Hadlock Estimated Fetal Weight Calculator

All four Hadlock EFW regressions, computed exactly from their published coefficients, with their live disagreement and the honest accuracy: a 95% interval wider than 14% of birth weight, which is why an EFW is not a weight.

Last updated 8 Oct 2026

Hadlock estimated fetal weight

4 published regressions

Which Hadlock equationThese are four separate published regressions, not one equation with optional terms, and they give different weights. Enter all four measurements whatever you choose: the selected equation uses the ones it needs and the figure below shows how far the four disagree on your numbers. Record which equation a reported EFW came from — reporting software does not always say.

Biparietal diameter (mm)Used by equations 2 and 4 only. Measured outer-to-inner in most conventions; which convention your unit uses shifts the figure by a millimetre or two.

Head circumference (mm)Used by equations 3 and 4 only. An ellipse fitted to the outer skull margin, not a derived figure from the biparietal and occipitofrontal diameters unless your protocol says so.

Abdominal circumference (mm)Used by all four. This is the measurement that carries the weight estimate and also the one with the largest operator variability, measured at the level of the stomach and the umbilical portion of the left portal vein. An abdominal circumference below the 10th centile is itself one of the consensus growth-restriction criteria, independently of the weight.

Femur length (mm)Used by all four. The ossified diaphysis only, excluding the distal epiphysis. The 85 mm maximum here is not arbitrary: every one of the four equations has a negative abdominal-by-femur interaction term, and above a femur length of about 134 mm the HC, AC and FL form would start DECREASING in abdominal circumference. No fetus reaches that, and the bound keeps the equation inside the region where it behaves.

2307g estimated fetal weightExample

BPD 85 mm, HC 310 mm, AC 300 mm, FL 65 mm, HC and AC and FL equation

The four regressions

log₁₀(EFW in grams) = a polynomial in the biometry in CENTIMETRES
HC, AC, FL form: log₁₀ W = 1.326 − 0.00326·AC·FL + 0.0107·HC + 0.0438·AC + 0.158·FL

centimetres, not millimetres: every one of Hadlock’s regressions is published with the biometry in centimetres. The fields above take millimetres, because that is what a report prints, and divide by ten before the arithmetic. An implementation that feeds millimetres straight in is wrong by orders of magnitude, which at least fails loudly
there are four, and they differ: AC with FL; BPD with AC and FL; HC with AC and FL; and all four together. They are separate published regressions with separate coefficients. On the measurements above they span 27 g, which is 1.2% of the estimate — much less than the error any of them carries against the actual birth weight
the negative interaction term: all four subtract a multiple of AC × FL. That is what makes them regressions rather than volume models, and it is also why the equations are only monotonic in each measurement over the physiological range. For the HC, AC, FL form the derivative with respect to AC is 0.0438 − 0.00326·FL, which turns negative at a femur length of 13.4 cm
plus or minus 14 to 15%: Dudley’s systematic review found 95% confidence intervals exceeding 14% of birth weight for every method; GLOWM gives about plus or minus 15%. About 80% of estimates land within 10% of birth weight and about 50% within 5%. The error is the headline fact about an EFW, not a footnote to it
weight is not a centile: this page gives grams. A centile needs a reference distribution for the gestational age, which is a table rather than a formula. The z-score and centile page does that from a mean and SD you supply

Worked example

BPD 85 mm, HC 310 mm, AC 300 mm, FL 65 mm, HC and AC and FL equation
Convert to centimetres: BPD 8.5, HC 31.0, AC 30.0, FL 6.5
AC × FL = 30.0 × 6.5 = 195.0
log₁₀ W = 1.326 − 0.00326×195.0 + 0.0107×31.0 + 0.0438×30.0 + 0.158×6.5
= 1.326 − 0.6357 + 0.3317 + 1.3140 + 1.0270 = 3.3630
EFW = 10^3.3630 = 2,307 g
The other three equations on the same biometry give 2,333 g (AC and FL), 2,326 g (BPD, AC, FL) and 2,311 g (all four) — a spread of 27 g, or 1.2%
Now the figure that matters: a 95% interval of plus or minus 15% on 2,307 g runs from 1,961 g to 2,653 g. The disagreement between the four equations is 1.2%; the disagreement between any of them and the baby is around fifteen times that. Choosing the equation is not the problem

The four Hadlock regressions, with the weight each gives at BPD 85 / HC 310 / AC 300 / FL 65 mm

Coefficients taken from a statement of all four and corroborated term for term against LOINC’s coded versions of the HC+AC+FL and four-parameter forms and against GLOWM’s statement of the HC+AC+FL form. Measurements in centimetres, weight in grams. The four span 27 g here, so which equation a report used is worth knowing and is rarely stated.
Equation	Measurements	Published form	EFW (g) at this biometry
1	AC, FL	log₁₀ W = 1.304 + 0.05281·AC + 0.1938·FL − 0.004·AC·FL	2,333
2	BPD, AC, FL	log₁₀ W = 1.335 − 0.0034·AC·FL + 0.0316·BPD + 0.0457·AC + 0.1623·FL	2,326
3	HC, AC, FL	log₁₀ W = 1.326 − 0.00326·AC·FL + 0.0107·HC + 0.0438·AC + 0.158·FL	2,307
4	BPD, HC, AC, FL	log₁₀ W = 1.3596 − 0.00386·AC·FL + 0.0064·HC + 0.00061·BPD·AC + 0.0424·AC + 0.174·FL	2,311

How accurate is an estimated fetal weight?

