Acute non-traumatic intracerebral haemorrhage (ICH) has a poor prognosis and is the least treatable form of acute stroke. Although less common than acute ischaemic stroke, ICH causes greater premature loss of productive life years on a global scale due to its predilection to affect people at younger ages with devastating consequences. Prognosis from ICH is related to location, initial volume and speed of expansion of the haematoma, and associated intraventricular haemorrhage. Care in a specialised stroke or neurointensive care unit improves outcome. Surgical haematoma evacuation should be pursued as for patients with cerebellar haemorrhage with neurologic deterioration, hydrocephalus, orsd brainstem compression. Haematoma evacuation may be considered, as a life-saving measure, in patients with coma or large haematoma with mass effect. Minimally-invasive surgery in stable patients is of uncertain benefit and is being evaluated in RCTs. Clinicians should not routinely use haemostatic therapies where there is no evidence of coagulopathy or anticoagulant use. When coagulopathy is present, early corrective measures should be taken. Early moderate intensity BP lowering to a systolic BP target of 140 mmHg is reasonable. Medical therapies to reduce mass effect and intracranial pressure should not be used routinely, but hyperventilation and hypertonic saline or colloidal osmotic agents are reasonable in patients with imminent herniation as a bridge to definitive neurosurgical intervention. Corticosteroids should be avoided. Novel neuroprotective approaches hold promise.
Acute non-traumatic intracerebral haemorrhage (ICH) is the most common manifestation of haemorrhagic stroke, which occurs as an isolated haematoma within the brain or extends directly (or indirectly) internally into the ventricular system or externally across the meninges (Figure 13.1). In general, ICH accounts for approximately 10% of strokes, but the proportion is much greater (20–50%) in particular racial-ethnic groups, including Asians, Blacks, and Hispanics, with leading risk factors of older age, male sex, hypertension, poorer socio-economic status, and colder environmental temperature (Feigin et al., 2009; Poon et al., 2015). As haemorrhagic forms of stroke have greater severity than acute ischaemic stroke, the adverse outcomes are disproportionately higher: 49% of all deaths and ~42% (47 million) of all ‘disability-adjusted life years lost’ due to stroke worldwide each year (Feigin et al., 2015).
Figure 13.1 Types of intracranial haemorrhage (Al-Shahi Salman et al., 2009).
There are several reasons why prevention should be the priority of management for ICH, including poor prognosis and limited acute treatments, with case fatality not dramatically changing over recent decades (van Asch et al., 2010; Poon et al., 2014); and, apart from age, the most important risk factor – high blood pressure (BP) – can be effectively modified (O’Donnell et al., 2010). Since two-thirds of incident ICH occurs in people aged ≥75 years in high-income countries (Samarasekera et al., 2015), the burden is likely to rise further as the populations age and primary prevention strategies remain inadequate.
ICH is a critical illness requiring rapid assessment and management, as patients are often unstable, have altered level of consciousness, and may rapidly deteriorate. An assessment of clinical severity and certain imaging features (i.e. site, volume, and presence of intraventricular haemorrhage [IVH]) of the haematoma are useful in triaging patients for intensive monitoring, respiratory support, and surgical intervention. Early use of palliative care plans have been shown to adversely influence outcome (Parry-Jones et al., 2016) and should be used cautiously. An active management plan is recommended unless there are clear clinical (e.g. massive ICH with deep coma), comorbid (e.g. advanced dementia or malignancy), or patient autonomy (e.g. advance directive) circumstances to suggest otherwise. Active care allows time to review the rapidity of any progression, for potential complications (e.g. seizures and dehydration) to be resolved, and for family members to adjust to the crisis and receive counselling relevant to their cultural, religious, and personal beliefs. Active care includes monitoring and timely intervention for neurological deterioration and adverse events, which is ideally organized in an intensive care or high dependency unit, where there is a high nurse:patient ratio of care and expertise. Well-organized, acute stroke unit care benefits patients with ICH to the same extent as those with acute ischaemic stroke (Langhorne et al., 2013) by directing effective management according to the type and severity of neurological impairment.
