Thrombus formation occurs as the end result of a series of events and reactions. Damage to the endothelium initiates the coagulation cascade, aggregation of activated platelet and the generation of thrombin. Thrombin causes fibrinogen cleavage and the formation of fibrin to stabilise clot. Thrombin-mediated fibrin formation occurs in direct relation to platelet activation.

Platelets promote the activation of the early stages of the intrinsic coagulation pathway by a process that involves a factor XI receptor and high-molecular-weight kininogen, as well as factors V and VIII, which interact with platelet membrane phospholipids to facilitate the activation of factor X to Xa and the conversion of prothrombin to thrombin (see Tutorial 6).

Conversely, fibrinolysis is the process of dissolving clots through activation of the fibrinolytic system. Fibrin degradation is catalysed by plasmin, the product of plasminogen activation. Thrombus dissolution is, in large part, mediated by fibrinolysis inside the thrombus. In consolidating thrombus, plasminogen binds to fibrin and to platelets, allowing local release of plasmin within the thrombus. In addition, the circulating plasminogen activators, tissue plasminogen activator (tPA) and single-chain urokinase plasminogen activator (scuPA), catalyse plasmin formation and the degradation of fibrin to small fibrin degradation products (FDPs) and thereby clot lysis [1].

Early attempts to treat stroke patients using thrombolytic drugs were made in the 1970s by Fletcher et al. [2], but urokinase given 24 h after the onset of symptoms caused additional morbidity due to secondary haemorrhage, and its use in stroke stopped. Interest restarted 10 years later, when the benefit of thrombolytics was reported for patients with acute coronary thrombosis. They were initially only then used in stoke patients with vertebrobasilar thrombosis [3].

What was apparent by this stage was that most strokes were caused by arterial thrombosis. Solis et al. [4] reported that angiography, if performed within 12 h of the onset of symptoms, showed occluded arteries (appropriate to any neurological deficit) in 80–90% of patients. The percentage of identifiable thrombotic occlusions decreased markedly, if angiography was performed later, because spontaneous thrombolysis occurs. Thus, leaving aside the issue of neuron vulnerability, therapeutic thrombolysis needs to be performed early.

The aim of therapeutic thrombolysis is to stimulate fibrinolysis by plasmin, through infusion of analogues of tissue plasminogen activator (tPA). The ideal thrombolytic drug would activate only fibrin-bound plasminogen (plasminogen has a high affinity for fibrin) and not circulating plasminogen. This specificity is important because otherwise they risk causing consumption of plasma fibrinogen and generalised hypocoagulability because once activated, the clotting process continues to produce fibrinogen to make more clot and so becomes exhausted. These drugs are discussed in Tutorial 6.

Thrombolysis with drugs alone can potentially be used to treat any intravascular thrombosis. However, it may be unnecessary (because of spontaneous thrombolysis), dangerous (causing secondary haemorrhage of infarcted brain tissue) or cause general complications (by interfering with overall clotting). Thus, its indications in the nervous system are limited to: acute ischaemic stroke, central retinal artery occlusion, symptomatic dural sinus thrombosis and in other parts of the body to acute myocardial infarction, selective cases of pulmonary embolism, large artery thrombosis and Budd–Chiari syndrome with hepatic vein thrombosis. Restricting the use of these drugs to serious medical conditions acknowledges their dangers and leaves vascular surgeons, cardiologists and endovascular therapists to perform recanalisation procedures by thrombectomy (e.g. stent retriever, suction clot extraction), mechanical clot disruption or other techniques (e.g. angioplasty, stenting). Procedures that aim to re-establish blood flow to tissues supplied by occluded vessels without or with reduced doses of thrombolytic drugs.

Early randomised controlled trials (RCT) used streptokinase, but more recent trials of intravenous thrombolysis (IV-TLS) have used either urokinase or tPA analogues (i.e. recombinant tissue plasminogen activator or rtPA). They showed that treatment is effective if instigated in the first few hours after the stroke.

