Injured patients are at significant risk of both hemorrhagic and thrombotic complications. These divergent risks create a therapeutic conundrum for trauma surgeons. Use of anticoagulation can lead to potential exsanguination and death, while avoidance of anticoagulation can lead to thrombotic complications and death . Our data represents a novel report that suggests that therapeutic anticoagulation can be safely accomplished in select patients with intracranial hemorrhage.
There is very little to guide trauma surgeons in the safety profile of therapeutic anticoagulation. A recent review by Golob, et. al. evaluated the safety of initiating therapeutic anticoagulation in multi-injured trauma patients . They noted that 21% of patients had complications from the therapy. The most common complication was an acute drop in hemoglobin requiring a blood transfusion; three patients died as a result of hemorrhage. Clinical factors associated with a higher risk of complications were COPD, low platelet count before therapy, and the use of unfractionated hemorrhage. This study, however, did not include any patients with head injuries, so extrapolation to this population is difficult.
Injured patients are at significant risk of thrombotic complications. Patients with multisystem trauma may develop DVT at a rate of 58%, while a quarter of patients with isolated intracranial hemorrhage may develop DVT . This has led to significant study evaluating medical DVT prophylaxis in head injured patients. These studies have evaluated both low dose heparin and low molecular weight heparin. Norwood, et.al. noted that enoxaparin could be safely administered to select patients within 24 h of craniotomy for trauma . In a separate report, this group noted a 3.4% progression rate of intracranial hemorrhage after institution of prophylactic doses of anticoagulants .
These reports were highly important in that they dispelled the traditional viewpoint that prophylactic anticoagulation is unsafe after brain trauma. They do not, however, speak to the safety profile of therapeutic anticoagulation. Traditional recommendations suggest that therapeutic anticoagulation is unsafe after traumatic intracranial hemorrhage. Textbooks have noted that anticoagulation should be delayed for 3 days to 6 weeks after injury “depending on local customs” (although no references were cited to support this recommendation) . Our data suggests that anticoagulation in the earlier portion of this window may be safe.
Much of the hesitation to use therapeutic anticoagulation after brain trauma likely stems from studies on pre-injury use of anticoagulants. Cohen, et.al. reported mortality rates of 84%–91% among patients who were anticoagulated prior to an intracranial bleed . Mina, et.al. compared anticoagulated patients to matched controls and found an absolute increase in mortality of 30% among the anticoagulated patients . Another study evaluated the effect of rapid reversal of coagulopathy. Patients who underwent a rapid, protocolized reversal of coagulopathy had a 38% absolute reduction in mortality compared to historical controls . Although these studies clearly indicated higher risks of death and disability among patients exposed to anticoagulants before the time of injury, they do not speak to the risks of administration of anticoagulants in a delayed fashion.
While many thrombotic complications can be treated without anticoagulation, there are specific scenarios in which anticoagulation has the potential to markedly improve a treatment regimen. Inferior vena cava (IVC) filters are the mainstay of treatment of both DVT and PE in patients with a contraindication to anticoagulation . There are certain situations, however, in which IVC filters are not adequate. The filters do not prevent propagation of a thrombus that has already embolized to the pulmonary vasculature. A saddle PE requires very little propagation to result in lethal shock, so anticoagulation in this population is critical. Similarly, the long term morbidity of phlegmasia cerulean dolens is reduced with anticoagulation. Further, there is a small, but defined, risk of thrombosis of the IVC after placement of a filter . This situation also requires anticoagulation. A final venous thrombosis that that is not amenable to treatment with an intravascular filter is an upper extremity DVT. Superior vena cava filters are uncommon and would lead to fatal intracranial swelling in the event of filter thrombosis.
There is only one report that has attempted to define the optimal treatment regimen of DVT or PE after intracranial hemorrhage . This report focused on non-traumatic hemorrhage, so the generalizability may be limited. The authors conducted a review of the literature and were unable to develop firm recommendations.
Blunt cerebrovascular injury is another event that may require anticoagulation despite the presence of an intracranial hemorrhage . Dissection of the carotid or vertebral arteries can lead to disabling or fatal stroke events, which may be prevented by adequate anticoagulation. Although much of the focus of treatment has shifted to antiplatelet regimens, there is a role for heparin in select cases. Our data suggests that therapeutic anticoagulation can be safely given to select patients with blunt cerebrovascular injury and intracranial hemorrhage.
Patients with mechanical cardiac valves represent a significant challenge to trauma surgeons [14–17]. The risk of artificial valves appears to be the highest in patients with a cage/ball valve in the mitral position. Atrial fibrillation and reduced left ventricular function add to the risk of stroke without anticoagulation. The natural history of these patients is unclear, as they are generally on anticoagulants, but we can glean some estimate of risk from studies that have evaluated temporarily discontinuing anticoagulation after intracranial hemorrhage. It appears safe to discontinue anticoagulation for brief periods of time [14, 15]. Most of this work has been conducted in patients with spontaneous intracranial hemorrhage. It is possible that traumatic hemorrhage is a different entity, as injured patients are more hypercoaguable than then general population. Our data represents an important adjunct to these studies, in that we have demonstrated that early reintroduction of anticoagulation can be safely accomplished.
There are limitations of this study worth noting. We did not have a protocolized approach to management of anticoagulation. Rather, we consulted with the neurosurgeons on a daily basis and we started anticoagulation when their clinical judgment indicated it was safe to do so. As such, we are likely dealing with a highly select patient population. Additionally, our sample size is limited. It is possible that we would have yielded different results with a larger sample size. Finally, some of our patients received anticoagulation for uncomplicated PE rather than the extreme examples listed in this discussion. This does not detract from our results demonstrating safety of anticoagulation, however.
In conclusion, selected patients with brain injury may safely be anticoagulated to prevent the propagation of thrombotic complications. Our data does not provide definitive evidence of the safety profile. Rather, this manuscript provides initial evidence that suggests that traditional beliefs about anticoagulation in patients with brain injuries may be incorrect. Our data should be used a springboard to develop further study on this issue, so that the specific groups that would most benefit from anticoagulation could be defined.