Chapter 14 — Minerals: Iron, Calcium, Magnesium, Zinc, Selenium, Iodine — The Essential Elements and Their Real-World Deficiency Patterns

The Hook: "Your bloods came back normal"

Devi Raghunathan had been told her blood tests were normal three times in two years.

Twenty-one, Division I 5000 metre runner, sixty miles a week, vegetarian since she was twelve. Three stress fractures. Nine months without a period. And a fatigue she described in a way that has stayed with me:

"It's not that I'm tired. It's that at about eighteen minutes there's nothing there, and then I have to run the rest of it on nothing."

Her hemoglobin was 12.6 g/dL.

That is a normal hemoglobin. It sits inside the reference range. Every clinician who looked at it was correct to say so, and every one of them stopped there — because in most contexts, "not anemic" is the question a hemoglobin answers.

Her ferritin was 11 ng/mL.

Nobody had measured it until we did.

Ferritin measures iron stores. Hemoglobin measures the consequence of running out.

They are different variables, and they fail in a specific order. When iron intake is inadequate, you don't become anemic on day one. You draw down your stores — ferritin falls — for months or years, functioning on a shrinking reserve. Only when the reserve is exhausted does hemoglobin start to drop.

So there is a long window, sometimes years wide, in which a person is genuinely iron-deficient without being anemic — and in which the standard screening test says normal.

This state has a name (iron deficiency without anemia), a plausible mechanism — iron is required for oxygen transport, for mitochondrial electron transport (Chapter 6 §6.4), and for muscle myoglobin — and a well-documented association with fatigue and reduced endurance performance.

It affects a large number of menstruating women and endurance athletes. Devi was both.

And notice the shape, because it is Chapter 13's threshold arriving in a new form: her intake wasn't the immediately obvious problem, her standard test was normal, and her status was not. The test everyone ran answered a question nobody should have been asking.

🏃 Fast Track: §14.2 (iron — the longest and most consequential section), §14.7 (iodine), and §14.12 (what to do). Twenty-five minutes.

🔬 Deep Dive: §14.9 (sodium — the most genuinely contested number in nutrition, and a masterclass in how two bodies of evidence can disagree) and §14.10 (absorption and competition) are where students and clinicians should spend time.

14.1 Why minerals behave differently from vitamins

Vitamins are organic molecules that can be destroyed — by heat, light, oxygen, time. Minerals are elements. You cannot destroy iron by overcooking spinach. It's an atom; it's still there.

That single difference produces four consequences that structure this whole chapter.

1. Losses happen through leaching, not degradation. Boil vegetables and water-soluble minerals go into the water. Cook them in less water, or use the water, and you keep them. Nothing is destroyed — it just moves.

2. Absorption is where the action is, and it's tightly regulated. For most vitamins, what you eat is roughly what you get. For minerals, absorption varies enormously — with your current status, with the chemical form, and with what else is in the meal. Iron absorption can range from a few percent to over a third depending on circumstances. This is why §14.10 matters more here than the equivalent section did in Chapter 13.

3. Minerals compete with each other. They share transporters. High zinc impairs copper absorption. High calcium impairs iron and zinc absorption. High iron impairs zinc. This is the mechanism behind several supplement harms, and it's why megadosing a single mineral is a genuinely different proposition from megadosing a water-soluble vitamin.

4. The window between deficiency and toxicity is narrower. Selenium is the extreme case. Iron overload is a real clinical entity. "It'll just be excreted" is much less true for minerals.

The essential minerals

This chapter covers the ones where deficiency is genuinely common, where the evidence is genuinely interesting, or where the supplement market is genuinely misleading. That's iron, calcium, magnesium, zinc, selenium, iodine, potassium and sodium.

14.2 Iron

The most clinically important mineral in this chapter, and the one with the widest gap between how often it's discussed and how often it's actually measured properly.

Two kinds, absorbed very differently

Non-heme iron absorption is the variable one, and it is powerfully modified by what's in the meal:

This is why the same 18 mg of dietary iron can deliver wildly different amounts to the bloodstream, and it is the practical core of the whole section.

Regulation: hepcidin

Your body has no mechanism for excreting excess iron. Losses are essentially passive — shed cells, blood loss, menstruation. So absorption is where iron is controlled, via a hormone called hepcidin.

When iron stores are high, hepcidin rises and absorption falls. When stores are low, hepcidin falls and absorption increases. A genuinely iron-deficient person absorbs iron far more efficiently than a replete one — which is elegant, and which also means that supplementing a replete person is both useless and, over time, potentially harmful.

And hepcidin rises with inflammation, which matters clinically: infection or inflammatory disease suppresses iron absorption and sequesters iron, producing the anemia of chronic disease — which looks like iron deficiency and is not treated the same way.

