Frequency and Definition
Intracerebral hemorrhage after systemic thrombolysis typically occurs within the first 24–36 hours after initiation of treatment. In the landmark NINDS tPA Stroke Trial, most hemorrhages occurred within the first 6 hours from tPA infusion. Several definitions of thrombolysis-induced symptomatic ICH (sICH) currently exist, depending on the clinical or radiological characteristics considered. According to the NINDS definition, a hemorrhage is considered symptomatic if it is associated with any decline in neurological status. An alternative definition of sICH frequently used in many clinical trials is deterioration of 4 or more points in the NIHSS score. When comparing the frequency of sICH among trials evaluating different revascularization strategies, it should be taken into account that the rates can vary significantly, depending on the definition chosen. For example, in a recent study of 985 ischemic strokes, the frequency of sICH ranged from 2.1% to 9.4% when different definitions of sICH were applied. Likewise, the ECASS III trial, which tested IV tPA administered between 3 and 4.5 hours after stroke onset, demonstrated a wide range of sICH rates (2.4%–7.9%), depending on specific criteria used to define sICH.
Table 1 summarizes the rates of sICH and sICH-related mortality in major international stroke trials conducted during the last 2 decades. The rate of sICH following administration of IV tPA within the first 3 hours of acute stroke was 6.4% in the original NINDS tPA trial. The SITS-MOST provided valuable information about the safety of administering IV tPA within 0–3 hours in routine clinical practice. In this large observational study, a total of 6483 patients were enrolled from 285 centers in the European Union (of which half had little previous experience with stroke thrombolysis). The rate of sICH (according to the NINDS definition) was 7.3%, indicating that systemic thrombolysis administered outside a clinical trial was associated with a similar complication rate.
Accurate comparison of sICH rates for different IA revascularization approaches is complicated because often more than one treatment modality is used to achieve successful revascularization. In many large prospective studies, the trial design allows patient enrollment after "failure" of systemic thrombolysis, whereas other patients can be taken directly for intervention because of ineligibility for systemic IV tPA administration. Trials of IA pharmacological thrombolysis (PROACT I and II and IMS I and II) demonstrated a rate of sICH in the range of 6.3%–15.4%. When mechanical thrombectomy with the Merci retriever device (Concentric Medical) was tested for acute stroke treatment in the MERCI and Multi-MERCI trials, sICH occurred in 7.8% and 9.8% of patients, respectively. Aspiration thrombectomy with the Penumbra system (Penumbra, Inc.) utilizes continuous aspiration together with mechanical fragmentation of the clot. In the Penumbra Pivotal Stroke trial, which was designed to assess safety of the Penumbra system for treatment of acute stroke due to large-vessel occlusion within the first 8 hours of symptom onset, sICH occurred in 11.2% of patients. Although the overall rates of sICH following pharmacological and mechanical endovascular approaches to acute stroke treatment are higher compared with IV thrombolysis with tPA alone, these results should be interpreted with caution. Endovascular therapy is typically reserved for strokes in patients with higher NIHSS scores (scores of 8–10 and above, depending on the trial design) and within a more extended time window (up to 6–8 hours of stroke onset). Both severity of baseline NIHSS score and extended treatment window are well-known risk factors for sICH (which we discuss in further detail later in this article), raising an argument that patients eligible for endovascular treatment are more prone to develop sICH due to the natural history of these strokes.
Intracerebral hemorrhage as a result of acute stroke treatment (either with IV or IA approaches) should be distinguished from ICH that can occur from a hyperperfusion syndrome following endarterectomy or carotid artery stenting in patients with carotid artery stenosis. This syndrome is thought to be caused by impaired cerebral autoregulatory mechanisms and usually occurs a few days after the surgery, although delayed presentations for up to several weeks have been described. The most catastrophic presentation associated with this syndrome—ICH—occurs in up to 1%–2% of patients who undergo carotid artery interventions, according to previously published retrospective studies of case series summarized in a review article by Moulakakis et al.
Radiographically, ICH can be classified on the basis of size of hemorrhage, as well as extent of ischemic infarct. This principle is applied in the ECASS classification scheme, in which hemorrhagic transformation after IV thrombolysis can be divided into 4 categories. Examples of these categories are shown in Fig. 1. Hemorrhagic infarction types 1 and 2 (HI-1 and HI-2) are defined as small petechiae along the margins of the infarct and larger more confluent petechiae without a space-occupying effect, respectively. Parenchymal hematoma Type 1 (PH-1) is defined as hemorrhage in less than 30% of the infarcted area, with some mild space-occupying effect. In parenchymal hematoma Type 2 (PH-2), hemorrhage is seen in more than 30% of the infarcted area, and there is significant space-occupying effect. Analysis of neurological outcomes in patients with each subtype of hemorrhagic transformation shows that PH-2 is associated with a poor chance for neurological recovery and a high rate of death at 3 months, whereas other subtypes do not have a significant effect on clinical outcomes following IV thrombolysis.
(Enlarge Image)
Figure 1.
Noncontrast CT scans showing examples of radiographic types of revascularization-induced ICH, according to the ECASS classification. A: Right MCA stroke with small petechial hemorrhage representing hemorrhagic infarction Type 1 (HI-1). B: Right internal capsule stroke with larger confluent petechial hemorrhages within the infarct core representing HI-2. C: Left parietooccipital stroke with hemorrhage in less than 30% of the infarcted area and some mild space-occupying effect representing parenchymal hematoma Type 1 (PH-1) hemorrhage. D: A large hematoma with significant space-occupying effect is characteristic of PH-2 hemorrhage.