Intracerebral hemorrhage due to thrombolytic therapy

Article section 7 of 16.  Previous  Next

By Kathleen Burger DO

The rate of intracerebral hemorrhage following thrombolytic therapy depends on the agent, dose, timing, route of administration, patient population, concomitant treatments, and the definition of hemorrhage used.

Thrombolytic therapy in myocardial infarction. Intracranial hemorrhage occurs in approximately 1% of patients who receive thrombolytic therapy for myocardial infarction (Hunt et al 1992; Maggioni et al 1992; Simoons et al 1993; Gore et al 1995). In the GUSTO-I trial of 41,021 patients with acute myocardial infarction, 268 patients had a symptomatic intracranial hemorrhage following administration of a thrombolytic agent (Gore et al 1995).

Pilot studies of intravenous thrombolytic therapy in acute stroke. The pilot studies of rtPA in acute ischemic stroke were accomplished in the late 1980s. Both the NINDS pilot study and the Burroughs Wellcome tissue plasminogen activator studies were dose-escalation safety studies (Brott et al 1992; del Zoppo et al 1992). The NINDS pilot study reported 3 symptomatic hemorrhages in 74 patients treated with doses of rtPA greater than 0.9 mg/kg within 90 minutes of symptom onset (Brott et al 1992). Three patients had asymptomatic hemorrhagic transformation. In the Burroughs-Wellcome study treatment had to be started within 8 hours of stroke onset (Wolpert et al 1993). Of the 98 patients treated with varying doses of rtPA, 11 had a symptomatic hemorrhage. The only significant baseline predictor of hemorrhage in this study was treatment greater than 6 hours from symptom onset.

Randomized clinical trials of intravenous thrombolytic therapy in acute stroke. A meta-analysis published by the Cochrane Database of Systematic Reviews in 2009 summarized the data of 26 randomized controlled trials completed by 2009 in 7152 acute stroke patients (Wardlaw 2009). Most of these trials tested intravenous thrombolysis up to 6 hours after symptom onset. Fifty-five percent of trials tested rtPA. Four trials evaluated intra-arterial thrombolysis. Overall, the odds of symptomatic intracranial hemorrhage were increased 3-fold (odds ratio 3.49, 95% confidence interval 2.81 to 4.33) in patients who received thrombolysis. In spite of this, there was a significantly reduced chance of poor outcome (combined death and dependency) during follow-up in patients who received thrombolytic therapy.

Intravenous thrombolytic therapy.

Recombinant tissue plasminogen activator. The NINDS rtPA Stroke Study formed the basis for FDA approval of intravenous rtPA for acute ischemic stroke within 3 hours of symptom onset (National Institute of Neurological Disorders and Stroke r-TPA Stroke Study Group 1995). In this study, the dose of rtPA was 0.9 mg/kg over 1 hour, with 10% of the dose given as a bolus. All patients were treated within 3 hours of onset of symptoms, and half of them within 90 minutes. Persistently elevated blood pressure before treatment (systolic >185 mm Hg and diastolic >110 mm Hg) was a study exclusion. Symptomatic intracerebral hemorrhage attributable to study drug was defined as a CT-documented hemorrhage that was temporally related to a change in the patient's clinical condition in the judgment of the clinical investigator within 36 hours from treatment onset. Of the 312 patients treated with rtPA, 6.4% of patients had a symptomatic intracerebral hemorrhage compared to 0.6% of placebo-treated patients. An additional 4.2% of patients treated with rtPA had an asymptomatic hemorrhage on a protocol-mandatory CT scan at 24 hours as compared to 2.8% of placebo patients. These rates of hemorrhage are very similar to the rates in the pilot NINDS studies in which patients were treated within 3 hours, but lower than in the other randomized thrombolytic studies in which the large majority of patients were treated after 3 hours (Table 1). A severe baseline neurologic deficit, as measured by the NIHSS, and the presence of edema or mass effect on the baseline CT scan were independently associated with an increased risk of symptomatic intracerebral hemorrhage in the NINDS trial (NINDS t-PA Stroke Study Group 1997). However, these patients were still likely to benefit from use of the medication.

