Blood cells

One factory, three products

All three lines come from the same place, and from the same ancestor cell in the bone marrow. That single fact gives you a reading rule that works across the whole panel.

If all three lines have moved together, the likely problem is the factory. If one line has moved on its own, the likely problem is something happening to that cell type after it left. Being used up, being destroyed, being held somewhere, or simply not having the raw material it needs.

Red cell production is not running at a constant rate either. It is controlled, and the signal comes from the kidney, which is why this section and the Kidneys section above are describing two halves of one system.

Why a blood count is slower than it looks

When the kidney senses low oxygen it releases erythropoietin and the marrow responds, but the response takes time. From the point a cell is committed to becoming a red cell to the point it is released into the blood is around a week. The cells already in circulation live about 4 months and are not affected at all.

So a haemoglobin taken today is reporting on a production decision made weeks ago, averaged across cells manufactured over the preceding 4 months. Whatever changed last week is mostly not in the number yet. That cuts both ways, though it does not mean a fortnight is too soon to look. UK guidance on iron deficiency expects haemoglobin to have risen by at least 10 g/L after 2 weeks of daily oral iron, and treats a smaller rise as a sign the treatment is not working. What the lag does mean is that full correction takes months, so a haemoglobin that has moved but not yet normalised at 2 weeks is a response rather than a failure.

There is a fast readout, although it is not on this panel. Reticulocytes are red cells released within the last day or two, still carrying traces of the machinery they used to build haemoglobin, and their number rises within days of the marrow being pushed. They answer the one question the rest of the count cannot, which is whether the marrow is responding. It is a separate test and not one we offer. Your GP can add it onto a blood count, and private labs sell it too, though usually priced as a test in its own right rather than as a few pounds on top. So if you are treating a deficiency and want an answer before the 2 week mark, that is the number to ask for, because it moves within days.

What the profile looks at

Red cells carry oxygen. Haemoglobin is the amount of oxygen-carrying protein, haematocrit is the proportion of your blood volume made of red cells, and the red cell count is how many there are. The rest are the indices: MCV is average cell size, MCH and MCHC describe how much haemoglobin they contain, and RDW is how variable their size is.

White cells are the immune system in transit, and the differential splits them into neutrophils, lymphocytes, monocytes, eosinophils and basophils. Which type has moved is more informative than the total, since the total can look unremarkable while the balance inside it has shifted.

Platelets are not cells. They are fragments shed from very large marrow cells, without a nucleus, and they live about 7 to 10 days. The count and the average size are both reported.

Only some of these are measurements

This is a pattern that runs across the whole panel rather than sitting in one marker. Two things are always genuinely measured: the analyser counts the red cells one at a time, and it measures the haemoglobin by breaking the cells open and reading the colour of what comes out.

After that it is arithmetic. Of haematocrit and average cell size, one is measured and the other is worked out from it, and which way round depends on the machine. The MCH and MCHC on your report are calculated. Haemoglobin divided by the cell count, and haemoglobin divided by haematocrit. Neither is a measurement of anything.

There are analysers that exist which can shine a laser through each intact red cell and read how much haemoglobin is inside it directly, one cell at a time, which gives a genuinely measured version of both numbers. Those machines report the calculated figure and the measured one side by side and compare them. If the two disagree by more than about 2 g/dL, something is wrong with the sample. But most labs only print the calculated one.

Which matters because a number derived from two others carries the error in both, and it is why an impossible-looking MCHC almost always says something about the specimen rather than about the person. The Body Guide page gathers every calculated number on a report together, because the same rule applies to all of them.

Size, and the number most people skip

Average size points in different directions. Anaemia with small cells suggests something different from anaemia with large cells, and the count alone would not distinguish them. Small points broadly towards iron, or towards an inherited difference in how haemoglobin is built. Large points towards B12 or folate, covered in the Nutrients section below.

