[Establishment method and significance of birthweight curve and reference in single center].

Objective: To establish neonatal birthweight percentile curves based on single-center cohort database using different methods, compare them with the current national birthweight curves and discuss the appropriateness and significance of single-center birthweight standard. Methods: Based on a prospective first-trimester screening cohort at Nanjing Drum Tower Hospital from January 2017 to February 2022, the generalized additive models for location, scale and shape (GAMLSS) and semi-customized method were applied to generate local birthweight percentile curves (hereinafter referred to as the local GAMLSS curves, semi-customized curves) for 3 894 cases who were at low risk of small for gestation age (SGA) and large for gestation age (LGA). Infants were categorized as SGA (birth weight<10th centile) by both semi-customized and local GAMLSS curves, semi-customized curves only, or not SGA (met neither criteria). The incidence of adverse perinatal outcome between different groups was compared. The same method was used to compare the semi-customized curves with the Chinese national birthweight curves (established by GAMLSS method as well, hereinafter referred to as the national GAMLSS curves). Results: (1) Among the 7 044 live births, 404 (5.74%, 404/7 044), 774 (10.99%, 774/7 044) and 868 (12.32%, 868/7 044) cases were diagnosed as SGA according to the national GAMLSS curves, the local GAMLSS curves and the semi-customized curves respectively. The birth weight of the 10th percentile of the semi-customized curves was higher than that of the local GAMLSS curves and the national GAMLSS curves at all gestational age. (2) When comparing semi-customized curves and the local GAMLSS curves, the incidence of admission to neonatal intensive care unit (NICU) for more than 24 hours of infants identified as SGA by semi-customized curves only (94 cases) and both semi-customized and local GAMLSS curves (774 cases) was 10.64% (10/94) and 5.68% (44/774) respectively, both significantly higher than that in non SGA group [6 176 cases, 1.34% (83/6 176); P<0.001]. The incidence of preeclampsia, pregnancy<34 weeks, and pregnancy<37 weeks of infants identified as SGA by the semi-customized curves only and both semi-customized and local GAMLSS curves was 12.77% (12/94) and 9.43% (73/774), 9.57% (9/94) and 2.71% (21/774), 24.47% (23/94) and 7.24% (56/774) respectively, which were significantly higher than those of the non SGA group [4.37% (270/6 176), 0.83% (51/6 176), 4.23% (261/6 176); all P<0.001]. (3) When comparing semi-customized curves and the national GAMLSS curves, the incidence of admission to NICU for more than 24 hours of infants identified as SGA by semi-customized curves only (464 cases) and both semi-customized and national GAMLSS curves (404 cases) was 5.60% (26/464) and 6.93% (28/404) respectively, both significantly higher than that in non SGA group [6 176 cases, 1.34% (83/6 176); all P<0.001]. The incidence of emergency cesarean section or forceps delivery for non-reassuring fetal status (NRFS) in infants identified as SGA by semi-customized curves only and both semi-customized and national GAMLSS curves was 4.96% (23/464) and 12.38% (50/404), both significantly higher than that in the non SGA group [2.57% (159/6 176); all P<0.001]. The incidence of preeclampsia, pregnancy<34 weeks, and pregnancy<37 weeks in the semi-customized curves only group and both semi-customized and national GAMLSS curves group was 8.84% (41/464) and 10.89% (44/404), 4.31% (20/464) and 2.48% (10/404), 10.56% (49/464) and 7.43% (30/404) respectively, all significantly higher than those in the non SGA group [4.37% (270/6 176), 0.83% (51/6 176), 4.23% (261/6 176); all P<0.001]. Conclusion: Compared with the national GAMLSS birthweight curves and the local GAMLSS curves, the birth weight curves established by semi-customized method based on our single center database is in line with our center' SGA screening, which is helpful to identify and strengthen the management of high-risk infants.目的： 依据本中心队列数据采用不同方法建立新生儿出生体重曲线，并与目前国内常用的协作网G曲线［采用基于偏度系数-中位数-变异系数法的广义可加模型（GAMLSS）拟合方法制定］进行比较，探讨适宜本中心的新生儿出生体重标准。 方法： 基于2017年1月1日至2022年2月28日于南京大学医学院附属鼓楼医院行妊娠早期胎儿结构超声筛查的前瞻性出生缺陷研究队列中的活产儿7 044例，对其中无小于胎龄儿（SGA）及大于胎龄儿（LGA）高危因素的低风险孕妇3 894例，分别采用GAMLSS法、基于低风险孕妇孕40周分娩新生儿平均出生体重和标准差采用半定制法建立本中心GAMLSS曲线（简称本中心G曲线）、本中心半定制曲线（简称本中心S曲线）。依据本中心S曲线、本中心G曲线对7 044例活产儿是否为SGA进行诊断，分为非SGA组（6 176例）、S较G额外诊断组（94例）、S及G均诊断组（774例），比较3组孕妇不良围产结局的发生率；依据本中心S曲线、协作网G曲线对7 044例活产儿是否为SGA进行诊断，分为非SGA组（6 176例）、S较协作网G额外诊断组（464例）、S及协作网G均诊断组（404例），比较3组孕妇不良围产结局的发生率。 结果： （1）7 044例活产儿中，依据协作网G曲线、本中心G曲线及本中心S曲线诊断SGA分别为404例（5.74%，404/7 044）、774例（10.99%，774/7 044）、868例（12.32%，868/7 044）。本中心S曲线新生儿出生体重的第10百分位数值在所有胎龄均高于本中心G曲线和协作网G曲线。（2）S较G额外诊断组、S及G均诊断组入住新生儿重症监护病房（NICU）时间>24 h的发生率［分别为10.64%（10/94）、5.68%（44/774）］均显著高于非SGA组［1.34%（83/6 176）；P均<0.001］；S较G额外诊断组、S及G均诊断组孕妇子痫前期（PE）的发生率分别为12.77%（12/94）、9.43%（73/774），孕34周前早产发生率分别为9.57%（9/94）、2.71%（21/774），孕37周前早产的发生率分别为24.47%（23/94）、7.24%（56/774），均显著高于非SGA组［分别为4.37%（270/6 176）、0.83%（51/6 176）、4.23%（261/6 176）；P均<0.001］。（3）S较协作网G额外诊断组、S及协作网G均诊断组新生儿入住NICU时间>24 h的发生率［分别为5.60%（26/464）、6.93%（28/404）］均显著高于非SGA组（P均<0.001）；S较协作网G额外诊断组、S及协作网G均诊断组孕妇因不确定的胎儿状态（NRFS）行紧急剖宫产术或产钳助产的发生率［分别为4.96%（23/464）、12.38%（50/404）］均高于非SGA组［2.57%（159/6 176）；P均<0.001］；S较协作网G额外诊断组、S及协作网G均诊断组孕妇PE的发生率分别为8.84%（41/464）和10.89%（44/404），孕34周前早产发生率分别为4.31%（20/464）和2.48%（10/404），孕37周前早产的发生率分别为10.56%（49/464）和7.43%（30/404），均显著高于非SGA组（P均<0.001）。 结论： 相较于协作网G曲线和本中心G曲线，依据本中心低风险活产儿出生体重，采用半定制法建立的本中心S曲线更有利于本中心SGA的筛查，有助于高危儿的识别和管理。.