The room you never enter
Almost everything a patient experiences in IVF happens in rooms you sit in: the consultation, the scans, the injections, the blood tests, the theatre. Then a warmed container is carried out through a small hatch, and the treatment continues somewhere you will probably never see. Patients tell me the not-knowing is its own kind of distress — separate from the waiting, and easier to fix. So this is what is on the other side of that hatch.
An embryology laboratory is a deliberately dull room, usually built directly beside the operating theatre so that eggs travel a few metres rather than down a corridor. Inside are incubators (sealed boxes that hold temperature and gas at fixed values), microscopes with heated stages, laminar flow hoods (workbenches with filtered air flowing over them so dust and microbes do not settle into open dishes), stocks of culture media (the liquids that eggs and embryos live in), and tanks of liquid nitrogen for freezing. There is nothing dramatic to look at.
The people working in it are embryologists, laboratory scientists rather than doctors. We do not prescribe your drugs or choose your protocol. What we do is handle the cells: locate your eggs in the fluid that arrives from theatre, prepare the sperm, bring the two together, keep the resulting embryos alive, assess them, load the one that goes back, and freeze the ones that remain. Everything described below is standard practice in modern IVF laboratories. Details differ between clinics, and where they differ, your own clinic's laboratory is the authority on what was done in your cycle.
The first hours: finding the eggs, preparing the sperm
Egg collection (oocyte retrieval) is done with pain relief, most often sedation, though the exact technique varies between clinics. A needle is passed through the vaginal wall under ultrasound guidance to drain each follicle, the fluid-filled sac in the ovary that may contain an egg. What actually reaches the laboratory is not a tray of eggs. It is test tubes of follicular fluid, usually blood-stained, with nothing visible in them. An egg is the largest cell in the human body and still only about a tenth of a millimetre across.
The fluid is poured into a dish on a heated stage and searched under a low-power microscope. What the embryologist is looking for is the cumulus-oocyte complex: the egg surrounded by a cloud of supporting cells, which in the fluid looks like a small pale, fluffy sunburst. Each one found is washed through clean medium and placed in a labelled dish inside an incubator. Not every follicle seen on the scan yields an egg. Some are empty, and in some the egg does not detach, so the number collected is often lower than the number of follicles counted. That gap is biology rather than anyone's carelessness. The work is done quickly, because outside the incubator an egg cools and the medium's acidity shifts within minutes.
The semen sample is usually produced the same morning. It must first liquefy. Freshly produced semen is gel-like and thins into a more watery liquid, usually within about a quarter of an hour and occasionally taking considerably longer; laboratories generally let this happen in an incubator at body temperature. Semen contains a good deal besides sperm: seminal fluid, immature cells, debris and sometimes white blood cells, none of which belong anywhere near an egg. The laboratory therefore prepares the sample, commonly by density gradient centrifugation (spinning it through layers of fluid of graded density, so that the denser, more mature and more normally formed sperm pass through and can be collected clear of the seminal fluid and debris) or by a swim-up method (clean medium is layered over the sample and the sperm that swim up into it are harvested). What comes out is a small volume of concentrated, motile sperm in culture medium.
Where there is no sperm at all in the ejaculate (azoospermia), sperm may instead be obtained surgically from the epididymis or the testis, and the laboratory then searches the tissue under the microscope for individual moving sperm. That search can take hours and does not always find any. Frozen sperm (banked earlier, or donor sperm where that is used and legally permitted) is thawed and assessed the same morning, because a thawed sample never performs exactly as its pre-freeze report did.
Conventional IVF or ICSI: how the decision is made
There are two ways to bring egg and sperm together. In conventional IVF (technically called insemination), each egg, still wrapped in its cumulus cells, is placed in a droplet of medium together with many thousands of prepared sperm, and the sperm must do the rest themselves: bind to the zona pellucida (the shell around the egg), work through it, and fuse with the egg. The laboratory creates the conditions; it does not perform the fertilisation.
In ICSI (intracytoplasmic sperm injection), the surrounding cells are stripped away with a brief enzyme treatment and a fine pipette so that the egg itself is visible. The embryologist then checks maturity, because only a mature egg can be fertilised: maturity is recognised by a small structure called the first polar body, which the egg extrudes when it has completed the right stage of its own division. Under high magnification, a single sperm is immobilised, drawn tail-first into a glass needle finer than a human hair, and injected through the shell and the egg's membrane into the egg's interior. Each egg is done individually, by hand.
ICSI is generally chosen where sperm count, movement or shape make it unlikely that conventional insemination would work reliably; where sperm were retrieved surgically or thawed from a limited frozen sample; after a previous cycle with failed or very low fertilisation; and where the eggs have to be handled individually anyway, such as when embryo biopsy for genetic testing is planned. Some laboratories use ICSI as their default for most cases; others reserve it. Both are defensible positions. The decision rests with the clinical team, working from the laboratory's assessment of that morning's sample rather than from a patient's preference for the more technical-sounding option.