Table 1. Incidence of Symptomatic Intracerebral Hemorrhage in Randomized Clinical Trials Using Thrombolytic Therapy in Acute Ischemic Stroke


No. of




Drug and dose




in treatment






in control







3 hours

rtPA 0.9 mg/kg






6 hours


1.5 million units








6 hours

rtPA 1.1 mg/kg




Australian Streptokinase Study**




4 hours


1.5 million units







6 hours


1.5 million units







0-3 hours

3-6 hours


rtPA 0.9 mg/kg







6 hours


6 mg +


500 IU/hr

for 4 hrs








6 hours


9 mg








3-5 hours


0.9 mg/kg








6 hours


0.9 mg/kg








3 hours


72 hour infusion








6 hours


72-hour infusion








3 to 4.5 hours


0.9 mg/kg
















3 to 9 hours





6 hours



90 ìg/kg


125 ìg/kg

0.1 mg/kg in 25 patients

0.25 mg/kg in 25 patients













* Symptomatic hemorrhage was defined as a CT-documented hemorrhage that was temporarily related to a change in the patient’s clinical condition in the judgment of the clinical investigator.
** Parenchymal hemorrhage type 2 is considered symptomatic/major hemorrhage and is defined as greater than 30% of the infarcted area affected by hemorrhage with mass effect or extension outside the infarct.
*** Parenchymal hematoma-homogenous area of circumscribed hyperdensity, usually with a mass effect and sometimes with ventricular extension.
**** Hemorrhagic transformation associated with neurologic deterioration of 4 or more points on the NIHSS or 1 point deterioration in level of consciousness.
***** Hemorrhagic transformation associated with neurologic deterioration of 4 or more points on the NIHSS at 72 hours or causing death by 90 days.

‡ Streptokinase and aspirin in the treatment group and aspirin alone in the control group.
† The percentages reflect the parenchymal hematoma of larger extent in the intention-to-treat population.

Table 2 lists the most important potential risk factors for intracerebral hemorrhage associated with the use of thrombolytic therapy in stroke based on the various studies that are summarized in this review. A systematic review of the literature reported several risk factors consistently associated with post-thrombolysis intracerebral hemorrhage. Of these, the most frequently found to predict symptomatic intracerebral hemorrhage across multiple studies are early CT hypodensity, elevated serum glucose or history of diabetes mellitus, and severe baseline neurologic deficit, as measured by the NIHSS (Lansberg et al 2007a).

Table 2. Risk Factors for Intracerebral Hemorrhage in Patients with Acute Ischemic Stroke


  Use of a thrombolytic agent

  Dose of thrombolytic agent



  Type of thrombolytic agent

  CT findings of early infarction prior to treatment

  Deeply ischemic brain tissue prior to treatment

  Hyperglycemia or history of diabetes mellitus

  High severity of stroke as determined by the NIHSS score

  Concomitant antithrombotic drug use



  Older age

  Longer time to treatment

  Deviation from established treatment protocols

  Elevated systolic blood pressure prior to treatment

  Prior antithrombotic drug use

  Prior stroke

  History of cardiac disease

  History of hypertension

  Non-smoking status

  Magnetic resonance imaging characteristics

  CT with leukoaraiosis

  Increased matrix metalloproteinase -9 levels

  Low platelet count

  Low LDL levels

  Low plasminogen activator inhibitor levels

In the European Acute Stroke Study (ECASS) of intravenous rtPA (1.1mg/kg) within 6 hours of symptom onset the risk of intracerebral hemorrhage was increased by 8.8% in the treatment group in the intention-to-treat analysis (Hacke et al 1995). Patients with changes on the baseline CT consistent with moderate or large regions of early cerebral infarction and those with a severe clinical deficit had an increased risk on hemorrhagic transformation, whereas advanced age increased the risk of parenchymal hematoma (Hacke et al 1995; Larrue et al 1997). There was no correlation between hemorrhage risk and time to treatment for either type of hemorrhage. The rate of hemorrhage in ECASS was much higher than in the NINDS rtPA study, which may well be related to the rtPA dose (1.1 mg/kg, 100 mg maximum) that was substantially higher than in the NINDS rtPA study. This is further supported by the decreased symptomatic hemorrhage rate in ECASS II, which used a lower rtPA dose than ECASS.

In the European Acute Stroke Study II (ECASS II) 800 patients were randomized to placebo or 0.9 mg/kg intravenous rtPA in the 6 hour time window (Larrue et al 2001). Only 158 patients were enrolled in the 0 to 3 hour time window. The rate of parenchymal hemorrhage was increased significantly in the treatment group as compared to the placebo group (8.1% vs. 0.8%). Risk factors for intracerebral hemorrhage in ECASS II included the extent of hypodensity on the baseline CT scan, congestive heart failure, advanced age, and aspirin use prior to thrombolysis (Larrue et al 2001).

