SEARCH

SEARCH BY CITATION

Keywords:

  • birth records;
  • gestational age;
  • accuracy;
  • time trends;
  • preterm;
  • post-term;
  • SGA;
  • LGA

Summary

  1. Top of page
  2. Summary
  3. Introduction
  4. Methods
  5. Results
  6. Discussion
  7. Conclusions
  8. Acknowledgements
  9. References

Accurate estimation of gestational age early in pregnancy is paramount for obstetric care decisions and for determining fetal growth and other conditions that may necessitate timing the iatrogenic intervention or delivery. We sought to examine temporal changes in the distributions of two measures of gestational age, namely, those based on menstrual dating and a clinical estimate. We further sought to evaluate relative comparisons and variability in indices of perinatal outcomes. We utilised the Natality data files in the US, 1990–2002 comprising women that delivered a singleton livebirth between 22 and 44 weeks gestation (n = 42 689 603).

Changes were shown in the distributions of gestational age based on menstrual vs. clinical estimate between 1990 and 2002, as well as changes in the proportions of preterm (<37, <32 and <28 weeks) and post-term (≥42 weeks) birth, and small- (SGA; <10th percentile) and large-for-gestational-age (LGA; birthweight >90th percentile) births. While the absolute rates of preterm birth <37 weeks, SGA and LGA births were lower based on the clinical estimate of gestational age relative to that based on menstrual dating, the increases in preterm birth rate between 1990 and 2002 were fairly similar between the two measures of gestational dating. However, the decline in post-term births was larger, based on the clinical estimate (−73.8%), than on the menstrual estimate (−36.6%) between 1990 and 2002. While the clinical estimate of gestational age appears to provide a reasonably good approximation to the menstrual estimate, disregarding the clinical estimate of gestational age may ignore the advantages of gestational age assessment in modern obstetrics.


Introduction

  1. Top of page
  2. Summary
  3. Introduction
  4. Methods
  5. Results
  6. Discussion
  7. Conclusions
  8. Acknowledgements
  9. References

Gestational age based on the last menstrual period (LMP) is widely used to derive the date of delivery based on the assumption that the window between the date of last menstrual period and ovulation is 14 days. Gestational age based on menstrual dating is, however, fallible in at least four aspects: (i) normal cycle length can vary considerably between women; (ii) women with irregular menstrual cycles or anovulation may not adhere to the often presumed 14-day interval for onset of ovulation;1 in fact, irregular bleeding episodes may sometimes reflect spontaneous, yet unrecognised, miscarriages;2 (iii) bleeding early in pregnancy may often be mistaken for a delayed menstrual period thereby offsetting the date of last menstrual period by as much as 4 weeks;3 and finally (iv) errors in the woman's recall of her date of last menstrual period.4–14 For these reasons, early sonographic assessment of pregnancy dating is generally regarded as an important adjunct to dating based on menstrual dates.

Gestational age assessment based on menstrual dates, however, remains the most widely used approach in population-based studies. The use of ultrasound for dating has become an important part of the clinical assessment of gestational age, particularly in industrialised countries where most pregnant women have one or more ultrasound examinations during pregnancy. Ultrasound examinations are primarily used to confirm the date of the LMP or to assign an estimated date of confinement when there is considerable discrepancy in the fetal measurements (e.g. fundal height) as compared with gestational age based on menstrual dates.

The US Natality data files contain, in addition to the menstrual estimate of gestational age, a clinical estimate (CE) of gestational age.15 In spite of the availability of the CE of gestation on these data files, the usefulness of this relative to that based on menstrual dates in the US population remains largely unexplored.

The objectives of our study, therefore, were to assess in singleton livebirths (i) the comparability of menstrual vs. CE of gestation, and (ii) to contrast temporal trends in indices of perinatal outcomes, including preterm and post-term, small-for-gestational-age (SGA) and large-for-gestational-age (LGA) births.

Methods

  1. Top of page
  2. Summary
  3. Introduction
  4. Methods
  5. Results
  6. Discussion
  7. Conclusions
  8. Acknowledgements
  9. References

We used data from the US Natality data files from 1990 to 2002, with the analyses restricted to singleton livebirths. The data contained in these files were derived from livebirth certificates in all 50 States and the District of Columbia, and were assembled and compiled by the National Center for Health Statistics (NCHS) of the Centers for Disease Control and Prevention.16 The data are coded according to uniform coding specifications, have passed rigorous statistical quality checks, have been reviewed and carefully edited by the NCHS, and form the basis for official birth statistics reporting in the US.

