Introduction: Acute Heart Failure is frequently decompensated by pulmonary infection, but the diagnosis of pulmonary infection sometimes is difficult in these patients due to similar signals and clinical symptoms in both pathologies. Furthermore, when it is possible the diagnosis of pulmonary infection, physicians may have difficult to determine etiology and delaying antibiotic therapy. Procalcitonin (PCT) have been used like a biomarker to determine the period of use of antibiotics in patients with acute respiratory infections. It is specific for bacterial infections and it have showed as a marker of severity infection and may help to determine interruption period of antibiotic therapy in a safety way for the patient. Aim: Evaluate levels of PCT related to interruption of antibiotics in patients with decompensated acute heart failure (DAHF) with suspected bacterial pulmonary infection. Methods: In this pilot project will be included around 100 patients, randomized in two groups: group A (PCT levels may guide the interruption of antibiotic at day 5) or group B (antibiotic period will be determined by the physician without the knowledge of PCT levels). Will be collected laboratorial and clinical data at days 0,3 and 5. Both groups will be compared to evaluate PCT levels and total period of antibiotic therapy, hospitalization and readmission in 30 days. This study will determine the sensibility/specificity of PCT in patients with DAHF.

It is a single-center study, randomized, controlled, prospective cohort and interventional not blind. DAHF diagnosis will consider symptoms, physical examination, x ray and other image methods and BNP ou NTproBNP levels. Pulmonary infection diagnosis will "Infectious Disease Society of America / American Thoracic Society Consensus Guidelines on the Management of Community-Acquired Pneumonia in Adults" criteria, that include fever, cough, sputum, pleuritic pain, positive pulmonary image and physical examination. PCT and NTproBNP levels will be measured in miniVIDAS® (bioMérieux Clinical Diagnostics, France). Statistics includes descriptive analysis. Parametric data will be evaluated using T test and non parametric data with Mann-Whitney-Wilcoxon test. Multivariative analysis related to diagnosis, mortality and morbidity will be performed by Cox regression model. Sensibility and specificity will be calculated by ROC curve. Statistical significancy will be considered when p<0.05.

Inclusion Criteria:

• DAHF diagnosis and suspected pulmonary infection
• BNP ≥ 500 pg/mL or NTproBNP ≥ 450 pg/mL for patients until 50 years old.
• BNP ≥ 500 pg/mL or NTproBNP ≥ 900 pg/mL for patients with 51 - 75 years old.
• BNP ≥ 500 pg/mL or NTproBNP ≥ 1800 pg/mL for patients older than 75 years.

Exclusion Criteria:

• Antibiotic use (oral or endovenous) in the last 15 days;
• Acute coronary syndrome;
• Creatinine > 3,0 mg/dL or hemodialysis;
• Pregnancy;
• Second suspectec infection;
• Suspected pulmonary thromboembolism;
• cancer;
• myocarditis.