The Alteplase Thrombolysis for Acute Noninterventional Therapy in Ischemic Stroke (ATLANTIS A) study evaluated the safety and efficacy of rtPA given between 0 and 6 hours after stroke onset. The first 142 patients enrolled in the ATLANTIS trial who were treated between 0 and 6 hours were included and evaluated separately as part A whereas part B evaluated those treated in the time window of 0 to 5 hours. The rate of symptomatic intracerebral hemorrhage was 11% in the treatment group and 0% in the placebo group, and 90 day mortality rates were 23% vs. 7%, respectively. The relatively high hemorrhage rate in this study has been attributed to the prolonged treatment window; however, the small number of patients in the study makes the observed hemorrhage rates fairly unstable (Clark et al 2000). The Alteplase Thrombolysis for Acute Noninterventional Therapy in Ischemic Stroke (ATLANTIS B) study enrolled 613 patients, 547 of these were treated in the 3- to 5-hour window. Treatment with rtPA increased the rate of symptomatic intracerebral hemorrhage in the first 10 days (7.0 vs. 1.1%) without a clear clinical benefit of treatment at 90 days (Clark et al 1999).

A pooled analysis including 2775 patients from the above mentioned randomized placebo-controlled trials of intravenous rtPA indicated that there may be a clinical benefit up to 4.5 hours from symptom onset (Hacke et al 2004).

The ECASS III trial evaluated the efficacy and safety of intravenous rtPA between 3 and 4.5 hours after stroke symptom onset (Hacke et al 2008). Treatment with rtPA significantly increased odds of a favorable outcome at 3 months (odds ratio 1.34, 95% confidence interval 1.02 to 1.76). Symptomatic intracerebral hemorrhage, defined as hemorrhage causally associated with a clinical deterioration of 4 or more points on the NIHSS or death, was 2.4% in the rtPA treatment arm and 0.2% in the placebo arm. Using the NINDS trial definition, the rate of symptomatic intracerebral hemorrhage in ECASS was 7.9% and 3.5% in the treatment and placebo arms, respectively. Of note, subcutaneous heparin for venous thromboembolism prophylaxis was allowed in the first 24 hours after rtPA, which was not allowed in prior trials. ECASS III also excluded patients older than 80 years, on oral anticoagulants, with NIHSS greater than 25, or with a history of diabetes and prior stroke. An observational study of patients treated with intravenous rtPA between 3 and 4.5 hours reported a 2.2% rate of symptomatic intracerebral hemorrhage at 7 days using the ECASS III definition (Wahlgren et al 2008a). Based on these trials, the American Heart Association/American Stroke Association updated its guidelines in 2009 recommending intravenous rtPA for eligible patients with acute ischemic stroke up to 4.5 hours after symptom onset (del Zoppo et al 2009).

rtPA in clinical practice. Many clinical studies of intravenous thrombolytic therapy in acute stroke have reported a rate of hemorrhage complications comparable to the NINDS rtPA trial (Albers et al 2000; Tanne et al 2002), but some studies have reported increased hemorrhage and mortality rates (Lopez-Yunez et al 2001). In a study of 50 stroke patients treated in 10 acute-care hospitals in Indianapolis, the authors found that protocol violations were common (16%) and were associated with an increased risk of symptomatic hemorrhage (Lopez-Yunez et al 2001). The SITS-MOST registry reported that patients treated with intravenous rtPA for acute strokes in clinical practice were similar to those enrolled in clinical trials. The rate of symptomatic intracerebral hemorrhage, using the NINDS rtPA trial definition, was 7.3% at 7 days (Wahlgren et al 2007). The STARS study, a large phase 4 study mandated by the FDA, similarly assessed the safety and efficacy of intravenous t-PA therapy in the 3-hour window in clinical practice (Albers et al 2000). The study included 389 patients with a median NIHSS of 13. Protocol violations occurred in 32% of patients but were not associated with an increased symptomatic hemorrhage rate. The overall symptomatic intracerebral hemorrhage rate was only 3.3%. Finally, the Canadian Alteplase for Stroke Effectiveness Study (CASES) reported on 1112 patients treated with rtPA at 60 centers throughout Canada. Patients treated between 3 and 4.5 hours had a trend towards increased rate of sICH compared to those treated from 0 to 3 hours (7.8 vs. 3.8%, p = 0.06), but no difference in functional outcome at 3 months was reported (Shobha et al 2010). Despite early concerns of a potentially higher risk of intracerebral hemorrhage associated with intravenous rtPA outside the clinical trial setting, these studies support a safety profile of rtPA in clinical practice that is comparable to the NINDS study.

