Luo Jun, Xu Fen, Lu Guang-Jin, Lin Hung-Chih, Feng Zhi-Chun
Department of Neonatology, BaYi Children's Hospital Affiliated to Clinical Medical College in Beijing Military General Hospital of Southern Medical University, Beijing, China; Department of Neonatology, Bao'an Maternity and Child Health Hospital of Shenzhen, Guangdong, China.
Department of Neonatology, Bao'an Maternity and Child Health Hospital of Shenzhen, Guangdong, China.
Early Hum Dev. 2014 Oct;90(10):557-64. doi: 10.1016/j.earlhumdev.2014.07.007. Epub 2014 Aug 7.
Relatively low serum mannose-binding lectin (MBL) levels and MBL genetic polymorphisms have been implicated as high risk factors for neonatal sepsis. However, different studies have reported conflicting findings and have generally been underpowered to exclude modest effect sizes.
Standard methodology of systematic reviews and meta-analyses was followed. PubMed, Embase, Cochrane, Web of Science, and Scopus databases were searched from January 1996 to December 2013. The eligible studies were collected and analyzed using Review Manager 5.2. Meta-Disc version 1.4 was used to describe and calculate sensitivity, specificity, summary receiver operator characteristic (SROC) curves and area under the curve. SROC curve analysis was used to summarize the overall performance. Funnel plots, Egger's test and Begg's test were used to investigate publication bias.
Seven studies addressing low MBL levels and MBL genetic polymorphisms (structure variant A/O, A/B of Exon1) were analyzed for susceptibility to neonatal sepsis, respectively. All of these control studies were of reasonable methodological quality. The pooled unadjusted odds ratio showed that low MBL levels were significantly associated with neonatal sepsis (P=0.0002; odds ratio=4.94, 95% confidence interval=2.16-11.29) and MBL genetic polymorphisms were also significantly associated with neonatal sepsis (P=0.03; odds ratio=1.41, 95% confidence interval=1.03-1.94). In subgroup analysis based on gestational age, increased risk was found in the preterm infants in the dominant model (RR 2.33, 95%CI 1.06-5.13, P=0.03). However, no association was observed for term infants in subgroup analysis. Additionally, the SROC curve of low MBL levels in the prediction of neonatal sepsis indicated a poor predictive ability. The area under curve was 0.80 (95% confidence interval=0.74-0.86).
Currently available evidence shows that neonates with low serum MBL levels are more than four times more likely to have neonatal sepsis compared to those with higher serum MBL levels. Neonates with MBL genetic polymorphisms are also susceptible to developing neonatal sepsis. However, a low serum MBL level was only of moderate value in detecting neonatal sepsis.
血清甘露糖结合凝集素（MBL）水平相对较低以及MBL基因多态性被认为是新生儿败血症的高风险因素。然而，不同研究报告的结果相互矛盾，而且通常样本量不足，无法排除中等效应量的影响。
遵循系统评价和荟萃分析的标准方法。检索了1996年1月至2013年12月期间的PubMed、Embase、Cochrane、Web of Science和Scopus数据库。使用Review Manager 5.2收集并分析符合条件的研究。使用Meta-Disc 1.4版描述并计算敏感性、特异性、汇总接受者操作特征（SROC）曲线及曲线下面积。采用SROC曲线分析总结总体表现。使用漏斗图、Egger检验和Begg检验来研究发表偏倚。
分别分析了7项关于低MBL水平和MBL基因多态性（外显子1的结构变体A/O、A/B）与新生儿败血症易感性的研究。所有这些对照研究的方法学质量都较为合理。汇总的未调整优势比显示，低MBL水平与新生儿败血症显著相关（P = 0.0002；优势比 = 4.94，95%置信区间 = 2.16 - 11.29），MBL基因多态性也与新生儿败血症显著相关（P = 0.03；优势比 = 1.41，95%置信区间 = 1.03 - 1.94）。在基于胎龄的亚组分析中，在显性模型中早产儿的风险增加（RR 2.33，95%CI 1.06 - 5.13，P = 0.03）。然而，亚组分析中足月儿未观察到相关性。此外，低MBL水平预测新生儿败血症的SROC曲线显示预测能力较差。曲线下面积为0.80（95%置信区间 = 0.74 - 0.86）。
现有证据表明，与血清MBL水平较高的新生儿相比，血清MBL水平低的新生儿患新生儿败血症的可能性要高出四倍多。具有MBL基因多态性的新生儿也易患新生儿败血症。然而，低血清MBL水平在检测新生儿败血症方面仅具有中等价值。
