A review of the revised ACEP Clinical Policy – IV tPA downgraded to Level B evidence
Nothing seems to stir the emotions of an emergency physician more than a discussion of tPA for stroke: who, why, and when. Skeptics cite problems in the data, and conflicts of interest of the authors. Supporters trust the literature that passed peer review and became policy of emergency medicine and neurology professional societies around the world.
2012 and 2015
The 2012 ACEP Clinical Policy on the use of tissue plasminogen activator for acute ischemic stroke  raised plenty of questions. It resulted in concerns about the effectiveness of the drug, the magnitude of its effectiveness, the risks of cerebral hemorrhage, effects of the drug on stroke mimics, legal implications of the policy, the policy’s impact in different practice settings, hospital differences in resources for stroke diagnosis and emergency care, and the lack of tools for patient/family shared decision making [2,3,4,5]. The key question asked was ‘Is IV tPA safe and effective for acute ischemic stroke if given within 3, and within 3-4.5 hours after symptom onset? The level A recommendation stated: IVtPA should be offered …to patients who meet …(NINDS) inclusion/exclusion criteria and can be treated within 3 hours after symptom onset. The level B recommendation stated: IVtPA should be considered in patients who meet…(ECASS III) inclusion/exclusion criteria and can be treated between 3 to 4.5 hours after symptom onset.
Criticism of this policy was swift, in these pages and elsewhere. Level A recommendations imply generally accepted principles of care and as such seemed to ignore or shut down years of well-intentioned, well-reasoned debate on tPA’s safety and efficacy. Subsequently, ACEP took the unprecedented step of reconsidering their policy, which had been co-authored with the American Academy of Neurology.
In June 2015, the ACEP Board approved a revised clinical policy on the use of tPA for ischemic stroke . The question asked was the same as in 2012 – whether IV tPA is safe and effective within 3 hours of symptom onset, and within 3-4.5 hours of symptom onset.
Is IV tPA safe and effective for patients with acute ischemic stroke if given within 3 hours, and within 3-4.5 hours of symptom onset? Recommendations for <3 and 3-4.5 hours are combined below for simplicity.
- Level A recommendation: None.
Note that the 2015 policy no longer provides level A recommendations for IVtPA.
- Level B recommendation for <3 hr administration: IV tPA should be offered and may be given …within 3 hours after symptom onset at institutions where systems are in place to safely administer the medication. Consider the increased risk of sICH.
- Level B recommendation for 3-4.5 hr administration: despite the known risk of sICH and the variability in the degree of benefit in functional outcomes, IV tPA may be offered and may be given to carefully selected patients…within 3-4.5 hours after symptom onset at institutions where systems are in place to safely administer the medication.
- Level C recommendation: When feasible, shared decisionmaking…should include a discussion of potential benefits and harms…
Risks and Benefits
Here are selected comments from Appendix D and text, using the most robust analyses that were provided for risk-benefit analysis:
IVtPA given < 3 hrs
- NNT = 8 (95% CI 4-31) NINDS
- NNH=17 (95% CI 12-34) NINDS
- sICH 5-7% sICH overall (could be lower with adherence to protocol; sICH definitions also vary with study)
- Proportion who improved: 13% achieving Rankin Scale 0-1 (NINDS 39% IVtPA vs 26% control)
IVtPA given 3-4.5 hrs
- NNT = 14 (95% CI 7-244) ECASS III
- NNH= 23 (95% 31-78) ECASS III
- sICH 3-8% and 2-6% depending on NINDS definition of sICH
- Proportion who improved: 7% achieving Rankin Scale 0-1, ECASS III, 52% for IVtPA vs 45% for control)
Who benefits and when?
Published data is somewhat population-based and not at all individual-based, so we are still in the dark about the patients most likely to benefit from IVtPA. If only 13% benefit from IVtPA if given within 3 hours, or 7% in 3-4.5 hours, what about the rest? What about the cost and resources used? Who exactly are the ‘carefully-selected’ patients who may receive IVtPA at 3-4.5 hours? The benefits of giving tPA in <3 hours, or ≤90 minutes also needs discussion .
