Skip to main content
UKPMC Funders Author Manuscripts logoLink to UKPMC Funders Author Manuscripts
. Author manuscript; available in PMC: 2013 Jan 4.
Published in final edited form as: J Ultrasound Med. 2010 Feb;29(2):215–223. doi: 10.7863/jum.2010.29.2.215

Patterns of fetal growth in a rural Indian cohort and a comparison with a western European population, data from the Pune Maternal Nutrition Study

Arun S KINARE 1, Manoj C CHINCHWADKAR 1, Asit S NATEKAR 1, Kurus J COYAJI 1, Andrew K WILLS 2, Charudatta V JOGLEKAR 1, Chittaranjan S YAJNIK 1, Caroline HD FALL 2
PMCID: PMC3537223  EMSID: EMS37954  PMID: 20103791

Abstract

Objective

To describe fetal size in a rural Indian population and compare it with European and urban Indian populations using ultrasound.

Methods

Participants were from the Pune Maternal Nutrition Study, India. Fetal growth curves were constructed from serial ultrasound scans at ~18, 30 and 36 weeks gestation in 653 singleton pregnancies. Measurements included femur length (FL) and abdominal circumference (AC), and biparietal diameter (BPD) and occipito-frontal diameter (OFD) from which head circumference (HC) was estimated. Measurements were compared with data from a large population-based study in France and a study of urban mothers in Vellore, South India.

Results

Fetal AC and BPD were smaller than the French reference at 18 weeks gestation (−1.38 SD and −1.30 SD respectively), while FL and HC were more comparable (−0.77 SD and −0.59 SD). The deficit remained similar at 36 weeks for AC (−0.97 SD), FL (−0.43 SD) and HC (−0.52 SD) and increased for BPD (−2.3 SD). Ultrasound at 18 weeks under-estimated gestational age, compared with LMP date, by a median of −1.4 (IQR −4.6, 1.8) days. The Pune fetuses were smaller, even at the 1st scan, than the urban Vellore sample.

Conclusions

Fetal size is smaller in a rural Indian population than in European or urban Indian populations, even in mid pregnancy. The deficit varied for different fetal measurements; it was greatest for abdominal circumference and biparietal diameter and least for femur length and head circumference.

Keywords: Fetal ultrasound, fetal growth, population differences, India

Introduction

Research into the developmental origins of health and disease has focused attention on fetal development as a determinant of lifelong health and capacity. Size and body proportions at birth predict short- and long-term outcomes, from infant mortality1, through childhood growth and cognitive ability2,3, to diseases in adult life such as type 2 diabetes and cardiovascular disease4,5. Research in this field has relied mainly on studies linking birthweight to outcomes in later life. Birth size, however, is a crude summary measure of fetal growth, and two babies of identical birthweight may have followed different fetal growth trajectories6. Prospective studies incorporating serial ultrasound measurements of fetal size may increase understanding of the fetal origins of health and disease.

Fetal growth differs not only between individuals but also between populations. Indian babies are among the smallest in the world; mean full term birthweight is 2.6-2.9 kg compared with 3.5-3.7 kg in white caucasian populations in high-income countries7. Indian babies are not proportionately smaller in all body measurements. A comparison of newborn anthropometry between UK and India showed that while birthweight and abdominal circumference were lower in the Indian babies by 1.7 and 2.4 standard deviations (SD), birth length and subscapular skinfolds were lower by only 1.0 and 0.5 SD8. Studies of fetal growth and its determinants in different populations may therefore provide information on population-specific health problems, such as the current epidemic of type 2 diabetes in India9.

The Pune Maternal Nutrition Study (PMNS), based in rural villages near Pune, India, is one of the first prospective studies established specifically to study associations between maternal nutritional status and long-term outcomes in the offspring8,10. Fetal size was measured serially using ultrasound in over 800 pregnancies. This paper is the first in a series exploring fetal growth patterns, their determinants, and their associations with health outcomes in the children. Since there is little published data from India, we first present simple descriptive data on fetal ultrasound measurements, and a comparison with a European population.

