Distribution of Birth Weight for Gestational Age in Japanese Infants Delivered by Cesarean Section

Background Neonatal anthropometric charts of the distribution of measurements, mainly birth weight, taken at different gestational ages are widely used by obstetricians and pediatricians. However, the relationship between delivery mode and neonatal anthropometric data has not been investigated in Japan or other countries. Methods The subjects were selected from the registration database of the Japan Society of Obstetrics and Gynecology (2003–2005). Tenth centile, median, and 90th centile of birth weight by sex, birth order, and delivery mode were observed by gestational age from 22 to 42 weeks among eligible singleton births. Results After excluding 248 outliers and 5243 births that did not satisfy the inclusion criteria, 144 980 births were included in the analysis. The distribution of 10th centile curves was skewed toward lower birth weights during the preterm period among both first live births and second and later live births delivered by cesarean section. More than 40% of both male and female live births were delivered by cesarean section at 37 weeks or earlier. Conclusions The large proportion of cesarean sections influenced the skewness of the birth weight distribution in the preterm period.


INTRODUCTION
Neonatal anthropometric charts are based on the distribution of measurements, mainly birth weight, of neonates at different gestational ages. 1 The Japanese neonatal anthropometric charts, which were revised in 1995, 2 are widely distributed to Japanese obstetricians and pediatricians for managing pregnancy and newborns.
Because more than 10 years had passed since publishing the revised charts, the research committee of the Ministry of Health, Welfare, and Labour for Multicenter Benchmark Research on Neonatal Outcomes in Japan attempted to develop new anthropometric charts. Due to the small sample size, the 1995 charts only contained data classified by sex and birth order. Using the registration database of the Japan Society of Obstetrics and Gynecology (JSOG), which includes a large number of pregnant women and their babies, we attempted to construct charts by mode of delivery, ie, vaginal delivery and cesarean section, as well as sex and birth order. This delivery mode-specific chart is unique to Japan, as no such chart exists in other countries. [3][4][5][6][7] In this study, we describe the different birth-weight distributions by gestational age and mode of delivery and discuss the factors that influenced this distribution.

METHODS
JSOG manages a registration system for pregnant women and their infants. To construct new neonatal anthropometric charts, we collected data from 2003 to 2005 on gestational age, birth weight, sex, birth order, and information on complications of singleton births from this database. Because JSOG approved the use of their database for the purpose of creating new neonatal anthropometric charts, this study was not subject to institutional review. Stillborn infants and those with severe asphyxia (Apgar score of 0 at 1 and 5 minutes after delivery), hydrops, or malformations were excluded from the analysis. Infants with missing information on sex or gestational age were also excluded.
Regarding mode of delivery, 6 modes were reported in the registration database: natural vaginal delivery, vacuumassisted vaginal delivery, forceps-assisted vaginal delivery, elective cesarean section, emergency cesarean section, and others. Natural vaginal delivery, vacuum-assisted vaginal delivery, and forceps-assisted vaginal delivery were defined as vaginal delivery, and elective and emergency cesarean sections were defined as cesarean delivery in this study. Because more than 80% of births delivered by elective cesarean section were delivered from 37 to 41 gestational weeks and approximately 60% of those delivered by emergency cesarean section were delivered at 36 week or earlier, we combined these modes of delivery in the analysis. Pregnant women for whom mode of delivery was classified as "others" were excluded from this analysis.
First, 10th centile, median, and 90th centile of birth weight by sex and birth order (first live births or second and later live births) were observed by gestational age from 22 to 42 weeks among all eligible births. Then, a similar observation was made by delivery mode. The values obtained were then plotted and fitted to cubic curves using the least squares method.

