Abstract
Subdural haematoma (SDH) is a common injury sustained by older people living with frailty and multimorbidity, and typically following falls from a standing height. Anticoagulant and antiplatelet use are commonly indicated in older people with SDH, but few data inform decision-making surrounding these agents in the context of intracranial bleeding. Opposing risks of rebleeding and thrombosis must therefore be weighed judiciously. Decision-making can be complex and requires detailed awareness of the epidemiology to ensure the safest course of action is selected for each patient. Outcomes of surgical decompression in acute SDH are very poor in older people. However, burr hole drainage can be safe and effective in older adults with symptomatic chronic SDH (cSDH). Such patients need careful assessment to ensure symptoms arise from cSDH and not from coexisting medical pathology. Furthermore, the emerging treatment of middle meningeal artery embolisation offers a well-tolerated, minimally invasive intervention which may reduce the risks of rebleeding in older adults. Nonetheless, UK SDH management is heterogenous, and no accepted UK or European guidelines exist at present. Further randomised trial evidence is required to move away from clinical practice based on historic observational data.
Key Points
Subdural haematoma (SDH) is a common type of traumatic brain injury and is associated with high levels of morbidity and mortality.
Decision making around anticoagulants and antiplatelet in a patient with subdural haematoma (SDH) requires appreciation of the risks versus benefits.
Outcomes of acute subdural haematoma in people aged >65 years appear to be poor even with surgery.
Age alone should not be considered a barrier to surgery in patients symptomatic with chronic subdural haematoma.
Physicians have a role in supporting complex decision making in older people with SDH.
Introduction
In England and Wales, head injury is the primary cause of 1.4 million hospital attendances annually [1], traumatic brain injury (TBI) is the most common type of injury sustained by older people with major trauma defining injuries [2]. When head injuries occur in the context of anticoagulant or antiplatelet use, haemorrhagic complications may occur in up to 15.9% [3]. Furthermore, prescriptions of antiplatelets and anticoagulants have also risen dramatically in the last decade, with the number of people taking oral anticoagulation for atrial fibrillation increasing by 86.5% between 2011–2014 and 2014–2017 [4].
Subdural haematoma (SDH) is a common subtype of TBI in older people and is associated with high levels of morbidity and mortality.
Acute SDH (aSDH) is present in 10–20% of TBIs and has high rates of poor outcome in older people, where inpatient mortality has been shown to approximate to 30–40% [5, 6]. Symptoms normally develop over minutes to hours and result from a rapid rise in intracranial pressure causing around 50% of patients to have a GCS of <13 by time of arrival to hospital. [7] Primary brain injury has already occurred at this point: the main aim of treatment is to reduce secondary pressure-related neurological injury mediated through tissue ischaemia.
In contrast, chronic subdural haematoma (cSDH) takes a more insidious course. cSDH is characterised by the accumulation of degraded blood products, combined with a localised inflammatory response and neovascularisation in the subdural membrane; this leads to an osmotic accumulation of fluid with resultant mass effect. Typically, the transformation from acute to chronic SDH occurs by 14–21 days post primary haemorrhage, although the initial onset date may be spontaneous or unclear [8].
The incidence of cSDH increases with age, with 8.2 cases per 100,000 (age > 70) per year [9]. By 2030, cSDH is predicted to be the most prevalent cranial neurosurgical condition in adults [10]. Epidemiological studies indicate that the most common presenting symptoms of cSDH are cognitive impairment, headache and focal limb weakness; rates of symptoms vary in current literature [11]. The presentation of SDH is often delayed in older people, whose cerebral atrophy can partially protect against mass effect on the brain parenchyma. While Glasgow Coma Scale (GCS) may be mildly reduced, coma is uncommon [12]. Chronic SDH is increasingly conceptualised as a sentinel health event, broadly equivalent to hip fracture, where frailty and multimorbidity typically affect patients with this condition [8]. In one study, 1-year mortality rate was comparable to hip fracture, at around 30% at 1 year [13], though lower rates are more commonly reported.
