文献速递儿童呼吸儿童感染

后疫情时期昆明RSV感染住院患儿临床特征及重症风险预测

[Clinical characteristics of hospitalized children with respiratory syncytial virus infection and risk prediction of severe illness during the post-COVID-19 era in Kunming].
PubMed 文献 · Zhonghua Er Ke Za Zhi · 2024-01-01

该回顾性研究比较了昆明市儿童医院2019年与2023年RSV感染住院患儿的临床特征,发现后疫情时期住院高峰为5至10月,喘息比例下降而发热、呼吸急促、抽搐、重症及多重混合感染比例升高。研究识别年龄、基础疾病、早产、NLR及混合感染为重症独立危险因素,并建立了预测性能良好的重症风险模型。

要点

  • 后疫情时期RSV住院高峰为5至10月,喘息减少但发热、呼吸急促、抽搐及重症比例增加。
  • 年龄、基础疾病、早产、NLR、混合感染是重症RSV感染的独立危险因素。
  • 基于五因素的重症预测模型AUC为0.85,最佳截断值0.21,灵敏度0.83,特异度0.80。

核心结论

  • 后疫情时期昆明地区儿童RSV感染住院高峰为5至10月,临床喘息减少但发热、呼吸急促、抽搐、重症及三重和四重混合感染比例增加。
  • 年龄、基础疾病、早产、NLR、混合感染是后疫情时期RSV感染导致重症的独立危险因素。
  • 建立的儿童RSV重症风险模型具有良好的预测能力,AUC为0.85,校准曲线P=0.319。

研究设计与样本

  • 回顾性研究收集昆明市儿童医院2019年1至12月和2023年1至12月RSV感染住院患儿共959例,男555例、女404例,就诊年龄15.4(7.3,28.5)月龄。
  • 2019年入院331例为疫情前组,2023年入院628例为后疫情组,基于倾向性评分匹配(PSM)比较两组病情严重度,匹配后各267例。
  • 组间比较采用χ2检验及Mann-Whitney U检验,多因素Logistic回归分析重症危险因素并建立预测模型,ROC曲线及校准曲线评价模型性能。

主要发现

  • 后疫情组住院高峰为2023年5至10月,各月分别为72例(11.5%)、98例(15.6%)、128例(20.4%)、101例(16.1%)、65例(10.4%)、61例(9.7%)。
  • PSM后后疫情组喘息比例低于疫情前组[109例(40.8%)比161例(60.3%),χ2=20.26,P<0.001]。
  • 后疫情组发热、呼吸急促、抽搐、重症比例及NLR、CRP、IL-6水平均高于疫情前组(均P<0.05)。
  • 后疫情组混合感染252例,三重和四重感染比例均高于疫情前组[59例(23.4%)比13例(12.1%)、30例(11.9%)比5例(4.7%),均P<0.05]。
  • 后疫情组混合感染常见病原依次为肺炎支原体56例(22.2%)、甲型流感病毒53例(21.0%)、鼻病毒48例(19.0%)、副流感病毒35例(13.9%)、腺病毒28例(11.1%)。

关键数值与推荐

  • 多因素Logistic回归示就诊年龄(OR=0.70,95%CI 0.62~0.78,P<0.001)、合并基础疾病(OR=10.03,95%CI 4.10~24.55,P<0.001)、早产(OR=6.78,95%CI 3.53~13.04,P<0.001)、NLR(OR=1.85,95%CI 1.09~3.15,P=0.023)、混合感染(OR=1.28,95%CI 1.18~1.38,P<0.001)与重症RSV感染独立相关。
  • 预测模型ROC曲线下面积0.85(95%CI 0.80~0.89,P<0.001),最佳预测截断值0.21,灵敏度0.83,特异度0.80。
  • 校准曲线显示预测概率与实际概率差异无统计学意义(P=0.319),模型校准度良好。

临床怎么用

  • 后疫情时期应关注5至10月RSV住院高峰,对发热、呼吸急促、抽搐及混合感染患儿提高重症警惕。
  • 对存在基础疾病、早产、低龄、NLR升高及混合感染的RSV患儿,可结合预测模型评估重症风险。
  • 预测模型最佳截断值0.21,灵敏度0.83、特异度0.80,可用于辅助识别需加强监护的重症高危患儿。

