Pregnancy Dating Methods Compared: LMP vs. Ultrasound vs. IVF

A due date sounds like it should be one fixed fact about a pregnancy, the same way a birthday is one fixed fact about a person. In practice, it is an estimate, and the pregnancy dating methods compared in this guide each infer the same underlying true age from a different, imperfect proxy. Two expectant parents with genuinely identical pregnancies, timed the same way down to the day, can walk out of two different clinics with two different due dates, simply because one clinic dated from the last menstrual period and the other dated from an early ultrasound. Neither number is a lie. They are two different estimates of the same unknown fact, built from two different kinds of evidence.

This article has the four main pregnancy dating methods compared side by side: last menstrual period (LMP), ultrasound measurement, conception-date tracking, and, for pregnancies conceived that way, a known IVF transfer date. It also covers the specific, published rules doctors actually follow for deciding which method wins when two estimates disagree, and why the due date that comes out the other end of all of this is better understood as the middle of a range than as a deadline circled on a calendar.

Pregnancy dating methods compared at a glance

The table sets the four methods side by side before each one is taken apart below.

LMP dating: the default, built on two big assumptions

Last menstrual period dating, the method behind Naegele's rule and behind this site's own pregnancy calculator, starts from the first day of your last period and adds 280 days, the standard 40-week full term. It became the default starting point for a simple, practical reason: most people know the date their last period started far more precisely than they know the date they ovulated or conceived, so it is the one date a dating method can reliably ask for at the very first prenatal visit, before any test has been run.

LMP dating rests on two assumptions that are true on average across a population but not necessarily true for any one individual: a 28-day menstrual cycle, and ovulation occurring around day 14 of that cycle. Plenty of real cycles do not match either assumption closely. A naturally longer or shorter cycle shifts the real ovulation, and therefore the real conception date, earlier or later than day 14 without LMP dating having any way to know that shift happened. Irregular cycles, whether from polycystic ovary syndrome, recent use of hormonal birth control, breastfeeding, perimenopause, or simple individual variation, widen this gap further, sometimes by a week or more. None of that makes LMP dating useless; it makes it a starting estimate that a more precise method, when available, is specifically meant to refine.

Ultrasound dating: measuring the embryo or fetus directly

Ultrasound dating works by directly measuring the embryo or fetus and comparing that measurement against standardized growth curves built from large reference populations. In the first trimester, the key measurement is crown-rump length (CRL), the length from the top of the head to the bottom of the torso, which correlates very tightly with true gestational age at that early stage. Later in the second trimester, dating relies on a combination of measurements, commonly biparietal diameter (the width of the head), head circumference, femur length, and abdominal circumference, combined into a single estimated gestational age.

The reason a first-trimester ultrasound is generally considered the single most accurate dating method available comes down to biology, not technology: in the first trimester, embryos and early fetuses of the same true gestational age are remarkably similar in size to one another. There simply has not been enough time yet for individual growth differences, genetics, sex, or anything else to meaningfully diverge. A CRL measurement taken before about 14 weeks can typically pin down gestational age within roughly 3 to 5 days.

That precision degrades steadily as pregnancy progresses, and this is worth understanding clearly because it directly explains why a late ultrasound is not used for dating even though ultrasounds keep happening throughout pregnancy for other reasons. By the second trimester, individual size variation has grown enough that a dating estimate from ultrasound alone is typically only accurate to within a week or two. By the third trimester, genuine differences in growth rate, influenced by genetics, fetal sex, and individual variation that has had months to compound, can put a size-based age estimate off by two to three weeks or more, which is far too imprecise to use for dating a pregnancy, even though the same measurements remain genuinely useful for a completely different purpose: tracking whether growth is on track for a fetus whose age is already established.

Conception-based dating: precise if you know it, useless if you do not

Some people track ovulation directly, using basal body temperature charting, ovulation predictor kits that detect the LH (luteinizing hormone) surge, or fertility-awareness methods, and know their approximate day of ovulation and intercourse or insemination with more precision than a generic "day 14 of a 28-day cycle" assumption provides. In that situation, a due date can be calculated by adding 266 days (38 weeks) to the estimated date of ovulation or conception, rather than 280 days from an LMP, since conception itself, not the start of the prior period, is being used as the reference point.

Going the other way, the conception calculator subtracts those 266 days from a due date, and the ovulation calculator predicts the ovulation date from your last period and average cycle length when there is no tracking data. Why the gap is 266 days and not 280 is worked out in conception date versus due date minus 40 weeks.

