Caveats in the monitoring of fetal growth using ultrasound estimated fetal weight

Abstract

Introduction

Ultrasound estimated fetal weight is increasingly being used in the monitoring of fetal growth. Differences between estimated fetal weight formulae, curves and measurement methods could lead to significant differences in results. The aim of this study was to investigate the potential impact of these differences on estimated fetal weight and its use in monitoring fetal growth, both by modelling and by analysis of ultrasound scan data.

Methods

Four estimated fetal weight curves were compared in their original form and also normalised to term weight. Estimated fetal weight was calculated from 50th centiles of widely used charts of abdominal and head circumference and femur length and plotted on a widely used estimated fetal weight curve. Fetal measurement data were used to assess the impact of fetal proportions on estimated fetal weight error and on growth trajectory when different estimated fetal weight formulae are used.

Results

Estimated fetal weight curves differ significantly, but after normalisation there is closer agreement. Estimated fetal weight modelled using modern measurement methods differs from the widely used estimated fetal weight growth curve. Errors in estimated fetal weight are correlated with differences in fetal proportions and this can lead to significant changes in estimated fetal weight growth trajectory if different estimated fetal weight formulae are used.

Conclusions

Choice of measurement methods, estimated fetal weight formulae and growth curves have a significant effect on estimated fetal weight growth trajectories relative to normal ranges. It is important to understand these caveats when using estimated fetal weight to monitor fetal growth.

Keywords

Ultrasound estimated fetal weight fetal growth

Introduction

Ultrasound estimated fetal weight (EFW) is increasingly being used in the monitoring of fetal growth.¹,² The EFW growth curve developed by Hadlock et al.³ has been widely used, but is 30 years old. New EFW growth charts have been developed by INTERGROWTH-21st, the World Health Organisation (WHO) and the Fetal Medicine Foundation (FMF).²,⁴,⁵

Figure 1 shows all four curves. With the exception of the INTERGROWTH-21st chart, which falls approximately 9% lower, the 50th percentiles are, on average, within 2%. More importantly, however, the 10th percentiles differ, on average in the third trimester, by −11%, −2.1% and 4.7%, respectively from Hadlock et al.³ Even in the context of the known systematic and random errors in EFW, these differences could lead to large differences in sensitivity and specificity for detecting the small or large fetus, resulting in inappropriate obstetric management decisions and delivery choices for women.⁶

Figure 1.

EFW growth curves developed by Hadlock et al.,³ Kiserud et al. (WHO),⁴ Stirnemann et al. (INTERGROWTH)² and Nicolaides et al. (FMF).⁵ 50th, 90th and 10th centiles shown.

In the development of the customised gestation-related optimal weight (GROW) curves used in the majority of maternity units in the United Kingdom the growth curve of Hadlock et al.³ was converted to a percentage of term weight by gestational age curve, with the 40-week weight assigned 100% to produce a proportionality curve.⁷ A proportionality curve shows the percentage of term weight, rather than absolute EFW, plotted against gestational age. This allows the growth curve to be adapted by using the average term weight for a specific population or a term weight customised for the individual patient as the 100% value. Gardosi et al. showed 50th percentile growth trajectories for the Hadlock et al., WHO and INTERGROWTH-21st curves to be very similar after conversion to a proportionality curve.⁷ These curves were developed using different populations and methods and the similarity of growth trajectories after normalisation suggests the possibility that differences between the original curves are due to populations and methods.

Most EFW formulae are based on a combination of circumference and length measurements. A number of methods are available for circumference measurement, the most commonly used being ellipse fitting, calculation from orthogonal diameters and tracing. The ellipse fitting method is widely used in modern practice and involves generating and fitting an on-screen ellipse to the fetal outline; the circumference is automatically calculated. Calculation from orthogonal diameters requires measurement of antero-posterior and transverse diameters and calculation using the mean diameter in the equation for circumference of a circle; some ultrasound scanners perform the calculation in this manner, others by applying the ellipse equation to the diameters. The tracing method was widely used when first available and involves guiding the on-screen cursor around the fetal outline; the ultrasound scanner calculates the length of the resulting trace.

