Date: May 25th, 2021

Guest Skeptic: Dr. Garreth Debiegun is an emergency physician at Maine Medical Center in Portland, ME and clinical assistant professor with Tufts University School of Medicine. He also works at an urgent care and a rural critical access hospital. Garreth is interested in wilderness medicine and is the co-director of the wilderness medicine clerkship at Maine Med, and the medical director for Saddleback Ski Patrol and for Maine Region NSP. At work Garreth imagines himself as a student of evidence-based medicine trying to provide the best care based on the best evidence.

References:

  • Suzuki et al. Effect of Mechanical Thrombectomy Without vs With Intravenous Thrombolysis on Functional Outcome Among Patients With Acute Ischemic Stroke: The SKIP Randomized Clinical Trial. JAMA 2021
  • Zi et al. Effect of Endovascular Treatment Alone vs Intravenous Alteplase Plus Endovascular Treatment on Functional Independence in Patients With Acute Ischemic Stroke: The DEVT Randomized Clinical Trial. JAMA 2021

Case: A 74-year-old woman presents to your emergency department with 1.5 hours of right-sided weakness, aphasia, and neglect. On rapid bedside assessment you calculate the National Institute of Health Stroke Score/Scale (NIHSS) of 11 and a Field Assessment for Stroke Triage for Emergency Destination (FAST-ED) score of 7; you suspect a large vessel occlusion (LVO) based on the high NIHSS and FAST-ED score >3. A non-contrast CT shows no evidence of intracranial hemorrhage. A CT angiogram plus CT perfusion demonstrate a clot in the left proximal middle cerebral artery (MCA) with a small infarcted area and a large penumbra. Based on your institution’s current guidelines, the patient is a candidate for endovascular therapy, but they are also within the current window for the administration of thrombolytics. You wonder if you gotta be starting something?  Specifically, you wonder if you should give the thrombolytics while waiting for your neurointerventional/endovascular team?

Background: Management of acute ischemic stroke has been discussed on the SGEM ever since this knowledge translation project was launched in 2012. My position remains that there is uncertainty as to whether tPA provides a patient-oriented outcome and this was summarized in the downgrading of the NNT recommendation to “yellow” with Dr. Justin Morgenstern from First10EM.

The world of stroke management has changed in the last few years the more robust evidence to support endovascular therapy (EVT) for large vessel occlusion (LVO) strokes. The tipping point came with the publication of MR. CLEAN in 2015. It was a multicenter, randomized, unblinded trial of patients with an LVO stroke in the anterior circulation treated in less than six hours after onset of symptoms. The primary outcome was good neurologic function defined as a modified Rankin Scale (mRS) score of 0-2 at 90 days. The trial included 500 patients and reported an absolute risk reduction of 14% (33% EVT vs 19% control) with a NNT of 7.

Four other RCTs were stopped early due to the publication of MR. CLEAN.  All of these were published in the NEJM in the following six months.

  • SWIFT PRIME (n=196) NNT of 4
  • EXTEND-IA (n=70) NNT of 2-5 depending on which outcome you picked
  • REVASCAT (n=206) extended the window to eight hours NNT 6.5
  • ESCAPE (n=316) extended to 12 hours NNT 4

There are dangers with stopping trials early. Dr. Gordon Guyatt wrote in the BMJ that it can introduce bias towards efficacy. He said you should have a high level of skepticism regarding the findings of trials stopped early for benefit, particularly when those trials are relatively small, and replication is limited or absent.

A systematic review and meta-analysis (SRMA) was published by Badhiwala, JH et al in JAMA 2015 looking at EVT for stroke. The first three RCTs reported no superiority. Then came MR. CLEAN and those four additional studies that were stopped early. The forest plot shows a point estimate that favors EVT and the 95% confidence interval does not cross the line of no statistical difference. We do need to be careful in not over interpreting this data because the bias of stopping trials early could have been compounded in the SRMA process.