Materials and Methods

The PMNS methodology has been described previously8,10. In brief, married women of childbearing age (15-40yrs, n=2675) were identified by a survey of 6 villages located 40-50 km from Pune City. Most families lived by subsistence farming on small landholdings, and most of the women were vegetarian, had low energy and protein intakes, and did farming work in addition to domestic chores. Non-pregnant women were enrolled in 1994-1996, and 2466 (92%) agreed to participate. Their weight and height were measured, and they were visited monthly by trained health workers to record menstrual period dates. The health workers were girls with 8-10 years of schooling, recruited from the same villages, to ensure rapport with the women. They were trained to ask probing questions, using the religious calendar and local events, to aid the women’s recall and obtain the most accurate possible menstrual dates. Women who missed a period had an ultrasound scan 15-18 weeks after the last menstrual period (LMP) date, to confirm pregnancy.

Ultrasound measurements

Ultrasound scans were carried out by one of two trained sonologists (ASN and MCC) using a portable machine with a curvilinear array 5MHz transducer (ALOKA SSD 500, version 8.1, Osaka, Japan) carried in a customised van that visited each village weekly. Biparietal diameter (BPD) was measured at the level of the thalami and cavum septum pellucidum from the outer table of the proximal calvarium to the inner table of the distal calvarium11,12. Occipito-frontal diameter (OFD) was measured at the same level, and head circumference (HC) was calculated using the formula (BPD+OFC) ×1.62. Abdominal circumference (AC) was calculated as (transverse diameter+anteroposteror diameter)×1.57 at the level of umbilical vein-ductus venosus complex13. Femur length (FL) was measured along the long axis of the ossified femoral diaphysis with both the femoral head and distal epiphysis visible, though not included in the measurement14. The variation attributable to observers ranged from 0.004-0.04% for these measurements. Sonographic gestation was calculated as an average of the predicted age derived from the fetal BPD, HC, AC and FL15.

The PMNS was designed to collect fetal growth data at fixed times during gestation (approximately 18±2 weeks, 28±2 weeks and 36±2 weeks). For a study of fetal growth, it is important to use LMP dates to derive fetal age, rather than sonographic gestation, because the latter assumes identical growth in all fetuses, and simply translates a measure of size into a gestational age, using reference data. Despite monthly visits by the health workers, LMP dates were clearly inaccurate in some women, leading to large discrepancies between LMP-derived and sonography-derived gestational age. For the practical purposes of running the study and scheduling later scans, gestational age at the first visit after a missed period was derived from the LMP date, unless it differed from the sonographic estimate by more than 2 weeks, in which case the latter was used. For the purpose of this analysis, we used only gestational ages derived from LMP dates, and excluded women whose LMP-derived gestation differed by more than 2 weeks from sonographic gestation (N=144). Seventeen women were excluded because they had no ultrasound or LMP data. Of 1,102 women with a confirmed pregnancy, 288 were excluded because of spontaneous or medical abortion, multiple pregnancy, fetal anomalies on ultrasound, or because the pregnancy was too advanced (>20 weeks). The final sample size was 653. Of these, 372 mothers (57%) had 3 scans, 228 (35%) had 2 scans and 53 (8%) had only 1 scan; 653 attended the 1st scheduled appointment, 587 the 2nd and 385 the third. The median (interquartile range [IQR]) gestational age at each examination was 17 (17,18), 29 (29,30) and 35 (34,36) weeks.

Most deliveries occurred at home in the villages. Health workers performed detailed newborn anthropometry within 72 hours of birth, using standardised protocols, adapted from reference techniques used in children16,17 to measure weight, crown-rump length, triceps and subscapular skinfold thickness, and mid-upper-arm, occipito-frontal head, and abdominal circumferences. Inter- and intra-observer variation studies and re-training sessions were performed for all health workers 6-monthly.

Permission for the study was granted by village elders and by the research ethics committee of the KEM Hospital, Pune. Informed consent was obtained from the women.

