儿童重症腺病毒性肺炎风险模型的构建及静脉注射免疫球蛋白治疗时机的前瞻性研究

Establishment of a risk model for severe adenovirus pneumonia and prospective study of the timing of intravenous immunoglobulin therapy in children

摘要:

目的 构建儿童重症腺病毒性肺炎（adenoviral pneumonia，AVP）的风险评分模型，并探讨静脉注射免疫球蛋白（intravenous immunoglobulin，IVIG）治疗重症AVP的合适时机。方法 回顾性分析1 046例AVP患儿的临床资料，运用多因素logistic回归法建立重症AVP风险预测模型。选取102例AVP患儿验证该模型。再前瞻性纳入75例年龄≤14岁且经该模型判定为有可能发展为重症AVP的患儿，按就诊顺序分为A、B、C 3组，每组各25例。A组：仅予对症支持治疗；B组：除对症支持治疗外，在进展为重症AVP前予IVIG治疗，1 g/(kg·d)，连续2 d；C组：除对症支持治疗外，在进展为重症AVP后予IVIG治疗，1 g/(kg·d)，连续2 d。比较A、B、C 3组治疗后的疗效及相关实验室指标。结果 年龄<18.5个月、基础病、热程>6.5 d、血红蛋白<84.5 g/L、谷丙转氨酶>113.5 U/L、合并细菌感染6个变量进入重症AVP风险预测模型。该模型的受试者操作特征曲线下面积为0.862，灵敏度为0.878，特异度为0.848。Hosmer-Lemeshow检验显示模型预测值和实际观测值间的一致性较好（P>0.05）。治疗后，B组发热时间和住院时间最短，住院费用最低，治疗有效率最高，并发症发生率最低，白细胞计数最低，白细胞介素（interleukin，IL）-1、IL-2、IL-6、IL-8、IL-10水平最低，肿瘤坏死因子α水平最高（P<0.05）。结论 该研究构建的重症AVP风险模型对预测重症AVP的发生具有良好的价值；在AVP患儿进展为重症前予IVIG治疗效果更好。

Abstract:

Objective To develop a risk prediction model for severe adenovirus pneumonia (AVP) in children, and to explore the appropriate timing for intravenous immunoglobulin (IVIG) therapy for severe AVP.Methods Medical data of 1 046 children with AVP were retrospectively analyzed, and a risk prediction model for severe AVP was established using multivariate logistic regression. The model was validated with 102 children with AVP. Then, 75 children aged ≤14 years who were considered at risk of developing severe AVP by the model were prospectively enrolled and divided into three groups (A, B and C) in order of visit, with 25 children in each group. Group A received symptomatic supportive therapy only. With the exception of symptomatic supportive therapy, group B received IVIG treatment at a dose of 1g/(kg·d) for 2 consecutive days, before progressing to severe AVP. With the exception of symptomatic supportive therapy, group C received IVIG treatment at a dose of 1 g/(kg·d) for 2 consecutive days after progressing to severe AVP. Efficacy and related laboratory indicators were compared among the three groups after treatment.Results Age<18.5 months, underlying diseases, fever duration >6.5 days, hemoglobin level <84.5 g/L, alanine transaminase level >113.5 U/L, and co-infection with bacteria were the six variables that entered into the risk prediction model for severe AVP. The model had an area under the receiver operating characteristic curve of 0.862, sensitivity of 0.878, and specificity of 0.848. The Hosmer-Lemeshow test showed good consistency between the predicted values and the actual observations (P>0.05). After treatment, group B had the shortest fever duration and hospital stay, the lowest hospitalization costs, the highest effective rate of treatment, the lowest incidence of complications, the lowest white blood cell count and interleukin (IL)-1, IL-2, IL-6, IL-8, IL-10 levels, and the highest level of tumor necrosis factor alpha (P<0.05).Conclusions The risk model for severe AVP established in this study has good value in predicting the development of severe AVP. IVIG therapy before progression to severe AVP is more effective in treating AVP in children.