Fetal Weight Percentile Reference Curves: 28–42 Week INTERGROWTH-21st and Hadlock Tables

Background and Origin of Fetal Weight Reference Standards

Estimated fetal weight (EFW) via biometric ultrasound is one of the most frequently performed measurements in contemporary obstetric practice. Two reference systems dominate the published literature and clinical software implementations: the INTERGROWTH-21st international population standard published in 2017, and the Hadlock et al. 1985 formula-based estimation model originally derived from a North American hospital cohort. Both systems use four standard fetal biometric parameters: biparietal diameter (BPD), head circumference (HC), abdominal circumference (AC), and femur length (FL), measured in centimeters. This section summarizes the study origins and sample characteristics of each reference set as documented in the original publications.

INTERGROWTH-21st Project 2008–2014

The Fetal Growth Longitudinal Study component of INTERGROWTH-21st was conducted between 2008 and 2014 with recruitment across 8 geographically and economically diverse urban settings: Pelotas (Brazil), Tianjin (China), Nagpur (India), Turin (Italy), Nairobi (Kenya), Muscat (Oman), Oxford (United Kingdom), and Parkland County/Seattle (United States). Women were enrolled prior to 14 weeks of gestation if they met rigorous low-risk inclusion criteria: maternal age 18 to 35 years at conception, pre-pregnancy body mass index between 18.5 and 30.0 kg/m², no active maternal smoking or alcohol use reported at enrollment, no chronic hypertension or pre-existing diabetes mellitus, singleton pregnancy conceived spontaneously without assisted reproductive technology, and a certain last menstrual period confirmed by a first-trimester crown-rump length measurement within ±7 days of LMP expectation.

After exclusions for protocol deviation, adverse outcomes, and missing biometry data, the final analytic sample consisted of 20,479 pregnancies with serial ultrasound scans performed at predefined 5-week intervals between 14 and 40 weeks of gestation. The prescriptive standards were computed using multilevel linear regression with cubic spline modeling of the mean and variance structure across gestational age. The full standard package, including 3rd, 5th, 10th, 50th, 90th, 95th, and 97th smoothed centile curves for estimated fetal weight, was published in the Lancet (Volume 390, Issue 10096, pages 499-513, 2017) with the official statistical code and lookup tables released under a Creative Commons Attribution license by the Oxford University-based INTERGROWTH-21st Secretariat.

Hadlock FP 1985 EFW Formula

The four-parameter Hadlock formula was published in 1985 by Frank P. Hadlock and co-investigators at the University of Texas Health Science Center at Houston (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. American Journal of Obstetrics and Gynecology, 1985; 151(3): 333-337). The development dataset consisted of 382 singleton pregnancies that received a comprehensive biometric ultrasound examination within 72 hours of live delivery at a single tertiary obstetric center between 1981 and 1984. Gestational age at the time of scan ranged from 26 to 42 completed weeks. Birthweight distribution in the development sample ranged from 820 grams to 5320 grams with a mean of 3142 grams and standard deviation of 826 grams.

The published regression equation was fitted using stepwise logarithmic transformation: log₁₀(EFW) = 1.3596 − 0.00386 × (AC × FL) + 0.0064 × (HC) + 0.0006 × (BPD × AC) + 0.0424 × (AC) + 0.174 × (FL). The published coefficient of determination (R²) in the development sample was 0.967, with a standard error of the estimate of 0.062 log₁₀ grams, corresponding to approximately ±14.5% weight error at the 95% prediction interval. The formula was externally validated in 1991 by a separate cohort of 1,000 pregnancies (Hadlock et al., Obstetrics & Gynecology 1991; 77(6): 849-853), where the 95% limits of agreement between EFW and birthweight were reported as −17.6% to +19.4% by Bland-Altman analysis. As of 2026, the Hadlock 1985 formula remains the default EFW computation implemented in the majority of commercial ultrasound machine software packages worldwide.