One thing here is widely misunderstood. ICSI solves exactly one problem: getting a sperm inside an egg. It does not improve egg quality, does not repair damaged sperm DNA, does not guarantee fertilisation, and does not produce a better embryo than the same egg and sperm would have produced on their own. It is a bypass, not an upgrade. The injection itself also carries a small risk that an egg does not survive the procedure, which is one reason it is not treated as automatically preferable.
The fertilisation check, and the days that follow
Roughly sixteen to eighteen hours after insemination or injection, each egg is examined for the signs of normal fertilisation: two pronuclei together with two polar bodies. The pronuclei are two round structures sitting inside the egg, one holding your genetic material and one the sperm's, visible in the window before they merge. An egg with two pronuclei is now a zygote, a one-celled embryo, and two pronuclei is what laboratories treat as the standard picture of normal fertilisation. Other patterns are read more carefully than they once were. Three pronuclei means an extra set of chromosomes, and those eggs are not used. A single pronucleus, or none visible at the moment of the check, is genuinely ambiguous. Pronuclei form and fade on their own schedule, so some of these eggs have in fact fertilised normally and were simply looked at slightly late, and current European guidance supports culturing them on cautiously under a laboratory's own written policy rather than discarding them automatically. Others have truly not fertilised. Which applies is a judgement for the laboratory that saw your eggs, working to a policy it should be able to describe to you.
The window is narrow, which is why the check happens at an awkward hour. For most patients it is also the first drop in the numbers: not every egg collected is mature, and not every mature egg fertilises, even with ICSI, even with a normal semen report. If the call that morning gave you a smaller number than you were expecting, that experience is extremely common, and it is not a sign that something was mishandled.
From there the embryos are cultured and assessed on a schedule. On day two they are typically around four cells (each cell is called a blastomere). On day three, around eight cells; the embryologist records cell number, how even the cells are, and how much fragmentation there is, meaning small pieces of cytoplasm that have broken off. On day four the cells flatten against each other and the boundaries blur, a stage called compaction, producing a morula. On day five or six a good embryo becomes a blastocyst: a fluid-filled cavity (the blastocoel), an inner cell mass (the small cluster of cells that would go on to form the baby), and an outer layer called the trophectoderm (which would form the placenta). To implant, it must then break out of its shell.
Loss along the way is expected rather than exceptional. Many embryos slow down or stop entirely, most often because of chromosomal errors already present in the egg or the sperm before anything was done in the laboratory. An embryo that arrests on day three was not starved or dropped; it reached the point where its own genome had to take over the work and could not. I state this bluntly because patients so often read a smaller number on day five than on day one as evidence that someone made a mistake.
What the laboratory controls, and why the conditions matter
For those few days, the embryo has no body around it. The laboratory is the body. Temperature is held near 37°C not only inside the incubators but on every microscope stage and work surface an egg touches. Acidity is held steady by the carbon dioxide in the incubator's atmosphere acting on the bicarbonate in the medium, which is why an incubator door left open matters. Culture media come in quality-checked batches with defined expiry dates, and are warmed and equilibrated (left in the incubator long enough to reach the right temperature and gas balance) before an egg is ever placed in them. Many laboratories also culture embryos under reduced oxygen, commonly around 5% rather than the roughly 20% of room air, because the fallopian tube and uterus are naturally low-oxygen environments. European good-practice recommendations for IVF laboratories say a low oxygen concentration should be used. The equipment for it is not universal and the exact setting differs between clinics, so this is one of the details worth asking your own laboratory about rather than assuming.
The less obvious controls are just as real. Volatile organic compounds are toxic to embryos: fumes from paint, solvents, cleaning products, new furniture, perfume. Laboratories therefore use air filtration, choose low-emission materials, and staff work without scent. Light exposure is kept short. Dishes are outside the incubator only as long as the task requires. And every dish, tube and freezing device is labelled and identity-checked at each step where two patients' material could conceivably be in the same room, in many laboratories by a second person or an electronic witnessing system that records each check.
Then there is monitoring: incubators alarm when they drift, gas cylinders are kept in reserve, and the laboratory needs power that does not stop. That last point is a genuine design constraint in Nepal, where laboratories plan for supply interruption with backup power, because an incubator that has drifted for hours cannot be undone afterwards. None of this makes an egg better. All of it exists to avoid making things worse, which is most of what quality means in an embryology laboratory.
Grading: what it does and does not predict
Grading is a written description of how an embryo looks down a microscope at a particular moment. On day three, the grade summarises cell number, how symmetrical the cells are, and how much fragmentation is present. On day five or six, the widely used blastocyst systems combine a number for how expanded the cavity is with two letters, commonly A, B or C, for the quality of the inner cell mass and of the outer trophectoderm, giving labels such as 4AA or 3BB. It is assessed by a human eye, or from the images of a time-lapse incubator that photographs embryos without removing them, and different experienced embryologists will sometimes grade the same embryo a step apart.
