Abstract
Background: Cardiac troponin T (cTnT) has been reported to be a predictor of stroke outcome. We hypothesized that the high-sensitivity cTnT (hs-cTnT) assay is superior to its previous assay generation for stroke outcome prediction.
Methods: cTnT was measured on emergency department admission in 60 consecutive patients (35 males, age 69.4±13.9 years) with ischemic stroke. Adverse 90-day clinical outcome was defined as death or dependence (modified Rankin scale ≥3 or Barthel index <75).
Results: Stroke aetiology was microangiopathy in three, macroangiopathy in 17, cardiac embolism in 26, dissection in one and unknown in 13 patients. At 90-day follow-up 16 (27%) patients had an adverse outcome. Receiver operating characteristics (ROC) curve analysis yielded a significantly better performance of hs-cTnT vs. cTnT (area under ROC curve: 0.80 [95% CI 0.68–0.89] vs. 0.70 [95% CI 0.57–0.82], p=0.017). The optimal discriminator values according to ROC analysis were 5.1 ng/L (hs-cTnT assay) and 11 ng/L (4th generation cTnT assay) with: sensitivities 94% vs. 56%, specificities 57% vs. 84%, positive predictive values 44% vs. 56%, and negative predictive values 96% vs. 84%.
Conclusions: Improvements in cTnT assay sensitivity and precision at the low measuring range resulted in a more accurate prediction of ischemic stroke outcome, particularly for ruling out worse outcome.
©2012 by Walter de Gruyter Berlin Boston