Statistical analysis

The approach outlined by Royston18 was used to construct fetal growth curves. The model formulation is a sequential process. A power (λ^) for the transformation of each fetal dependent variable (Y) was estimated using a Box-Cox regression procedure. We then used regression to estimate a suitable function of gestational age X=g(T) for Y and Yλ^, such that each individual’s response variable was approximately linear. For the function g(T), a family of 2nd degree fractional polynomial functions was considered and the model with the lowest deviance selected19. The usefulness of the Y transformation in reducing residual non-normality and heteroscedasticity was assessed using a pseudo F-test. The following multilevel model was then fitted using a restricted maximum likelihood algorithm. The following multilevel model (multilevel because we had repeated measures from each fetus) for each fetal size component was then fitted using a restricted maximum likelihood algorithm:

Yijλ^=β0+μ0i+(β1+μli)g(Tij)+εij (1)

Where, Yijλ^ is the transformed response of HC, BPD, AC and FL and g(Tij) is the covariate function of gestational age for fetus i at observation j. β0 and β1 are the fixed intercept and slope, μ0i and μ1i are the random intercept and slope coefficients for each fetus (j), and eij is the leftover error term for fetus i, at observation point j. Predicted means and growth curves with a 95% reference interval were plotted from the final models. Results from the model fitting process and parameter estimates to recreate the growth curves are presented in Table 1.

Table 1. Results from the model fitting process and parameter estimates to recreate the growth charts.

Y
HC BPD AC FL
MODEL FITTING:
λ^ 0.73 0.58 0.32 1.0
95% CI (0.66, 0.81) (0.51, 0.65) (0.26, 0.39) (0.96, 1.06)
λ^ rounded 0.75 0.6 0.3 1.0
λ^ final 0.75 0.6 0.3 1.0
Untransformed Y
g(T) T2-0.01752 T3 T3 (1-0.25328 ln T) T3 (1-0.24754 ln T) T−2 (1-0.45866 ln T)
Transformed Y (Yλ)
g(T) T2-0.01789 T3 T2-0.01787 T3 T−2 (1-0.48246 ln T) -
For transformation
Ftrans 18.97 50.83 114.7 -
PARAMETER ESTIMATES
μ^ 2.3352 1.0889 3.2996 10.9676
β^ 0.02302 5.259 × 10−3 986.5059 8410.0829
σ^μ2 0.1284 0.01184 2.8832 × 10−3 0.095506
σ^β2 5.004 × 10−7 6.375 × 10−8 2296.5523 97444.34
σ^μβ2 1.8721 × 10−4 2.206 × 10−5 1.9713 77.5027
σ^ε2 0.05187 2.8262 × 10−3 7.6017 × 10−4 0.01283

Notation is as per Royston 199516. Powers (λ^) for a transformation of fetal measurements (Y) were estimated using the Box-Cox regression procedure in Stata v10, and rounded to the nearest .05 (λ^ rounded) or 1 if the 95% CI overlapped 1 (no transformation). The covariate functions for gestational age in weeks (g(T)), selected from a family of 2nd order fractional polynomials19, are presented. The effect of λ^ rounded in reducing residual non-normality and heteroscedasticity was assessed using an F-test (Ftrans). High values of F (>3) indicate that the transformation of Y improves the residual diagnostics, and so dictated the choice of a transform of Y (λ^ final). Bold values indicate the final covariate function used to transform gestational age.

The parameter estimates can be used to recreate the growth profiles and reference intervals. The mean E(Yλ) and variance var (Yλ) of the transformed fetal measurement is given by:
E(Yλ)=μZ=μ+βg(T)
var(Yλ)=σZ2=σ^μ2+g(T)2σ^β2+2g(T)σ^μβ2+σ^ε2

The desired reference interval can then be calculated using the normal distribution function, for example a 95% reference range would be given by (μZ ± 1.96 σZ)1/λ

To test the reliability of the growth curves, models were fitted in another sample of 153 fetuses from mothers in the same community who had ultrasound scans within 10 weeks of their LMP as part of another study. The growth curves in this subset were similar to those reported in this study. Further, the coefficients to indicate the data source were non significant (p>0.05, Wald test) in the growth models using the pooled datasets.

