Advertisement

A customized standard to assess fetal growth in a US population

      Objective

      The objective of the study was to assess the factors that affect fetal growth and birthweight, and to derive coefficients for a customized growth chart applicable in an American population.

      Study Design

      This was a prospective cohort study of 35,235 pregnancies. Coefficients for physiological and pathological variables were derived by backward multiple regression.

      Results

      The expected birthweight at 40.0 weeks for a standard-size primiparous mother of European origin in an uncomplicated pregnancy was 3453.4 g, very similar to the standardized birthweight observed in other populations. Physiological coefficients were derived for maternal height, weight, parity, ethnic origin, and sex of the baby. Smoking, history of preterm delivery, and hypertensive diseases in the current pregnancy all had negative effects on birthweight, whereas babies of diabetic mothers weighed more. Low as well as high body mass index was associated with birthweight deficit at term.

      Conclusion

      Coefficients that allow determination of the customized growth potential, individually adjusted and excluding known pathological factors, have been derived. Babies of obese mothers have an increased risk of not reaching their fetal growth potential.

      Key words

      Accurate assessment of intrauterine growth is an essential part of antenatal care and perinatal research. It requires a standard that can be individually adjusted or customized to reflect the growth potential of the fetus in each pregnancy.
      • Gardosi J.
      Customised fetal growth standards: rationale and clinical application.
      For Editors' Commentary, see Table of Contents
      To determine the customized growth potential, a predicted weight at term for a pregnancy in optimal conditions is firstly calculated, using adjustment coefficients derived from the local population.
      • Gardosi J.
      • Mongelli M.
      • Wilcox M.
      • Chang A.
      An adjustable fetal weight standard.
      Physiological or constitutional variables such as maternal size, parity and ethnic origin are adjusted for, whereas pathological factors such as smoking, hypertensive diseases and diabetes are excluded, even if they are known to be present, to set the expected standard so as to better recognize if fetal growth has been affected. The predicted “term optimal weight” is combined with a proportionality function derived from an ultrasound-based fetal weight curve to determine the optimal and normal range of fetal weight for each point in gestation.
      • Gardosi J.
      • Mongelli M.
      • Wilcox M.
      • Chang A.
      An adjustable fetal weight standard.
      Such a customized standard has been found to improve the distinction between normal and abnormal growth and to enhance our understanding of the factors associated with fetal growth restriction.
      • de Jong C.L.D.
      • Gardosi J.
      • Dekker G.A.
      • Colenbrander G.J.
      • van Geijn H.P.
      Application of a customised birthweight standard in the assessment of perinatal outcome in a high risk population.
      • Clausson B.
      • Gardosi J.
      • Francis A.
      • Cnattingius S.
      Perinatal outcome in SGA births defined by customised versus population-based birthweight standards.
      • McCowan L.
      • Harding J.E.
      • Stewart A.W.
      Customised birthweight centiles predict SGA pregnancies with perinatal morbidity.
      • Groom K.M.
      • Poppe K.K.
      • North R.A.
      • McCowan M.E.
      Small-for-gestational age infants classified by customized or population birthweight centiles: impact of gestational age at delivery.
      • Figueras F.
      • Figueras J.
      • Meier E.
      • et al.
      Customised birthweight percentiles accurately predict perinatal morbidity.
      It is recommended by Royal College of Obstetricians and Gynaecologists Guidelines

      Royal College of Obstetricians and Gynaecologists. The investigation and management of the small-for-gestational age fetus. Royal College of Obstetricians and Gynaecologists Green Top Guideline 2002 (#31).

      and is already in widespread use clinically and in ongoing research. There has recently been a call for customized growth charts to be adopted by obstetricians in the United States.
      • Resnik R.
      One size does not fit all.
      Locally derived standards with appropriate coefficients for adjusting the expected term weight according to physiological variables have been published for maternity populations in the United Kingdom,
      • Gardosi J.
      • Mongelli M.
      • Wilcox M.
      • Chang A.
      An adjustable fetal weight standard.
      New Zealand,
      • McCowan L.
      • Stewart A.W.
      • Francis A.
      • Gardosi J.
      A customised birthweight centile calculator developed for a New Zealand population.
      France,
      • Ego A.
      • Subtil D.
      • Grange G.
      • et al.
      Customised versus population-based birth weight standards for identifying growth restricted infants: French multicenter study.
      Spain,
      • Figueras F.
      • Meler E.
      • Iraola A.
      • et al.
      Customised birthweight standards for a Spanish population.
      and Australia.
      • Mongelli M.
      • Figueras F.
      • Francis A.
      • Gardosi J.
      A customised birthweight centile calculator developed for an Australian population.
      The main purpose of this study was to derive coefficients that can be used to determine a customized fetal growth potential in an American population.

