Meta-analysis of prognostic tests in neonates over 35-week gestational age with hypoxic-ischemic encephalopathy
摘要
目的 评价临床上用于>35周胎龄新生儿缺氧缺血性脑病远期神经发育预后评价体系的预测价值.方法 检索时间区间从1980年1月至2012年12月.用“新生儿缺氧缺血性脑病”“远期预后”“预测”和“neonates with hypoxic-ischemic encephalopathy”“prognostic test”为检索词分别检索中国知网-中国学术期刊网络出版总库、万方数据库、Pubmed、Embase数据库中的相关文献.纳入研究需同时满足:(1)胎龄>35周患有围产期窒息或HIE的新生儿的预后研究;(2)神经发育结局至少随访到18月龄及以上;(3)神经发育结局明确预后良好或不良.采用RevMan 5.2软件对满足纳入标准的相关研究进行荟萃分析,评价各种检测方法的敏感度和特异度.结果 共检索出260篇相关文献,对20篇符合纳入标准的研究进行荟萃分析.符合条件的评价指标包括振幅整合脑电图(aEEG),磁共振成像(MRI)[T1/T2加权、弥散加权像(DWI)、表观弥散系数(ADC)和磁共振波谱(MRS)],颅脑超声和临床分度.结果显示,目前最有前景的是aEEG敏感度0.936(95%CI:0.897~0.964);特异度0.884(95%CI:0.836~0.921).影像学上,T1/T2加权像灵敏度0.913(95% CI:0.850~0.956),特异度0.630(95%CI:0.544~0.711);DWI灵敏度0.933(95% CI:0.817 ~0.956),特异度0.694(95% CI:0.546 ~0.817);ADC灵敏度0.778(95% CI:0.664~0.867),特异度0.971(95%CI:0.898 ~0.996);MRS灵敏度为0.918(95% CI:0.804 ~0.977),而特异度0.642(95% CI:0.515~0.755).颅脑超声灵敏度0.683(95%CI:0.519 ~0.819),特异度0.459(95% CI:0.295~0.631).临床分度灵敏度0.967(95%CI:0.924~0.989),特异度0.314(95%CI:0.241 ~0.394).结论 证据表明,aEEG,MRI(T1/T2、DWI、ADC、MRS),颅脑超声,临床分度对新生儿HIE远期预后均有重要作用,其中aEEG灵敏度和特异度都较高;临床分度的灵敏度较高;影像学中以T1/T2加权像、DWI以及MRS灵敏度较好,而ADC特异度较高.亚低温治疗可降低评价系统的预测作用.每种检查都有其优缺点,以临床神经评估为中心的包含各种脑监护脑影像技术的综合运用是目前判断预后的最佳办法.
Objective To explore the prognostic value of currently used clinical tests in neonatal patients over 35-week gestational age with perinatal asphyxia and hypoxic-ischemic encephalopathy (HIE).Methods " Neonate hypoxic-ischemic encephalopathy " and " prognostic test" were used as key words.Searches were made on the databases of PubMed,EMBASE,CNKI and WanFang for studies published between January 1980 and December 2012.Studies were included if they (1) evaluated outcome in over 35week gestational age neonates with perinatal asphyxia or HIE; (2) reported outcomes at minimal follow-up age of 18 months;(3) confirmed neurodevelopment with favorable or adverse outcomes.RcvMan 5.2 software 1.4 was used for meta-analysis.And the evaluation criteria of each study were employed to assess and analyze the results comprehensively.Results Among 260 relevant studies,20 were included.There are 7 tests contented our inclusion criteria:amplitude-integrated electroencephalography (aEEG),magnetic resonance imaging (MRI:T1/T2 weighted imaging,diffusion weighted imaging (DWI),apparent diffusion coefficient (ADC),magnetic resonance spectroscopy (MRS)),cerebral ultrasound (CUS) and HIE stage.The most promising tests were aEEG:sensitivity 0.936 (95% CI:0.897-0.964) ; specificity 0.884 (95% CI:0.836-0.921).In imaging,T1/T2-weighted MRI sensitivity 0.913 (95% CI:0.850-0.956),specificity 0.630(95 % CI:0.544-0.711) ; diffusion weighted MRI sensitivity 0.933 (95 % CI:0.817-0.956),specificity 0.694(95% CI:0.546-0.817) ; ADC sensitivity 0.778 (95% CI:0.664-0.867),specificity 0.971(95% CI:0.898-0.996) ; MRS sensitivity 0.918(95% CI:0.804-0.977),specificity 0.642(95% CI:0.515-0.755).CUS sensitivity 0.683 (95% CI:0.519-0.819),specificity 0.459 (95% CI:0.295-0.631).HIE stage sensitivity 0.967 (95% CI:0.924-0.989),specificity 0.314 (95% CI:0.241-0.394).Conclusions These evidence suggests an important role for aEEG,MRI (T1/T2-weight,DWI,ADC),CUS and HIE stage.The sensitivity and specificity of aEEG are both relatively high; HIE stage has good sensitivity; in imaging,T1/T2-weight,DWI and MRS yield better sensitivity while ADC has the best specificity.Mild hypothermia treatment reduces the predictive value of evaluation system.Each test has its advantages and disadvantages.Thus an integrated use of a variety of prognostic tests centering on clinical neurological assessment is optimal.