The primary hypothesis is that, in patients with advanced heart failure, an outpatient fluid management strategy guided by BNP levels and clinical targets will lead to fewer days hospitalized or dead over a 3-month period compared to an outpatient fluid management strategy using clinical targets alone.

STARBRITE will test the hypothesis that a defined fluid management strategy, tailored to specific symptoms and physiological targets, will improve morbidity and mortality in the advanced heart failure population. Individual targets for each patient will be based on the outcome of the index hospitalization, during which therapy is adjusted to optimize clinical status, blood pressure, and renal function. Identifying a standard fluid management strategy may be an important way to limit the complications of diuretic therapy and the duration of hospitalization for these patients. In addition, a standard approach may provide objective criteria that can be used to systematically deliver care in many types of clinical settings.

In this study, patients will be randomized to two strategies of outpatient fluid management: 1) the Congestion Score Strategy and, 2) the BNP Strategy.

Inclusion Criteria:

1. Patients admitted with NYHA class III/IV symptoms of heart failure
2. Left ventricular ejection fraction ≤35%
3. Will have received at least one dose of intravenous diuretics during the current hospitalization
4. Prior to discharge, has received heart failure education and has learned about daily weights, and fluid, sodium, and dietary restrictions
5. Able to return for follow-up care at the Brigham and Women's Hospital Cardiomyopathy Clinic or the Duke Heart Failure Management Program
6. Has regular access to a telephone
7. ≥18 years of age

Exclusion Criteria:

1. Moderate to severe valvular stenosis
2. Admitted with acute coronary syndrome
3. Creatinine >3.5
4. Renal failure on hemodialysis
5. Pregnant or lactating