Approximately 1% of the UK population receives anticoagulation with warfarin. Head injury accounts for some 1.4 million emergency department attendances in the country. Therefore, significant numbers of patients with head injury have a therapeutic coagulopathy. This review aims to examine the existing evidence for optimal management of warfarinised head injured patients, particularly with respect to the need for early CT imaging and the use of reversal agents in cases of proved haemorrhage.
- mental health
Statistics from Altmetric.com
Head injury accounts for a large number of consultations in the emergency department (ED), with some 1.4 million patients attending each year in the UK with minor head injury.1 Thus, head injury represents a significant part of the ED workload. Roughly 1% of the UK populous receives anticoagulation, most commonly warfarin, the chief indications being embolic prophylaxis in atrial fibrillation and management of venous thrombotic disease.2
Warfarin use is likely to increase in the UK with the recent introduction of National Institute for Health and Clinical Excellence (NICE) guidelines reiterating anticoagulation as a central tenet for management of this condition.3
The risk of intracranial haematoma is increased 10-fold in patients taking warfarin, with the majority of bleeds being spontaneous. It is acknowledged that intracranial haemorrhage in warfarinised patients is preceded by apparently innocuous trauma in a minority of patients, although large studies to calculate this risk are not yet available.4 5
There is a great deal of controversy regarding the appropriate management of warfarinised patients with minor head injury, largely stemming from the fact that a substantive evidence base for best practice for this subset of patients is lacking. The introduction of NICE guidelines for managing head injury has had a significant impact upon practice in the emergency department and has allowed for more consistent management.6 The evidence base in this guideline for appropriate use of CT scanning to detect intracranial pathology is derived from large studies in Canada and New Orleans, USA.7 8 However, the Canadian group excluded warfarinised patients from their study outright while the New Orleans group included only one patient with coagulopathy in their study.
The contentious areas in managing anticoagulated patients who incur minor head injury include the appropriate use of CT scanning, necessity for admission and observation, and advice on withholding warfarin post-injury for a specified period of time.9 10
The NICE guidelines advise that CT head scan should be immediately requested in all warfarinised patients with head injury who suffer any loss of consciousness or post-traumatic amnesia. However, as stated above, there is no evidence base to suggest the most appropriate threshold for obtaining a CT scan in this group of patients. Therefore, there is uncertainty as to whether the NICE guidance provides a suitable protocol for managing anticoagulated patients. This review of the literature aims to outline what is currently known about the management of this patient group.
A literature search was conducted from the databases of Medline (via Pubmed; http://www.pubmed.gov) and the Cochrane Database of Systematic Reviews (via the OVID Technologies interface; http://gateway.ovid.com). All searches were limited to the English language. The PubMed search was limited to the year range 1998–2008 and used the combinations of the following words: ‘head injury’, ‘head trauma’, ‘warfarin’ and ‘anticoagulation’.
The initial search yielded 56 articles. Further review was performed to eliminate those articles that did not have sufficient focus on pre-injury anticoagulation and management, narrowing the yield to 15 articles which are further discussed below. This was not a systematic literature review and therefore the papers were not scored.
A retrospective analysis conducted by Li et al11 included 144 anticoagulated patients attending the ED with head injury. Of these patients, 134 (93%) were described as receiving their injury from a ‘simple fall’. This study included patients taking warfarin who had incurred head injury and subsequently had CT head scan. Patients excluded were those with new neurology, altered mentation or who were deemed high or moderate risk. This group found the incidence of intracranial haemorrhage in their study to be 6.2% (9/144), with six parenchymal haemorrhages, two subdural haematomas and a single subarachnoid haemorrhage being described. The authors of this study concluded that significant numbers of warfarinised patients develop intracranial pathology with even minor head injury. The authors postulate that this subset of patients should be considered at moderate risk and go on to recommend intracranial imaging for all warfarinised head injury patients during emergency assessment, regardless of neurological symptoms.
These results are largely in agreement with a study by Fabbri et al,12 which included 501 patients who had CT scan following head injury but who do not fulfil criteria for CT scanning as described in the NICE guidance (so called ‘NICE negative’ patients). This group found the incidence of intracranial bleeding to be 8% in their described patient group. Forty per cent of the patients with intracranial haemorrhage noted were taking anticoagulants. Fabbri et al calculated an OR of 5.48 for developing intracranial haemorrhage for coagulopathic patients and go on to suggest mandatory CT scanning for all anticoagulated patients.
Further evidence for the potentially catastrophic clinical course of anticoagulated patients comes from a study by Lavoie et al.13 This group conducted a retrospective analysis of 384 patients who presented to the ED with head injury, 9% of whom were anticoagulated. This group demonstrated a higher rate of mortality in the anticoagulated patients with a calculated OR of 2.73 for risk of death. Although this study does not stratify patients into risk categories based on neurological status, it does provide further evidence for negative outcome for head trauma as a consequence of anticoagulation. It would suggest that particular caution should be observed during the emergency assessment of such patients.