The last row against the first four is the whole point. The four Hadlock equations disagree with each other by about one percent at this biometry and with the actual birth weight by fifteen. A reader shown a bare figure of 2,307 g needs the interval, not a fifth decimal place.
Finding	Figure	Source
95% confidence interval, every method examined	exceeds 14% of birth weight	Dudley 2005, via Lappen and Myers 2017
Estimates within 10% of birth weight	about 80%	Lee et al., via Lappen and Myers 2017
Estimates within 5% of birth weight	about 50%	Lee et al., via Lappen and Myers 2017
Accuracy of most published equations	about plus or minus 15%	GLOWM biometry chapter
Direction of error	overestimation is the commoner direction	Milner and Arezina 2018
Spread between the four Hadlock equations here	27 g, or 1.2%	this calculator

An exact equation for an inexact quantity

Hadlock’s weight regressions are closed-form log-polynomials in the biometry, so they can be computed exactly, and this page does. That is worth separating from the thing they are often confused with: Hadlock also published weight-for-gestational-age CENTILES, and those are a table, not a formula. The equations here turn measurements into grams. Turning grams into a centile needs a reference distribution for the gestational age, and no such distribution — Hadlock’s, INTERGROWTH-21st’s, the WHO’s or a customised one — is a formula this engine could hold.

There are four regressions and they are genuinely separate: abdominal circumference with femur length; biparietal diameter added; head circumference added instead; and all four measurements together. Each has its own coefficients and each includes a negative abdominal-by-femur interaction term, which is what makes them statistical fits rather than geometric volume models. On the default biometry above they span about 1.2% of the estimate. That is small, and it is the smaller of the two numbers a reader needs. The larger one is the error against the baby: Dudley’s systematic review found 95% confidence intervals exceeding 14% of birth weight for every method, GLOWM gives about plus or minus 15%, and roughly one estimate in five is more than 10% out. A reported weight of 2,307 g is consistent with a baby anywhere from about 1,961 to 2,653 g.

That interval is why an estimated fetal weight must not be read as a weight, and the error matters most exactly where the estimate is used most — at the thresholds. Lappen and Myers’ paper is about precisely this: systematic error in EFW leads to the underestimation of growth restriction, because an overestimated weight lifts a small fetus above a centile threshold, and overestimation is the commoner direction. If the number is to be compared with a threshold, compare the interval with it too, and read the centile arithmetic and the consensus growth-restriction criteria with that in mind. This page computes a number, or applies published criteria, and names the body that publishes them. It renders no clinical decision: nothing here says whether this fetus is growth restricted, whether a scan should be repeated, whether labour should be induced or when a pregnancy should be delivered. An obstetric measurement is read alongside the whole clinical picture — history, serial trend, the other biometry — and never instead of it. It supports a clinician’s judgement rather than replacing it.

Frequently asked questions

Which Hadlock equation should I use?

Whichever your unit’s reporting software uses, and record it. The HC, AC and FL form is the one LOINC codes as the headline estimate, so it is the default here. Milner and Arezina’s systematic review found ‘Hadlock A’ most accurate with the lowest random error but does not say which of the four that is, so this page does not claim one is best. On typical third-trimester biometry they differ by about one percent.

Do the measurements go in as millimetres or centimetres?

The published equations take centimetres. The fields here take millimetres, because that is what an ultrasound report prints, and divide by ten internally. Feeding millimetres into the published form directly gives an answer wrong by many orders of magnitude.

How wrong can an estimated fetal weight be?

Dudley’s systematic review found 95% confidence intervals exceeding 14% of birth weight for every method examined, and GLOWM puts most equations at about plus or minus 15%. About 80% of estimates fall within 10% of birth weight and about half within 5%. Overestimation is the commoner direction.

Why does this page not give a centile?

Because a centile requires a published reference distribution for the gestational age, which is a table or a skewed parametric model rather than a formula, and fitting a curve to one would misclassify fetuses at the extremes — the only place the answer changes anything. Enter your standard’s own mean and standard deviation on the z-score and centile page instead.

Does the choice of growth standard change the centile?

Substantially. In 117,027 UK singleton pregnancies, 7.2% were small for gestational age by GROW and 2.4% by INTERGROWTH-21st, with two-thirds of the GROW-identified cases missed by INTERGROWTH-21st. The same fetus with the same weight crosses the threshold or does not depending on the standard, so always record which one was used.

Related calculators

References

Hadlock FP, Harrist RB, Sharman RS, Deter RL, Park SK. Estimation of fetal weight with the use of head, body, and femur measurements – a prospective study. Am J Obstet Gynecol. 1985;151(3):333-7.
Regenstrief Institute. LOINC 11746-5 (EFW from AC & FL & HC, Hadlock 1985) and 11732-5 (EFW from AC & BPD & FL & HC, Hadlock 1985). Indianapolis: Regenstrief Institute.
Dudley NJ. A systematic review of the ultrasound estimation of fetal weight. Ultrasound Obstet Gynecol. 2005;25(1):80-9.
Lappen JR, Myers SA. The systematic error in the estimation of fetal weight and the underestimation of fetal growth restriction. Am J Obstet Gynecol. 2017;216(5):477-83.
Milner J, Arezina J. The accuracy of ultrasound estimation of fetal weight in comparison to birth weight: a systematic review. Ultrasound. 2018;26(1):32-41.

Not medical advice. For healthcare professionals and education. Reference intervals vary by laboratory and assay — always use your own laboratory's. Never base a dose or a treatment decision on this page alone. Full disclaimer at calcengines.com/disclaimer/