      Materials and Methods

      Study population

      Anonymized data were obtained from a National Institutes of Health-sponsored study conducted at 15 centers across the United States to evaluate screening tests in singleton pregnancies, conducted between October 1999 and December 2002 with institutional review board approval and participants' informed consent. All pregnancies were dated by ultrasound, and entry into the study was between gestational age of 10 weeks 3 days through 13 weeks 6 days according to fetal crown rump measurement. Details of the database have been described elsewhere.
      • Malone F.D.
      • Canick J.A.
      • Ball R.H.
      • et al.
      First-trimester or second-trimester screening, or both, for Down's syndrome.
      By agreement with the original study team, the data obtained for this study were analyzed by us independently.
      Of a total of 38,033 cases, 2798 were excluded because of missing or inconsistent values for gestation and/or birthweight, leaving 35,255 cases for univariate analysis. To derive coefficients for customized centiles, a further 4398 cases were excluded, which consisted of preterm deliveries (< 37 weeks) and/or cases with an incomplete set of variables, resulting in 30,837 cases for the multiple regression analysis.

      Statistical analysis

      The covariates for the multiple regression model are listed in Table 1. They include physiological variables such as maternal height and weight, parity, ethnicity, and the sex of the baby; in addition, pathological factors were assessed, relating to past history or complications in the current pregnancy. Maternal characteristics such as weight and smoking status were collected at time of recruitment.
      TABLE 1Characteristics of study population
      Characteristicn with datan%MeanSDMedianIQR
      Maternal age (y)35,25530.15.830.38.7
      Maternal height (cm)34,895164.57.1165.110.2
      Maternal weight (kg)35,22867.614.964.115.9
      BMI (kg/m2)34,87125.05.223.76.0
       < 20389011.2
       20-29.925,86474.2
       ≥ 30511714.7
       ≥ 3519055.5
      Parity35,247
       015,94945.2
       111,52532.7
       2508414.4
       317835.1
       ≥ 49062.6
      Ethnic origin35,232
       African American17414.9
       Asian/Pacific Islander13913.9
       European23,89867.8
       Hispanic786722.3
       Native American/Alaskan2100.6
       Other1250.4
      Smoking (n/d)35,224
       1-911203.2
       10-194311.2
       ≥ 20900.3
      Alcohol (drinks/wk)35,226
       0.5-25781.6
       ≥ 31800.5
      Marijuana use35,2433250.9
      Cocaine use35,246330.1
      History of
       Abortion35,235608317.3
       Miscarriage35,216920726.1
       Preterm birth35,230
        120235.7
        2+3120.9
      Diabetes35,2033130.9
      Gestational diabetes35,13512323.5
      Antepartum hemorrhage35,234
       Spotting444112.6
       Bleeding5551.6
      Placental abruption35,1512470.7
      Pregnancy induced hypertension35,14215904.5
      Preeclampsia35,1458292.4
      Threatened preterm labor35,14817925.1
      Premature delivery (< 37 wks)35,23525307.2
      Gestation at delivery (d)35,255274.813.1277.013.0
      Birthweight (g)35,2553345.3541.83373.7630.5
      Sex35,202
       Male17,91150.9
       Female17,29149.1
      Delivery mode35,117
       Vaginal unassisted24,06768.5
       Vaginal operative26397.5
       Cesarean841124.0
      Stillbirth35,235700.2
      Neonatal death35,235310.1
      BMI, body mass index; IQR, interquartile range.
      Gardosi. Customized standard to assess fetal growth. Am J Obstet Gynecol 2009.
