Table 3

Unadjusted and adjusted effects on mortality

VariablesIQRCrude HR (95% CI)p ValueC-statisticAdj. HR* (95% CI)p ValueC-statisticAdj. HR† (95% CI)p ValueC-statistic
hsTnT, ng/mL0.0071.27 (1.20 to 1.34)<0.0010.651.21 (1.13 to 1.32)<0.0010.80
ln(NT-proBNP), pg/mL1.4651.56 (1.34 to 1.83)<0.0010.631.54 (1.24 to 1.90)<0.0010.801.33 (1.06 to 1.68)0.0130.81
ln(MR-proANP), pmol/L0.8211.45 (1.21 to 1.72)<0.0010.591.40 (1.10 to 1.79)0.0060.801.21 (0.94 to 1.55)0.1420.81
MR-proADM, nmol/L0.2801.23 (1.20 to 1.36)<0.0010.651.31 (1.19 to 1.44)<0.0010.811.24 (1.12 to 1.38)<0.0010.81
CT-proET-1, pmol/L27.8581.27 (1.20 to 1.35)<0.0010.611.21 (1.14 to 1.30)<0.0010.801.19 (1.11 to 1.28)<0.0010.81
ln(copeptin), pmol/L1.0491.49 (1.28 to 1.75)<0.0010.601.22 (1.04 to 1.42)0.0140.791.21 (1.03 to 1.42)0.0210.80
  • Cox proportional hazard models for cardiovascular hormones are shown. HRs refer to a 1-IQR increase in continuous variables. HRs are adjusted (adj.) for the null model built from all variables selected by bootstrapping, that is, for age, tumour entity (tumours of categories 1, 2, 3, 6, 10 and 18) and condition of tumour stage 4 and additionally for GFR and cardiac status. The discriminatory power of the respective variables was assessed using Harrell's c-index.

  • *HR adjusted to age, tumour stage 4, tumour entity, cardiac status and GFR.

  • †HR adjusted to age, tumour stage 4, tumour entity, cardiac status, GFR and hsTnT.

  • Fonts in bold indicate statistical significance (p<0.05).

  • CT-proET-1, C-terminal pro endothelin-1; GFR, glomerular filtration rate; hsTnT, high-sensitive troponin T; MR-proADM, mid-regional pro-adrenomedullin; MR-proANP, mid-regional pro-atrial natriuretic peptide; NT-proBNP, N-terminal pro B-type natriuretic peptide.