A retrospective study by Reynolds et al14 attempts to describe the post injury clinical course in a small number of anticoagulated patients with head injury. The study included 32 patients, all of whom were taking warfarin, were aged 65 or over and had suffered clinically defined minor head injury. Twenty-four patients were discharged home directly from the ED, of these 22 were alive at follow-up while two were uncontactable. Of the remaining eight patients, four were admitted following CT scan having developed parenchymal bleeds and were alive at follow-up. The remaining four patients developed serious neurological deterioration in the ED, despite an initial GCS of 15. The mean time for deterioration was observed to be 3.83 hours. Again, all of these patients suffered parenchymal haemorrhages. Of these four patients, three died as a consequence of their head injury. This group calculates an incidence of intracranial bleed of 25% in their study group, much higher than incidences described by other groups.
A retrospective review of a level 1 trauma registry by Cohen et al15 included 77 anticoagulated patients presenting with a Glasgow Coma Scale (GCS) of 13–15. The international normalised ratio (INR) had been measured in just 57% of these patients, and of these 47% had an INR above four. Of these patients, 26% were discharged home directly from the ED, with only 35% of these having had a CT. The remaining patients were all admitted for neurological observation, with 70% having CT scans. Of the discharged group, 18/20 returned with intracranial haemorrhage. Among the admitted group, 80% deteriorated to GCS <10 with abnormal CT, with a mortality of 84%. The high mortality observed in this study is likely to be due to some element of selection bias as recruitment was restricted to patients who died, were admitted to ITU or required hospital admission for longer than 3 days. Furthermore, the study included patients with GCS of 13. Such a GCS score combined with coagulopathy would stratify these patients into a high-risk category for intracranial bleed, where mortality would be expected to be high. Despite these limitations, this study highlights important inconsistencies in the management of this group of patients, particularly with respect to measuring INR, obtaining intracranial imaging and the timing of that imaging. The authors suggest many improvements including the routine measurement of INR in all anticoagulated patients, while routine CT scanning should be considered. The authors suggest the involvement of haematologists in circumstances where INR is supra-therapeutic, with a view to considering reversal of anticoagulation.
A study by Gittleman et al16 retrospectively reviewed a series of 89 patients who were all anticoagulated, either with heparin or warfarin and attended the ED with head injury and subsequently had CT scans. Of the 89 patients, 77 were taking warfarin while the remainder were prescribed heparin. All patients sustained their injury as the result of a fall. On admission, 66 patients had a GCS of 15 while the remaining 23 all had a GCS of 14. Of the patients with a presenting GCS of 14, 7/23 (30.4%) had a demonstrable bleed evident on CT scanning, while none of the patients with GCS 15 had an abnormal scan. The authors calculated an overall intracranial bleed rate of 7.8% in their study population, with the rate exceeding 30% in those with GCS 14. The authors of the study conclude by suggesting that anticoagulated patients with CS 15 are much less likely to develop significant intracranial pathology than patients with lower GCS and advocate not scanning anticoagulated patients with a GCS of 15 routinely. However, beyond the initial presentation, there is no follow-up of these patients, therefore this study has not taken into account delayed complications and is likely to have underestimated the incidence of intracranial insult in patients with initially normal scans. Aside from this, the study further confirms the high incidence of intracranial bleeds in anticoagulated patients following seemingly trivial injury. The studies and their findings are summarised in table 1.
Reversal of anticoagulation
No randomised controlled trials assessing outcomes appear to exist on the management of warfarin-associated ICH. The literature in this area largely comprises small case series and retrospective data analyses. Consequently, the management of patients between studies varies widely, with little consensus on the reversal regimens employed, and indeed their dosing.
A study by Ivascu et al17 entered 82 warfarinised patients with head injury into their study protocol, which involved early routine scanning of anticoagulated patients and expedient initiation of reversal therapy in cases of haemorrhage. Of these patients, 19 were found to have suffered ICH on CT. Of the ICH-positive patients, 84% presented with an initial GCS of 14 or above. The median INR for patients with haemorrhage was 2.7 compared to 2.5 for those without. All patients with haemorrhage were given fresh frozen plasma, with reversal being initiated at a mean time of 1.9 hours after presenting to the ED. Two of the 19 protocol patients who received FFP died, giving a mortality rate of roughly 10% to this group of patients. The authors point out that both fatalities had presented more than 10 hours after their initial injury and had developed very large haemorrhages by the time they were scanned.
The authors compare this mortality rate to a prospective cohort study they had conducted within the same institution. This study had demonstrated an overall 48% mortality rate of anticoagulated head injury patients, with a significant proportion of patients suffering haematoma expansion despite receiving fresh frozen plasma (FFP) therapy. In this study, however, reversal therapy was instituted at a mean time of 4.3 hours after presentation.
The authors of the study conclude that neither GCS nor level of anticoagulation can reliably predict the presence of ICH and that expeditious CT scanning combined with prompt warfarin reversal with FFP can reduce ICH progression and improve mortality.