      Coefficients for customized birthweight centiles were derived according to methods described previously,
      • Gardosi J.
      • Mongelli M.
      • Wilcox M.
      • Chang A.
      An adjustable fetal weight standard.
      using multiple regression with covariate selection by backward elimination and input and removal significance levels of 0.05. Both physiological and pathological variables were used to calculate the respective coefficients. However, only the nonpathological variables are used in an additive model to adjust the predicted term weight and growth potential. In addition to maternal height and weight, we also included low and high body mass index (BMI) as pathological categories, based on less than the 10th and greater than the 90th centiles of the BMI distribution in this population (Table 2).
      TABLE 2Coefficients from multiple regression model
      VariableCoefficientSEP value
      Gestational age (from 280 d)
       Linear term22.860.45< .001
       Quadratic term-0.3110.026< .001
       Cubic term-0.0070.002< .001
      Sex
       Male66.02.2< .001
       Female-66.02.2< .001
      Maternal height (from 163 cm)
       Linear6.3980.434< .001
       Cubic-0.0030.001.0015
      Maternal weight (from 64 kg)
       Linear7.5780.338< .001
       Quadratic-0.0870.015< .001
       Cubic0.00050.0002< .0041
      Ethnic origin
       African American-161.011.0< .001
       Hispanic-38.65.7< .001
       Other-140.839.6< .001
      Parity
       Para 196.25.2< .001
       Para 2121.96.9< .001
       Para 3125.910.7< .001
       Para ≥ 4122.714.6< .001
      Past history
       Miscarriage12.95.1.0118
       Preterm delivery 1-55.110.4< .001
       Preterm delivery ≥ 2-77.629.1.0077
      Smoking (n/d)
       1-9-99.212.8< .001
       10-19-174.920.2< .001
       ≥ 20-246.345.8< .001
      Low BMI (< 19.83 kg/m2)
      Limits represent 10th and 90th centiles for BMI in this population.
      -20.49.3.0283
      High BMI (> 31.71 kg/m2)
      Limits represent 10th and 90th centiles for BMI in this population.
      -63.412.4< .001
      Diabetes241.726.1< .001
      Antepartum hemorrhage-41.218.7.0271
      Pregnancy induced hypertension-26.811.4.0184
      Preeclampsia-60.718.3< .001
      Analysis centered on 280 days' gestation, for a standard mother (height 163 cm, weight 64 kg at first visit, para 0, European origin). Coefficients of model: constant, 3453.4g; SE, 382.6; R2, 0.27.
      BMI, body mass index, SE, standard error.
      Gardosi. Customized standard to assess fetal growth. Am J Obstet Gynecol 2009.
      a Limits represent 10th and 90th centiles for BMI in this population.
      A standard ANOVA test yielded P < .001, giving the model as significant. Test of residuals confirmed assumptions of normality, linearity and uniformity of variance.
      To allow comparison with previous studies,
      • Gardosi J.
      • Mongelli M.
      • Wilcox M.
      • Chang A.
      An adjustable fetal weight standard.
      • McCowan L.
      • Stewart A.W.
      • Francis A.
      • Gardosi J.
      A customised birthweight centile calculator developed for a New Zealand population.
      • Mongelli M.
      • Figueras F.
      • Francis A.
      • Gardosi J.
      A customised birthweight centile calculator developed for an Australian population.
      the birthweight constant was calculated for a gestation length of 280 days and a standard mother, defined as of Anglo-European origin, in her first pregnancy, height 163 cm, weight 64 kg, and the baby's sex unspecified or neutral.
      All analyses were performed using either SPSS (version 14.0; SPSS Inc, Chicago, IL) or Excel 2003 SP3 (Microsoft, Redmond, WA).

