Objective: To introduce the controversy in the classification of hemorrhagic transformation (HT) and to identify risk factors for poor outcome due to HT after reperfusion therapy for acute ischemic stroke.
Data Source: A MEDLINE search was performed (1995 – September, 2000) to identify relevant English-language publications, including clinical trials and review articles with the key words of hemorrhagic transformation, tissue plasminogen activator, streptokinase and stroke.
Study Selection and Data Extraction: Controlled clinical trials of reperfusion therapy in acute ischemic stroke patients were selected to identify the known risk factors for HT.
Data Synthesis and Results: There are two methods of classification of HT. One method is based upon purely radiological criteria, while the other relies on a combination of clinical symptoms and radiological findings. In three clinical trials that describe successful reperfusion strategies in human stroke patients, the following risk factors for the development of HT associated with poor outcome were identified: fibrinolysis, baseline CT changes consistent with ischemia, and severity of symptoms. The rates of symptomatic HT range from 5-10% in these trials. The relative risk (treatment event rate versus placebo event rate) varies based on the placebo rate, which is important to consider in assessing the safety of the therapeutic intervention.
Conclusion: HT can worsen the outcome of an ischemic stroke and is clearly associated with reperfusion strategies such as fibrinolytics and ancrod. Patients with large ischemic strokes, characterized by early CT findings of ischemia at baseline and poor clinical presentation, are at the highest risk of HT. However, these same patients are at high risk for a poor outcome regardless of the treatment they receive and may still benefit from reperfusion. Protocol violations with tPA use for acute ischemic stroke has been identified as an important contributor to the development of symptomatic HT. Each patient must be individually assessed for the relative risks and benefits of reperfusion therapy before being treated according to a carefully executed protocol.
Hemorrhagic transformation (HT) of an ischemic lesion is a common occurrence in stroke patients, occurring in up to 70% of patients in one published series. In fact, it has only been with the advent of sensitive imaging techniques, performed at prescribed times after stroke onset, that the true incidence of HT is emerging. For the most part, HT is thought to be without clinical consequence, or asymptomatic, and some investigators have gone as far as to say that some HT may be a good prognostic sign as it demonstrates that reperfusion has occurred and the tissue may return to normal function.
Interest in HT has been heightened by the introduction of reperfusion strategies designed to dislodge the offending clot and reestablish cerebral blood flow. All of the reperfusion strategies proven effective in improving functional outcome after ischemic stroke have done so through a manipulation of the endogenous fibrinolytic system. Because of this, all of them carry increased the risk of HT, which is sometimes fatal.
The purpose of this paper is to introduce the controversy in the classification of HT and to identify risk factors for poor outcome due to HT following reperfusion therapy.
Hemorrhagic Transformation after Acute Ischemic Stroke
HT describes a wide range of radiological phenomena. The European Cooperative Acute Stroke Study (ECASS) investigators, have classified HT into hemorrhagic infarction (HI) and parenchymal hematoma (PH). HI1 represents petechial (spotty) hemorrhages in the infarcted area while HI2 represents confluent petechiae. In PH1, a clot is evident and the hemorrhage is confined to <33% of the area of the infarct and in PH2 the hemorrhage is larger, is usually associated with mass effect, and can be in an area distal to the ischemic lesion. It is generally accepted that PH2, especially when extended to the ventricles, is associated with a high early mortality rate.
There are two main schools of thought regarding the classification of HT after ischemic stroke. The first is that only the radiological characteristics should be used to describe the main types of HT. This is the method used by the ECASS investigators and is based on the observation that HT often occurs in patients who deteriorate due to cerebral edema or other phenomena. Using pure radiological criteria, the ECASS investigators were able to demonstrate that only PH2 (not HI1, HI2 or even PH1) significantly increased the risk of early deterioration and death.
The second approach to HT classification embraces the concept that the radiological characteristics of this phenomenon cannot be interpreted without considering the clinical course of the patient. In this strategy, HT is considered to be either symptomatic or asymptomatic. This second strategy has been popularized by the National Institute of Neurological Disorders and Stroke (NINDS) tPA Stroke Study Group. It recognizes that small hemorrhages in crucial areas of the brain can be devastating and assumes that clinicians are the most qualified to assess whether an HT is “significant” or not.