The two-test problem

This is Devi's section and it deserves precision.

Depletion runs in order: stores fall → ferritin falls → (long interval) → hemoglobin falls → anemia.

So a normal hemoglobin excludes anemia and says almost nothing about iron status.

Thresholds, roughly: WHO uses ferritin below about 15 ng/mL to indicate depleted stores in adults; many clinicians use below 30 ng/mL as functionally deficient, and in athletes some use higher cut-offs still, on the basis that performance and fatigue outcomes appear to respond at levels above the anemia threshold.

Devi at 11 was below every one of those, with a normal hemoglobin, for at least two years.

📊 Diagram (described). Picture a staircase descending in three steps, and a single alarm bell mounted at the bottom.

Step one — replete. Iron stores in liver, spleen and bone marrow are full. Ferritin is comfortably in range. Hemoglobin normal. Everything works.

Step two — storage depletion. Intake is below losses, so the body draws on stores. Ferritin falls. Hemoglobin is still normal, because the marrow is being supplied — just from the reserve rather than from the diet. This step can last months or years. The person may feel increasingly tired, may find training harder, may notice their performance flattening. Every hemoglobin test in this period comes back normal.

Step three — iron-deficient erythropoiesis. Stores are essentially gone. Red cell production begins to be limited. Hemoglobin starts drifting down but may still sit inside the reference range, because reference ranges are wide.

Then the bell rings — anemia. Hemoglobin falls below the cut-off. Now the standard test says something.

Two features of this picture matter more than anything else in the chapter.

The alarm is mounted at the bottom of the staircase. It tells you that you have already descended three steps. It is not an early warning system; it's a floor sensor. Ferritin is the only thing that tells you which step you're on.

And the descent is asymmetric with the recovery. Repleting stores takes months of consistent intake, because absorption is capped even when hepcidin is low. You can fall down this staircase in a season and climb back up over half a year — which is why catching it at step two rather than at the bell is worth so much.

Devi had been on step two for at least two years, being told the bell hadn't rung.

⚠️ The critical caveat: ferritin is an acute phase reactant. It rises with inflammation, infection, and liver disease — so a normal or high ferritin does not exclude iron deficiency in someone who is inflamed. Where there's doubt, CRP is measured alongside, and transferrin saturation or soluble transferrin receptor may be used. This is a genuine trap and it catches people.

🔬 Claim → Evidence → Verdict

The claim: "Iron supplementation helps people who are iron-deficient — including those who aren't yet anemic."

Where it comes from: Iron's roles are unambiguous: oxygen transport in hemoglobin, oxygen storage in myoglobin, and multiple roles in the mitochondrial electron transport chain (Chapter 6 §6.4). Deficiency anemia and its correction have been understood for a century.

What the evidence actually shows: For iron deficiency anemia, supplementation is definitively effective — this isn't contested. For iron deficiency without anemia, the evidence is more recent and genuinely supportive: randomized trials in iron-deficient non-anemic women with fatigue have shown improvement in fatigue with iron repletion, and trials in iron-deficient endurance athletes have shown improvements in performance markers. The effects are not universal and the trials vary, but the direction is consistent and mechanistically coherent.

📉 Evidence quality: ✅ Rung 6–8 for iron deficiency anemia. Rung 6 and reasonably consistent for deficiency without anemia.

Verdict: ✅ Well supported — in people who are actually deficient, which requires measuring ferritin, not hemoglobin alone.

⚠️ And the counterpart matters as much as the verdict. Do not take iron supplements without testing. Iron has no excretion route, absorption rises when you're low and falls when you're replete but not to zero, and hereditary hemochromatosis — a relatively common genetic iron overload disorder — is actively worsened by supplementation. Iron supplements also commonly cause constipation and GI distress, and iron overdose is a leading cause of poisoning death in young children. This is the clearest "test, don't guess" nutrient in the book.

Who's actually at risk

That last row is the most important clinical line in this chapter. Iron deficiency is a finding, not a diagnosis, and in some groups it's the presenting sign of something that needs looking for.

🧩 Productive struggle. Five minutes before reading on.

Devi is vegetarian, runs sixty miles a week, and has a ferritin of 11. Her diet contains a reasonable amount of iron on paper — lentils, spinach, fortified cereal, tofu.

Why is she deficient anyway? Generate at least five contributing factors, using §14.2 and what you know about her from earlier chapters.

What I'd list

1. Non-heme only. As a vegetarian, every milligram is non-heme — 2–20% absorbed rather than 15–35%. Her paper intake and her delivered intake are different numbers, and requirements for vegetarians are often cited as roughly 1.8× higher for exactly this reason.

2. Phytate. Her main iron sources — lentils, wholegrains, tofu — are also her main phytate sources. The iron and its principal inhibitor arrive in the same mouthful.

