METHODS AND RESULTS: In this prospective single-hospital study, we recruited 980 consecutive post-acute
myocardial infarction patients (718 men, median [range] age 66 [24 to 95] years), with follow-up over 342 (range 0 to 764) days. Plasma copeptin was highest on admission (n=132, P<0.001, day 1 versus days 2 to 5) and reached a plateau at days 3 to 5. In the 980 patients, copeptin (measured at days 3 to 5) was elevated in patients who died (n=101) or were readmitted with
heart failure (n=49) compared with survivors (median [range] 18.5 [0.6 to 441.0] versus 6.5 [0.3 to 267.0] pmol/L, P<0.0005). With logistic regression analysis, copeptin (odds ratio, 4.14, P<0.0005) and
NTproBNP (odds ratio, 2.26, P<0.003) were significant independent predictors of death or
heart failure at 60 days. The area under the receiver operating characteristic curves for copeptin (0.75) and
NTproBNP (0.76) were similar. The logistic model with both markers yielded a larger area under the curve (0.84) than for
NTproBNP (P<0.013) or copeptin (P<0.003) alone, respectively. Cox modeling predicted death or
heart failure with both
biomarkers (log copeptin [hazard ratio, 2.33], log
NTproBNP [hazard ratio, 2.70]). In patients stratified by
NTproBNP (above the median of approximately 900 pmol/L), copeptin above the median (approximately 7 pmol/L) was associated with poorer outcome (P<0.0005). Findings were similar for death and
heart failure as individual end points.
CONCLUSIONS: