目的 构建重度窒息新生儿院内死亡的风险预测模型。 方法 回顾性纳入2022年1月—2025年12月河南省20家三级医疗机构收治的836例重度窒息新生儿，根据结局分为死亡组（267例）和存活组（569例），分析患儿的出生情况、复苏手段、早期并发症等。采用多因素logistic回归筛选危险因素，构建列线图预测模型；绘制受试者操作特征曲线、校准曲线，并进行决策曲线分析，分别评价模型的区分度、校准准确度与临床应用价值。 结果 多因素logistic回归分析显示，乳酸水平高、出生体重低、5 min Apgar评分≤5分、使用肾上腺素复苏、使用扩容药物复苏、新生儿持续性肺动脉高压、肺出血、休克、Ⅲ/Ⅳ级脑室内出血是重度窒息患儿院内死亡的独立危险因素（P<0.05）。基于上述危险因素建立的列线图模型预测重度窒息患儿院内死亡风险的曲线下面积为0.795（95%CI：0.762~0.827）；校准曲线和决策曲线分析提示模型准确度高，临床适用性良好。 结论 重度窒息患儿院内死亡与出生体重、复苏方式及肺出血、休克等早期并发症密切相关，基于这些因素构建的风险模型具有良好的预测价值。

Objective To construct a risk prediction model for in-hospital mortality in neonates with severe asphyxia. Methods A total of 836 neonates with severe asphyxia admitted to 20 tertiary hospitals in Henan Province between January 2022 and December 2025 were retrospectively enrolled and divided into a death group (n=267) and a survival group (n=569) according to outcome. Birth characteristics, resuscitation measures, and early complications were analyzed. Multivariable logistic regression was used to identify risk factors, and a nomogram prediction model was developed. Receiver operating characteristic curves, calibration curves, and decision curve analysis were used to evaluate the model's discrimination, calibration accuracy, and clinical applicability. Results Multivariable logistic regression showed that high lactate level, low birth weight, a 5-minute Apgar score of ≤5, use of epinephrine, volume expansion during resuscitation, persistent pulmonary hypertension of the newborn, pulmonary hemorrhage, shock, and grade III/IV intraventricular hemorrhage were independent risk factors for in-hospital mortality in neonates with severe asphyxia (P<0.05). The nomogram based on these risk factors yielded an area under the curve of 0.795 (95%CI: 0.762-0.827). Calibration and decision curve analyses indicated that the model had high accuracy and good clinical applicability. Conclusions In-hospital mortality among severely asphyxiated neonates is closely associated with birth weight, resuscitation modalities, and early complications such as pulmonary hemorrhage and shock. The risk model constructed based on these factors shows good predictive value.