Hadlock et al.⁸,⁹ made circumference measurements either by tracing the perimeter or by derivation from two orthogonal diameters and these methods were regarded as interchangeable.¹⁰ Their EFW formula was developed using biparietal diameter (BPD), abdominal circumference (AC), head circumference (HC) and femur length (FL) measurements on 276 unselected, mainly middle class, Caucasian subjects⁹; their growth curve was developed from 392 healthy, mainly middle class, White subjects.³ The INTERGROWTH-21st EFW growth curve was developed using modern ultrasound equipment and measurement methods (ellipse fitting) on 4321 healthy subjects and employed methodologies that reduced bias; an EFW formula was developed using measurements of AC and HC on 2404 fetuses delivered within 14 days of their final scan.²,¹¹ The WHO EFW growth curve was developed using modern ultrasound scanners and measurement methods (ellipse fitting) on 1362 healthy subjects⁴; the EFW formula of Hadlock et al.⁹ using HC, AC and FL was used. The FMF EFW growth curve was developed using unspecified equipment and measurement methods on an unselected population of 95,579 patients⁵; the EFW formula of Hadlock et al.⁹ using HC, AC and FL was used.

Different measurement methods have been used in the development of EFW growth curves and this may have an impact on current practice. Hadlock et al. concluded that tracing and derivation methods were interchangeable.¹⁰ More recently evidence has been provided that the methods give different results, with traced circumferences being approximately 3% larger than derived circumferences.^12–14 Currently, derived or ellipse fitted circumferences are widely recommended,¹¹,¹⁵,¹⁶ and some authors suggest that these two methods may be used interchangeably.¹⁷,¹⁸ This results in a potential caveat in the use of EFW for monitoring fetal growth, that current measurement methods may result in a systematic underestimation of fetal weight with respect to growth charts developed using traced circumferences or a mixture of methods, such as that of Hadlock et al.³ The potential consequences are inappropriate management due to incorrect diagnosis of fetal growth restriction or underestimation of EFW of the larger fetus.

EFW formulae are available that incorporate all or a selection of regularly performed fetal measurements, i.e. BPD, AC, HC and FL. It is not uncommon to use a formula incorporating AC, HC and FL throughout most of pregnancy and then change to a formula without HC if an accurate head measurement is not obtained in late pregnancy. It is well known that fetal proportions vary between subjects; the ratios of FL to AC and HC to AC have been reported and have been used in attempts to differentiate between different patterns of fetal growth restriction.¹⁹,²⁰ There may, therefore, be a further caveat if formulae with and without HC give different results in some circumstances, which may result in misdiagnosis of fetal growth restriction due to an artefactual change in growth trajectory.

The aim of this study was to investigate the potential impact of these caveats on EFW and its use in monitoring fetal growth, both by modelling and by analysis of ultrasound scan data.

Methods

Since the publication of the analysis of Gardosi et al.⁷ showing similarities between the 50th percentiles of EFW proportionality curves from three different sources a further EFW curve has been published by the FMF.⁵ We generated proportionality curves for all four EFW curves, including the 10th and 90th percentiles, to assess agreement.

In order to assess the potential systematic error arising from the use of circumference measurement methods differing from those used in development of EFW growth charts, mean values from well-known derived and traced growth charts¹²,¹³,²¹ were used to calculate EFW using the AC, HC and FL method of Hadlock et al.⁹ This method is widely used and its accuracy has not been surpassed by more recent formulae.²²,²³ Results were plotted on the EFW growth chart of Hadlock et al.³

Data from a previous study were used to assess the effect of fetal proportions on EFW accuracy.²⁴ The measurements had been recorded from clinically required examinations and had been retained without patient identification. No ethical approval was required as no patient identifiable data were used. Subjects were scanned within 10 days of delivery (mean: four days) and birthweight was recorded at delivery. EFW values were corrected for time to delivery using the proportionality curve of Mongelli and Gardosi.²⁵ Errors in EFW with respect to birthweight were plotted against FL/AC and HC/AC, looking for a relation between errors and fetal proportions.