RDW is the one that gets ignored, and it often moves first. An average conceals a mixture. Someone becoming iron deficient is producing small cells while still carrying a full complement of normal-sized older ones, so for a long stretch the average sits inside the range while the spread widens underneath it. RDW measures that spread, and in developing iron deficiency it commonly rises before average size falls, and well before haemoglobin does. It is the earliest thing on a full blood count to move, and it is the step before the sequence described in the Iron section.

The same logic separates two things that otherwise look alike. Iron deficiency produces a mixture, so small cells with a wide spread. The inherited traits that produce small cells make them uniformly small, so small cells with a narrow spread. Same average, different spread, different answer.

That second one has a test, and a lot of people carrying it never find out. Thalassaemia trait is common in people of Mediterranean, Middle Eastern, South Asian, African and Southeast Asian background. It produces small red cells for life, with no symptoms and no consequences for the person carrying it. The trap is that it looks like iron deficiency on a blood count, so people spend years taking iron they do not need and wondering why nothing changes.

The gene behind it makes the beta chain of haemoglobin. Adult haemoglobin is 4 protein chains, 2 alpha and 2 beta, folded around the iron atoms that do the oxygen carrying, and the beta chains come from a pair of genes, a copy inherited from each parent. Beta thalassaemia trait means a copy is faulty and makes little or no beta chain. The other copy carries on working, so the cell still makes haemoglobin, just more slowly and less of it.

The small cells follow from that. A developing red cell divides while it fills with haemoglobin and stops dividing once the concentration inside it is right, so filling more slowly means it fits in an extra division on the way and leaves the marrow smaller and paler. That is why the trait shows up as small cells rather than as too few of them.

Why it is common where it is comes down to malaria. A single faulty beta gene makes a red cell a worse place for the parasite to grow, enough that where malaria was endemic the faulty gene spread. Inheriting a faulty copy from both parents is beta thalassaemia major, a serious lifelong condition, and that is the price the population pays for the protection. It is the same bargain the Biomarker Guide sets out for the Duffy blood group.

Haemoglobinopathy screening usually starts with a haemoglobin electrophoresis or an HbA2. That settles the beta thalassaemia traits, but there is no marker on it for the alpha ones, so those need DNA analysis to confirm. It is not something we sell. A GP will arrange it for persistently small cells with a normal iron panel, and private labs sell it too. It counts twice over, because the thing is inherited, so a result tells you something about what your children could inherit as well.

Sickle cell trait is not a mild anaemia

Sickle cell trait sits in the same gene and behaves nothing like thalassaemia trait. Rather than switching the beta chain off, the sickle change swaps a single building block inside it. The haemoglobin still assembles and still carries oxygen. What it also does, once it has handed its oxygen over, is stack into long fibres that stiffen the cell and pull it out of shape.

Whether that happens depends on how much of the sickle form is in the cell. A carrier has a working copy of the gene as well, so roughly 30 to 45% of the haemoglobin is the sickle form and the rest is ordinary adult haemoglobin. That is too dilute for the fibres to build at the oxygen levels a body normally runs at. The cells keep their shape, keep their size, and last as long as anybody else's.

So sickle cell trait does not cause anaemia and does not make the cells small, whatever consumer health writing says about it being a mild version of the disease. A person with sickle cell trait has an ordinary blood count, and where that count is not ordinary there is a separate reason for it. The recognised cautions for carriers are about settings where oxygen genuinely runs short, such as a general anaesthetic or deep sea diving. Nothing about the trait itself needs treating.

The alpha side of the same molecule

Haemoglobin needs alpha chains as well, and the alpha side is built differently. Where the beta chain has a pair of genes, the alpha chain has 4, which turns alpha thalassaemia into a sliding scale rather than a carrier state and a disease. Lose the use of a single gene and there is usually nothing to see. With 2 gone the cells come out small, a picture close to beta thalassaemia trait. Losing 3 is a genuine lifelong anaemia, and where all 4 are gone a fetus cannot make working haemoglobin at all.

Alpha thalassaemia cannot be confirmed without looking at the DNA. A faulty beta gene shifts the proportions of the different haemoglobins in the blood, which is what the standard test measures. Missing alpha genes do not, so the usual test reads much the same either way. The milder alpha plus form is the one most often carried by people of African, Caribbean, South Asian and Middle Eastern background. The alpha zero form, most common in people of Mediterranean and East or Southeast Asian background, is the version that matters most in pregnancy.

