Due Date Accuracy by Method: LMP vs. Ultrasound vs. IVF

Every pregnancy gets a due date, and they are all presented with the same confidence: a single day, written down, repeated at every appointment. What is rarely made clear is that the number behind it might be accurate to within five days or to within two weeks, depending entirely on how it was worked out.

The American College of Obstetricians and Gynecologists sets out the hierarchy explicitly, and it is worth knowing where your own date sits in it — particularly if you have been given two different dates and nobody explained why one won.

An ultrasound in the first trimester dates a pregnancy to within about five to seven days, and it is the method every guideline prefers.

The reason it works so well is a genuinely elegant piece of biology. In the first weeks after conception, embryos grow at almost exactly the same rate regardless of genetics, maternal size or anything else. A crown-rump length measurement at eight weeks therefore predicts gestational age very tightly, because there is very little natural variation to get in the way.

That advantage disappears as pregnancy progresses. By the second and third trimesters, genetic and environmental differences have taken hold and babies of the same age differ substantially in size. A scan at 30 weeks measures how big the baby is with reasonable accuracy and how old it is with an error of two to three weeks. This is why late scans are used to assess growth, not to redate.

In vitro fertilisation is the single case where conception is not estimated at all. The date of fertilisation is a matter of laboratory record, and the due date follows to within about a day.

The arithmetic differs slightly by transfer day, and the reason confuses people. A day-5 blastocyst transfer is dated at 261 days from transfer; a day-3 embryo transfer at 263. It looks backwards until you notice that the count is always to the same endpoint of development. The day-5 embryo is two days further along at the moment of transfer, so it has two fewer days left to travel.

If you conceived through IVF, your due date is the most precise one on this page and you can largely stop worrying about which method is best.

Naegele's rule — due date equals the first day of the last period plus 280 days — dates from the early nineteenth century and remains the default when no early scan is available. It is accurate to within one to two weeks for people with regular cycles, and considerably less for everyone else.

Its weakness is a pair of assumptions baked in and rarely stated. It assumes a 28-day cycle, and it assumes ovulation on day 14. Neither is true for a great many people. An analysis of more than 600,000 cycles from a tracking app found that only around 13 per cent were exactly 28 days long, and ovulation day varied widely even within that group.

This is why a good calculator adjusts for cycle length. If you consistently run 32-day cycles, you probably ovulate around four days later than the formula assumes, and your due date should move four days later with it. The Due Date Calculator applies that adjustment.

What it cannot fix is irregularity. If your cycles vary from 25 to 40 days, there is no single number to adjust by, and an early scan is the only way to get a dependable date.

This is the situation that generates the most anxiety, and the rule is simpler than it feels.

ACOG's guidance is that when a first-trimester ultrasound differs from LMP-based dating by more than a defined margin — about five to seven days before nine weeks, and a week or so beyond that — the pregnancy is redated to the scan. The scan wins, and it is not a close call.

The usual explanation is late ovulation rather than anything being wrong. If you ovulated on day 20 rather than day 14, your period dates will read as six days further along than you actually are, and a scan will correctly place you six days earlier. Nothing has gone wrong; the formula's assumption simply did not hold for you.

Occasionally the difference points at something worth investigating, which is another reason the scan takes precedence — it is measuring the pregnancy rather than inferring it.

Once a due date is established, usually at the first scan, it generally stays put for the rest of the pregnancy. People sometimes find this frustrating when a later scan suggests a different age.

The logic is sound. The margin of error on ultrasound dating grows as the baby grows, so a 32-week scan is a much worse dating tool than an 11-week one. Redating on the basis of a later, less accurate measurement would replace a good estimate with a worse one.

There is also a clinical reason. Decisions about induction, monitoring and preterm management all hinge on gestational age, and a date that moves around undermines all of them. A fixed, early, well-founded date is more useful than a continuously updated one.

It is worth ending on the limit of the whole exercise. Even with an exact conception date, pregnancy length varies enormously.

One study measured pregnancies from the day of implantation, removing ovulation-timing noise entirely, and still found a spread of 37 days between the shortest and longest normal pregnancies. Some of that variation was systematic — older mothers carried slightly longer, as did women who had themselves been heavier at birth — but most of it was simply variation.

Only about four per cent of babies are born on their due date. The useful output is not the day but the window: anywhere from 37 to 42 weeks is a normal length of pregnancy, and the dated week-by-week table on the calculator shows exactly when those fall for you.