Noncontrast computed tomography (CT) and brain magnetic resonance imaging (MRI) are appropriate initial investigations to diagnose ICH. CT is generally more readily available, but is less useful in establishing any underlying structural cause, which should be suspected when the haematoma is in a lobar or an atypical deep location; when there is a disproportionate amount of subarachnoid haemorrhage or perihaematomal oedema; and in younger patients without a history of hypertension or illicit drug use as predisposing factors. The initial brain imaging may identify the likely cause, such as tumours, large vascular malformations, or underlying infarctions with haemorrhagic transformation. CT or MRI angiography should be the next step to further screen for any underlying vascular anomaly, such as an intracranial aneurysm, arteriovenous malformation, or cavernous angioma; acquired arteriopathies, such as vasculitis or moyamoya disease; reversible vasoconstriction syndrome; cerebral venous sinus thrombosis; and other causes. Catheter cerebral angiography is required where prior investigations identify a potential macrovascular abnormality; suspicion remains high despite negative screening tests (macrovascular abnormalities can go undetected by CT and MR [van Asch et al., 2015]); or interventional treatment is being considered.
Although early mobilization of patients is generally considered beneficial for maintaining muscle strength and to reduce risks associated with immobility, particularly venous thromboembolism, very early and intensive mobilization is not indicated in ICH patients. A large multicentre randomized trial compared very early (median start 18.5 hours after onset), high intensity mobilization with conventional early (median start 22.4 hours after onset), standard intensity mobilization, and found that standard early mobilization was associated with improved outcomes, particularly in those with ICH (AVERT Trial Collaboration Group, 2015). Although elevating the head of patients with large ICH may reduce intracranial pressure, one large randomized trial showed no benefits (or harms) of positioning the patient to be sitting up versus lying flat after acute ICH (Anderson et al., 2017a).
The manifestations of ICH depend on the location of the initial site/tract of haemorrhage and subsequent mass effect from the expanding haematoma on adjacent structures (see Box 13.1). Haematoma expansion or ‘growth’ occurs in most patients with ICH: its recognition depends in part on the velocity of growth and in part on the time intervals from symptom onset to initial, and subsequent, brain imaging. Haematoma growth immediately after haemorrhage onset occurs in all patients and contributes to prehospital and early post-arrival neurological deterioration, which occurs in 31% of ICH patients (Shkirkova et al., 2018). Continued radiological haematoma growth between initial post-arrival brain imaging and follow-up imaging affects a fifth of patients with acute ICH, is more common in patients arriving earlier after onset, taking antiplatelet or anticoagulant agents, and with larger haematoma volumes on initial imaging, and is associated with a poor outcome (Davis et al., 2006; Al-Shahi Salman et al., 2018). When ICH is related to anticoagulation therapy, haematoma growth often continues over a more protracted time period. In addition to contributing to local pressure and midline shift, haematoma growth can track into the ventricular system (IVH), causing hydrocephalus, further mass effect, increased intracranial pressure (ICP), and reduced cerebral perfusion. Whether or not ICH growth occurs, any extravasation of blood into the brain parenchyma promotes perihaematomal oedema, initially from infiltration of blood plasma and later from neurotoxicity related to haemoglobin breakdown products causing inflammation (Xi et al., 2006; Yu et al., 2017), and greater oedema worsens prognosis (Yang et al., 2015). Each of the phases of the pathophysiology of ICH influences the degree and rate of recovery and potential therapeutic targets (Poon et al., 2014; Specogna et al., 2014).
Loss of consciousness at presentation (Glasgow Coma Scale, GCS)
Progressive neurological deterioration (National Institutes of Health Stroke Scale, NIHSS)
Location of haematoma
– internal capsule – damage to motor tracts
– basal ganglia/thalamus – risk of IVH
– brainstem/cerebellum – greater risk from mass effect
– volume – more blood, worse outcome
– volume increase – ongoing bleeding, worse outcome
– morphology – irregular borders and heterogeneous consistency (e.g. swirl1 sign) indicate active bleeding
– spot sign2 – positivity on CT angiography indicates active bleeding
Presence of IVH – worse outcome
Hypertensive response – higher systolic BP, worse outcome
Hyperglycaemia – higher serum glucose, worse outcome
Older age – the older brain (with cerebral atrophy) can accommodate greater mass effect from the ICH, but elders have reduced brain/physiological reserve, greater comorbidities, and higher risk of cardiovascular complications
In this chapter we will outline available treatments for acute ICH, rationale for their use, level of supporting evidence for efficacy and safety, and implications for practice and research. A hierarchical approach has been taken in the selection of Level 1 or 2 of evidence for recommendations; a lower level of evidence is based on expert consensus and summarized in current guidelines (Steiner et al., 2014; Hemphill et al., 2015).
Stroke units provide an integrated package of supportive care potentially beneficial for ICH, including active management of physiological variables relating to fluid balance, pyrexia, oxygenation, and glycaemia; early mobilization out of bed; skilled nursing; and multidisciplinary rehabilitation.