Other intravenous thrombolytic therapies. Multiple other thrombolytic agents – streptokinase, desmoteplase, tenecteplase, and ancrod – have been studied in randomized trials (Candelise et al 1995; Hommel et al 1995; Donnan et al 1996; Sherman et al 2000; Hacke et al 2005; Haley et al 2005; Furlan et al 2006; Hennerici et al 2006; Hacke et al 2009). However, these medications have not been approved for use in stroke patients because of safety concerns or a lack of established efficacy. Tenecteplase deserves special mention following recent completion of its phase 2B trial. After thrombolysis treatment of ischemic stroke, patients were carefully selected based on CT perfusion imaging (Parsons et al 2012). Tenecteplase was associated with significantly better reperfusion, improved clinical outcomes, and reduced symptomatic hemorrhagic conversion when compared to alteplase.

Intra-arterial thrombolytic therapy. Several teams of investigators have studied whether intra-arterial administration of thrombolytic agents could be more effective and safer than intravenous administration (Zeumer et al 1989; Brott et al 1992; Bollaert et al 1995; Furlan et al 1999). A thrombolytic agent can be delivered via a small arterial catheter at the site of the clot. Theoretically, this intra-arterial delivery increases the chance for lysis and decreases the likelihood of hemorrhage because of a lower total dose and less widespread effects of the agent. Although the reported lysis rates using intra-arterial delivery are higher than those reported for patients treated with intravenous thrombolytic agents, the rates of hemorrhage are similar to or greater than those in studies using intravenous rtPA within 3 hours of onset. A retrospective multicenter analysis comparing stroke patients treated with intravenous tPA to those treated with either intra-arterial thrombolysis or bridging intravenous plus intra-arterial therapy found an increased odds of sICH in patients receiving endovascular treatment (OR 3.47; 95% CI 1.19–10.01, p<0.05) after adjusting for baseline age, stroke severity, and stroke location (Singer et al 2009). Three good-sized randomized controlled studies of intra-arterial therapy have been completed.

Pro-urokinase. The Prolyse in Acute Cerebral Thromboembolism Trial Study (PROACT I), compared 6 mg of intra-arterial pro-urokinase and high-dose heparin (100 units/kg bolus followed by 1000 units/hour for 4 hours) versus high-dose heparin alone (del Zoppo et al 1998). The medications were infused at the site of the clot. The rate of symptomatic intracerebral hemorrhage in the combined treatment group was 27% versus 20% for high-dose heparin alone in the first 16 patients. For this reason the heparin dose was decreased to 2000 units bolus followed by 500 units per hour. The subsequent rate of symptomatic hemorrhage with combined pro-urokinase and heparin was 7% versus 11% for low-dose heparin alone (del Zoppo et al 1998).

The PROACT II study randomized 180 patients with an acute middle cerebral artery occlusion of less than 6 hours duration (Furlan et al 1999). The treatment group received 9 mg intra-arterial pro-urokinase over 2 hours plus intravenous heparin 2000 units bolus and 500 units per hour infusion for 4 hours. The placebo group received heparin only. Symptomatic intracerebral hemorrhage was observed in 10% of the treatment group and 2% of the control group. The higher rates of hemorrhage compared to other stroke trials probably reflect the greater baseline stroke severity in this study. Nevertheless, the results of this trial support the use of intra-arterial thrombolysis in the 6-hour time window in patients with a middle cerebral artery occlusion based on a 15% absolute increase in favorable clinical outcome in the treatment group in spite of the increased bleeding rate. With that said, it is uncertain whether intra-arterial thrombolysis would be superior to intravenous thrombolysis in the 3- to 4.5-hour window.

Urokinase. The Middle Cerebral Artery Local Fibrinolytic Intervention Trial evaluated the safety and efficacy of intra-arterial urokinase in patients with a middle cerebral artery occlusion of less than 6 hours duration (Ogawa et al 2007). A total of 114 patients were randomized to either the treatment arm, in which subjects were treated with heparinization followed by local delivery of urokinase at the site of occlusion, or the control arm, in which no specific therapy was indicated. Wire clot maceration was allowed in the treatment arm. There was a nonsignificant increase in favorable outcomes in the urokinase treatment group. The 90-day cumulative mortality was 5.3% in the urokinase group and 3.5% in the control group, and symptomatic intracerebral hemorrhage within 24 hours of treatment occurred in 9% and 2%, respectively.