Assessment of gestational age

Two measures of gestation are currently available on the Natality data files. These include gestational age based on menstrual dates, and a CE of gestation. Gestational age in these data files are largely based on menstrual dating (over 95% of births), calculated as the interval (in completed weeks) between the dates of the LMP and birth. In the remainder of births, when the menstrual estimate of gestational age was incompatible with the reported birthweight, the CE of gestation was instead substituted.17 In addition, missing gestational age based on menstrual date was statistically imputed.17 This imputation was done when a valid month and year of the LMP was available but the day was missing. The replacement of clinically estimated gestational age and the imputation of missing gestational age were both performed by the NCHS prior to release of the data.

Indices of perinatal outcomes

We assessed several perinatal outcomes in this study. These included preterm birth <37 completed weeks, as well as its severity defined by delivery at <28, <32 and 34–36 weeks, and post-term births defined as ≥42 weeks. In addition, we also examined changes in proportions of SGA and LGA births. SGA and LGA were defined as sex-specific birthweight <10th and >90th percentiles, respectively, at each week of gestation. The 10th and 90th percentile cut points of birthweight were derived from the 1990 livebirths in the US as a baseline (i.e. internal reference norms). Perinatal outcomes were examined in relation to both LMP and CE of gestational age.

Cohort composition

The analysis was restricted to singleton livebirths that occurred between 22 and 44 completed weeks' gestation based on menstrual dates. Of all livebirths delivered between 1990 and 2002, we excluded births with missing birthweight (n = 43 267), gestational age <22 weeks (60 767 and 13 729 based on menstrual date and CE, respectively), gestational age ≥45 weeks (430 117 and 1550 based on menstrual date and CE, respectively), and missing data on gestational age (489 017 and 6 992 666 based on menstrual date and CE, respectively). Data on CE of gestational age are not reported by the State of California to the NCHS/CDC prior to data compilation.17 Therefore, births from California were excluded from the analysis that involved the CE of gestational age. After all exclusions, 42 689 603 singleton livebirths delivered between 22 and 44 weeks remained for analysis.

Statistical analysis

Temporal trends in perinatal outcomes (preterm and post-term births, SGA and LGA births) between 1990 and 2002 based on menstrual and CE of gestational age were assessed through logistic regression models. In order to directly estimate the period effect denoting changes in the proportion of the perinatal outcomes between 1990 and 2002, we fitted the models using a log-link function (instead of the traditional ‘logit’ link function). From these models, the relative risk and 95% confidence interval [95% CI] were derived to denote trends.

This study was approved by the ethics review committee of the Institutional Review Board of UMDNJ-Robert Wood Johnson Medical School, New Brunswick, NJ, USA.

Results

  1. Top of page
  2. Summary
  3. Introduction
  4. Methods
  5. Results
  6. Discussion
  7. Conclusions
  8. Acknowledgements
  9. References

Characteristics of menstrual and CEs of gestational age and birthweight in 1990 and 2002 are shown in Table 1.

Table 1.  Distribution of gestational age based on menstrual estimate, clinical estimate and birthweight, US singleton livebirths, 1990 and 2002
 Menstrual estimate of gestational ageClinical estimate of gestational age
1990200219902002
Gestational age (weeks)
 Mean ± standard deviation39.1 ± 2.439.7 ± 2.439.2 ± 2.239.7 ± 2.1
 Median (range)39 (22–44)39 (22–44)40 (22–44)39 (22–44)
Birthweight (g)
 Mean ± standard deviation3356 ± 5783323 ± 582

Birthweight distributions

Shifts in mean birthweight-for-gestational age between 1990 and 2002 based on menstrual and CE of gestational age are shown in Fig. 1. At 22–32 weeks' gestation, the mean birthweight was, on average, 60 g lower in 2002 compared with 1990 based on the menstrual estimate (left panel), and 20 g lower in 2002 compared with 1990 based on the CE (right panel). At ≥32 weeks, the distributions of mean birthweight were fairly similar between the two measures of gestational age.

image

Figure 1. Changes in the distribution of mean birthweight between 1990 and 2002 for gestational age based on last menstrual period (left panel) and clinical estimate of gestational age (right panel), US singleton livebirths.

