Comparison of two point-of-care measurements of hemoglobin against standard laboratory values during pediatric liver transplantation: A retrospective review

Abstract

Background: Point-of-care hemoglobin testing devices play an important role in intraoperative anesthetic management where significant hemorrhage is anticipated; however, the reliability of these devices has not been examined in the context of pediatric liver transplantation. In this retrospective observational study, we aimed to determine whether 95% of results from two point-of-care hemoglobinometers, the HemoCue and iSTAT, would fall within a difference of ±1 g/dl, our a priori-defined clinically acceptable level of agreement, of the hemoglobin measures on a core laboratory complete blood count.

Methods: We retrospectively collected data from 70 patients presenting for a liver transplant at a single center, tertiary care pediatric hospital over a 3.5-year period. We analyzed 92 contemporaneous pairs of hemoglobin values from the HemoCue and complete blood count, and 252 pairs of hemoglobin values from the iSTAT and complete blood count. Agreement between the point-of-care devices and complete blood count was assessed using Bland-Altman analysis, which was the primary outcome. Secondary analyses included an error grid analysis and Cohen's kappa statistic.

Results: Both point-of-care devices underestimated complete blood count hemoglobin values and neither device satisfied our a priori-defined clinically acceptable level of agreement that 95% of values would fall within ±1 g/dl of the complete blood count measurement. The mean difference [limits of agreement] of the HemoCue was 0.4 g/dl (p < .001) [-0.9 to 1.6 g/dl] and of the iSTAT was 0.6 g/dl (p < .001) [-1.4 to 2.6 g/dl]. Secondary error grid analysis revealed that neither device performed well enough to replace a complete blood count at critical thresholds of hemoglobin values.

Conclusions: While the HemoCue and iSTAT contribute information in a timely manner during dynamic intraoperative situations, there is significant imprecision compared to the gold standard complete blood count. If clinical stability allows, the results of these point-of-care hemoglobinometers should be confirmed with a complete blood count, rather than being used as the sole factor in determining transfusion needs during pediatric liver transplantation.

Keywords: hemoglobin; liver transplant; pediatric; point-of-care; retrospective.