rtPA. Neither pro-urokinase or urokinase are readily clinically available, resulting in intra-arterial administration of rtPA to treat acute ischemic stroke patients with large artery occlusion at many centers. No published randomized controlled trials compare the intra-arterial administration of rtPA to placebo in acute ischemic stroke. However, the Interventional Management of Stroke Study evaluated intra-arterial rtPA given after a reduced “bridging” dose (0.6 mg/kg) of intravenous rtPA (IMS Study Investigators 2004). Eighty subjects were enrolled and comparisons were made to matched controls and rtPA-treated patients from the NINDS stroke study. The 6.6% rate of symptomatic intracerebral hemorrhage in treated patients was similar to those treated with intravenous rtPA in the NINDS trial. A second study evaluated this combined intravenous and intra-arterial approach to treatment of moderate to severe strokes within 3 hours of symptom onset (IMS II Trial Investigators 2007). Eighty-one patients were treated and compared to historical controls from the NINDS trial. This trial reported a symptomatic intracerebral hemorrhage rate of 9.9%, similar to other studies. Results of the IMS-3 trial are not yet available. The trial was recently stopped early, though not due to safety concerns (Anderson 2012).

Mechanical thrombolytic therapy. Mechanical thrombolysis and thrombectomy has been shown in several uncontrolled trials to promote arterial recanalization. These endovascular techniques may be used alone or in combination with intravenous or intra-arterial administration of thrombolytic drugs. Currently there are 2 catheter systems, the Merci Retriever® (Concentric Medical, Inc., Mountain View, CA) and the Penumbra System® (Penumbra, Inc., Alameda, CA), cleared by the U.S. FDA for revascularization of patients with acute ischemic stroke secondary to large vessel occlusive disease within 8 hours of symptom onset.

Merci retriever. The MERCI trial evaluated the safety and efficacy of the Merci clot retriever system in the treatment of ischemic stroke patients with a large vessel occlusion within 8 hours of symptom onset not eligible for intravenous rtPA (Smith et al 2005). Embolectomy using the Merci catheter was associated with a 7.8% rate of symptomatic intracerebral hemorrhage in the MERCI trial, defined as intracranial bleeding within 24 hours of treatment leading to a clinical worsening of greater than 4 points on the NIHSS or death or any subarachnoid hemorrhage. A second trial (Multi MERCI) reported a 9.8% rate of symptomatic intracerebral hemorrhage after embolectomy (Smith et al 2008). The Multi MERCI trial included patients that had persistent occlusions after standard dose intravenous rtPA and allowed adjuvant intra-arterial rtPA. Neither of these trials had a control arm. However, mechanical thrombectomy using the Merci Retriever was associated with significantly higher rates of recanalization (MERCI 46%, Multi MERCI 69.5%) than reported in the placebo arm of PROACT II (18%).

Penumbra system. The Penumbra catheter was evaluated in a single-arm study of patients with acute ischemic stroke within 8 hours of symptom onset with a treatable intracranial large vessel occlusion (Penumbra Pivotal Stroke Trial Investigators 2009). Patients that were refractory to intravenous rtPA were included. Symptomatic intracerebral hemorrhage was reported in 11.2% of patients. Successful recanalization was reported in 81.6% of treated vessels.

Other mechanical devices. Both the Solitaire™ FR Revascularization Device and Trevo Device retrievable stents show promise given the reportedly low 6.7% incidence of symptomatic hemorrhage when used to treat acute ischemic stroke. However, only a limited number of case series are available to demonstrate this (Novakovic et al 2012).