3. Tea or coffee with meals. Polyphenols substantially inhibit non-heme absorption, and this is one of the largest and most fixable effects in the whole section. (She drank tea with dinner.)

4. Menstrual losses — until they stopped. Which brings us to:

5. She's an endurance athlete, and the athlete-specific losses stack: foot-strike hemolysis, GI microbleeding, sweat losses, and — importantly — exercise-induced inflammation raises hepcidin, which suppresses absorption for hours after hard training. Her training is simultaneously increasing her requirement and reducing her absorption.

6. And the one from Chapter 4: she is in low energy availability (EA 31.8 kcal/kg FFM). When total intake is inadequate, every micronutrient is under-supplied — you cannot eat enough iron inside 2,100 kcal while running sixty miles a week. Her iron problem is partly an energy problem wearing a mineral costume.

If you got four, you've understood the section. And notice that the fix has to address several of these at once — which is why "take an iron tablet" would have been an incomplete answer even though she did need one.

14.3 Calcium

Ninety-nine percent of your body's calcium is in bone, where it serves as both structure and a reserve. The remaining one percent runs muscle contraction, nerve signalling, and clotting — and blood calcium is defended tightly, at the expense of bone if necessary.

Requirements: roughly 1,000 mg/day for most adults, rising to around 1,200 mg for older adults and post-menopausal women (check current DRIs for your group).

🔍 Why this works. Calcium is worth understanding as a regulated nutrient rather than a stored one, because it explains why the supplement evidence disappoints.

Blood calcium is defended within an extremely narrow range, because muscle contraction, nerve signalling and clotting all depend on it. If blood calcium starts to fall, parathyroid hormone rises, and it does three things: increases calcium absorption from the gut (via activated vitamin D), reduces urinary loss, and — if needed — releases calcium from bone.

Bone is the buffer. It is a structural tissue that doubles as a reservoir, and when the two roles conflict, the reservoir function wins, because a fatal arrhythmia arrives faster than a fracture.

Two things follow. First, blood calcium is useless as a measure of calcium nutrition — it stays normal until something is badly wrong, which is a familiar shape by now (§14.4's magnesium, Ch 14's ferritin, Ch 13's threshold). Second, and this is the key to §14.3's verdict: adding calcium to someone whose regulation is working just gives the kidneys more to excrete. The system was never short of input; it was managing a balance. That's why bone density moves slightly and fractures mostly don't — and why the interventions that do move fracture risk are the ones that change the demand side: loading the skeleton through resistance training, and supplying protein (Chapter 8 §8.9).

Sources: dairy, fortified plant milks, tinned fish with bones, tofu set with calcium sulfate, leafy greens (with a caveat — spinach's calcium is poorly absorbed because of its oxalate content; kale, bok choy and broccoli are much better absorbed).

🔬 Claim → Evidence → Verdict

The claim: "Calcium supplements prevent osteoporosis and fractures. Everyone over 50 should take them."

Where it comes from: Impeccable logic. Bone is calcium; osteoporosis is loss of bone; therefore more calcium. Plus decades of enthusiastic promotion.

What the evidence actually shows: Considerably weaker than the logic predicts. Large trials and meta-analyses of calcium supplementation in community-dwelling adults have generally found small or no reductions in fracture risk — the Women's Health Initiative calcium and vitamin D arm being the most prominent. Bone mineral density improves slightly; fractures largely don't follow, which is another surrogate-endpoint gap (Chapter 2 §2.7).

Where it does work: calcium plus vitamin D in institutionalized or frail older adults, and in people with genuinely low intake, where the evidence is better. Population matters enormously here.

And there is a safety question, genuinely unresolved: some analyses have suggested calcium supplements (not dietary calcium) may be associated with cardiovascular risk, possibly through transient post-dose spikes in serum calcium. Other analyses find no such signal. The honest state is unresolved, and it's part of why the guidance has shifted toward food first.

📉 Evidence quality: Multiple large RCTs; consistent modest-to-null in community-dwelling adults; better in frail and institutionalized populations.

Verdict: 🟠 Probably false as a general recommendation for healthy adults. 🟢 Reasonable for people with genuinely low intake, and for frail or institutionalized older adults with vitamin D. Food first, and remember Chapter 8: protein and resistance training do more for fracture risk than calcium does.

14.4 Magnesium

The mineral with the widest gap between marketing enthusiasm and evidential support — and, awkwardly, one where the enthusiasm has a real basis.

What's genuinely true: magnesium is a cofactor for hundreds of enzymatic reactions, including essentially everything involving ATP (Chapter 6 §6.1). It's involved in muscle and nerve function, blood pressure regulation, and glucose metabolism. And population intake data does consistently show a substantial fraction of adults below the RDA (~310–420 mg/day depending on age and sex).

