Years ago, a friend called by and asked us to look after her two children while she rushed into hospital to give birth. Labour pains had begun four weeks before her “due date”. Ironically, she is a physician who has since had a distinguished medical career; but her dilemma graphically reflected medicine’s casual approach to estimating birth dates.

How do physicians do it?

Because human ovulation lacks external signs, it is admittedly difficult to estimate reliable birth dates. Physicians generally calculate the due date from the first day of a woman’s last menstrual period (LMP). Many still use a rule of thumb originally proposed in 1812 by German obstetrician Franz Nägele: Subtract three months from the first day of LMP and add seven days. Curiously, physicians prefer this to reckoning from a standard duration of 40 weeks, yet Nägele’s Rule actually increases error because months differ in length!

Large-scale studies have consistently revealed enormous variability in pregnancy lengths calculated from LMP. The average is indeed close to 40 weeks (280 days), but birth dates for normal pregnancies can be more than a month astray on either side. In other words, Nägele’s Rule puts pregnant women on standby for two whole months. Ultrasound monitoring (if used) does allow physicians to refine predicted birth dates, but researchers have shown little interest in achieving greater accuracy.


Distribution of pregnancy durations for a large sample of births in Sweden, adapted from Bergsjø et al. (1990).

To say the least, it is inconvenient for an expectant mother to cope with such uncertainty about her due date, although the Latin verb expectare does mean “to wait”. But other things are at stake. Most seriously, physicians deciding dates for induction or Caesarean sections need reliable information on pregnancy duration to decide with confidence whether approaching births will be premature or overdue. Because half of normal births take place up to a month after the due date based on Nägele’s Rule, there is an inherent tendency to induce some births without real medical need.

Why are human pregnancy lengths so variable?

Ovulation, followed by conception, typically occurs halfway through a woman’s menstrual cycle. On average, then, the real duration of pregnancy  —  from conception to birth  —  is about two weeks less than the interval from LMP to birth: 266 days instead of 280. Substantial variation in ovulation time relative to LMP is the main reason for the very approximate due dates calculated by physicians.

But another crucial factor leads to even greater variation. Thus far, only single births have been considered. With more than one baby, things differ radically for a very simple reason: The womb cannot expand indefinitely, so babies are born increasingly smaller after ever shorter pregnancies as their number increases. In 1952, epidemiologists Thomas McKeown and Reginald Record published a landmark paper on fetal growth in multiple pregnancies, reporting a regular decline in pregnancy duration as number of babies increased. The average of 40 weeks from LMP for single births decreases to around 37 weeks with twins, 35 weeks with triplets, and 34 weeks with quadruplets. So births with multiple pregnancies are far more likely to occur before the medically recognized threshold of 37 weeks for prematurity.

McKeown and Record reported the relationship between multiple births and pregnancy duration over 60 years ago. So do physicians now routinely use modified versions of Nägele’s Rule to predict birth dates for multiple pregnancies? Unfortunately, this does not seem to be the case. Certainly, online calculators designed to allow women to reckon their own due dates generally fail to allow for the reduction in pregnancy length with multiple births. Our friend who rushed into hospital “a month early” in fact gave birth to her twins at about the expected time!

Doing a better job

Accurate information on human pregnancy length, dated from conception, is sorely needed. Mary Leakey’s autobiography recounts her discovery of a partial skeleton of the 20-million-year-old fossil ape Proconsul on Rusinga Island in Kenya. She rightly saw this as a “wildly exciting find which would delight human paleontologists all over the world”. Her major find also had another significant outcome: “When the magnitude of our discovery had sunk in, back in our camp at Kathwanga, Louis and I wanted to celebrate. We were exhilarated and also utterly content with each other and we thought that quite the best celebration would be to have another baby.” Their son Philip was born 262 days later. Here, it seems, is one fairly secure record of the interval between conception and birth. If nothing else, it graphically shows how exploring our origins can contribute to human reproduction.

But anecdotal evidence is not enough. At last, we now have comprehensive information on real pregnancy lengths. A team of epidemiologists has just published a paper reporting on an analysis using hormonal evidence of ovulation time. The average interval from ovulation to term birth was 268 days, and most normal pregnancies were less than two weeks astray on either side. So timing pregnancy from ovulation itself almost halves the variation. Translating this into routine testing for mothers would greatly reduce the current uncertainty of due dates.

Yet there is still much to be done. Although over four million babies have now been born through in vitro fertilization, I have been unable to find a single medical publication reviewing data for intervals between fertilization and birth. This eloquently underscores continued lack of interest in precise calculation of pregnancy lengths.


Bergsjø, P., Denman, D.W., Hoffman, H.J. & Meirik, O. (1990) Duration of human singleton pregnancy. A population-based study. Acta Obstet. Gynecol. Scand. 69:197-207.

Gibson, J.R. & McKeown, T. (1950) Observations on all births (23,970) in Birmingham, 1947. I: Duration of gestation. Brit. J. Soc. Med. 4:221-233.

Jukic, A.M., Baird, D.D., Weinberg, C.R., McConnaughey, D.R. & Wilcox, A.J. (2013) Length of human pregnancy and contributors to its natural variation. Hum. Reprod. 28: 2848-2855.

Leakey, M. D. 1984. Disclosing the Past. London: Weidenfeld & Nicolson.

McKeown, T. & Record, R.G. (1952) Observations on foetal growth in multiple pregnancy in man. J. Endocrinol. 8:386-401.

Nägele, F.K. (1843) Lehrbuch der Geburtshülfe. Erster Theil: Physiologie und Diätetik der Geburt. Mainz: Theodor von Zabern.

You are reading

How We Do It

Fake Sex Hormones: Chemicals Damage Health and Reproduction

Synthetic substances that mimic steroids have insidious side-effects.

Masturbation: Self-Abuse or Biological Necessity?

A dispassionate view of the natural history of self-gratification.

Bigfoot Parents Have Small Brains

The cautionary tale of Australia’s peculiar incubator birds