Background: Blood loss is a major concern in hip replacement, and may lead to allogeneic red blood cells (RBC) transfusion, associated with increased mortality and morbidity (1). Current guidelines recommend a restrictive RBC transfusion threshold of 8 g/dL to patients undergoing orthopedic surgery (2). To decide when to transfuse, accurate haemoglobin concentrations values are required. Traditionally they are obtained through the laboratory automated analysis (Lab Hb), but new technologies have allowed for noninvasive and continuous haemoglobin measurement (SpHb). In this study we compared the SpHb with the laboratory haemoglobin and assessed if there would be a different decision on whether to transfuse, based on the different values obtained.
Methods and Materials: We present preliminary data from a prospective observational study with EC and IRB approval. Lab Hb was determined in the hospital laboratory using an automated analyser, and the non-invasive haemoglobin measurements were done using the Masimo's Radical-7® Pulse CO-Oximeter and rainbow® R1 25 revision F. Randomly selected patients undergoing elective total hip replacement were monitored with the SpO2, and whenever blood was collected for Lab Hb, the concurrent SpHb was recorded. Correlation analysis was done using Pearson test. Bland-Altman was also employed.
Results and Discussion: 59 paired measurements were obtained. There was a good correlation (fig.1) between SpHb and Lab Hb (Pearson R= 0,742, p<0,001). Bland-Altman test (fig.2) revealed a bias of -1,15, meaning Lab Hb was recurrently higher than the SpHb. The limits of agreement were [-3,89; 1,59].
Considering the restrictive RBC transfusion threshold of 8 g/dL for these patient population, we found that in two cases, the decision to transfuse would differ based on the measurement taken into account. In one case, the SpHb was below the threshold while the Lab Hb was above, and in the other case, it was the reverse.
Conclusions: We found a good correlation between SpHb and Lab Hb. The bias was superior to one found in the literature (3), although we were using a more recent version of the sensors. However, this difference between measurements would have a clinical significant impact in two situations (3,4%). Solely based on the Hb threshold (2), this would have resulted in the decision to transfuse one patient that didn’t need, and not to transfuse one that needed.

09 / Clinical implications of the accuracy of non-invasive haemoglobin monitoring