Computer analysis of the human embryo growth curve: differences between published ultrasound findings on living embryos in utero and data on fixed specimens.

Accurate information on the normal growth rate of the human embryo is fundamental to a better understanding of the embryonic period of pregnancy. Crown-rump length measured previously in utero (N = 227) with vaginal ultrasound in 107 in vitro fertilization (IVF) or gamete intrafallopian transfer (GIFT) singleton pregnancies was compared to the greatest length of fixed human embryos from the Carnegie collection, of known developmental stage whose postovulatory ages were estimated from menstrual histories. Average crown-rump length in utero was 60% of the greatest length of the fixed specimens prior to postovulation day 33, but were equal after postovulation day 40. The growth rate of in utero embryos and fixed specimens, analyzed by computer using exponential equations, was compared to linear and polynomial equations used in previously published embryo growth tables. The exponential equation, length = exp(a + B/age), fit in utero measurements best, while the equation length = exp[a + b/exp(age)] fit the fixed specimens best. Differences between length in utero and in fixed specimens may be related to distortion of the fixed embryos resulting from the formalin fixation, to ultrasound distortion, to curling of the embryo, or to incorrectly estimated ages of the fixed specimens. Study of human embryos in utero is now practical with vaginal ultrasound.