We compared the PMNS data with fetal ultrasound data from a large French population-based study, in which fetal measurements were made using similar techniques20. The Pune fetal measurements were estimated and plotted in z-standardized units referenced to this cohort. We also made a comparison with an urban South Indian cohort (Vellore)21. No equations for the growth curves were provided in this paper, so we restricted our comparisons to the tabulated median, 10th and 90th centiles at 18, 30 and 36 weeks provided by the authors. Information on HC was also not available for this cohort. Patterns of fetal growth were similar in male and female fetuses, so the data were pooled. Pooling also allowed a direct comparison with the reference populations.

Mean values for birthweight and birth length at full term (40 weeks) in this population were obtained by adjusting measurements at birth for gestation in completed weeks, using linear regression.

Results

The PMNS mothers were short, light and thin (mean height and body mass index (BMI): 152 cm and17.9 kg/m2) (Table 2); 64% were underweight (<18.5 kg/m2)22. The majority were under 22 years old and approximately one-third were primiparous. The mean birth weight was 2609g, 32% were classified as low birth weight (<2500g) and 11% were born pre-term (<37 weeks gestation).

Table 2. Characteristics of the cohort.

N Mean (SD) (unless stated)
Fetus at birth
Sex (male) 626 340 (54.3%)
Birth weight (g) 576 2609 (398)
Low birth weight (<2500g) 576 183 (32%)
Birth length (cm) 596 47.4 (2.34)
Head circumference (cm) 597 32.9 (1.47)
Abdominal circumference (cm) 597 28.4 (2.1)
Gestational age (weeks) 627 39.1 (38.1, 40.3)
Premature (<37 weeks) 627 69 (11%)
Mother:
Age (years) 653 21 (19, 23) (range: 15-40)
Parity (0) 653 210 (32%)
(1-3) 653 415 (64%)
(≥4) 28 (4.3%)
Height (cm) 653 151.9 (5.0)
Weight (kg) 647 41.7 (5.0)
BMI (kg/m2) 647 17.8 (16.7, 19.1)
Father:
Height (cm) 610 164.5 (6.1)
Weight (kg) 614 52.6 (7.8)
BMI (kg/m2) 609 19.0 (17.6, 20.7)

n (%);

median (interquartile range)

The mean birth weight adjusted to 40 weeks was 2718g (SD 337). While direct comparisons against the reference populations are not possible, in a contemporaneous cohort of 58,834 newborns from the same region as the French fetal data23, the mean birth weight at 40 weeks was 3477 g (SD 409). In the South Indian urban study the median birth weight was 3000 g (this estimate is likely to be larger if we had the adjusted value at 40 weeks). Newborn size was therefore smaller in Pune than in both the French and Vellore populations. The mean adjusted birth length at 40 weeks was 48.1 cm (SD 1.99) in the PMNS sample and 50.5 cm (SD 1.82) in the French geographically matched sample. In standardized units, the Pune babies were 1.86 SD and 1.32 SD below the French weight and length reference respectively at birth.

Tests for bias in the sample

There were no significant differences in any of the fetal size variables at the 1st visit when comparing fetuses that had 3 scans versus those which had <3 scans (p=0.6-0.95 for all comparisons, adjusted for gestational age), and no significant difference in birth weight, length or abdominal circumference between those with differing number of scans (p=0.084-0.25).

Statistical description of models

A transformation was deemed necessary to improve the model-fit of all the fetal measurements except femur length (Table 1). The models and residuals were a good fit; the proportion of observations outside the 95% reference interval ranged from 4.2 – 4.7%.

Whilst there was tracking in the growth curves, there was still substantial centile crossing. This was evident in the intra-class correlation coefficients, which reflect the amount of within-subject correlation in SD-scores across gestation. The ICC’s (95% CI) for HC, BPD, AC and FL were 0.47 (0.42, 0.53), 0.49 (0.43, 0.54), 0.51 (0.46, 0.56) and 0.62 (0.58, 0.66).

Comparison of growth curves with the European population

Mean BPD and AC in the PMNS were smaller than in the European sample, even at the first ultrasound scan (Figure 1, Table 3), while HC and FL values were closer to the European values. All measurements were smaller in late pregnancy and there was a hierarchy - the greatest relative disparities were in AC (median 29.0 cm v 31.0 cm) and BPD (8.1 cm v 9.0 cm) at 36 weeks.

