探讨新生儿C－反应蛋白(CRP)的影响因素。

选择2009年1月至2010年12月北京军区总医院八一儿童医院新生儿重症监护中心收治住院的新生儿中，入院时血清CRP >8 mg/mL的220例新生儿为研究对象，纳入研究组。将其按照《实用新生儿学》关于新生儿疾病诊断标准，结合患儿入院诊断及入院时检验结果将其分为：败血症亚组(n＝19)、早产低体重亚组(n＝50)、新生儿肺炎亚组(n＝151)、低血糖亚组(n＝14)、溶血病亚组(n＝20)、新生儿呼吸窘迫综合征(NRDS)亚组(n＝8)、病理性黄疸亚组(n＝81)、颅内出血亚组(n＝13)、新生儿缺氧缺血性脑病(HIE)亚组(n＝29)、窒息亚组(n＝20)、羊水污染亚组(n＝62)及胎膜早破亚组(n＝44)。同时，选择同期在本院住院治疗无上述相应疾病的新生儿2129例，纳入对照组，为了研究时与个研究亚组不交叉，分别纳入与各研究亚组相对应的对照亚组(n＝201, 170, 69, 206, 200, 212, 139, 207, 191, 200, 158, 176)，血清CRP平均秩次采用两样本秩和检验Mann Whitney U检验，相关性分析采用简单相关分析(Pearson correlation)多重线性回归分析方法(本研究遵循的程序符合本院人体试验委员会制定的伦理学标准，得到该委员会批准，分组征得受试对象监护人的知情同意，并与其签署临床研究知情同意书)。研究组与对照组新生儿性别、日龄、喂养方式及分娩方式等比较，差异无统计学意义(P>0.05)。

采用Mann Whitney U检验及简单相关分析进行初步筛选结果显示，血清CRP平均秩次：败血症(是：138.32，否：107.86；U＝1381.0，P＝0.046)，NRDS(是：66.44，否：112.16；U＝495.5，P＝0.046)；简单相关分析结果：血清CRP与血红蛋白(Hb)，红细胞压积(Hct)，总胆红素(TBIL)，直接胆红素(DBIL)及血糖(Glu)可能相关(r＝－0.149，－0.143，－0.164，0.246，0.316；P<0.05)。进一步以血清CRP为因变量，以初步筛选有意义指标为自变量，采用逐步回归分析方法进行多重线性回归分析，DBIL和Glu进入回归方程的结果显示，DBIL和Glu与血清CRP均呈正相关关系(b1＝0.477, b2＝0.311；95%CI：0.608～2.265，0.588～9.445；P＝0.001，0.028)，该模型经检验有统计学意义(F＝9.298，P＝0.001)，但拟合优度略差(R²＝0.341)。

DBIL，Glu与血清CRP均呈正相关关系，并且DBIL对血清CRP的影响程度大于Glu，但拟合优度略差。对其他影响血清CRP的重要因素，尚待发现。

To study the influential factors of C-reactive protein (CRP) in neonates.

A total of 220 neonates whose serum CRP exceeded 8 mg/mL were recruited in the department of Neonate Intensive Care Unit (NICU) of BaYi Children's Hospital of General Military Hospital of Beijing from January 2009 to December 2010. They were selected to study group According to the diagnostic criteria for neonatal disease of Practical Neonatology and in accordance with the diagnosis and examination results on admission, they were divided into following sub-groups: sepsis sub-group (n=19), premature low birth weight sub-group (n=50), neonatal pneumonia sub-group (n=151), hypoglycemia sub-group (n=14), hemolytic disease sub-group (n=20), respiratory distress syndrome of newborn (NRDS) sub-group (n=8), pathological jaundice sub-group (n=81), intracranial hemorrhage sub-group (n=13), hypoxic ischemic encephalopathy (HIE) sub-group (n=29), asphyxia sub-group (n=20), amniotic fluid pollution sub-group (n=62) and premature rupture of membranes sub-group (n=44). At the same time, a total of 2129 inpatient neonates without corresponding diseases were chosen as corresponding control group (n=201, 170, 69, 206, 200, 212, 139, 207, 191, 200, 158, 176), respectively. Procedures were performed according to the Helsinki Declaration and approved by institutional Review Board (Institutional Review Board of General Military Hospital of Beijing PLA). The blood sampling of neonates were collected according to patient condition after approval was obtained from their families during hospitalization period. There were no statistically significant difference about gender, age, feeding and delivery methods between two groups (P>0.05). The average rank of serum CRP took two-sample rank test (Mann Whitney U test), correlation analysis took Pearson correlation of multiple linear regression analysis.

The preliminary screening results by Mann Whitney U test showed mean rank: Serum CRP of neonates septicemia (Yes: 138.32, No: 107.86; U=1381.0, P=0.046), NRDS (Yes: 66.44, No: 112.16; U=495.5, P=0.046), results of Pearson correlation showed possible correlation among serum CRP and hemoglobin (Hb), hematocrit (Hct), total bilirubin (TBIL), direct bilirubin (DBIL), glucose (Glu) (r=-0.149, -0.143, -0.164, 0.246, 0.316, P<0.05). Stepwise regression analysis was used further, dependent variable: Serum CRP, covariates: Neonatal septicemia, NRDS, Hb, Hct, TBIL, DBIL, Glu, only DBIL and Glu entered regression equation, standardized partial regression coefficient were 0.477, 0.311, respectively. The 95% confidence intervals were 0.608-2.265 and 0.588-9.445 (P=0.001, 0.028), respectively, so DBIL and Glu were positively correlated with serum CRP, the model had statistical significance (F=9.298, P=0.001), goodness of fit was slightly poorer (R²=0.341).

Serum CRP are positively correlated with DBIL and Glu, furthermore, DBIL has more positive effects on serum CRP than Glu. Some others important influential factors on serum CRP need to be detected.