In a recent editorial (Vol 104, November 1997)1 it is argued that antenatal caregivers should use Nägele's rule, rather than ultrasound, to estimate the expected date of delivery. Conversely, a good argument can be made out to estimate the delivery date by ultrasound and to use menstrual dating only to decide when to do an ultrasound examination2.
This is an argument which is unlikely to be resolved–two significant digits of standard deviation is unlikely to convince anyone in the ultrasound camp to switch allegiance to the menstrual daters, or even to the ‘middle ground’ of only altering the menstrual dating if it differs more than seven, ten or fourteen days (whichever the case may be) from the ultrasound estimation.
Is such accuracy necessary? In the second trimester, ultrasound estimation of the gestational age improves the accuracy of biochemical screening for spina bifida and Down's syndrome. On the other hand, it can be argued that ultrasound screening can replace biochemical screening of both these conditions, without prior knowledge of the exact gestational age.
In the third trimester gestational age is of use to decide on expectant management or delivery in a complicated pregnancy. On the other hand, other factors (such as the neonatal care facilities and even the obstetrician's opinion whether the fetus is viable) have even greater influence on the neonatal outcome3. The ‘accurate’ estimation of the expected date of delivery by ultrasound has led to the virtual disappearance of amniocentesis to confirm lung maturity prior to elective caesarean section. Conversely, an elective delivery three days earlier or later is unlikely to have a major impact on neonatal morbidity, or on the number of women who go in to spontaneous labour before their planned caesarean section. That leaves two arguments for calculating the expected date of delivery. The first is to provide the woman (and her in-laws) with an estimation of the approximate time when she can expect to deliver. Obviously, we cannot pinpoint the spontaneous onset of delivery in advance. The second argument is to avoid unnecessary inductions of ‘post term’ pregnancies. Post term pregnancy is arguably an iatrogenic disease. If we did not create the false expectation of an ‘expected date of delivery’ and the anxiety this develops in the woman, her family and her obstetrician if she has not yet delivered by the ‘expected date’, the pressure for induction might be decreased.
Would any obstetrician who declines to provide a pregnant patient with an ‘expected date’ have any patients? An alternative might be to provide the woman with a ‘Cumulative frequency of delivery’ graph4. This graph illustrates her chance of spontaneous onset of labour by a certain date. This might avoid the anxiety generated by the tyranny of the ‘expected date of delivery’, while also graphically illustrating the time scale when she might reasonably expect to deliver.