Hemoglobin measurement in venous blood versus single-drop and pooled capillary blood: A method-comparison study in a controlled and survey setting in Uganda among children and women

Background Standard practice for estimating anemia in population-based surveys is to use a point-of-care device to measure hemoglobin (Hb) in a single drop of capillary blood. Emerging evidence points to larger than expected differences in Hb concentration depending on the blood source. Objective We evaluated use of different blood sources to measure Hb with a HemoCue 201+ analyzer compared to the reference method of venous blood tested with a Sysmex XN-450 hematology analyzer. Design Hb concentration in venous, pooled, and single drop capillary blood were collected in controlled (laboratory) and survey (Demographic Health Survey-8 pilot) settings in Uganda among children 6-59 months and non-pregnant women 15-49 years. Capillary and venous blood collected from the same individual was tested using a HemoCue 201+ analyzer and the venous blood was also measured with a Sysmex XN-450 hematology analyzer. Agreement between measures was estimated using Lin's concordance correlation coefficient, Bland-Altman plots, and Deming regression. Means and prevalences were compared using paired t-tests and McNemar's tests, respectively. Results The limits of agreement between Hb measured using a HemoCue 201+ analyzer and the reference method was lowest for venous (1.1 to 1.96 g/dL), followed by pooled capillary (1.45 to 2.27 g/dL), and single drop capillary blood (2.23 to 3.41 g/dL). Mean differences were less than 0.5 g/dL across comparators. There were statistically significant differences in Hb concentration from both types of capillary blood. Anemia prevalence was lower in pooled capillary blood compared to the reference method. Conclusion The variability of Hb measured by capillary blood using the HemoCue 201+ analyzer was higher than venous blood but the extent to which this impacts the validity of Hb and anemia estimates requires further exploration. Future research is also needed to evaluate the implications of using venous versus capillary blood in population-based surveys.