MEDLINE searches (1995 – September 2000) were conducted to identify randomized clinical trials and subgroup analyses of hemorrhagic transformation after reperfusion therapy in ischemic stroke patients. The following key words were used: hemorrhagic transformation, tissue plasminogen activator, streptokinase and stroke.
Three hallmark randomized clinical trials using streptokinase for the treatment of acute ischemic stroke were identified. Although discontinued early due to excess deaths in the treatment groups, the streptokinase studies contributed important information regarding the patients at risk of symptomatic intracerebral hemorrhage after reperfusion therapy. All three studies included in Table 1 identified streptokinase as an important contributor to HT. Other contributors to HT are also included in the table. For example, the use of early ASA after thrombolytic treatment, as seen in the Multicentre Acute Stroke Trial – Italy (MAST-I), significantly increased the risk of symptomatic intracerebral hemorrhage (ICH). These studies support the strict prohibition of antithrombotic treatment for the first 24 hours after fibrinolysis. However, whether the agent (streptokinase), the time of administration (4-6 hours from onset of symptoms) or the dose of streptokinase (1.5 million units) is the most important factor contributing to the high mortality from ICH in these studies remains unclear.
Table 1. Streptokinase Hemorrhagic Transformation
|Risk Factors||Placebo Rates||Treatment Rates|
SK + ASA
Baseline CT changes
SK with decreased LOC
Time >3 hours
MAST-I=Multicenter Acute Stroke Trial – Italy; SK=Streptokinase; ASA=aspirin; MAST-E=Multicenter Acute Stroke Trial – Europe; CT=computed tomography; LOC=level of consciousness; ASK=Australian Streptokinase Trial
Successful Reperfusion Strategies
In the three clinical trials in which reperfusion of the ischemic brain was shown to be effective in reducing disability at 90 days, the incidence of symptomatic HT ranged from 5-10% (2.5- to 10-fold the placebo rate for this event) and the definitions varied with each study. In a multivariable analysis of the NINDS study, only severity of stroke and CT changes at baseline were shown to be important predictors of symptomatic HT. Increased age was associated with symptomatic HT, but much of the increased risk was accounted for by an increased severity of stroke in these patients.
Table 2. Successful Reperfusion Strategies – Hemorrhagic Transformation
|Risk Factors||Placebo Rates||Treatment Rates|
|NINDS (4)||3 hours||Severity
Edema or mass effect on CT
|PROACT II (5)||6 hours||Severity
(CT exclusion criteria)
|STAT (6)||3 hours||Fibrinogen level at 9-72 hours||Symptomatic=2%
ECASS=European Cooperative Acute Stroke Study; HI=hemorrhagic infarction; PH=parenchymal hematoma; CT=computed tomography; TPA=tissue plasminogen activator; NINDS=National Institute of Neurologic Disorders and Stroke tPA Stroke Trial; PROACT=Prolyse in Acute Cerebral Thromboembolism; STAT=Stroke Treatment with Ancrod Trial
Unsuccessful tPA Trials
Attempts to widen the therapeutic window beyond three hours for intravenous tPA in acute ischemic stroke have not been successful. ECASS I and II, along with Atlantis and A0276g all involved the use of tPA after 3 hours. ECASS I and II and A0276g each had 6 hour time windows , while Atlantis had a 3–5 hour window. The symptomatic HT rates in tPA-treated patients were 7% in Atlantis and 8.8% in ECASS II, just slightly higher than the 6.4% rate seen in NINDS. However, the A0276g trial reported an 11% symptomatic HT rate, which was substantially influenced by a rate of 18.2% in the patients treated five to six hours after symptom onset. A direct influence of time of treatment on the incidence of HT was confounded by a nine-fold increase in the number of severe strokes (NIHSS > 20) in the patients who received tPA between 5 and 6 hours after onset.