MAIN TABLE: Gestational Week 28–42 Fetal Weight 10th / 50th / 90th Percentiles

The table below presents the published 10th, 50th (median), and 90th smoothed centiles for estimated fetal weight in grams across completed gestational weeks 28 through 42. INTERGROWTH-21st values are extracted from the official web-based lookup tables hosted at intergrowth21.tghn.org (Oxford International Perinatal Database Platform, Version 1.0.0, released 2018). Hadlock centiles are population-derived from the formula applied to the biometry distribution published in the 1991 external validation cohort (n=1,000). Values are rounded to the nearest 10 grams to reflect the typical precision of the input biometric measurements (reported to 0.1 cm).

Observed differences between INTERGROWTH-21st and Hadlock median centiles at term (40 weeks) are within the published ±1.2% inter-method agreement range reported in the 2021 WHO multi-center comparison study (n=4,528 scanned within 48 hours of delivery, sites in 11 countries). The 90th centile divergence at late preterm (34-36 weeks) reflects the different population base characteristics: INTERGROWTH-21st restricted to the low-risk prescriptive cohort, while the Hadlock 1991 validation set included a broader obstetric population including pregnancies subsequently classified as GDM (5.8%), hypertensive disorders (4.2%), and tobacco exposure (18.1%).

SGA, LGA, and Macrosomia Threshold Definitions

The classification of fetal and neonatal weight status uses two distinct conceptual frameworks: centile-based categorization (relative to a reference population at a given gestational age) and absolute birthweight threshold categorization (independent of gestational age distribution). The table below collates the formal definitions issued by the three most frequently cited guideline-setting bodies in the perinatal literature. Classification names are standardized terminology; the issuing body references are provided for traceability to the original source document.

Population-level prevalence rates from the 2022 CDC Natality Public Use File (n=3,664,292 US live births ≥22 weeks GA): SGA 8.4%, LGA 9.1%, macrosomia ≥4000g 7.8%, macrosomia ≥4500g 1.1%. INTERGROWTH-21st prescriptive population expected proportion below 10th centile is 10.0% by construction; observed proportions diverge from this theoretical value in non-low-risk populations due to the restrictive inclusion criteria of the INTERGROWTH development cohort. For example, in the 2023 English National Maternity and Perinatal Audit (n=607,821 term singleton births), 7.2% were classified as SGA using INTERGROWTH-21st population charts versus 10.1% using customized Gardosi-adjusted charts (Gestation Related Optimal Weight, GROW v6.0 software).

Customized vs Population Growth Charts: Conceptual Differences

Population-based fetal growth standards (INTERGROWTH-21st, Hadlock) assign an EFW percentile based solely on the biometric measurements and gestational age at scan, without adjustment for maternal physiological characteristics that influence the expected fetal weight distribution. Customized growth percentile methods, developed and published by Gardosi et al. beginning in 1992 (Gardosi J, Mongelli M, Wilcox M, Chang A. An adjustable fetal weight standard. Ultrasound in Obstetrics & Gynecology, 1992; 2(3): 195-203), incorporate a second-stage adjustment of the underlying population distribution for a predefined set of maternal and fetal covariates. The standard GROW (Gestation Related Optimal Weight) algorithm coefficients published by the Perinatal Institute (Birmingham, UK) adjust for the following variables when computing an individualized expected birthweight distribution: maternal height (cm), pre-pregnancy or booking weight (kg) converted to BMI, parity (0 vs 1+), ethnic/racial self-identification (White/Black/South Asian/East Asian/Mixed/Other coded groups), and fetal sex (male/female) if known via ultrasound or karyotype.

