2011. Fifth, the post-infection state — persistent inflammatory cytokine signaling (IL-6, IL-1β) acting on hypothalamic and brainstem fatigue circuits, which the Dubbo Infection Outcomes Study documented at 11% incidence of CFS-criteria fatigue at six months after EBV, Q fever, or Ross River virus infection, with risk tracking acute illness severity rather than premorbid psychiatry Hickie 2006.

Two adjuncts that are real but less common as primary drivers: vitamin D deficiency, with Nowak 2016 showing a self-perceived fatigue benefit in 25(OH)D-deficient adults; and glycemic dysregulation, where post-prandial blood-sugar swings and reactive hypoglycemia produce a felt-fatigue pattern (the afternoon crash) that resolves when the underlying carbohydrate load or insulin resistance is treated.

Evidence

What is settled. Iron deficiency without anemia causes fatigue and responds to oral iron — Vaucher 2012 randomised 198 menstruating women aged 18–53 with fatigue, ferritin <50 µg/L, and hemoglobin >12.0 g/dL to 80 mg elemental iron or placebo for 12 weeks; the fatigue score fell by 47.7% on iron versus 28.8% on placebo (difference −18.9%, p=0.02), with hemoglobin and ferritin rising in the treated arm. The applicability is large because the population (menstruating women with normal CBC but low ferritin) is common and routinely missed by clinicians anchoring on hemoglobin.

B12 deficiency. Stabler 2013 reviews the canonical biology: serum B12 measures total cobalamin, but functional intracellular adequacy is better captured by methylmalonic acid (MMA) and homocysteine, which rise before serum B12 leaves the reference range. Two iatrogenic contributors are documented at population scale. de Jager 2010 randomised 390 insulin-treated type-2 diabetics to metformin 850 mg three times daily or placebo over 4.3 years, with a mean B12 decrease of 19% in the metformin arm and a number-needed-to-harm of 14 for clinical B12 deficiency. Lam 2013 case-controlled 25,956 incident B12-deficiency cases against 184,199 controls in Kaiser Permanente, finding odds ratio 1.65 for ≥2 years of PPI use and 1.25 for ≥2 years of H2-receptor antagonist use, with a dose response.

Overt hypothyroidism. Levothyroxine replacement is one of the most-studied drug interventions in medicine and reliably resolves the fatigue, weight, and metabolic signature when TSH is elevated and free T4 is low, with guideline-grade evidence and clinical-community alignment Garber 2012.

Sleep-disordered breathing. Peppard 2013 recalibrated US population prevalence using the Wisconsin Sleep Cohort: moderate-to-severe OSA in roughly 10% of middle-aged men and 17% of older bands; the older Sleep Heart Health and HypnoLaus data sets converge. CPAP treatment of moderate-to-severe OSA improves Epworth sleepiness scores and self-reported daytime energy across multiple RCTs; the cardiovascular endpoint literature is more contested but the daytime-symptom literature is not.

Depression. Fatigue is one of the nine DSM-IV/5 MDD criteria; it is also among the most common residual symptoms in partially-treated depression, with implications for treatment-switching decisions Targum & Fava 2011. Wessely's primary-care cohort showed psychological distress as the dominant predictor of long-term fatigue persistence when standard medical workup is unremarkable Wessely 1997.

Post-acute infection fatigue. Hickie 2006 established the prospective rate at ≈11% across three different acute infections, robust to pathogen biology. Evans 2021 (PHOSP-COVID, n=1077 hospitalised UK survivors at 6 months) reported chronic fatigue as the single most common persistent symptom, with female sex, middle age, and acute illness severity as predictors of non-recovery; the cohort design captures the symptom rate but does not yet identify a treatment.

Vitamin D. Nowak 2016 randomised 120 healthy adults with 25(OH)D <20 ng/mL to 100,000 IU single-dose vitamin D3 or placebo, with significant fatigue-score improvement at 4 weeks in the treated arm. Benefit appears confined to the deficient subset.