The major RCTs are ECASS I and II (European Cooperative Acute Stroke Study) [7, 8] and the NINDS trial [9]. In ECASS, tPA was given intravenously within 6 h of the onset of symptoms. It showed a statistically nonsignificant trend towards a benefit from treatment. The NINDS trial was a double-blind, placebo-controlled trial of tPA in doses of 0.9 mg/kg IV within the first 3 h of the onset of symptoms (10% given as an initial bolus and the remainder over 60 min). The result was a statistically significant benefit in long-term outcomes in the drug-treated group, who were 30% more likely to have minimal or no disability at 3 months than non-treated controls. Though the rates of neurological improvement at 24 h were the same and the haemorrhage rate at 36 h was slightly higher for treated patients (6.4 vs. 0.6%), the overall mortality rate (at 3 months) was still lower in the treated group (17 vs. 21%). The results of this study supported the use of tPA in patients who can be evaluated and treated within 3 h of the onset of symptoms. This conclusion was emphasised by the findings of another, similar double-blind trial (ATLANTIS), which failed to show a favourable result from treatment with tPA in the same intravenous doses and an increased frequency of symptomatic haemorrhage when given between 3 and 6 h following symptom onset [10].

The NINDS trial result was not supported by a meta-analysis of data from previous trials [1]. The difference was attributed to the very strict inclusion criteria in the NINDS trial (i.e. treatment within <3 h of symptom onset, mild deficit, blood pressure limits and no previous stroke) [11]. As a result, the uptake of IV-tPA in clinical practice has been slow because only a minority of stroke patients are eligible on these criteria [12]. Solutions are to extend the time-to-treat (TTT) limit and improve patient selection [13]. The more recent ECASS III showed that for IV-tPA TTT could be extended to 4.5–6 h [8]. Imaging has been a key element in selecting patients for thrombolysis, and it has become the focus for a great deal of research to develop better triage criteria. Its contribution is discussed below.

An important lesson in the evolution of the treatment guidelines for acute stroke is not to rely on data from small trials. An example is a trial to assess the efficacy and safety of abciximab. An initial trial showed some benefits with a lower risk of secondary haemorrhage in stroke patients [14], but a subsequent larger trial was stopped after 808 participants because of higher secondary haemorrhage rates in the treated patients [15].

The rationale of intra-arterial thrombolysis (IA-TLS) is to target the drug to the site of thrombosis. Its goal is the same as intravenous treatment, i.e. to limit the area of infarcted parenchyma and to enhance the survival of any functionally disabled cerebral tissue in the surrounding ischaemic penumbra. It is obviously only suitable for patients with proven intra-arterial thrombosis and in practice to those with thrombus in the internal carotid and middle cerebral arteries or vertebral and basilar arteries. The advantages over IV-TLS are higher recanalisation rates, only patients with demonstrated arterial occlusion are treated and a reduction in the amount of administered drug. It can also be combined with mechanical clot disruption.

There have been two RCTs of IA-TLS which have confirmed most of these advantages. They were PROACT I [16] and PROACT II [17] and performed comparing IA-TLS with pro-urokinase to IA placebo infusion in PROACT 1. In both studies, the TTT interval between symptom onset and starting AI-TLS was limited to 6 h, and patients in both arms were given heparin. IA-TLS was performed via a microcatheter placed in the proximal middle cerebral artery as close to thrombus as possible and pro-urokinase infused over 2 h to a maximum dose of 9 mg. Mechanical clot disruption was not allowed. Despite an increased rate of early symptomatic haemorrhagic transformation in the treated group (10.2 vs. 2%), the ultimate mortality rates were similar and functional outcomes (mRS ≤ 2) were better (40 vs. 25%) which represented a 15% absolute difference. This benefit over intravenous thrombolysis has been confirmed in observational studies [18].

The disadvantages of IA-TLS are the substantial numbers of potential patients that need to be assessed for possible treatment, and the optimum time to treatment remains <3–6 h of symptom onset. This presents a substantial logistic challenge since arterial imaging and selection (together with transfer to a specialist centre) have to be completed within this period. Other considerations are the risks associated with cerebral catheterisation and, if heparin is given during procedures, the potential risk of increased morbidity in the event of a haemorrhagic complication.