Download figure to PowerPoint

The distributions of birthweight between menstrual and CE of gestational age at <28, 28–31, 32–33, 34–36, 37–41 and ≥42 weeks are shown in Figs 2–4. At <28 weeks, the birthweight distributions based on the two measures of gestational age appear fairly similar but with the estimate based on menstrual date exhibiting a small ‘bump’ or second mode around 2500 g (Fig. 2, left panel). At 28–31 weeks, the menstrual estimate of gestational age exhibits a clear bimodal distribution. At 32–33 weeks (Fig. 3, left panel), the bimodal distribution of gestational age based on menstrualdates is apparent. At 34–36 weeks (Fig. 3, right panel), the overall birthweight distributions between the two measures of gestational age appear fairly similar, but with the distribution based on menstrual estimates slightly shifted more towards the right in comparison with the CE of gestational age. At 37–41 weeks, birthweight distributions between the clinical and menstrual estimate of gestational age are virtually identical (Fig. 4, left panel), whereas at post-term gestations (42–44 weeks), the menstrual estimate of gestational age appears shifted more towards the left (Fig. 4, right panel).

image

Figure 2. Distribution of birthweight by gestational age based on last menstrual period and clinical estimate of gestational age at 22–27 weeks (left panel), and at 28–31 weeks (right panel), US singleton livebirths, 2002.

Download figure to PowerPoint

image

Figure 3. Distribution of birthweight by gestational age based on last menstrual period and clinical estimate of gestational age at 32–33 weeks (left panel), and at 34–36 weeks (right panel), US singleton livebirths, 2002.

Download figure to PowerPoint

image

Figure 4. Distribution of birthweight by gestational age based on last menstrual period and clinical estimate of gestational age at 37–41 weeks (left panel), and at 42–44 weeks (right panel), US singleton livebirths, 2002.

Download figure to PowerPoint

Trends in indices of perinatal outcomes

Temporal changes in the proportions of preterm and post-term births between 1990 and 2002 based on the two measures of gestational age assessments are shown in Fig. 5. Proportionate increases in preterm birth rate (<37 weeks) between 1990 and 2002 were fairly similar between menstrual-based and CE of gestational age (21.2% and 17.3%, respectively), although the absolute rates of preterm birth were generally lower based on the CE in comparison with the menstrual estimate (Table 2). Increases in preterm birth at 34–36, <32 and <28 weeks were consistently larger based on the CE than the menstrual estimate, despite the overall lower absolute rates. In contrast, the decline in post-term birth rate was 36.6% using the menstrual estimate of gestational age, whereas the decline was 73.8% using the CE of gestational age.

image

Figure 5. Distribution of gestational age-specific mean birthweight based on menstrual estimate and clinical estimate of gestational age (left panel), and average difference (in weeks) between gestational age based on menstrual dates and clinical estimate of gestational age (right panel), US singleton livebirths, 2002.

Download figure to PowerPoint

Table 2.  Trends in preterm and post-term births based on menstrual and clinical estimates of gestational age, US livebirths, 1990 and 2002
 Menstrual estimate of gestational ageClinical estimate of gestational age
1990 (%)2002 (%)Change (%) [95% CI]1990 (%)2002 (%)Change (%) [95% CI]
Preterm birth (weeks)
 <3710.012.121.1 [20.6, 21.6]7.48.617.3 [16.6, 17.9]
 34–368.18.77.2 [6.6, 7.8]5.06.220.9 [20.1, 21.7]
 <322.02.14.9 [4.2, 5.5]1.41.614.5 [13.0, 16.1]
 <280.70.822.3 [19.9, 24.7]0.50.739.5 [36.6, 42.4]
Post-term birth ≥42 weeks10.76.8−36.6 [−37.0, −36.3]4.41.4−73.8 [−74.1, −73.5]

Gestational age-specific mean birthweight in 2002 based on the two measures of gestational age (left panel) and average difference (in weeks) between the two measures of gestational age (right panel) are shown in Fig. 6. The mean birthweight between 28 and 34 weeks is, on average, 200 g higher for the menstrual than CE of gestational age. At preterm gestational ages, the CE of gestational age is higher in comparison with that based on menstrual dates, with the difference reaching almost 2.5 weeks between 28 and 32 weeks, and begins to decrease thereafter. At term and beyond, the LMP-based gestational age is higher than the CE, the difference reaching almost 4 weeks at post-term gestations.

image

Figure 6. Temporal trends in preterm birth <37 weeks (left panel) and post-term birth 42–44 weeks (right panel) based on gestational age using menstrual dates and clinical estimates of gestational age, US singleton livebirths, 1990–2002.

