If you don't remember your password, you can reset it by entering your email address and clicking the Reset Password button. You will then receive an email that contains a secure link for resetting your password
If the address matches a valid account an email will be sent to __email__ with instructions for resetting your password
Evidence of placental disease and poor perinatal outcome is more common in infants who are small by customized centiles, compared with population centiles. Because preterm births are more likely to be associated with placental pathology, a greater proportion of preterm births are likely to be small for gestational age (SGA) by customized centiles, compared with population centiles. Our objective was to compare the proportion of infants classified as SGA by customized and population birthweight centiles at different gestational ages at delivery.
This was a retrospective observational study of 17,855 nulliparous women delivering between 1992 and 1999 at National Women’s Hospital, Auckland, New Zealand. The proportion of SGA infants (birthweight less than the 10th centile) classified by customized and population birthweight centiles delivering at less than 34 weeks, 34-36+6 weeks, and 37 weeks or longer were compared.
A total of 1847 infants (10.3%) were customized SGA, compared with 2111 (11.8%) who were population SGA (relative risk [RR] 0.9, 95% confidence interval [CI] 0.8 to 0.9). Of preterm deliveries less than 34 weeks (n = 392), 29.1% were customized SGA and 17.1% were population SGA (RR 1.7, 95% CI 1.3 to 2.2). Of deliveries at 34-36+6 weeks (n = 946), 18.0% were customized SGA and 13.7% were population SGA (RR 1.3, 95% CI 1.1 to 1.6). The converse was observed at term (n = 16,517), 9.5% classified as customized SGA and 11.5% as population SGA (RR 0.82, 95% CI 0.77 to 0.87). Of all early preterm perinatal deaths (less than 34 weeks) 31 of 72 infants (43%) were customized SGA and 23 of 72 infants (32%) were population SGA. There were no perinatal deaths or deliveries less than 34 weeks in infants who were classified as SGA by population criteria only.
Customized centiles classified more infants as SGA, compared with population centiles, in preterm births but not for term births in nulliparous women.
The terms small for gestational age (SGA) and intrauterine growth restriction (IUGR) are often used interchangeably by obstetric and pediatric clinicians and in the obstetric and pediatric literature. Although there is considerable overlap between these conditions, these terms are not synonymous.
SGA has been defined in many ways but most commonly is used to describe a birthweight below the 10th centile for gestational age. This definition can be applied only after birth and will include some small but normal babies. The term IUGR suggests a fetus failing to reach its full growth potential because of placental dysfunction. These growth-restricted babies are often also small for gestational age and if they are diagnosed antenatally are likely to have abnormal umbilical and/or uterine artery Doppler waveforms.
Because failure to reach full fetal growth potential is difficult to measure and define, the term SGA is often used as an alternative. However, this can be misleading because many infants are small but not growth restricted and therefore not at risk of morbidity or mortality, and some infants are not conventionally SGA but are growth restricted.
Population birthweight centiles, which account for fetal sex and gestational age at delivery, are typically used to classify size at birth. However, customized birthweight centiles, which in addition adjust newborn size at birth for the maternal variables parity, ethnicity, height, and weight,
may provide a better estimate of infants that are SGA as well as growth restricted. Customized centiles identify a subgroup of infants born to bigger mothers, who are SGA by customized but not population centiles and have high rates of morbidity and mortality.
Conversely, approximately 30% of infants who are SGA by population centiles and born to small mothers are reclassified as normally grown.
Comparisons of customized and population birthweight centiles have demonstrated that infants classified as SGA by customized centiles have substantially higher rates of perinatal morbidity and mortality than those classified as SGA by population centiles only.
Prospective studies in which SGA was diagnosed before birth have shown that approximately two thirds of pregnancies with babies that are SGA by customized centiles have abnormal umbilical and/or uterine artery Doppler studies antenatally,
Approximately one third of all preterm deliveries are iatrogenic, and this is most commonly indicated for fetal or maternal disease related to placental insufficiency. High rates of SGA would be expected and have been demonstrated in this subgroup of iatrogenic preterm births.
Because both iatrogenic and spontaneous preterm births are thought to be linked with placental disease and because customized centiles are associated with abnormal umbilical and uterine Doppler waveforms as well as poor perinatal outcome, it may be expected that customized centiles will classify more preterm infants as SGA than population centiles. Although it has been demonstrated that customized SGA is more common preterm than at term, there are only very few data regarding the incidence of SGA classified by population and customized centiles in preterm births.