For the above reasons, combination treatment or ‘bridging’ therapy in which IV-tPA is given during investigations and triage for possible IA-TLS, is widely practiced [19]. It has been shown that a substantial proportion of patients still have arterial occlusions after the IV-tPA (about 70%) and that recanalisation rates are better when the two treatments were combined [20]. However, despite improving recanalisation rates, trails of combined therapy generally failed to show a significant clinical benefit at 3 months [21, 22]. The main drawback to combined therapy was high haemorrhage rates; in both the RECANALISE [21] and IMS II [23] studies, symptomatic intracranial haemorrhage occurred in about 10% of patients, whereas IV trials that exclude patients with substantial arterial occlusion reported symptomatichaemorrhage rates of 2–8%, e.g. ECASS III [23] and the Safe Implementation of Thrombolysis in Stroke study [24]. The role of combination therapy became blurred with the commercial introduction of clot extraction devices and their permitted use in trials designed to test the efficacy of IA-TLS. Though consistent in identifying early recanalisation as a benefit most were ambivalent over any long-term outcome advantage. The process culminated with the stopping of the largest of such trials, IMS III (Interventional Management of Stroke) study in 2013 on the grounds of futility after 656 participants randomised to endovascular treatment with IV-tPA (0.9 mg alteplase per kilogramme) or IV-tPA alone achieved mRS ≤ 2 function at 90 days in 41% and 39% of patients, respectively [25]. Poor outcomes in the IA-TLS treated group was related to the length of time to recanalisation (mean 325 min) which was achieved in 76% [26].

The role of thrombectomy was dramatically increased in 2015 following the publication of positive reports of five RCTs comparing IV-tPA with endovascular clot extraction (i.e. mechanical thrombectomy) against IV-tPA alone. These resolved any lingering controversy over its value and repudiated the conclusions drawn from IMS III. The first RCT to report was the MR CLEAN trial [27] conducted in Holland, which randomised 500 acute stroke patients to intra-arterial thrombolysis and/or stent-retriever thrombectomy with IV-tPA versus IV-tPA and standard medical care. This protocol was repeated in the other trials (see Table 17.2) with minor modifications in patient selection criteria and TTT thresholds for instigating IV-tPA and thrombectomy. In MR CLEAN, intra-arterial therapy had to be started within 6 h of symptom onset. The number of participants recovering to be functionally independent (defined as mRS ≤ 2) measured outcome. The result was that 32% of thrombectomy and 19.1% of standard care patients achieved function independence. Rates of adverse events were similar in the two groups. The online publication of data from MR CLEAN leads to data monitoring committees stopping other recruiting trials with similar protocols; these included EXTEND-IA, ESCAPE and REVASCAT.

There have since been several data review publications (i.e. meta-analysis), which confirm the efficacy of mechanical thrombectomy. Rodrigues et al. [32] calculated a risk ratio of 1.56 (95% CI 1.38–1.75) for a good functional outcome after adjuvant mechanical thrombectomy and 0.86 (95% CI 0.69–1.06) for mortality. Badhiwala et al. [33] used pooled data from eight RCTs (2423 patients) comparing IV-tPA with and without mechanical thrombectomy and documented a higher revascularisation rate at 24 h, 75.8% versus 34.1% for adjuvant thrombectomy; similar haemorrhage rates, 5.7% versus 5.1%; and higher good functional outcome rates at 90 days, 44.6% versus 31.8%. A common feature of these trials was the use of stent-retrieval devices for thrombectomy.

The initial neurological status of the acute stroke patient is important and related to their likelihood of recovery. In trials, it is generally compared between patients, using the US National Institute of Health Stroke Scale (NIHSS) [34]. Scores greater than 15 indicate a very poor prognosis with a high probability of death or severe disability and below 6 that recovery is likely [35]. However, patients with low scores may have established cerebral infarction and need to be distinguished from those experiencing transient ischaemic attacks (TIA). This is a dilemma for researchers designing clinical trials since a proportion of patients presenting with the clinical signs of acute stroke will recover spontaneously (presumably because of natural thrombolysis), and therefore for them, intervention is unnecessary. In practice, this seems less of a problem, possibly because of imaging.