      Results

      Table 1 describes the characteristics of the study population and lists the covariates entered into the multivariate model. The results of the multiple regression analysis are presented in Table 2, listing coefficients for the significant variables together with their standard error and P value. The overall R2 of the model was 0.27. Maternal height, weight at first visit, parity, and the baby's sex were significant variables. For ethnic origin, only African American, Hispanic, and a miscellaneous “other” group reached significance.
      Several pathological factors were significant, including past history and complications during the index pregnancy (Table 2). Both high and low BMI had a negative effect on birthweight, with high BMI the stronger factor.
      In Table 3, the physiological coefficients for a standard mother are compared with previous analyses in England, New Zealand, and Australia
      • Gardosi J.
      • Mongelli M.
      • Wilcox M.
      • Chang A.
      An adjustable fetal weight standard.
      • McCowan L.
      • Stewart A.W.
      • Francis A.
      • Gardosi J.
      A customised birthweight centile calculator developed for a New Zealand population.
      • Mongelli M.
      • Figueras F.
      • Francis A.
      • Gardosi J.
      A customised birthweight centile calculator developed for an Australian population.
      and show close similarities, with no statistical difference in any of the categories with the exception of some variation in the higher parity groups. The constants (ie, the potential term birthweights for a standard mother) were also remarkably similar: 3453.4 g (United States), 3455.6 g (United Kingdom), 3464.4 g (New Zealand), and 3463.6 g (Australia).
      TABLE 3Comparison of coefficients for standard mother
      VariableUnited States (current study)EnglandNew ZealandAustralia
      Constant3453.43455.63464.43463.6
      SE of model382.6389.0420.4410.4
      Gestational age (from 280 d)
       Linear term22.8620.719.519.1
       Quadratic term-0.311-0.213-0.28-0.34
       Cubic term-0.007-0.000170.0006
      Sex
       Male66.048.957.766.9
       Female-66.0-48.9-57.7-66.9
      Maternal height (from 163 cm)
       Linear6.46.79.67.8
       Cubic-0.003
      Maternal weight (from 64 kg)
       Linear7.589.188.449.0
       Quadratic-0.087-0.151-0.114-0.15
       Cubic0.00050.0010.000650.001
      Parity
       Para 196.2101.9101.694.8
       Para 2121.9133.7101.8115.2
       Para 3125.9140.2123.3116.0
       Para ≥ 4122.7162.7175.599.2
      Ethnic origin
       African American-161.0
       African Caribbean-127.5
       African-218.5-297.4
       Hispanic-38.6
       Middle Eastern-89.9-110.0
       Bangladeshi-79.3
       Indian/Pakistani-162.0
       Indian-149.4-149.5
       Pakistani-187.3
       Chinese100.9
       Maori-66.8
       Samoan84.2
       Tongan124.1
       Other-140.8
      Standard mother is defined as being of European origin, height 163 cm, weight 64 kg, first pregnancy, with baby sex averaged between male and female; current study compared with previous findings from England,
      • Gardosi J.
      • Mongelli M.
      • Wilcox M.
      • Chang A.
      An adjustable fetal weight standard.
      New Zealand,
      • McCowan L.
      • Stewart A.W.
      • Francis A.
      • Gardosi J.
      A customised birthweight centile calculator developed for a New Zealand population.
      and Australia
      • Mongelli M.
      • Figueras F.
      • Francis A.
      • Gardosi J.
      A customised birthweight centile calculator developed for an Australian population.
      .
      SE, standard error.
      Gardosi. Customized standard to assess fetal growth. Am J Obstet Gynecol 2009.