Nationally, there is considerable variability in the parent specialty responsible for the care of older patients with both acute and chronic SDH. Tertiary neurosurgical services are highly specialised and unlikely to develop sufficient capacity to be able to support the care of all older people with SDH, especially where no neurosurgical intervention is appropriate [14].
Furthermore, co-existent acute medical pathology causing falls, frailty and multimorbidity often requires continued care from geriatric medicine services, which have the expertise in managing frailty and multimorbidity [14]. It is therefore essential that geriatricians are equipped with the relevant knowledge and skills required to manage patients with SDH under their care.
Reversal and suspension of anticoagulation and antiplatelets
Decision making
The decision to suspend anticoagulants or antiplatelets in a patient with SDH requires careful appreciation of the precise opposing risks of haemorrhage and thrombosis, and this varies depending on the acuity of bleeding.
If anticoagulation is continued during acute intracranial bleeding, mortality is reported to be as high as 80% [15]. Whilst this level of mortality is not applicable to chronic SDH, it does illustrate the potential risks of continued catastrophic intracranial haemorrhage which must be considered when bleeding is acute. In contrast, the risks of thrombosis (Table 1) can be relatively small for short periods of suspended anticoagulation. Even in the context of mechanical heart valves, thrombosis risk is 8.6% in patients not subjected to anticoagulation for 12 months [16]. Other data indicate that thromboembolic risk in a metallic mitral valve is 1.6 in 1,000 at 5 days [16].
Thrombosis risk in conditions indicating anticoagulation
| Indication for anticoagulation | Thrombosis risk |
|---|---|
| Atrial fibrillation | CHA2DS2VASC >3 = stroke risk 1.6 in 1000 at 5 days [17]. Up to 12% at 1 year for CHA2DS2VASC 9. |
| Venous Thromboembolism | Recent VTE (<3 months) = high risk; discuss with haematology and consider IVC filter [18] Unprovoked VTE, >3 months ago = 22.6% recurrence after 5 years [18] |
| Mechanical heart valves | Off anticoagulation, total risk is 8.6% per annum; major embolism 4% [16] Bileaflet AVR 0.5 in 1000 at 5 days [18] MVR 1.6 in 1000 at 5 days [18] |
| Antiphospholipid syndrome | Prior thrombosis history in APS patients can convey aggressive thrombotic tendencies, with concurrent thrombocytopaenia [19] Management is complex; these patients should be discussed urgently with Haematology. |
| Polycythaemia vera | 2.5–5% VTE risk per annum [20] |
| Essential Thrombocytosis | 1.9–3% VTE risk per annum [20] |
| Indication for anticoagulation | Thrombosis risk |
|---|---|
| Atrial fibrillation | CHA2DS2VASC >3 = stroke risk 1.6 in 1000 at 5 days [17]. Up to 12% at 1 year for CHA2DS2VASC 9. |
| Venous Thromboembolism | Recent VTE (<3 months) = high risk; discuss with haematology and consider IVC filter [18] Unprovoked VTE, >3 months ago = 22.6% recurrence after 5 years [18] |
| Mechanical heart valves | Off anticoagulation, total risk is 8.6% per annum; major embolism 4% [16] Bileaflet AVR 0.5 in 1000 at 5 days [18] MVR 1.6 in 1000 at 5 days [18] |
| Antiphospholipid syndrome | Prior thrombosis history in APS patients can convey aggressive thrombotic tendencies, with concurrent thrombocytopaenia [19] Management is complex; these patients should be discussed urgently with Haematology. |
| Polycythaemia vera | 2.5–5% VTE risk per annum [20] |
| Essential Thrombocytosis | 1.9–3% VTE risk per annum [20] |
Thrombosis risk in conditions indicating anticoagulation
| Indication for anticoagulation | Thrombosis risk |
|---|---|
| Atrial fibrillation | CHA2DS2VASC >3 = stroke risk 1.6 in 1000 at 5 days [17]. Up to 12% at 1 year for CHA2DS2VASC 9. |
