目的 寻找预测极低出生体质量（VLBW）儿和超低出生体质量（ELBW）儿出院前结局的敏感的评分指标。方法 收集2018年7月1日至2021年1月31日收治的VLBW儿和ELBW儿的临床资料。评估新生儿急性生理学评分-Ⅱ（SNAP-Ⅱ）、新生儿急性生理学评分围生期补充-Ⅱ（SNAPPE-Ⅱ）、新生儿临床危险指数（CRIB）及新生儿危重病例评分（NCIS）对VLBW儿和ELBW儿出院前死亡、坏死性小肠结肠炎、支气管肺发育不良、肺出血、脑室旁白质软化及视网膜病变的预测价值。结果 共收治VLBW儿 491例,经筛选最终纳入223例VLBW儿（含56例ELBW儿）。无论VLBW儿或ELBW儿,存活组的NCIS评分高于死亡组,SNAP-Ⅱ、SNAPPE-Ⅱ以及CRIB评分均低于死亡组,差异均有统计学意义（P<0.05）。在VLBW儿中,经ROC曲线分析发现,CRIB评分预测VLBW儿死亡的AUC最大,AUC为0.888,95%CI为0.827~0.949,当CRIB评分为1.5时,其预测VLBW儿死亡的约登指数为0.672,灵敏度0.944,特异度0.728。在ELBW儿中,CRIB评分预测ELBW儿死亡的AUC最大,AUC为0.901,95%CI为0.819~0.982,当CRIB评分为5.5时,其预测ELBW儿死亡约登指数为0.673,灵敏度0.895,特异度0.778。四种评分中只有CRIB对VLBW儿支气管肺发育不良和视网膜病变有预测区分价值;四种评分对VLBW儿肺出血的发生均有预测价值,但只有CRIB对ELBW儿肺出血有一定预测区分度。结论 四种危重评分中CRIB对VLBW儿和ELBW儿死亡风险具有更敏感的预测价值。但各评分对VLBW儿和ELBW儿出院前其他结局的预测价值仍较为有限,亟需研发更为合适的评分体系。

Objective To find sensitive scoring indicators of pre-discharge outcomes in very low birth weight (VLBW) and extremely low birth weight (ELBW) infants. Methods Clinical data of VLBW and ELBW infants admitted from July 1, 2018 to January 31, 2021 were collected. The values of score for neonatal acute physiology version II (SNAP-Ⅱ), simplified version of the score for neonatal acute physiology perinatal extension (SNAPPE-Ⅱ), clinical risk index for babies (CRIB) and neonatal critical illness score (NCIS) were evaluated to predict predischarge death, necrotizing enterocolitis, bronchopulmonary dysplasia, pulmonary hemorrhage, periventricular leukomalacia, and retinopathy in VLBW and ELBW infants. Results A total of 491 VLBW infants were admitted and 223 VLBW infants (including 56 ELBW infants) were selected. In either VLBW or ELBW infants, NCIS score of survival group was higher than that of death group, SNAP-Ⅱ, SNAPPE-Ⅱ and CRIB scores were lower than those of death group, the differences were statistically significant (P<0.05). ROC curve analysis showed that CRIB score had the highest AUC (AUC=0.888, 95%CI: 0.827-0.949) in predicting VLBW death. When CRIB score was 1.5, the Jorden index of predicting VLBW death was 0.672, the sensitivity was 0.944 and the specificity was 0.728. The CRIB score had the highest AUC (AUC=0.901, 95%CI: 0.819-0.982) in predicting the death of ELBW infants. When CRIB score was 5.5, the Jorden index was 0.673, the sensitivity was 0.895 and the specificity was 0.778. Among the four scores, only CRIB had predictive value for bronchopulmonary dysplasia and retinopathy in VLBW infants. All the four scores had predictive value for pulmonary hemorrhage in VLBW infants, but only CRIB had certain predictive value for pulmonary hemorrhage in ELBW infants. Conclusions Among the four kinds of critical score, CRIB has more sensitive predictive value to death risk of VLBW and ELBW infants. However, the predictive value of each score for other predischarge outcomes in VLBW and ELBW infants is still limited, so it is urgent to develop a more appropriate scoring system.