IGFBP7 and GDF-15, but not P1NP, are associated with cardiac alterations and 10-year outcome in an elderly community-based study

Table 5 Results of Cox proportional hazard regression models

	N	All-cause mortality		CV mortality		Hospitalization		CV hospitalization
	N	HR[95CI]	p	HR[95CI]	p	HR[95CI]	p	HR[95CI]	p
IGFBP7
Univariate	1634	17.17 [11.31–26.05]	1.1 × 10⁻⁴⁰	31.7 [15.06–66.87]	9.8 × 10⁻²⁰	2.31 [1.66–3.21]	7.5 × 10⁻⁷	4.48 [2.89–6.96]	2.5 × 10⁻¹¹
Multivariable	1486	5.20 [2.79–9.68]	2.0 × 10⁻⁷	7.73 [2.28–26.26]	0.001	1.50 [0.99–2.28]	0.057	2.26 [1.27–4.01]	0.006
Multivariable + echo	809	4.25 [1.81–9.99]	0.001	2.52 [0.47–13.39]	0.279	1.11 [0.61–2.01]	0.733	1.65 [0.76–3.57]	0.204
Multivariable + biomarkers	1486	2.13 [1.08–4.22]	0.029	1.97 [0.51–7.66]	0.329	0.94 [0.60–1.48]	0.794	1.04 [0.55–1.94]	0.911
GDF-15
Univariate	1705	3.44 [2.95–4.02]	4.0 × 10⁻⁵⁶	3.96 [2.92–5.36]	6.0 × 10⁻¹⁹	1.60 [1.42–1.80]	3.0 × 10⁻¹⁴	1.91 [1.63–2.24]	2.3 × 10⁻¹⁵
Multivariable	1621	2.58 [2.09–3.20]	2.5 × 10⁻¹⁸	2.84 [1.82–4.43]	4.0 × 10⁻⁶	1.32 [1.13–1.53]	3.3 × 10⁻⁴	1.50 [1.21–1.84]	1.2 × 10⁻⁴
Multivariable + echo	843	2.50 [1.82–3.44]	1.9 × 10⁻⁸	1.89 [0.96–3.73]	0.067	1.42 [1.15–1.77]	0.001	1.44 [1.09–1.92]	0.011
Multivariable + biomarkers	1621	2.03 [1.62–2.56]	1.5 × 10⁻⁹	1.75 [1.06–2.88]	0.028	1.17 [0.99–1.37]	0.059	1.22 [0.98–1.53]	0.079

Cox proportional hazard models for all-cause mortality or hospital admission for cardiovascular reason. HR [95%CI] for Ln-transformed concentration of the biomarker
For IGFBP7 the multivariable regression analysis was corrected for age, sex, eGFR, diabetes, smoking, dyslipidaemia, MI, AF, heart failure as identified in Table 1
For GDF-15 the multivariable regression analysis was corrected for age, sex, eGFR, diabetes, smoking, dyslipidaemia, angina, MI, AF and COPD, as identified in Table 2
Echo: included MFS and enlarged left atrial area in the regression analyses
Biomarkers: included log-transformed hsTnT and log-transformed NTproBNP in the regression analysis

ISSN: 1471-2261