Date: May 6th, 2021

Guest Skeptic: Dr. Daniel Schwerin is employed with Prisma Health-Upstate as a clinical assistant professor, emergency medicine GME director for emergency medical services and medical director for several local EMS agencies and has lectured on prehospital stroke management.

Reference: Fatima et al. Mobile stroke unit versus standard medical care in the management of patients with acute stroke: A systematic review and meta-analysis. International Journal of Stroke 2020

Case: A 70-year-old man develops sudden right-sided weakness beginning shortly after breakfast and his partner appropriately calls emergency medical services (EMS). Their local EMS service arrives quickly with a conventional ambulance. He has heard about these special ambulances with CT Scanners and wonders if that will make an important difference for his partner.

Background: We have discussed stroke so many times on the SGEM. It is one of the five most popular topics like TXA, PE, POCUS and ketamine. Justin Morgenstern from First10EM and I recently downgraded the NNT website recommendation for tPA in acute ischemic stroke to “yellow”. A yellow recommendation means the benefits and harms are unclear due to the uncertainty in data. But something that often comes up when discussing stroke treatment is we need to go fast because time is brain.

The term “time is brain” was coined by Dr. Camilo Gomez back in 1993. He modified his position in 2018 and said:

“It is no longer reasonable to believe that the effect of time on the ischaemic process represents an absolute paradigm. It is increasingly evident that the volume of injured tissue within a given interval after the estimated time of onset shows considerable variability in large part due to the beneficial effect of a robust collateral circulation.”

We never did have high-quality evidence to support the position that treating stroke patients earlier was better. All we had was an association because there were no RCTs that randomized stroke patients into getting thrombolytics early or late. This means there could have been unmeasured confounders responsible for the observed effect.

The largest placebo controlled RCT looking at tPA for acute ischemic stroke was IST-3 which was covered on SGEM#29. There were several serious problems with that trial including:

  • Largely unblinded trial (91%)
  • Stopped early
  • Self-reported outcome by telephone or mailed questionnaire
  • No superiority for primary outcome
  • 4% absolute increase in early mortality

Another interesting point about IST-3 is the subgroup analysis did not support the claim that time was brain. There was no statistical difference between the <3hrs, 3-4.5hrs and >4.5hrs. However, the point estimate favored tPA in <3hrs, then placebo between 3-4.5hrs and then back to tPA in >4.5hrs? You also need to look very carefully at the figure to see they used the 99% confidence interval instead of the standard 95% confidence intervals. If calculating the Odds Ratio for the 3-4.5hr group you find it is statistically significant favouring the placebo group.

Clinical Question: Does a mobile stroke unit (MSU) with earlier imaging and delivery of tPA improve outcomes, or is the downstream effect of improved resources at a comprehensive stroke center that improves outcomes for patients with strokes?

Reference: Fatima et al. Mobile stroke unit versus standard medical care in the management of patients with acute stroke: A systematic review and meta-analysis. International Journal of Stroke 2020

  • Population: This was a systematic search that found 11 articles that were either randomized controlled trials (RCTs), retrospective or prospective studies that compared the clinical outcomes among patients treated in either a mobile stroke unit or through conventional care/standard medical care for the acute stroke.
    • Exclusions: Case–control studies, case series, and case reports
  • Intervention: MSU that is a specialized ambulance equipped with a CT-scanner, point-of-care laboratory, and thrombolysis is started immediately within the MSU vehicle.
  • Comparison: Conventional care that consists of transferring to the patient the emergency department or specialized stroke centres and given thrombolysis in-hospital according to the imaging report.
  • Outcome:
    • Primary Outcomes: Neurologic outcome as defined by modified Rankin scale (mRS) score at day 7 and day 1 post treatment. Good neurologic outcome was an mRS of 0-2 while a poor neurologic outcome was an mRS of 3-6
    • Secondary Outcomes: All-cause mortality, stroke related-neurological death, other adverse events, and mean time gains

Authors’ Conclusions: Our results corroborate that patients treated in mobile stroke unit lead to short-term recovery following acute stroke without influencing the mortality rate. Further prospective studies are needed to validate our results.”

Quality Checklist for Therapeutic Systematic Reviews:

  1. The clinical question is sensible and answerable. Unsure
  2. The search for studies was detailed and exhaustive. Yes
  3. The primary studies were of high methodological quality. No
  4. The assessment of studies were reproducible. Yes
  5. The outcomes were clinically relevant. No
  6. There was low statistical heterogeneity for the primary outcomes. No
  7. The treatment effect was large enough and precise enough to be clinically significant. Unsure

Results: Eleven publications (seven RCTs and four observational studies) were included in this SRMA with a total of 21,297. Most patients (97%) were from Germany. There were 28% (n=6,065) in the MSU group and 72% (n=15,232) in the conventional care (CC) group. The mean age was 70 years and the mean NIHSS score was 9.8 MSU and 8.4 CC.

Key Result: Better neurologic outcome was not observed at one day post treatment but was at seven days in patients treated by the MSU compared to conventional care.