What is contested. The ferritin threshold for symptomatic iron deficiency without anemia — US labs commonly flag <15 or 30 µg/L as deficient, but the trial evidence and emerging hepcidin-anchored consensus support 50 µg/L as the physiologic floor for symptomatic effect Vaucher 2012. Treatment of subclinical hypothyroidism for non-specific symptoms including fatigue — the TRUST trial (Stott 2017, n=737 adults ≥65 with TSH 4.60–19.99 and normal free T4) showed levothyroxine produced no improvement in tiredness or hypothyroid symptom scores at one year despite achieving biochemical correction; the older-adult result does not necessarily generalise to younger patients but it raised the bar for empirical treatment. Adrenal fatigue as a discrete diagnosis — a systematic review of 58 studies found no consistent biomarker basis and no validated link between marketed salivary cortisol curves and the symptom complex sold Cadegiani & Kater 2016; primary adrenal insufficiency (Addison's) is a real but rare condition diagnosed by morning cortisol and ACTH stimulation, not by wellness-clinic panels Bornstein 2016.

What is preliminary. Optimal dose and population for vitamin D fatigue benefit; whether mitochondrial cofactor supplementation (CoQ10, NADH, NAD+ precursors) helps post-viral fatigue beyond placebo (small trials, mixed); the mechanistic substrate of long COVID fatigue (viral persistence vs autoimmunity vs microvascular vs autonomic — all hypotheses, none treatment-actionable yet).

Protocol

First-pass lab panel — the standard workup when "I'm tired all the time" persists across several weeks of better sleep hygiene:

Complete blood count with differential — anemia screen.
Ferritin, with the operational interpretation that <50 µg/L can be symptomatic even when hemoglobin is in range Vaucher 2012.
TSH, reflex to free T4 if abnormal Garber 2012.
Vitamin B₁₂, with MMA as a confirmatory marker if serum B12 falls in the low-normal band (200–400 pg/mL) Stabler 2013.
HbA_1c and fasting glucose — diabetes and pre-diabetes.
25(OH) vitamin D — common, treatable, cheap to check Nowak 2016.
Comprehensive metabolic panel — liver, kidney, electrolytes, calcium.

Second-pass if the first pass is unremarkable and the fatigue persists:

Home sleep apnea test or in-lab polysomnography — strongly indicated even without snoring or obesity, especially in women, where fatigue is the dominant OSA presentation Peppard 2013.
Validated mental-health screening: PHQ-9 for depression, GAD-7 for anxiety.
Medication review — sedating antihistamines, beta-blockers, gabapentinoids, opioids, benzodiazepines, statins, alcohol, cannabis, cumulative anticholinergic burden.
Post-infection history — symptom onset within six months of an acute viral illness shifts the differential toward the post-viral path Hickie 2006 Evans 2021; the NICE 2021 ME/CFS criteria apply at the six-month mark NICE 2021.

What the protocol explicitly avoids: salivary cortisol curves marketed as adrenal fatigue panels Cadegiani & Kater 2016; empirical levothyroxine for subclinical hypothyroidism in the absence of TPO antibodies or severe symptoms Stott 2017; empirical B12 injections without testing; and any IV-drip protocol delivered without a workup.

Contraindications

This entry is a workup framework rather than a specific intervention; contraindications attach to the conditions the workup uncovers, not to the workup itself. Empirical iron supplementation without testing is contraindicated in hemochromatosis (genetic iron overload, more common than usually appreciated) and counterproductive in anemia of chronic disease, where iron is mobilisation-restricted rather than absolutely deficient Weiss & Goodnough 2005. Empirical levothyroxine without confirmed TSH derangement risks subclinical hyperthyroidism, atrial fibrillation, and accelerated bone loss Garber 2012.

Red flags that warrant urgent clinician evaluation rather than a routine outpatient lab panel: unintentional weight loss (>5% body weight over 6 months), drenching night sweats, fever without an obvious source, palpable lymphadenopathy, hematuria or blood in stool, unexplained bruising or bleeding, severe new headache, vision changes, focal neurological deficits, exertional chest pain or new shortness of breath on minor exertion. These shift the differential toward malignancy, occult infection, or cardiac disease, where the workup pathway is different.

