Freedom from Disability at 3 Months

Among these generally early-presenting acute ischaemic stroke (AIS) – LVO patients participating in trials predominantly testing highly effective reperfusion devices, random allocation to endovascular intervention was associated with an increase in disability-free outcome (mRS, 0–1) at 3 months after randomization (29.0% thrombolysis, 16.3% control; RR 1.72, 95% CI: 1.44–2.06; p < 0.00001). This represents 127 more disability-free patients per 1000 treated with endovascular reperfusion compared with control (see Figure 7.1). There was no heterogeneity across individual trials (heterogeneity p = 0.49), indicating consistency in evidence of treatment benefit.

Functional Independence at 3 Months

Endovascular therapy predominantly with highly effective mechanical devices was also associated with an increase in functional independence (mRS 0–2) at 3 months after randomization (47.2% endovascular therapy, 30.1% control; RR 1.58, 95% CI: 1.34–1.86; p < 0.00001). This represents 171 more functionally independent patients per 1000 treated with endovascular reperfusion compared with control (see Figure 7.2). Indications of benefit were consistent across studies, without evidence of variability (heterogeneity p = 0.51).

Death by 3 Months

By 3 months, patients allocated to endovascular therapy predominantly with highly effective devices had a non-significant trend toward lower mortality (14.5% endovascular therapy, 17.6% control; RR 0.83, 95% CI: 0.66–1.05; p = 0.13), representing a potential 31 fewer deaths per 1000 patients treated with endovascular therapy compared with control (see Figure 7.3). Signals of benefit were consistent across studies (heterogeneity p = 0.34).

Early Symptomatic Intracranial Haemorrhage

Endovascular therapy predominantly with highly effective devices was not associated with alteration in rates of early symptomatic intracranial haemorrhage within 7–10 days of treatment (4.2% endovascular therapy, 4.0% control; RR 1.10, 95% CI: 0.70–1.73; p < 0.68) (see Figure 7.4). Effect findings were consistent across trials (heterogeneity p = 0.68).

Other Complications

Endovascular therapy with highly effective thrombectomy devices is associated with additional, less-common complications, including, most notably, infarcts in new territories and femoral artery access site haematoma or pseudoaneurysm. These were not uniformly recorded in all the major randomized trials, but information on their frequency is available from subsets of the trials. An infarct in a new territory most commonly arises when control is lost over a thrombus or a fragment of a thrombus being retrieved from the cerebral circulation, and the released thrombus embolizes to a new territory. For example, a clot grasped and withdrawn from the M1 MCA may escape device control during passage through the ICA siphon and embolize to the anterior cerebral artery, causing ischaemia and infarction in the previously uninvolved anterior cerebral artery. Across 2 of the trials, infarcts in a new territory occurred in 5.4% (18/336) of endovascular versus 0.3% (1/370) of medical patients, RR 19.8, 95% CI: 2.7–147.7; p = 0.004 (Berkhemer et al., 2015; Jovin et al., 2015). A haematoma or pseudoaneurysm at the femoral arterial access site is a potential complication of all endovascular procedures. With endovascular thrombectomy, across 2 of the trials, femoral artery haematoma or pseudoaneurysm occurred in 5.6% (15/268) of endovascular versus 0% (0/253) of medical patients, RR 29.3, 95% CI: 1.8–486.7; p = 0.02 (Goyal et al., 2015; Jovin et al., 2015).

Degree of Disability at 3 Months

Analyses of dichotomized outcomes consider only a single health state transition that treatment may affect. For example, the outcome of being alive and disability-free counts only transitions across the border between mRS level 2 and mRS level 1, while the outcome of being alive and independent counts only transitions between mRS level 3 and mRS level 2. In contrast, analyses of ordinal outcomes consider simultaneously multiple valuable health state transitions that treatment may affect (Saver, 2011; Bath et al., 2012). For example, analysis of the distribution of outcomes over the entire mRS counts all transitions across all seven degrees of post-stroke disability the scale assesses, from asymptomatic, through varying degrees of impairment, to death.

An analysis has been conducted of the effect of highly effective reperfusion devices on the distribution of patient outcomes across all 7 levels of disability assessed by the mRS using participant-level data from the first 5 completed trials, enrolling 1287 patients (Goyal et al., 2016) (Figure 7.5). In these trials, allocation to endovascular thrombectomy resulted in increased proportions of patients at all disability ranks, including more patients who are asymptomatic (mRS 0), disability-free (mRS 0–1), functionally independent (mRS 0–2), ambulatory (mRS 0–3), not needing continuous care (mRS 0–4), and alive (mRS 0–5). Overall, treatment with highly effective endovascular thrombectomy versus control increased the odds of a better (less disabled) outcome state, adjusted common odds ratio (cOR) 2.49 (95% CI: 1.76–3.53). Using the automated algorithmic min–max joint outcome table method, for every 1000 patients treated with endovascular thrombectomy, a net of 423 patients will have a lower level of disability as a result of treatment (Tokunboh et al., 2018).

