Comment

Intravenous thrombolysis before thrombectomy for acute ischaemic stroke

Pooja Khatri

Lancet 2022; 400: 76-78

https://doi.org/10.1016/S0140-6736(22)01286-7

Before endovascular therapy, patients with ischaemic stroke due to occlusion of large arteries were treated with only intravenous thrombolysis, specifically alteplase. Intravenous thrombolysis recanalised about 25% of occluded large arteries, resulting in fewer than 30% of patients achieving functional independence (modified Rankin Score [mRS] 0–2) at 90 days.¹ In 2015, second-generation thrombectomy devices, in combination with intravenous thrombolysis, were proven to recanalise about 70–80% of arteries and improve rates of functional independence by 20–30% compared with intravenous thrombolysis alone.² Some in the field began to ask if intravenous thrombolysis is even necessary for these patients. Intravenous thrombolysis can delay the more often definitive therapy of endovascular therapy. Moreover, intravenous thrombolysis could increase symptomatic intracerebral haemorrhage, or distal migration of thrombi, rendering them inaccessible to thrombectomy. Intravenous thrombolysis certainly incurs substantial cost.³

Four previous trials have addressed this question of endovascular therapy versus intravenous thrombolysis and endovascular therapy. The SKIP trial in Japan and the MR CLEAN-NO IV trial in Europe did not show non-inferiority of endovascular therapy alone.4, 5 The DEVT and DIRECT-MT trials, both in China, showed non-inferiority but used wide margins inclusive of clinically meaningful effects.6, 7DEVT allowed for up to a 10% reduced rate of functional independence, and DIRECT-MT allowed for an adjusted common odds ratio of as low as 0·80 for a favourable functional level, for direct endovascular therapy to be declared non-inferior. Both trials also had methodological issues that increased the risk of bias, including long arrival to intravenous thrombolysis start times and significant protocol deviations for DIRECT-MT.⁸ Nevertheless, the results generated the hypothesis of differential treatment effects in Chinese patients, or perhaps Asian patients, compared with others.

In The Lancet two further trials by Urs Fischer and colleagues (SWIFT-DIRECT)⁹ and Peter J Mitchell and colleagues (DIRECT-SAFE)¹⁰ have each further tested the non-inferiority of bypassing intravenous thrombolysis. To address the question most ethically, as with the previous trials, these two trials exclusively enrolled the subset of endovascular therapy-eligible patients who arrived for medical care at thrombectomy-capable centres and who were, therefore, least likely to benefit from preceding intravenous thrombolysis. Despite providing this competitive advantage to the direct endovascular therapy group, delaying initiation of intravenous thrombolysis while determining endovascular therapy and trial eligibility, and using generous non-inferiority margins (12% in SWIFT-DIRECT and 10% for DIRECT-SAFE), both well conducted trials did not show the non-inferiority of omitting intravenous thrombolysis (SWIFT-DIRECT⁹ adjusted risk difference −7·3%, 95% CI −16·6 to 2·1, p=0·12; DIRECT-SAFE¹⁰ adjusted risk difference −5·1%, 95% CI −16 to 5·9, p=0·19). In fact, directions of effect favoured intravenous thrombolysis use; in SWIFT-DIRECT,⁹ 90-day functional independence was observed in 57% of the randomly assigned participants in the endovascular therapy group versus 65% in the intravenous thrombolysis and endovascular therapy group, and 55% versus 61% in DIRECT-SAFE.¹⁰ Furthermore, symptomatic intracerebral haemorrhage rates were similar between the treatment groups of each trial, and angiographic reperfusion rates were nominally higher in the intravenous thrombolysis-randomised group of the SWIFT-DIRECT trial. In addition to the aforementioned design aspects that favoured the direct endovascular therapy group, DIRECT-SAFE was underpowered due to early termination and thereby favoured intravenous thrombolysis use in this regard.

SWIFT-DIRECT participants were enrolled in Europe and Canada, had a median age of 72 years, and 51% were women.⁹ DIRECT-SAFE participants were enrolled in Australia, New Zealand, China, and Vietnam, had a median age of 69 years, and 43% were women; 46% were enrolled in China and Vietnam.¹⁰ Heterogeneity was not shown within subgroups in either trial, regarding inability to show non-inferiority of bypassing intravenous thrombolysis. In fact, superiority of using intravenous thrombolysis was suggested for those younger than 70 years (than those 70 years or older) in SWIFT-Direct (interaction p=0·0051) and those enrolled in China and Vietnam (vs Australia and New Zealand) in DIRECT-SAFE (interaction p=0·024). Although these subgroup results might be spurious, due to being underpowered, and specific ethnicities of patients enrolled in each country were not available, DIRECT-SAFE showed no evidence of the lower treatment effect in populations from the Asia region suggested by comparing results of previous completed trials.

In an expedited guideline of the European Stroke Organisation, a study-level meta-analysis of all six trials (including publicly presented data from the current two trials) showed that a prespecified 1·3% benefit on functional independence of intravenous thrombolysis could not be ruled out.⁸ Moreover, a 5% benefit of intravenous thrombolysis remained possible, similar to the magnitude observed in the practice-changing ECASS-3 trial of alteplase compared with placebo at 3 h to 4·5 h from stroke onset.¹¹ Furthermore, successful reperfusion was more frequent with combined therapy.⁸

Why would intravenous thrombolysis potentially add benefit to highly effective endovascular therapy? First, intravenous thrombolysis allows immediate administration of a potentially beneficial reperfusion therapy before endovascular therapy eligibility is established and while preparation for the procedure is under way. When intravenous thrombolysis eliminates the inciting thrombus before endovascular therapy is even initiated, time to cerebral reperfusion and consequent ischaemic injury is minimised. In SWIFT-DIRECT and DIRECT-SAFE, despite short times from intravenous thrombolysis start to endovascular therapy start by design, 3–4% fewer patients randomly assigned to intravenous thrombolysis still harboured large vessel occlusions on the initial diagnostic angiogram. Second, intravenous thrombolysis might be the only reperfusion therapy received if intracranial thrombi prove inaccessible due to vessel tortuosity or resistant to thrombectomy attempts. Third, residual thrombi that are too distal for thrombectomy, and even non-visualised microthrombi, might be recanalised by systemic thrombolysis.¹² Finally, intravenous thrombolysis might improve the rate of technically successful thrombectomy.⁸ Most benefits would be amplified when the time gap between intravenous thrombolysis and endovascular therapy is larger, as in the more common scenario when patients receive intravenous thrombolysis at outlying hospitals before transfer to an endovascular-capable hospital. Faster acting or more effective intravenous therapies would possibly favour combined therapy as well.¹³

The case appears closed. Bypassing intravenous thrombolysis is highly unlikely to be non-inferior to standard care by a clinically acceptable margin for most patients. More effective intravenous drug therapies are needed.