A couple more RCTs were published since the 2015 SRMA:

  • DAWN 2018 (n=206) NNT 3 between 6-24 hours, stopped early
  • DEFUSE 2018 (n=182) NNT 4 between 6-16 hours, stopped early

Here is a lists all the studies and the number of patients enrolled/center/year put together by friend of the SGEM, Dr. Salim Rezaie (REBEL EM). The average was 4.6 patients/year from these stroke centres.

While the NNT for EVT looks very impressive, it is suffering from denominator neglect. Once the patient is getting the intervention the NNT is in the single digits. However, how many possible stroke patients were screened, and code strokes were called to get the impressive NNTs of the EVT trials?

Most of these trials didn’t tell us how many patients they screened to get the patients they registered.  Only one did – EXTEND-IA.  In that RCT, they screened ~7,800 patients, and registered 70. That’s 0.9%, or about 1/110, that were eligible for the treatment. The NNT in MR CLEAN, the largest and arguably the most methodologically sound of these trials, was 7.  If you must screen 110 patients to find one eligible, and need to treat 7 to benefit one, then about one out of every 770 patients having a potential stroke would benefit from this therapy.

Therefore, we think the NNT provided by the RCTs suffer from denominator neglect. An NNT of 770 is not nearly as impressive as an NNT of 7. The agenda is in the numerator, but the devil is in the denominator.

One major component of LVO management is the use of systemic thrombolytics in patients presenting within the current thrombolytic treatment window prior to endovascular intervention. However, it’s unclear if systemic thrombolytic administration results in better outcomes or if it simply exposes the patient to increased risks at a higher cost.

There is a SRMA by Mistry et al published in Stroke 2017 on the issue of EVT +/- tPA. They included 13 studies (n=1,769) with three RCTs (which accounted for only 25% of patients) and ten observational studies (which accounted for the other 75% of the patients). They reported an NNT of 17 for good neurologic outcome (defined as mRS 0-2) with EVT + tPA compared to EVT alone. However, it was not statistically significant if you just looked at the RCT data. This is a problem of combining lower quality observational studies with higher-quality RCTs.

We looked at a 2020 RCT by Yang P et al published in the NEJM on SGEM#292. This trial asked if EVT alone was non-inferior to EVT + tPA in stroke patients with LVOs presenting withing 4.5 hours of symptoms. The cohort consisted of 654 patients and the key results was EVT alone was not non-inferior to EVT plus tPA. Our SGEM bottom line was that there does not appear to be a role for systemic thrombolysis in acute ischemic stroke for appropriate patients when EVT is readily available.

Two more trials have been published and that is what we are going to review today on the SGEM.


Clinical Question: Does intravenous thrombolysis prior to mechanical thrombectomy increase the percentage of patients with LVO who achieve mRS of 0-2 (functional independence) at 90 days?


Reference #1:  Suzuki et al. Effect of Mechanical Thrombectomy Without vs With Intravenous Thrombolysis on Functional Outcome Among Patients With Acute Ischemic Stroke: The SKIP Randomized Clinical Trial. JAMA 2021

  • Population: Patients aged 18-85 evaluated at one of 23 stroke centers in Japan within 4.5 hours of onset, who were found to have: NIHSS ≥6, CTA or MRA proven ICA or M1 occlusion, minimal imaging changes of ischemic stroke. Patients had to be previously independent (mRS 0-2) and meet Japanese criteria for thrombolysis.
    • Exclusions: Contraindication for contrast agent or EVT, contraindication for IVT, presence of severe renal disorder, pregnancy or possibility of pregnancy, unlikely to complete the study, such as due to progressive malignant tumor, judged incompatible with the study by the investigators
  • Intervention: Mechanical thrombectomy of LVO without additional thrombolytics
  • Comparison: Alteplase given at a dose of 0.6 mg/kg in addition to mechanical thrombectomy performed within 30 minutes of randomization
  • Outcome:
    • Primary Outcome: mRS evaluated by physical exam or phone interview at 90 days.
    • Secondary Outcomes: Ordinal shift analysis of mRS, mRS 5-6, mRS 0-1, mRS 0-3, 90-day mortality, successful reperfusion, recanalization and adverse events (ICH)

Authors’ Conclusions: Among patients with acute large vessel occlusion, mechanical thrombectomy alone, compared with combined intravenous thrombolysis plus mechanical thrombectomy, failed to demonstrate noninferiority regarding favorable functional outcome. However, the wide confidence intervals around the effect estimate also did not allow a conclusion of inferiority.”