Where it gets complicated:

Serum magnesium is a poor marker. Only about 1% of body magnesium is in blood, and levels are maintained tightly by drawing from bone and tissue. So a normal serum magnesium does not exclude inadequacy — which is a genuine problem, because it means we can't easily identify who would benefit.

And "below the RDA" is not "deficient" — Chapter 13 §13.2's whole point. Frank magnesium deficiency with clinical signs is uncommon in healthy people, and is seen mainly with GI losses, alcohol use disorder, certain diuretics, PPIs, and poorly controlled diabetes.

🔬 Claim → Evidence → Verdict

The claim: "Almost everyone is magnesium deficient. Supplementing improves sleep, anxiety, muscle cramps, and blood pressure."

Where it comes from: Real intake shortfalls relative to the RDA, a genuine biochemical role in hundreds of reactions, a poor biomarker that makes the claim hard to disprove, and — this is the important part — magnesium is cheap and well tolerated, so the cost of being wrong is low.

What the evidence actually shows: It splits by outcome. Blood pressure: modest reductions in meta-analyses of supplementation trials — real but small. Migraine prophylaxis and eclampsia treatment: genuine clinical uses with reasonable evidence. Sleep and anxiety: trials are small, heterogeneous, and inconsistent; the popular confidence far exceeds the data. Muscle cramps: generally null in trials, including in older adults and pregnancy, despite being the single most common reason people buy it. And the "everyone is deficient" framing rests on RDA comparison (Ch 13 §13.2) plus an uninterpretable biomarker.

📉 Evidence quality: Modest RCT support for blood pressure; specific clinical uses established; weak and inconsistent for the popular claims.

Verdict: 🟡 Unclear / it depends. Magnesium is a reasonable, cheap, low-risk supplement for someone with genuinely low dietary intake — legumes, nuts, seeds, whole grains and leafy greens are the food answer — and the specific popular claims are much less supported than the marketing suggests. (This is why Walt's magnesium was one of the two supplements kept in Chapter 11: cheap, plausible, low risk, low intake — not because the evidence is strong.)

14.5 Zinc

Roles: immune function, wound healing, protein synthesis, taste and smell, and over a hundred enzymes.

Requirements: roughly 8–11 mg/day for adults. Sources: oysters (extraordinarily high), meat, shellfish, legumes, nuts, seeds, whole grains — with the same phytate absorption problem as iron, meaning plant-based eaters are advised to aim higher, often cited as roughly 1.5× the RDA.

Genuine deficiency causes impaired immunity, poor wound healing, loss of taste, hair loss, and in children, growth impairment. It is common globally and uncommon in well-fed populations — with exceptions in malabsorption, alcohol use disorder, and strictly plant-based diets without attention.

🔬 Claim → Evidence → Verdict

The claim: "Zinc lozenges shorten colds. Take zinc at the first sign of symptoms."

Where it comes from: Zinc has real roles in immune function, and there's a plausible local mechanism — zinc ions may interfere with rhinovirus replication in the nasal passage.

What the evidence actually shows: Meta-analyses and Cochrane reviews have generally found a modest reduction in cold duration with zinc lozenges started early — commonly reported as roughly a day, though estimates vary and the trials are notably heterogeneous in dose, formulation (acetate vs gluconate), and timing. Formulation appears to matter, and many commercial products don't match what was tested. Adverse effects — bad taste, nausea — are common enough to affect adherence and blinding.

What it does not show: prevention. And ⚠️ intranasal zinc has been associated with permanent loss of smell and should be avoided.

📉 Evidence quality: Multiple RCTs, meta-analyzed, heterogeneous, modest effect on duration.

Verdict: 🟡 Unclear / it depends — a real but modest effect on duration when started early with the right formulation; nothing for prevention; ⚠️ avoid intranasal.

⚠️ And the harm that matters more. High-dose zinc causes copper deficiency. Zinc and copper compete for absorption, and sustained high zinc intake — the UL for adults is around 40 mg/day — induces intestinal metallothionein, which traps copper. The result can be anemia, neutropenia, and neurological damage that may be irreversible.

Walt was taking 50 mg a day, in an "immune blend," indefinitely, with nobody monitoring copper. That is above the UL, sustained, and it is one of the two genuinely dangerous items in his cabinet (Chapter 16).

14.6 Selenium

Short section, one big point: selenium has one of the narrowest windows between deficiency and toxicity of any nutrient.

Requirement: around 55 µg/day for adults. UL: around 400 µg/day.

That's a factor of roughly seven — narrow by nutritional standards.