Conception-based dating can meaningfully outperform standard LMP dating for someone with real ovulation-tracking data, precisely because it sidesteps the "assume day 14" guess entirely. It is not, however, more precise than a first-trimester ultrasound for most people, since even careful ovulation tracking still carries some uncertainty about the exact day fertilization occurred, and it is obviously not available at all to anyone who has not been tracking ovulation.

There is a specific, technical reason ovulation tracking tends to correct LMP dating more for some people than others, and it comes down to which half of the cycle is actually variable. A menstrual cycle has two phases: the follicular phase, from the start of a period until ovulation, and the luteal phase, from ovulation until the next period. The follicular phase is the one that varies a lot from person to person and even cycle to cycle, sometimes stretching well beyond or falling well short of the 14 days LMP dating assumes. The luteal phase, by contrast, tends to be far more consistent for a given individual, typically staying within a narrow band of around 12 to 16 days regardless of how long the front half of that same cycle ran. That asymmetry is exactly why a longer or shorter total cycle usually means a shifted ovulation date, not a shifted length of time from ovulation to the next period, and it is why tracking ovulation directly, through basal body temperature (which rises measurably after ovulation) or an ovulation predictor kit (which detects the LH surge roughly 24 to 36 hours before ovulation), corrects for exactly the part of LMP dating's assumption that is most likely to be wrong.

How ultrasound growth curves were actually built

Ultrasound dating is not derived from a formula worked out on paper; it is built from reference growth curves compiled by measuring thousands of pregnancies whose gestational age was already independently known, historically from a clear LMP or, later, from a large IVF dataset with a known transfer date, and plotting how embryo or fetal size at each measured week of confirmed gestational age tended to cluster. The most widely used modern formulas for translating fetal measurements into a gestational age estimate trace back to work published by Frank Hadlock and colleagues in the 1980s, which combined measurements like biparietal diameter, head circumference, abdominal circumference, and femur length into a single estimated age and, separately, an estimated fetal weight.

Population reference curves like these carry an inherent limitation worth naming plainly: they describe an average drawn from whichever population was actually measured to build them, and any individual fetus, from any population, can reasonably fall somewhere in the normal spread around that average rather than sitting exactly on it. This is part of an ongoing, genuine discussion in obstetric research about whether a single universal growth standard fits every population equally well, or whether growth curves should eventually be adjusted for factors such as parental height, ethnicity, or regional population differences. None of that changes the basic reliability of first-trimester CRL for dating specifically, since size variation is smallest precisely when dating ultrasounds are performed, but it is a fair caveat about later-pregnancy growth assessment, which is a different use of the same measurements than dating is.

Naegele's rule is over two hundred years old

The specific "add 280 days" calculation behind LMP dating is not a modern statistical model; it is generally attributed to Franz Karl Naegele, a German obstetrician who described the rule in the early 19th century, and it has remained in everyday clinical use, essentially unchanged, ever since. That kind of longevity is worth pausing on for a moment. Naegele's rule has stayed in use for two centuries not because a more sophisticated statistical alternative has never been proposed, but because it is simple enough to calculate by hand from information every pregnant person already has on the very first prenatal visit, and it turns out to be a reasonably good starting estimate across a large population even though it is a known simplification for any one individual. Ultrasound dating, conception tracking, and IVF transfer dates all exist as refinements on top of that same 200-year-old starting point, not replacements for the idea behind it.

Why the dating method matters for more than the due date

Gestational age is not just a countdown to delivery; it is a load-bearing number that several distinct parts of prenatal care are built directly on top of. First-trimester combined screening, which estimates the likelihood of certain chromosomal conditions from a blood test (measuring hormones including hCG and PAPP-A) combined with an ultrasound measurement called nuchal translucency, reports its result as a risk figure that is calculated specifically for the gestational age entered; get the age meaningfully wrong, and the reported risk figure shifts even though nothing about the actual pregnancy changed. Non-invasive prenatal testing (NIPT), a blood test that screens for certain conditions from cell-free fetal DNA circulating in a pregnant person's blood, has its own recommended earliest testing window, typically starting around 9 to 10 weeks, which depends on gestational age being reasonably accurate to avoid testing too early for a reliable result. The detailed mid-pregnancy anatomy ultrasound, usually scheduled in an 18-to-22-week window specifically because that window is when most major fetal structures are developed enough to be clearly visualized, is timed against gestational age as well. Later on, whether a birth is classified as preterm, early term, full term, late term, or postterm, a classification that can affect everything from delivery-room staffing decisions to how a newborn's readiness for certain routine procedures is judged, depends directly on which gestational age the pregnancy is actually being tracked against. Getting the dating method right, and correctly updating it through the standard redating rules when better information becomes available, ripples forward through essentially every date-dependent decision the rest of the pregnancy involves.