What grading is for is ranking. Across large numbers of cycles, better-looking blastocysts implant more often than poorer-looking ones, so grading is used to decide which of your embryos to transfer first and which to freeze. That is the whole of its job: putting your own embryos in an order, on that day.
What grading cannot do is read chromosomes. A confident-looking embryo can be chromosomally abnormal, and an unremarkable-looking one can be entirely normal and become a healthy child. Chromosomes are examined only by a genetic test performed on a small biopsy of cells, preimplantation genetic testing for chromosome number, or PGT-A, and even then only in the handful of cells removed rather than in the whole embryo. It is a separate procedure with its own limitations, not part of routine grading. A grade is also not a report card on you or your partner, not a measure of your worth as parents, and not a prediction about a future child's health or intelligence. Patients ask me for the grade as though it were a verdict. It is a laboratory's best judgement about which embryo to try first, and it can be wrong in either direction.
Transfer, and freezing what remains
For transfer, the chosen embryo is drawn into a fine, soft catheter in a tiny volume of medium. The embryologist carries it to the theatre side, the doctor passes the catheter through the cervix under ultrasound guidance and releases it, and the embryologist then examines the catheter and dish under the microscope to confirm the embryo actually left. The procedure itself takes minutes and usually needs no anaesthesia. The embryo is not attached or glued to anything; it must implant on its own over the following days.
Not every cycle ends in a fresh transfer. Some clinics transfer on day three, some on day five, and in many cycles everything is frozen and transferred in a later month instead: for example when the ovaries have responded too strongly and transferring would be unsafe, when genetic test results are pending, or when the lining of the uterus is not suitable that month. That decision is clinical, not a laboratory judgement, and a frozen transfer is a normal plan rather than a setback.
Surplus embryos that are developing well enough are preserved by vitrification. The embryo passes through solutions of cryoprotectants, which replace much of the water inside its cells, and is then cooled extremely fast, at rates of the order of thousands of degrees per minute, into liquid nitrogen at −196°C. At that speed the remaining water becomes glass-like instead of forming ice crystals, which would tear the cell apart from the inside. Each device is labelled with patient identifiers and stored in a recorded position in the tank, and how long embryos may be stored and what may be done with them later are legal and consent matters your clinic handles directly.
The honest limits: not every embryo is suitable for freezing, because a slow or heavily fragmented embryo is unlikely to survive the process; survival on warming is high with vitrification in experienced hands but is not universal; and a frozen embryo is a chance, not a reservation. Being told "we have frozen two" is real and valuable news. It is not the same as being told that two children are waiting, and I would rather say so than let you discover it later.
Why a normal semen analysis can still be followed by fertilisation failure
A semen analysis measures volume, concentration, motility and morphology, sometimes with vitality and a white-cell count. Those are descriptive measures compared against population reference values, and a single sample varies considerably from week to week in the same man. The report tells you how many sperm there are and how they look and move: a headcount and a fitness check.
It does not test what fertilisation actually requires. It does not measure the integrity of the DNA inside the sperm head, whether the sperm can bind to and penetrate the egg's shell, whether its acrosome (the enzyme-carrying cap at the front of the head) releases properly, what it contributes to the machinery of the first cell division, or whether it can switch the egg on, a process called oocyte activation. A sperm can look flawless and swim beautifully and still fail at any one of these. Deficiency in the sperm factor that triggers oocyte activation, with phospholipase C zeta the protein most often implicated in the research literature, is one described cause of repeated total fertilisation failure after ICSI.
It is also a mistake to treat fertilisation as purely a sperm exam. The egg supplies almost everything the embryo starts with: its cytoplasm, its machinery, the spindle that segregates the chromosomes, the capacity to complete its own division and to respond to the sperm at all. Egg-side factors cause fertilisation failure too, among them maturity, cytoplasmic quality and effects that increase with age, and after a single cycle it is frequently impossible to say which side was the limiting one. An egg that looked mature under the microscope can still be immature at the level that matters, and that is not visible.
Total fertilisation failure, where no egg fertilises at all, is uncommon and is devastating when it happens. What it changes is what the laboratory and clinical team do next: moving from conventional IVF to ICSI, reviewing how the sperm was prepared, revisiting the timing of the trigger and egg maturity, and, where fertilisation failed completely or almost completely, considering whether assisted oocyte activation applies. That is deliberately not a routine technique: European good-practice guidance does not recommend it for general use and confines it to a narrow set of situations such as complete activation failure or globozoospermia, long-term safety data on it are still limited, and it is neither offered nor permitted everywhere. It belongs to a specialist review of a cycle that has already failed, not to planning a first one. There is some relief buried in all this, and it is narrow. Fertilisation failure after a "normal" semen report does not mean the report was wrong, that the laboratory was careless, or that anything was concealed from you. It means the test measured what that test measures.