Download figure to PowerPoint

Temporal changes in the proportions of SGA births between 1990 and 2002 are shown in Fig. 7. Proportions of SGA births declined between 1990 and 2002 by 4.3% and 12.9% based on menstrual estimate and CE of gestational age, respectively (Table 3). The proportion of LGA births also declined between the two periods by 4.4% using the menstrual estimate of gestational age, whereas the decline was 1.8% based on the CE of gestational age.

image

Figure 7. Temporal trends in SGA births (left panel) and LGA births (right panel) based on gestational age using menstrual dates and clinical estimate of gestational age, US singleton livebirths, 1990–2002.

Download figure to PowerPoint

Table 3.  Trends in small-for-gestational (SGA) and large-for-gestational (LGA) age births based on menstrual and clinical estimates of gestational age, US livebirths, 1990 and 2002
 Menstrual estimate of gestational ageClinical estimate of gestational age
1990 (%)2002 (%)Change (%) [95% CI]1990 (%)2002 (%)Change (%) [95% CI]
SGA10.09.5−4.3 [−4.7, −3.8]10.59.1−12.9 [−13.3, −21.5]
LGA10.09.4−4.4 [−4.8, −3.7]8.78.5−1.8 [−2.3, −1.3]

The efficacy of menstrual estimate of gestational age in predicting adverse perinatal outcomes, with the CE of gestational age considered the ‘gold standard’ is shown in Table 4. The sensitivity of menstrual estimate of gestational age for very preterm births (<28 weeks) was high (98.9%) whereas the sensitivity declined for preterm birth <32 and 34–36 weeks. The sensitivity for post-term births was poor (51.6%) based on the menstrual estimate of gestational age.

Table 4.  Efficacy of menstrual estimate of gestational age in predicting adverse perinatal outcomes, using the clinical estimate of gestational age as the ‘gold standard’, US livebirths, 2002
Perinatal outcomes based on clinical estimate of gestational ageSensitivity (%)Specificity (%)Predictive value
Positive (%)Negative (%)
Preterm birth (weeks)
 <3780.395.262.498.0
 34–3673.295.973.298.1
 <3296.299.782.899.9
 <2898.9100.096.5100.0
Post-term birth ≥42 weeks51.693.89.099.4
Small-for-gestational age birth81.897.778.198.2
Large-for-gestational age birth91.598.282.899.2

Discussion

  1. Top of page
  2. Summary
  3. Introduction
  4. Methods
  5. Results
  6. Discussion
  7. Conclusions
  8. Acknowledgements
  9. References

Clinical estimate of gestation duration in singleton, livebirths in the US appears to closely approximate to gestational age based on menstrual dates at <28 and 37–41 weeks. Concordance between menstrual and CE of gestational age at 28–36 and at ≥42 weeks is, however, poor. Previous attempts to compare perinatal health outcomes based on clinical and menstrual estimate of gestational age have largely resulted in notable discrepancies. Specifically, Mustafa and David18 reported that the overall concordance between menstrual and CE of gestational age was 46%, and that the concordance increased to 78% for 1-week difference, and 87% for 2-week differences. They opined that ‘use of different methods of determining gestation in different years or geographic populations will result in artefactual differences in important indicators such as prematurity rate’. In addition, Alexander et al.19 suggested that the discrepant results from these two methods of estimating gestational age could lead to discontinuities in tracking population outcome measures.

The basis for assigning the CE of gestational age does not exist nor is it standardised in these data. It further remains unclear as to how clinicians assign the gestational age at birth. It is likely that the obstetric estimate of gestational age is the major contributor to determination of gestational age at birth,20 as most infant medical records from the nursery have gestational age assigned on admission based on obstetric information prior to a paediatric evaluation. While the Dubowitz and Ballard methods are two viable methods used by paediatricians to assess gestational age in newborns,21,22 these estimates are known to have wide variations in accuracy and are rarely used as primary gestational age determination. Ultrasound estimate of gestational age has been suggested to be biased by variations in fetal growth,23,24 but this effect is more important in the latter part of pregnancy. In addition, systematic variations in the CE of gestational age by hospital size and trimester of initiation of prenatal care, makes this method of gestational age assignment less favourable than that based on menstrual dating.19 On the contrary, while such systematic errors in menstrual estimate of gestational age are of less concern, the latter method is more affected by random errors.19

Is clinically estimated gestational age ‘better’ for population-based studies?