      Comment

      This analysis shows that in an American population, birthweight varies with similar physiological factors to those found in maternity populations elsewhere,
      • Gardosi J.
      • Mongelli M.
      • Wilcox M.
      • Chang A.
      An adjustable fetal weight standard.
      • McCowan L.
      • Stewart A.W.
      • Francis A.
      • Gardosi J.
      A customised birthweight centile calculator developed for a New Zealand population.
      • Mongelli M.
      • Figueras F.
      • Francis A.
      • Gardosi J.
      A customised birthweight centile calculator developed for an Australian population.
      including maternal height, weight, parity, and ethnic origin as well as gestational age and sex of the baby. Furthermore, the magnitude of effect of these variables on birthweight is similar, suggesting that they apply universally. Once such variables are adjusted for, and pathological factors excluded, comparisons across geographical boundaries are possible. Our results show that populations of similar Anglo-European origin show striking similarities in the constant (ie, the birthweight expected for a standard size mother at the end of a normal pregnancy) (Table 3).
      The standard mother principle allows interesting international comparisons of the effect of ethnicity on birthweight. In the current study, babies of African American mothers weigh 161.0 g less, which lies between the coefficients for African Caribbean (-127.5) and sub-Saharan African (-218.5g) ethnic groups in England.
      • Gardosi J.
      • Mongelli M.
      • Wilcox M.
      • Chang A.
      An adjustable fetal weight standard.
      Babies of Hispanic mothers were found to weigh 38.6 g less but neither of the other 2 ethnicities specified in this database, Asian and Native American, showed statistically significant differences. This could be a result of too few cases, heterogeneity, a weak effect on birthweight, or a combination of factors.
      A recent study
      • Bukowski R.
      • Uchida T.
      • Smith G.C.S.
      • et al.
      Individualised norms of optimal fetal growth.
      using the same database but different methodology found that babies of Asian mothers weigh 73.8 g less than the white (European) reference. However, this is likely to be a composite figure of a heterogeneous Asian group. We know from other studies (Table 3) that south Asian babies can weigh up to 187 g less and Chinese babies 100 g more than their European counterparts.
      Further work is needed to improve our understanding of the effect of ethnicity on birthweight. First, the database here was not large enough to assess growth in other ethnic groups living in America. Second, there are intergenerational differences, and the respective contributions of ethnicity and maternal size on increased birthweight in second or third generation migrants has yet to be quantified. Third, more information is required to study potential hidden confounders such as social deprivation that may affect fetal growth. However, differences in weight for gestation between ethnic groups persist when low-risk groups are compared, after exclusion of disproportionate social and other factors.
      • Alexander G.R.
      • Kogan M.D.
      • Himes J.H.
      • Mor J.M.
      • Goldenberg R.
      Racial differences in birthweight for gestational age and infant mortality in extremely low risk US populations.
      The database provides information about several pathological factors (Table 2). Their effect on birthweight can be assessed better when physiological factors are also included in the multivariate analysis: this results in a predicted term weight that is free from such pathology and thus represents an optimal growth potential. Even though adverse factors such as past obstetric history or maternal smoking are known at the beginning of pregnancy, they are not used to calculate the term optimal weight (TOW) to which a baby is expected to grow: the aim of customized growth curves is not to predict the actual birthweight but to produce the optimal standard which is attainable, and against which the effects of any pathology can then be measured.
      The adverse effect of smoking on fetal growth and birthweight is well known.
      • Wu Wen S.
      • Goldenberg R.L.
      • Cutter G.R.
      • et al.
      Smoking, maternal age, fetal growth, and gestational age at delivery.
      The current analysis allows this effect to be quantified after adjustment for other maternal and pregnancy variables (Table 2). Comparison with other analyses using similar standardization and multiple regression techniques again shows remarkable similarities and demonstrates a dose-dependent deficit, which increases to about 250 g at term (current study: 246.3 g; England: 246.0 g2; Spain: 256.2 g.
      • Figueras F.
      • Meler E.
      • Iraola A.
      • et al.
      Customised birthweight standards for a Spanish population.
      The inclusion of high and low BMI in the multiple regression allowed us to assess the effect of pathological BMI in addition to the effect that maternal height and weight exert within normal BMI limits. Because BMI has been increasing in many societies in modern times, it is uncertain what the correct limits are today to distinguish normal from abnormal. Pragmatically, we used the 10th and 90th centiles for BMI in this population sample, which ensured that we had sufficient numbers in each category for the regression analysis. The results show first that mothers who have a low BMI and are potentially malnourished may have smaller babies, even after adjusting for maternal size and other constitutional variables, although the effect is relatively small (-20.4 g).