A formal meta-analysis identified 8 randomized controlled trials (RCTs) specifically reporting outcomes in 483 enrolled patients with ICH. Among ICH patients, stroke unit care reduced death or dependency, 46.8% vs 74.2%, risk ratio (RR) 0.79 (95% confidence interval [CI]: 0.61–1.00) (Figure 13.2). There was evidence of heterogeneity across trials (I2 = 62%), but this appeared to arise from bidirectional scatter among small sample size trials without evidence of biasing of the effect estimate. Stroke unit care reduced all-cause mortality, 24.4% vs 50.0%, RR 0.73 (95% CI: 0.54–0.97), without evidence of trial heterogeneity (I2 = 21%).
Figure 13.2 Effect of stroke unit vs general medical ward management of acute ICH on death or dependency at end of scheduled follow-up months (Langhorne et al., 2013).
Implications for Practice
Patients with ICH should be managed in dedicated stroke units or neurocritical care units that actively monitor and manage physiological variables relating to fluid balance, body temperature, and oxygenation; pursue early mobilization out of bed; have skilled, stroke-knowledgeable nursing; and deliver multidisciplinary rehabilitation.
Neurosurgery has the potential to improve outcomes from acute ICH through:
evacuation of the haematoma to reduce early mass effect and elevated intracranial pressure (ICP), and later haematoma-related neurotoxicity, via craniectomy, craniotomy, or minimally invasive approaches (e.g. endoscopic suction or catheter-based drainage)
This section considers both open craniotomy and a variety of minimally invasive neurosurgical approaches to haematoma evacuation. The potential physiological benefits of each approach towards clearance of ICH (± IVH) in relieving mass effect are counterbalanced by the risks of anaesthesia, particularly in an older multi-morbid patient group (Lovelock et al., 2007; Samarasekera et al., 2015); damage to healthy brain tissue trespassed upon to access the haematoma cavity; and post-operative complications (e.g. re-bleeding, infection, thromboembolism). Accordingly, evidence-based analysis and individual case clinical judgement are required. Importantly, it is desirable that improved survival does not solely result in more patients alive but severely disabled, with poor physical function and quality of life.
The latest Cochrane systematic review included 10 RCTs with a total of 2059 patients randomized to surgery with medical management versus medical management alone (Prasad et al., 2008). Patients undergoing surgery received either craniotomy (open surgery) or stereotactic/endoscopic (catheter-guided) haematoma evacuation. Among 9 trials enrolling 1996 patients contributing to this outcome, surgery combined with medical therapy compared with medical therapy alone reduced death or dependence, 63.1% vs 70.6%, odds ratio (OR) 0.71 (95% CI: 0.58–0.88; p = 0.001). No significant heterogeneity of effect across trials was noted (I2 = 25%). Among all 10 trials, surgery combined with medical therapy compared with medical therapy alone reduced mortality, 27.0% vs 33.6%, OR 0.74 (95% CI: 0.61–0.90; p = 0.003). For the mortality outcome, moderate heterogeneity across trials was observed (I2 = 49%), due at least in part to enhanced mortality benefit in some trials of stereotactic or endoscopic surgery compared with open craniotomy.
A subsequent meta-analysis additionally incorporating later reported trials included 15 RCTs enrolling 3366 patients (Mendelow et al., 2013). In this further analysis, surgery combined with medical therapy compared with medical therapy alone reduced death or unfavourable functional outcome, 59.3% vs 66.5%, OR 0.74 (95% CI: 0.64–0.86). However, there was substantial heterogeneity of effects across trials (I2 = 67%), heterogeneity p value = 0.0002), and both of the two largest trials, STICH 1 and STICH 2, were individually non-positive and had point estimates for treatment effect less than the aggregate, raising concerns regarding risk of bias of the included RCTs, in addition to uncertainties arising from heterogeneous designs, participants, interventions, and comparators.
In the Cochrane review, data on the effect of surgery by type of surgical procedure were available from 8 RCTs enrolling 1335 patients, including 566 (42.4%) enrolled in trials of stereotactic or endoscopic surgery plus medical therapy versus medical therapy alone, and 769 (57.6%) enrolled in trials of open craniotomy plus medical therapy versus medical therapy alone (Prasad et al., 2008). A non-significantly greater benefit with minimally invasive surgery than with open craniotomy was noted. For stereotactic or endoscopic surgery plus medical therapy compared with medical therapy alone, rates of death or dependence were 62.3% vs 71.0%, OR 0.66 (95% CI: 0.46–0.95); for open craniotomy plus medical therapy compared with medical therapy alone, 74.9% vs 78.4%, OR 0.82 (95% CI: 0.59–1.15) (Figure 13.3).