  • Primary Outcomes:
    • Good neurologic outcome day 7: OR 1.46 (95% CI; 1.306–2.03, p=0.02)
      • 1 RCT (23%) and 2 Obs (77%) n=885
    • Good neurologic outcome day 1: OR 1.18 (95% CI; 0.88-1.57, p=0.26)
      • 1 RCT (16%) and 1 Obs (84%) n=758
  • Secondary Outcomes: There was no statistical difference in mortality, stroke related death or other serious adverse events. Patients were treated 13 min faster with MSU compared to CC which was statistically significant.
    • Mortality: OR 0.98 (95% CI; 0.81–1.18, p=0.80)
    • Stroke-related or neurological death (OR: 1.37, 95% CI: 0.81–2.32, p=0.24)
    • Stroke related neurological deficits: OR 1.37 (95% CI; 0.81–2.32, p=0.24)
    • Other serious adverse events: OR 0.69 (95% CI; 0.39–1.20, p=0.19)
    • Mean time-to-treatment MSU 62 min vs 75 min CC; mobile stroke unit compared to conventional care (62.0 min vs. 75.0 min; p=0.03

 1. External Validity: The majority (97%) of the patients included in the SRMA were from Germany. Europe has a different pre-hospital system than North America. It is unclear if these results could be applied to our practice. Only 618 patients out of the 21,297 patients were from studies done in the USA.

2. GIGO: This is the concept of garbage-in, garbage-out. It means if you combine observational studies which are of lower methodological quality with higher-quality RCTs, mashing them all up into a meat grinder does not get you closer to the “truth”. A case could be made for whether it was even appropriate to meta-analyze some of the data. The secondary outcomes of time to scan and treatment all had an I2 test of heterogeneity of 99%. While they did appropriately use a random effects model for the analysis that level of heterogeneity it would be reasonable to suggest that the data should not have been combined.

Looking at the primary outcomes there were only three studies meta-analyzed for the seven-day result and more than ¾ of the data came from two observational studies. The heterogeneity was also moderate at 44% using the I2 metric. It was even worse for the one-day outcome where there were only two studies included with more than 80% of the data coming from the one observational study.

I grew up on an apple farm. To make great apple pies you need great ingredients. A cow pie is something that comes out of the back of a cow and when it lands on the ground it is about the size of a pie. Adding a cow pie to an apple pie does not make the apple pie taste any better. Combining observational studies to RCTs does not increase my confidence in the results. 

3. Primary Outcome: They had two primary outcomes and SGEM listeners know…there can be only one, primary outcome. However, the outcomes were at one- and seven-days post treatment. This is a very short time frame for stroke studies. Usually, the primary outcome for thrombolytic use in acute ischemic stroke studies is at 3 or 6 months. We know from the NINDS-Part 1 trial published in 1995 that there was not statistical difference in their primary outcome (an improvement of 4 points over base-line values in NIHSS score or the resolution of the neurologic deficit within 24 hours of the onset of stroke symptoms).

4. Misleading: While the study did include over 20,000 patients, less than 1,000 were meta-analyzed for the two primary outcomes. Just reading the abstract could give the impression that this was a much bigger study for the summary statistic odds ratio (OR) of good neurologic outcome.

5. Cost: This is an important aspect to consider. MSU ambulances are much more expensive to buy (initial startup price tag around 1 million dollars) and operate (cost of a critical care paramedic/nurse/CT tech along with having a neurologist on standby to review the head CT) than regular ambulances. If we do not have high-quality evidence that giving thrombolytics 13 minutes earlier makes a patient-oriented outcome difference then the expense cannot be justified.

It is even worse; we do not have high-quality evidence that tPA is superior to placebo. Out of the eight foundational RCTs on this issue, only two out of the eight claimed benefit for their primary outcome (NINDS-II and ECASS-III). Neither trial was ever replicated and both have been reanalyzed (Hoffman and Schriger and Alper et al) and failed to confirm superiority of tPA for acute ischemic stroke. So, should we be spending lots of money and resources to deliver a drug faster that we are unsure if it even works.

Comment on Authors’ Conclusion Compared to SGEM Conclusion: We would amend the conclusions to say that weak evidence provides mixed results on MSU being associated with better short-term outcomes while no statistical difference was observed in mortality rates. Further high-quality results would help clarify questions around the utility of MSU.

SGEM Bottom Line: We cannot recommend the use of MSU based on the available evidence.

Case Resolution: The man is taken to the closest stroke center and is worked up for acute ischemic stroke.

Dr. Daniel Schwerin

Clinical Application: The evidence does not support local EMS agencies to purchase specialized ambulances with CT scanners to set up mobile stroke units.

What Do I Tell My Patient? Your partner appears to be having a stroke. There are some places that have these special ambulances with CT scanners. The evidence shows they get people treated a few minutes faster. However, we don’t have good evidence those minutes makes a clinical difference.

Keener Kontest: There was no winner last week. The correct answer was William Kouwenhoven (1886–1975) who is credited as the founder of modern CPR.

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

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