Figure 1. Plots of HC, BPD, AC and FL with the fitted prediction line and 95% reference range. The 95% reference and mean from the western European cohort is in grey.

Figure 1

Table 3. Predicted 10th, 50th and 90th centiles of fetal size 20, 28 and 36 weeks of gestation.

Pune European Vellore*
Fetal Measure 10th 50th 90th 10th 50th 90th 10th 50th 90th
20 weeks:
Head circumference (cm) 15.5 16.7 17.9 15.9 17.1 18.4 - - -
Biparietal diameter (cm) 4.1 4.4 4.8 4.4 4.8 5.1 4.4 4.8 5.3
Abdominal circumference (cm) 12.7 13.9 15.2 13.8 15.2 16.5 12.9 14.6 16.8
Femur length (cm) 2.8 3.1 3.4 3.0 3.2 3.5 3.0 3.3 3.8
28 weeks:
Head circumference (cm) 24.2 25.5 26.8 24.1 25.7 27.3 - - -
Biparietal diameter (cm) 6.4 6.8 7.2 6.8 7.2 7.7 6.6 7.4 7.7
Abdominal circumference (cm) 20.6 22.2 23.9 21.7 23.7 25.7 19.9 22.9 25.5
Femur length (cm) 5.0 5.3 5.6 5.0 5.3 5.7 5.0 5.5 6.1
36 weeks:
Head circumference (cm) 29.1 30.4 31.8 29.3 31.2 33.2 - - -
Biparietal diameter (cm) 7.7 8.1 8.6 8.5 9.0 9.5 8.2 8.7 9.3
Abdominal circumference (cm) 26.9 29.0 31.1 28.3 31.0 33.7 27.3 29.5 32.6
Femur length (cm) 6.5 6.8 7.1 6.5 6.9 7.3 6.4 7.0 7.3
*

Actual centiles

The differences in the relative size of the fetal components are shown more clearly when the PMNS values are plotted as standardised scores on the European reference (Figure 2). Mean HC and FL tracked between the 22nd and 45th percentiles of the European reference from 18 to 36 weeks. Mean BPD tracked below the 10th percentile of the European distribution over gestation and AC between the 8th and 16th percentile.

Figure 2. Predicted mean HC, BPD, AC, FL and their 95% confidence intervals at 18, 30 and 35 weeks of gestation in the Indian cohort plotted as a percentile on the western European reference.

Figure 2

Comparison with South Indian urban cohort

Compared with data from Vellore, the PMNS fetuses were smaller at all time points in gestation, with marked differences in BPD and AC evident even at 20 weeks (Table 2).

Gestational dating and intra-uterine growth restriction

The Hadlock equations to date pregnancies from 4 ultrasound measurements15 systematically underestimated gestational age compared to LMP dating (median difference at 18 weeks: −1.35 days; IQR: −4.60, 1.77, signed ranks test: p<0.001. Defined as an abdominal circumference below the 10th percentile of the European reference, 54% of the fetuses had intra-uterine growth restriction (IUGR) at the 1st visit (~18 weeks), 36% at the 2nd (~28 weeks) and 30% at the 3rd (~35 weeks) visit.

Discussion

This is the first study to present ultrasound-derived measurements of fetal growth from a rural Indian cohort, and to compare them against a western population. Fetal abdominal circumference and biparietal diameter were markedly smaller than the western reference at 18 weeks gestation, while femur length and head circumference were comparable. In late pregnancy (30 and 36 weeks), all measurements were smaller than the European reference. The deficit was greatest for abdominal circumference and biparietal diameter (the latter becoming smaller relative to the European population as pregnancy progressed). PMNS fetal size measurements were also small compared with an urban South Indian study.