Quality Checklist for Randomized Clinical Trials:

  1. The study population included or focused on those in the emergency department. Unsure
  2. The patients were adequately randomized. Yes
  3. The randomization process was concealed. Yes
  4. The patients were analyzed in the groups to which they were treated (per-protocol analysis for non-inferiority trials). Yes
  5. The study patients were recruited consecutively (i.e. no selection bias). Unsure
  6. The patients in both groups were similar with respect to prognostic factors. Unsure
  7. All participants (patients, clinicians, outcome assessors) were unaware of group allocation. No
  8. All groups were treated equally except for the intervention. Unsure
  9. Follow-up was complete (i.e. at least 80% for both groups). Yes
  10. All patient-important outcomes were considered. Yes
  11. The treatment effect was large enough and precise enough to be clinically significant. No

Results: They recruited 204 patients into this trial with a mean age of 74 years, 63% male and a median NIHSS score of 18.


Key Result: EVT alone was not non-inferior to EVT + tPA


  • Primary Outcome: Favorable neurologic outcome (mRS 0-2) at 90 days
    • Intention-to-Treat (ITT) Analysis: 59.4% EVT vs 57.3% EVT+ tPA. Absolute difference of 2.1% (95% CI; -11.4 to ꝏ), unable to conclude noninferiority.
    • Per-Protocol (PP) Analysis: 60.8% EVT vs 58.8% EVT + tPA. Absolute difference of 2.1% (95% CI; -13.7% to ꝏ), unable conclude noninferiority.
  • Secondary Outcomes: There was no statistical difference between the groups in any of the secondary outcome except for adverse events. There was an 13% absolute increase in ICH with EVT + tPA compared to EVT alone. However, symptomatic ICH was not statistically different.

1. Non-Inferiority Trials: This was a non-inferiority trial, designed to show that one therapy is not significantly worse than the other. This study failed to show this based on their delta of an odds ratio (OR) of 0.74. Theyreported an OR of 1.09 favoring EVT alone but the lower part of the 95% confidence interval was 0.63. This trial also does not demonstrate superiority or equivalence.

2. External Validity: This trial was conducted in Japan which introduces two differences. First, Japan uses a dose of alteplase at 0.6 mg/kg rather than the 0.9 mg/kg dose used in North America. This was discussed on SGEM#325 when critically appraising the THAWS trial with Professor Fatovich. One could argue that a higher dose may have made a bigger difference, but it also could have increased the risk of bleeding. It is not clear how this would have affected the results.  The reason a lower dose was used is because Asians are reported as having a higher rate of intracerebral atherosclerosis (Kim JS, 2014), so it is not clear how this would change the results if applied to non-asian populations.

3. Secondary Outcomes: The study found no difference on most outcomes. The only difference was higher rate of bleed at 36 hours in the group which received tPA and yet symptomatic head bleed was not different between the two groups. Despite this, we know that ICH can predict a worse prognosis. Further study would be needed to clarify the relationship between thrombolytics, ICH and worse outcomes.

4. Subgroup Analysis: As mentioned in the checklist there was a larger percentage of men in the combination therapy group than the thrombectomy only group. However, in subgroup analysis of gender, also of age, atrial fibrillation, blood glucose, antithrombotic agent, NIHSS score and randomization time there was no difference between groups.  This suggests that gender is not a major factor in outcomes.

5. Power Calculations: The authors cite in their limitations that they calculated a non-inferiority margin on 0.74 (that combination therapy is not more than 26% worse than thrombectomy therapy alone). The authors state three reasons this may not be the correct margin. First, this margin is based on prior data using 0.9 mg/kg rather than 0.6 mg/kg used in this study.  Second, the margin was selected using the fixed-margin method rather than the minimal clinically important difference.  Third, favorable outcomes in this study were more frequent than prior data.  Finally, we would add that being an open label trial may alter the data results also affecting the power calculation. At the end of the day, it does not matter because we do not need to estimate effect size as it has already been determined. What is left is the interpretation of a 2.1% absolute difference between the two interventions that were not statistically different and an OR that favored EVT alone with the lower end of the 95% CI being 0.63.

