Drops of capillary blood are not appropriate for hemoglobin measurement with HemoCue: A comparative study using drop capillary, pooled capillary, and venous blood samples

Abstract

Abstract Background: Population-based surveys matched by time and in the same country, but which use different methodologies for determining hemoglobin (Hb) concentration, have shown inconsistencies in estimating anemia prevalence. This study aims to estimate measurement errors in Hb using the HemoCue 201+ in capillary blood, both in drops and pooled, compared with venous blood analyzed both with the HemoCue 201+ and clinical methodologies as references. Methods: 149 participants (children, women, and older adults) interviewed in their homes in Jojutla, Morelos, Mexico, were randomly allocated to donate either a drop of capillary blood or 350 µL of pooled capillary blood. Each group also donated 4 mL of venous blood. Hb was analyzed in duplicate using the HemoCue 201+ and by a reference method using cyanmethemoglobin. We also compared the latter method with a clinical hemocounter. We performed Bland Altman and Lyn’s concordance analyses. We calculated 95% confidence intervals (±95% CI).Results: A positive bias was found in HemoCue 201+ (0.314 g/dL) which was used as an adjustment factor. After adjustment, both venous blood (0.000±0.33 95%CI -0.649, 0.649) and pooled capillary blood (-0-02± 0.36; 95%CI -0.75, 0.70) using HemoCue reported the same average Hb concentration as that estimated using venous blood with the cyanmethemoglobin method, and within a variation below 0.7 g/dL in the same individuals. Single drops of capillary blood produced a higher positive bias (0.425±0.81; 95%CI -1.176, 2.026) despite adjustment, and the variation between duplicate of hemoglobin was ±1.6 g/dL (95%CI -1.176, 2.026), and with some results as high as 2-3 g/dL. Comparatively, the variation of the cyanmethemoglobin method and the clinical hemocounter with venous blood was ±0.3 g/dL (95%CI -0.23, 0.38 ug/dL). Conclusions: Random variation of results from drops of capillary blood cannot be corrected, affecting the accurate and precise measurement of Hb concentration from this source. Pooled capillary blood, in our hands, performed similarly to venous blood and may be useful as an alternative to venous blood when the latter is difficult to obtain. In any case, the popular use of drops of capillary blood for estimating Hb concentration both in individuals and in populations should be discontinued.