Drey, Eleanor A. MD, EdM*; Kang, Mi-Suk MPH*; McFarland, Willi MD, PhD†; Darney, Philip D. MD, MSc*
Correct dating of pregnancies is important in the fields of obstetrics, gynecology, neonatology, and pathology. For those pregnancies that will be terminated, correct dating reduces risks of abortion procedures and determines abortion techniques. Abortion technique changes as gestational duration advances. For example, greater preoperative cervical dilation is required as pregnancy progresses.
Because the date of the last menstrual period (LMP) is not a consistently valid predictor of gestational duration, most abortion providers confirm or determine gestational duration by ultrasonography. In the second trimester, biparietal diameter (BPD) has been shown to be a reliable single measurement for estimating gestation. However, it is common practice for physicians to confirm dating after pregnancy termination by measuring the fetal foot length. Thus, obstetricians, neonatologists, and pathologists use foot length tables not only for quality assessment of ultrasound dating, but also to determine gestational duration.
Although previous logarithmic equations and tables of foot lengths according to gestation have been developed, the majority of studies either dated pregnancies using LMP (usually confirmed by physical examination) or did not specify how gestational duration was established.1–7 Currently used reference tables do not reflect modern, more precise dating of pregnancies allowed by ultrasound confirmation.
Most studies either examined fetal foot length in predominantly white populations or did not specify ethnicity. The two studies that examined ethnic differences did not find significant differences in measurements between groups.8,9
The high volume and multiethnic population of our clinic offered us the opportunity to assess ranges of fetal foot length by gestational duration in a setting with routine use of ultrasound dating by BPD. The goals were to establish normative fetal foot length ranges using LMP and ultrasound dating by BPD and to examine variations in these ranges by ethnicity.
MATERIALS AND METHODS
San Francisco General Hospital is a 530-bed public hospital serving an ethnically diverse population. Approximately 2,000 pregnancies are terminated yearly at the Women's Options Center, an outpatient abortion clinic located within the hospital. About half of these are second-trimester terminations by dilation and evacuation (D&E). Of the D&Es, a majority are performed between 19 and 23 weeks of gestation. The patient population is a mix of lower-income San Francisco residents, who include many recent immigrants, and patients with more advanced gestations, genetic anomalies, drug dependence, or pregnancies complicated by illnesses, who are referred from throughout Northern California.
As part of a study of human immunodeficiency virus prevalence approved by the Committee on Human Research of the University of California, San Francisco, a consecutive series of 2,025 women were selected for blinded serologic testing and chart abstraction during 2001 and 2002. Chart abstraction included demographic characteristics and data that could be used to determine gestational duration, including LMP, sonographic determination of BPD, and direct measurement of fetal foot length immediately after pregnancy termination. Records were excluded from our analysis if they had known chromosomal or morphologic abnormalities, fetal demise, unspecified dating criteria, or no record of fetal foot length measurement. Data from multiple gestations were included. Gestations less than 10 weeks by ultrasound examination were excluded because of imprecision in measuring foot length in early pregnancy.
All patients having D&Es had gestational duration determined by using ultrasound measurements of BPD on GE RT3200 or GE Logiq α200 machines (General Electric, Milwaukee, WI) with a 3.5-mHz linear-array transducer. Biparietal diameter was measured in a standard fashion from the skull's leading edge to leading edge (from the outer to the inner margin of the skull) at the standard level for intracranial landmarks,10 including the falx, cavum septum pellucidum, third ventricle, and thalamic nuclei. Our clinic's BPD limit for elective terminations is 56 mm (equivalent to 23 weeks, 1 day). When BPDs are close to that limit (within 1 mm), our protocol requires confirmation of gestational duration by a femur length measurement. Femur lengths were not included, however, in the data abstracted for this study. Both ultrasound machines used Hadlock's growth tables to determine gestational duration.11
Fetal foot length was measured from the heel to the first or second toe, whichever was longer, using a translucent ruler immediately after the abortion procedure and before formalin fixation. A second investigator reviewed all chart abstraction forms for accuracy, and we double-entered all data to catch entry errors.
To determine the most reliable measurement of gestational duration, we established a definition of “best estimate” for those records in which the gestation by LMP was within 1 standard deviation from the ultrasonographic estimate of gestation based on the growth charts used to calculate the ultrasound gestational duration.11 In those cases, the LMP-determined gestational duration was used. Unadjusted medians and interquartile ranges for foot length were calculated by LMP-determined gestation for the entire data set and for measurements satisfying the “best estimate” criteria. Linear regression modeled the relationship between fetal foot length and gestation based on the 3 different methods of calculating gestation: LMP dating alone, ultrasound dating alone, and “best estimate” dating. To determine the impact of ethnicity on the relationship between foot length and gestational duration, separate linear regression models were performed for whites, blacks, Latinas, and Asian and Pacific Islanders. To avoid making assumptions about data beyond the range collected here, we present data only for pregnancies between 10 and 25 weeks gestation as determined by LMP. All data were analyzed using SAS 8.2 (SAS Institute, Cary, NC).