Comment on Authors’ Conclusion Compared to SGEM Conclusion: We generally agree with the authors conclusion.


Reference #2:  Zi et al. Effect of Endovascular Treatment Alone vs Intravenous Alteplase Plus Endovascular Treatment on Functional Independence in Patients With Acute Ischemic Stroke: The DEVT Randomized Clinical Trial. JAMA 2021

  • Population: Patients 18 years of age or older with CTA or MRA proven ICA or M1 occlusion stroke eligible for treatment with tPA within 4.5 hours from last known well.
    • Exclusions: ICH, contraindications to tPA mRS of 0-2 at baseline, pregnant or lactating, contraindication to contrast agents, arterial disease that would prevent EVT, pre-existing neurologic or psychiatric disorders that would confound the neurologic exam, multiple occlusions, mass, cerebral vasculitis, AVM or aneurysms, terminal illness, unlikely to follow up at 90 days, and currently in another trial.
  • Intervention: EVT of LVO without additional tPA
  • Comparison: Standard dose tPA (0.9 mg/kg) in addition to EVT
  • Outcome:
    • Primary Outcome: Proportion of patients achieves mRS 0-2 at 90 days evaluated by two blinded neurologists based on voice/video recordings.
    • Secondary Outcomes: Distribution of mRs at 90 days, mRS 0-1, reperfusion on angiography during procedure, reperfusion on CTA or MRA, change in NIHSS at 24 hours and between 5-7 days and EQual-5D (European Quality of life 5-Dimension scale) score at 90 days

Authors’ Conclusions: Among patients with ischemic stroke due to proximal anterior circulation occlusion within 4.5 hours from onset, endovascular treatment alone, compared with intravenous alteplase plus endovascular treatment, met the pre-specified statistical threshold for noninferiority for the outcome of 90-day functional independence.”

Quality Checklist for Randomized Clinical Trials:

  1. The study population included or focused on those in the emergency department. Unsure
  2. The patients were adequately randomized. Yes
  3. The randomization process was concealed. Yes
  4. The patients were analyzed in the groups to which they were treated (ie a per-protocol analysis for non-inferiority trials). Yes
  5. The study patients were recruited consecutively (i.e. no selection bias). Unsure
  6. The patients in both groups were similar with respect to prognostic factors. Yes
  7. All participants (patients, clinicians, outcome assessors) were unaware of group allocation. No
  8. All groups were treated equally except for the intervention. Unsure
  9. Follow-up was complete (i.e. at least 80% for both groups). Yes
  10. All patient-important outcomes were considered. Yes
  11. The treatment effect was large enough and precise enough to be clinically significant. No/Yes

Results: They assessed 509 patients for eligibility. More than half were excluded with 235/509 (46%) ultimately being randomized to EVT alone or EVT + tPA. The median age was 70 years, 56% were male and the median NIHSS score was 16.


Key Result: EVT alone was non-inferior to EVT + tPA


  • Primary Outcome: Favorable neurologic outcome (mRS 0-2) at 90 days
    • Per-Protocol (PP) Analysis: 53.2% EVT vs 46.0% EVT + tPA, absolute difference of 7.1% (95% CI; -5.9 to ꝏ), able to conclude non-inferiority because less than their pre-specified 10% delta.
  • Secondary Outcomes: There was no statistical difference in any secondary outcome except for adverse events. As in the other RCT, there was a statistical increase in ICH with EVT+tPA but the difference in symptomatic ICH was not statistically.

1. TIAs: One of the secondary outcomes measured in this study was NIHSS at 24 hours. We know from NINDS part 1 that thrombolytics was not superior to placebo for NIHSS score at 24 hours. Furthermore, in this study successful reperfusion was found in 3 patients (2.6%) who received thrombolytics prior to thrombectomy and in 2 patients (1.7%) who did not.  This clearance of clot suggests that thrombolytics do not have a major role in clot clearance.