Strengths of the study were that it was population-based, and by using a portable machine collected serial ultrasound data on a representative sample from a rural population. Measurements were made using standardized protocols by 2 experienced medical sonologists, ensuring high-quality measurements and minimizing ‘noise’ due to inter-observer variation. There are few studies with carefully collected LMP data based on regular monthly visits, in a population like this. However, despite the care taken, 144 women were excluded from the analysis because of an implausibly discrepant date relative to ultrasound measurements of fetal size. Conception date during the first cycle after stopping the oral contraceptive pill (OCP) can be unreliable, but none of the women in the study were taking the OCP. The study was not designed to generate fetal growth reference curves, which are ideally based on cross-sectional data collected evenly throughout pregnancy24,25, rather than at 3 time points as in the PMNS. However, there was no evidence of bias in terms of the contribution of repeated measurements to the growth plots, and so we have presented the fetal growth equations (Table 1) as a potentially useful reference.

This is the first data showing that the Indian fetus, at least in this rural population, is smaller than the European fetus even at 18 weeks gestation. It is generally thought that the small size of Indian babies at birth is attributable to small maternal size and/or an inadequate nutrient supply in mid-late pregnancy (due to maternal undernutrition and/or placental insufficiency), but that early fetal growth, when nutrient requirements are very small and when there are no constraints on space for growth, is similar to other populations21. This suggests that any interventions to increase fetal growth in rural Indian populations would need to be pre-conceptional or in early pregnancy.

The data suggest a hierarchy within the different fetal components or tissues; femur growth is relatively preserved in Indian fetuses, while abdominal circumference (a combined measure of visceral size and subcutaneous fat) grows more slowly than in European fetuses from early pregnancy. This is consistent with our findings at birth in Indian populations8,26. Compared with UK newborns, birth length in the PMNS was relatively preserved (−1.0 SD) compared with birthweight (−1.74 SD) and abdominal circumference (−2.38 SD). Newborn length was still 4 cm lower at birth than in the UK babies8, suggesting a greater relative deficit in components of fetal length other than femoral length, for example spine or head.

The pattern of growth differed for the two measures of fetal head size in Pune. While head circumference was similar to the European population in early pregnancy, biparietal diameter was markedly smaller. This suggests that the head shape of these fetuses may differ and that the occipito-frontal diameter may be larger than in European fetuses in early pregnancy. Fetal occipito-frontal diameter (OFD) is not usually reported in the ultrasound literature and so we are unable to confirm this. We do not know the implications of any differences in fetal head shape for brain growth and function.

The PMNS was established to study early-life exposures (maternal nutrition, fetal growth and newborn phenotype) in relation to long-term outcomes (risk for cardiovascular disease and type 2 diabetes). It will therefore take time before the significance of our findings is known in terms of these clinical outcomes. The main importance of our findings for obstetricians is that the growth of the rural Indian fetus differs from the Western ultrasound reference that is generally used in clinical practice in India. Hence, gestational age tends to be underestimated, and intra-uterine growth restriction is diagnosed very frequently. In our study, the mean difference in gestation between the ultrasound and LMP estimates was 1.4 days at ~18 weeks, which is unlikely to be of obstetric significance (though it may be of significance for long-term outcomes). The incidence of intra-uterine growth restriction was very high (≥30%) throughout gestation, and this is likely to influence pregnancy management, although other ultrasound features of IUGR would usually also be considered, such as liquor volume and umbilical blood flow patterns. Our study was not large enough to relate fetal growth patterns to obstetric and perinatal complications (out of 770 births after 28 weeks gestation there were 8 stillbirths and 8 early neonatal deaths). However, our data suggest the need for a locally-generated fetal growth reference, along with prospective data on obstetric and perinatal outcomes, to enable the development of better clinical guidelines for rural populations, which constitute a high proportion of many developing country populations (~70% in India).

Conclusions

Fetal size is smaller in a rural Indian population than in European or urban Indian populations, even in mid-pregnancy. The deficit varies in different fetal ‘components’, being greatest for abdominal circumference and biparietal diameter and least for femur length and head circumference. Interventions designed to increase fetal growth should start pre-conceptionally or in early pregnancy. Gestational age derived using ultrasound measurements and Western reference equations are underestimated in this population. Our findings need to be replicated in other Indian populations with data collected earlier in pregnancy. Future analysis of this data will examine determinants of fetal growth, including parental size and maternal nutritional status, and relationships of fetal growth to newborn condition, post-natal growth and childhood metabolic status.