Table 3. Unsuccessful tPA Trials – Hemorrhagic Transformation
|Risk Factors||Placebo Rates||Treatment Rates|
|6 hours||Severity (HI)
Baseline CT changes (HI)
Increased age (PH)
|ATLANTIS (17)||3-5 hours||TPA||Symptomatic=1.1%
|A0276g (18)||6 hours||TPA
Time to treatment > 5 hours
ECASS=European Cooperative Acute Stroke Study; TPA=tissue plasminogen activator; HI=hemorrhagic infarction; PH=parenchymal hematoma
Postmarketing Experience with tPA
Since the first evidence of tPA safety and efficacy in treating patients with acute ischemic stroke, various groups have published their experiences with using the therapy in clinical practice. In 5 independent case series involving from 30 to 389 patients per series, the rates of symptomatic HT ranged from 3-7%. This was consistent with event rates reported in the NINDS trial and reflective of acceptable risk. However, in a report describing the experience of several hospitals from Cleveland, Ohio, investigators reported an alarming symptomatic HT rate of 15%, accompanied by a relatively high 15% in-hospital mortality rate. Of note, the investigators also reported a 50% incidence of protocol violations, the most common of which was an inappropriate use of antithrombotic agents during the first 24-hour period following tPA administration.
A consistent theme in all of the published reports of tPA for stroke in clinical practice is the importance of utilizing a strict protocol, based on the NINDS tPA Stroke trial. Deviation from the protocol can have disastrous results in terms of symptomatic HT and early mortality.
HT can be a devastating consequence of thrombolytic therapy in patients with acute ischemic stroke. Using radiological criteria alone to categorize HT as either parenchymal hematoma or hemorrhagic infarction has a high degree of objectivity and inter-rater reliability and appears to enable a prediction of early deterioration and mortality. Most experts agree; however, that a combination of clinical and radiological criteria must be used to adequately judge the significance of HT to an individual patient. For this reason, most clinical trials have enumerated HT as being either symptomatic or asymptomatic and clinical practice has followed this lead.
The patients at highest risk of symptomatic HT are those individuals who present with a severe stroke (i.e. a National Institute of Health Stroke Scale (NIHSS) greater than 20) and those with early evidence of edema and mass effect on the initial CT. These parameters are probably markers for the patients with the most severe damage to the microvasculature. Del Zoppo maintains that damage to the endothelium and basal lamina of the cerebral vasculature is necessary for the leakage of blood into the brain. Symptomatic HT rates of up to 30% were observed in the highest risk patients. This includes patients who receive tPA outside of a carefully executed protocol. An alarmingly high incidence of protocol violations (50%), the most common being antithrombotic treatment within 24 hours of receiving tPA, was reported by Cleveland investigators.
Although no randomized clinical trial evidence exists, another factor thought to be important in reducing the risk of HT after reperfusion in stroke patients is management of elevated blood pressure. Experimental stroke studies have demonstrated a direct correlation between blood pressure and risk of HT and treatment of blood pressure elevation has been shown to decrease HT size. Guidelines for the management of blood pressure elevation in patients who have received tPA have been published and should be followed.
Patients at high risk of symptomatic HT after tPA therapy for stroke are also at high risk of a poor outcome without HT. In fact, despite their increased risk of HT, the NINDS investigators reported an increased likelihood of a good outcome (albeit small) when patients with severe stroke received tPA. It is clear that each ischemic stroke patient must be evaluated individually for the relative risks and benefits of reperfusion therapy.
HT can worsen the outcome of an ischemic stroke and is clearly associated with reperfusion strategies such as fibrinolytics and ancrod. Patients with large ischemic strokes, characterized by early CT findings of ischemia at baseline and poor clinical presentation, are at the highest risk of HT. However, these same patients are at high risk for a poor outcome regardless of the treatment they receive and may still benefit from reperfusion. Protocol violations with tPA use for acute ischemic stroke has been identified as an important contributor to the development of symptomatic HT. Each patient must be individually assessed for the relative risks and benefits of reperfusion therapy before being treated according to a carefully executed protocol.