This study shows that cooling blood to 33°C and haemodilution with HES interacts to impair clot formation. We also found that fibrinogen substitution was effective also at 33°C, alone or in combination with haemodilution. The type of solution in haemodilution, however, did affect the reversal effect of fibrinogen in dilutional coagulopathy. After adding fibrinogen, clot propagation (COT2) improved more in HES haemodilution, while fibrinogen-dependent clot strength (Fib2 G'max) increased in RA haemodilution only, and not at all in HES haemodilution. Furthermore, combining fibrinogen with FXIII had an additional effect on fibrinogen-dependent clot strength (Fib2 G'max) with RA haemodilution only. Apart from these principle findings, we also confirmed the results of previous studies with other viscoelastic haemostatic assays that showed that both hypothermia and haemodilution independently impaired coagulation, and that HES impaired coagulation more than RA.
VHAs such as thrombelastography, measure coagulation in whole blood rather than in plasmaand they are therefore able to detect interactions between platelets, red blood cells and coagulation factors. Thus, VHAs more accurately describe coagulation in vivo and they are better predictors of bleeding than traditional measures of coagulation . FOR’s ability to measure viscosity changes enables detection of the start of clot formation (COT1), which is in contrast to with thrombelastography. FOR uses free oscillation, which results in less strain on the clot as compared with thrombelastography. Compared to maximal clot strength measured with thrombelastography, the corresponding measure with FOR (Fib1 G'max) is more dependent on platelets than on factors affecting the fibrinogen polymerization, i.e. fibrinogen concentration and FXIII . Thus, FOR may provide new information on the effect of hemodilution and hypothermia on coagulation.
In concordance with earlier studies with VHAs, we found that hypothermia impaired clot formation variables [18, 26]. In addition, we observed that clot strength was decreased even by mild hypothermia, which other studies failed to detect [19, 27]. Previous studies only noted this decrease in clot strength during moderate to severe hypothermia [18, 26]. The different measuring principles of ROTEM and FOR might be the cause of the different effects observed on clot strength by hypothermia. The possible higher sensitivity of clot strength variables measured with the FOR than with other methods is interesting. However, the decrease in clot strength observed from 37° to 33°C is moderate and unlikely to have any clinical implications.
Our results demonstrated that haemodilution with either HES or RA compromised coagulation, and that the effect of HES was more pronounced. This is in accordance with previous studies where HES was shown to affect coagulation more than RA both in vitro[28, 29] and in vivo[30, 31]. It is believed that HES, in contrast to crystalloids, also impair fibrinogen/fibrin polymerization .
Not only platelets are important cellular components of global haemostasis, but also erythrocytes play an important role . The role of the haematocrit level has been previously studied. In vitro studies with FOR and TEG, where the platelet counts were held constant, showed increased clot strengths with decreasing haematocrit levels [21, 33]. This increase in clot strength may be due to increasing fibrinogen levels after haemodilution with plasma. Ogawa recently also showed an increase in clot strength after haemodilution with plasma . However, at a given fibrinogen concentration, clot strength instead improved with falling haematocrit levels. Thus, fibrinogen levels seem to be far more important than haematocrit level for the final clot strength.
Other studies have not found any interaction effects between hypothermia and haemodilution, as assessed by coagulation time and clot strength [35, 36]. We discerned no interactions for the clot strength variables, but we did find significant interaction effects between hypothermia and haemodilution with HES in the time to complete clot formation (COT2). There was also a tendency towards interaction between hypothermia and haemodilution with RA for the time to complete clot formation (COT2), as well as between hypothermia and haemodilution with both solutions for the initiation of clot formation (COT1). FOR and thrombelastography are not completely comparable methods. For example, in contrast to thrombelastography, FOR can measure viscosity changes and discerns the coagulation time from the changes in viscosity. This may explain the inability of thrombelastography to detect any interaction effects on variables describing the early phases of coagulation.
Substitution with fibrinogen concentrate in dilutional coagulopathy has been demonstrated to effectively increase clot strength [9, 10, 13] and to reduce bleeding [11, 37]. A high ratio of fibrinogen to red blood cell transfusion is associated with improved survival rates . In this study, we found that fibrinogen increased clot strength and shortened the coagulation time (COT1 and COT2). The reduction in COT1, though, was very small and significant only when fibrinogen was combined with FXIII. These results are in agreement with the aforementioned studies with ROTEM, where addition of fibrinogen increased clot strength, while the coagulation time (CT) failed to improve after addition of fibrinogen [9, 13]. Interestingly, fibrinogen supplementation decreased COT2 significantly more in haemodilution with HES than in that with RA, while fibrinogen-dependent clot strength increased more in haemodilution with RA. In fact, Fib2 G'max did not increase at all after adding fibrinogen or fibrinogen combined with FXIII in haemodilution with HES. In other words, fibrinogen substitution reversed the early phases of the coagulation regardless of the presence of starch, while haemodilution with HES made fibrinogen unable to improve clot strength. One may speculate that higher doses of fibrinogen may be required to improve also clot strength after HES haemodilution.
FXIII’s activity is diminished by haemodilution . FXIII is ineffective at correcting dilutional coagulopathy in vitro, but has been shown to decrease post-operative chest-drain bleeding after cardiac surgery . Combining fibrinogen with FXIII in RA haemodilution had an additional on Fib2 G'max but no effect on Fib1 G'max, as discussed above. The additional increase of Fib1 G'max after adding FXIII together with fibrinogen is small (not significant) and will be masked by the platelets when assessing Fib1 G'max. Therefore, analysis of Fib2 G'max will add important information on factors affecting the fibrinogen polymerization.
The findings of this study emphasise the importance of reversing hypothermia and avoiding haemodilution. As previously concluded, it is reasonable to avoid HES in cases of major haemorrhages. Our study showed that the addition of fibrinogen was also beneficial in mild hypothermia and thus, patients are not required to be warmed before infusing fibrinogen. Moreover, if HES has been used in seriously bleeding patients, higher doses than those given in this study are probably needed to re-establish adequate clot strength. Finally, combining fibrinogen with FXIII may be worthwhile in RA induced coagulopathy, since this combination improved fibrinogen-dependent clot strength (Fib2 G'max) more than fibrinogen alone in this in vitro study. However, controlled clinical studies need to be undertaken with FOR, to establish correct doses of fibrinogen and factor XIII in dilutional coagulopathy.
In contrast to other studies, we studied the combination of concurrent hypothermia, haemodilution and coagulation factors. We chose to study mild hypothermia and moderate haemodilution, since we wished to study a common and clinically relevant situation. Unfortunately, this decreased the power to detect any statistical differences. Furthermore, FOR is not a widely used method in investigations of coagulopathy in bleeding patients. However, comparison with studies using thrombelastography (ROTEM and TEG) is logical, since FOR is very similar to thrombelastography and these methods are good predictors of perioperative bleeding. There are several other limitations to performing in vitro haemodilution studies using citrated blood, as is the case in this study. The situation in vitro may not accurately represent the situation in vivo. For example, in vitro models do not account for shear stress, release of tissue factor by the endothelium, and activation of procoagulation or fibrinolytic pathways in response to tissue trauma. Our study was performed in blood from healthy volunteers, whereas trauma patients may exhibit a wide range of coagulopathies.