Finally, EFW growth was modelled assuming average growth velocities in AC, HC and FL, using an HC/AC ratio two standard deviations²⁶ (approximately 12%) above average. Measurements at three week intervals from 28 to 37 weeks gestation were modelled, with EFW being calculated from AC, HC and FL for the first three measurements and from AC and FL for the final measurement to assess the impact of changing formula.

Results

Figure 2 shows the Hadlock et al.,³ WHO,⁴ INTERGROWTH² and FMF⁵ EFW growth curves normalised to a percentage of term weight by gestational age curve, with the 40-week weight assigned 100% to produce a proportionality curve.

Figure 2.

EFW growth curves developed by Hadlock et al.,³ Kiserud et al. (WHO),⁴ Stirnemann et al. (INTERGROWTH)² and Nicolaides et al. (FMF)⁵ normalised to a percentage of term weight by gestational age curve, with the 40-week weight assigned 100% to produce a proportionality curve. 50th, 90th and 10th centiles shown.

Figure 3 shows the EFW growth chart of Hadlock et al.³ with EFW growth trajectories calculated from 50th percentile values of the traced and derived AC and HC curves, together with the FL curves, of Chitty et al.¹²,¹³,²¹ The difference between EFW from traced and derived circumference measurements grows from 5% at 20 weeks to 7% at 40 weeks. Table 1 and Figure 4 show the modelled effect on growth trajectory of this 5–7% negative systematic error on plotting EFW for a fetus growing just above the 10th percentile.

Figure 3.

EFW growth trajectories calculated from 50th percentile values of traced and derived AC and HC curves and the 50th percentile of FL (Chitty et al.¹²,¹³,²¹), plotted on the growth curve of Hadlock et al.³

Table 1.

The modelled effect of the difference between estimated fetal weight (EFW) estimated from traced and derived circumferences for a fetus growing just above the 10th percentile.

Gestational age (weeks)	10th Percentile	Traced EFW (g)	Derived EFW (g)	Difference (%)
28	1040	1050	988	5.9
31	1505	1520	1427	6.1
34	2043	2063	1933	6.3
37	2602	2628	2458	6.5
40	3110	3141	2932	6.6

Figure 4.

The modelled effect of the difference between EFW estimated from traced and derived circumferences for a fetus growing just above the 10th percentile.

Figures 5 and 6 show the error in EFW compared to birthweight plotted against HC/AC and FL/AC respectively for 250 subjects scanned within 10 days of delivery (mean: four days). EFW values were corrected for time to delivery using the proportionality curve of Mongelli and Gardosi.²⁵ There is a weak but statistically significant correlation between the percentage error and the ratios (r = −0.22, p < 0.001; r = −0.25, p < 0.001, respectively).

Figure 5.

The error in EFW compared to birthweight plotted against HC/AC.

Figure 6.

The error in EFW compared to birthweight plotted against FL/AC.

Table 2 and Figure 7 show the result of modelling EFW growth assuming average growth velocities in AC, HC and FL, using an HC/AC ratio two standard deviations (approximately 12%) above average, with EFW calculated from AC, HC and FL for the first three measurements and from AC and FL for the final measurement. This resulted in an apparent fall in growth trajectory, which could be interpreted as fetal growth restriction.

Table 2.

Modelling estimated fetal weight (EFW) growth assuming average growth velocities in abdominal circumference (AC), head circumference (HC) and femur length (FL), using an HC/AC ratio two standard deviations above average, with EFW calculated from AC, HC and FL and from AC and FL.

Gestational age (weeks)	AC (mm)	HC (mm)	FL (mm)	HC/AC	EFW (g) (AC, HC, FL)	EFW (g) (AC, FL)
28	231	292	53	1.27	1215	1140
31	261	323	59	1.24	1766	1628
34	290	350	65	1.21	2407	2182
37	315	371	70	1.18	3077	2757

Figure 7.

Modelling EFW growth assuming average growth velocities in AC, HC and FL, using an HC/AC ratio two standard deviations above average, with EFW calculated from AC, HC and FL for the first three measurements and from AC and FL for the final measurement.