Potential risk factors for post-thrombolysis symptomatic intracerebral hemorrhage. Several studies have evaluated clinical predictors of post-thrombolysis symptomatic intracerebral hemorrhage. However, a systematic review of the literature identified a few risk factors that have repeatedly been associated with symptomatic intracerebral hemorrhage after thrombolytic treatment of acute ischemic stroke (Lansberg et al 2007a). Lansberg and colleagues identified 12 studies that met their inclusion criteria, of which CT characteristics, elevated serum glucose or history of diabetes, and elevated NIHSS repeatedly were found to be independently associated with post-thrombolysis symptomatic intracerebral hemorrhage (NINDS t-PA Stroke Study Group 1997; Demchuk et al 1999; Jaillard et al 1999; Barber et al 2000; Dubey et al 2001; Kase et al 2001; Larrue et al 2001; Tanne et al 2002; Hill and Buchan 2005; Cocho et al 2006). Advanced age, increased time to treatment, high systolic blood pressure, low platelet count, history of congestive heart failure, low plasminogen activator inhibitor levels, prior antiplatelet use, non-smoking status, low density lipoprotein levels, MRI characteristics, and deviations from treatment protocols have also been associated with post-thrombolysis in single studies (Larrue et al 2001; Gilligan et al 2002; Selim et al 2002; Tanne et al 2002; Ribo et al 2004; Hill and Buchan 2005; Bang et al 2007; Marti-Fabregas et al 2007; Uyttenboogaart et al 2008; Wahlgren et al 2008b; Mokin et al 2012). However, the results of these studies remain unclear. Differences in study design, population, symptomatic intracerebral hemorrhage definition, statistical model, thrombolytic therapy, and route of administration have lead to variability of findings, and some reports are conflicting. Advanced age was reported to independently predict post-thrombolysis hemorrhage in a post-hoc analysis of ECASS II (Larrue et al 2001); however, in the multicenter rtPA trial the association of age and intracerebral hemorrhage disappeared when CT and lab findings were included in the multivariate model (Tanne et al 2002). It is also difficult to truly determine whether these different factors are truly independent. The extent of early CT changes, advanced age, elevated NIHSS, and hyperglycemia all may just be markers of severe stroke rather than independent predictors of symptomatic intracerebral hemorrhage.

CT characteristics. Several studies have demonstrated an increased risk of intracerebral hemorrhage following intravenous rtPA treatment in patients with evidence of early infarct signs on CT (NINDS t-PA Stroke Study Group 1997; Demchuk et al 1999; Jaillard et al 1999; Barber et al 2000; Dubey et al 2001; Larrue et al 2001; Tanne et al 2002; Cucchiara et al 2009). These early ischemic changes include hypoattenuation of the brain parenchyma, loss of cortical grey-white junction differentiation, and swelling with sulcal effacement. The risk of symptomatic intracerebral hemorrhage was increased in rtPA-treated patients with prominent signs of edema or mass effect on the baseline CT in the NINDS rtPA Stroke Trial (Patel et al 2001). Despite this increased risk, patients in this group had a more favorable clinical outcome if they received rtPA than if they received placebo.

The Alberta stroke program developed a quantitative CT score, the Alberta Stroke Programme Early CT Score (ASPECTS), to help assess functional outcome and intracerebral hemorrhage following thrombolytic therapy for acute ischemic stroke (Barber et al 2000). This score quantifies early ischemic CT changes in patients with anterior circulation strokes by subtracting points for each area of hypodensity. Two studies have correlated lower ASPECTS with higher rates of post-thrombolysis hemorrhage. Patients enrolled in the ECASS II trial with lower ASPECTS had significantly higher rates of symptomatic intracerebral hemorrhage, approaching 40% in the those with the most significant CT changes (Dzialowski et al 2006). Other studies have shown that early CT changes involving more than 33% of the middle cerebral artery territory are particularly predictive of post-thrombolysis symptomatic intracerebral hemorrhage. Tanne and colleagues reported that patients with hypodensity involving more than 33% of the middle cerebral artery territory have 6 times higher rate of symptomatic intracerebral hemorrhage (Tanne et al 2002).

Leukoaraiosis on CT of the brain is associated with an increased risk of symptomatic hemorrhagic transformation and worse clinical outcomes in tPA-treated stroke patients when compared to patients without leukoaraiosis (Aries et al 2010); however, rtPA should not be withheld in such cases. It should be kept in mind that early widespread CT findings of infarction are uncommon and warrant confirmation of the symptom onset time. In contrast, subtle early infarct signs on CT are common in acute ischemic stroke patients and should not prevent an otherwise eligible patient from being treated with thrombolytic therapy.

Hyperglycemia and history of diabetes. Elevated serum glucose and history of diabetes mellitus were also identified to independently predict hemorrhage in several thrombolytic trials (NINDS t-PA Stroke Study Group 1997; Demchuk et al 1999; Jaillard et al 1999; Kase et al 2001; Selim et al 2002; Tanne et al 2002; Arnold et al 2012). The exact pathophysiologic connection between glucose and post-thrombolysis hemorrhage is unclear. Long-standing hyperglycemia in diabetics leads to damaged blood vessels, which in turn may be more prone to rupture in the setting of ischemic stroke. Acute hyperglycemia may promote or exacerbate brain injury or simply be a marker of stress in severe stroke.