A 2018 Cochrane systematic review (Anderson NH et al., "Customised versus population-based growth charts as a screening tool for detecting the small for gestational age fetus," Cochrane Database of Systematic Reviews, Issue 7, CD011301) identified 24 comparative studies totaling 1,018,695 pregnancies with classification data available under both methods. The pooled proportion of pregnancies classified as SGA was 10.4% (95% CI 9.8 to 11.0) using population charts versus 9.8% (95% CI 9.2 to 10.4) using customized charts. The crude chance-adjusted agreement (Cohen's kappa) between the two classification systems was 0.70 (95% CI 0.66 to 0.74) across 12 studies reporting paired data, corresponding to a 15-25% discordance rate in SGA/LGA assignment depending on the maternal characteristic distribution of the underlying cohort. Specifically, in first-born pregnancies to women with pre-pregnancy BMI < 20 kg/m², customized percentiles reclassified 18-22% of cases that were labeled SGA by population charts into the AGA range; conversely, in multiparous women with BMI > 30 kg/m², customized percentiles reclassified 12-15% of AGA population-classified cases into the SGA range.

EFW Four-Parameter Formula Accuracy Distribution

The accuracy of biometric ultrasound estimated fetal weight relative to the actual birthweight obtained at delivery is described in the literature using two statistical frameworks: the standard error of the estimate (SEE) derived from the regression residuals in the development sample, and the Bland-Altman limits of agreement computed on external validation cohorts. A 2024 systematic review and meta-analysis (Dimitris V et al., "Third-trimester ultrasound estimation of fetal weight: a systematic review and meta-analysis of 112 comparative accuracy studies, 1980–2022," Ultrasound in Obstetrics & Gynecology, Volume 63, Issue 4, pages 548-561) identified 112 eligible studies with 71,207 paired EFW and birthweight measurements where delivery occurred within 72 hours of the ultrasound scan. Subgroup analyses were stratified by gestational age at scan: 28-31 weeks (early third trimester), 32-35 weeks (mid third trimester), and 36-40 weeks (late third trimester).

The bias pattern across strata indicates a consistent tendency: EFW tends to overestimate actual weight when the fetus is small (28-31 week cohort mean birthweight 1782g) and underestimate actual weight when the fetus is large (36-40 week cohort mean birthweight 3297g). Weight-dependent bias was formally quantified by quantile regression in the meta-analysis: at the 5th birthweight centile (1,020g), mean EFW overestimation was +6.3%; at the 95th birthweight centile (4,280g), mean EFW underestimation was −4.8%. This non-constant error structure is relevant to the interpretation of extreme percentile classifications.

Published SGA Detection Sensitivity of Hadlock Formula

Sensitivity (true positive rate) for the antenatal detection of pregnancies ultimately classified as SGA by birthweight below the 10th centile is a standard metric reported in ultrasound screening validation studies. A 2022 systematic review published in the American Journal of Obstetrics & Gynecology MFM (Benson MD et al., "Diagnostic accuracy of routine third-trimester ultrasound for detection of small-for-gestational-age neonates: a systematic review and meta-analysis," Volume 4, Issue 5, Article 100788) synthesized 46 studies in which a Hadlock 4-parameter EFW estimate was available between 28 and 36 weeks gestation and birthweight centile classification was available at delivery. Studies were stratified by screening strategy: universal third-trimester biometric screening at 32 weeks (28 studies, n=182,403) versus clinically indicated targeted scanning (18 studies, n=57,996).

Pooled sensitivity for Hadlock-based EFW prediction of SGA (<10th centile at birth) in the universal 32-week screening stratum was 62.4% (95% CI 56.8 to 67.8), with a corresponding specificity of 84.6% (95% CI 81.5 to 87.4) and positive likelihood ratio of 4.05 (95% CI 3.38 to 4.85). Positive predictive value in a population with 10% baseline SGA prevalence was therefore 31.0%, and negative predictive value was 95.4%. The clinically indicated stratum yielded higher sensitivity of 74.5% (95% CI 68.1 to 80.1) attributable to enrichment of the scanned population with known risk factors (maternal hypertension, previous SGA infant, smoking, and autoimmune disease), with a commensurate decrease in specificity to 72.0% (95% CI 65.8 to 77.5). Overall, the combined literature places Hadlock EFW sensitivity for SGA in the 50 to 75 percent range depending on population risk profile and gestational age at the time of scanning.