Figure 7.5 Level of disability at 3–6 months, endovascular thrombectomy with highly effective devices versus control, pooled trial analysis, based on data in Goyal et al. (2016).

Results for Highly Effective Endovascular Interventions against Non-Endovascular Controls, among Imaging-Selected Patients Late after Onset

In acute cerebral ischaemia, when an artery is abruptly occluded, collateral vessels provide compensating blood flow to the supplied region. The robustness of the collateral supply varies widely from individual to individual, depending on variations in circle of Willis anatomy, stenoses and occlusions in the alternative channels themselves, the site of the new occlusion, and systemic blood pressure. Shortly after the inciting vascular occlusion, a small brain region within the supplied field may experience complete loss of blood flow, causing cellular death within 1 to a few minutes. But a much larger, surrounding zone will experience moderate reductions in blood flow that the brain cells can tolerate for tens of minutes to several hours. As time from onset lengthens, the zone of the completed, irreversible infarction – the core – expands and the rim of threatened but still salvageable tissue – the penumbra – shrinks.

CT and MRI techniques indexing core and penumbra volume permit identification of the subset of patients with well-defined onset times who are ‘slow progressors’ and still harbour rescuable tissue in later time windows (Wheeler et al., 2015). They also enable assessment of tissue status in patients with uncertain onset times, due to stroke onset some time during sleep or awake onset in an unaccompanied patient, with aphasia or confusion precluding patient report or observation of onset time. With CT, perfusion CT imaging identifies as core regions with extreme reduction in blood flow or blood volume, and as penumbra regions with moderate blood flow reduction (perfusion–core mismatch). With MRI, diffusion MR sequences identify as core regions with substantial diffusion abnormality, indicating advanced bioenergetic compromise, and perfusion MRI sequences identify as penumbra regions with moderate-to-severe blood flow reduction not yet showing diffusion abnormality (perfusion–diffusion mismatch). Alternative, less precise but more readily obtained, approaches to identifying patients with penumbra estimate the volume of perfusion reduction using the presence of LVO (vessel–core mismatch), the extent of collateral vessels (collateral–core mismatch), or the severity of neurological deficits (clinical–core mismatch).

Using imaging evidence of persisting salvageable tissue to select a subset of late-presenting patients has been tested in 2 trials comparing highly effective endovascular reperfusion devices with non-endovascular treatment. Across both trials, random allocation to endovascular reperfusion therapy was associated with increased freedom from disability (mRS 0–1 at 3–6 months, 31.2% vs 11.1%; RR 2.76, 95% CI: 1.75–4.35; p < 0.0001) and increased functional independence (mRS 0–2 at 3–6 months, 46.7% vs 14.8%; RR 3.12, 95% CI: 2.15–4.53: p < 0.00001) (see Figures 7.1 and 7.2). In addition, patients allocated to endovascular therapy had no alteration in mortality (16.6% vs 21.7%; RR 0.76, 95% CI: 0.41–1.40; p = 0.38), or symptomatic haemorrhage rate (6.3% vs 3.7%; RR 1.63, 95% CI: 0.65–4.06; p = 0.29) (see Figures 7.3 and 7.4).

Effect of Onset to Treatment Time and Door-to-Treatment Time

Among the 1287 patients, 194 (15.1%) were treated within between 0.5 and 2 hours of onset, 657 (51.0%) between 2 and 4 hours, 352 (27.4%) between 4 and 6 hours, and 79 (6.1%) between 6 and 12 hours.

Alive and Functionally Independent at the End of Patient Follow-up

Earlier treatment was associated with greater benefit in increasing functional independence (mRS 0–2) at 3 months. With time from onset (last known well) to arterial puncture of 2 hours, allocation to endovascular reperfusion therapy was associated with more frequent alive and independent outcome (52.0% endovascular therapy, 27.0%, control; RR 1.93, intercept of full population ordinal model) (Figure 7.6). This represents 250 more alive and functionally independent patients per 1000 treated with endovascular reperfusion compared with control. In contrast, with time from onset (last known well) to arterial puncture of 5 hours, allocation to endovascular reperfusion therapy increased alive and independent outcome to a lesser degree (43.8% endovascular therapy, 27.0% control; RR 1.62). This represents 168 more alive and non-disabled patients per 1000 treated with endovascular therapy compared with control.

Figure 7.6 Modification of endovascular thrombectomy benefit by onset to expected arterial puncture. (A) Common odds ratio for reduced level of disability at 3 months; (B) 3-month outcome rates for each of the 7 levels of the mRS. Reproduced from Saver et al. (2016), with permission from the American Medical Association.

