Thrombolysis-Related Hemorrhage: Can We Make Thrombolysis Safer?

Abstract

Intravenous tissue plasminogen activator (tPA) improves outcomeswhenadministeredwithin4.5hoursof symptomonset of ischemic stroke.1 Symptomatic intracranial hemorrhage (sICH) is the most feared complication after administration of intravenous tPA. The percentage of patients with a good functional outcome after sICH (as defined by the Safe Implementation of Thrombolysis in StrokeMonitoring Study2) following administration of tPA has been shown to be less than 7%, and mortality rates can be greater than50%.3Almost2decadesafter approvalof intravenous tPA by the US Food and Drug Administration, our ability to prevent thrombolysis-related sICH has not advanced much beyondadhering to inclusion andexclusion criteria and controlling blood pressure after thrombolysis. Moreover, there is no clearconsensusonhowtomanage thrombolysis-relatedsICH.4 Although theoccurrenceof sICH isuncommon, theknownrisk of this complicationmayweigh heavily on the decision to administer thrombolysis. The high morbidity and mortality associated with sICH represent an area where even a small improvement in prevention or treatment could have a dramatic effect. In this issueof JAMANeurology, Yaghi et al5 address an important gap in our knowledge and treatment of this complication.TheypooleddataonsICHafter thrombolysis from10 large strokecenters in theUnitedStates.TheydefinedsICHusing the definition from the Safe Implementation of Thrombolysis in Stroke-MonitoringStudy,2which includedthepresenceofaparenchymalhematomaaccountingformorethanone-thirdof the infarctvolume(parenchymalhematomatype2)detectedbycomputed tomography, alongwith at least a 4-point increase from thebaselineNationalInstitutesofHealthStrokeScalescore.They examined “real-world” treatments for sICH and also included codestatuschangetocomfortcaremeasures inthefirst24hours intheiranalysis.Theirprimaryoutcomewas in-hospitalmortality, andtheir secondaryoutcomewashematomaexpansiondetected by subsequent imaging. Multivariable analysis demonstrated that treatmentof thrombolysis-related sICHdidnot reducemortalityorhematomaexpansion.Codestatuschangewas the only predictor significantly associated with increased inhospitalmortality. In this study, Yaghi et al5 have begun to unravel a complex topic that is fraught with confounding factors. The strengths of their study include addressing the relative rarity of sICH by pooling data from large stroke centers. Combined, thestrokecentershada totalof 3894patientswhowere treated with intravenous tPA, and of these patients, 128 had sICH (3.3%). Inaddition,Yaghi et al5 included important clinical and radiographicpredictorsof sICH.They foundanassociationbetween hematoma expansion and hypofibrinogenemia, supporting the rationale for administering blood products to reverse the coagulopathy induced by thrombolysis; however, hypofibrinogenemia was found only in a minority of patients. Yaghi et al5 noteddelays in time todiagnosis and in the initiationof treatment for sICH,which they felt couldhavecontributed to the lackof benefit of treatment. Evenwith the limitations inherent in retrospective analysis, this studyprovides the first comprehensive description of how thrombolysisrelated sICH ismanaged in representative stroke centers in the United States. Although the studymay have been underpowered to detect the benefit of any particular treatment, it highlights the complexity of thepatient population, treatment options, and differences in clinical practice for sICH across the UnitedStates. Importantly, this study informs futureprospectivework,providesdata that call for a reevaluationofour treatment of postthrombolysis sICH, and calls on the stroke community to reevaluatewhat an acceptable risk of thrombolysis should be. Predictors of hemorrhagic transformation after the administration of tPA include advancing age, diabetes mellitus, elevated blood pressure, low platelet count, and increased time to thrombolysis.6,7 Predictors of sICH after tPA administration have also been described and include a high National Institutes of Health Stroke Scale score, large hypodensity detected on admission by computed tomography of the head, and low cerebral blood volume detected by magnetic resonance imaging.7,8 Some of these predictors also influence patient outcome, which highlights the importance of accounting for confounders in analyses. While Yaghi et al5 did include important confounding variables, the retrospective nature of their study did not allow for the capture of other important factors such as blood pressure management and recanalization, which are known to play a role in the development of sICH.9,10 Interestingly, in this study,5 more than 50% of patients with sICH following administration of tPA did not receive treatment. We can only speculate that the lack of treatment might be explained by code status change, advanced directives, or the perception of futility by clinicians and/or family members. While it is intuitive that the rapid reversal of thrombolysis-related coagulopathy could limit the damage due to sICH, there may be other factors that weigh heavily against treatment of sICH for clinicians. It is important to identify these factors and determine whether they are justified. In addition, the variability in treatment for sICH is guided by each individual clinical situation. An attempt to standardize treatment may be warranted and, ultimately, would help in determining whether treatment for sICH is actually beneficial. Lastly, Yaghi et al5 conclude that the lack of benefit in the treated groupmay be partially due to a delay in time to recognition and a delay in the initiation of treatment. They propose continued frequent neurological monitoring of patients Related article page 1451 Opinion Editorial