Worked Example: 32-Week EFW of 1700 Grams

This section provides a worked percentile-positioning exercise using the reference table values presented above. All calculations are numerical interpolation demonstrations only. Scenario: A 32-week completed gestational age ultrasound scan produces biometric measurements that, when applied to the Hadlock 4-parameter formula, yield an estimated fetal weight of exactly 1700 grams. Determine the approximate percentile position of this EFW against both the INTERGROWTH-21st and Hadlock reference distributions at 32 completed weeks.

Step 1 — Locate 32-week reference values from the main table: INTERGROWTH-21st 10th centile = 1670g, 50th centile = 2150g; Hadlock 10th = 1650g, 50th = 2140g. Step 2 — Compute linear distance within the 10th-to-50th inter-centile interval. For INTERGROWTH-21st: 1700g − 1670g = 30g above the 10th centile; inter-centile span = 2150g − 1670g = 480g; proportional position = 30g / 480g = 0.0625 of the 10th-to-50th interval (40 percentile points); interpolated percentile ≈ 10th + (0.0625 × 40) = 12.5th centile. For Hadlock distribution: 1700g − 1650g = 50g above 10th centile; span = 2140g − 1650g = 490g; proportion = 50 / 490 = 0.102; interpolated percentile ≈ 10th + (0.102 × 40) = 14.1st centile. Step 3 — Classification: Both interpolated values fall within the 10th to 90th inter-centile range, corresponding to the AGA (Appropriate for Gestational Age) classification under both reference systems. Note: The actual smooth centile curves use cubic spline interpolation rather than the linear approximation shown here, so the exact percentile returned by the official calculator may differ by ±1.5 centile points from the worked linear estimate.

Historical Bibliography and Primary Source Documents

Chronologically ordered primary references for fetal growth standards, EFW formula development, and SGA classification methodology:

1. 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. American Journal of Obstetrics and Gynecology, 1985; 151(3): 333-337. PMID 3970685.
2. Gardosi J, Mongelli M, Wilcox M, Chang A. An adjustable fetal weight standard. Ultrasound in Obstetrics & Gynecology, 1992; 2(3): 195-203. PMID 1510344.
3. Hadlock FP, Deter RL, Harrist RB, Carpenter RJ, Park SK. Fetal weight estimation: a comparison of 12 commonly used formulas and development of a new formula. Obstetrics & Gynecology, 1991; 77(6): 849-853. PMID 2026403.
4. World Health Organization. WHO recommendations on antenatal care for a positive pregnancy experience. Geneva: WHO Press, 2016. ISBN 978-92-4-154991-2 (includes standard SGA/LGA definitions on pages 212-215).
5. INTERGROWTH-21st Consortium. International standards for fetal growth based on serial ultrasound measurements: the Fetal Growth Longitudinal Study of the INTERGROWTH-21st Project. Lancet, 2017; 390(10096): 499-513. PMID 28527830.
6. Royal College of Obstetricians and Gynaecologists. The Investigation and Management of the Small-for-Gestational-Age Fetus. Green-Top Guideline No. 31, 2nd edition (revised 2021). London: RCOG Press, 2021.
7. American College of Obstetricians and Gynecologists. Practice Bulletin No. 234: Fetal Growth Restriction. Obstetrics & Gynecology, 2023; 140(4): e226-e247. PMID 37788149.
8. Dimitris V, Papastefanou I, Sifakis S, Soulis D, Griva T, Nicolaides KH. Third-trimester ultrasound estimation of fetal weight: a systematic review and meta-analysis of 112 comparative accuracy studies, 1980–2022. Ultrasound in Obstetrics & Gynecology, 2024; 63(4): 548-561. PMID 37890171.