Gestational Age Explained: Why Pregnancy Is Dated From Before Conception

At the moment you are told you are four weeks pregnant, conception happened roughly two weeks earlier. The first fortnight of a standard forty-week pregnancy contains no pregnancy at all. That is not an error in the counting; it is the counting working as designed.

Gestational age is measured from the first day of the last menstrual period, not from fertilisation, and the reason is entirely practical. Almost nobody knows the day they conceived. Almost everybody can name the day their last period started. Obstetric convention anchors itself to the observable event and accepts a systematic two-week offset as the price. What follows is where that convention came from, what the arithmetic inside it assumes, how much the assumption can be wrong, and why the resulting date is better understood as the centre of a distribution than as a deadline.

The clock starts before conception, and that is deliberate

Two different quantities are in circulation, and confusing them accounts for most of the puzzlement around pregnancy dating.

Gestational age, sometimes called menstrual age, is counted from the first day of the last menstrual period. It is the figure used in every antenatal appointment, on every scan report, in every clinical guideline and in the definitions of preterm, term and post-term birth. Fetal age, or conceptional age, is counted from fertilisation and runs roughly two weeks behind. Embryology textbooks use the second; clinical medicine uses the first, near-universally.

The offset of two weeks is not a measurement. It is an assumption, imported from a model cycle in which menstruation begins on day one, ovulation occurs on day fourteen, and fertilisation follows within a day. Every conventional due date carries that assumption inside it, and much of what looks like imprecision in pregnancy dating is really the gap between that model cycle and a real one.

The practical consequence is that a full-term pregnancy of forty weeks' gestational age represents about thirty-eight weeks of actual development. When a scan report states a gestational age, it is reporting where the measurements place the pregnancy on the menstrual clock, not how long the fetus has existed.

Naegele's rule, and the assumptions inside it

The formula in near-universal use is attributed to Franz Karl Naegele, a German obstetrician of the early nineteenth century, though the underlying rule of thumb predates him. In its arithmetic form it adds 280 days to the first day of the last menstrual period. The older mnemonic version takes the first day of the last period, subtracts three months, then adds seven days and a year. It produces almost the same answer, drifting by a day or two depending on which months are involved.

Two hundred and eighty days is forty weeks. It is built from two components: fourteen days from the start of menstruation to ovulation, and 266 days from ovulation to birth. Neither number is a biological constant. The second is a population average with real spread around it. The first is the weaker of the two, and it is the one that fails most often.

Our pregnancy due date calculator implements exactly this. Enter a last menstrual period of 5 January 2026 with a standard cycle and it returns an estimated date of 12 October 2026, which is 280 days later, along with the gestational age today and the milestone dates that follow from it.

The rule's real weakness is that ovulation on day fourteen describes a minority of cycles. Ovulation timing varies substantially between people and between cycles in the same person. Most of that variability sits in the follicular phase, the stretch from menstruation to ovulation. The luteal phase after ovulation is comparatively stable. So a person with a consistently long cycle is not usually spending longer between ovulation and their next period; they are usually taking longer to ovulate. Naegele's rule cannot see any of this, because it never observes ovulation. It infers it.

Cycle length is the one adjustment a formula can make

Because the variability sits before ovulation, a formula can partially correct for it if it knows your usual cycle length. The standard adjustment shifts the estimated date by the difference between your cycle and twenty-eight days, on the reasoning that a thirty-two-day cycle ovulates roughly four days later than a twenty-eight-day one.

Our calculator applies that correction. Holding the same last menstrual period of 5 January 2026, the estimated date moves as follows.

Fourteen days separate the ends of that table, from a single input that most calculators do not even ask for. If your cycles run long or short and you have been using a tool that assumes twenty-eight days, your estimate is systematically displaced, and the displacement is large enough to matter for decisions about induction and for the interpretation of growth measurements.

The correction has an obvious limitation: it presumes you have a usual cycle length. Irregular cycles, recent hormonal contraception, breastfeeding, perimenopause, thyroid disease and polycystic ovary syndrome all make the input unreliable, and when the input is unreliable the adjustment cannot rescue the estimate. In those situations the formula is not the right instrument, and a clinician will say so.

Dating by conception, IVF transfer and ultrasound

When the date of conception is genuinely known, the two-week assumption can be discarded and the estimate improves.