| Venous Thromboembolism | Recent VTE (<3 months) = high risk; discuss with haematology and consider IVC filter [18] Unprovoked VTE, >3 months ago = 22.6% recurrence after 5 years [18] |
| Mechanical heart valves | Off anticoagulation, total risk is 8.6% per annum; major embolism 4% [16] Bileaflet AVR 0.5 in 1000 at 5 days [18] MVR 1.6 in 1000 at 5 days [18] |
| Antiphospholipid syndrome | Prior thrombosis history in APS patients can convey aggressive thrombotic tendencies, with concurrent thrombocytopaenia [19] Management is complex; these patients should be discussed urgently with Haematology. |
| Polycythaemia vera | 2.5–5% VTE risk per annum [20] |
| Essential Thrombocytosis | 1.9–3% VTE risk per annum [20] |
| Indication for anticoagulation | Thrombosis risk |
|---|---|
| Atrial fibrillation | CHA2DS2VASC >3 = stroke risk 1.6 in 1000 at 5 days [17]. Up to 12% at 1 year for CHA2DS2VASC 9. |
| Venous Thromboembolism | Recent VTE (<3 months) = high risk; discuss with haematology and consider IVC filter [18] Unprovoked VTE, >3 months ago = 22.6% recurrence after 5 years [18] |
| Mechanical heart valves | Off anticoagulation, total risk is 8.6% per annum; major embolism 4% [16] Bileaflet AVR 0.5 in 1000 at 5 days [18] MVR 1.6 in 1000 at 5 days [18] |
| Antiphospholipid syndrome | Prior thrombosis history in APS patients can convey aggressive thrombotic tendencies, with concurrent thrombocytopaenia [19] Management is complex; these patients should be discussed urgently with Haematology. |
| Polycythaemia vera | 2.5–5% VTE risk per annum [20] |
| Essential Thrombocytosis | 1.9–3% VTE risk per annum [20] |
It is therefore clear that even in patients at the highest risk of thrombosis, almost all patients with acute intracranial bleeding will benefit from stopping anticoagulation for a period. At the present time, few data inform decisions surrounding anticoagulation cessation or reversal in the context of cSDH, where there is no acute bleeding. Clinical practice therefore varies on this point.
The same principles apply for antiplatelet therapy, where continued antiplatelet use is also associated with incremental morbidity and mortality after TBI [21]. Pooled RCT data have shown in-stent thrombosis arises in up to 48.2% after cessation of dual antiplatelet therapy within 30 days of percutaneous coronary everolimus-eluting stent deployment; management of this group of patients is therefore particularly complex [22]. However, observational evidence suggests that discontinuation of antiplatelet therapy after 30 days following drug-eluting stent deployment does not appear to have a large impact on risk of in-stent thrombosis or major cardiac event [23].
Therefore, in patients with acute intracranial bleeding who are also within 30–90 days of coronary drug-eluting stent deployment, some clinicians will continue aspirin monotherapy to minimise risk of cardiac ischaemic events. Discussion with the treating cardiology team is advised to determine the precise stent deployed to allow a personalised risk assessment.
The risk of recurrence with antiplatelet discontinuation <30 days following ischaemic stroke appears to be significantly lower, with recurrent events seen in only 0.69% [24].
Specific Interventions
When a decision has been made to reverse an anticoagulant, immediate action depends on the nature of the anticoagulant or antiplatelet used; discussion with haematology is advised.
Warfarin
Patients on warfarin should undergo immediate reversal with prothrombin complex concentrate (PCC) and receive 5–10 mg of intravenous vitamin K, targeting an INR <1.3. INR should be rechecked at 30 minutes, 4–6 hours and 24 hours, with further doses of PCC given as necessary to achieve full reversal.