Deficiency is geographically determined, because selenium content in food tracks soil selenium. It was historically endemic in parts of China, where it contributed to Keshan disease, a cardiomyopathy. In most Western diets, intake is adequate.

Toxicity — selenosis — causes hair and nail loss, garlic breath, GI symptoms, and neurological effects.

And the practical trap: Brazil nuts are extraordinarily high in selenium, variably so depending on where they grew. A few Brazil nuts a day can approach or exceed the UL. This is the one food in this book where "a handful of nuts" is genuinely inadvisable as a daily habit.

On supplementation: the SELECT trial (Chapter 13 §13.7) tested selenium and vitamin E for prostate cancer prevention and found no benefit — with a prostate cancer signal in the vitamin E arm. Selenium supplementation in replete populations has not delivered.

14.7 Iodine: the forgotten success, and the returning problem

Iodine deserves more attention than it gets, because it contains both one of public health's greatest victories and a genuine re-emerging problem.

What it does: iodine is required to make thyroid hormone. That's essentially its whole job, and that job is enormous.

Deficiency causes: goitre; hypothyroidism; and — critically — impaired neurodevelopment in the fetus and infant. Severe deficiency in pregnancy causes cretinism: profound, irreversible intellectual disability. Iodine deficiency has been described as the world's leading preventable cause of intellectual disability.

The victory: iodized salt. Introduced progressively from the 1920s onward and adopted in most countries, universal salt iodization has dramatically reduced iodine deficiency disorders worldwide. Like folic acid fortification (Chapter 13, Case Study 2), it works without requiring any behaviour change, reaches the people advice doesn't, and is invisible precisely because it succeeded.

The returning problem, and this is genuinely current:

• Specialty salts — sea salt, Himalayan pink salt, kosher salt — are generally not iodized. As these have displaced iodized table salt, iodine intake has fallen in some populations.
• Most salt in the food supply comes from processed food, and manufacturers typically use non-iodized salt.
• Reduced dairy consumption matters in countries where dairy is a major iodine source (partly from iodine-containing disinfectants used in milking).
• Plant-based milk alternatives are frequently not fortified with iodine, unlike dairy.

Result: mild-to-moderate iodine insufficiency has been documented as re-emerging in several countries, particularly among women of childbearing age.

🔬 Claim → Evidence → Verdict

The claim: "Adequate iodine intake in pregnancy matters for the child's neurodevelopment."

Where it comes from: Iodine is required for thyroid hormone; thyroid hormone is required for fetal brain development. Severe deficiency causes cretinism — this has been understood for a century and is not in dispute.

What the evidence actually shows: The severe end is unambiguous. The more recent and more relevant question is mild-to-moderate deficiency, and here observational evidence has associated lower maternal iodine status with modestly lower cognitive scores in offspring. Trial evidence in mild deficiency is more limited and less consistent. But the requirement in pregnancy is genuinely elevated — commonly cited as around 220 µg/day versus 150 µg for other adults — and the downside risk of inadequacy is severe and irreversible.

📉 Evidence quality: ✅ Unambiguous for severe deficiency. Observational and mechanistically strong for mild deficiency; trial evidence thinner.

Verdict: ✅ Well supported that iodine adequacy in pregnancy matters. Many countries recommend an iodine-containing prenatal supplement, and most standard prenatal formulations include it — worth checking, because not all do. ⚠️ This is particularly important for women who avoid dairy, use non-iodized specialty salt, or eat plant-based — which is an increasingly common combination and is exactly the group least likely to be aware of it.

⚠️ And note the other direction: excess iodine can also cause thyroid dysfunction. Kelp supplements can contain very high and variable amounts. More is not better here either — Chapter 5's Case Study 2 flagged the "thyroid support" shelf for exactly this reason.

14.8 Potassium: the one nobody mentions

Requirement: the Adequate Intake sits around 2,600–3,400 mg/day depending on age and sex.

Typical intake: substantially below that in most Western populations. Most people get roughly half to two-thirds of the AI.

Why it matters: potassium is central to blood pressure regulation, and higher potassium intake is consistently associated with lower blood pressure and lower stroke risk. It works in part by promoting sodium excretion — which is why §14.9's argument is really about a ratio, not a single number.

Sources: potatoes (with skin), beans and lentils, bananas (less than their reputation suggests — a banana is around 400 mg, and a baked potato has roughly twice that), leafy greens, tomatoes, yogurt, avocado, dried fruit.

⚠️ When to see a professional. Potassium supplements are a genuine hazard and are restricted for good reason — hyperkalemia can cause fatal arrhythmias. This matters especially for anyone with kidney disease or taking ACE inhibitors, ARBs, or potassium-sparing diuretics, where high potassium intake — including from potassium-based "low sodium" salt substitutes — can be dangerous. Get potassium from food, and if you have kidney disease or take those medications, ask before changing anything.