Figure 13.3 Comparison of surgery type vs conservative management of acute ICH on death or dependency at 6 months (Prasad et al., 2008).
A distinctive minimally invasive surgical technique for evacuation of deep ICH is aspiration via catheter assisted by the instillation of fibrinolytic drugs into the haematoma cavity to dissolve the clot for enhanced clearance. The Minimally Invasive Surgery plus recombinant Tissue plasminogen activator (rt-PA) for ICH Evacuation (MISTIE) phase 2 RCT compared imaged-guided catheter placement with local rt-PA administration (0.3 mg or 1.0 mg, every 8 hours for up to 9 doses) added to standard medical care compared to standard medical care alone in 96 randomly allocated patients (Hanley et al., 2019). Patients were enrolled with ICH of at least 20 millilitres (mL) in volume that had been stable in size for at least 6 hours. Haematoma volume at end of treatment, typically around day 4, was reduced in the aspiration group, 20 mL vs 41 mL, p < 0.0001, and there was a nonsignificant increase in the rate of being capable of bodily self-care or better (modified Rankin Scale [mRS] 0–3) at 6 months, 33% vs 21%, hazard ratio (HR) 1.32 (95% CI: 0.62–2.82). No significant differences were noted in the lead safety outcomes of 30-day case fatality, 7-day procedure-related case fatality, 72-hour symptomatic bleeding, and 30-day brain infections, although asymptomatic haemorrhages were more common in the surgical group. These results provided support for the phase 3, MISTIE III trial that randomly assigned 506 ICH patients to catheter placement with local administration of 1.0 mg rt-PA every 8 hours for up to 9 doses or standard medical care. Fibrinolytic-assisted catheter aspiration was associated with lower mortality at 1 year (20% vs 27%, HR 0.67, p = 0.04), but also with nominally increased severely disabled outcomes (mRS 4–5: 42% vs 38%), with resulting no net significant effect on the primary outcome of ambulatory and capable of bodily self-care or better (mRS 0–3), 45% vs 41%, p = 0.33.
Additionally, the MISTIE investigators reported results of a parallel pilot RCT, the Intraoperative Stereotactic Computed Tomography-Guided Endoscopic Surgery (ICES) for Brain Haemorrhage study, that tested the safety and efficacy of image-guided endoscopic haematoma evacuation via a burr hole (Vespa et al., 2016). The investigators compared 14 patients receiving the endoscopic intervention with 42 patients receiving medical care only, with the medical care group comprised of both patients randomized to medical therapy within ICES and patients randomized to medical therapy in MISTIE II. On average, the procedure achieved an immediate reduction in haemorrhage volume of 68%, and a non-significantly higher proportion of patients who received surgical intervention made a good functional recovery (mRS score 0–3) at 180 and 365 days (42.9% vs 23.7%).
There have been no RCTs of surgical management of infratentorial ICH. A systematic review of observational studies included 792 patients with primary cerebellar haemorrhage: the overall case fatality was 31%, and surgery (with or without ventriculostomy) was non-significantly associated with a lower case fatality compared to conservative management (29% vs 33%) (Witsch et al., 2013).
A 2017 systematic review and meta-analysis of decompressive hemicraniectomy for spontaneous supratentorial ICH identified one small RCT (n = 40) and 227 cases from 7 observational studies (Yao et al., 2017). Outcomes of hemicraniectomy with or without haematoma evacuation were compared with outcomes of haematoma evacuation alone in the randomized trial and 6 of the 7 non-randomized observational studies, and with medical therapy alone in the remaining non-randomized observational study. Decompressive craniectomy compared with haematoma evacuation or medical therapy alone was associated with reduced frequency of poor outcome (Glasgow Outcome Scale [GOS] score 1–3, or mRS score 4–5), RR 0.91 (95% CI: 0.84–0.99). This result was heavily influenced by the single RCT (Moussa and Khedr, 2017).
Decompressive craniectomy was also associated with reduced mortality, RR 0.67 (95% CI: 0.53–0.85), which was more stable when subjected to sensitivity analyses but the authors were cautious in their interpretation from being unable to adjust for confounding factors. There was no excess of re-bleeding or hydrocephalus among decompressive craniectomy patients.
A multi-centre RCT, the Swiss trial of decompressive craniectomy versus best medical treatment of spontaneous supratentorial intracerebral haemorrhage (SWITCH) (NCT02258919), is currently ongoing with the aim of recruiting 300 patients to compare decompressive craniectomy with best medical management.
