Advertisement
Canadian Journal of Cardiology

N-terminal Pro B-Type Natriuretic Peptide, High-Sensitivity Cardiac Troponin T, and Hibernating Myocardium in Patients With Ischemic Heart Failure

Published:January 12, 2018DOI:https://doi.org/10.1016/j.cjca.2018.01.005
      To the Editor:
      Zelt et al. examined the relationship between N-terminal pro-B-type natriuretic peptide (NT-proBNP), high-sensitivity cardiac troponin T (hs-cTnT), and the extent of hibernating myocardium and scar tissue in 39 patients with ischemic left ventricular dysfunction and heart failure.
      • Zelt J.G.
      • Liu P.P.
      • Erthal F.
      • et al.
      N-terminal pro B-type natriuretic peptide and high-sensitivity cardiac troponin T levels are related to the extent of hibernating myocardium in patients with ischemic heart failure.
      For the prediction of hibernation, the area under the receiver operating characteristic curve for NT-proBNP and hs-cTnT was 0.76 and 0.78, respectively. The adjusted odds ratio (95% confidence intervals) of Log (NT-proBNP) and Log (hs-cTnT) for > 10% hibernation was 8.83 (0.15-515.59) and 8.57 (0.90-82), respectively. I have some concerns about the study.
      First, the authors concluded that serum NT-proBNP and hs-cTnT were associated with the presence and extent of hibernating myocardium, especially in patients with significant scar tissue. They adjusted for ejection fraction, age, estimated glomerular filtration rate, and reported 7 hibernation events. The low number of events makes it difficult to establish stable estimates in logistic regression analysis.
      • Austin P.C.
      • Steyerberg E.W.
      Events per variable (EPV) and the relative performance of different strategies for estimating the out-of-sample validity of logistic regression models.
      In addition, the coefficient of determination in the multivariate regression analysis, which is the adjusted R2 value, was not presented for the prediction of hibernation. Furthermore, multiple regression analysis requires ≥ 10 subjects per variable for stable estimation.
      • Marill K.A.
      Advanced statistics: linear regression, part II: multiple linear regression.
      Taken together, further study in which an adequate enough number of events can be analyzed is needed to assess the association between NT-proBNP, hs-cTnT, and the extent of hibernating myocardium.
      Additionally, receiver operating characteristic curve analysis is a univariate procedure. Although the area under the curve by NT-proBNP and hs-cTnT produced moderate accuracy, a significant value does not guarantee the ability to predict the hibernation of myocardium.
      Finally, the authors did not simultaneously include NT-proBNP and hs-cTnT in their multivariate analyses. Because the presence of these biological markers might have multiple meanings, the combination of NT-proBNP and hs-cTnT would be useful in the analysis for predicting hibernation. There is a critical report on how to use logistic regression analysis for prediction
      • Meurer W.J.
      • Tolles J.
      Logistic regression diagnostics: understanding how well a model predicts outcomes.
      ; the application of the statistical procedures used in this study should be addressed to allow the authors to be able to specify the association.

      Disclosures

      The author has no conflicts of interest to disclose.

      References

        • Zelt J.G.
        • Liu P.P.
        • Erthal F.
        • et al.
        N-terminal pro B-type natriuretic peptide and high-sensitivity cardiac troponin T levels are related to the extent of hibernating myocardium in patients with ischemic heart failure.
        Can J Cardiol. 2017; 33: 1478-1488
        • Austin P.C.
        • Steyerberg E.W.
        Events per variable (EPV) and the relative performance of different strategies for estimating the out-of-sample validity of logistic regression models.
        Stat Methods Med Res. 2017; 26: 796-808
        • Marill K.A.
        Advanced statistics: linear regression, part II: multiple linear regression.
        Acad Emerg Med. 2004; 11: 94-102
        • Meurer W.J.
        • Tolles J.
        Logistic regression diagnostics: understanding how well a model predicts outcomes.
        JAMA. 2017; 317: 1068-1069

      Linked Article