Misconceptions

"It's just stress." Stress amplifies fatigue and can be the trigger, but it rarely sustains a rest-resistant chronic version on its own — when it does, the accurate name is depression or burnout, both of which warrant their own treatment paths Targum & Fava 2011.
"My labs are normal, so there is nothing wrong." Reference ranges are population-level descriptive statistics, not clinical decision thresholds. Ferritin in the 15–30 µg/L band is "within range" in many US labs but is symptomatic in a measurable fraction of menstruating women Vaucher 2012. TSH of 4–6 mIU/L lies on the high tail of the population distribution; whether to treat is genuinely contested Stott 2017. Serum B12 in the 250–350 pg/mL band can mask functional deficiency that MMA exposes Stabler 2013.
"Adrenal fatigue — your cortisol curve is off." The diagnosis has no biomarker basis; the systematic review of 58 studies found no validated link between marketed salivary cortisol patterns and the symptom complex sold Cadegiani & Kater 2016. The real condition (primary adrenal insufficiency, Addison's) is rare, dangerous, and diagnosed by morning serum cortisol and ACTH stimulation, never by a wellness-clinic panel Bornstein 2016.
"More caffeine fixes it." Caffeine antagonises adenosine receptors and masks the fatigue signal; if the reader needs caffeine to function, the underlying debt is unpaid, not resolved.
"Sleep more." Correct if the problem is insufficient duration; useless or counterproductive if the problem is sleep-disordered breathing — more hours of fragmented sleep is not restorative. Time in bed is not time asleep, and time asleep is not the same as restorative sleep architecture.

Audience-specific patterns

Menstruating women. Iron deficiency without anemia is the highest-yield finding by population fraction — Vaucher's RCT population was exactly this group, and prevalence estimates of ferritin <30 µg/L in menstruating women cluster in the 20–30% range. Order ferritin even when hemoglobin is normal Vaucher 2012.
Adults ≥50. Parietal cell loss reduces intrinsic factor and B12 absorption; 10–15% prevalence of B12 deficiency in this band, often masked by serum B12 in the low-normal range. Add MMA when serum B12 is between 200 and 400 pg/mL Stabler 2013.
Adults on metformin. Annual B12 check is reasonable after 2+ years of use; deficiency in ~7–30% of long-term users, with NNH of roughly 14 over 4.3 years in the de Jager RCT de Jager 2010.
Adults on chronic PPIs. Same logic — OR 1.65 for ≥2 years of use in the Lam case-control Lam 2013. Either supplement or de-prescribe (most chronic PPI use is stronger than the indication warrants).
Post-viral, especially post-COVID. Fatigue is the most common persistent symptom at 6+ months in hospitalised cohorts Evans 2021; NICE 2021 ME/CFS criteria become applicable at the six-month mark NICE 2021.
Women presenting with fatigue but no loud snoring. OSA under-detection is the consistent finding — female presentation is more often fatigue plus insomnia rather than the classic male snoring + apneic episodes Peppard 2013.
Adolescents and young adults after mononucleosis. ≈11% develop persistent CFS-criteria fatigue post-EBV Hickie 2006; pacing rather than push-through is the current evidence-aligned approach NICE 2021.
Vegetarians and vegans. B12 deficiency without supplementation; iron deficiency in menstruating vegetarians.

Alternatives

Non-workup defaults are not strategies but acts of resignation: rest more, caffeine more, push through. Each is a reasonable short-term response to a bad week, none is a multi-month strategy. The genuinely competing approach is the integrative or functional medicine workup, which adds adrenal-fatigue salivary cortisol panels, hair mineral analysis, food-sensitivity IgG arrays, leaky-gut panels, and proprietary energy supplements. These tests do not validate against outcomes; the diagnostic claims are not supported by the systematic literature Cadegiani & Kater 2016; the protocol monetises diagnostic uncertainty. Worth naming because the reader will encounter aggressive marketing for it, particularly when mainstream workup returns unremarkable.

Failure modes

Lab-and-forget. Reader orders the panel, receives a ferritin of 22 µg/L, hears "labs are normal" from a clinician anchored on hemoglobin, and never acts on the finding. Years pass.
Single-finding tunnel vision. One result is treated; other contributors are missed. TSH of 5.2 gets levothyroxine; the actual problem is also OSA.
Iron repletion abandoned for GI side effects. Oral iron causes nausea or constipation in ~40% of users; many stop before repletion completes. Alternate-day dosing (every other day rather than daily) improves absorption and tolerance.
Levothyroxine for subclinical hypothyroidism with no symptomatic response. Continued indefinitely despite TRUST-grade evidence of no benefit Stott 2017.
Treating "adrenal fatigue" while the real cause progresses. Months of licorice extract, adaptogens, and salivary cortisol re-tests while untreated OSA continues to fragment sleep or iron deficiency continues to compress oxygen delivery.
Push-through exercise in post-viral fatigue. NICE 2021 explicitly removed graded exercise therapy from the ME/CFS guideline because post-exertional malaise worsens rather than improves with forced activity NICE 2021.
Misreading "tired" as "sleepy." Sleepy points toward sleep debt or OSA; tired-but-can't-sleep more often points toward depression, anxiety, hyperthyroidism, or a circadian phase problem. The clinical interview distinguishes them; the lab panel does not.