Of the 2,025 women whose charts were abstracted, 871 were excluded because no fetal foot length was recorded, and 55 were excluded for other reasons. Some records had multiple exclusion criteria. Records without a fetal foot length were primarily those from early first-trimester pregnancies in which measurement of the fetal foot is impossible or inaccurate. Our study sample was ethnically diverse, with 25% white, 36% black, 22% Latina, and 12% Asian or Pacific Islander (Table 1). Reflecting the limitations of LMP dating in establishing gestational duration, only 30% of the sample reported being sure of their LMP dates. Using our “best estimate” criteria, 45% of the sample had agreement between LMP-established and ultrasonographically established gestational duration.
Figure 1 describes the median foot length values by each dating method: LMP alone, ultrasonography alone, and “best estimate.” Although the median foot length values by the LMP-determined and “best estimate” samples were not very different, the ranges of values around the medians were much greater in the sample dated by LMP. When using ultrasonography only, the range of values was almost identical to that found using the “best estimate” sample.
In fitting the linear model, simple linear regression demonstrated the best fit to the foot length data. Although squared and cubic terms were introduced, neither improved the model fit and were dropped from the final models. Gestational duration by ultrasonography alone produced a better model fit than duration by LMP alone (R2 of 0.92 compared with 0.27). Using the ultrasound dating that we had available for our entire sample (n = 1,099) yielded the following equation: foot length = −30.3 + days of gestation × 0.458. Regression by LMP-determined gestational duration using the “best estimate” sample (n = 491) provided an almost identical equation (foot length = −29.8 + days of gestation × 0.45) and a similar R2 value of 0.87, although the standard errors were larger. Regression by LMP-determined gestational duration alone (n = 875) gave this equation: foot length = 8.2 + days of gestation × 0.159 (R2 = 0.27).
Figure 2 shows the regression line established using the records fitting the “best estimate” criteria and the raw values measured in the full sample, illustrating the spread of the “best estimate” values relative to that of the full sample. Using the regression equation generated by the model using the “best estimate” sample, we have predicted ranges for fetal foot length based on LMP-determined gestational duration (Table 2).
Regressions by ethnicity were not significantly different compared with the simple regression, regardless of method used to determine gestational duration. Including ethnicity in the model, we obtained the following R2 values for the different ways of determining gestational duration: 0.92 for ultrasonography alone, 0.87 for “best estimate,” and 0.28 for LMP.
Many authors have worked to confirm Streeter's original foot length measurements.1–7,12 However, our study differs from these because we used ultrasound dating in a large, diverse population. We thus were able to recalibrate normal ranges of foot length by the best-available methods of dating pregnancies, giving us more reliable dating estimates than earlier tables.
We also are able to present estimated ranges based on a regression model calculated in units of exact days of gestational duration and provide the ranges with clear definitions of durations not specified in earlier studies. We therefore provide a statistically modern “gold standard.” A limitation of many earlier constructions of normative fetal measurements is their lack of specificity in discussing durations. For example, ranges for completed weeks compared with exact weeks would be expected to be different, because measurements truncated to completed weeks would include data for fetuses up to 6 days older than the recorded value. The truncation to a measurement of “completed weeks” of gestation also can be problematic in constructing normal ranges for fetal measurements, because many fetal measurements increase rapidly, particularly as the gestation advances.13
From these data we conclude that ultrasonography yields the estimate of gestational duration that best fits our data, and that it is similar to a “best estimate” measurement of duration. Not surprisingly, LMP is less reliable than either ultrasound or “best estimate” approaches. Using ultrasonography alone, we were able to explain 92% of the variation in fetal foot length, compared with only 27% of the variation by LMP data alone. That this fit is so remarkably close reflects the unsurprising close correlation between fetal foot and head sizes. Ultrasonography alone is the best available predictor of fetal foot length and is not statistically different from the “best estimate” approach commonly used in obstetrics for calculating gestational duration.
Our calculated range of dates for fetal foot length measurements is closest to that of Mercer et al1 despite their use of ultrasound-determined foot lengths rather than direct measurements, although at upper gestational ages, their measurements overestimate gestational ages compared with our data. Our foot length measurements are consistently greater than those of studies relying primarily on LMP dating.2,7,9,12 Our model thus could decrease any potential overestimation of gestational duration by foot length that might occur by using these earlier data for reference values.