以上解读依据论文摘要(全文未开放获取或许可不允许翻译转载)。

原文

Objective: To compare the epidemiological and clinical characteristics of hospitalized children with respiratory syncytial virus (RSV) infection in Kunming among the pre-and post-COVID-19 era, and to establish a prediction model for severe RSV infection in children during the post-COVID-19 period. Methods: This was a retrospective study. Clinical and laboratory data were collected from 959 children hospitalized with RSV infection in the Department of Pulmonary and Critical Care Medicine at Kunming Children's Hospital during January to December 2019 and January to December 2023. Patients admitted in 2019 were defined as the pre-COVID-19 group, while those admitted in 2023 were classified as the post-COVID-19 group. Epidemiological and clinical characteristics were compared between the two groups. Subsequently, comparison of the clinical severity among the two groups was performed based on propensity score matching (PSM). Furthermore, the subjects in the post-COVID-19 group were divided into severe and non-severe groups based on clinical severity. Chi-square test and Mann-Whitney U test were used for pairwise comparison between groups, and multivariate Logistic regression was applied for the identification of independent risk factors and construction of the prediction model. The receiver operating characteristic (ROC) curve and calibration curve were employed to evaluate the predictive performance of this model. Results: Among the 959 children hospitalized with RSV infection, there were 555 males and 404 females, with an onset age of 15.4 (7.3, 28.5) months. Of which, there were 331 cases in the pre-COVID-19 group and 628 cases in the post-COVID-19 group. The peak period of RSV hospitalization in the post-COVID-19 group were from May to October 2023, and the monthly number of inpatients for each of these months were as follows: 72 cases (11.5%), 98 cases (15.6%), 128 cases (20.4%), 101 cases (16.1%), 65 cases (10.4%), and 61 cases (9.7%), respectively. After PSM for general data, 267 cases were matched in each group. The proportion of wheezing in the post-COVID-19 group was lower than that in the pre-COVID-19 group (109 cases (40.8%) vs. 161 cases (60.3%), χ2=20.26, P<0.001), while the incidences of fever, tachypnea, seizures, severe case, neutrophil-to-lymphocyte ratio (NLR), C-reactive protein and interleukin-6 levels were all higher than those in the pre-COVID-19 group (146 cases (54.7%) vs. 119 cases (44.6%), 117 cases (43.8%) vs. 89 cases (33.3%), 37 cases (13.9%) vs. 14 cases (5.2%), 69 cases (25.8%) vs. 45 cases (16.9%), 3.6 (1.9, 6.4) vs. 2.3 (1.8, 4.6), 9.9 (7.1, 15.2) vs. 7.8 (4.5, 13.9) mg/L, 20.5 (15.7, 30.4) vs. 17.2 (11.0, 26.9) ng/L, χ2=5.46, 6.36, 11.47, 6.42, Z=4.13, 3.06, 2.96, all P<0.05). There were 252 cases and 107 cases with co-infection in the post-and pre-COVID-19 groups, respectively. The proportion of triple and quadruple infection in the post-COVID-19 group was higher than that in the pre-COVID-19 group (59 cases (23.4%) vs. 13 cases (12.1%), 30 cases (11.9%) vs. 5 cases (4.7%), χ2=5.94, 4.46, both P<0.05). Among the 252 cases with co-infection in post-COVID-19 group, the most prevalent pathogens involving in co-infections, in order, were Mycoplasma pneumoniae 56 cases (22.2%), Influenza A virus 53 cases (21.0%), Rhinovirus 48 cases (19.0%), Parainfluenza