Stakes

A treatable driver that goes unidentified compresses a reader's life for the duration. Three to five years of "I'm too tired to do that" is the common arc by the time the workup finally happens. The second-order costs accumulate: missed exercise and reactive eating drive weight gain, which worsens OSA in a feedback loop; social withdrawal and irritability erode relationships; under-performance at work gets coded as lack of ambition or commitment; the silent organ-level damage from untreated OSA (cardiovascular and metabolic), untreated hypothyroidism (lipid panel deterioration, cognitive slowing), and chronic untreated depression (suicide risk, cardiovascular mortality) continues. Most readers carrying chronic fatigue have already been carrying it for one to three years before they take it seriously enough to test.

Payoff

Recovery curves vary by driver. Iron repletion produces felt improvement in 4–12 weeks once ferritin rises above the symptomatic threshold Vaucher 2012. B12 replacement (oral high-dose or intramuscular) lifts cognitive and energy symptoms within weeks, though neurological recovery is slower and incomplete if deficiency was prolonged Stabler 2013. Levothyroxine response is 6–12 weeks for energy in confirmed overt hypothyroidism. CPAP for moderate-to-severe OSA can be transformative within nights to weeks for daytime energy and morning headache, with longer-arc cardiovascular benefit. SSRI response in depressed patients is 4–8 weeks. Post-viral fatigue is the hardest case: pacing rather than push-through, with most cases improving over months to a year but a substantial minority persisting longer NICE 2021 Evans 2021.

The credibility range

Optimist case

Most chronic fatigue presenting in primary care has a findable, treatable cause. The relevant prevalences are large: iron deficiency without anemia in 20–30% of menstruating women, with a clean RCT supporting treatment Vaucher 2012; moderate-to-severe OSA in 10–17% of middle-aged adults, with CPAP benefit on daytime symptoms Peppard 2013; B12 deficiency in 10–15% of adults over 50 and in long-term metformin or PPI users Stabler 2013 de Jager 2010 Lam 2013; hypothyroidism in 4–7% of adults with reliable response to replacement Garber 2012; depression as a major contributor with effective pharmacological and behavioural treatment Targum & Fava 2011. The cumulative coverage is most of the symptomatic population; the labs are inexpensive; the treatments work. The clinical-guideline literature is mature, and the differential is teachable.

Skeptic case

A substantial fraction of chronic fatigue does not have a findable lab abnormality. Wessely's primary-care cohort showed the dominant predictor of persistent fatigue after exhaustive workup was psychological distress and somatic symptom severity rather than any specific biomarker Wessely 1997. The workup risks giving the reader false confidence that lab tests will identify a fixable problem, when the answer for many is depression, anxiety, post-viral state, or true ME/CFS without a current treatment. Some flagship interventions have collapsed under their own trials — empirical treatment of subclinical hypothyroidism for fatigue showed no benefit in TRUST Stott 2017. The ferritin threshold for symptomatic deficiency is contested, and a lab whose decision threshold is in flux is a lab whose interpretation is fragile. Post-viral fatigue has no consistent treatment beyond pacing NICE 2021. A "find the cause" framing can become a treadmill of testing that monetises uncertainty without resolving symptoms.

Author's call

Both positions describe true features of the territory. The synthesis is workup-as-triage: run the panel, treat what is both findable and treatable (which is the majority of cases by the population prevalences above), and accept that an unremarkable workup is itself a useful result — it rules out the dangerous and treatable conditions and shifts the remaining differential toward depression, post-viral states, ME/CFS, and lifestyle / sleep-architecture work. The framing is "rule out the things worth ruling out," not "guaranteed to find the cause." Evidence is rated 4: each individual workup component is well-evidenced, and the workup-as-a-whole is clinical-guideline territory, but the heterogeneity of "chronic fatigue" as a presenting complaint prevents 5. Controversy is rated 3: real disagreement on subclinical hypothyroidism treatment, ferritin thresholds, and the adrenal-fatigue ecosystem.

