European Stroke Journal Comment

The quest of a surrogate safety outcome for carotid revascularisation to boost new trials: a new hope?

Author: Giuseppe Reale, MD, Institute of Neurology, Università Cattolica del Sacro Cuore – Rome. ESO-YSPR Committee.

Original Article: Silent brain infarcts on diffusion-weighted imaging after carotid revascularisation: a surrogate outcome measure for procedural stroke? A systematic review and meta-analysis (Traenka et al, 2018)

ESJ Comment

The quest of a surrogate safety outcome for carotid revascularisation to boost new trials: a new hope?

The enrolment of patients in the most important randomized controlled trials (RCT) of carotid endarterectomy (CEA) started more than 30 years ago. In the meantime, the development of carotid artery stenting (CAS) has changed the clinical practice, being the procedure as effective as CEA in preventing stroke recurrence or TIA after the post-procedural period1,2,3.

The main safety outcome of the aforementioned revascularization techniques is procedural stroke (clinical diagnosis of stroke based on WHO definition within 30 days from the procedure) and long-term data seem to favour CEA over CAS (5.0% vs. 8.2% of procedural stroke risk, respectively). It is clear that the procedural risk is not marginal and many efforts have been made to achieve new surgical and interventional techniques that improve the procedural safety of CEA and CAS4. In order to establish the safety of new techniques, there is an urgent necessity of new randomized controlled trials (RCT), but the necessity of a large sample size tailored on the frequency of procedural stroke would slow-down the process.

This is why the Authors of a Paper published today in the European Stroke Journal investigated whether the presence of lesions on diffusion-weighted imaging (DWI+) after CAS or CEA might provide a good surrogate outcome measure for procedural stroke. The assumption is that the higher frequency of DWI+ might reduce the sample size needed for future RCT, boosting the collection of new evidence on the safety of new CEA/CAS protocols.

With this in mind, the Authors performed a systematic review of the literature, including studies with symptomatic or asymptomatic patients, any degree of stenosis, who underwent CEA or CAS and had a pre- and post-procedural MRI with DWI sequences.

About a half of the 4871 reported CAS were done in symptomatic carotid stenosis and DWI+ was reported in 37% of CAS, while procedural ischemic stroke occurred in 2.6%. About 65% of the 2099 CEA were performed in patients with symptomatic carotid stenosis and DWI+ were detected in 10.8% of cases, while procedural ischemic stroke complicated 1.4% of CEA.

Then the Authors had to prove that DWI+ satisfies the criteria for surrogate outcomes postulated in the “ICH Harmonised Tripartite Guideline: Statistical Principals for Clinical Trials, the Authors”: (i) the biological plausibility of the relationship, (ii) the demonstration of the prognostic value of the surrogate for the clinical outcome and (iii) evidence that treatment effects on the surrogate correspond to effects on the clinical outcome.

About the first criterion, the reliability of MRI DWI sequences in detecting ischemic lesions due to cellular hypoxia is nowadays out of question.

To explore the second criterion, the Authors found a significant correlation between DWI+ and procedural stroke (CAS or CEA), both in a crude correlation model and -more strongly- in a bivariate random effects logistic regression model, that takes into account the heterogeneity of the subjects included in the studies.

Because of the variability of sample sizes of the available studies and the width of confidence intervals, a regression analysis was not feasible in order to explore correlation between treatment effect on DWI+ and procedural stroke. In this view, the Authors performed a metanalysis of nineteen studies (including 2 RCT) to assess whether the treatment effect for DWI+ had the same trend of the treatment effect for procedural stroke. In particular, they demonstrated that the comparison between CAS and CEA using DWI+ as the outcome favoured the same treatment (CEA) as the comparison using stroke as the outcome.

Although a rigorous and elegant statistical analysis, the study did not fulfil the aforementioned surrogate variable criteria, but it still supports the use of DWI+ as a surrogate outcome measure for procedural stroke in carotid revascularisation procedures. Future randomized trials comparing treatment effects on DWI+ and procedural stroke will finally assess its surrogacy. If fully demonstrated, this surrogate outcome might contribute to a 10-time decrease of the sample size needed for future trials on CEA and CAS safety. Moreover, the rate of procedural stroke reported in this article is lower than that described in previous reviews: this is probably due to the new advances in surgical and neurointerventional procedures, but only further RCT will confirm this hypothesis.


  1. North American Symptomatic Carotid Endarterectomy Trial Collaborators. Beneficial effect of carotid endarterectomy in symptomatic patients with high-grade carotid stenosis. N Engl J Med 1991;325: 445–453.
  2. Randomised trial of endarterectomy for recently symptomatic carotid stenosis: final results of the MRC European Carotid Surgery Trial (ECST). Lancet 1998; 351:1379–1387
  3. Bonati LH, Lyrer P, Ederle J, et al. Percutaneous trans- luminal balloon angioplasty and stenting for carotid artery stenosis. Cochrane Database Syst Rev 2012; 9:CD000515.
  4. Bonati LH, Dobson J, Algra A, et al.; Carotid Stenting Trialists’ Collaboration. Short-term outcome after stenting versus endarterectomy for symptomatic carotid stenosis: a preplanned meta-analysis of individual patient data. Lancet 2010;376: 1062–1073.