Department of Pediatrics, School of Medicine, University of Alabama at Birmingham, Birmingham, Alabama;
Department of Pediatrics, School of Medicine, Wayne State University, Detroit, Michigan.
Pediatrics. 2021 Jun;147(6). doi: 10.1542/peds.2020-048678. Epub 2021 May 13.
Early determination of prognosis is important in neonates with hypoxic-ischemic encephalopathy (HIE). Our objective was to test scoring systems developed earlier (original scoring system) and develop new prognostic models.
Secondary analysis of data from the multicenter randomized controlled trial of longer, deeper, or usual care cooling in neonatal HIE (NCT01192776) that enrolled 364 neonates diagnosed with moderate or severe HIE. The primary outcome was death or moderate or severe disability at 18 to 22 months, and secondary outcome was death during initial hospitalization. Testing of early neurologic clinical examination (<6 hours of age) and the original scoring system for prognostic ability was done, followed by development of new scoring systems and classification and regression tree (CART) models by using early clinical variables (<6 hours of age).
For death or disability, the original scoring system correctly classified 75% (95% confidence interval: 69%-81%), whereas the new scoring system correctly classified 78% (73%-82%), and the CART model correctly classified 76% (72%-81%). Early neurologic clinical examination also had a correct classification rate of 76% (71%-80%). Depth and duration of cooling did not affect prediction. Only a few components of the early neurologic examination were associated with poor outcome. For death, the original scoring system correctly classified 72% (66%-77%), the new scoring system 68% (63%-72%), the new CART model 87% (83%-90%), and early neurologic evaluation 81% (77%-85%).
The 3 models (scoring system, CART, and early neurologic evaluation) were comparable in predicting death or disability. For in-hospital death, CART models were superior to scoring systems and early neurologic examination.
在患有缺氧缺血性脑病（HIE）的新生儿中，早期确定预后非常重要。我们的目的是测试早期开发的评分系统（原始评分系统）并开发新的预测模型。
对多中心随机对照试验中接受更长时间、更深程度或常规治疗的新生儿 HIE（NCT01192776）的数据进行二次分析，该试验纳入了 364 名被诊断为中重度 HIE 的新生儿。主要结局为 18 至 22 个月时死亡或中重度残疾，次要结局为住院期间死亡。对早期神经临床检查（<6 小时）和原始评分系统进行了预后能力的测试，随后使用早期临床变量（<6 小时）开发了新的评分系统和分类回归树（CART）模型。
对于死亡或残疾，原始评分系统正确分类 75%（95%置信区间：69%-81%），而新评分系统正确分类 78%（73%-82%），CART 模型正确分类 76%（72%-81%）。早期神经临床检查的正确分类率也为 76%（71%-80%）。冷却的深度和持续时间不影响预测。只有少数早期神经检查的组成部分与不良结局相关。对于死亡，原始评分系统正确分类 72%（66%-77%），新评分系统 68%（63%-72%），新 CART 模型 87%（83%-90%），早期神经评估 81%（77%-85%）。
3 种模型（评分系统、CART 和早期神经评估）在预测死亡或残疾方面具有可比性。对于院内死亡，CART 模型优于评分系统和早期神经检查。