Stroke severity. Several studies have independently shown that stroke severity, as rated by the NIHSS, is associated with an increased risk of post-thrombolysis hemorrhage (NINDS t-PA Stroke Study Group 1997; Dubey et al 2001; Cocho et al 2006; Cucchiara et al 2009). It is likely that in other thrombolytic studies, where stroke severity failed to correlate with hemorrhage risk, closely related covariates, such as infarct volume or CT changes, masked any potential relationship. Despite a higher risk of hemorrhage, patients with severe strokes are still more likely than not to benefit from thrombolytic therapies (NINDS t-PA Stroke Study Group 1997).

Prior antiplatelet use. Prior antiplatelet use may increase the rate of thrombolytic-associated intracerebral hemorrhage. Reports of acetylsalicylic acid (ASA) use preceding intravenous tPA-treated strokes have provided conflicting results (Anonymous 1997; Larrue et al 2001; Tanne et al 2002; Uyttenboogaart et al 2008). Patients receiving dual antiplatelet therapy with acetylsalicylic acid and clopidogrel may have particularly greater risk of thrombolysis-related sICH (Cucchiara et al 2009). In the SITS-MOST database, the rates of sICH were reported to be: 1.1% for antiplatelet naïve patients, 2.5% for any antiplatelet use, 2.5% for acetylsalicylic acid monotherapy, 1.7% for clopidogrel monotherapy, 2.3% for acetylsalicylic acid plus dipyridamole, and 4.1% for acetylsalicylic acid and clopidogrel (Diedler et al 2010). Although the rate of sICH is higher in the acetylsalicylic acid plus clopidogrel cohort, this is still below the sICH rate reported in prior clinical trials and should not exclude otherwise appropriate patients from acute therapy.

Other radiographic risk factors. Similar to CT studies, investigators have been interested in specific MRI parameters that might identify patients who are at an increased risk for hemorrhagic transformation of an ischemic infarct in the setting of thrombolytic therapy. One such parameter is a low ADC value in the infarcted tissue on diffusion-weighted imaging (DWI). Studies have demonstrated that the volume of ischemic tissue with an ADC value of less than 500x10(-6)mm2/sec is an independent predictor of intracerebral hemorrhage (Tong et al 2000; Selim et al 2002). Larger infarct volumes, as measured by DWI, and reperfusion status have been shown to independently predict symptomatic intracerebral hemorrhage in patients enrolled in a study of rtPA between 3 and 6 hours identified to have a perfusion-diffusion mismatch on pretreatment MRI (Lansberg et al 2007b). DWI lesion volumes also predicted symptomatic intracerebral hemorrhage in a retrospective analysis of patients treated with both intravenous and intra-arterial thrombolytics (Singer et al 2009).

A recent stroke within 3 months of an acute infarction is a contraindication to treatment with intravenous thrombolysis because of a perceived risk of increased hemorrhage risk into necrotic tissue. Up to 50% of patients who’ve experienced a transient ischemic attack have diffusion-weighted positive areas of subclinical infarction on MRI (Redgrave et al 2007). There may an increased risk of hemorrhagic transformation with thrombolysis in stroke patients who report a recent transient ischemic attack; however, one study reported the rate of sICH in this group to be 8.3% (McKinney et al 2012), and another reported no influence of recent transient ischemic attack on post-thrombolysis outcomes (de Lecinana et al 2010).

Other biochemical risk factors. In recent years, studies have focused on the role of biochemical variables that might play a role in predicting thrombolysis-associated hemorrhagic transformation.

Plasminogen activator inhibitor (PAI-1) inhibits rtPA. Low levels of PAI-1 have been shown to potentiate the action of rtPA and are a theoretical risk factor for hemorrhagic complications. Ribo and colleagues demonstrated an association between circulating levels of PAI-1 and subsequent hemorrhage risk in patients treated with rtPA (Ribo et al 2004). A subsequent study did not replicate this finding (Cocho et al 2006).

Matrix metalloproteinases (MMPs) belong to a family of endopeptidases that are released from vascular endothelium and white blood cells during ischemia and cleave extracellular matrix proteins leading to microvascular damage and blood-brain barrier disruption. MMP inhibitors appear to decrease hemorrhagic transformation in animal models of thrombolysis in acute ischemic stroke. In a study of 250 patients with hemispheric ischemic stroke, MMP-9 of 140 ng/mL or higher was associated with hemorrhagic transformation (odds ratio 12, p < 0.0001) after adjustment for potential confounders and final infarct volume (Castellanos et al 2003). Another study by the same authors showed that the serum cellular fibronectin concentration also was an independent predictor of hemorrhagic transformation in ischemic stroke patients (odds ratio 2.1, p = 0.0002) (Castellanos et al 2004). Cellular fibronectin levels in this study correlated positively with MMP-9 levels and the hypodensity volume on the CT scan.

