Comparison of the accuracy of hemoglobin point of care testing using HemoCue and GEM Premier 3000 with automated hematology analyzer in emergency room

The laboratory analysis provides accurate, but time consuming hemoglobin level estimation especially in the emergency setting. The reliability of time-sparing point of care devices (POCT) remains uncertain. We tested two POCT devices accuracy (HemoCue^®201⁺ and Gem^®Premier™3000) in routine emergency department workflow. Blood samples taken from patients admitted to the emergency department were analyzed for hemoglobin concentration using a laboratory reference Beckman Coulter LH 750 (HB_LAB), the HemoCue (HB_HC) and the Gem Premier 3000 (HB_GEM). Pairwise comparison for each device and Hb_LAB was performed using correlation and the Bland–Altman methods. The reliability of transfusion decision was assessed using three-zone error grid. A total of 292 measurements were performed in 99 patients. Mean hemoglobin level were 115 ± 33, 110 ± 28 and 111 ± 30 g/l for Hb_HC, Hb_GEM and Hb_LAB respectively. A significant correlation was observed for both devices: Hb_HC versus Hb_LAB (r² = 0.93, p < 0.001) and HB_GEM versus HB_LAB (r² = 0.86, p < 0.001). The Bland–Altman method revealed bias of −3.7 g/l (limits of agreement −20.9 to 13.5) for HB_HC and HB_LAB and 2.5 g/l (−18.6 to 23.5) for HB_GEM and HB_LAB, which significantly differed between POCT devices (p < 0.001). Using the error grid methodology: 94 or 91 % of values (Hb_HC and Hb_GEM) fell in the zone of acceptable difference (A), whereas 0 and 1 % (Hb_HC and Hb_GEM) were unacceptable (zone C). The absolute accuracy of tested POCT devices was low though reaching a high level of correlation with laboratory measurement. The results of the Morey´s error grid were unfavorable for both POCT devices.