The length of pregnancy: an echographic reappraisal.

A study of 998 pregnant women was conducted to estimate the random errors in dating pregnancy by menstrual history and by ultrasound measurements of the fetus (crown-rump length or biparietal diameter before 20 weeks, menstrual age) as well as the biological variability of pregnancy length. The latter was found to be 7.74 days. The random errors made when dating pregnancy by menstrual history and by ultrasound measurements were 7.16 days and 4.26 days, respectively. (All variabilities are expressed as standard deviations). The differences between the two methods were mainly attributable to some large errors occurring when dating pregnancy on the basis of menstrual history. Differences between methods of more than 10 days were observed in about 10% of our population. When the two methods were used on the same patients to define a pregnancy as preterm, at term, or postterm, discrepancies were found in about 7% of cases.