Supported by NIH grants RR00166 and HD02274.
Introduction Fetal ultrasound measures of growth have a variety of scientific and clinical uses. These include studying normal and deviant fetal growth in nonexperimental populations, assessing effects of drugs, poisons, and disease agents on fetal growth, assessing relationships between abnormal fetal growth and poor pregnancy outcomes, and identifying fetuses exhibiting intrauterine growth retardation or acceleration. With the exception of identifying abnormal growth, all of these uses depend on having a reliable estimate of conception dates, independent of ultrasound measures themselves (Kramer et al, 1988; Pryor, 1997).
Ultrasound measures are also used by some investigators and clinicians to estimate gestational age and provide estimates of conception dates. However, unless the investigator or clinician makes the assumption that the fetus is growing normally— e.g., within the 90% confidence limits for a particular parameter— ultrasound measures cannot provide reliable estimates of gestational age or conception dates for an individual fetus (Oechsli, 1990). This potential misuse of ultrasound growth measures is especially important in experimental and epidemiological research where the variables under study are expected to produce alterations in growth rate. Under such circumstances, individuals that actually grow fast or slow will always have inaccurately estimated gestational ages from ultrasound growth norms, providing misleading or even false conclusions.
The purposes of the data provided here are to (1) summarize selected fetal growth parameters representing norms for our M. nemestrina colony, and (2) provide equations for users to estimate gestational age based on these norms. The data and methodology are based on a report by Conrad et al (1995). We assume that an independent estimate of the conception date is available when these equations are being used to identify deviant fetal growth rates. However, even without an independent conception date estimate, these equations may be useful for identifying deviant growth patterns when different fetal growth measures may yield markedly different gestational age estimates.
Methods Longitudinal fetal ultrasound measurements were taken on 39 pigtailed macaque dams selected from the Washington Regional Primate Research Center colony. The table below shows the sample sizes at each measurement period. Sample sizes decreased progressively across gestational ages, as two fetus were delivered by C-section at selected ages for neuroanatomical studies. The dams and sires bred for this study had average or above average reproductive histories for this colony, with no evidence of prior fetal growth retardation (Sackett, 1990). A time-mating procedure was used to estimate conception dates. Dams were placed in a male's cage for 24-72 hours starting on the day of perineal skin swelling detumescence. This procedure yields an average conception date accuracy within plus 3 days (White, Blaine, & Blakley, 1973). Pregnancies were confirmed by progesterone and NIH pregnancy kit tests 12-20 days after mating.
Ultrasound measurements were taken at 7-10 day intervals, beginning on day 30. Real time imaging was done using a 5 MHz linear array probe, except when structures exceeded the 45mm frame size. In those cases static imaging was performed with a 5 MHz static transducer. All measurements were made by an ultrasound specialist. Two global measurements, crown-rump length and total uterine volume, were recorded along with measures from the abdomen, head, limbs, and thorax. Eighteen different measures were recorded. A test-retest reliability study revealed an average error across all measures of 2%, with a maximum error of 9%, and all other error estimates less than 5%.
Number of pregnant females assessed at each gestational day by ultrasound exam.
Trimester 1
Day
N
Trimester 2
Trimester 3
Data were analyzed using both linear and non-linear growth models. For linear models, linear and quadratic regression lines were fit to individual fetal growth curves within trimesters. Trimester 1 went from day 30-50, trimester 2 from 60-110, and trimester 3 from 120-165. Non-linear models were fit to each individual growth curve over the gestation period 30, 40, 45, or 50-165 days, depending on the earliest age at which a structure could be visualized and accurately measured.
The best parameters obtained by non-linear estimation were bi-parietal diameter, head area, head circumference, and femur length. These parameters together explained over 98% of the variance in gestational age. Gestational age estimates based on the non-linear equations can be obtained below for each of these parameters. In addition, gestation age estimates for crown-rump length can be obtained from a first trimester linear equation. Each of the other four measures can also be estimated from the linear equations within trimesters two or three. Crown-rump length could not be estimated accurately after the first trimester because total fetal size was too large for our transducers to fit onto a single scan.
We present both linear and non-linear estimates for the following reasons. The linear equations provide slightly, but significantly, better fits to the data with smaller error estimates than do the nonlinear equations. To use the linear equations it is necessary to have an independent estimate of the conception date because the user must know the correct trimester (see the table above for the trimester age ranges) to validly employ these equations. The nonlinear equations can be used without an independent conception date estimate. However, this use is to determine if all parameters are growing at equal rates—i.e., have about equal gestational estimates—as markedly different gestation estimates are one indicator of abnormal fetal development. As discussed in the introduction section, the nonlinear equations without an independent conception date estimate cannot be used to estimate gestational age unless it is assumed that the fetus is growing at a normal rate.
Select the links below to see figures showing 95% confidence bands for Crown-Rump Length from the first trimester linear equation, and for the other four parameters from the non-linear equations. All of the raw data points are also shown in each figure. Although small sex differences probably occur by mid-term or shortly thereafter (Tarantal & Hendrickx, 1988), our sample size was too small to detect any statistically reliable differences between male and female fetuses.
Gestation Age Equations Enter your measurement value in the box below and check the appropriate measurement description. Linear Equations
References
Conrad, S., Ha,J., Lohr, C., & Sackett, G. (1995). Ultrasound measurement of fetal growth in Macaca nemestrina. American Journal of Primatology, 36, 15-35.
Kramer, M.S., McLean, F.H., Boyd, M.E., & Usher, R.H. (1988) The validity of gestational age estimation by menstrual dating in term, preterm, and postterm gestations. Journal of the American Medical Association, 260, 3306-3308.
Oechsli, F.W. (1990) Ultrasound fetoscopy and intrauterine growth retardation: two misused fashionable ideas. Paediatric and Perinatal Epidemiology, 4, 8-21.
Pryor, J. (1996) The identification and long term effects of fetal growth restriction. British Journal of Obstetrics and Gynaecology, 103, 1116-1122.
Sackett, G.P. (1990) Sires influence fetal death in pigtailed macaques (Macaca nemestrina). American Journal of Primatology, 20, 13-22.
Tarantal, A.F., & Hendrickx, A.G. (1988) Prenatal growth in the cynomolgus and rhesus macaques (Macaca mulatta and Macaca fascicularis): A comparison by ultrasonography. American Journal of Primatology, 15, 308-323.
White, R.J., Blaine, C.R., & Blakley, G.A. (1973) Detecting ovulation in Macaca nemestrina by correlation of vaginal cytology, body temperature, and perineal tumescence with laparoscopy. American Journal of Physical Anthropology, 189-195.