After assisted reproduction, the timing is documented rather than inferred. A day-5 blastocyst transfer places the embryo at a known developmental stage on a known calendar day, so the arithmetic runs backwards with confidence. Our calculator handles the common cases separately: from the same anchor date of 5 January 2026, the last-menstrual-period method returns 12 October 2026, a known conception date returns 28 September 2026, a day-3 embryo transfer returns 25 September 2026 and a day-5 transfer returns 23 September 2026. These are not competing estimates of the same thing. Each method interprets the date you entered as a different event, so entering the same day under different methods should give different answers.

Ultrasound is the other route, and where it disagrees with a formula, ultrasound generally wins. ACOG Committee Opinion 700, published in Obstetrics and Gynecology (2017), sets out the current approach: an ultrasound measurement of crown-rump length in the first trimester is the most accurate method of establishing gestational age, and when it differs from the menstrual dating by more than a defined margin, the ultrasound date should be adopted. The margin widens as pregnancy advances, because fetal growth becomes more variable between individuals and a measurement of size becomes a progressively weaker proxy for age. The committee is explicit that once a due date has been established by the best available method, it should not be revised on the basis of later scans.

The WHO recommendations on antenatal care for a positive pregnancy experience (2016) make a related point from a global health perspective, recommending one ultrasound scan before twenty-four weeks in part for accurate dating, since so much of what follows in antenatal care depends on knowing where a pregnancy actually is.

The due date is a midpoint, not an appointment

The single most useful thing to understand about an estimated due date is how few births occur on it. Fewer than one in twenty arrive on the estimated date itself. The date is the centre of a distribution, and the distribution is wide.

Jukic and colleagues, in Human Reproduction 28:2848-2855 (2013), addressed this with unusual precision. Working from a cohort in which ovulation had been identified using daily hormone measurements rather than inferred from cycle dates, they measured the interval from ovulation to birth directly. Even with conception timing known, the length of gestation varied by more than five weeks across spontaneous pregnancies with no obvious complication. Some of that variation was associated with maternal characteristics and with events in early implantation, but a substantial part of it was simply variation.

That study removes the usual excuse for imprecise dating. It is tempting to assume that estimates would be sharp if only we knew the conception date, and Jukic's data show otherwise. Human gestation length is genuinely variable. A formula cannot be more precise than the biology it is describing.

This is why the clinically meaningful language is about ranges rather than points. Term is a period, not a day, and our calculator marks the start of the full-term window at thirty-seven weeks rather than presenting the estimated date in isolation. Anything about induction, post-term management or a pregnancy that is not progressing as expected is a conversation with your midwife or obstetrician, who is weighing the dating alongside growth, movements, blood pressure and your history.

What the weeks actually mark

Gestational age is the index that organises antenatal care, which is why the number is quoted at every appointment. Screening tests have windows measured in weeks, and several are only valid inside them.

Our calculator lays out the standard landmarks against your own dates. For a last menstrual period of 5 January 2026 with a standard cycle, the first trimester ends on 6 April, the second begins on 13 April, the anatomy scan window opens around 25 May at twenty weeks, the third trimester begins on 20 July at twenty-eight weeks, and the full-term window opens on 21 September at thirty-seven weeks.

Trimester boundaries deserve a note, because they are conventions rather than physiological transitions. Different sources place the start of the third trimester at twenty-seven or twenty-eight weeks, and both appear in reputable references; we use twenty-eight for consistency with the milestone list. Nothing changes in a pregnancy on the day a trimester label does. The boundaries are useful for organising information, not for interpreting how a pregnancy is going.

Using the number well

Know your dating method, and know that not all estimates are equivalent. An estimate derived from a first-trimester scan is a stronger number than one derived from a remembered period date, which is in turn stronger than one derived from an uncertain period date and an assumed cycle length.

Give a calculator the inputs it needs. If your cycles are not twenty-eight days, entering your actual length moves the estimate by up to a fortnight, and a tool that does not ask is quietly assuming.

Treat the estimate as a window. Planning around a five-week span rather than a single day is a more accurate reflection of the evidence and, in practice, a less frustrating way to approach the last month.

And keep the estimate in its place. The dating drives which tests happen when, and it will be revisited by the people caring for you if the clinical picture calls for it. Some warning signs need prompt contact with your maternity unit: vaginal bleeding, a reduction or change in fetal movements, severe or persistent headache, visual disturbance, abdominal pain, or fluid loss. These are urgent regardless of what week a calculator says you are in, and none of them should wait for a scheduled appointment.

For what happens next in the pregnancy, pregnancy weight gain: what the guidelines actually say covers how gain is assessed against your starting point, and nutrition during pregnancy covers the intake changes the guidelines describe. If you are working out where your cycle sits in all of this, the menstrual cycle and training explains why cycle length varies as much as it does in the first place.