Dabigatran
This can be reversed with idarucizumab.
Factor Xa inhibitors
Acute reversal of apixaban, rivaroxaban and edoxaban is currently recommended with activated prothrombin complex concentrate. Andexanet alfa, a recombinant human form of factor Xa, is a novel reversal agent for factor Xa inhibitors. Trial evidence is encouraging, with effective haemostasis achieved in up to 79% of cases [25].
Unfractionated (UFH) and Low Molecular Weight Heparin (LMWH)
Protamine sulfate is used for reversal of heparins.
Reversal of antiplatelets
Role of Platelet Transfusion
Platelet transfusion in the context of acute intracranial bleeding should be reserved for thrombocytopaenia, where platelets should be kept>100 109/L. Some centres advocate for platelet transfusion prior to surgery in all patients receiving antiplatelet therapy, though data to support this practice are scarce. Some observational data suggest no difference in recurrence risk of cSDH between platelet transfusion and non-platelet transfusion groups [26]. Other data have observed increased morbidity and mortality in patients undergoing platelet transfusion [27]; clinical trials are therefore required.
Desmopressin and Tranexamic Acid
The effectiveness of these agents in reversal of antiplatelet effect remains unclear; their use is therefore not currently recommended [28].
The role of delaying surgery after cessation of antiplatelet therapy in cSDH
There is significant variation in practice around timing of burr hole drainage of cSDH following cessation of antiplatelet therapy. One multicentre UK study showed that burr hole drainage for cSDH was carried out a median of 3 days after cessation of aspirin and 6 days after cessation of clopidogrel. This study found no significant difference in recurrence rate between groups operated on after a washout period of <3 days versus a washout period of 3 days or more [26].
Venous thromboembolic (VTE) prophylaxis
The risk of VTE following traumatic SDH has been previously been shown to vary between 1 and 20%, with a more recent large observational study showing a rate of 4.9% [29]. Risk is higher in older people, especially in the presence of other traumatic injuries and longer lengths of hospital stay.
Mechanical VTE prophylaxis should be commenced at the earliest opportunity in the absence of lower limb contraindications. However, there is a paucity of evidence to guide timing of chemoprophylaxis in the context of SDH to guide practice. A recent systematic review in the context of acute TBI concluded that chemoprophylaxis with low molecular weight heparin (commenced 24–72 hours after injury, following a repeat CT head showing stable appearances) was associated with reduced VTE incidence without increased risk of haematoma expansion [30]. Whilst this evidence is most relevant to acute SDH, the risks of bleeding in chronic SDH are likely to be considerably lower. This may influence decisions on when chemoprophylaxis can safely be provided.
The role of inferior vena cava (IVC) filters
Indiscriminate use of IVC filters is associated with net patient harm. UK guidelines recommend consideration of IVC filters in patients with proximal (above knee) DVT or PE when anticoagulation is contraindicated [31]. IVC filters should be used for the shortest possible duration and removed as soon as anticoagulation can be restarted.
The role of neurosurgical intervention
Acute SDH and chronic SDH are very different entities from a neurosurgical perspective. In the context of trauma, aSDH is usually suggestive of a significant primary brain injury. Surgical intervention for acute SDH requires either craniotomy with haematoma evacuation or occasionally decompressive craniectomy. A recent randomised control trial has shown comparable disability and quality of life data between craniotomy and craniectomy. Rates of reoperation are higher in those undergoing craniotomy, although wound complications are around three times as high in those undergoing craniectomy [32].
In contrast, the liquified blood products of chronic subdural haematoma are amenable to burr hole drainage, which is substantially less invasive. To date, no randomised control trials have examined the outcome of surgical versus conservative management of either acute or chronic SDH. Decision-making is therefore based on observational data.
Acute SDH
Unsurprisingly, aSDH has a significantly poorer prognosis than cSDH or acute on chronic SDH; inpatient death following aSDH is approximately 15 times higher than cSDH [33]. Increasing frailty is associated with significantly higher risk of 30-day mortality and 6-month unfavourable neurological outcome [34].
