B-type natriuretic peptide is one of the most sensitive and specific biohumoral markers of heart failure. We hypothesized that B-type natriuretic peptide changes during treatment of heart failure may provide independent information on disease progression and outcome in patients enrolled in the Val-HeFT trial.

Patients were divided into four groups according to concentrations of B-type natriuretic peptide at baseline versus 4 months (n = 3740) or 12 months (n = 3343), with respect to the baseline median (97 pg/mL): low→low (stable below median, 44%-46%), high→high (stable above median, 32%-37%), high→low (above to below median, 12%-14%), and low→high (below to above median, 6%-9%). Cox multivariate regression analysis was used to assess the risk of death and morbidity, with adjustment for baseline B-type natriuretic peptide concentrations.

Patients who improved their B-type natriuretic peptide at 4 months (high→low) had a similar risk for mortality (hazard ratio = 1.191, 95% confidence interval [CI] 0.870-1.631, P =.2746) compared with the low→low patients. Conversely, patients who worsened in their B-type natriuretic peptide (low→high) had a risk for mortality (hazard ratio 2.578, CI, 1.861-3.571, P <.0001) higher than patients in the low→low group, and indistinguishable from the high→high group. Worsening of B-type natriuretic peptide (low→high) was associated with 0.03 cm/m² increase in left ventricular end-diastolic diameter, whereas it decreased by 0.10 cm/m² in high→low and low→low groups (P <.001).

Changes in B-type natriuretic peptide over time with respect to a threshold value of 97 pg/mL convey an independent and additional prognostic value compared with a single determination of B-type natriuretic peptide in a large population of patients with chronic symptomatic heart failure and might be helpful in the management of these patients.