Mark R. Etherton, MD PhD

Charidimou A, Turc G, Oppenheim C, Yan S, Scheitz JF, Erdur H, et al. Microbleeds, Cerebral Hemorrhage, and Functional Outcome After Stroke Thrombolysis: Individual Patient Data Meta-Analysis. Stroke. 2017

In this entry, I discuss the recent meta-analysis, using individual patient data, by Andreas Charidimou and colleagues on cerebral microbleeds (CMB) and the risk of intracerebral hemorrhage (ICH) and poor functional outcomes after intravenous thrombolytic therapy for acute ischemic stroke.

Prior to this study, a recent meta-analysis had demonstrated that the mere presence of pre-treatment CMBs was associated with increased odds of symptomatic intracerebral hemorrhage (sICH) after intravenous thrombolytic therapy for acute ischemic stroke (Charidimou et al. Stroke. 2015). Building on this study, the authors performed a pooled, individual patient data meta-analysis to evaluate several hypotheses pertaining to the presence, quantity, and location of pre-treatment CMBs in relation to ICH risk and post-stroke outcomes.

The authors obtained data from multiple prospective or retrospective studies that evaluated ICH and 3- to 6-month functional outcomes after acute ischemic stroke in patients with antecedent MRI-defined CMBs treated only with intravenous tPA. A total of 2048 participants from 8 centers were used for the analysis. ICH was defined using the ECASS-2 criteria for hemorrhagic infarction, parenchymal hematoma (PH) and sICH (CH accompanied by a worsening of the NIHSS score by 4 or more points). Poor outcome at 3 to 6 months post-stroke was defined by a modified Rankin score >2.

CMBs were prevalent in 26.7% of the participants. Adjusting for treatment delay, age, and baseline NIHSS severity, participants with any CMBs had an increased odds of PH (OR 1.50; 95%CI 1.09-2.07) and remote PH (rPH, OR 3.04; 95%CI 1.73-5.35) compared to those without CMBs. In this model, the presence of CMBs was not associated with an increased risk of sICH (OR 1.42; 95%CI 0.86-2.35). When the CMBs were categorized according to the burden, however, increasing CMB number was associated with sICH, PH, and rPH. In participants with >10 CMBs on pre-treatment MRI, a >3-fold increase in risk of sICH was seen with IV tPA treatment (OR 3.65; 95%CI 1.17-11.42).

The authors then explored whether the location of the CMB had any influence on ICH risk or functional outcomes with the assumption being that the presence of strictly lobar CMBs was suggestive of Cerebral Amyloid Angiopathy (CAA). In both the CAA-related (strictly lobar CMBs) and non-CAA-related (mixed or strictly deep CMBs), the presence of CMBs was shown to increase the odds of both PH and rPH. Similar to the overall analysis, increasing burden of strictly lobar CMBs was associated with increased risk of rPH (OR 16.40, 95%CI 2.87-93.58). In patients with strictly deep or mixed CMBs, increased CMB burden was also associated with risk of rPH as well as PH.

The last major question of this analysis was whether the observations of increased risk of ICH after IV tPA treatment in patients with CMBs translated to worse functional outcomes at 3 to 6 months. Of note, patients with >10 CMBs had an increased risk of poor outcome, defined as a mRS >2 (OR 3.99; 95%CI 1.55-10.22). When CMBs were organized according to location, non-CAA-related CMBs also predicted poor outcome (OR 1.30; 95%CI 1.06-1.58).

The findings in this manuscript are of high interest to practicing Neurologists and Emergency Medicine clinicians. In patients with acute ischemic stroke, those with pre-treatment CMBs are at increased risk of ICH and poor outcomes after intravenous thrombolysis. Based on these findings, the outstanding question is, how are these results incorporated into clinical practice?

The current recommendations from the AHA/ASA are that intravenous tPA is reasonable in patients with CMBs (Class IIa, Level of Evidence B; Demaerschalk et al. Stroke. 2016). Of note, the Bleeding Risk Analysis in Stroke in Thrombolysis (BRASIL) study, which was excluded from this analysis because of lack of individual patient data, showed no difference in sICH rate between the CMB and non-CMB groups. Additionally, it is overall reassuring that in these analyses, the overall rates of sICH are low. At present, it would, therefore, seem that a prudent approach, given the marked benefit of intravenous thrombolysis, would be to incorporate the presence of CMBs into the overall clinical awareness and stratification of risk in this patient population.