2. Stopped Early: This study was stopped early after enrolling only 235/970 planned patients because of a statistical finding of likely futility. However, we know that stopping studies early can introduce bias. This bias can be compounded when included in a SRMA. It would have nice to continue enrolling more patients unless harm was clear evidence of harm. However, stopping studies early usually makes a study more likely to find a non-true benefit rather which this study did not find so we are more willing to believe this data.

3. M1 vs M2 Occlusion: In both this study and the prior study only patients with ICA (internal carotid artery) or M1 occlusion were considered candidates for enrollment. Many institutions will perform thrombectomy on patients with M2 occlusions or ACA (anterior cerebral artery) occlusions so it is not clear if patients with smaller, more distal occlusions would be more or less likely to benefit from intravenous thrombolytics.

4. Intracranial Hemorrhage (ICH) Rates: NINDS and subsequent other trials reported ICH rates around 2-6%. This study found ICH rates to be 21.7% for thrombectomy only group and 32.5% for the combination group. This is a much higher rate and is similar to but lower than the SKIP trial (33.7% and 50.5%).  Clearly this data support that LVO has a much higher ICH rate than other strokes and therefore thrombolytics may carry even higher risk to patients than we have historically considered.

5. Comparison to Prior Data: Prior evidence has suggested that patients who received intravenous thrombolytic therapy had better outcomes.  However, this came from observational data.  It is possible, even probable, that the patients who were not given intravenous thrombolytics had contraindications which also were predictors of poor outcome.  This is now a third study with data suggesting no benefit from adding thrombolytics to thrombectomy.  As further data come out, practice may shift.  Also, this is further data raising the possibility that intravenous thrombolytics may not provide benefit to any stroke patients and may provide impetus to re-perform NINDS or ECASS III with the newer imaging technology. 

Comment on Authors’ Conclusion Compared to SGEM Conclusion: We generally agree with the authors conclusions.


SGEM Bottom Line: Three studies have now suggested there may be no significant difference in patients with LVO who are treated with only thrombectomy vs those who are treated with combined thrombectomy and intravenous thrombolytics.


Case Resolution: Per your hospital protocols you call a code stroke. Neurology comes and assesses the patient. They order intravenous alteplase and await their neuro-interventional colleagues to evaluate the patient to see if she may benefit from thrombectomy.  After your shift, you email your facility’s stroke neurologist to open a discussion about your hospital protocols and consider if thrombolytics should be removed from protocols for patients with LVO.

Dr. Garreth Debiegun

Clinical Application: Current data support careful consideration of potential benefits and potential harms of systemic thrombolytics in patients with LVO, consideration must be given to patients who are candidates for interventional procedures.  As more data is available current protocols should be reconsidered. Further studies are pending: MR CLEAN NO IV, SWIFT DIRECT, and DIRECT-SAFE.

What Do I Tell My Patient?  Your symptoms are due to a stroke.  The CT scan shows there is a clot in an artery in your brain. We have a team that may be able to remove this clot like in heart attacks.  There is also a clot busting drug that may help to break up these clots.  Data suggests that adding this clot busting drug to the clot removal procedure may not improve your outcome. However, we do not have definitive evidence and current guidelines still support using this clot busting drug.  Let’s discuss the potential benefits and potential harms of this drug with you and your family to see if it is something you would want to try.

Keener Kontest: Last weeks’ winner was Dr. Kristen Johannessen an ED physician from Williamsburg, Virginia. She knew President Ronald was shot in the left side of his chest.

Listen to the podcast this week to hear the trivia question. Email your answer to TheSGEM@gmail.com with “keener” in the subject line. The first correct answer will receive a cool skeptical prize.

Other FOAMed:

  • REBEL EM: The DEVT + SKIP Trials: Does Systemic Thrombolysis Prior to Endovascular Treatment Improve Outcomes in Large Vessel Occlusion Strokes?


Remember to be skeptical of anything you learn, even if you heard it on the Skeptics Guide to Emergency Medicine.