This issue has previously been discussed with regards to heparin and thrombolysis for submissive pulmonary embolism.  Heparin was adopted as treatment for PE before the era of evidence-based medicine.  Although heparin prevents further clot formation, it probably doesn't have any immediate effect on hemodynamic stability.  Regardless, it has become dogma that all patients with PE must receive heparin.  Consequently, trials of thrombolysis have all compared heparin vs. heparin plusthrombolysis (rather than heparin versus thrombolysis alone).  Any excessive bleeding is blamed on the thrombolytic (rather than the synergistic combinationof heparin plus thrombolytic), tending to lead us back to heparin monotherapy in a perpetual loop of circular logic:

Currently we may be facing a similar phenomenon with regards to ischemic CVA.  Thrombolysis for CVA has been broadly adopted despite questionable evidence regarding its efficacy (e.g. cage-match debate here).  The dogma that all eligible stroke patients must receive thrombolysis may eventually impede progress with endovascular therapies.

Recently, several studies have found a benefit of endovascular therapy for proximal ischemic CVA (ESCAPE, EXTEND, and Mr. Clean).  These studies evaluated the efficacy of endovascular therapy when combined with systemic thrombolysis, thus assuming rather than testing the value of systemic thrombolysis in this scenario.

It seems counter-intuitive to intentionally treat an arterial occlusion with both systemic thrombolysis and mechanical intervention.  For example, thrombolysis would never be given to a STEMI patient who was going immediately for percutaneous coronary intervention.  

The benefit of thrombolysis in this situation is unclear.  The benefit of thrombolysis for ischemic stroke overall is unclear, with most studies showing no benefit or harm (see theNNT.com).  Large proximal occlusions in particular seem to be poorly responsive to thrombolysis.  Furthermore, with plans for emergent thrombectomy, at best a patient receiving thrombolysis would be revascularized at a slightly earlier time-point.  Given that patients appear to benefit from mechanical thrombectomy for up to 6-12 hours after stroke onset, the incremental benefit of revascularization 1-2 hours earlier is unclear.  Unfortunately studies correlating time-to-intervention with outcome in stroke tend to be biased, as discussed previously here.  

It may be beneficial to use some form of anticoagulation while awaiting neurointervention to prevent extension of the thrombus.  A titratable and reversible agent would be ideal for this purpose, such that it could be discontinued promptly in case of intracranial hemorrhage or perhaps after thrombectomy.  Heparin infusions have previously been investigated for this purpose, but this was not beneficial when combined with older interventional techniques (SYNTHESIS 2013).

Performing catheter-directed thrombectomy following loading with aspirin and systemic thrombolysis could worsen hemorrhagic complications compared to using a less aggressive anticoagulant regimen.  Although uncommon, wire perforation of a cranial artery does happens.  In the setting of systemic thrombolysis, any intracranial hemorrhage will be exacerbated.

Unfortunately, the better-powered studies (Mr. Clean and ESCAPE) weren't entirely forthcoming with regards to intracranial hemorrhage (ICH) rates.  Rather than reporting the actual ICH rates, both of these studies only reported the rate of symptomatic intracranial hemorrhage.  Rates of asymptomatic intracranial hemorrhage are typically higher, making asymptomatic hemorrhage rate a more sensitive measurement.  For example, Mr Rescue 2013 found a significant increase in asymptomatic hemorrhage with thrombectomy, despite no difference in symptomatic hemorrhage rates (table below).

In the setting of a patient with a large ischemic stroke, judging whether a superimposed ICH is “symptomatic” is highly subjective.  ESCAPE lacked a uniform definition of what constituted a symptomatic ICH, but instead “symptomatic ICH was clinically determined at the study site.”  The definition of symptomatic ICH used in Mr. Clean was an ICH accompanied by an increase of four or more points in the NIH stroke scale (by this definition, an ICH could result in paralysis of one arm and still be deemed asymptomatic).  Clinicians reporting these events were not blinded to treatment-group assignment.  This may explain the relative absence of small symptomatic ICH (table below): any small bleeds were unlikely to produce enough neurologic findings to be deemed symptomatic.
0

Lack of information about asymptomatic hemorrhage rates doesn't affect the overall conclusion of these trials.  However, this benchmark would be useful information to understand going forward, in order to compare complication rates between various studies and adjust the intensity of anticoagulation.

Overall, new studies on endovascular interventions are exciting and offer hope to patients with proximal infarcts who otherwise tend to fare poorly.   Rather than applying systemic thrombolysis to every infarct, these authors have advanced the concept of personalized stroke therapy.   Neuroimaging may be used to determine which patients may benefit from specific interventions.  

A remaining question is what the optimal anticoagulant strategy is for patients undergoing mechanical thrombectomy.   Systemic thrombolysis may not be the best choice for this situation, because it could increase the risk of peri-procedural hemorrhage and it is difficult to reverse in the event of an intracranial hemorrhage.  Heparin infusions have previously been investigated as a bridge to thrombectomy and may deserve re-evaluation in combination with current interventional technology.   Compared to percutaneous coronary intervention, very little is known about this topic and critical questions remain to be answered.