US guidelines recommend consideration of surgical intervention for aSDH where there is CT evidence of clot thickness > 10 mm, or midline shift >5 mm [35], where there is considered ‘potential for good recovery’.
However, a 2021 systematic review and meta-analysis evaluated outcomes in patients aged >65 following surgery for aSDH [36]. Mortality was 40% at the point of discharge and 49% at long term follow up. In patients who survived, poor neurological outcome (Glasgow Outcome Scale 1–3, Table 2), was 81% at discharge and 79% at long term follow up (Figure 1). Outcomes of aSDH in people aged >65 therefore appear to be poor even with surgery.
Glasgow Outcomes Scale (GOS)
| Glasgow Outcome Scale | Definition |
|---|---|
| 1 | Death |
| 2 | Persistent vegetative state |
| 3 | Dependent for support daily |
| 4 | Independent but some disability |
| 5 | Good recovery |
| Glasgow Outcome Scale | Definition |
|---|---|
| 1 | Death |
| 2 | Persistent vegetative state |
| 3 | Dependent for support daily |
| 4 | Independent but some disability |
| 5 | Good recovery |
Glasgow Outcomes Scale (GOS)
| Glasgow Outcome Scale | Definition |
|---|---|
| 1 | Death |
| 2 | Persistent vegetative state |
| 3 | Dependent for support daily |
| 4 | Independent but some disability |
| 5 | Good recovery |
| Glasgow Outcome Scale | Definition |
|---|---|
| 1 | Death |
| 2 | Persistent vegetative state |
| 3 | Dependent for support daily |
| 4 | Independent but some disability |
| 5 | Good recovery |
Outcomes of surgical intervention for aSDH in patients aged >65
Age > 80 has also been shown to be predictive of mortality in traumatic aSDH [37]. A UK observational study demonstrated that poor neurological outcome (GOS 1–3) was predicted by age > 85 years; moreover, all patients aged >90 years with GCS <10 at presentation had poor outcome [33]. These data inevitably influence case selection: in the UK, few patients aged >80 undergo surgical decompressive haemicraniectomy for traumatic aSDH, though a small proportion with spontaneous aSDH some may have better outcomes.
More commonly, older patients with aSDH are clinically observed and supported for a 2-week period. During this interim period, the acute haematoma typically evolves to cSDH consistency, potentially rendering the cSDH amenable in surviving patients to burr hole treatment.
Chronic SDH
Definitive neurosurgical treatment of symptomatic cSDH is typically indicated when cSDH >10 mm, or midline shift >5 mm. Data demonstrate that even age > 90 is not associated with higher rates of perioperative morbidity or mortality following burr hole decompression [38].
Case selection is complicated by challenges in determining the precise impact of cSDH, because older people may not display characteristic focal neurology. Epidemiological studies indicate that the most common presenting symptoms of cSDH are cognitive impairment and headache. These are common symptoms in sick older adults and can often be attributed to myriad acute medical pathologies, especially in patients with coexistent background cognitive impairment.
Therefore, where cSDH and acute medical pathology co-exist, it can be difficult to determine the relative contribution of cSDH to cognitive decline. A careful history from a collateral source can be very influential, especially where decline in cognition has been sub-acute, or where the accompanying medical illness is mild and incongruent with the scale of cognitive deterioration. Iterative evaluation after treatment of other potential causes of delirium may be required to ascertain the relative contribution of the cSDH.
Patients who are considered asymptomatic, or only very mildly affected by their cSDH, should be managed conservatively, as the benefits of surgery (albeit moderate) are outweighed by the risks.