In addition, because customized centiles are more closely associated with placental disease, it may be expected that they will classify more infants born to mothers with hypertensive disease of pregnancy as SGA than are classified as SGA by population centiles.
The aims of our study in a population of nulliparous women were to compare the following: (1) the proportion of infants classified as SGA by customized and population birthweight centiles at different delivery gestations and (2) the proportion of infants classified as SGA by customized and population birthweight centiles in women with hypertensive disease of pregnancy at different delivery gestations.
Materials and Methods
Data was extracted from the Auckland Maternity Services Information System (AMSIS) database of births at National Women’s Hospital (Auckland, New Zealand) from 1992 to 1999 (n = 70,416). Data were collected prospectively for each pregnancy and birth and individual patient data checked for errors and completeness against each medical record after delivery. Validation of AMSIS has been reported previously in a study using the same data set.
Nulliparous women were included in this study. Multiple pregnancies (n = 2962), delivery less than 22 weeks’ gestation (n = 179), major fetal chromosomal or congenital anomalies (n = 794), transfer into the hospital after delivery (n = 1908), and women with underlying medical disorders of diabetes, chronic hypertension, renal disease, and autoimmune conditions (n = 4543) were excluded. These underlying medical conditions were excluded because the diagnosis of preeclampsia and gestational hypertension is less reliable made in this context. Women with incomplete data for maternal ethnicity, fetal sex, birthweight, gestational age at delivery (n = 21), or missing data for both maternal height and weight (n = 23,858) were also excluded. The final data set included 17,855 women.
Gestational age was calculated from menstrual history and, in the large majority, confirmed or adjusted by ultrasound scan performed less than 24 weeks. Ethnicity was self-determined by the mother. SGA was defined as a birthweight below the 10th centile for gestational age by sex-adjusted population centiles
(referred to as “customized SGA”) (www.gestation.net). Infants who were SGA by customized centiles but not population centiles were referred to as “customized SGA only,” and infants who were SGA by population centiles but not customized centiles were referred to as “population SGA only.” Because of the nature of this study, antenatal Doppler data were not available. Gestational hypertension (GH) was defined in a previously normotensive woman as a systolic blood pressure (BP) 140 mm Hg or greater and/or a diastolic BP 90 mm Hg or greater after 20 weeks’ gestation with no significant proteinuria. Preeclampsia was defined as GH with significant proteinuria
National High Blood Pressure Education Program Working Group on High Blood Pressure in Pregnancy Report of the National High Blood Pressure Education Program Working Group on High Blood Pressure in Pregnancy.
(greater than 300 mg per 24 hours or urine dipstick 2+ or greater). Perinatal death included fetal death from 20 weeks’ gestation and neonatal death in the first 28 days of life.
In women with data for maternal height but no weight (n = 728) or weight but no height (n = 9945), the corresponding missing value was imputed based on a multivariate normal distribution within each appropriate ethnic group. For each individual, the average value of 30 imputations was calculated and used as the final value for subsequent analyses.
All continuous data were normally distributed and expressed as mean values (SD). Comparisons were made between the proportion of SGA infants delivered using each method to calculate birthweight centiles. Comparisons among three gestational age at delivery categories were made: less than 34 weeks (reflecting severe and moderate prematurity), 34-36+6 weeks (reflecting mild prematurity), and 37 weeks or longer. Relative risk (RR) and associated 95% confidence intervals (CIs) were calculated. Pearson’s χ2 test was used to compare categorical data with Fisher’s exact test substituted when more than 1 cell within the comparison had a value less than 5. Statistical significance has been determined when P < .05.
A total of 17,855 nulliparous women were studied. Ethnicity comprised 9624 European (53.9%), 1575 Maori (8.8%), 1648 Chinese (9.2%), 1433 Samoan (8%), 797 Indian (4.4%), 712 Tongan (4%), and 2066 of other ethnic origins (11.6%). The mean maternal age was 27.3 years (range 13-46), and the mean body mass index was 25.1 (range 12.8 to 61.0). The overall rate of preterm delivery was 7.5% (n = 1338) with 2.2% (n = 392) delivering less than 34 weeks and 5.3% (n = 946) at 34-36+6 weeks. Hypertensive disease occurred in 1881 women (10.5%), 520 women (2.9%) had preeclampsia, and 1361 (7.6%) had GH.
Of the total population, 1847 (10.3%) infants were customized SGA and 2111 (11.8%) were population SGA (RR 0.9, 95% CI 0.8 to 0.9). A total of 1523 infants (8.5%) were SGA by both customized and population centiles, 324 infants (1.8%) were SGA by customized centiles alone, and 588 infants (3.3%) were SGA by population centiles alone.
