Prognostic value of D-dimer in acute heart failure
© BioMed Central Ltd 2005
Published: 9 June 2005
Several factors associated with the pathophysiology of heart failure (HF) contributed to the occurrence of thrombo-embolic events, such as vascular disease, hypercoagulability and venous stasis. Many studies showed elevation of coagulation markers, including D-dimer, in advanced stages of HF. The role of D-dimer is still unknown as a long-term prognostic marker in HF patients.
To evaluate the best value of D-dimer that can predict inhospital death. To determine the prognostic role of D-dimer after 1 year of follow-up in patients with decompensated HF.
Materials and methods
A cohort of 70 patients with decompensated HF (85.7% in NYHA class IV) admitted to a coronary care unit during the year 2003. The D-dimer was measured in 53 patients (77.2 ± 10.2 years old, 54.7% male, 84.9% NYHA class IV) at hospital admission; and it was correlated with inhospital death and event-free survival (1 year of follow-up after baseline hospitalization). We used the ROC curve to establish the best cutoff for sensibility and specificity for inhospital death, followed by the chi-square test; and also the log rank test to analyze the Kaplan–Meier curve. We consider P ≤ 0.05 statistically significant.
The best cutoff point of D-dimer in the ROC curve to predict inhospital death was 1433 mg/dl (P = 0.03), with sensibility = 80%, specificity = 69% and negative predictive value = 97%. After 1 year of follow-up we observed that patients with D-dimer ≥ 2000 mg/dl during initial hospitalization had the worst prognosis (event-free survival median = 295 days when D-dimer <2000 mg/dl vs 70 days when D-dimer ≥ 2000 mg/dl, P = 0.03).
An elevated D-dimer at hospital admission in patients with decompensated HF seems to have clinical importance, indicating a higher probability of inhospital death and worse event-free survival after 1 year.