Determination of capillary hemoglobin levels using the HemoCue system in intensive care patients

Purpose: The study aimed to compare hemoglobin (Hb) values determined using the portable HemoCue system (HemoCue Hb 201+; HemoCue AB, Angelholm, Sweden) with laboratory Hb level determination. Materials and Methods: Adult patients hospitalized in our surgical intensive care unit who required an Hb level determination were included. To determine Hb level, one drop of arterial (A) or venous (V) blood was analyzed using HemoCue (HemoCue([A/V])), and also with an automated analyzer in the laboratory (Hb reference method, or Hb Lab([A/V])). Capillary blood (Cap) sample obtained simultaneously by fingerstick was analyzed using HemoCue (HemoCue([Cap])). Factors that could interfere with the accuracy of fingerstick measurements were also studied. Paired Hb level measurements were compared by Bland and Altman analysis (Hb Lab([A/V]) vs HemoCue([A/V]) and HemoCue([Cap])). Results: One hundred fifty blood samples were obtained from 79 patients. The mean absolute differences between Hb Lab [A] and HemoCue [A], Hb Lab [V] and HemoCue [V] and Hb Lab [A/V] and HemoCue [Cap] were 0.1 g/dl (95% confidence interval, -1.9 to + 2.2 g/dl), 0.1 g/dl (95% CI, -2.5 to + 2.6 g/dl) and 1.1 g/dl (95% CI, -3.6 to + 5.8 g/dl, respectively. Edema was the sole independent risk factor for discordance between HemoCue([Cap]) and Hb Lab([A/V]) (odds ratio, 6.65; 95% CI, 1.99-22.21; P < .001]. Conclusions: Hemoglobin level determination using HemoCue should not be used in critically patients, especially when capillary blood samples are used and/or in presence of edema. (C) 2011 Elsevier Inc. All rights reserved.