Predicting response to carvedilol for the treatment of heart failure: a multivariate retrospective analysis.

BACKGROUND

Carvedilol has been shown to decrease the progression of heart failure and improve left ventricular function and survival in patients with a left ventricular ejection fraction (LVEF) less than 35%. However, not all patients respond uniformly to this therapy. We proposed to identify variables that could, potentially, be used to predict response to carvedilol therapy as measured by the change in LVEF after treatment (Delta LVEF), and to identify pretreatment variables associated with hospitalization for heart failure after carvedilol therapy.

METHODS AND RESULTS

A retrospective analysis of 98 patients treated with open-label carvedilol for a mean period of 16 months was performed by using bivariate and step-wise multivariate analyses. Bivariate analysis showed a positive correlation of Delta LVEF with heart rate at baseline (P =.001). There was a negative correlation of Delta LVEF with baseline LVEF (P <.01), diabetes mellitus (P =.04), and ischemic cardiomyopathy (P =.0002). Multivariate analysis showed a positive correlation of Delta LVEF with heart rate at baseline (P =.01) and a negative correlation with initial LVEF (P =.02) and ischemic cardiomyopathy (P =.006). Variables associated with hospitalization after initiation of carvedilol therapy were New York Heart Association (NYHA) classification (P =.001), lower extremity edema (P =.001), presence of an S3 (P =.02), hyponatremia (P =.02), elevated blood urea nitrogen (BUN) (P =.002), atrial fibrillation (P =.001), diabetes mellitus (P =.02), and obstructive sleep apnea (P =.009).

CONCLUSIONS

Heart failure patients with the lowest LVEF or the highest heart rate at baseline had the greatest gain in LVEF after treatment with carvedilol. Patients with ischemic cardiomyopathy derived less benefit. Patients with clinical evidence of decompensated heart failure were at greater risk for hospitalization after initiation of carvedilol therapy.