The rates of SGA infants were higher by both criteria for preterm births, compared with term births. For customized birthweight centiles, 29.1% of women delivering at less than 34 weeks (RR 3.1, 95% CI 2.6 to 3.6) and 18.0% delivering at 34-36+6 weeks had an SGA infant (RR 1.9, 95% CI 1.6 to 2.2), compared with 9.5% delivering at term (RR 1.0). For population birthweight centiles, 17.1% of women delivering at less than 34 weeks (RR 1.5, 95% CI 1.2 to 1.8) and 13.7% at 34-36+6 weeks had an SGA infant (RR 1.2, 95% CI 1.0 to 1.4), compared with 11.5% at term (RR 1.0). Even though the large majority of SGA infants by either criteria was delivered at 37 weeks or longer, a greater proportion of all population SGA infants were delivered at term, compared with customized SGA infants, 90.6% and 84.6%, respectively (P < .0001) (see Figure 1).
The use of customized birthweight centiles increased the proportion of preterm infants classified as SGA, compared with those classified as SGA by population birthweight centiles. The converse was observed at term (Table 1). No infants delivered less than 34 weeks were classified as population SGA only, and only 5 at 34-36+6 weeks (0.5%) were population SGA only (Table 2). Of all early preterm perinatal deaths (n = 72 at less than 34 weeks), 31 (43%) were classified as customized SGA and 23 (32%) were classified as population SGA (P = .23). At 34-36+6 weeks and 37 weeks or longer, all perinatal deaths of SGA infants were SGA by both customized and population criteria. There were no perinatal deaths at any gestation in infants who were classified as population SGA only.
TABLE 1Incidence of customised and population SGA according to gestation at delivery
See text for relative risks comparing customised and population SGA for each gestational age at delivery group.
SGA both, SGA by customised and population centiles; customised SGA only, SGA by customised but not population centiles; population SGA only, SGA by population but not customized centiles; non-SGA both, not SGA by either customized or population centiles.
Women delivering preterm with preeclampsia or GH were more likely to have infants classified as SGA by customized rather than population birthweight centiles (Figure 2). Of all births to women with preeclampsia or GH, at less than 34 weeks (n = 110), 52 infants (47.3%) were customized SGA, compared with 38 infants (34.5%) who were population SGA (RR 1.4, 95% CI 1.0 to 1.9), and at 34-36+6 weeks (n = 215), 67 infants (31.2%) were customized SGA and 44 infants (20.5%) were population SGA (RR 1.5, 95% CI 1.1 to 2.1). There was little difference in the incidence of SGA by customized or population centiles in women with preeclampsia or GH delivering at term (n = 1556), 208 customized SGA (13.4%), and 219 population SGA (14.1%) (RR 1.0, 95% CI 0.8 to 1.1).
This is the first study in a large population to assess the rates of customized SGA with population SGA in relation to gestational age at delivery. Preterm infants were more likely to be customized SGA than population SGA, and the converse was true for infants born at term. Previous population-based studies comparing customized and population SGA have provided very limited data for rates of SGA by each criteria in relation to gestational age at delivery.
A single small study of 217 high-risk pregnancies reported rates of SGA infants by customized and population criteria in 26 women who delivered preterm. Of these, 54% delivered customized SGA infants, compared with 27% who delivered population SGA infants (P < .05).
The higher rate of customized SGA infants, compared with population SGA infants, in preterm births may reflect a closer association among customized SGA, placental disease, and true growth restriction. Further evidence to support this association is our finding of higher rates of customized SGA infants, compared with population SGA infants, in women with preeclampsia or GH delivering preterm but not in those with preeclampsia or GH at term. This is consistent with the concept that abnormal placentation is more common in early-onset preeclampsia, compared with term disease.
Fetal Medicine Foundation Second Trimester Screening Group An integrated model for the prediction of preeclampsia using maternal factors and uterine artery Doppler velocimetry in unselected low-risk women.
The methods used to generate population and customized centiles differ and may partly explain the large difference seen in the prevalence of SGA by each criterion in preterm deliveries. The sex-specific population centiles used in this study were calculated from birthweight data for all live births including preterm deliveries in a predominantly Caucasian population.
The number of infants at very preterm gestations were small, and the mean birthweights will have been influenced by the high proportion of pathological pregnancies and accompanying growth-restricted infants, which are delivered at preterm gestations. Population centiles are therefore likely to underestimate SGA at early gestations.
