Modification of the THRIVE Score by Endovascular Stroke Treatment: The THRIVE-c Calculation in VISTA Endovascular

Dr Alexander C Flint
Dr Geoffrey Donnan
Dr Steve Davis
Dr Philip Bath
Dr Vivek A Rao
Dr Sheila L Chan
Dr Sean P Cullen
Dr Jonathan Hartman
Dr Werner Hacke

The THRIVE Score and related THRIVE-c calculation are widely used tools for prognostication after acute ischemic stroke, factoring in patient age, presentation NIH Stroke Scale score, and comorbidities (hypertension, diabetes mellitus, and atrial fibrillation). After our group initially developed the THRIVE Score, we performed further external validations in VISTA Acute and SITS-MOST. We later developed the THRIVE-c calculation using the VISTA Acute and SITS-MOST data sources in order to include individual patient continuous predictors in our estimation of 90-day post-stroke outcome, and we showed that the THRIVE-c approach is superior to the parent THRIVE score. We have also shown that the relationship between THRIVE and outcomes is statistically independent from the relationship between recanalization therapy (tPA or endovascular stroke treatment [EST]) and outcomes. We now propose to use data from VISTA Endovascular to determine how EST and degree of EST success as measured by the end-of-case TICI score modify the pre-intervention THRIVE-c calculation in the prediction of long-term outcome.

The RCTs in VISTA Endovascular have established that EST improves long-term outcome after large artery occlusion ischemic stroke. We will use both Bayesian and traditional statistical methods to quantify the impact of EST on outcomes in the context of the THRIVE-c calculation, with assigns a pre-intervention probability of good outcome after acute stroke.

To perform the analysis, we will use the following variables from VISTA Endovascular:

age (de-identified by transformation of 90+ into the mean for 90+)
initial NIHSS
last NIHSS during hospitalization
hypertension (0/1)
diabetes mellitus (0/1)
atrial fibrillation (0/1)
female gender (0/1)
mRS at 90 days
Death by 90 days
IV tPA administration (0/1)
time from onset to IV tPA administration
TICI (or mTICI) score { for EST arm }
recanalization (0/1) - { for EST arm; TICI 2b/3 or equiv }
randomized to endovascular stroke treatment
time from onset to groin puncture { for EST arm }
trial name
device name { for EST arm }
intra-arterial tPA used { for EST arm; 0 or # mg tPA IA administered }

For Bayesian analysis we will calculate THRIVE-c to determine the pre-intervention probability of good outcome at 90 days, then determine the likelihood ratio for EST or mechanically successful EST (e.g. TICI 2b/3) in order to determine how EST modifies the THRIVE-c pre-interevention probability into a post-intervention probability of good outcome.
For traditional analyses we will build multivariable logistic regression models of good outcome and ordinal logistic regression models of mRS and quantify the improvement of such models obtained by adding EST (or EST mechanical success) to the models, as measured by standard metrics of model fit and by the area under the curve (AUC) for the received-operator characteristics (ROC) curve for the models, an approach we have used extensively in our prior THRIVE work using VISTA Acute and SITS-MOST data.

Internal funding arranged from the PI\'s institution (Kaiser Permanente Northern California).

Flint AC, Rao VA, Chan SL, Cullen SP, Faigeles BS, Smith WS, Bath PM, Wahlgren N, Ahmed N, Donnan GA, and Johnston SC. (2015) Improved ischemic stroke outcome prediction using model estimation of outcome probability: the THRIVE-c calculation. International Journal of Stroke, 10: 815.

Flint AC, Cullen SP, Rao VA, Faigeles BS, Pereira VM, Levy EI, Jovin TG, Liebeskind DS, Nogueira RG, Jahan R, Saver JL on behalf of the SWIFT and STAR Trialists. (2014) The THRIVE score strongly predicts outcomes in patients treated with the Solitaire device in the SWIFT and STAR trials. International Journal of Stroke, 9: 698.

Flint AC, Gupta R, Smith WS, Kamel K, Faigeles BS, Cullen SP, Rao VA, Bath PM, Wahlgren N, Ahmed N, Donnan GA on behalf of the SITS International and VISTA-plus Investigators. (2014) The THRIVE score predicts symptomatic intracerebral hemorrhage after intravenous tPA administration in SITS-MOST. International Journal of Stroke, 9: 705.

Flint AC, Faigeles BS, Cullen SC, Kamel H, Rao V, Gupta R, Smith WS, Bath PM, and Donnan GA, on behalf of the VISTA Collaboration. (2013) The THRIVE Score predicts ischemic stroke outcomes and thrombolytic hemorrhage risk in the Virtual International Stroke Trials Archive (VISTA) (2013). Stroke, 44: 3365.

Flint AC, Gupta R, Nogueira RG, Lutsep HL, Jovin TG, Albers GW, Liebeskind DS, Sanossian N, and Smith WS, for the TREVO-2 Trialists. (2013) The THRIVE Score predicts outcomes with a third-generation endovascular stroke treatment device in the TREVO-2 trial. Stroke, 44: 3370.

Kamel H, Patel N, Rao VA, Cullen SP, Faigeles BS, Smith WS, and Flint AC. (2013) The THRIVE score predicts ischemic stroke outcomes independent of thrombolytic therapy in the NINDS tPA trial. Journal of Stroke and Cerebrovascular Diseases, 22(7): 1111.

Flint AC, Kamel H, Rao VA, Cullen SP, Faigeles BS, and Smith WS. (2014) Validation of the Totaled Health Risks in Vascular Events (THRIVE) score for outcome prediction in endovascular stroke treatment. International Journal of Stroke, 9(1): 32.

THRIVE score / THRIVE-c website: www.thrivescore.org

THRIVE score and THRIVE-c calculation on mdcalc: https://www.mdcalc.com/thrive-score-stroke-outcome