Acknowledgements

We thank the women and their families and the late Dr Banoo Coyaji, Director of the KEM Hospital, Pune, and initiator twenty-five years ago, of the rural health care programme in the study area. We acknowledge major contributions to the study by Dr Shobha Rao, Dr VN Rao, Dr Monesh Shah, Dr Binu John, Dr Anuja Bisht, Dr Mahananda Bhavikatti, Dr Asawari Kanade, Mrs Punam Gupta, Mrs Parveen Bharucha. The study was funded by the Wellcome Trust and Medical Research Council, UK. We thank Sneha-India (Society for the study of Natal Effects on Health in Adults) for its support.

References

  • 1.Williams RL, Creasy RK, Cunningham GC, Hawes WE, Norris FD, Tashiro M. Fetal growth and perinatal viability in California. Obstetrics and Gynecology. 1982;59:624–632. [PubMed] [Google Scholar]
  • 2.Goldenberg RL, Hoffman HJ, Cliver SP. Neurodevelopmental outcome of small-for-gestational age infants. Eur J Clin Nutr. 1998;52(Suppl 1):S54–58. [PubMed] [Google Scholar]
  • 3.Martorell R, Ramakrishnan U, Schroeder DG, Melgar P, Neufeld L. Intrauterine growth retardation, body size, body composition and physical performance in adolescence. Eur J Clin Nutr. 1998;52(S1):S43–53. [PubMed] [Google Scholar]
  • 4.Hales CN, Barker DJP, Clark PMS, Cox LJ, Fall C, Osmond C, Winter PD. Fetal and infant growth and impaired glucose tolerance at age 64. BMJ. 1991;303:1019–1022. doi: 10.1136/bmj.303.6809.1019. [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 5.Leon D, Lithell HO, Vagero D, Koupilova I, Mohsen R, Berglund L, et al. Reduced fetal growth rate and increased risk of death from ischaemic heart disease: cohort study of 15000 Swedish men and women born 1915-29. BMJ. 1998;317:241–245. doi: 10.1136/bmj.317.7153.241. [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 6.Harding JE. The nutritional basis of the fetal origins of adult disease. Int J Epidemiol. 2001;30:15–25. doi: 10.1093/ije/30.1.15. [DOI] [PubMed] [Google Scholar]
  • 7.Sachdev HPS. Low birthweight in South Asia. In: Gillespie S, editor. Malnutrition in South Asia; a regional profile. Unicef; 1997. [Google Scholar]
  • 8.Yajnik CS, Fall CHD, Coyaji KJ, Hirve SS, Rao S, Barker DJP, Joglekar C, Kellingray S. Neonatal anthropometry: the thin-fat Indian baby; the Pune Maternal Nutrition Study. Int J Obesity. 2003;27:173–180. doi: 10.1038/sj.ijo.802219. [DOI] [PubMed] [Google Scholar]
  • 9.Wild S, Roglic G, Green A, Sicree R, King H. Global prevalence of diabetes; estimates for the year 2000 and projections for 2030. Diabetes Care. 2004;27:1047–1053. doi: 10.2337/diacare.27.5.1047. [DOI] [PubMed] [Google Scholar]
  • 10.Rao S, Yajnik CS, Kanade A, Fall CHD, Margetts BM, Jackson AA, Shier R, Joshi S, Rege S, Lubree H, Desai B. Intake of micronutrient-rich foods in rural Indian mothers is associated with the size of their babies at birth; the Pune Maternal Nutrition Study. J Nutr. 2001;131:1217–1224. doi: 10.1093/jn/131.4.1217. [DOI] [PubMed] [Google Scholar]
  • 11.Hadlock FP, Deter RL, Harrist RB, Park SK. Fetal biparietal diameter: rational choice of plane of section for sonographic measurement. Am J Roentgenol. 1982;138:871–874. doi: 10.2214/ajr.138.5.871. [DOI] [PubMed] [Google Scholar]