Customized centiles are gestation and fetal sex specific but also adjusted for maternal variables of ethnicity, parity, height, and weight. The centiles are calculated from an adjusted birthweight range expected at 40 weeks and extrapolated back using a standard, longitudinal, ultrasound-derived curve of intrauterine weight gain.
This method therefore estimates normal expected weights at premature gestations from subjects with ongoing pregnancies and results in higher mean weight at preterm gestations. Infants born prematurely are more likely to have associated pathology including placental disease and customized centiles are therefore likely to estimate a more accurate incidence of growth restriction at preterm gestations.
At term the use of customized birthweight centiles reduced the number of infants who were classified as SGA, compared with those classified as SGA by population centiles. Previous studies have already demonstrated that up to 30% of population SGA infants are reclassified as normally grown after application of customized centiles. This subgroup of small babies are likely to be small but normal infants because they have a very low incidence of abnormal umbilical and uterine Doppler waveforms
and very low rates of morbidity, suggesting they are not growth-restricted infants. The lower incidence of SGA infants at term using customized birthweight centiles is probably due to the reclassification of these constitutionally small normal babies once ethnicity, parity, and maternal size have been accounted for.
There were no perinatal deaths and very low rates of preterm birth (0.9%) in population SGA only infants. These findings are reassuring and are consistent with previous studies that have shown population SGA only infants have low rates of morbidity and mortality with perinatal outcomes similar to non-SGA infants and again provides reassurance that they are not growth restricted.
Our results again highlight the important contribution of maternal factors to fetal size. Therefore, when considering a diagnosis of growth restriction, either antenatally or after birth, these maternal factors must be considered. To illustrate this point, a male infant born at 40 weeks with a birthweight of 3200 g to a Caucasian mother of small stature (weight 50 kg, height 155 cm) has a customized birthweight centile of 39; however, a male infant born at 40 weeks to a larger Caucasian mother (weight 98 kg, height 180 cm) has a customized birthweight centile of 7, is SGA, and is likely to be growth restricted. Population growth standards would not have identified the second baby as being growth restricted.
Although the use of customized centiles may identify more growth-restricted infants than population centiles, they can not be applied until after birth and hence can not be used for clinical management antenatally. However, customized antenatal growth charts that incorporate the same maternal factors as those used in customized birthweight centiles are available (GROW, software available as free download, www.gestation.net). GROW generates an individualized fetal weight standard for each week of gestation from 24 weeks. A chart is generated for each woman with fundal height on 1 axis and estimated fetal weight on the other. Fundal height is plotted serially, and if an ultrasound is performed, estimated fetal weight is recorded on the graph. Use of GROW has been shown to increase antenatal detection of SGA babies,
Use of GROW has the potential to improve clinical decision making by both reducing the number of ultrasound scans required for small women with appropriately sized babies and prompting further investigation with Doppler studies for those with suspected growth restriction.
The use of customized birthweight centiles significantly increases the number of infants delivering preterm who are classified as SGA, compared with those classified SGA after application of population centiles. This may reflect the association of preterm birth with placental disease and growth restriction. The reduction in the number of infants classified as SGA by customized centiles at term is likely to be due to the exclusion of small but normal non–growth-restricted infants. Our findings reinforce the importance of using customized birthweight centiles when classifying infants as SGA, in particular when considering preterm births, because they are more likely to represent truly growth-restricted infants.
This work was supported in part by a grant from the Mercia Barnes Trust (to K.M.G.).
Cite this article as: Groom KM, Poppe KK, North RA, et al. Small-for-gestational-age infants classified by customized or population birthweight centiles: impact of gestational age at delivery. Am J Obstet Gynecol 2007;197:239.e1-239.e5.
We applaud Groom et al1 for their recent work to better understand the properties of customized birthweight percentiles. However, their conclusions that their findings “highlight the important contribution of maternal factors to fetal size” and “reinforce the importance of using customized birthweight centiles when classifying infants as small for gestational age (SGA), in particular when considering preterm births” deserve further consideration. It is important to recognize that there are 2 differences between the customized and population standards at preterm ages: (1) the customized percentiles include a regression-based adjustment for maternal factors that are associated with birthweight (ie, the “customization” to identify babies expected to be “small-but-healthy”), and (2) the customized percentiles use an intrauterine standard (via backward extrapolation of Hadlock's in utero standard2), in contrast to the population reference that is based on the distribution of live births.