Burr hole drainage is commonly performed as the first line surgical intervention for symptomatic cSDH in patients considered to have potential for recovery. This procedure involves the drilling of two holes through the skull vault, usually under general anaesthetic, prior to haematoma washout. Concurrent subdural or subgaleal drain placement has become common practice, having been shown to reduce recurrence rate with no increase in post-operative complications [39]. A recent meta-analysis has compared outcomes of subdural versus subgaleal drain placement; this study showed no significant difference in recurrence of cSDH, nor clinical neither radiological outcome [40]. Subgaleal drains are less invasive than their subdural counterparts with lower risk of parenchymal tissue injury [41]. They may therefore become preferred in the light of these very recent data.
US guidelines recommend surgical intervention for cSDH in cases where there is radiological mass effect or midline shift and where there is ‘moderate to severe cognitive impairment, or progressive neurological deterioration attributable to the cSDH, in patients with potential for recovery’ [35].
Few data have examined associations between frailty and outcomes in cSDH. One retrospective cohort study indicated that patients with frailty had higher risk of poor functional outcomes than those without frailty. Only 35% of frail patients included were discharged home following surgical intervention versus 91% of the non-frail. However, no long-term follow up status was reported [42]. A subsequent study has also shown frailty to be associated with lower rates of functional independence and recovery following burr hole drainage [43].
These data echo previously demonstrated associations between frailty and poor functional outcomes in older trauma patients [44]. Comparative data are needed to evaluate outcomes in matched older patients who did not undergo surgery. One small scale study has shown that patients aged >90 years managed surgically have higher rates of inpatient and long-term survival, and higher levels of functional independence than matched patients managed conservatively [45].
Age alone should therefore not be considered a barrier to surgery in patients symptomatic with cSDH, who may have potentially reversible pathology.
Middle Meningeal Artery Embolisation (MMAE)
MMAE is an emerging treatment for cSDH and acute-on-chronic SDH. The MMA provides blood supply to the newly neovascularised dural membrane in cSDH: this process is believed to contribute to haematoma expansion and recurrence of SDH. MMAE halts the blood flow to this friable capillary nexus and is strongly associated with haematoma resolution [46]. MMAE can be performed in isolation, pre- or post-burr hole surgery, depending on severity of symptoms and level of surgical risk. It is a low risk and minimally invasive procedure performed by interventional radiology. Complication rates of MMAE have shown to be low, at around 1% [47].
Observational data for MMAE are encouraging. A recent meta-analysis and systematic review shows improved rates of treatment failure, surgical rescue and cSDH resolution with MMAE as an adjunct to conventional treatment. No increase in complications, mortality, or long-term disability was found. Number needed to treat (NNT) for surgical rescue is 9 and NNT for complete resolution of cSDH is 3 [46].
Randomised trials are underway and NICE guidance is anticipated [48]. In the interim, potential treatment algorithms have been suggested [49]; Figure 2 outlines our current approach.
Proposed treatment algorithm for management of cSDH. MMAE: middle meningeal artery embolization.
Supportive medical care
Management of thrombocytopaenia
Thrombocytopaenia is associated with a significantly increased risk of haematoma expansion; patients with a platelet count of <135 × 109/L at the time of injury are >30 times more likely to require neurosurgery [15]. Thrombocytopaenia should therefore be avoided in patients with acute SDH, and platelets should be kept >100 × 109/L, during the acute phase of intracranial bleeding, transfusing as necessary [15]. This will be impractical in some patients (e.g. those with chronic profound thrombocytopaenia in the context of myelodysplastic disorders); such cases should be discussed with haematology regarding realistic platelet targets, and for how long aggressive maintenance of platelet targets should be pursued.
Blood pressure management
Blood pressure targets in patients with SDH are controversial. Brain Trauma Foundation guidelines recommend targeting systolic blood pressure (SBP) >100 mmHg in patients aged 50–69, and SBP >110 mmHg in patients aged >70 [50]. A recent retrospective study has compared outcomes in SDH patients treated with SBP targets of 100–150 mmHg or SBP < 180 mmHg within the first 24 hours of injury. This study found no significant difference in 30-day mortality, or rates of subsequent surgical intervention between the two groups [51].
