Prognostic Value of Combined Measurement of Brain Natriuretic Peptide and Triiodothyronine in Heart Failure

Abstract

Background Both low free triiodothyronine (fT3) and high brain natriuretic peptide (BNP) have been separately described as prognostic predictors for mortality in heart failure (HF). We investigated whether their prognostic value is independent. Methods and Results From January of 2001 to December of 2006, we prospectively evaluated 442 consecutive patients with systolic HF and no thyroid disease or treatment with drugs affecting thyroid function (age 65 ± 12 years, mean ± standard deviation, 75% were male, left ventricular ejection fraction 33% ± 10%, New York Heart Association (NYHA) class I and II: 63%, NYHA class III and IV: 37%). All patients underwent full clinical and echocardiographic evaluation and assessment of BNP and thyroid function. Both cardiac and all-cause mortality (cumulative) were considered as end points. During a median 36-month follow-up (range 1–86 months), 110 patients (24.8%) died, 64 (14.4%) of cardiac causes. Univariate Cox model predictors of all-cause mortality and cardiac death were age, body mass index, creatinine, hemoglobin, ejection fraction, NYHA class, BNP, fT3, and thyroxine level. Multivariate analysis selected age, NYHA class, hemoglobin, BNP, and fT3 as independent predictors for all-cause mortality and NYHA class, BNP, and fT3 as independent predictors for cardiac mortality. Patients with low fT3 and higher BNP showed the highest risk of all-cause and cardiac death (odds ratio 11.6, confidence interval, 5.8–22.9; odds ratio 13.8, confidence interval, 5.4–35.2, respectively, compared with patients with normal fT3 and low BNP). Conclusion fT3 and BNP hold an independent and additive prognostic value in HF.