Main findings
In 9 610 TBI patients, the present study showed the main following findings : 1/ coagulopathy was present at hospital admission in 28.5% of the patients, 2/ the higher the AIShead (i.e. the more severe the TBI), the higher the proportion of coagulopathic patients, 3/ this increased incidence of coagulopathy was observed at an earlier stage of TBI severity when severe extracranial lesions were present, 4/ for each level of TBI severity, the proportion of patients with coagulopathy was significantly higher in patients with severe extracranial associated lesions, 5/ when compared to head injuries, extra-head injuries were associated with higher incidences of coagulopathy, 6/ the presence of severe extracranial injuries was one of the strongest risk factors for coagulopathy following TBI (OR 2 (1.8–2.3), P < 0.001).
As far as we know, this study is the largest cohort to date to examine the role of head and extra-head injuries and their association in the development of coagulopathy in trauma patients. All degrees of TBI severity, from mild (AIShead = 1) to currently untreatable (AIShead = 6) were included herein, whereas many previous studies restricted their analysis to moderate-to-severe TBI patients (excluding AIShead values of 0–1 and 6)(10,21,22). Moreover, all types of extracranial lesions associated to TBI, from none to extremely severe, were also included in the present study. Previous studies on TBI-related coagulopathy often excluded patients with important extracranial lesions (generally AIS score in any extracranial region ≥ 3), or even with any extracranial lesion (9,10,12). Thus, the contribution of concomitant extracranial lesions to coagulopathy following TBI was not studied at all or only partially studied.
Since injury types and severities are very diverse among reported series, as are the methods and times used to define coagulopathy, it is not surprising that the incidence of coagulopathy at hospital admission can range from 7 to 97.5% of TBI patients (6,10,23). In the present study, the overall incidence of coagulopathy at admission was 28.5%, which is in line with the overall incidences of 32.7% and 35.2% reported in two meta-analyses including 5357 and 7037 patients, respectively (6,23). These pooled proportions of patients developing coagulopathy after TBI obviously do not reflect the wide disparities related to several factors highlighted in the present study, including head and extra-head injury severities.
First, a continuously graded association between the severity of head injury and coagulopathy has been shown herein, as the higher was AIShead, the higher was the proportion of coagulopathic patients. While this might appear self-evident, Wafaisade et al. (8) and Lustenberger et al. (24) found AIShead as one of the independent risk factors for coagulopathy after TBI, but only for AIShead ≥ 5. More recently, Böhm et al. (9) did not identify AISbrain as an independent predictor of coagulopathy in a prospective cohort of 598 patients with isolated TBI (AISbrain ≥1 and extra-head AIS = 0), although a larger proportion of patients with severe head injury (AISbrain ≥ 5) had coagulopathy as compared to those with lower severity level. In the present study, inclusion of a large cohort of patients with all magnitudes of TBI allowed showing the stepwise and independent increased risk of coagulopathy with increasing head injury severity. Second, we hypothesized that the presence of severe extracranial lesions associated to the TBI could significantly contribute to the occurrence of coagulopathy. Previous studies investigating coagulopathy following head and/or extra-head injuries mainly assessed whether TBI induces a different coagulopathy than non-TBI does. In a prospective cohort of 80 trauma patients, Genet et al. showed that 46% of patients with combined TBI and extra-head injuries had coagulopathy, compared with only 13% and 5% of patients with isolated TBI or isolated injuries in other body regions (P < 0.001), but found no difference with regard to coagulation parameters whether patients had TBI or not (25). In 795 consecutive major trauma patients with head AIS ≥ 3 or other body areas with AIS score of ≥ 3, the dominant region of injury did not influence coagulation changes, and coagulopathy resulted from a combination of tissue injury and shock rather than a particular injury pattern (26). Similarly, de Oliveira et al. classified 345 patients in isolated TBI (AIShead≥3 and AISextra−head<3), and multisystem trauma with or without severe TBI (defined by AISextra−head≥3 with or without AIShead≥3, respectively) (27). Isolated severe TBI was not an independent risk factor for the development of coagulopathy; however, isolated severe TBI patients who developed coagulopathy had high mortality rates (66 vs. 16.6%, P < 0.05) (27). Samuels et al. compared 48 patients with isolated severe TBI, 45 patients with severe TBI and extracranial injuries, and 479 patients with isolated extracranial injuries by using CCA and viscoelastic hemostatic assays (VHA) (13). When compared to isolated extracranial injuries, isolated severe TBI seemed to be independently associated with a specific coagulopathy phenotype (mainly abnormalities in the rapidity of clot initiation, fibrin cross-linking, and hypofibrinogenemia), conversely to the aforementioned series (22, 23). Of note, proportions of patients with INR > 1.3 were similar between the groups and data on fibrinogen concentration as measured by conventional Clauss method were not provided (13). The present large cohort of patients with all magnitudes of head (from mild to untreatable) and extra-head (from none to untreatable) injury severity allowed specifying the contribution of severe extracranial lesions in the context of TBI: the increased incidence of coagulopathy with head injury severity was observed at a lower level of TBI severity when severe extracranial lesions were present ; for each level of TBI severity, the proportion of patients with coagulopathy was significantly higher in patients with concomitant severe extracranial lesions. The presence of severe extracranial injuries was one of the most strongly independent risk factors for admission coagulopathy observed in the present cohort.