Discussion

It is well known that EFW is subject to systematic and random errors that vary between formulae, observers and centres.²⁷ Wright et al. concluded that errors in individual fetal measurement cause substantial errors in EFW, resulting in misclassification of small and large-for-gestational ages fetuses.²⁸ In interpreting EFW measurements in the context of fetal growth, it is important to understand and acknowledge these errors since many obstetric decisions such as timing and mode of delivery are based on EFW.

Figure 1 shows that EFW centiles on the INTERGROWTH-21st chart are lower than those on the Hadlock et al., WHO and FMF charts.^2–5 The consequences of this were highlighted by Francis et al. who showed that in a normal population the INTERGROWTH-21st chart identified only 4.4% of babies as small-for-gestational-age (below 10th percentile) and 20.6% of babies as large-for-gestational-age (above 90th percentile).²⁹

Normalisation of EFW growth curves to a percentage of term weight by gestational age curve, with the 40-week weight assigned 100%, as described by Gardosi et al.,⁷ showed close agreement between Hadlock et al.,³ INTERGROWTH² and WHO⁴ curves (Figure 2). The maximum differences between any two of these three curves in the third trimester were 1.8% for the 50th centile, 2.7% for the 10th percentile and 5.6% for the 90th percentile. The maximum differences between the normalised FMF⁵ curve and any other were 5.1% for the 50th centile, 7.9% for the 10th percentile and 3.7% for the 90th percentile. These differences are due to the reduction in slope of the FMF curve relative to the others after 36 weeks, resulting in higher percentages throughout pregnancy; the reason for this difference in slope is unknown.

The close agreement in Figure 2, after normalisation, between the former three curves suggests that normalisation to a customised term birthweight or, for those who prefer population-based charts, to a population mean term birthweight removes the systematic differences between curves generated by different populations and different EFW formulae used in their production. There is already a body of evidence for customised growth charts.^1,7,29–34 The key factors in the use of charts then become the absolute accuracy of measurement methods and formulae and the interpretation of size and growth patterns.

The use of derived or ellipse fitted circumference methods, rather than the mixture of derived and traced methods used by Hadlock et al.,⁹ could lead to systematic underestimation of fetal weight by up to 7% as shown in Figure 3. Further, derived and ellipse fitted measurements are not interchangeable, as there are significant systematic and random differences between them and there may be inter-observer differences in the perception of the best fit of an ellipse.³⁵,³⁶ It would be a retrograde step to return to tracing circumferences as it is both more time-consuming and subject to greater variability than the derived or ellipse fitting methods. The possibility of systematic underestimation by current methods is the subject of a further study.

The relationship between error in EFW and FL/AC and HC/AC ratios could lead to systematic errors in EFW in fetuses with larger or smaller ratios. The practical consequences are that the EFW growth trajectories in these cases will follow a higher or lower centile than the true trajectory with the potential consequence that measurements may fall artefactually above or below a threshold for intervention. A new EFW formula is required that takes account of the variability of fetal proportions.

EFW formulae are not interchangeable, even when derived from the same patient cohort. Figure 7 shows that changing formula during growth monitoring can produce an artefactual change in growth rate. It is good practice in these circumstances to recalculate all EFWs using the same formula.

Conclusions

This study provides support for the concept of customisation or normalisation of EFW growth curves. The image planes for fetal measurement are widely understood and accepted. However, standardisation of circumference measurement methods is required; the derived method may be the most reproducible as the measurement points are objectively defined. A new EFW formula that is robust in the presence of variability of fetal proportions is required. Improvements to current methods and practice are essential in supporting evidence-based decision-making in obstetric management.

Supplemental Material

sj-pdf-1-ult-10.1177_1742271X20954508 - Supplemental material for Caveats in the monitoring of fetal growth using ultrasound estimated fetal weight

Supplemental material, sj-pdf-1-ult-10.1177_1742271X20954508 for Caveats in the monitoring of fetal growth using ultrasound estimated fetal weight by Nicholas John Dudley and Helen Varley in Ultrasound

Footnotes

Contributors

NJD and HV conceived the study. NJD performed modelling and data analysis and drafted the manuscript. NJD and HV reviewed and revised the manuscript and approved the final version.