Stakeholder and incentive map

Mainstream endocrinology, sleep medicine, hematology, internal medicine. The lab-driven workup is their standard practice. Incentive aligned with identifying and treating real disease; guideline literature reflects this Garber 2012 Bornstein 2016 NICE 2021.
Integrative and functional medicine clinics. Major commercial driver behind the adrenal-fatigue diagnosis, salivary cortisol panels, proprietary supplement stacks, and IV vitamin drips. Revenue from a population mainstream medicine often fails to satisfy. The diagnostic claims are not validated Cadegiani & Kater 2016.
Supplement industry. Iron and B12 are honest, cheap, effective wins here; proprietary adrenal blends, mitochondrial cocktails, and "energy" stacks are sold on much weaker evidence. Vitamin D sits in between — defensible in deficient adults, marketed beyond the data.
Primary care under time pressure. Brief workup, "labs are normal, sleep more" disposition. Readers often have to push for the second-pass workup (sleep study, MMA, mental health screen).
ME/CFS and long COVID patient advocacy. Strong, organised, and justifiably skeptical of "it's psychological" framings. Played a substantial role in the NICE 2021 reversal on graded exercise therapy NICE 2021.
Wellness influencer ecosystem. Fatigue is high-engagement content; quality ranges from useful (sleep hygiene, circadian alignment) to predatory (DIY desiccated thyroid, "morning cortisol cocktails").

Population variability

Sex. Iron deficiency dominates premenopausal women; hypothyroidism is roughly 5–8× more common in women; depression prevalence is roughly 2× higher in women; OSA is more common in men by absolute count but under-diagnosed in women because the presentation differs from the textbook male pattern Peppard 2013.
Age. B12 deficiency rises sharply after 50; hypothyroidism prevalence rises with age; OSA prevalence rises through middle age; post-mononucleosis fatigue is concentrated in adolescents and young adults Hickie 2006 Stabler 2013.
Race and ethnicity. Iron deficiency higher in populations with menorrhagia (notably from fibroids) and in groups with greater dietary plant-iron reliance; OSA prevalence elevated at given BMI in Asian populations (craniofacial morphology) and in African American populations.
Medication exposure. Metformin and chronic PPI use both touch large and growing populations de Jager 2010 Lam 2013. Beta-blockers, sedating antihistamines, gabapentinoids, opioids, benzodiazepines, and cannabis all contribute in subsets.
Recent infection history. Post-COVID, post-EBV, post-Q fever, post-Lyme cohorts each carry a defined risk window for persistent fatigue Hickie 2006 Evans 2021.
Diet. Vegetarian and vegan diets without B12 supplementation are at high deficiency risk; menstruating vegetarians also at iron deficiency risk.

Knowledge gaps

Optimal ferritin threshold for symptomatic iron deficiency in non-anemic adults. Vaucher used <50 µg/L; many guidelines still anchor on <15–30 µg/L. RCTs spanning the intermediate band are limited; the threshold for symptomatic effect remains a moving target as hepcidin and soluble transferrin receptor data accumulate.
Treatment of post-viral fatigue beyond pacing. Low-dose naltrexone, mitochondrial cofactors (CoQ10, NADH, NAD+ precursors), antivirals — all preliminary, none guideline-grade.
Mechanistic substrate of long COVID fatigue. Viral persistence, autoimmunity, microvascular dysfunction, autonomic dysregulation including POTS — competing hypotheses, none yet treatment-actionable.
Younger-adult subclinical hypothyroidism. TRUST evidence is older-adults-only; whether younger adults with TSH 4–10 and fatigue benefit from levothyroxine is not as well-tested.
Whether vitamin D supplementation helps fatigue in non-deficient adults. Nowak's effect was in 25(OH)D-deficient adults; benefit at sufficient baseline levels has not been established.
Screening cadence in asymptomatic adults. No consensus on whether to screen ferritin, B12, or TSH in well, asymptomatic adults — and at what interval. The workup described here is symptom-triggered, not screening.

Brief coverage

The input brief named eight relationships to cover: sleep quality, thyroid and adrenal function, iron and B₁₂ status, blood sugar, mood, infection recovery, and daily functioning. Each appears in the body. Sleep gets primary treatment under mechanism, evidence, the second-pass protocol (home sleep test), and the audience pattern for women. Thyroid is covered as overt hypothyroidism (treatable, guideline-anchored) with the subclinical band explicitly flagged as contested per TRUST. Iron and B₁₂ are covered in depth as the two highest-yield findings; the medication-induced B₁₂ story (metformin and PPIs) is broken out as its own audience pattern. Blood sugar is covered through HbA_1c + fasting glucose in the first-pass panel and woven into mechanism through reactive hypoglycemia; not given its own addressing section because the trial-grade fatigue evidence specific to non-diabetic blood sugar is weaker than for the other drivers. Mood is the depression strand; infection recovery is the post-viral / post-COVID strand; daily functioning is the through-line of stakes, payoff, and the audience pattern.