While no randomized trials have assessed use versus non-use of EVD in ICH, physiological reasoning and observational series provide support. In non-communicating hydrocephalus, the insertion of an EVD in a trapped ventricle and performance of cerebrospinal fluid (CSF) drainage will relieve local pressure; in communicating hydrocephalus, EVD placement and CSF drainage will reduce generalized intracranial pressure. In patients with IVH with or without ICH, a retrospective, single-centre case series of 183 patients noted that EVDs were placed in 37%. EVD placement was an independent predictor of reduced mortality (OR 0.31, 95% CI: 0.10–0.94) and discharge alive and capable of bodily self-care (mRS 0–3) (OR 15.7, 95% CI: 1.83–134.2) (Herrick et al., 2014). In a multicentre analysis of 563 patients with ICH with or without IVH, EVDs were placed in 25%. In propensity score adjusted analysis, EVD placement was associated with improved outcomes among ICH patients who also had IVH, with increased frequency of functional independence (mRS 0–2) at discharge (OR 8.43, 95% CI: 2.39–29.79) and a non-significant reduction in mortality at 30 days (OR 0.60, 95% CI: 0.34–1.05) (Lovasik et al., 2016).
Haemorrhages that are solely or predominantly intraventricular do not produce focal mass effect but can result in damaging hydrocephalus, including non-communicating (obstructive) hydrocephalus due to formation of large clots that plug a particular ventricular conduit and communicating (non-obstructive) hydrocephalus due to diffuse injury by blood degradation products to arachnoid granulations. A systematic review and meta-analysis of 8 small RCTs, 3 prospective cohorts, and 12 retrospective cohorts including 418 patients with ICH and IVH receiving intraventricular fibrinolysis versus 367 controls found that intraventricular fibrinolysis was associated with a lower case fatality (RR 0.55, 95% CI: 0.42–0.71), though benefit did not reach statistical significance in analyses confined to RCTs alone (RR 0.60, 95% CI: 0.35–1.01) (Khan et al., 2014). Intraventricular fibrinolysis was associated with increased good functional outcome (RR 1.66, 95% CI: 1.27–2.19), though not with statistical significance when analyzing only RCTs (RR 1.57, 95% CI: 0.89–2.80). Intraventricular instillation of fibrinolytics was not associated with increased risks of re-haemorrhage or hydrocephalus and was associated with reduced shunt dependence.
More recently, the Clot Lysis: Evaluating Accelerated Resolution of intraventricular haemorrhage phase 3 (CLEAR III) RCT enrolled 500 patients with IVH causing third or fourth ventricle obstruction and no more than 30 mL of accompanying supratentorial ICH (Hanley et al., 2017). All patients received clot removal via EVD, and were randomly allocated to 1 mg doses of rt-PA or placebo saline via EVD, up to a maximum of 12 doses, 8 hours apart. The primary outcome of capable of bodily self-care or better (mRS 0–3) at 180 days was similar in both groups, 48% vs 45%, RR 1.06 (95% CI: 0.88–1.28; p = 0.55). The alteplase group had reduced mortality through 180 days, 18% vs 29%, HR 0.60 (95% CI: 0.41–0.86; p = 0.006), but also a higher rate of severe disability (mRS 5), 17% vs 9%, RR 1.99 (95% CI: 1.22–3.26; p = 0.007). Symptomatic bleeding was similar between groups, and there were fewer cases of ventriculitis in the treatment group.
One RCT has evaluated lumbar drains as an add-on therapy to intraventricular fibrinolysis (IVF) for patients with severe IVH. The placement of lumbar drains may help restore physiological CSF circulation and prevent permanent shunt dependency. Among 30 randomized patients, lumbar drain placement reduced permanent shunt placement, 0% vs 43% (p = 0.007) (Staykov et al., 2017).
Interpretation of the Evidence
In view of the lack of clear benefit in individual large RCTs of surgical evacuation and heterogeneity across trials in meta-analyses, early surgery is not clearly beneficial as a treatment strategy for broad, relatively unselected groups of patients with ICH. Minimally invasive endoscopic and stereotactic approaches hold promise, and the results of further RCTs of these interventions and decompressive craniectomy as an adjunctive or alternative surgical manoeuvre are awaited (see below for list of ongoing trials).
Implications for Clinical Practice
At present, based on the available trials and observational series, for patients with supratentorial ICH, guidelines do not recommend early surgical intervention as a broad intervention for individuals who are clinically stable, but do recognize surgical treatment as of potential benefit in select circumstances (Steiner et al., 2014; Hemphill et al., 2015). European guidelines (Steiner et al., 2014) suggest surgical intervention may be of value for patients with a GCS score of 9–12. US guidelines recognize haematoma evacuation as having the potential to reduce mortality, but to increase severe disability, for patients experiencing neurological deterioration. Decompressive hemicraniectomy is also recognized as potentially life-saving, but with increased disability, for patients with coma, large haematomas with midline shift, or elevated intracranial pressure refractory to medical management (Hemphill et al., 2015).