      At the other end of the BMI spectrum, however, the finding of a more substantial, negative effect (-63.4 g) of obesity on birthweight seems at first surprising and contradicts previous observations that high maternal weight protects the baby from being small for gestational age (SGA).
      • Cnattingius S.
      • Bergstrom R.
      • Lipworth L.
      • Kramer M.S.
      Prepregnancy weight and the risk of adverse pregnancy outcomes.
      However, we believe that this finding demonstrates the value of adjusting the definition of the SGA limit according to physiological variables, in order to better identify pathology. Recent preliminary analysis of a large Swedish database has shown that application of customized centiles in a population of large BMI mothers helps to identify a group of SGA babies not recognised by population centiles, which are at significantly increased risk of stillbirth.
      • Gardosi J.
      • Clausson B.
      • Francis A.
      The use of customised versus population-based birthweight standards in predicting perinatal mortality.
      The more pathological factors are identified and recorded in a database, the more likely that a term optimal weight can be determined that is free from all such pathology. To test their effect, we ran the multiple regression without pathological variables and included physiological factors plus smoking only. This resulted in a minimal difference in the constant (the birthweight at term expected for a standard mother), from 3453.4 (Table 2) to 3452.4, and an SE of the model that increased slightly, from 382.6 to 385.0.
      In practice, pathological factors may have negative or positive effects on birthweight that may balance out or, with the exception of smoking, do not occur frequently enough in an unselected population to affect the constant. It also suggests that physiological factors act on birthweight independently of pathology, and it is valid to compare databases, even though they contain pathological variables to a different degree (Table 3).
      Despite adjustment for many factors shown to affect birthweight, the R2 of the model seems modest at 0.27, consistent with other studies.
      • McCowan L.
      • Stewart A.W.
      • Francis A.
      • Gardosi J.
      A customised birthweight centile calculator developed for a New Zealand population.
      The R2 is 0.16 if only gestational age is adjusted and increases only slightly to 0.18 if fetal sex is added. Our model therefore represents an improvement in R2 from 0.16 to 0.27 (ie, 68% increase) over the use of birthweight-for-gestational age centiles only.
      The reason for the overall low correlation as expressed by R2 may be because of other factors affecting birthweight that we do not know about. However, it would be difficult to imagine what additional physiological or pathological variables could raise the correlation and predictive value much higher. We suggest that it is more likely that birthweight, as other biological measures, is subject to considerable random variation, whereas only the systematic factors can be predicted in any population sample. This is consistent with the finding that the SE in our model, with all variables included, is 382.6 g (Table 2), smaller than if only gestational age is adjusted for (SE 410.5 g).
      Whereas random variation relates generally to the whole birthweight distribution, it mostly affects values around the mean, where most measurements are, and where small differences have little clinical significance. A systematic shift, on the other hand, would be most notable at the extremes such as the 10th and 90th centile limits defining small and large for gestational age, respectively. The Figure shows the shift across these limits, which would result from variations in the expected mean birthweight in either direction, in a population with an SE of 382.6 g. For example, for a mother who is 10 cm taller and 10 kg heavier than average, her BMI would still be normal, but the expected birthweight would be about (10 × 6.4 g) + (10 × 7.6 g) = 140 g heavier than that of a standard size mother (Table 2). According to the Figure, a 140 g shift would mean that 50% of babies who should be considered SGA would not be identified as such, if adjustment for maternal stature is not used.
      Figure thumbnail gr1
      FIGUREBirthweights reclassified as SGA
      Proportion of birthweights reclassified as SGA (less than the 10th centile) or LGA (greater than the 90th centile) because of adjustment of the term optimal weight (ΔTOW) in a standardised normal distribution with standard error 382.6 (). The example used in the text is marked (ΔTOW = 140 g).
      LGA, large for gestational age; SGA, small for gestational age.
      Gardosi. Customized standard to assess fetal growth. Am J Obstet Gynecol 2009.
      In conclusion, analysis of the main physiological variables affecting fetal growth shows that they have a similar effect on normal fetal growth in this population as elsewhere, further adding to the concept that the best local and international standard for fetal growth is one that is individually adjustable. Larger multiethnic American databases will provide further opportunity to study the influences on birthweight in different parts of the population. In the meantime, the coefficients derived in this analysis have been added as an American version of the Gestation Related Optimal Weight (GROW) program (available for free download from www.gestation.net).
      Gestation Network
      GROW: Gestation Related Optimal Weight Customised growth chart software versions 5.x-7.x, 2000-2009.
      The program includes software to generate customized growth charts for individualized prediction of fetal growth potential for prospective surveillance, and a centile calculator for the retrospective assessment of birthweight.