Gestational age based on menstrual dates clearly exhibits a bimodal distribution at very preterm and moderately preterm gestational ages, whereas such a phenomenon is not evident in the distribution of gestational age based on a CE. The second mode (also referred to as the ‘second bump’) in the gestational age distribution has often been attributed to errors in gestational age.13 Clearly, gestational age based on a CE appears more biologically plausible in comparison with the menstrual estimate. However, it has been argued that such an estimate (of clinical gestational age) would probably be biased as assignment of gestational age is made in conjunction with the observed birthweight. Nevertheless, clinical assignment of gestational age by obstetricians generally is independent of birthweight in practice, as the expected date of confinement is usually established during pregnancy prior to the infant's birth. In fact, it is very likely that the disappearance of the second mode in the CE of gestational age is the result of assigning a gestational age estimate that closely approximates to the birthweight of the newborn.

Strengths and limitations

This study is perhaps the most comprehensive, population-based study in the US that assessed the comparability of menstrual vs. CE of gestation duration, and contrasts temporal trends in indices of perinatal outcomes, including preterm and post-term birth, SGA and LGA births.

One of the limitations of the present study relates to generalisability of the findings. Data on the CE of gestational age is not reported by California, and so, approximately 12% to 13% of births have been excluded from the analyses. Nevertheless, there are no compelling data or reasons to believe that exclusion of California births would have introduced a bias in our findings.

Conclusions

  1. Top of page
  2. Summary
  3. Introduction
  4. Methods
  5. Results
  6. Discussion
  7. Conclusions
  8. Acknowledgements
  9. References

Notwithstanding these limitations, and in the absence of a true gold standard for pregnancy dating, it may be premature to conclude that one single approach to pregnancy dating is preferable to another. With that said, it may be a reasonable proposition to consider a hybrid combination of the menstrual and CE of gestational age as currently available on the NCHS Natality and Mortality data files, with our suggestion mirroring those previously proposed.18,19 Disregarding clinical estimate of gestational age is at best misleading, and at worst ignores the advantages of gestational age assessment in modern obstetrics.

Acknowledgements

  1. Top of page
  2. Summary
  3. Introduction
  4. Methods
  5. Results
  6. Discussion
  7. Conclusions
  8. Acknowledgements
  9. References

Dr Ananth is partially supported through a grant (R01 HD038902) awarded to him from the National Institutes of Health.

The author thanks Darios Getahun, K.S. Joseph, Michael S. Kramer, Yinka Oyelese, Joyce Martin, Morgan Peltier, John Smulian and Anthony Vintzileos for their generous effort in reviewing an earlier draft of the manuscript, and for offering valuable comments.