  • 12.Hadlock FP, Deter RL, Harrist RB, Park SK. Fetal head circumference: relation to menstrual age. Am J Roentgenol. 1982;138:649–653. doi: 10.2214/ajr.138.4.649. [DOI] [PubMed] [Google Scholar]
  • 13.Hadlock FP, Deter RL, Harrist RB, Park SK. Fetal abdominal circumference as a predictor of menstrual age. Am J Roentgenol. 1982;139:367–370. doi: 10.2214/ajr.139.2.367. [DOI] [PubMed] [Google Scholar]
  • 14.Hadlock FP, Deter RL, Harrist RB, Park SK. Fetal femur length as a predictor of menstrual age: sonographically measured. Am J Roentgenol. 1982;138:875–8. doi: 10.2214/ajr.138.5.875. [DOI] [PubMed] [Google Scholar]
  • 15.Hadlock FP, Deter RL, Harrist RB, Park SK. Estimating fetal age: computer-assisted analysis of multiple fetal growth parameters. Radiology. 1984;152:497. doi: 10.1148/radiology.152.2.6739822. [DOI] [PubMed] [Google Scholar]
  • 16.Tanner JM. Chapter 11: Standards of normal growth. In: Tanner JM, editor. Foetus into man; physical growth from conception to maturity. Second edition Castlemead publications; Ware, UK: 1989. [Google Scholar]
  • 17.Garrow JS. Chapter 2: Composition of the body. In: Garrow J, James WPT, Ralph A, editors. Human nutrition and dietetics. 10th edition Churchill Livingstone; Edinburgh, UK: 2000. [Google Scholar]
  • 18.Royston P. Calculation of unconditional and conditional reference intervals for foetal size and growth from longitudinal measurements. Stat Med. 1995;14:1417–1436. doi: 10.1002/sim.4780141303. [DOI] [PubMed] [Google Scholar]
  • 19.Royston P, Altman DG. Regression using fractional polynomials of continuous covariates: parsimonious parametric modelling (with discussion) Applied Statistics. 1994;43:429–467. [Google Scholar]
  • 20.Salomon LJ, Bernard JP, Ville Y. Estimation of fetal weight: reference range at 20-36 weeks’ gestation and comparison with actual birth-weight reference range. Ultrasound Obstet Gynecol. 2007;29:550–5. doi: 10.1002/uog.4019. [DOI] [PubMed] [Google Scholar]
  • 21.Mathai M, Thomas S, Peedicayil A, Regi A, Jasper P, Joseph R. Growth pattern of the Indian fetus. Int J Gynaecol Obstet. 1995;48:21–24. doi: 10.1016/0020-7292(94)02237-2. [DOI] [PubMed] [Google Scholar]
  • 22.World Health Organization . Physical status: the use and interpretation of anthropometry. World Health Organization; Geneva: 1995. (WHO Technical Report Series No. 854). [PubMed] [Google Scholar]
  • 23.Salomon LJ, de Stavola Bernard JP, Kenward BM, Ville Y. Poids et taille de naissance: courbes et equations. [Birth weight and size: charts and equations] Journal de Gynecologie Obstetrique et Biologie de la Reproduction. 2007;36:50–56. doi: 10.1016/j.jgyn.2006.09.001. [DOI] [PubMed] [Google Scholar]
  • 24.Altman DG, Chitty LS. Charts of fetal size: 1. Methodology. Br J Obstet Gynaecol. 1994;101:29–34. doi: 10.1111/j.1471-0528.1994.tb13006.x. [DOI] [PubMed] [Google Scholar]
  • 25.Silverwood RJ, Cole TJ. Statistical methods for constructing gestational age-related reference intervals and centile charts for fetal size. Ultrasound Obstet Gynecol. 2007;29:6–13. doi: 10.1002/uog.3911. [DOI] [PubMed] [Google Scholar]
  • 26.Krishnaveni GV, Hill JC, Veena SR, Leary SD, Saperia J, Chachyamma KJ, Karat SC, Fall CHD. Truncal adiposity is present at birth and in early childhood in South Indian children. Indian Pediatrics. 2005;42:527–538. [PubMed] [Google Scholar]

RESOURCES