Abstract：
BACKGROUND: Anemia prevalence estimates reported in population surveys can vary based on the blood specimen source (capillary or venous) and analytic device (hematology autoanalyzers or portable hemoglobinometers) used for hemoglobin (Hb) determination.
OBJECTIVE: This study aimed to compare accuracy and precision of Hb measurement in three blood specimen types on three models of hemoglobinometers against the results from venous blood from the same individuals measured on automated analyzers (AAs).
METHODS: This multisite (Cambodia, Ethiopia, Guatemala, Lebanon, Nigeria, and Tanzania) study assessed Hb measurements in paired venous and capillary blood specimens from apparently healthy women (aged 15-49 y) and children (aged 12-59 mo) using three HemoCue® Hb models (201+, 301, and 801). Measurements were compared against reference values: venous blood in hematology AA and adjusted via regression calibration or mean difference in HemoCue® Hb. Venous, capillary pool, and single-drop capillary blood specimens were assessed for accuracy and precision.
RESULTS: Venous blood measured using HemoCue® Hb 301 exhibited a positive mean error, whereas responses in HemoCue® Hb 201+ and 801 were nondirectional compared with the reference. Adjustment with the reference harmonized mean errors for all devices across study sites to <1.0 g/L using venous blood. Precision was highest for venous blood (±5-16 g/L) in all sites, lowest for single-drop capillary (±9-37 g/L), and intermediate (±9-28 g/L) for capillary pool blood specimen. Imprecision differed across sites, especially with both capillary blood specimens, suggesting different levels of personnel skills.
CONCLUSIONS: Findings suggest that venous blood is needed for accurate and precise Hb determination. Single-drop capillary blood use should be discouraged owing to high measurement variability. Further research should evaluate the viability and reliability of capillary pool blood for this purpose. Accuracy of HemoCue® Hb devices can be improved via standardization against results from venous blood assessed using AA.
OBJECTIVE: This study aimed to compare accuracy and precision of Hb measurement in three blood specimen types on three models of hemoglobinometers against the results from venous blood from the same individuals measured on automated analyzers (AAs).
METHODS: This multisite (Cambodia, Ethiopia, Guatemala, Lebanon, Nigeria, and Tanzania) study assessed Hb measurements in paired venous and capillary blood specimens from apparently healthy women (aged 15-49 y) and children (aged 12-59 mo) using three HemoCue® Hb models (201+, 301, and 801). Measurements were compared against reference values: venous blood in hematology AA and adjusted via regression calibration or mean difference in HemoCue® Hb. Venous, capillary pool, and single-drop capillary blood specimens were assessed for accuracy and precision.
RESULTS: Venous blood measured using HemoCue® Hb 301 exhibited a positive mean error, whereas responses in HemoCue® Hb 201+ and 801 were nondirectional compared with the reference. Adjustment with the reference harmonized mean errors for all devices across study sites to <1.0 g/L using venous blood. Precision was highest for venous blood (±5-16 g/L) in all sites, lowest for single-drop capillary (±9-37 g/L), and intermediate (±9-28 g/L) for capillary pool blood specimen. Imprecision differed across sites, especially with both capillary blood specimens, suggesting different levels of personnel skills.
CONCLUSIONS: Findings suggest that venous blood is needed for accurate and precise Hb determination. Single-drop capillary blood use should be discouraged owing to high measurement variability. Further research should evaluate the viability and reliability of capillary pool blood for this purpose. Accuracy of HemoCue® Hb devices can be improved via standardization against results from venous blood assessed using AA.
摘要：
背景：人口调查中报告的贫血患病率估计值可能因血液样本来源（毛细血管或静脉）和用于血红蛋白（Hb）测定的分析设备（血液学自动分析仪或便携式血红蛋白计）而异。
方法：这种多站点（柬埔寨，埃塞俄比亚,危地马拉,黎巴嫩,尼日利亚,坦桑尼亚）研究使用三种HemoCue®Hb模型（201，301、801)。将测量值与参考值进行比较：血液学自动分析仪（AA）中的静脉血，并通过回归校准或HemoCue®Hb的平均差异进行调整。静脉,毛细管池,和单滴毛细血管血液标本的准确性和精密度进行评估。
结果：通过HemoCue®Hb301测量的静脉血表现出正平均误差，而HemoCue®Hb201+和801的反应与参考相比是非定向的。使用静脉血将所有设备的参考协调平均误差调整为<1·0g/L。所有部位的静脉血精度最高（±5至16g/L），单滴毛细管最低（±9至37g/L），和中间（±9至28g/L）的毛细血管池血液标本。不同地点的不精确性不同，尤其是两个毛细血管血液样本，建议不同水平的人员技能。
结论：研究结果表明，准确和精确的Hb测定需要静脉血。由于测量变异性高，应禁止使用单滴毛细血管血液。进一步的研究应评估用于此目的的毛细血管池血液的可行性和可靠性。HemoCue®Hb装置的准确性可以通过针对通过AA评估的静脉血结果的标准化来提高。
方法：这种多站点（柬埔寨，埃塞俄比亚,危地马拉,黎巴嫩,尼日利亚,坦桑尼亚）研究使用三种HemoCue®Hb模型（201，301、801)。将测量值与参考值进行比较：血液学自动分析仪（AA）中的静脉血，并通过回归校准或HemoCue®Hb的平均差异进行调整。静脉,毛细管池,和单滴毛细血管血液标本的准确性和精密度进行评估。
结果：通过HemoCue®Hb301测量的静脉血表现出正平均误差，而HemoCue®Hb201+和801的反应与参考相比是非定向的。使用静脉血将所有设备的参考协调平均误差调整为<1·0g/L。所有部位的静脉血精度最高（±5至16g/L），单滴毛细管最低（±9至37g/L），和中间（±9至28g/L）的毛细血管池血液标本。不同地点的不精确性不同，尤其是两个毛细血管血液样本，建议不同水平的人员技能。
结论：研究结果表明，准确和精确的Hb测定需要静脉血。由于测量变异性高，应禁止使用单滴毛细血管血液。进一步的研究应评估用于此目的的毛细血管池血液的可行性和可靠性。HemoCue®Hb装置的准确性可以通过针对通过AA评估的静脉血结果的标准化来提高。