Three cell lines, three different clocks

Red cells live about 4 months. Platelets last a week to 10 days. Neutrophils turn over in about a day, while some lymphocytes persist for years. So a single full blood count is mixing timescales as well as cell types, and a neutrophil count is a snapshot of now in a way that a haemoglobin is not.

And a large share of your neutrophils are not in the count at all. At any moment a substantial proportion are not flowing freely but rolling slowly along the walls of small vessels. Adrenaline detaches them into the circulation within minutes, so a white cell count is describing a distribution as much as a production rate, and it can move sharply without a single new cell having been made.

The average platelet size carries information for the same structural reason. Newly made platelets tend to be larger, so a low count with a high average size suggests the marrow is replacing them, and a low count with a small average size suggests it is not.

The enzyme that only matters under stress

A red cell has no nucleus and no mitochondria. Both were thrown out to make room for haemoglobin, and the cost of that is a cell which cannot make new proteins and cannot repair much. It has to last its 4 months on the equipment it was issued with.

Carrying oxygen is chemically hazardous work. Oxygen leaks electrons, and the products of that damage haemoglobin and the membrane holding it. The cell's defence runs on glutathione, which is kept in its active form by a supply of NADPH, and the only route a red cell has to NADPH begins with an enzyme called glucose-6-phosphate dehydrogenase, or G6PD. Every other cell in the body has alternatives. The red cell does not.

G6PD deficiency makes that supply line thin rather than absent, which is why it is silent almost all the time. Under ordinary conditions there is enough enzyme to keep up. Add a heavy oxidant load and the glutathione cannot be recycled fast enough, haemoglobin denatures and clumps against the inside of the membrane, and the spleen either bites the damaged portions out of the cells or takes the cells away altogether. That is the haemolysis. It builds over hours to days after the trigger and then stops on its own once the vulnerable cells have gone.

There are 3 kinds of trigger. Broad beans, also called fava beans, which are why the condition has been recognised around the Mediterranean for as long as it has. Infection, which is probably the most common trigger of all and is routinely under-stated. And a small number of medicines. The long lists of drugs to avoid that circulate are mostly not evidence based, and much of what is on them was blamed for episodes the infection being treated had caused. What governs UK prescribing is the BNF list, and it belongs with whoever is doing the prescribing rather than with a website.

The gene sits on the X chromosome, which is why it shows up differently in men and women. A man has a single X, so a faulty copy leaves all of his red cells short of the enzyme. Women have 2, and each cell shuts down whichever of the pair it is not using, at random, so a woman with a single faulty copy ends up with a mixture of deficient and normal red cells. That is why a measurement in a woman can read normal while a real share of her cells stay vulnerable.

Nor is it a single condition. The version common in people of African background leaves more residual enzyme than the version common in Mediterranean, Middle Eastern and South Asian populations, so episodes with it tend to be milder and shorter.

It is common, and almost nobody carrying it knows. It is most common in men of Black, South Asian, Middle Eastern and Mediterranean background, its distribution follows malaria for the same reason thalassaemia's does, and fewer than 1 in 50 of the men who carry it have it recorded anywhere in their health record. How common it is, and why it matters for an HbA1c, is in HbA1c, in the Biomarker Guide.

The testosterone link

Testosterone raises haemoglobin and haematocrit. It is the most common adverse effect of testosterone treatment and the most common reason for pausing it, and the mechanism runs through three of the systems on this page at once.

It stimulates erythropoietin, so the kidney's signal to the marrow goes up. It suppresses hepcidin, which as the Iron section describes opens the exit doors on the body's iron stores and makes iron available for building haemoglobin. And it appears to reset the relationship between the two, so the body stops turning erythropoietin down at a haemoglobin where it normally would.

