Kevin S. Attenhofer, MD
Although tissue plasminogen activator (tPA) has been approved in the United States for treatment of acute ischemic stroke since the mid-90s, there persists a sharp divide between neurologists and the emergency medicine community regarding the safety and efficacy of tPA. Imbalances between the treatment and control groups in the NINDS tPA study, criticisms of the subjective nature of functional outcome scores, and allegations of conflicts of interest have all contributed to the lingering controversy surrounding tPA. I have personally encountered significant resistance and hostility to IV tPA from other providers over 20 years after its approval. In this article, Muruet et al. add to a growing volume of literature that should help the community neurologist assuage the concerns of fellow providers.
Let’s start with what’s already available in the literature. The NINDS tPA study was the first RCT to succeed in demonstrating a significant improvement in functional outcome (ordinal shift of modified Rankin Scale, among other metrics) in stroke patients treated with tPA within 3 hours. Other RCTs at the time failed because of usage of other thrombolytics and/or a more expanded timeline. Some of these trials, such as ECASS-1 and ATLANTIS, demonstrated a positive outcome consistent with NINDS when the patients treated in the 0-3 hour window were pooled and reanalyzed. ECASS-3 extended the time window for treatment with tPA by demonstrating a sustained but diminishing benefit out to 4.5 hours. Multiple prospective, open-label registries and associated meta-analyses have externally validated these findings over the past two decades.
The two dreaded complications of tPA are hemorrhage and death. The consistent functional benefit of tPA seen in these trials was dampened — but not offset — by a statistically significant increase in the rate of symptomatic intracranial hemorrhage (sICH). Proportions of patients with sICH vary in these trials, partially because the definitions of sICH were not uniform throughout. Most of these trials had limited follow-up time (usually 90 days), so it has remained unclear if tPA has any effect, positive or negative, on long-term mortality. At 1 year, the NINDS tPA study assessed mortality without finding any significant difference between the two groups. IST-3 found a small reduction in 3-year mortality in the treatment arm, which was non-significant.
Muruet et al. present only the second propensity score-matched observational study on the long-term outcomes of IV tPA. The first, a Danish study, found a 34% decrease in mortality in tPA treated patients compared to controls at a median follow-up of 1.4 years. Muruet et al. have used a propensity score-matched cohort study design to determine whether thrombolysis with IV tPA improves long-term survival up to 10 years (median 5.4 years) after an acute ischemic stroke. They chose a propensity score method because this reduces the effect of confounding variables and allows the authors to replicate the measures of effect commonly reported in randomized clinical trials.
2052 patients were recruited in the South London Stroke Register. 334 had missing data, leaving 1718 remaining. 246 tPA treated (0-4.5 hours) patients were paired with 492 controls. The standard difference of means of propensity scores between groups was 0.14, and the ratio of variances of propensity scores was 1.27. Both of these measures indicated that the samples were adequately matched.
Figure. Survival curves for intravenous alteplase treated group (darker) and control group (lighter) groups. Median follow-up time 5.45 years. Median survival for treated group 5.72 years. Median survival for control group 4.98 years. Stratified log-rank test: P<0.001.
Their results are clear. The Kaplan–Meier estimate shows a higher survival for patients treated with IV tPA than for those in the control group. The median survival time for the treated group was 5.72, compared to 4.98 years for the control group. The absolute risk reduction at 5 years was 8.33% (95% CI, 8.19–8.47; NNT=12) and 5.07% (95% CI, 4.92–5.22) at 10 years (NNT=20).
After adjusting for multiple variables, treatment with IV tPA was associated with a 28% decrease in mortality at 5 years and a 37% decrease in mortality at 10 years. Those treated earlier (within the first 3 hours) had even better results: 32% mortality reduction at 5 years and 42% at 10 years.
Muruet et al. have found evidence of reduced long-term mortality after thrombolysis. While it would be unethical to study this in the setting of an RCT, they have used a propensity-score method to attempt to recreate an RCT design in a real-world stroke registry, suggesting an external validation to their findings. In doing this study, they have also addressed some of the issues leading to the controversy over tPA. First of all, mortality is not a subjective outcome. Second, buried in the text is an interesting finding: The mortality benefit seemed to be driven by older patients with NIHSS>16, the population believed by some to be under-represented in the NINDS tPA study. As this is a registry of patients pre-2015, it is unlikely that this effect can be explained away by thrombectomy (although exact numbers of patients receiving thrombectomy — if any — is not stated).
Armed with this information, clinicians can comfortably tell their patients and fellow providers that despite the risk of ICH, patients treated with tPA will be more likely to live a longer, more functional life than those not treated with tPA. What more can you ask for? I believe Muruet et al. have put the nail in the coffin regarding the tPA controversy. There is no argument for clinical equipoise on the matter of tPA and stroke.