Declaration of Conflicting Interests

The author(s) declared no potential conflicts of interest with respect to the research, authorship, and/or publication of this article.

Ethics approval

Not required. Only one element of the study used patient data. These data had been collected from clinically required examinations 25 years ago and had been held since then with no patient identification. No ethical approval was required as no patient identifiable data were used, as per The Declaration of Helsinki.

Funding

This research received no specific grant from any funding agency in the public, commercial, or not-for-profit sectors.

Guarantor

NJD

Acknowledgements

None

ORCID iD

Nicholas John Dudley

References

Williams

Turner

Butler

, et al. Fetal growth surveillance – current guidelines, practices and challenges. Ultrasound 2018; 26: 69–79.

Stirnemann

Villar

Salomon , et al., for The International Fetal and Newborn Growth Consortium for The 21st Century (INTERGROWTH-21st). International estimated fetal weight standards of the Intergrowth-21st project. Ultrasound Obstet Gynecol 2017; 49: 478–486.

Hadlock

Harrist

Martinez-Poyer

In utero analysis of fetal growth: a sonographic weight standard.

Radiology 1991; 181:129–133.

Kiserud

Piaggio

Carroli

, et al. The World Health Organization fetal growth charts: a multinational longitudinal study of ultrasound biometric measurements and estimated fetal weight. PLoS Med 2017; 14: e1002220.

Nicolaides

Wright

Syngelaki

, et al. Fetal Medicine Foundation fetal and neonatal population weight charts. Ultrasound Obstet Gynecol 2018; 52: 44–51.

Dudley

NJ.

A review of ultrasound fetal weight estimation in the early prediction of low birthweight. Ultrasound 2013; 21: 181–186.

Gardosi

Francis

Turner

, et al. Customised growth charts: rationale, validation and clinical benefits. Am J Obstet Gynecol 2018; S609–S618.

Hadlock

Harrist

Carpenter

, et al. Sonographic estimation of fetal weight. The value of femur length in addition to head and abdomen measurement. Radiology 1984; 150: 535–540.

Hadlock

Harrist

Sharman

, et al. Estimation of fetal weight with the use of head, body and femur measurements: a prospective study. Am J Obstet Gynecol 1985; 151: 333–337.

10.

Hadlock

Kent

Loyd

, et al. An evaluation of two methods for measuring fetal head and body circumferences. J Ultrasound Med 1982; 1: 359–360.

11.

Papageorghiou

Ohuma

Altman

, et al.; International Fetal and Newborn Growth Consortium for the 21st Century (INTERGROWTH-21st). International standards for fetal growth based on serial ultrasound measurements: the Fetal Growth Longitudinal Study of the INTERGROWTH-21st Project. Lancet 2014; 384: 869–879.

12.

Chitty

Altman

Henderson

, et al. Charts of fetal size: 2. Head measurements. Br J Obstet Gynaecol 1994; 101: 35–43.

13.

Chitty

Altman

Henderson

, et al. Charts of fetal size: 3. Abdominal measurements. Br J Obstet Gynaecol 1994; 101: 125–131.

14.

Dudley

Chapman

The importance of quality management in fetal measurement.

Ultrasound Obstet Gynecol 2002; 19: 190–196.

15.

Loughna

Chitty

Evans

, et al. Fetal size and dating: charts recommended for clinical obstetric practice. Ultrasound 2009; 17: 161–167.

16.

McCarthy

Walker

SP.

International fetal growth standards: one size fits all. Lancet 2014; 384: 835–836.

17.

Sarris

Ioannou

Dighe

, et al.; International Fetal and Newborn Growth Consortium for the 21st Century (INTERGROWTH-21st). Standardisation of fetal ultrasound biometry measurements: improving the quality and consistency of measurements. Ultrasound Obstet Gynecol 2011; 38: 681–687.

18.