Imaging is critical in the selection of patients and is usually performed with CT because it is generally available in hospital receiving areas and is easier to perform than MRI in the acute setting when patients may be neurologically impaired and details of past medical history are incomplete.

CT: This will normally show no parenchymal abnormality in patients with ischaemic stroke. Extensive areas of hypodensity are relative contraindications to thrombolysis. Low-density change exceeding one-third of the middle cerebral artery territory is a contraindication to intravenous thrombolysis. The ECASS trialists set this threshold, but a more sophisticated assessment has been proposed as the Alberta Stroke Program Early CT Score (ASPECTS) [36]. This schema divides the middle cerebral artery territory (as displayed on axial CT scans at specified levels) into 10 segments, and the reader subtracts a point for each area that is hypointense relative to normal brain. Scores less than 7 are associated with poor outcomes [37]. CT will show primary haemorrhage and any secondary haemorrhagic changes in areas of infarction (thereby contraindicating thrombolysis). Hyperintense vessels may be seen and suggest the presence of intravascular thrombosis.

CTA: This is an effective method of imaging arteries and can be combined with CT perfusion imaging to identify areas of reduced cerebral perfusion and intra-arterial thrombosis. It offers the potential of quickly identifying patients who may benefit from IA-TLS or clot removal.

MRI: MR scanning poses some logistical problems for examining the acute stroke patients and is generally less available than CT. Concerns about its inability to identify intracerebral haemorrhage have been exaggerated, and additional susceptibility-weighted scans are sufficiently accurate to exclude primary haemorrhage and anything but the most insignificant secondary bleeding. However, increasing the number of sequences leads to longer scan times and delays in starting therapy. Its main advantage is the capability of diffusion-weighted imaging (DWI). This technique is highly sensitive to early cerebral ischaemia and will show changes within minutes of cerebral infarction [38]. MR perfusion will show areas of absent or reduced perfusion and is used in combination with DWI to assess the viability of brain affected by arterial occlusions. Areas of abnormality demonstrated by DWI which are shown to be still perfused on MR-perfusion scans are described as penumbra regions which are potentially recoverable if blood flow can be re-established before infarction. This is a potentially useful objective parameter for selecting patients for thrombolysis, but cerebral ischaemia is a rapidly evolving pathology and imaging is by necessity only a ‘snap-shot’ of the process. The practical application of the technique has proved difficult in trials [39], and a recent meta-analysis was unable to show a benefit that warranted delaying thrombolysis for the additional data it provided [40].

DSA: It is used only to guide interventions and assess outcomes, i.e. extent of recanalisation. The latter is quantified using the Thrombosis in myocardial infarction (TIMI) trial perfusion grades adopted from cardiology (see Table 17.3).

Imaging guidelines were recently published by the American Heart Association, and all the imaging options for imaging the acute stroke patient were critically reviewed [42]. There remains much that is uncertain in the field of functional imaging but for triage of patients for thrombolysis or thrombectomy demonstrating large artery occlusion is all that really matters, and this can be done by CTA or MRA.

The benefit of rapidly re-establishing flow in a thrombosed artery has been recognised by many endovascular therapists since acute intravascular thrombosis is one of the many hazards that may complicate endovascular navigation. Clot retrieval is one solution, but balloon angioplasty to compress clot against the vessel wall can re-establish blood flow and assist endogenous or therapeutic thrombolysis. Successful angioplasty has been reported in case studies, alone [43, 44] and combined with stents [45].

Small clinical studies of the use of stents in acute stroke have been reported [46, 47]. The SARIS study [46] treated patients up to 8 h after onset of symptomatic stroke and achieved TIMI 2 or 3 grades of recanalisation (see Table 17.3) in all cases with a symptomatic haemorrhage rate of 5%. The value of additional stenting was apparent in the Brekenfeld et al. study [47] with recanalisation rates reaching 90% without complicating secondary haemorrhage or vessel perforation, attributed to avoidance of repetitive mechanical disruption [47] This initiative was backed by animal studies showing the potential of angioplasty with stenting [48].