目的 对比外周血无创总血红蛋白测定与静脉血血红蛋白检测在肾病患儿血红蛋白测量中的相关性及一致性,并分析其可能的影响因素,为外周血无创总血红蛋白检测技术应用于肾病儿童血红蛋白的监测提供参考依据。方法 选择北京大学第一医院儿科2014年10月至2015年2月85名肾病住院患儿作为研究对象,其中男50例(58.8%),女35例(41.2%),年龄(9.35±4.29)岁(3~18岁), 非惯用手的无名指作为检测部位,用无创血红蛋白检测仪(PRONTO-7)检测无创血红蛋白浓度(noninvasive hemoglobin monitoring by spectrophotometry, SpHb),然后在 5 min内抽取静脉血2 mL,通过血液分析仪(Beckman coulter DXH-800)测定静脉血红蛋白浓度(true hemoglobin, tHb)。对所测得的SpHb和tHb两组数据进行比较分析。结果 对85对数据进行统计学分析,相关性分析显示,SpHb与tHb的相关系数为0.85(P<0.05), 有显著相关性。Bland-Altman一致性分析显示, SpHb与tHb的差值的平均值为-1.3 g/dL,标准差为1.47 g/dL,两者差值的95%CI为-4.2~1.5 g/dL,提示SpHb的平均测量结果低于tHb测量结果。根据SpHb与tHb的差值百分比绘制Bland-Altman一致性描点图,显示差值/均值百分比的均值为-9.8%,一致性界限为(-35.9%,16.2%), 超过了《全国临床检验操作规程》中允许的±6%的可接受范围。外周血无创血红蛋白检测与静脉血血红蛋白检测的一致率为31.8%,一致率的95%CI为21.7%~41.9%。四格表卡方检验显示采用SpHb值诊断贫血灵敏度较高,但特异性低,假阳性率高,且差异具有统计学意义 (P=0.000)。结论 肾病患儿的无创血红蛋白检测值与静脉血红蛋白值之间具有显著相关性,可以用于动态监测血红蛋白浓度的变化趋势,但一致性不足,无创血红蛋白测定尚不能取代静脉血红蛋白测定,尚不能用于贫血的诊断,其在肾病患儿血红蛋白浓度测量中的准确度有待于进一步探讨。

Objective: To assess the correlation and consistency between continuous noninvasive hemoglobin detection and venous blood hemoglobin detection in children with kidney disease, and try to analyze the affecting factors. Try to provide a reference for the monitoring of hemoglobin in children with kidney disease by continuous noninvasive hemoglobin detection technique. Methods: Eighty-five inpatient children with kidney disease, 50 boys (58.8%) and 35 girls (41.2%), aged from 3 years old to 18 years old (9.35±4.29) were included finally. Noninvasive hemoglobin monitoring by spectrophotometry (SpHb) was stably read by PRONTO-7, selecting the ring finger of the non handedness as the detection site. And then the venous blood hemoglobin of the same patient was collected in 5 minutes as the true hemoglobin (tHb) by Beckman coulter DXH-800. The data of SpHb and tHb were compared and analyzed using SPSS 17.0 and MedCalc. Results: Correlation analysis showed data of SpHb and tHb were with significant correlation, the correlation coefficient between SpHb and tHb was 0.85 (P<0.05). Bland-Altman plot points suggested that the mean of differences between SpHb and tHb was -1.3 g/dL. The 95% CI of agreement of SpHb-tHb was -4.2-1.5 g/dL, suggesting that the average measurement result of SpHb was lower than that of tHb. The mean of differences as percent between SpHb and tHb was -9.8%, 95% CI of agreement was (-35.9%, 16.2%) exceeding the acceptable range of true value ±6%. The consistent rate of non-invasive hemoglobin detection and venous blood hemoglobin detection was 31.8%, the 95% CI of consistent rate was (21.7%, 41.9%). The chi square test of the fourfold table showed that the diagnosis of anemia with SpHb was of high sensitivity, but the specificity was low, the false positive rate was high, and the difference was statistically significant (P=0.000). Conclusion: There was a significant correlation between SpHb and tHb in the children with kidney disease. Noninvasive hemoglobin measurement can be used for monitoring of changes of hemoglobin in children with kidney diseases. But the consistency between SpHb and tHb needs to be improved. Noninvasive hemoglobin measurement could not replace the venous hemoglobin measurement. It could not be used for the diagnosis of anemia, and the accuracy of hemoglobin concentration measurement in children with kidney disease should be further explored.