RESULTS
During the study period, 147 medical facilities participated in the JSOG registration system, and 150 471 singleton births were reported to the registration database. A total of 5243 births were excluded from the analysis; thus, the study population comprised 145 228 births. Then, an additional 248 clinical outliers were excluded from this population. Consequently, 144 980 singleton births (74 740 boys and 70 240 girls) were included in the analysis (Table 1). Among  Second and later live births  Total  First live births  Second and later live births  Total   22  26  30  56  21  30  51  23  76  63  139  48  52  100  24  92  107  199  73  84  157  25  103  129  232  96  125  221  26  140  122  262  97  152  249  27  156  185  341  135  129  264  28  203  202  405  151  175  326  29  197  209  406  161  170 Figure 1 shows the birth weight distribution of singleton male infants by gestational age and birth order. The 10th centile curves of first live births and second and later live births were skewed to lower birth weights in the preterm period. When the birth weight distributions are classified by delivery mode, the 10th centile curves were skewed to lower birth weights among both first live births and second and later live births delivered by cesarean section ( Table 2, Figures 2  and 3). Coefficients of determination of all fitted curves were higher than 0.98, and the skewness was similar in 10th centile curves of birth weight of female infants who were delivered by cesarean section (data not shown).
Values were plotted and fitted to cubic curves by using the least squares method in Figures 2 and 3.

DISCUSSION
The 10th centile birth weight curves of Japanese singleton infants by gestational age were skewed toward low values during the preterm period. Cesarean section influenced this distribution because the proportion of births delivered by cesarean section was large during the preterm period, especially from 26 to 29 weeks. As curves for 24 to 26 gestational weeks appeared to be markedly skewed toward low values, there was a difference in gestational period between the area of the curves with the most skewness and that representing the largest proportion of cesarean sections. We were unable to determine the reason for this, as no country has included delivery mode in neonatal anthropometric charts. [2][3][4][5][6][7] Due to this uncertainty, the research committee for creating new neonatal anthropometric charts in Japan decided to eliminate cesarean deliveries from the charts. The new Japanese neonatal anthropometric chart will thus include only the birth weight of singleton infants born by vaginal delivery as standard curves, which are created after excluding factors related to fetal growth. We used the least squares methods to calculate the distribution of birth weights in this study because it was also employed in the revised charts in 1995. 2 The LMS (λ, μ, σ) method, however, will be used to create the new Japanese charts. 8 More than 40% of preterm infants were delivered by cesarean section in Japan. The proportion of cesarean sections was reported to be increasing among preterm infants in the United States. [9][10][11] The reasons for cesarean section are not available in the JSOG registration database; however, one known reason is fetal growth restriction (FGR), which is a decrease in the fetal growth rate that inhibits an infant from obtaining its complete genetic growth potential. FGR is caused by placental dysfunction or maternal complications such as pre-eclampsia. 12,13 It is associated with increased perinatal mortality and morbidity, as well as with increased risk of long-term complications such as impaired neurodevelopment, adult type 2 diabetes, and hypertension. 13 Ultrasonography techniques, including the non-stress test, biophysical profile scoring, and pulse Doppler methods, enable obstetricians to carefully evaluate fetal growth. 14 Due to these methods of fetal management, especially observation of growth in fetal head circumference, obstetricians are more likely to deliver fetuses with FGR during preterm in the event of non-reassuring fetal status. Indeed, approximately 80% of fetuses with FGR were delivered by cesarean section in European countries. 15 Cesarean section is also likely to be selected in cases of preterm premature rupture of membranes. 16 Because the JSOG database mainly includes tertiary hospitals, low birth weight infants were overrepresented in our study population as compared with the general population. It has been reported that whereas 8.5% of male births and 10.8% of female births were less than 2500 grams in the general population, approximately 25% of births were less than 2500 grams in some tertiary hospitals. [17][18][19] In addition, pregnant women with complications might be more likely to be admitted to, and undergo cesarean section in, tertiary hospitals. Due to this selection bias, 10th centile birth weights of cesarean section births may be less than those of the general population. The reliability of gestational age is the most important issue in creating neonatal anthropometric charts. We were unable to confirm whether gestational age was assessed by ultrasonography during first trimester among pregnant women registered in the JSOG system. Many Japanese clinics and hospitals that treat pregnant women have ultrasonography equipment. However, because estimation of gestational age by ultrasonography was not mentioned in Japanese guidelines for obstetrical practice, some facilities may have calculated gestational age by asking pregnant women about their last menstrual period. 20 In conclusion, the 10th centile birth weight curves of Japanese singleton infants delivered by cesarean section by gestational age were skewed toward low values during the preterm period. This might reflect the fact that fetuses with FGR were more likely to be delivered by cesarean section to prevent worsening fetal growth. Thus, the birth weights of singleton infants born by vaginal delivery were used as standard curves to develop new Japanese neonatal anthropometric charts.