For infratentorial haemorrhage, in the absence of RCT data, the most recent American Heart Association (AHA)/American Stroke Association (ASA) guidelines endorse early neurosurgical removal of haematoma in patients with cerebellar haemorrhage who are deteriorating neurologically, have brainstem compression, or have hydrocephalus, and emphasize that initial management with an EVD alone is insufficient (Hemphill et al., 2015). Evacuation of brainstem haemorrhage is considered harmful and is not recommended.
AHA/ASA guidelines recommend that insertion of an EVD should be considered for the monitoring and management of raised ICP in patients with a GCS score of ≤8, those with evidence of transtentorial herniation, and those with significant IVH or hydrocephalus.
Implications for Research
1. Minimally invasive surgery versus craniotomy in patients with supratentorial hypertensive intracerebral haemorrhage (MISICH) (NCT02811614)
2. Minimally invasive endoscopic surgery vs medical management in supratentorial intraparenchymal haemorrhage (INVEST) (Fiorella, 2016)
3. Minimally invasive surgery treatment for patients with spontaneous supratentorial intracerebral haemorrhage (MISTICH) (ChiCTR-TRC-12002026)
4. Decompressive hemicraniectomy in intracerebral haemorrhage (SWITCH) (NCT02258919)
Active ongoing bleeding, causing continued haematoma growth with additional tissue destruction and mass effect, is a mediator of poor outcome in acute ICH. The presence of the CT-angiography spot sign on early imaging, indicating ongoing bleeding, is associated with a 2.4-fold increase in mortality at 3 months (Demchuk et al., 2012). The rationale for using haemostatic drugs is to achieve early resolution of bleeding and attenuate haematoma growth. Several haemostatic therapies have been tested in acute ICH patients:
o Tranexamic acid – competitively inhibits activation of plasminogen to plasmin
o Provides functioning platelets to treat antiplatelet therapy-related ICH
A Cochrane systematic review identified 12 RCTs enrolling 1732 patients (1150 treated vs 582 controls or active comparators) (Al-Shahi Salman et al., 2018).
Three small RCTs of intravenous antifibrinolytic drugs (tranexamic acid in two and aminocaproic acid in the other) enrolled 57 patients (33 active, 24 control). Active antifibrinolytic therapy was associated with a non-significant reduction in ICH growth (RR 0.76, 95% CI: 0.56–1.05). However, antifibrinolytic therapy was also associated with non-significant increases in death or dependency (RR 1.25, 95% CI: 0.57–2.75), death (RR 1.16, 95% CI: 0.31–4.39), and any serious adverse events (SAEs) (RR 1.50, 95% CI: 0.39–5.83) and thromboembolic SAEs (RR 1.59, 95% CI: 0.07–35.15). More recently, the Tranexamic acid for Intracerebral Haemorrhage (TICH-2) RCT randomized 2325 ICH patients within 8 hours of onset to intravenous (IV) tranexamic acid, 1 g loading dose followed by another 1 g infused over 8 hours, or to matching placebo. On the primary outcome of global disability distribution over all 7 levels of the mRS at 3 months, IV tranexamic acid showed a non-significant shift to reduced disability, adjusted common odds ratio (acOR) 0.88 (95% CI: 0.76–1.03; p = 0.11). Although there were fewer early deaths by day 7 with IV tranexamic acid (aOR 0.73) deaths by 3 months were not lower (aHR 0.92, 95% CI: 0.77–1.10). No increase in venous or arterial thromboembolism was noted.
Blood Clotting Factors
A recent meta-analysis identified 7 RCTs, enrolling 1480 patients, that have tested blood clotting factors (predominantly rFVIIa) versus placebo/open control (Al-Shahi Salman et al., 2018). Treatment with blood clotting factors was associated with a non-significant reduction in death or incapacity for body self-care (mRS 4–6) in 6 trials that included 1390 patients, RR 0.87 (95% CI: 0.70–1.07) (Figure 13.4). Blood clotting factors were also associated with non-significant reductions in death (7 trials, 1480 patients: RR 0.75, 95% CI: 0.51–1.09), all SAEs (2 trials, 81 patients: RR 0.81 95% CI: 0.30–2.22), and ICH growth (3 trials, 151 patients: RR 0.74, 95% CI: 0.36–1.48), but a nonsignificant increase in thromboembolic SAEs (5 trials, 1398 patients: RR 1.24, 95% CI: 0.80–1.91).
