Table 4. Diagnostic parameters of hs-cTnT thresholds for inpatient mortality.
Hs-cTnT (ng/L) | Sensitivity | 95% CI | Specificity | 95% CI | PPV | 95% CI | NPV | 95% CI | +LR | 95% CI | -LR | 95% CI |
---|---|---|---|---|---|---|---|---|---|---|---|---|
3 | 100 | 93–100 | 5 | 2–10 | 28 | 21–35 | 100 | 59–100 | 1.05 | 1.0–1.1 | 0 | - |
5 | 100 | 93–100 | 13 | 8–20 | 30 | 23–37 | 100 | 81–100 | 1.15 | 1.1–1.2 | 0 | - |
10 | 94 | 84–99 | 39 | 31–48 | 36 | 28–45 | 95 | 86–99 | 1.55 | 1.3–1.8 | 0.15 | 0.05–0.5 |
14 (Normal) | 92 | 81–98 | 45 | 37–54 | 38 | 29–47 | 94 | 85–98 | 1.68 | 1.4–2.0 | 0.17 | 0.07–0.5 |
20 | 86 | 74–94 | 55 | 46–63 | 41 | 32–51 | 92 | 84–97 | 1.92 | 1.5–2.4 | 0.25 | 0.1–0.5 |
23 (Youden) | 84 | 71–93 | 61 | 53–69 | 44 | 34–55 | 91 | 84–96 | 2.19 | 1.7–2.8 | 0.26 | 0.1–0.5 |
25 | 78 | 65–89 | 64 | 56–72 | 44 | 34–55 | 89 | 81–94 | 2.20 | 1.7–2.9 | 0.34 | 0.2–0.6 |
30 | 69 | 54–81 | 67 | 59–75 | 43 | 32–55 | 86 | 78–92 | 2.09 | 1.5–2.8 | 0.47 | 0.3–0.7 |
35 | 65 | 50–78 | 68 | 59–76 | 42 | 31–54 | 84 | 76–90 | 2.01 | 1.5–2.8 | 0.52 | 0.4–0.8 |
40 | 59 | 44–72 | 72 | 64–79 | 44 | 32–56 | 83 | 75–89 | 2.11 | 1.5–3.0 | 0.57 | 0.4–0.8 |
45 | 57 | 42–71 | 74 | 66–81 | 45 | 32–58 | 83 | 75–89 | 2.21 | 1.5–3.2 | 0.58 | 0.4–0.8 |
50 | 45 | 31–60 | 79 | 71–85 | 43 | 30–58 | 80 | 72–86 | 2.10 | 1.4–3.3 | 0.70 | 0.5–0.9 |
55 | 41 | 28–56 | 79 | 71–85 | 41 | 28–56 | 79 | 71–85 | 1.92 | 1.2–3.0 | 0.75 | 0.6–1.0 |
60 | 41 | 28–56 | 81 | 74–88 | 45 | 30–60 | 79 | 72–86 | 2.22 | 1.4–3.6 | 0.72 | 0.6–0.9 |
CI: Confidence interval; hs-cTnT: High sensitivity cardiac troponin T; +LR: Positive likelihood ratio; -LR: Negative likelihood ratio; NPV: Negative predictive value; PPV: Positive predictive value. Normal denotes the hs-cTnT threshold derived from the 99th percentile of healthy individuals widely used in clinical practice for detecting cardiac injury. Youden denotes the optimal threshold based on the receiver operating characteristics curve shown in Fig 3.