14.9 Sodium: the most contested number in nutrition

I've saved this for last because it is the one place in Part III where I genuinely cannot give you a clean answer, and I'd rather show you the argument than pick a side and hide the other one.

What isn't disputed

Sodium raises blood pressure, and reducing it lowers blood pressure. This is about as well-established as anything in this book. DASH-Sodium — a controlled feeding trial — demonstrated clear, dose-dependent blood pressure reductions with sodium reduction, with larger effects in people with higher baseline blood pressure and in the DASH dietary pattern.

And blood pressure causes cardiovascular disease. Also not disputed.

Most people eat far more sodium than they need. Physiological requirement is low — a few hundred milligrams a day. Typical Western intake is around 3,000–4,000 mg/day (Theo's is 4,100). The vast majority comes from processed and restaurant food, not the salt shaker — commonly cited at around 70–75%.

What is disputed

Whether reducing sodium at a population level reduces cardiovascular events, and whether there's a lower bound below which it stops helping or starts harming.

The case for aggressive reduction: blood pressure is a strong causal intermediate; the relationship is dose-dependent; modelling suggests population-level reduction would prevent substantial numbers of events; WHO recommends under 2,000 mg/day.

The case for caution: some large observational studies — the PURE study most prominently — have reported J-shaped associations, with higher event rates at both high and low estimated sodium intake. If real, that would imply a floor below which reduction is counterproductive.

And here is where the methodological fight lives, because it's genuinely decisive:

🔬 Claim → Evidence → Verdict

The claim: "Everyone should reduce sodium to under 2,300 mg — or under 1,500 mg — per day."

Where it comes from: The unambiguous sodium–blood pressure relationship, the unambiguous blood-pressure–disease relationship, and population modelling connecting the two.

What the evidence actually shows: The intermediate steps are solid. The end-to-end evidence — that population sodium reduction reduces events — rests substantially on modelling rather than on trials with hard endpoints, because the definitive trial faces Chapter 1 §1.2's walls. Meanwhile the observational J-curve findings are real findings with serious methodological objections that have not been fully resolved either way.

What I think is defensible: reducing sodium is clearly beneficial for people with hypertension or elevated blood pressure, where the intermediate is the outcome you care about. For normotensive people at typical intakes, the benefit is more speculative and the population targets are more confident than the trial evidence supports.

📉 Evidence quality: ✅ for the blood pressure effect. Genuinely contested for hard outcomes at the population level.

Verdict: 🟡 Unclear / it depends — and unlike some 🟡s in this book, this one reflects a live disagreement among serious researchers rather than a shortage of data. If you have high blood pressure, reduce sodium — that one is well supported. If you don't, the honest answer is that the population targets are not as securely evidenced as their confidence implies, and increasing potassium (§14.8) may matter as much as reducing sodium.

🍽️ On your plate. Whatever you conclude about the target, one thing is uncontroversial: the salt shaker is not where your sodium comes from. Around 70–75% arrives in processed and restaurant food — bread, processed meat, sauces, soups, cheese, ready meals. Cooking more of your own food reduces sodium substantially without anyone tasting less salt, which is why Chapter 31's argument does more for sodium than any amount of shaker discipline.

14.10 Absorption, competition, and why megadosing minerals is different

Pulling §14.1's threads together, because this is where minerals genuinely differ from vitamins.

The competition map:

Three practical consequences.

1. Timing matters for supplements in a way it rarely does for vitamins. Iron is better absorbed away from calcium, tea and coffee; taking it with vitamin C helps. Calcium and iron supplements shouldn't be taken together.

2. Single-mineral megadosing risks inducing a different deficiency. This is not theoretical — it is Walt's 50 mg of zinc, and it is why "more of a good mineral" is a worse idea than "more of a good vitamin."

3. Food handles this automatically. Whole foods deliver minerals in proportions and quantities that don't trigger competition. The competition problem is almost entirely a supplement problem.

🔬 Claim → Evidence → Verdict

The claim: "Modern soil is depleted, so our food no longer contains the minerals it used to. You need a colloidal/ionic mineral supplement to make up the difference."

Where it comes from: A kernel of truth. Soil mineral content genuinely varies geographically — §14.6's selenium and §14.7's iodine are both soil-dependent. And some analyses comparing historical and modern food composition tables have reported declines in certain mineral concentrations, possibly related to cultivar selection for yield and to a dilution effect in faster-growing crops.

What the evidence actually shows: The historical comparisons are methodologically weak — analytical methods, cultivars, sampling, and food composition table construction all changed over the period, which makes like-for-like comparison very difficult. More importantly, plants cannot grow without the minerals they require; a plant grown in genuinely iron-deficient soil doesn't produce low-iron wheat, it produces less wheat. The nutrients most affected by soil are precisely the ones not required by the plant itself — selenium and iodine — which is why those two are genuinely geographically variable and the others largely aren't.