Figure 13.4 The effect of blood clotting factors versus placebo/open control for acute ICH not associated with anticoagulant use, on death or incapacity for body self-care (modified Rankin Scale score 4–6) at day 90 (Al-Shahi Salman et al., 2018).
Two additional trials have presented pooled preliminary results, both testing rFVIIa in acute ICH patients showing the CT-angiography spot sign indicating ongoing bleeding and higher risk for haematoma expansion. Patients within 6.5 hours of onset were randomized to 80 mg rFVIIa or placebo. Among 69 randomized patients, haematoma expansion did not differ between the treatment arms. In the rFVIIa group, median volume evolution was from 16 mL at baseline to 22 mL at 24 hours, and in the placebo group from 20 mL at baseline to 29 mL at 24 hours, p = 0.9. Long-term functional outcomes and mortality also did not differ between the two treatment groups (Jeffrey, 2017).
The systematic review identified 1 small RCT (n = 13) that evaluated FFP against FFP with factor IX complex concentrate in patients with warfarin-related ICH and showed the addition of factor IX complex concentrate accelerated correction of the international normalized ratio (INR); however, this was not associated with a difference in the underpowered analysis of neurological outcomes (Boulis et al., 1999).
Also of relevance is the INR Normalization in Coumadin Associated Intracerebral Hemorrhage (INCH) RCT, which randomized patients with intracranial haemorrhage to four-factor prothrombin complex concentrate (PCC) versus fresh frozen plasma (Steiner et al., 2016). Among 50 randomized patients, the predominant haemorrhagic subtype was ICH, present in 88%, while the remaining 12% had subdural haematomas. PCC was superior to FFP on the primary endpoint of normalizing the international normalized ratio (INR <1.2) within 3 hours, 67% vs 9%, OR 30.6 (95% CI: 4.7–197.9; p = 0.0003), and reduced the amount of haematoma growth at 24 hours, 8.3 vs 22.1 mL (p = 0.02). There was also a non-significant reduction in mortality by 3 months in the PCC group (19% vs 35%, p = 0.14), but no increase in being alive and capable of bodily self-care or better (mRS 0–3) at 3 months (37% vs 39%, p = 0.47).
ICH Associated with Other Anticoagulants
There are no RCTs confined to ICH patients testing specific antidotes to other anticoagulants. Treatment is based on studies in broader groups of patients with major bleeding or planned surgery at diverse organ sites, often including a small subset with ICH, for example in studies of recently developed antidotes for non-vitamin-K oral anticoagulants (NOACs). In the single arm RE-VERSE AD trial studying idarucizumab to reverse the direct thrombin inhibitor dabigatran, ICH accounted for 53/301 (18%) of patients enrolled with life-threatening bleeding (Pollack et al., 2017). Idarucizumab normalized coagulation parameters within 10–30 minutes in more than 95% of patients. Thrombotic events occurred in 4.7% of patients by day 30. In the single arm, ANNEXA-4 trial studying andexanet alfa to reverse any of the four factor Xa inhibitors (apixaban, rivaroxaban, edoxaban, or enoxaparin), ICH accounted for 14/67 (21%) of patients enrolled with life-threatening bleeding (Connolly, 2016). Within 15–30 minutes, andexanet alfa reduced anti-factor Xa activity by a relative 86%–93%. Thrombotic events occurred in 18% of patients by day 30.
The Platelet Transfusion in Cerebral Hemorrhage (PATCH) RCT recruited 190 patients with ICH while on antiplatelet therapy and with normal platelet counts (Baharoglu et al., 2016): 97 were randomized to receive a platelet transfusion as soon as possible (within 6 hours of ICH onset and 90 minutes of the diagnostic scan) and 93 received standard care. Allocation to the platelet transfusion group was associated with a shift towards greater disability across all 7 levels of the mRS at 3 months, adjusted common OR 2.05 (95% CI: 1.18–3.56) (Figure 13.5). Patients treated with platelet transfusion tended to be more likely to have an SAE during the hospital stay, 42% vs 29%, OR 1.79 (95% CI: 0.98–3.27), including nominally more brain oedema and intraventricular extension SAEs. Reflecting the small sample size, the treatment groups had some imbalances in baseline prognostic features, including numerically larger ICHs >30 mL in the platelet transfusion group. Nonetheless, the results suggest no benefit and even potential harm from routine use of platelet transfusion in ICH patients taking antiplatelet agents.
