Electronic supplementary material
The online version of this article (https://doi.org/10.1186/s13049-017-0463-0) contains supplementary material, which is available to authorized users.
At the University Hospital Zurich (USZ) and the Cantonal Hospital of Lucerne (LUKS) an individualized goal-directed coagulation and transfusion algorithm was introduced and implemented before 2012 (Coagulation algorithm of the USZ: USZ-Alg; of the LUKS: LUKS-Alg). Main differences between both algorithms are: 1) A target haematocrit-range of 0.21–0.24 (USZ-Alg) vs. a lower haematocrit limit only (LUKS-Alg). 2) Blind coagulation-package in selected cases (LUKS-Alg only). 3) Factor XIII substitution is considered earlier according to the USZ-Alg.
The Aim of this study was to analyse the impact of two different coagulation algorithms on the administration of allogeneic blood products, coagulation factors, the frequency of point of care measurements and haemoglobin level during resuscitation in trauma patients.
This retrospective, multicentre, observational study included all adult trauma patients with an injury severity score (ISS) ≥ 16 primarily admitted to the USZ or the LUKS in the period of 2012 to 2014. Referred patients and patients with missing/incomplete records of the initial treatment at the emergency department (ED) were excluded. Two propensity score matched groups were created using a non-parsimonious logistic regression to account for potential differences in patient and trauma epidemiology.
A total of 632 patients meeting the inclusion criteria were admitted to the two hospitals: 428 to the USZ and 204 to the LUKS. Two Propensity score matched groups (n = 172 per group) were created. Treatment with USZ-Alg compared with LUKS-Alg resulted in a lower number of patients receiving RBC transfusion (11.6% vs. 29.7%, OR 3.2, 95% CI 1.8–5.7, p < 0.001) and lower amount of RBC transfusion (0.5 SD 1.9 vs. 1.5 SD 3.9, p < 0.001). The different treatment algorithms resulted in lower mean haemoglobin levels in the USZ during resuscitation (8.0 SD 1.7 vs. 9.4 SD 1.8 g/dl, p < 0.001) and at admission to the ICU (8.3 SD 1.2 vs. 10.6 SD 1.9 g/dl, p < 0.001. Blood gas analyses to monitor treatment and haematocrit were made more frequently in the USZ (1.4 SD 0.8 vs. 1.0 SD 0.7 measurements per hour, p = 0.004).
A goal-directed coagulation algorithm including a target haematocrit-range including frequent and repeated haematocrit measurement may lead to less transfusion of RBC compared to only a lower haematocrit limit, when treating severely traumatized patients.