References

  1. Top of page
  2. Summary
  3. Introduction
  4. Methods
  5. Results
  6. Discussion
  7. Conclusions
  8. Acknowledgements
  9. References
  • 1
    Saito M, Yazawa K, Hashiguchi A, Kumasaka T, Nishi N, Kato K. Time of ovulation and prolonged pregnancy. American Journal of Obstetrics and Gynecology 1972; 112:3138.
  • 2
    Treloar AE, Behn BG, Cowan DW. Analysis of gestational interval. American Journal of Obstetrics and Gynecology 1967; 99:3445.
  • 3
    Gjessing HK, Skjaerven R, Wilcox AJ. Errors in gestational age: evidence of bleeding early in pregnancy. American Journal of Public Health 1999; 89:213218.
  • 4
    Hertz RH, Sokol RJ, Knoke JD, Rosen MG, Chik L, Hirsch VJ. Clinical estimation of gestational age: rules for avoiding preterm delivery. American Journal of Obstetrics and Gynecology 1978; 131:395402.
  • 5
    Blondel B, Morin I, Platt RW, Kramer MS, Usher R, Bréart G. Algorithms for combining menstrual and ultrasound estimates of gestational age: consequences for rates of preterm and postterm birth. BJOG 2002; 109:718720.
  • 6
    Emery ES, Eaton A, Grether JK, Nelson KB. Assessment of gestational age using birth certificate data compared with medical record data. Paediatric and Perinatal Epidemiology 1997; 11:313321.
  • 7
    Mongelli M, Gardosi J. Birth weight, prematurity and accuracy of gestational age. International Journal of Gynaecology and Obstetrics 1997; 56:251256.
  • 8
    Waller DK, Spears WD, Gu Y, Cunningham GC. Assessing number-specific error in the recall of onset of last menstrual period. Paediatric and Perinatal Epidemiology 2000; 14:263267.
  • 9
    David RJ. The quality and completeness of birthweight and gestational age data in computerized birth files. American Journal of Public Health 1980; 70:964973.
  • 10
    Alexander GR, Petersen DJ, Powell-Griner E, Tompkins ME. A comparison of gestational age reporting methods based on physician estimate and date of last normal menses from fetal death reports. American Journal of Public Health 1989; 79:600602.
  • 11
    Feresu SA, Gillespie BW, Sowers MF, Johnson TR, Welch K, Harlow SD. Improving the assessment of gestational age in a Zimbabwean population. International Journal of Gynaecology and Obstetrics 2002; 78:718.
  • 12
    Berg AT. Menstrual cycle length and the calculation of gestational age. American Journal of Epidemiology 1991; 133:585589.
  • 13
    Kramer MS, McLean FH, Boyd ME, Usher RH. The validity of gestational age estimation by menstrual dating in term, preterm, and postterm gestations. JAMA 1988; 260:33063308.
  • 14
    Savitz DA, Terry JW Jr, Dole N, Thorp JM Jr, Siega-Riz AM, Herring AH. Comparison of pregnancy dating by last menstrual period, ultrasound scanning, and their combination. American Journal of Obstetrics and Gynecology 2002; 187:16601666.
  • 15
    Taffel SM, Ventura SJ, Gay GA. Revised US certificate of birth – new opportunities for research on birth outcome. Birth 1989; 16:188193.
  • 16
    National Center for Health Statistics. 1995–2002 Linked Birth/infant Death Data Set. NCHS CD-ROM series 20, No 16–23. Hyattsville, MD: United States Department of Health and Human Services, Centers for Disease Control and Prevention. Access at http://www.cdc.gov/nchs/linked.htm/
  • 17
    Taffel S, Johnson D, Heuse R. A method of imputing length of gestation on birth certificates. Vital and Health Statistics, Series 2 1982; 93:111.
  • 18
    Mustafa G, David RJ. Comparative accuracy of clinical estimate versus menstrual gestational age in computerized birth certificates. Public Health Reports 2001; 116:1521.
  • 19
    Alexander GR, Tompkins ME, Petersen DJ, Hulsey TC, Mor J. Discordance between LMP-based and clinically estimated gestational age: implications for research, programs, and policy. Public Health Reports 1995; 110:395402.
  • 20
    Smulian JC, Ananth CV, Hanley ML, Knuppel RA, Donlen J, Kruse L. New Jersey's electronic birth certificate program: variations in data sources. American Journal of Public Health 2001; 91:814816.
  • 21
    Alexander GR, De Caunes F, Hulsey TC, Tompkins ME, Allen M. Validity of postnatal assessments of gestational age: a comparison of the method of Ballard et al. and early ultrasonography. American Journal of Obstetrics and Gynecology 1992; 166:891895.
  • 22
    Feresu SA, Harlow SD, Gillespie BW, Welch K, Johnson TR. Birthweight-adjusted Dubowitz methods: reducing misclassification of assessments of gestational age in a Zimbabwean population. Central African Journal of Medicine 2003; 49:4753.
  • 23
    Goldenberg RL, Davis RO, Cutter GR, Hoffman HJ, Brumfield CG, Foster JM. Prematurity, postdates, and growth retardation: the influence of use of ultrasonography on reported gestational age. American Journal of Obstetrics and Gynecology 1989; 160:462470.
  • 24
    Morin I, Morin L, Zhang X, Platt RW, Blondel B, Bréart G, et al. Determinants and consequences of discrepancies in menstrual and ultrasonographic gestational age estimates. BJOG 2005; 112:145152.