Brain natriuretic peptide predicts serious cardiac allograft rejection independent of hemodynamic measurements

Serum brain natriuretic peptide (BNP) has been reported to be elevated in heart transplant recipients with both cellular and vascular rejection. Whether BNP can be used to help predict the severity of rejection is not well established. We analyzed serial BNP measurements obtained during endomyocardial biopsy procedures in consecutive heart transplant patients occurring >45 days after transplantation. To eliminate potential confounding from prior rejection episodes, we included only observations in which the previous biopsy grade was 0 or 1A. Multivariable linear regression was performed examining the outcome of increasing seriousness of rejection, defined as grade 0 < 1A < 2 < 1B < 3A < vascular rejection. A univariable logistic regression model was performed using log-transformed BNP as a predictor of vascular rejection. There were 77 patients, with 161 separate observations. Median time between transplantation and first assessment was 6.0 months (interquartile range, 2.1, 31.6). Using multivariable linear regression, 3 factors were significantly associated with biopsy score: pulmonary capillary wedge pressure ( p < 0.0001), BNP ( p = 0.003), and heart rate ( p = 0.01). Even after other significant univariable predictors (including pulmonary capillary wedge pressure) were forced into the model, BNP remained a significant predictor of biopsy score ( p = 0.02). Log BNP was a significant univariable predictor of vascular rejection, with an odds ratio of 12.55 (per 1 unit increase, 95% confidence interval 3.43–45.84; p = 0.0001) and a model c-statistic of 0.91. BNP predicts new episodes of serious cardiac allograft rejection, particularly vascular rejection, independent of hemodynamic measurements, and may be a useful part of rejection surveillance.