virus 35 cases (13.9%), and Adenovirus 28 cases (11.1%).The result of multivariate Logistic regression showed that age (OR=0.70, 95%CI 0.62-0.78, P<0.001), underlying diseases (OR=10.03, 95%CI 4.10-24.55, P<0.001), premature birth (OR=6.78, 95%CI 3.53-13.04, P<0.001), NLR (OR=1.85, 95%CI 1.09-3.15, P=0.023), and co-infection (OR=1.28, 95%CI 1.18-1.38, P<0.001) were independently associated with the development of severe RSV infection in the post-COVID-19 group. The ROC curve of the prediction model integrating the above five factors indicated an area under the curve of 0.85 (95%CI 0.80-0.89, P<0.001), with an optimal cutoff of 0.21, a sensitivity of 0.83 and a specificity of 0.80. The calibration curve showed that the predicted probability in this model did not differ significantly from the actual probability (P=0.319). Conclusions: In the post-COVID-19 era in Kunming, the peak in pediatric hospitalizations for RSV infection was from May to October, with declined incidence of wheezing and increased incidence of fever, tachypnea, seizures, severe cases, and rates of triple and quadruple co-infections. Age, underlying diseases, premature birth, NLR, and co-infection were identified as independent risk factors for severe RSV infection in the post-COVID-19 period. In this study, a risk prediction model for severe pediatric RSV infection was established, which had a good predictive performance. 目的: 比较昆明地区新型冠状病毒感染疫情前及后疫情时期呼吸道合胞病毒(RSV)感染住院患儿流行病学及临床特征变化,并建立后疫情时期儿童重症RSV感染风险预测模型。 方法: 回顾性病例总结。收集昆明市儿童医院呼吸与危重症医学科2019年1至12月和2023年1至12月两个时期RSV感染住院的959例患儿的流行病学、临床及实验室特征。2019年入院患儿为疫情前组,2023年入院患儿为后疫情组。比较两组患儿流行病学及临床特征,同时基于倾向性评分匹配(PSM)比较两组患儿的病情严重度,并根据病情轻重进一步将初始后疫情组分为重症组和非重症组。组间比较采用χ2检验及Mann-Whitney U检验,多因素Logistic回归分析重症危险因素并建立预测模型,受试者工作特征(ROC)曲线及校准曲线评价模型性能。 结果: 959例RSV感染住院患儿中男555例、女404例,就诊年龄15.4(7.3,28.5)月龄,其中疫情前组331例、后疫情组628例。后疫情组住院高峰期为2023年5至10月,分别为72例(11.5%)、98例(15.6%)、128例(20.4%)、101例(16.1%)、65例(10.4%)、61例(9.7%)。PSM后,疫情前组和后疫情组各267例。后疫情组喘息比例低于疫情前组[109例(40.8%)比161例(60.3%),χ2=20.26,P<0.001],后疫情组发热、呼吸急促、抽搐、重症、中性粒与淋巴细胞比值(NLR)、C反应蛋白及白细胞介素6水平均高于疫情前组[146例(54.7%)比119例(44.6%)、117例(43.8%)比89例(33.3%)、37例(13.9%)比14例(5.2%)、69例(25.8%)比45例(16.9%)、3.6(1.9,6.4)比2.3(1.8,4.6)、9.9(7.1,15.2)比7.8(4.5,13.9)mg/L、20.5(15.7,30.4)比17.2(11.0,26.9)ng/L,χ2=5.46、6.36、11.47、6.42、Z=4.13、3.06、2.96,均P<0.05]。后疫情组及疫情前组中合并混合感染分别为252例和107例,后疫情组混合感染中三重、四重感染比例均高于疫情前组[59例(23.4%)比13例(12.1%)、30例(11.9%)比5例(4.7%),χ2=5.94、4.46,均P<0.05]。后疫情组252例混合感染常见于肺炎支原体56例(22.2%)、甲型流感病毒53例(21.0%)、鼻病毒48例(19.0%)、副流感病毒35例(13.9%)和腺病毒28例(11.1%)。多因素Logistic回归分析示后疫情组就诊年龄(OR=0.70,95%CI 0.62~0.78,P<0.001)、合并基础疾病(OR=10.03,95%CI 4.10~24.55,P<0.001)、早产(OR=6.78,95%CI 3.53~13.04,P<0.001)、NLR(OR=1.85,95%CI 1.09~3.15,P=0.023)、混合感染(OR=1.28,95%CI 1.18~1.38,P<0.001)均与重症RSV感染独立相关。ROC曲线结果显示基于独立危险因素建立的重症预测模型曲线下面积为0.85(95%CI 0.80~0.89,P<0.001),最佳预测截断值0.21,灵敏度0.83,特异度0.80;采用校准曲线对模型进行准确性验证,校准曲线接近理想曲线,具有良好校准度(P=0.319)。 结论: 后疫情时期昆明地区儿童RSV感染住院高峰为5至10月,临床上喘息减少,但发热、呼吸急促、抽搐、重症、三重及四重混合感染比例增加。年龄、基础疾病、早产、NLR、混合感染是后疫情时期RSV感染导致重症的独立危险因素。建立的儿童RSV重症风险模型其具有良好的预测能力。.

来源:Zhonghua Er Ke Za Zhi,查看原文。版权归原作者所有。

引用

  1. Liu HF, Feng QL, Huang RW, et al. [Clinical characteristics of hospitalized children with respiratory syncytial virus infection and risk prediction of severe illness during the post-COVID-19 era in Kunming]. Zhonghua Er Ke Za Zhi. 2024. PMID: 38527502. doi:10.3760/cma.j.cn112140-20240219-00109

供执业医师工作参考,不替代临床判断;数值与分级以指南原文为准。