E. Bradshaw Bunney, MD Stroke Care within the 3 Hour IV tPA Window: Why IV tPA, or What Alternatives?

E. Bradshaw Bunney, MD E. Bradshaw Bunney, MD Associate Professor Department of Emergency Medicine University of Illinois at Chicago University of Illinois Hospital Our Lady of the Resurrection Medical Center

E. Bradshaw Bunney, MD Disclosures AstraZeneca, advisory board Genentech, speakers bureau ACEP Scientific Review Committee Executive Board, Foundation for Education and Research in Neurologic Emergencies

E. Bradshaw Bunney, MD Objectives Review the evidence for the use of thrombolytics in ischemic stroke Discuss the Phase IV and re-analysis results Review the concerns about the use of thrombolytics Discuss other alternatives to thrombolytics.

E. Bradshaw Bunney, MD Case 19 yo female collapsed a work on Super Bowl Sunday 2006 EMS found her not moving her right side, aphasic, eyes deviated to the left Onset time 20 minutes prior to EMS arrival BP 120/62, HR 84, RR 14

E. Bradshaw Bunney, MD Case In ED – Friend confirms onset time Friend states no PMHx, no drug or alcohol use PE - R arm 0/5 strength, R leg 3/5, aphasic, eyes deviated to L No family available

E. Bradshaw Bunney, MD Case Glucose = 97 Not pregnant CBC, electrolytes, coagulation all normal CT head = normal

E. Bradshaw Bunney, MD Case Is this a stroke? Seizure? Hysteria? Drugs? What do you do next? Thrombolysis?

E. Bradshaw Bunney, MD NINDS 11 years ago 624 subjects 2 arms –24 hour follow up = no improvement –30 day follow up = improvement

E. Bradshaw Bunney, MD NINDS Trial Results Percentage with favorable outcome t-PA Placebo No. of patients: Modified Rankin Scale40%28% Glasgow Outcome Scale43%32% NIHSS34%20% Symptomatic ICH (within 36 hr) 6.4%0.6% Death (by 90 days)17%21%

E. Bradshaw Bunney, MD IV Thrombolysis  14% absolute increase for the best clinical outcomes as measured by an NIHSS of 0-1.  Benefit = Need to treat 8 patients with t-PA in order to have one additional patient with this best outcome.  6% absolute increase in the number of symptomatic ICH.  Harm = Will have one symptomatic ICH for every 16 patients treated with t-PA.  2 patients will have a minimal or no deficit for everyone patient with a symptomatic ICH

E. Bradshaw Bunney, MD Phase IV t-PA Trials Can the NINDS trial results be replicated? What happens in “the real world”? Do protocol violations make a difference?

E. Bradshaw Bunney, MD Phase IV t-PA trials AuthorEligible patients Patients receiving tPA(%) Mean time to Rx Median NIHSS score Favorable outcome % ICH% Symptom atic ICH % Protocol deviation NINDS %10.9%6.4% Chiu103530(2.9%)2’37”1463%10%6.6% Tanne189>2’11-159%5.8%30% Wang90057(6.3%)2’28” %9%5%9% Buchan154068(4.4%)1595%31%9%16% Albers3892’44” %11.5%3.3%33% Katzan394870(1.8%)1222%15.7%50% Chapman255646(1.8%)2’45” %9%2.2%17% Grotta (16%)2’17”1433%4.5%13% Bravata631517%6%67% Total12,282928(5.8%)2’25” %9.6%5.2%13-67%

E. Bradshaw Bunney, MD Meta-analyses

Meta-analyses Wardlaw et al. Net benefit despite hazards For 1000 treated up to 6hrs, 55 improve, 20 die Heterogeneity and wide CI make results unreliable Additional trial data required

E. Bradshaw Bunney, MD Meta-analyses Graham et al., 15 published reports ICH rate 5.2%, total death rate 13.4% All better than NINDS Lysis can be used safely across wide variety of practice settings

E. Bradshaw Bunney, MD Meta-analyses Hacke et al. 6 randomized trials Sooner thrombolytics given the greater the benefit Particularly when given within 90 min. of onset

E. Bradshaw Bunney, MD CONTROVERSY: Meta-analysis Hoffman and Cooper Pooled data can not replace new or confirmatory data Meta-analyses did not include streptokinase trials which were negative No reason to exclude streptokinase

E. Bradshaw Bunney, MD Re-analysis

NINDS Re-analysis Does the protocol work? Do subgroup imbalances invalidate the entire trial?