Beyond head and extra-head injury severity assessed by AIS scores, we identified several additional risk factors for coagulopathy upon hospital admission following TBI. Since acute traumatic coagulopathy develops in response to the combination of tissue damage and systemic hypoperfusion (28,29), it was consistent to identify variables previously described as predictors of hemorrhage or massive transfusion: unstable pelvic trauma, prehospital fluids > 1000 mL, intubation, shock index ≥ 1, vasopressors, high blood lactate, low temperature and hemoglobin at admission (17, 27, 28). We also identified gunshot as an independent risk factor for coagulopathy, and indeed, hemostatic disorders have previously been reported more frequently in penetrating than in blunt head injuries (7, 29). Osmotherapy and pupillary abnormality, markers of TBI severity as is AIShead, were associated to coagulopathy. Male sex and young age were also independent risk factors for coagulopathy, which might be related to greater trauma severity usually sustained by this population of patients.
Limitations.
Our study has some limitations. First, this is a retrospective study with the inherent risks of missing data and confounding factors. However, data collection was prospective as this study used data from the Traumabase®, a multicenter national registry that secures data processing, limits missing data through reactive data tracking, and incorporates control for biases inherent in retrospective data collection (15, 34). Among variables considered for the present analysis, there was little missing data. We used multivariate analysis to overcome the most significant confounding variables such as degree of shock, head and extra-head injury severities, though unmeasured confounders may have accounted for differences in the incidence of coagulopathy between subgroups of patients. Only associations but no causalities can be ascribed from the present data.
Second, we used prothrombin time ratio (PT Quick percent) < 70%, platelet count < 100 G.L− 1, and Clauss fibrinogen level < 1.5 g.L− 1 to define coagulopathy in our study. Trauma-induced coagulopathy is classically defined as a prothrombin ratio of 1.2 or higher (28), which corresponds to prothrombin ratio (Quick percent) < 70% (19). Prothrombin Time (PT) assesses the extrinsic and common pathways of coagulation, and results are commonly expressed in seconds, percentage activity (prothrombin ratio in Quick percent) or International Normalized Ratio (INR). We used prothrombin ratio in Quick percent because this is the preferred way to express PT in the participating centers, and more generally in France. Since PT as a single parameter could not describe the complex coagulation disorders following trauma, we also used two additional laboratory tests routinely measured on hospital admission and used as targets of the initial coagulation resuscitation (18): fibrinogen concentration and platelet count. These variables have already been used to define coagulopathy, especially in the TBI context (9, 10). By reporting these three criteria for coagulopathy, we aimed at assessing a larger spectrum of hemostatic disorders following trauma. Nevertheless, some contributors to the complex pathophysiology of trauma-induced coagulopathy were not captured by the present data. Finally, we used AIShead ≥ 3 and AISextra−head ≥ 3 to define severe TBI and severe extracranial injuries, respectively. AIS is a consensus derived anatomically-based injury severity scoring system commonly used and most studies use a cut-off value superior or equal to 3 to classify an injury as severe (9,10,22). Although AIS and its derivatives ISS (17), New Injury Severity Score (NISS) (34) and Trauma and Injury Severity Score (TRISS) (35) perform well as a predictor of mortality, they primarily measure the overall injury severity. It would be interesting to achieve a higher level of granularity, for example by investigating the influence of different types of head (sub- or epidural vs intraparenchymal hemorrhage) or extra-head injury (orthopedic vs thoracic injury for example) on the incidence of coagulopathy. The impact of multiple associated lesions of lesser severity (AIS = 2), might be investigated as well.