IVF transfer dates: the most precise dating available, when it applies

For a pregnancy conceived through in vitro fertilization, dating gets unusually precise, because the normal biggest source of uncertainty in every other method, not knowing exactly when fertilization happened, simply does not exist. The embryo's exact age at transfer is a known clinical fact: a "day-3" transfer places a 3-day-old embryo, and a "day-5" (blastocyst) transfer places a 5-day-old embryo, both recorded precisely by the fertility clinic.

Due dates for IVF pregnancies are calculated by working backward from that known age, using the same 280-day full term as any other method, just anchored to a different, more precise starting point. The standard 14-day gap between LMP and ovulation still applies, so an equivalent "LMP" date can be reconstructed as the transfer date minus 14 days minus the embryo's age at transfer. For a day-5 (blastocyst) transfer, that works out to a commonly used shortcut of adding 261 days to the transfer date (280 days minus 14 days for the ovulation offset minus 5 days for the embryo's age). For a day-3 transfer, the shortcut is 263 days from the transfer date (280 minus 14 minus 3).

As a formula, with a standing for the embryo's age at transfer in days:

The exact figure a specific fertility clinic uses can vary slightly depending on their own convention, so the number your clinic gives you, tied to your specific protocol and transfer day, should always take priority over a generic formula found online.

The actual rules for when a doctor changes your dates

With the four methods laid side by side, the redating decision itself is not a matter of individual preference; it follows published discrepancy thresholds that compare the ultrasound-estimated gestational age against the LMP-based gestational age, with the acceptable gap widening as pregnancy progresses (reflecting exactly the growing size-variation problem described above). The thresholds below come from ACOG Committee Opinion 700, Methods for Estimating the Due Date. The ultrasound date replaces the LMP date only when the gap is larger than the number shown:

Two things about this table are easy to miss. First, the threshold gets wider, not narrower, later in pregnancy, precisely because a late ultrasound's own dating precision is worse, so a bigger disagreement is needed before it is trusted over the existing estimate. Second, once a due date is set using these rules, standard practice is to keep it fixed for the rest of the pregnancy rather than continuously re-adjusting it every time a new ultrasound is performed, since later ultrasounds are measuring growth against an already-established age, not re-deriving that age from scratch.

Four scans show how the rule plays out. The gap has to be larger than the threshold, so a gap equal to the threshold keeps the LMP date.

The due date is a median, not a deadline

Every one of these methods ultimately produces the same kind of output: an estimated due date built from an average, 280 days from LMP, 266 from conception, or a clinic-specific IVF figure, applied to one specific, individual pregnancy that may not follow the average exactly. It is worth being explicit about what that means in practice: a due date marks the middle of a normal range of delivery timing, not a deadline a pregnancy is expected to hit exactly.

The commonly cited estimate is that only around 4 to 5 percent of babies are actually born on their calculated due date. The much more informative range is the one built into the ACOG-endorsed term windows: early term (37 weeks 0 days through 38 weeks 6 days), full term (39 weeks 0 days through 40 weeks 6 days), late term (41 weeks 0 days through 41 weeks 6 days), and postterm (42 weeks 0 days and beyond). A large share of deliveries happen across the early-term through late-term span, a roughly five-week window, and landing anywhere within it is well within normal variation, whichever single day a calculator happens to print as "the" due date. The numbers behind that 4 to 5 percent figure are worked through in the guide to babies born on their due date, and the due date calculator estimates a date from your last period or a known conception date.

What to actually do when your own dates disagree

Now that you have seen how the major methods differ, if an early ultrasound comes back with a date noticeably different from your own LMP-based estimate, the practical answer is straightforward: your provider is applying the redating rules above, not making an arbitrary call, and the resulting date is the one that should be used going forward for scheduling, screening timing, and every other date-dependent part of your care. If you conceived through IVF, your clinic's transfer-date-based due date is more precise than a generic LMP calculation and should take priority over any LMP-based estimate calculated separately. And if you are unsure which date is currently the one your care team is using, that is worth a direct, simple question at your next appointment rather than an assumption either way, for exactly the reasons the section above lays out: gestational age quietly drives far more of your prenatal care than the single number printed on a calendar.

To see the LMP-based date and your current week in one step, use the pregnancy calculator.