Prédiction du pronostic fonctionnel de l’infarctus cérébral traité par thrombolyse intraveineuse

Abstract : Intravenous thrombolysis (IVT) is the only licensed drug for acute ischemic stroke (AIS). However, about half of the treated patients do not achieve functional independence at 3 months (mRS>2), mostly due to lack of early recanalization or symptomatic intracranial hemorrhage (sICH). Firstly, we aimed to determine if 3-month outcome (mRS) after IVT can be reliably predicted based on clinical and MRI variables available at admission. Secondly, we assessed the relationships between the clinical course within 24 hours after IVT and 3-month mRS. We collected clinical and MRI data of all patients treated by IVT ≤4.5 hrs for AIS between 2003 and 2015 in Sainte-Anne hospital, Paris. (1) We derived the MRI-DRAGON score, a simple tool consisting of 7 clinical and MRI variables available at admission, which can reliably predict 3-month mRS>2 (c-statistic=0.83 [0.78-0.88]). (2) We then performed an external validation of this score in the Lille cohort, showing good discrimination and calibration of the model, despite an overestimation of the risk of mRS>2 in patients with a high MRI-DRAGON score. (3) Trying to find additional predictors of long-term outcome, we showed that the cerebral microbleed (CMB) burden at baseline was not an independent predictor of 3-month mRS after adjusting for confounding factors (age and hypertension).Furthermore, we assessed the relationships between early clinical course after IVT and 3-month mRS, based on two common clinical events: (4) Firstly, the lack of very early neurological improvement (VENI) 1 hour after IVT, which was observed in 77% patients and strongly associated with 3-month mRS, but did not improve the predictive ability of the model when incorporated into the MRI-DRAGON score. (5) Secondly, early neurological deterioration (END) within 24 hours after IVT, occuring in 14% patients in our systematic review and meta-analysis. (6) In our cohort, the positive predictive value of END for 3-month mRS>2 prediction was 90%. END of undetermined cause (ENDunexplained) accounted for 70% of ENDs, and was associated with no prior use of antiplatelets, proximal artery occlusion, DWI-PWI mismatch volume and lack of recanalization. We proposed a simple score to predict 3-month mRS soon after admission in patients treated by IVT for AIS. It may be used to help therapeutic decisions, by identifying patients likely to achieve 3-month mRS ≤2 after IVT alone. We have also shown that CMB burden before IVT is not an independent predictor or 3-month outcome. We participate in an ongoing international individual patient data meta-analysis to determine whether there is a subgroup of patients with CMBs, which seems to have an independent risk of poor 3-month outcome so important that it might outweigh the expected benefit of IVT. Although lack of VENI 1 hour after IVT is strongly associated with 3-month mRS>2, it doesn’t seem to be specific enough to guide decision-making regarding additional thrombectomy (bridging therapy), and should therefore not delay an endovascular procedure. Finally, our results suggest that a persistent cerebral hypoperfusion contributes to most ENDs. Therefore, many ENDs might be avoided in a near future, given the recent proof of the clear superiority of bridging therapy over IVT alone regarding recanalization. This revolution in acute stroke management leads the way to important clinical research perspectives, such as developing a tool to accurately predict 3-month mRS after bridging therapy. Important research efforts will be needed to develop a personalized treatment algorithm, helping to determine which therapeutic option (bridging therapy, IVT alone, thrombectomy alone, or no recanalization therapy) would be the best for each patient.
Document type :
Complete list of metadatas

Cited literature [181 references]  Display  Hide  Download
Contributor : Abes Star <>
Submitted on : Tuesday, December 6, 2016 - 11:20:11 AM
Last modification on : Saturday, June 15, 2019 - 3:26:16 AM
Long-term archiving on : Thursday, March 23, 2017 - 12:50:19 AM


Version validated by the jury (STAR)


  • HAL Id : tel-01409765, version 1



Guillaume Turc. Prédiction du pronostic fonctionnel de l’infarctus cérébral traité par thrombolyse intraveineuse. Neurosciences [q-bio.NC]. Université Sorbonne Paris Cité, 2015. Français. ⟨NNT : 2015USPCB078⟩. ⟨tel-01409765⟩



Record views


Files downloads