Bedrest
Bedrest following burr hole drainage of cSDH has historically been common practice in many centres until subdural drain removal. The GET-UP trial [52] demonstrated that early mobilisation was associated with preserved functional status and a significant reduction in post-operative infections. Furthermore, early mobilisation did not result in increased SDH recurrence. Bedrest is therefore not recommended.
Seizure Prophylaxis
Seizures in the aftermath of traumatic brain injury can result in secondary ischaemic injury; prophylactic anticonvulsants have therefore been proposed to reduce this complication. However, anticonvulsants convey a substantial side effect profile. Moreover, systematic review data have not demonstrated a significant reduction in seizure frequency with empirical anti-seizure medication in the context of isolated SDH, where the absence of parenchymal brain injury is less epileptogenic than cerebral contusions and subarachnoid haemorrhage [53]. Outside of a trial setting, anticonvulsants should therefore be reserved for patients with SDH demonstrating clinical or electroencephalography evidence of seizure activity.
Dexamethasone
Concepts around the pathophysiology of cSDH have evolved over the last decade; it is now believed that cSDH evolution is driven by a cascade of inflammatory processes triggered by bleeding and trauma. Attention has therefore been paid to the merits of corticosteroids in the management of SDH. Whilst trial data have demonstrated that dexamethasone is associated with reduced recurrence rate of SDH, they also demonstrated higher rates of mortality and poorer functional outcome at 6 months [54]. Dexamethosone use is therefore not currently recommended, although this may change with emerging data in future.
Atorvastatin
Observational data have indicated that low-dose atorvastatin (20 mg) was associated with improved clinical outcomes in patients with cSDH managed non-operatively. These outcomes included greater reduction in haematoma volume and lower rates of recurrence and surgery [55]. The pharmacological basis for this finding is currently unclear; more evidence is required before statin therapy forms part of routine UK management of cSDH.
Tranexamic Acid (TXA)
TXA is used in mild–moderate TBI, when given within 3 hours of injury [56]. However, evidence of benefit in management of SDH is limited. Routine use is therefore not recommended.
The role of repeat scanning
Much heterogeneity exists in relation to the decision to re-scan patients with known intracranial bleeding. In our practice, we tend to reserve repeat imaging for patients with neurological deterioration (GCS drop ≥2). Some UK centres routinely repeat CT scanning, including following burr hole drainage for cSDH, to evaluate the surgical result although some studies indicate that there is little advantage to this approach over alternatives driven by symptomatic deterioration [57].
The role of repeat CT scanning in medically managed patients is also controversial. One study of TBI patients with stable neurological signs found that only one patient required surgery following repeat routine scanning [58]. Routine rescanning in neurologically stable patients therefore has a low yield in identifying patients who require surgery.
However, repeat scanning to assess for stable intracranial appearances can be supportive prior to restarting anticoagulation or antiplatelet therapy: the presence of new acute bleeding can inform the risk versus benefit appraisal, and may support deferred initiation of anticoagulation.
Restarting anticoagulation/antiplatelet agents following SDH
Similar principles apply to the decision to restart anticoagulation following SDH as to the decision to suspend anticoagulation at the point of diagnosis. However, the risk-benefit profile changes with time, and therefore needs constant re-evaluation as the risk of rebleeding falls.
The greatest risk posed by reinstated anticoagulation is haematoma expansion, which can have devastating consequences causing secondary neurological injury. Haematoma expansion in acute TBI is very common in the first 24 hours, but falls by 48 hours and is rare by 72 hours [59]. However, SDH behaves differently to intraparenchymal and subarachnoid bleeding. Acute SDH commonly expands and/or rebleeds in the days and weeks after onset, prior to transforming into chronic SDH. A large cohort study demonstrated the cumulative risk of recurrent aSDH to be 9% during the first four weeks following initial aSDH. This subsequently increased to 14% at one year. Risk of recurrence is higher in patients treated with surgical intervention compared with conservative management. Age was also shown to be a significant predictor of recurrence, with highest risk in those aged 70 years and over [60].