Functionally independent outcome declined more steeply with each delay during the interval from Emergency Department (ED) arrival to puncture (door to puncture) than from last known well to puncture (onset to puncture). Each 60-minute delay in onset to puncture was associated with a reduction in likelihood of functionally independent (mRS 0–2) outcome of OR 0.87 (0.78–0.96), while each 60-minute delay in door to puncture was associated with a reduction of OR 0.55 (0.43–0.71). Likely the faster decline in independent outcome with longer door to puncture than onset to puncture intervals was related to (1) a more accurately identifiable start time for the interval (ED arrival vs last known well), and (2) the screening out from some trials of patients with very rapid progression during the onset to door interval, due to entry criteria in requiring absence of substantial infarct extent upon arrival.

Effect of Age

Randomized trials have shown that older patients compared with younger patients have worse functional outcomes from acute ischaemic stroke but benefit to the same relative degree from endovascular therapy over a wide range of ages, from 50 to over 80 years old. A tendency to a lesser degree of benefit was noted, however, for young adults aged 18 to 49, among whom control group patients have good outcomes related to greater capacity for neural repair and recovery. Among 5 pooled trials of highly effective reperfusion devices, enrolling 1287 patients, 12% were aged 18 to 49, 72% were aged 50 to 79, and 15% were 80 years and older.

Endovascular thrombectomy among broadly selected patients early after onset tended to reduce final level of disability (over 6 levels of the mRS) to a lesser degree among patients aged 18 to 49 (common odds ratio 1.36) than among patients aged 50 to 80 and above (cORs ranging from 2.41–3.68, heterogeneity p = 0.07). Over the broad range from age 50 to 80 and above, the relative benefit was similar in degree in different age groups but with worse overall outcomes in older patients. For example, for ages 60 to 69, functional independence (mRS 0–2) was 51.9% with endovascular therapy, 27.9% with control, OR 1.78 (95% CI: 1.25–2.55); while for age 80 and above, functional independence was 29.8% with endovascular therapy, 13.9% with control (OR 2.09, 95% CI: 1.03–4.25). Considering mortality, endovascular thrombectomy was associated with reduced death among the oldest old (aged ≥80: OR 0.60, 95% CI: 0.36–0.99) but not younger individuals (aged 18–79: OR 0.95, 95% CI: 0.69–1.37).

Effect of Presenting Deficit Severity

Randomized trials have shown that severe, compared with moderate, presenting neurological deficits are a prognostic, but not a treatment benefit-modifying, patient feature. Across a broad range of presenting deficit severity, patients with more severe deficits benefit to the same relative degree from highly effective endovascular thrombectomy, without evidence of heterogeneity (heterogeneity p = 0.45), but absolute rates of good outcome are lower when initial deficits are greater. The odds of a better level of disability across the entire mRS were improved by endovascular thrombectomy to a similar relative degree across 4 levels of presenting deficit severity (National Institutes of Health Stroke Scale [NIHSS] ranges of ≤10, 11–15, 16–20, and ≥21) (heterogeneity p = 0.45) (Figure 7.7). For example, rates of functional independence among moderate deficit patients (NIHSS 11–15) were 58.1% versus 27.1% (OR 1.70, 95% CI: 1.19–2.43), and among severe deficit patients (NIHSS ≥21) were 23.0% versus 13.8% (OR 1.80, 95% CI: 1.09–2.96).

Figure 7.7 Common odds ratio for more favourable 3-month disability level in patient subgroups of differing presenting deficit severity and differing extent of infarct signs on first imaging. Figure based on data from Goyal et al. (2016).

Patients with mild initial presenting deficits have not been enrolled in substantial numbers in trials of highly effective endovascular thrombectomy. Among the first five trials, three formally excluded patients with NIHSS scores of less than 6 or less than 8, and the remainder enrolled very few patients in this range. However, observational series have noted that patients with initially mild deficits associated with LVOs have a substantial rate of subsequent stroke progression and poor outcome (Rajajee et al., 2006). Randomized trials in these patients are needed.

Effect of Imaging Findings on Noncontrast CT

Randomized trials have suggested that endovascular thrombectomy benefits both patients with mild and patients with moderate early infarct changes on initial brain imaging to the same relative degree, but may benefit patients with extensive early infarct changes less or not at all. The extent of early ischaemic injury changes on initial brain CT and MRI scans has been most often quantified using the Alberta Stroke Program Early CT Score (ASPECTS). Among 5 trials enrolling 1278 patients testing highly effective mechanical thrombectomy, 53% had mild early infarct changes (ASPECTS 9–10), 37% moderate (ASPECTS 6–8), and 9% severe early infarct changes (ASPECTS 0–5, but predominantly 4–5). The odds of a better level of disability across the entire mRS were improved by endovascular thrombectomy to a similar relative degree at each of these 3 levels of presenting deficit severity (heterogeneity p = 0.49) (see Figure 7.7), though patients with less extensive infarcts had better outcomes in both treatment arms: no or small infarct signs (ASPECTS 9–10), 49.7% versus 30.2%, OR 0.84, 95% CI: 0.73–0.96; moderate infarct signs (ASPECTS 6–8), 44.4% vs 23.3%, OR 0.81 (0.61–1.08); extensive infarct signs (ASPECTS 0–5), 23.3% versus 13.3% (see Figure 7.7).