In one line
Anaemia in pregnancy is usually iron deficiency, it is usually preventable, and its most important consequence is that it removes the reserve a woman needs to survive obstetric haemorrhage. Diagnose it against a pregnancy-specific haemoglobin threshold, prove the iron deficiency with ferritin, and treat early enough that the haemoglobin has risen before she reaches labour.
Where this sits. The marrow, the red cell and the oxygen-carrying physiology underneath this chapter are taught in the Primary chapter on blood physiology. Where the anaemic woman actually comes to harm is the Finals chapter on postpartum haemorrhage, which is the reason this is a haemorrhage chapter.
Where anaemia actually kills
More than half the pregnant women in your clinic are anaemic. African women carry the highest burden in the world, 57.1%, against 15% in high-income countries. This is not an occasional complication. It is the background state of the clinic.
It is also almost never written down as a cause of death. Anaemia kills at one remove. A woman loses a litre of blood after delivery, a volume your labour ward manages routinely, and she decompensates while the woman in the next bed does not. The haemorrhage is what gets recorded. The reason she had no margin was decided months earlier at a booking visit, and is recorded nowhere.
Three things then go wrong, over and over, in every unit:
- Iron is prescribed correctly and the haemoglobin does not move.
- A low haemoglobin is waved through as physiological dilution when it is not.
- The window in which oral iron could still have worked closes before anyone notices it was open.
None of those is carelessness. Each has a mechanism, and the mechanism tells you exactly what to do instead.
Iron in the body: what comes in, what stays, what leaks out
An adult carries about 3 to 4 g of iron, and almost none of it is sitting idle. Roughly two-thirds is in circulating haemoglobin, doing the job iron exists for. Most of the rest is in storage as ferritin and haemosiderin in the liver, spleen and marrow macrophages, with a small amount in myoglobin and in the iron-containing enzymes. A tiny fraction, only about 3 mg, is in transit on transferrin at any moment. That transit pool is the one the marrow actually draws from, and it turns over completely several times a day.
Dietary iron arrives in two forms that behave quite differently. Haem iron, from meat, fish and poultry, is absorbed intact and efficiently, in the order of 15 to 35%. Non-haem iron, from plants, legumes and fortified staples, must first be reduced from ferric to ferrous before it can cross, and is absorbed at perhaps 2 to 20%. That range is wide because non-haem absorption is heavily modified by what is eaten with it: ascorbate helps, while phytates in maize and legumes, polyphenols in tea and coffee, and calcium all inhibit it. In a maize-staple population eating little red meat, most dietary iron is the inefficient kind, eaten with the inhibitors.
The practical consequence of all that is a single unimpressive number. A person on a mixed diet absorbs somewhere between 1 and 5 mg of iron a day, and that is the ceiling on what the gut can deliver however much is on the plate.
Losses are the other side of the ledger, and this is where iron stops behaving like other nutrients. A non-pregnant adult loses only about 1 to 2 mg a day, and loses it passively: in shed enterocytes and skin cells, in small amounts of blood, and in menstruation. There is no regulated excretion at all.
Because iron cannot be excreted, it must be policed at the door
The absence of an excretory route sounds like a detail. It dictates the entire design.
Sodium, potassium, water and calcium are all regulated at the kidney, which decides how much to throw away. Iron is toxic in excess, generates free radicals, and has no such tap. A substance like that must be controlled somewhere, and if it cannot be controlled at the exit, it has to be controlled at the entrance.
Iron balance is therefore regulated almost entirely at absorption, together with a second control on how much stored iron is allowed back out of the macrophages.
The body builds that control economically. Iron leaves an enterocyte, and leaves a macrophage, through one exporter and only one: ferroportin. A single exit is a cheap thing to police, because one control point governs the whole system.
The control is hepcidin, a peptide made by the liver. Hepcidin binds ferroportin and causes it to be internalised and degraded. Hepcidin high means the door is removed: dietary iron sits trapped inside the enterocyte and is shed when that cell turns over, and stored iron sits locked inside macrophages where the marrow cannot reach it.
That is the mechanism in full. Before reading on, hold it still and consider what looks like the obvious clinical move: if she is short of iron, give more of it, more often. Work out what hepcidin does to that plan.
Two consequences fall out, and both run against instinct.