The pattern over time fits that. Haemoglobin and haematocrit rise by roughly 7% to 10%, erythropoietin rises, and ferritin and hepcidin fall over the first few months. Then erythropoietin and hepcidin drift back towards where they started while the haemoglobin stays up and erythropoietin is still not being suppressed. That last part is what a new set point means.

The practical consequence is that a rising haematocrit and a falling ferritin during testosterone treatment are one finding, not two. The iron has not been lost. It has been spent, on the red cells that are pushing the haematocrit up. Reading the ferritin on its own, without the haematocrit next to it, is how that gets misunderstood.

When a high haematocrit is not the testosterone

If you are taking testosterone it is the obvious explanation, and usually the right one. It is not the only thing that pushes a haematocrit up, and two of the others are better ruled in or out than assumed.

Sleep apnoea is the common one, and it is badly underdiagnosed. It affects something like one in six men between 50 and 70. Breathing stops repeatedly through the night, oxygen dips each time, and the kidney reads that as low oxygen and turns erythropoietin up. So the haematocrit climbs in somebody with no idea anything is happening, because it happens while they are asleep. It also shares almost every risk factor with low testosterone, so the two travel together constantly. A home sleep test is a device you wear for one night in your own bed. A GP can refer you, and it is sold privately for far less than any scan.

The other is polycythaemia vera, which is a bone marrow condition rather than a response to anything. Two blood tests separate it from everything else. An erythropoietin level, which runs low in polycythaemia vera and normal or high in every oxygen-driven cause including sleep apnoea, smoking and altitude. And a JAK2 mutation test, positive in around 98% of cases. Neither is something we sell, both are standard, and a GP can request them.

Worth the effort because the three causes are managed in completely different directions, and because a raised haematocrit is not a cosmetic finding. In population data it tracks with higher cardiovascular and all-cause mortality independently of the usual risk factors.

One drug interaction to know about if you are on testosterone. The SGLT2 inhibitors, now prescribed widely for diabetes, kidney disease and heart failure, raise haematocrit in their own right. On both at once the effects add up.

Why we do not sell this by finger prick

A full blood count needs whole blood that has not clotted even slightly, and capillary blood sitting on the skin before it reaches the tube is prone to exactly that. Our lab puts the failure rate for a full blood count on a finger-prick sample at 20 to 30%, against under 1% for venous. Selling it and having that proportion of orders fail would not be a service.

What changes when these are disturbed

Falling red cells or haemoglobin produce fatigue, breathlessness, pallor, headaches and cold hands and feet. White cell changes accompany infection, inflammation, allergy and a long list of other things. Platelet changes can show up as easy bruising or bleeding at the low end.

And one combination is a single finding rather than three, though it rarely gets read that way. A low haemoglobin, an iron panel that looks adequate, and a reduced eGFR is not three problems. It is one. The kidney makes the signal that drives red cell production, so less kidney means less signal. And because hepcidin is cleared by the kidney, less kidney also means hepcidin accumulates and iron is held in storage. That is why the ferritin can look reassuring while the haemoglobin does not. Two effects, one cause, described across three different sections of this page.

Recessive conditions and related parents

Thalassaemia, sickle cell and G6PD deficiency share an inheritance pattern. A carrier is well. The condition itself turns up when a child inherits a faulty copy of the same gene from both parents, so it depends on 2 particular people happening to carry the same variant.

Being related raises that chance, because relatives have inherited stretches of the same DNA from a shared ancestor. The UK figures are settled. Parents who are not related have a 2 to 3% chance of a baby with a genetic condition or congenital anomaly. First cousins have a 4 to 6% chance. Put the other way round, 94 to 96% of babies born to first cousins have neither.

Part of that effect is not about cousins at all, because marrying within a long-settled community concentrates the same variants whether or not a given couple are related. Rates have also fallen a long way. In the Bradford population where this has been tracked longest, first cousin unions among women of Pakistani heritage fell from 39.3% to 27.0% over about a decade, with the drop sharpest among women born in the UK and women under 25.

None of this is visible on a blood count and no private blood test speaks to it. What speaks to it is NHS antenatal screening and NHS clinical genetics, both free, and both built around this exact question.