And "colloidal" and "ionic" mineral products are marketing terms with no established advantage; some such products have been found to contain heavy metal contaminants.

📉 Evidence quality: Weak historical comparisons; sound plant physiology pointing the other way.

Verdict: 🟠 Probably false as a general claim. Real for selenium and iodine specifically, which is why iodized salt exists — and which is a targeted fortification argument, not a colloidal mineral argument.

🔄 Check your understanding. A 34-year-old woman with heavy periods and persistent fatigue is told her "blood count is fine." She buys a high-street iron supplement and takes it with her morning tea and a calcium-fortified breakfast cereal. Identify everything wrong here.

Answer

The testing: a normal full blood count excludes anemia, not iron deficiency. She may well be at step two of the staircase. Ferritin was the test she needed — and with heavy periods plus fatigue, there was a clear indication for it.

The self-supplementation: ⚠️ taking iron without testing is the specific thing §14.2 warns against. In her case it may well be right, but she doesn't know, nobody has a baseline to measure improvement against, and if she doesn't improve there's now no way to tell whether that's because she was never deficient or because the dose is inadequate.

The timing — and this is the part almost everyone gets wrong: she is taking it with tea (polyphenols, a large inhibitory effect) and with a calcium-fortified cereal (calcium competes for absorption). She has selected almost the worst possible moment in her day. Iron with vitamin C, away from tea, coffee and calcium, would deliver substantially more from the same tablet.

And the missing question: heavy menstrual bleeding is itself worth addressing clinically. Iron deficiency is a finding, not a diagnosis — treating the consequence while ignoring the cause is the pattern §14.2 flags for men and post-menopausal women, and it applies here too.

What she should do: ask for ferritin with CRP, move the tablet away from tea and calcium, take it with vitamin C, and raise the heavy periods with her GP.

14.11 What's worth testing

The rule remains Chapter 13's: test when there's a reason to suspect, not as a screening ritual — and ferritin is the one that gets missed.

14.12 What to actually do

1. If you're a menstruating woman, an endurance athlete, or vegetarian — and especially if two or three apply — get ferritin checked. Not hemoglobin. Ferritin. This is the single highest-yield action in this chapter.

2. Improve non-heme iron absorption for free. Vitamin C with iron-containing meals; tea and coffee between meals rather than with them; soak or sprout legumes. These cost nothing and the effect is substantial.

3. Don't take iron without testing. ⚠️ No excretion route, hemochromatosis exists, and unexplained deficiency in men or post-menopausal women needs investigating rather than supplementing.

4. Get calcium from food, and remember that protein and resistance training do more for fracture risk than calcium supplements.

5. Check your salt is iodized — or that your prenatal supplement contains iodine. ⚠️ Especially if pregnant or planning pregnancy, and especially if you use specialty salt, avoid dairy, or eat plant-based.

6. Eat more potassium. Potatoes, beans, greens, yogurt. Most people are well below the AI, and it may matter as much as sodium reduction.

7. Reduce sodium if you have high blood pressure — that one's clear. And note that cooking your own food does more than the shaker.

8. Don't megadose single minerals. ⚠️ Zinc above the UL causes copper deficiency. This is the clearest supplement harm in Part III.

🧾 Cost check. A ferritin test typically costs very little and is often free within a health system when there's an indication. Devi's cost less than one month of the protein powder she'd been buying — and it was the finding that changed her career.

What we don't know

The sodium question is genuinely open and I've resisted resolving it. Whether population-level sodium reduction reduces events, and whether there is a lower bound, depends on a methodological dispute about urinary measurement that has not been settled. Anyone who tells you this is settled — in either direction — is telling you about their confidence.

We don't have a good biomarker for magnesium status, which means we cannot identify who would benefit from supplementation, which means the trials are diluted with people who were never going to respond.

And we don't know the optimal ferritin threshold for treating deficiency without anemia, particularly in athletes. Cut-offs of 15, 20, 30 and higher are all used, and the choice determines who gets treated.

Spaced Review

Answer before reading on.

1. (Chapter 13) The threshold concept was "status is not intake." Give the mineral version, using Devi.

Her hemoglobin was normal and her ferritin was 11. Hemoglobin measures the consequence of running out; ferritin measures the reserve. Depletion runs stores → ferritin → (long gap) → hemoglobin, so a normal hemoglobin excludes anemia and says almost nothing about iron status. The test everyone ran answered a question nobody should have been asking.

2. (Chapter 6) Why would low iron specifically impair endurance performance? Name two mechanisms.