Risk factors predicting symptomatic intracerebral hemorrhage after intra-arterial thrombolysis. Few data are available on predictors of hemorrhagic transformation in patients receiving intra-arterial thrombolysis for acute ischemic stroke. Based on univariate analyses of 24 variables from the PROACT II study, elevated glucose (baseline glucose >200 mg/dL) was the only independent risk factor for symptomatic hemorrhage (Kase et al 2001). In a retrospective study of 89 patients who were treated with intravenous and intra-arterial or pure intra-arterial thrombolysis, major symptomatic hemorrhage was 7% (Kidwell et al 2002). The rate of hemorrhagic transformation was not different in patients receiving combined intravenous-intra-arterial therapy as compared to intra-arterial therapy alone. Independent predictors for any hemorrhagic transformation in this series were higher NIHSS score, longer time to recanalization, lower platelet count, and higher glucose levels. Suarez and colleagues also reported a significant association between higher NIHSS score and elevated serum glucose levels and intracerebral hemorrhage in 54 patients treated with intra-arterial urokinase in the 6-hour time window (Suarez et al 1999). A large multicenter study retrospectively analyzed data from 645 stroke patients treated with intravenous, intra-arterial, and intravenous plus intra-arterial thrombolysis (Singer et al 2009). This study reported DWI lesion volume on pretreatment MRI, and intra-arterial or intravenous plus intra-arterial thrombolysis independently predicted symptomatic intracerebral hemorrhage. Patients treated with intra-arterial or intravenous plus intra-arterial were 3.5 times more likely to develop symptomatic intracerebral hemorrhage than those treated with intravenous alone (Singer et al 2009).

Risk scores for predicting post-thrombolysis intracerebral hemorrhage. Two groups have proposed risk-scoring systems to better quantify the hemorrhagic risk associated with thrombolytic treatment of acute ischemic stroke. The Multicenter Stroke Survey Scale was developed using markers identified in the Multicenter rtPA Stroke Survey of post-thrombolysis risk factors (Tanne et al 2002; Cucchiara et al 2008). The risk factors used in this scale are age greater than 60 years, NIHSS greater than 10, admission serum glucose greater than 150 mg/dL, and platelet count less than 150,000. This scoring system was both developed and tested using the Multicenter rtPA Stroke Survey data set and has not been validated in an independent prospective cohort. The rate of symptomatic intracerebral hemorrhage using this tool was 0%, 5%, 4%, and 18% for 0, 1, 2, and 3 or more risk factors present, respectively (Cucchiara et al 2008).

The Hemorrhage After Thrombolysis (HAT) Score was developed using a combination of previously published markers of increased post-thrombolysis hemorrhage risk that provided the highest predictive ability in their cohorts (Lou et al 2008). Using this method, the authors identified a history of diabetes or admission glucose above 200 mg/dL, pretreatment NIHSS, and early CT hypodensity as the most important risk factors. The rate of symptomatic intracerebral hemorrhage using this scale was 2%, 5%, 10%, 15%, and 44% for HAT Scores of 0, 1, 2, 3, and more than 3, respectively. This scale was largely derived and refined from the NINDS trial data set and has only independently been validated in a single small cohort of patients at the authors’ institution.

Finally, the GRASPS score assigns points for hemorrhagic risks, including glucose at presentation (G), race (Asian) (R), age (A), sex (male) (S), systolic blood pressure (P), and stroke severity (S) based on the NIHSS score. GRASPS is the first prediction tool validated in a large national data set that is available to assist clinicians in determination of symptomatic intracerebral hemorrhage risk following treatment with IV tPA (Menon et al 2012).

If these clinical risk-scoring schemes are validated in large, prospective cohorts, they could provide clinicians with an additional tool to better estimate the risks associated with thrombolytic treatment in an individual patient. It must be emphasized that these tools do not provide information about the potential benefits of thrombolysis. Regardless of the risk of symptomatic intracerebral hemorrhage, a patient may still have better odds of a meaningful recovery if treated with thrombolytics.

In This Article

Historical note and nomenclature
Clinical manifestations
Clinical vignette
Pathogenesis and pathophysiology
Differential diagnosis
Diagnostic workup
Prognosis and complications
References cited