We sought to assess the utility of serial BNP measurements in patients with severe heart failure and attempted to correlate values with invasively derived data.

In a retrospective study, we analyzed serial BNP levels in patients receiving hemodynamically guided therapy for severe heart failure and sought correlation with invasively derived data.

Thirty-nine patients with New York Heart Association Class III-IV, with an ejection fraction of 35% or less, who had a pulmonary artery catheter inserted for hemodynamically tailored heart failure therapy, were identified and serial BNP measurements reviewed. BNP was estimated on admission, at 12 and 36 hours. Normally distributed variables are expressed as mean ± SD and otherwise as median ± interquartile range. Mean ejection fraction was 16% ± 6%. Mean pulmonary artery occlusion pressures (PAOP) fell with therapy and were 25 ± 7 mmHg, 18 ± 7 mmHg and 19 ± 7 mmHg at admission, 12 hours and 36 hours respectively (P < 0.05). Median BNP levels fell from 1200 ± 641 to 771 ± 803 at 12 hours and to 805 ± 771 at 36 hours (P < .001). There was no correlation between BNP and any hemodynamically derived variable. A change in BNP was not associated with a change in PAOP in any individual patient. Only 42% remained alive on medical therapy at 30 days.

In patients with severe heart failure, BNP levels do not accurately predict serial hemodynamic changes and do not obviate the need for pulmonary artery catheterization.