First: iron shuts its own door. A dose of oral iron raises serum hepcidin, and it stays raised for roughly 24 hours. The next dose, taken that afternoon or the following morning, therefore arrives at a gut that has just been instructed to stop absorbing.
The consequence is worth stating flatly, because it inverts the intuition: you can prescribe more iron and deliver less of it. Absorption is greater from one daily dose than from the same amount split across the day, and greater from alternate-day than from consecutive-day dosing. Nobody would have guessed that from the clinical picture. It comes straight out of the single-exporter design.
Second: inflammation shuts the same door. Interleukin-6 raises hepcidin. In a woman with HIV, tuberculosis, or any chronic inflammatory illness, ferroportin is stripped from her macrophages and the iron she already owns is locked in the wrong compartment.
She is not short of iron. She has enough iron and cannot get to it. Prescribing more oral iron treats a shortage she does not have, using a door that is shut, and the clinic sees a woman who has failed three months of treatment and gets labelled non-adherent.
There is a third consequence, and it is the one that decides how you investigate.
Ferritin is the storage protein, and it is also an acute-phase reactant. Inflammation raises it.
Sit with what that means in one patient. Her inflammation is locking her iron away where the marrow cannot use it, and the same inflammation is pushing up the number you would have used to detect the problem. The test moves in the reassuring direction at precisely the moment it should be alarming you.
The rule that falls out is asymmetric, and the asymmetry is the point:
- A low ferritin proves iron deficiency. Nothing falsely lowers it.
- A normal or high ferritin does not exclude it, if there is inflammation.
Which tells you what to add to the request form: a C-reactive protein, whose job is not to diagnose anything but to tell you whether the ferritin can be believed. Where it cannot, the transferrin saturation carries the diagnosis instead, because saturation reflects iron in transit rather than iron in store, and inflammation does not inflate it.
What pregnancy does to that system
Pregnancy changes three things at once, and they pull in different directions.
The requirement rises steeply. Roughly 300 to 350 mg is transferred to the fetus and placenta. About 500 mg goes into expanding maternal red cell mass. Around 250 mg leaves with the blood lost at delivery. Against a background loss of 1 to 2 mg a day, the pregnancy adds up to about 1000 mg, and the demand is not spread evenly: it is modest in the first trimester and heaviest in the third, when fetal transfer peaks.
Absorption upregulates to help. Hepcidin falls through pregnancy, particularly from the second trimester, which is exactly the adaptation the mechanism above predicts: the door is opened because demand has risen. Fractional absorption can rise several-fold. This is genuine physiological compensation, and it is why many women manage without supplementation.
Plasma volume expands faster than red cell mass. Plasma rises by roughly 40 to 50%; red cell mass by about 20 to 30%. Both go up. They simply go up at different rates.
Those three changes are what make anaemia in pregnancy both common and genuinely difficult to diagnose, and it is worth seeing exactly why before going near a threshold.
Take the demand first. Upregulated absorption raises the ceiling, but it does not remove it. A woman who begins pregnancy with full stores, about 500 mg, and absorbs well, can meet a 1000 mg bill. A woman who begins with empty stores, or eats mostly non-haem iron with tea, or delivered fourteen months ago and never refilled, cannot. She does not become iron deficient because something went wrong in her pregnancy. She becomes iron deficient because the sum does not work, and it was never going to.
Iron deficiency in pregnancy is arithmetic, not misfortune, which is why prevention beats treatment here in a way it does not in most of medicine. You cannot make her absorb 20 mg a day. You can only send her into pregnancy with a fuller store, or add to the income side, and both decisions are made months before the deficiency shows up in a blood result.
Two different women, one identical haemoglobin
The plasma expansion creates a separate problem, and this one is diagnostic rather than nutritional.
Haemoglobin is measured as a concentration, which is a ratio. If the denominator grows faster than the numerator, the ratio falls. Follow that through and it produces something genuinely strange: a woman can be carrying more total haemoglobin than before she conceived, and have a lower haemoglobin concentration, at the same moment. Nothing is wrong with her. The number fell because the plasma rose.
The dilution is deepest around 28 to 32 weeks, which is exactly when we take the second antenatal sample.
Now put the two processes side by side. A woman running an iron deficit has a low haemoglobin. A woman with entirely normal physiology has a low haemoglobin. The measurement is the same. The clinical meaning is opposite.