Salomon

Alfirevic

Berghella

, et al.; ISUOG Clinical Standards Committee. Practice guidelines for performance of the routine mid-trimester fetal ultrasound scan. Ultrasound Obstet Gynecol 2011; 37: 116–126.

19.

Campbell

Thoms

Ultrasound measurement of the fetal head to abdomen circumference ratio in the assessment of growth retardation.

Br J Obstet Gynaecol 1977; 84: 165–174.

20.

Hadlock

Deter

Harrist

, et al. A date-independent predictor of intrauterine growth retardation: femur length/abdominal circumference ratio. Am J Roentgenol 1983; 141: 979–984.

21.

Chitty

Altman

Henderson

, et al. Charts of fetal size: 4. Femur length. Br J Obstet Gynaecol 1994; 101: 132–135.

22.

Hammami

Mazer Zumaeta

Syngelaki

, et al. Ultrasonographic estimation of fetal weight: development of new model and assessment of previous models. Ultrasound Obstet Gynecol 2018; 52: 35–43.

23.

Kong

WWK.

Comparison of the accuracy of INTERGROWTH-21 formula with other ultrasound formulae in fetal weight estimation. Taiwan J Obstet Gynecol 2019; 58: 273–277.

24.

Dudley

NJ.

Selection of appropriate ultrasound methods for the estimation of fetal weight.

Br J Radiol 1995; 68: 385–388.

25.

Mongelli

Gardosi

Gestation-adjusted projection of estimated fetal weight. Acta Obstet Gynecol Scand 1996; 75: 28–31.

26.

Snijders

RJM

Nicolaides

KH.

Fetal biometry at 14–40 weeks’ gestation.

Ultrasound Obstet Gynecol 1994; 4: 34–48.

27.

Dudley

NJ.

A systematic review of the ultrasound estimation of fetal weight.

Ultrasound Obstet Gynecol 2005; 25: 80–89.

28.

Wright

Smith

, et al. Impact of biometric measurement error on identification of small- and large-for-gestational-age fetuses. Ultrasound Obstet Gynecol 2020; 55: 170–176.

29.

Francis

Hugh

Gardosi

Customised vs INTERGROWTH-21st standards for the assessment of birthweight and stillbirth risk at term. Am J Obstet Gynecol 2018; S692–S699.

30.

Clausson

Gardosi

Francis

, et al. Perinatal outcome in SGA births defined by customized versus population-based birthweight standards. Br J Obstet Gynecol 2001; 108: 830–834.

31.

Ego

Subtil

Grange

, et al. Customized versus population-based birth weight standards for identifying growth restricted infants: a French multicenter study. Am J Obstet Gynecol 2006; 194: 1042–1049.

32.

Larkin

Speer

Simhan

HN.

A customized standard of large size for gestational age to predict intrapartum morbidity.

Am J Obstet Gynecol 2011; 204: 499.e1–10.

33.

Cha

H-H

Kim

J-Y

Choi

S-J

, et al. Can a customized standard for large for gestational age identify women at risk of operative delivery and shoulder dystocia? J Perinat Med 2012; 40: 483–488.

34.

González-González

González-Dávila

Cabrera

, et al. Application of customized birth weight curves in the assessment of perinatal outcomes in infants of diabetic mothers. Fetal Diagn Ther 2015; 37: 117–122.

35.

Schmidt

Temerinac

Bildstein

, et al. Finding the most accurate method to measure head circumference for fetal weight estimation. Eur J Obstet Gynecol Reprod Biol 2014; 178: 153–156.

36.

Dudley

NJ.

Are ultrasound foetal circumference measurement methods interchangeable?

Ultrasound 2019; 27: 176–182.

Supplementary Material

Please find the following supplemental material available below.

For Open Access articles published under a Creative Commons License, all supplemental material carries the same license as the article it is associated with.

For non-Open Access articles published, all supplemental material carries a non-exclusive license, and permission requests for re-use of supplemental material or any part of supplemental material shall be sent directly to the copyright owner as specified in the copyright notice associated with the article.

0.00 MB

0.09 MB