In vitro evaluation of a novel system for monitoring surgical hemoglobin loss

Background Accurate measurement of intraoperative blood loss is an important clinical variable in managing fluid resuscitation and avoiding unnecessary transfusion of blood products. In this study, we measured surgical blood loss using a tablet computer programmed with a unique algorithm modeled after facial recognition technology. The aim of the study was to assess the accuracy and performance of the system on surgical laparotomy sponges in vitro. Methods Whole blood samples of premeasured hemoglobin (Hb) and volume were reconstituted from units of human packed red blood cells and plasma and distributed across surgical laparotomy sponges. Normal saline was added to simulate the presence of varying levels of hemodilution and/or irrigation use. Soaked sponges from 4 different manufacturers were scanned using the Triton System with Feature Extraction Technology (Gauss Surgical, Inc., Palo Alto, CA) under 3 different ambient light conditions in an operating room. Accuracy of Hb loss measurement was evaluated relative to the premeasured values using linear regression and Bland-Altman analysis. Correlations between studied variables and measurement bias were analyzed using nonparametric tests. Results The overall mean percent error for measure of Hb loss for the Triton System was 12.3% (95% confidence interval [CI], 8.2%-16.4%). A strong positive linear correlation between the premeasured and actual Hb masses was noted across the full range of intraoperative lighting conditions, including (A) high (r = 0.95 [95% CI, 0.93-0.96]), (B) medium (r = 0.94 [95% CI, 0.93-0.96]), and (C) low (r = 0.90 [95% CI, 0.87-0.93]) mean ambient light intensity. Bland-Altman analysis revealed a bias of 0.01 g [95% CI, -0.03 to 0.06 g] of Hb per sponge between the 2 measures. The corresponding lower and upper limits of agreement were -1.16 g (95% CI, -1.21 to -1.12 g) per sponge and 1.19 g (95% CI, 1.15-1.24 g) per sponge, respectively. Measurement bias of estimated blood loss and Hb mass using the new system were not associated with the volume of saline used to reconstitute the samples (P = 0.506 and P = 0.469, respectively), suggesting that the system is robust under a wide range of sponge saturation conditions. Conclusions Mobile blood loss monitoring using the Triton system is accurate in assessing Hb mass on surgical sponges across a range of ambient light conditions, sponge saturation, saline contamination, and initial blood Hb. Utilization of this tool could significantly improve the accuracy of blood loss estimates.