E. Bradshaw Bunney, MD Baseline NIHSS Imbalance NIHSS Score > 20 No. of patients Placebo (n=312) t-Pa (n=310) Chi-square (4 DF) = 14.8; p = 0.005

E. Bradshaw Bunney, MD Favorable Outcome Related to Baseline NIHSS - Modified Rankin Scale Favorable Outcome Related to Baseline NIHSS - Modified Rankin Scale

E. Bradshaw Bunney, MD Baseline NIHSS - Specific Odds Ratios Test for equal ORs: Chi-square (4 DF) = 1.70; p = 0.79 Insufficient evidence was found to a declare a difference in treatment effects (ORs) across the five strata

E. Bradshaw Bunney, MD OTT Analysis Report Review Committee had concerns about analyzing OTT as a continuous variable Uncertainty about the exact time of stroke onset. OTT distribution was nonlinear with 25% of all the patients having OTT values of either 89 or 90 minutes.

E. Bradshaw Bunney, MD Symptom onset vs Cumulative % Time from symptom onset to treatment (minutes) Cumulative percentage

E. Bradshaw Bunney, MD NINDS ICH Analysis # of Risk Factors # of patients treated with t-PA (n=310) # Symptomatic ICHs (# of placebo patients with ICH) Percentage (%) (1) (1)4.9 > Risk Factors for ICH: Baseline NIHSS > 20 Age > 70 years Ischemic changes present on initial CT Glucose > 300 mg/dl (16.7 mmol/L)

E. Bradshaw Bunney, MD Re-analysis Conclusions  The independent reanalysis of the NINDS t-PA clinical trial confirms the results from the initial NEJM publication  Support the use of t-PA in stroke patients within three hours of symptom onset  Number needed to treat calculation based on this reanalysis confirms that approximately 8-10 patients need to be treated with t-PA in order to cause one extra patient to have the best clinical outcome.  2 patients will improve for every one that develops a symptomatic ICH

E. Bradshaw Bunney, MD EM Physicians and Lysis Brown et al. 1,105 of 2600 ACEP members responded 40% not likely to use thrombolytics –65% risk of ICH –23% perceived lack of benefit –12% both Upper limit ICH rate 3.4% Lowest acceptable relative improvement 40%

E. Bradshaw Bunney, MD If not t-PA, then what? Most therapies studied outside the 3 hour window – Intra-arterial thrombolysis – Mechanical clot removers – Neuroprotectants – Hypothermia Due to time needed to complete the procedure – may not be true for neuroprotectants

E. Bradshaw Bunney, MD If not t-PA, then what? When uncertain about the diagnosis other tests may be needed – CTA – MRI – Angiography This will frequently cause the 3 hour window to expire, but allows for other interventions – Triple play = stent, mechanical clot removal, intra-arterial thrombolytics

E. Bradshaw Bunney, MD If not t-PA, then what? Newer therapies have small trials compared to IV t-PA IV t-PA has been shown to be effective Stroke neurologists prefer IV t-PA and then a “second look” with further diagnostic tests – MRI – CTA Do not wait

E. Bradshaw Bunney, MD Informed Consent: Documentation With t-PA, there is a 30% greater chance of a good outcome at 3 months With t-PA use, there is 10x greater risk of a symptomatic ICH (severe bleeding stroke) Mortality rates at 3 months are the same regardless of whether t-PA is used 2 patients will have a minimal or no deficit for every one patient with a symptomatic ICH

E. Bradshaw Bunney, MD Documentation Just as important “The patient is NOT a candidate for t-PA because…”

E. Bradshaw Bunney, MD Case Small hospital, no neurologist interested in seeing the patient Called 2 Universities before finding one to accept the patient Family arrived, patient not improving

E. Bradshaw Bunney, MD Case Stroke neurologist = “Give IV t-PA” t-PA given at 2 hours 15 minutes from onset R arm movement and aphasia improving prior to transfer

E. Bradshaw Bunney, MD Case MRI at University = small infarct ECHO cardiogram = Patent foramen ovale, likely embolic stroke Outcome = normal except small vision loss.

E. Bradshaw Bunney, MD Conclusion Data supports the use of IV t-PA when the NINDS protocol is strictly followed Develop a protocol that allows patients to have the greatest chance of receiving therapy as quickly as possible Sooner is better Document well on all patients, t-PA or not

E. Bradshaw Bunney, MD Questions? Brad Bunney Ferne_eusem_2006_bunney_3hour_100606_finalcd 8/1/ :21 AM