Figure 13.5 Distribution of mRS scores at 3 months in PATCH trial (Baharoglu et al., 2016).
Thrombolysis with rt-PA for acute ischaemic stroke is associated with a 3–5% risk of major symptomatic ICH (sICH). However, there are no RCTs to guide management of this uncommon but potentially deadly complication. In 2015, a multicentre retrospective analysis of the treatment of thrombolysis-related ICH reported time to treatment, treatment modality, and clinical outcomes that included haematoma growth and in-hospital mortality (Yaghi et al., 2015). Of 3894 acute ischaemic stroke patients treated with thrombolysis, 128 (3.3%) had a symptomatic ICH. Overall, 49/128 (38%) received one or more haemostatic or surgical therapies, with the most common being cryoprecipitate in 40 (31%), platelet transfusion in 37 (29%), FFP in 26 (20%), and surgical haematoma evacuation or decompressive craniectomy in 18 (15%). Infrequent agents were vitamin K, PCC, aminocaproic acid, and rFVIIa. Median time from sICH diagnosis to treatment was 112 minutes. In-hospital mortality was 52% and was non-significantly lower in patients who received any form of treatment, 41% vs 59% (p = 0.11). Haematoma growth was more common in patients with severe hypofibrinogenaemia, 36% vs 25% (p = 0.01), lending support for the use of cryoprecipitate.
Interpretation of the Evidence
The available evidence does not support the routine use of haemostatic therapies after spontaneous ICH unrelated to antithrombotic drug use. Platelet transfusion seems harmful for ICH associated with antiplatelet drug use. For vitamin K antagonist-associated ICH, four-factor PCC seems superior to FFP for rapid INR reversal and averting haematoma expansion, but whether this translates into improved clinical outcomes has not been established. Reversal therapies for other anticoagulant agents are not supported by trials focused solely upon patients with ICH, but have some support from trials in patients with life-threatening bleeding of diverse types, among whom ICH patients constituted about one-fifth the study population.
Implications for Clinical Practice
Clinicians should not routinely use haemostatic therapies where there is no evidence of coagulopathy or anticoagulant use.
For ICH patients who are taking anticoagulation, a prompt assessment of type, last dose, and a coagulation screen are vital in the initial assessment. Immediate cessation of the anticoagulant followed by swift administration of a specific antidote or coagulation factors is recommended with guidance from a haematologist. Based on national guidelines drawing upon the evidence here reviewed (Steiner et al., 2014; Hemphill et al., 2015; Frontera et al., 2016) as well as subsequent trial data (Connolly, 2016; Pollack et al., 2017), recommended approaches include the following:
1. Unfractionated heparin (UHF): use protamine at a dose of 1 mg for every 100 units of UFH given in the previous 2–3 hours (maximum single dose 50 mg), repeat bloods for activated partial thromboplastin time (aPTT), and, if prolonged, give a repeat protamine dose of 0.5 mg per 100 units of UFH.
2. Low-molecular-weight heparin (LMWH): for enoxaparin, give 1 mg of protamine per 1 mg of enoxaparin if the dose was given within 8 hours; if the dose was given within 8–12 hours, then the protamine dose can be halved. Beyond 12 hours, no treatment is recommended. For other LMWH formulations, use 1 mg of protamine per 100 anti-Xa units administered in the past 3–5 half-lives of the drug. In renal insufficiency or ongoing bleeding, use half the dose of protamine and consider rFVIIa, if protamine is contraindicated.
4. Direct factor Xa inhibitors (e.g. rivaroxaban, apixaban): establish the timing of last dose. If within 48 hours (or longer if renal or hepatic impairment), administer andexanet alfa if available. If not available, give four-factor PCC. Also, if last dose was within 2 hours, consider activated charcoal for intubated patients.
5. Direct thrombin inhibitors (e.g. dabigatran): if the last dose was within 48 hours and there is no renal failure, administer idarucizumab to reverse dabigatran; where idarucizumab in unavailable, give PCC at a dose of 50 units/kg. If renal failure is present, consider reversal beyond this time, and re-dose if there is ongoing bleeding. Also, if last dose was within 2 hours, consider activated charcoal for intubated patients. Haemodialysis has a role where there is renal insufficiency and idarucizamab is unavailable, or when pharmacological therapy has been ineffective.
For patients with ICH after intravenous rt-PA, immediately discontinue rt-PA and administer cryoprecipitate (or tranexamic acid, if the former is contraindicated) as soon as possible (Frontera et al., 2016). Fibrinogen levels should be re-checked after administration of a reversal agent, and further cryoprecipitate given if serum levels remain low.