Prognostic Value of High-Sensitivity Cardiac Troponin I in Patients with Non-Ischaemic Heart Failure: Insights from China

Figure 1. Characteristics of baseline hs‐cTnI. (A) Distribution of hs‐cTnI cross study population. (B) Density plot and maximally selected rank statistics for best cut‐off of hs‐cTnI for all‐cause mortality prediction. (C) Correlation heatmap between hs‐cTnI and baseline variables. (A) The limit of detection (LoD) of hs‐cTnI was 0.002 ng/mL, the cut‐off value for outcome prediction of hs‐cTnI was 0.011 ng/mL, and the 99th upper level of hs‐cTnI was 0.034 ng/mL. (B) The optimal cut‐off of hs‐cTnI optimizing for all‐cause mortality was 0.011 ng/mL. (C) Spearman correlation showed that hs‐cTnI was mild positively correlated with NT‐proBNP, NYHA III/IV, big‐ET, RDW‐SD, RDW, WBC, NLR, hs‐CRP, LAD, and LVEDDi and was negatively correlated with sex, PLR, TP, sodium, eGFR, and LVEF. No correlation existed between hs‐cTnI and VD/DM. big‐ET, big endothelin; DM, diabetes mellitus; hs‐CRP, high‐sensitivity C‐reactive protein; LVEDDi, left ventricular end‐diastolic diameter index; LVEF, left ventricular ejection fraction; NLR, neutrophil‐to‐lymphocyte ratio; PLR, platelet‐to‐lymphocyte ratio; RDW, red blood cell distribution; RDW‐SD, red blood cell distribution width standard deviation; RVEDD, right…

Figure 2. Kaplan–Meier analysis for all‐cause mortality during 3‐year follow‐up. (A) Kaplan–Meier curve of different hs‐cTnI levels. (B,C) Kaplan–Meier curve of different NT‐proBNP levels in patients with hs‐cTnI ≤0.011 ng/mL (B) and hs‐cTnI >0.011 ng/mL (C). (D,E) Kaplan–Meier curve of different hs‐CRP levels in patients with hs‐cTnI ≤0.011 ng/mL (D) and hs‐cTnI >0.011 ng/mL (E). (F,G) Kaplan–Meier curve of valvular heart disease co‐morbidity in patients with hs‐cTnI ≤0.011 ng/mL (F) and hs‐cTnI >0.011 ng/mL (G). As shown in the figure, the risk of all‐cause mortality was significantly higher in patients with hs‐cTnI >0.011 ng/ml over a 3‐year follow‐up time (A,B). Notable discrimination could also be observed in patients with NT‐proBNP >1738 pg/mL than those not across different hs‐cTnI levels (B,C). No discrimination on survival probability was obtained across the different hs‐cTnI level in patients with/without hs‐CRP > 3 mg/mL (D,E) or concomitant with/without valvular heart disease (F,G) during the 3‐year follow‐up time. hs‐CRP, high‐sensitivity C‐reactive protein; VD, valvular heart disease.

Figure 3. Decision curve analysis for the hs‐cTnI‐incorporated model during 3‐year follow‐up. Decision curve analysis for the hs‐cTnI‐incorporated model (A) and its comparison with the well‐established model (B). The hs‐cTnI incorporated model showed a higher overall net benefit compared with the well‐established model across a threshold probability of 12.5–62.5%. Model 1 (well‐established model): age, sex, NYHA class, LVEF, Na, eGFR, diabetes mellitus history, usage of ACEI/ARB, usage of β‐blockers, Hb, NT‐proBNP. Model 2 (hs‐cTnI‐incorporated model): Model 1 + hs‐cTnI. ACEI angiotensin‐converting enzyme inhibitor; ARB, angiotensin II receptor blocker; Hb, haemoglobin; hs‐CRP, high‐sensitivity C‐reaction protein; LVEF, left ventricular ejection fraction; NYHA, New York Heart Association; RDW‐SD, red blood cell distribution width standard deviation.

Figure 4. hs‐cTnI for Prediction of all‐cause mortality: subgroup analysis. The prognostic impact of high hs‐cTnI seemed to be more notable in NIHF patients concomitant with diabetes mellitus. eGFR, estimated glomerular filtration rate; hs‐CRP, high‐sensitivity C‐reaction protein; LVEF, left ventricular ejection fraction; NYHA, New York Heart Association.