      References

        • Gardosi J.
        Customised fetal growth standards: rationale and clinical application.
        Sem Perinatol. 2004; 28: 33-40
        • Gardosi J.
        • Mongelli M.
        • Wilcox M.
        • Chang A.
        An adjustable fetal weight standard.
        Ultrasound Obstet Gynecol. 1995; 6: 168-174
        • de Jong C.L.D.
        • Gardosi J.
        • Dekker G.A.
        • Colenbrander G.J.
        • van Geijn H.P.
        Application of a customised birthweight standard in the assessment of perinatal outcome in a high risk population.
        BJOG. 1998; 105: 531-535
        • Clausson B.
        • Gardosi J.
        • Francis A.
        • Cnattingius S.
        Perinatal outcome in SGA births defined by customised versus population-based birthweight standards.
        BJOG. 2001; 108: 830-834
        • McCowan L.
        • Harding J.E.
        • Stewart A.W.
        Customised birthweight centiles predict SGA pregnancies with perinatal morbidity.
        BJOG. 2005; 112: 1026-1033
        • Groom K.M.
        • Poppe K.K.
        • North R.A.
        • McCowan M.E.
        Small-for-gestational age infants classified by customized or population birthweight centiles: impact of gestational age at delivery.
        Am J Obstet Gynecol. 2007; 197 (239e1-5)
        • Figueras F.
        • Figueras J.
        • Meier E.
        • et al.
        Customised birthweight percentiles accurately predict perinatal morbidity.
        Arch Dis Child Fetal Neonatal Ed. 2007; 92: 277-280
      1. Royal College of Obstetricians and Gynaecologists. The investigation and management of the small-for-gestational age fetus. Royal College of Obstetricians and Gynaecologists Green Top Guideline 2002 (#31).

        • Resnik R.
        One size does not fit all.
        Am J Obstet Gynecol. 2007; 197: 221-222
        • McCowan L.
        • Stewart A.W.
        • Francis A.
        • Gardosi J.
        A customised birthweight centile calculator developed for a New Zealand population.
        Aust N Z J Obstet Gynaecol. 2004; 44: 428-431
        • Ego A.
        • Subtil D.
        • Grange G.
        • et al.
        Customised versus population-based birth weight standards for identifying growth restricted infants: French multicenter study.
        Am J Obstet Gynecol. 2006; 194: 1042-1049
        • Figueras F.
        • Meler E.
        • Iraola A.
        • et al.
        Customised birthweight standards for a Spanish population.
        Eur J Obstet Gynaecol Reprod Biol. 2008; 136: 20-24
        • Mongelli M.
        • Figueras F.
        • Francis A.
        • Gardosi J.
        A customised birthweight centile calculator developed for an Australian population.
        Aust N Z J Obstet Gynaecol. 2007; 47: 128-131
        • Malone F.D.
        • Canick J.A.
        • Ball R.H.
        • et al.
        First-trimester or second-trimester screening, or both, for Down's syndrome.
        N Engl J Med. 2005; 353: 2001-2011
        • Bukowski R.
        • Uchida T.
        • Smith G.C.S.
        • et al.
        Individualised norms of optimal fetal growth.
        Obstet Gynecol. 2008; 111: 1065-1076
        • Alexander G.R.
        • Kogan M.D.
        • Himes J.H.
        • Mor J.M.
        • Goldenberg R.
        Racial differences in birthweight for gestational age and infant mortality in extremely low risk US populations.
        Pediatr Perinatal Epidemiol. 1999; 13: 205-217
        • Wu Wen S.
        • Goldenberg R.L.
        • Cutter G.R.
        • et al.
        Smoking, maternal age, fetal growth, and gestational age at delivery.
        Am J Obstet Gynecol. 1990; 162: 53-58
        • Cnattingius S.
        • Bergstrom R.
        • Lipworth L.
        • Kramer M.S.
        Prepregnancy weight and the risk of adverse pregnancy outcomes.
        N Engl J Med. 1998; 338: 147-152
        • Gardosi J.
        • Clausson B.
        • Francis A.
        The use of customised versus population-based birthweight standards in predicting perinatal mortality.
        BJOG. 2007; 114: 1301-1302
        • Gestation Network
        GROW: Gestation Related Optimal Weight.
        (Accessed May 12, 2009)