In our sample, regression by LMP alone weakened the relationship between gestational duration and foot length. Thus increasing inaccuracy flattens the true relationship between increasing gestation and foot length. This inaccuracy is magnified at the extremes of gestational duration. This weakness of LMP dating is most concerning at the later gestational durations, where the limits often are critical. Although ultrasonographic determinations (eg, BPD, femur length, abdominal circumference) are less precise in calculating gestational duration as pregnancies advance, for the purpose of predicting fetal size, BPD alone is far more useful than LMP and can be trusted at least as much as an LMP confirmed by ultrasonography (what we have referred to here as the “best estimate”). If clinicians rely on BPD dating, they rarely would be expected to abort a fetus of a greater size or more advanced gestation than predicted by the dating ultrasonogram.
Most physicians have definite gestational duration limits beyond which they do not provide abortions. If one were to rely primarily upon LMP dates to model foot length, 382 women of our 1,099-person sample would have been considered to be beyond our clinic's cutoff of 23 weeks 1 day, whereas only 19 would have been excluded using our “best estimates” model. Not only does this demonstrate the clinical importance of accurate estimates of gestational duration, but it also reflects a problem with relying on previous LMP-based tables, which similarly overestimate duration by fetal foot length.2,7,9,12 Our results also show that the “best estimates” dating is similar to BPD dating, which gives us confidence that BPD provides adequate estimation of gestational duration to rely upon as a single measurement for second-trimester pregnancy termination. The stability of our foot length regressions regardless of ethnicity makes us confident that our calculations are reliable in a variety of patient populations.
One of the most important determinants of safe abortion provision is the size of the fetus, and correct dating is known to decrease abortion complications.14 Our data provide reassurance that preoperative ultrasonographic estimates of gestational duration by BPD accurately predict fetal size and allow for safe surgical abortion.
1. Mercer BM, Sklar S, Shariatmadar A, Gillieson MS, D'Alton ME. Fetal foot length as a predictor of gestational age. Am J Obstet Gynecol 1987;156:350–5.
2. Hern WM. Correlation of fetal age and measurements between 10 and 26 weeks of gestation. Obstet Gynecol 1984;63:26–32.
3. Kellokumpu-Lehtinen P. Age determination of early human embryos and fetuses. Ann Hum Biol 1984;11:567–70.
4. Kaul SS, Babu A, Chopra SR. Fetal growth from 12 to 26 weeks of gestation. Ann Hum Biol 1986;13:563–70.
5. Birkbeck JA, Billewicz WZ, Thomson AM. Human foetal measurements between 50 and 150 days of gestation, in relation to crown-heel length. Ann Hum Biol 1975;2:173–8.
6. Mandarim-de-Lacerda CA. Foot length growth related to crown-rump length, gestational age and weight in human staged fresh fetuses. An index for anatomical and medical use. Surg Radiol Anat 1990;12:103–7.
7. Munsick RA. Human fetal extremity lengths in the interval from 9 to 21 menstrual weeks of pregnancy. Am J Obstet Gynecol 1984;149:883–7.
8. Jakobovits A, Westlake W, Iffy L, Wingate M, Caterini H, Chatterton R, et al. Early intrauterine development: II. The rate of growth in Black and Central American populations between 10 and 20 weeks’ gestation. Pediatrics 1976;58:833–41.
9. Munsick RA. Similarities of Negro and Caucasian fetal extremity lengths in the interval from 9 to 20 weeks of pregnancy. Am J Obstet Gynecol 1987;156:183–5.
10. Filly RA, Hadlock FP. Sonographic determination of menstrual age. In: Callen PW, editor. Ultrasonography in obstetrics and gynecology. Philadelphia (PA): W.B. Saunders Company; 2000. p. 146–70.
11. Hadlock FP, Deter RL, Harrist RB, Park SK. Estimating fetal age: computer-assisted analysis of multiple fetal growth parameters. Radiology 1984;152:497–501.
12. Streeter G. Weight, sitting height, head size, foot length, and menstrual age of the human embryo. Contrib Embryol 1920;11:143–70.
13. Royston P, Wright EM. How to construct “normal ranges” for fetal variables. Ultrasound Obstet Gynecol 1998;11:30–8.
14. Darney PD, Atkinson E, Hirabayashi K. Uterine perforation during second-trimester abortion by cervical dilation and instrumental extraction: a review of 15 cases. Obstet Gynecol 1990;75:441–4.
Figure. No caption available.