Oxygen transport (hemoglobin) and oxygen storage in muscle (myoglobin) — plus iron's role in the mitochondrial electron transport chain (Ch 6 §6.4), which is where aerobic ATP production happens. All three limit exactly the system a 5000 m runner depends on.

3. (Chapter 12) The sodium J-curve and the alcohol J-curve share a structural problem. What is it?

Reverse causation in the low-exposure group. In alcohol it was sick-quitter bias — abstainers enriched with people who stopped because they were ill. In sodium it's that sick people eat less, so low measured intake may mark illness, appetite loss, or frailty rather than cause harm. Both J-curves may be artefacts of who ends up at the low end.

Project Checkpoint: Your Mineral Gap Analysis

Component fourteen. Twenty-five minutes, and one part of it may be the most consequential thing you do with this book.

Step 1 — the iron question, first and separately.

Answer honestly:

• [ ] I menstruate
• [ ] I have heavy periods
• [ ] I do regular endurance training
• [ ] I'm vegetarian or vegan
• [ ] I'm pregnant or planning pregnancy
• [ ] I donate blood regularly
• [ ] I have unexplained fatigue

Two or more boxes → ask for a ferritin test. Not "a blood test" — ferritin specifically, with CRP, because a hemoglobin will very likely come back normal and close the question. The sentence is: "Could I have my ferritin checked as well as my full blood count?"

Step 2 — the absorption audit. Do you drink tea or coffee with meals? Take calcium supplements with iron-rich food? Eat legumes without soaking? These are free fixes with real effects.

Changes I could make at zero cost: ____

Step 3 — iodine. Check your salt. Is it iodized? If you're pregnant or planning, does your prenatal contain iodine? ⚠️ If you use sea salt or pink salt, avoid dairy, or eat plant-based, this is a genuine gap and most people in that position don't know it.

Step 4 — rough intakes. Using Appendix A and your diary, estimate calcium, potassium, and sodium. Order of magnitude only.

Step 5 — check your supplements against the UL. ⚠️ Particularly zinc (UL ~40 mg) and selenium (UL ~400 µg). Add up everything, including multivitamins and "immune" formulations.

Anything above a UL? ____

Non-tracking alternative. Do Steps 1, 2, 3 and 5 only. The iron question, the free absorption fixes, the iodine check and the UL check contain nearly all of the value, and none requires estimating intake.

Next checkpoint (Chapter 15): your hydration self-check — urine colour, a sweat rate estimate, and what you actually need.

Chapter Summary

Minerals are elements, not molecules — they can't be destroyed, only leached. Which means: absorption is where the action is and it's tightly regulated · minerals compete with each other · the deficiency-to-toxicity window is narrower than for vitamins.

Iron — the big one.

• Heme (~15–35% absorbed) vs non-heme (~2–20%), with absorption powerfully modified: vitamin C and meat enhance; phytate, tea/coffee polyphenols and calcium inhibit.
• Hepcidin regulates absorption; there's no excretion route.
• Ferritin = stores (falls early). Hemoglobin = consequence (falls late). A normal hemoglobin excludes anemia and says almost nothing about iron status.
• ⚠️ Ferritin is an acute phase reactant — normal or high doesn't exclude deficiency in the inflamed.
• ⚠️ Unexplained iron deficiency in men or post-menopausal women needs investigation for bleeding, not a supplement.

This chapter's verdicts:

Iodine: the victory (iodized salt) and the returning problem (specialty salts, processed-food salt, less dairy, unfortified plant milks) — concentrated in women of childbearing age.

Potassium: most people get half to two-thirds of the AI, it lowers blood pressure, and ⚠️ supplements and potassium-based salt substitutes are dangerous in kidney disease or on ACE inhibitors/ARBs.

Sodium: blood pressure effect ✅; population outcome evidence genuinely contested, turning on a methodological dispute about spot urine estimation and on reverse causation — the same structural problem as Chapter 12's J-curve. 70–75% comes from processed and restaurant food, not the shaker.

⚠️ The clearest supplement harm in Part III: zinc above the UL causes copper deficiency — anemia, neutropenia, and potentially irreversible neurological damage. Walt was taking 50 mg indefinitely.

The one thing to remember: Devi was told three times that her bloods were normal. They were. The normal test was the wrong test, and a ferritin measurement costing almost nothing was the finding that changed her career.

What's Next

Chapter 15 covers the nutrient people are most confident about and most wrong about.

Where did "eight glasses a day" come from? (The origin is itself a small masterclass in how a recommendation mutates.) Is thirst good enough? How do you estimate your own sweat rate? When is an electrolyte drink genuinely warranted, and when is it flavoured sugar with a marketing budget?

And hyponatremia — the risk from drinking too much, which has killed athletes, and which nobody sells a product to prevent.