The "adrenal function" call

The brief named adrenal function alongside thyroid. Treating these symmetrically would have been editorially misleading: hypothyroidism is a real, common, treatable cause of fatigue; "adrenal fatigue" is a marketed diagnosis with no biomarker basis per a systematic review of 58 studies Cadegiani and Kater 2016, and primary adrenal insufficiency (Addison's) is rare enough that it does not belong in a default workup but is dangerous enough that it warrants a mention in misconceptions and a pointer to Bornstein 2016. The entry covers adrenal function honestly — by debunking the popular-medicine framing and naming the real condition — rather than by giving it the same weight as thyroid. The misconceptions section carries the load.

Rating difficulties

Evidence at 4 rather than 5. Each individual workup component is RCT-grade and guideline-backed, but the workup-as-a-whole is a clinical pathway rather than a single intervention with Cochrane-level evidence of "running the panel improves outcomes," so 5 felt overclaimed.
Longevity at 3. The entry's longevity benefit is mediated through specific downstream diagnoses (OSA, severe iron deficiency, hypothyroidism, depression) rather than direct. 3 reflects "meaningful disease-prevention or mortality reduction" through those mediators; 4 would imply a more direct effect than the workup itself produces.
Beauty cumulative at 1. Genuinely small and indirect — energy enables exercise enables long-arc body composition; sleep apnea treatment improves cardiovascular and metabolic markers that touch aging trajectory. Not zero, but a stretch above 1.
Applicability at 4. Considered 5 because lifetime prevalence of weeks-to-months fatigue approaches universal, but the strict reach criterion ("universal substrate" like sleep, walking, water) did not fit; 4 ("most adults; broadly useful") was the honest call.
Action = test, cadence = as-needed. Could have been action = decide, since the reader's downstream choice is what to act on, but the entry's core ask is "gather your own data via labs," which is the textbook test case. Cadence is as-needed because the workup is symptom-triggered rather than scheduled.

Separate-entry candidates

Sleep apnea: workup, CPAP, oral appliances. Substantial enough to need its own page; this entry just points at the door.
Iron supplementation protocol: alternate-day dosing, ferrous sulfate vs bisglycinate, vitamin C timing, the with-coffee anti-pattern, GI tolerability.
B₁₂ supplementation: oral high-dose vs intramuscular, methylcobalamin vs cyanocobalamin, dosing strategies for metformin and PPI users.
ME/CFS management following the NICE 2021 guideline — pacing, post-exertional malaise, what not to do.
Long COVID — emerging treatment landscape (low-dose naltrexone trials, autonomic-focused care, microclot hypothesis).
Hashimoto's thyroiditis specifically, with the anti-TPO antibody workup and the wrinkles the basic TSH screen misses.
Subclinical hypothyroidism: when to treat, the TRUST evidence, the age-stratified picture.

Future link candidates

sleep-apnea — pair with this entry's audience section on women.
upper-airway-resistance-syndrome — referenced in headline spec examples; lives adjacent to apnea workup.
mouth-tape — adjacent to airway/sleep-quality territory.
creatine — exists; minor cross-link via vegetarian B₁₂ and energy.
morning-sunlight — adjacent circadian work for tiredness without a medical driver.
caffeine — adjacent given the masking dynamic called out in misconceptions.
alcohol — sleep-disrupting amounts as a fatigue contributor.

Other hard calls

NICE 2021 vs older graded exercise therapy guidance. Took the NICE 2021 position on ME/CFS directly — graded exercise therapy is not recommended — both in failure-modes and in the audience post-viral section. This is the current evidence-aligned position; the older PACE-trial-based push-through recommendation is the de-recommended one. Patient advocacy played a role in the reversal and the article does not relitigate it.
Ferritin threshold framing. Anchored on 50 µg/L per Vaucher 2012 and the emerging hepcidin-aligned consensus rather than the legacy 15–30 µg/L lab cutoff. Flagged for the reader explicitly so they can push back if their clinician anchors low.
Audience-scoped sub-blocks. Used inside the audience addressing section for menstruating women (18-39, 40-59) and adults 60+. The metformin / PPI / post-viral patterns are demographic-by-exposure rather than demographic-by-age-or-sex, so they live in plain paragraphs rather than typed audience sub-blocks.
Featured = true. The applicability and the score put this in flagship territory; the workup is the kind of entry the catalogue exists to deliver, and the dream-tier dek and tagline were written for that placement.