Some clinicians therefore argue that there is a need for greater caution in SDH when restarting anticoagulation. However, failure to restart anticoagulation (where indicated) in patients with TBI contributes to significant risk of thromboembolic events (Table 1), which can equally cause major neurological injury or even death. Personalised decision making must therefore be undertaken on a case-by-case basis after appraising the risks of reinstated anticoagulant against those of discontinuation. Middle meningeal artery embolisation is also worth considering in people needing to urgently restart anticoagulation, as this may reduce risk of rebleeding during the highest risk period.
Unfortunately, there is a paucity of data available to guide the timing of reintroduction of anticoagulant therapy. Clinical practice is therefore heterogenous. Clinicians often accept reinstated anticoagulation at time-points ranging from 72 hours following injury to several months later, though most clinicians consider reinstated anticoagulation to be safe in most circumstances after 2 weeks [61].
Timing of resumed anticoagulation following traumatic intracranial haemorrhage is currently being evaluated in the RESTART-ICH trial, where reinstatement of anticoagulation is randomised to 1–4 weeks post-injury. The trial is anticipated to report in 2027 [62].
High Risk Scenarios
In patients with a high-risk indication for anticoagulation (e.g. recent VTE or metallic heart valve), case series data and expert consensus support suspension of anticoagulation for 7–10 days. In a survey of over 500 experts, the range for restarting anticoagulation was 3–14 days after the bleed, with a median of 6–7 days.
In patients with a very high-risk indication for anticoagulation, where anticoagulation must be restarted <7 days post-injury to save life or limb (e.g. high-volume central pulmonary embolus), split- bolus treatment dose LMWH is a favoured approach in some centres, where it has the advantage of being reversible with protamine in the event of re-bleeding. Furthermore, the split dose regimen may offer more predictable anticoagulant effect than a continuous IV heparin infusion, which can be challenging to administer outside of a high-dependency environment.
UK DVLA Considerations
UK Driver and Vehicle Licensing Agency (DVLA) restrictions exist for SDH [63] and patients should always be advised to seek up to date advice, as regulations are constantly reviewed. At the time of writing, UK patients suffering traumatic SDH who hold a Group 1 licence must not drive and must inform the DVLA; time off driving totals at least 6 months. If surgical intervention has taken place, the restrictions for that procedure may also apply. Restrictions around spontaneous SDH vary slightly but where surgery is undertaken, a 6-month suspension is required; details can be found on the DVLA website.
However, patients suffering cSDH or acute-on-chronic subdural haematoma, even if treated surgically, can at the time of writing resume driving upon recovery (on a group 1 licence).
In patients who have high risk of SDH recurrence, for example in those with membrane formation, restrictions can be more nuanced; this requires case by case review, and where any doubt exists, direct consultation with the DVLA should be undertaken.
Conclusion
The incidence of SDH is rising due to an ageing population, increased use of anticoagulants and improved rates of diagnosis. Few high-quality studies exist upon which to base clinical practice. Physicians in the future will increasingly be called upon to manage the care of older patients who are not surgical candidates, or to help neurosurgical services determine which patients are likely to benefit from surgery. Physicians have a role in supporting complex decisions, especially in relation to anticoagulation suspension and reinstatement. A working knowledge of the epidemiology surrounding SDH is needed to contribute to these discussions and ensure that patients benefit from an accurate risk assessment and evidence-based care.
The ‘Improving Care in Elderly Neurosurgery Initiative’ comprises a cross-disciplinary group, seeking to improve care of older people with cSDH. They are in the process of forming consensus guidelines, which we anticipate will provide further invaluable guidance in this arena [64].
Declaration of Conflicts of Interest
None.
Declaration of Sources of Funding
None.
Comments