Who is most at risk for sICH?
Several scoring systems for assessing the likelihood of sICH after IVtPA have been developed . Predictive variables and scoring methods vary with the tool, but typically include early infarct signs on non-contrast CT scan; prestroke Rankin scale score; age; antithrombotic therapy; hyperglycemia; and NIHSS on admission.
What about the ED, the hospital and the hospital system?
Stroke diagnosis and emergency management is not a one-man emergency physician show. It requires a team to make what is often a difficult diagnosis and treatment plan. All members of the team should use the same unified protocol: indications for activating Code Stroke; NIHSS reporting; imaging protocols; scoring systems for risk-benefit analysis; and well-articulated and well-presented information and consent forms.
What about tools for shared decision-making?
Involving patients and family in decision-making is more complex than asking ‘do you want the clot-buster’? Patients who elect a treatment, whether a stent for STEMI, or IVtPA for stroke, assume they will personally benefit from the treatment. This is not the case with IVtPA where most do not benefit. Only a few do. And a small number are harmed. Graphs and pictorial images are an excellent way to communicate a risk-benefit analysis, but are in various stages of development [10,11]. The challenge for tool developers is to convey information simply, clearly, and quickly given a variety of clinical scenarios.
The 2015 ACEP clinical policy discussion is detailed. Study deficiencies are noted, the discussion uses updated data, and methodology, case definitions, heterogeneity and outcome differences are noted. IVtPA was lowered from a level A to level B recommendation, and qualifications on the proper environment for drug administration were added. However, these qualifications were not spelled out in any detail and key issues relative to decision-making remain unaddressed. The revised policy reminds us that data on risks and benefits are still a moving target. What is clear is that the pressure is on for emergency physicians who must be the initiators and leaders of the emergency stroke care team.
- Clinical Policy: Use of Intravenous tPA for the Management of Acute Ischemic Stroke in the Emergency Department . Annals of Emergency Medicine, 61:2, February 2013, 225- 243
- Ellison D, ‘The Lytic Lottery’ Annals of Emergency Medicine, 62:5, October 2013, 432-33
- Newman, David, ‘Thrombolytics for Acute Ischemic Stroke’ ,theNNT.com, March 25, 2013 accessed Aug 7, 2015
- Hoffman JR and Schriger DL ‘A Graphic Reanalysis of the NINDS Trial’ Annals of Emergency Medicine 54:3 September 2009, 329-336
- Klauer, K ‘tPA and the Problems of ‘Indication Creep. Emergency Physicians Monthly, May 29, 2013.
- Clinical Policy: Use of Intravenous Tissue Plasminogen Activator for the Management of Acute Ischemic Stroke in the Emergency Department. www.acep.org/workarea/DownloadAsset.aspx?id=102373
- Zinsser, William. On Writing Well. The Classic Guide to Writing Nonfiction’. 7e, 2006, HarperCollins, New York.
- Miller JB, Heitsch L, Siket M, et al ‘The Emergency Medicine Debate on tPA for Stroke: What is Best for our Patients? Efficacy in the First Three Hours’ Acad Emerg Med 2015 July; 22(7):852-5
- Asuzu D, Nystrom K, Amin H et al ‘Comparison of 8 scores for predicting symptomatic intracerebral hemorhage after IV thrombolysis’ NeurocritCare 2015 Apr; 22 (2):229-33. PMID 25168743
- Gadhia J, Starkman S, Ovbiagele B, et al ‘Assessment and Immprovement of Figures to Visually Convey Benefit and Risk of Stroke Thrombolysis’ Stroke 2010 February; 41 (2): 300-306
- Flynn D, Ford GA, Stobbart L et al ‘A review of decision support, risk communication and patient information tools for thrombolytic treatment in acute stroke: lessons for tool developers. BMC Health Services Research 2013, 13:225