A further study by Goldstein et al18 reiterates the importance of timely reversal of anticoagulation. This group performed a retrospective analysis of 69 warfarinised patients who attended with head injury and had ICH demonstrated by CT scanning. The authors observed that patients whose INR was successfully reversed within 24 hours had shorter median time from diagnosis of ICH to first dose of FFP. The authors demonstrated that every 30-minute delay in first dose of FFP was associated with a decreased odds of 20% of reversal of warfarin in the first 24 hours. However, earlier reversal was not associated with an evident improvement in neurological outcome.
The use of FFP is complicated by the large volumes required to reverse anticoagulation, particularly in cases where the INR is greatly prolonged. Rapid infusion acutely increases plasma protein levels and consequently increases the risk of circulatory overload.19
A multicentre Swedish study from 200120 set out to survey and assess current management of warfarin-associated ICH, with a particular emphasis on reversal therapy. This group retrospectively reviewed all anticoagulated head injured patients who had an ICH at 10 Swedish hospitals over a 4-year period. Their study identified groups of patients according to therapy received, including those treated with prothrombin complex, those to whom FFP was administered and, finally, those receiving no such therapy.
There were 23 patients in the prothrombin complex group, with a mortality of 17% at 24–48 hours and 60% at 6 months. Eighteen patients comprised the FFP group with a mortality rate at 24–48 hours of 11% and 33% at 6 months. Of the 42 patients receiving no reversal therapy 43% were dead at 24–48 hours, with mortality rising to 64% at 6 months. The authors of this study highlight the lack of an existing consensus with respect to reversal of anticoagulation and assert the need for randomised trials to discern optimum management for this vulnerable patient group.
The aim of this literature review was to examine the existing evidence for optimal management of warfarinised head injured patients, particularly with respect to the need for early CT imaging and the use of reversal agents in cases of proved haemorrhage.
Current NICE guidelines do not support the routine use of CT imaging in all anticoagulated patients presenting with head injury, suggesting selectively scanning those with GCS=15 who have a history of post-traumatic amnesia and/or loss of consciousness.6 However, our literature review indentified several retrospective analyses that report the development of ICH in patients that can be described as ‘NICE negative’—that is, do not fulfil NICE criteria for necessitating intracranial imaging. Many of the authors of the studies reviewed go on to advocate the routine use of CT scanning in this patient group.
A survey by the authors of 139 EDs demonstrates that current practice varies widely in the UK. Of those surveyed, 42% replied that all warfarinised head injured patients receive urgent CT head scans routinely, while the remaining 58% replied that urgent scanning is not employed for all such patients.21
Our review identified that no prospective studies or randomised trials exist assessing the proper use and timing of CT scanning in head injured patients taking warfarin and that trials informing NICE guidelines do not include sufficient numbers of anticoagulated patients to make firm assertions about how CT scanning should be employed in this patient group. The data that do exist, albeit limited to retrospective analyses and case series, advocate particular caution when dealing with this patient group and highlight their vulnerability to haemorrhage at a seemingly lower threshold.
Our findings support the need for a large-scale prospective study to determine the risk of haemorrhage following minor head injury and the most appropriate use of intracranial imaging and clinical observation. Similarly, our review found that solid evidence for the reversal of anticoagulation in cases of ICH is deficient. Again, no randomised trials exist to determine the optimum use and timing of reversal therapy. The existing literature is largely composed of small case series and retrospective data, with a variety of reversal agents employed in these studies, making comparison of findings difficult. There is good evidence that reversal improves outcome in patients with elevated INR who suffer haemorrhagic complications (not necessarily ICH). However, some of the data reviewed indicate that haematoma progression in ICH occurs despite reversal therapy, with no benefit in terms of neurological status observed. The existing data also suggest that timing of reversal therapy is important, with normalisation more likely to be achieved with early initiation of therapy. Furthermore, only one study exists comparing different reversal therapies and includes only small numbers of patients, making meaningful comparisons difficult. There are no studies that have investigated the role of withholding anticoagulation in patients withhead injury who may be supra-therapeutic, but with no neurological complications. Further prospective studies are required to evaluate the role of anticoagulation in the outcome of this patient group.
In summary, our literature review has demonstrated that the evidence for the management of warfarinised patients who incur minor injury is rather limited and confined to a small series of retrospective analyses. The incidence of ICH in minor head injury is unknown, but is likely to elevated by taking warfarin. There is a lack of consensus regarding the appropriate role of CT scanning for these patients. Similarly, there are few existing data evaluating the use of reversal agents and comparing the efficacies of different regimens in instances where ICH has occurred in warfarinised patients. Our group would advocate further large-scale studies to elucidate the most suitable management strategies for this vulnerable patient group.
Competing interests None.
Provenance and peer review Not commissioned; externally peer reviewed.
If you wish to reuse any or all of this article please use the link below which will take you to the Copyright Clearance Center’s RightsLink service. You will be able to get a quick price and instant permission to reuse the content in many different ways.