Bacterial Pulmonary Infection in PICU

Overview

The purpose of this study is to develop a scoring system to allow doctors to accurately identify children on a mechanical ventilator who have bacterial pneumonia. Currently this diagnosis is very difficult to make, resulting in the overuse of antibiotics and the promotion of antibiotic-resistant bacteria in the pediatric intensive care unit (ICU). Four ICUs at 3 children's hospitals will participate. Study participants will include 150 children, ages 2 months to 17 years old who require mechanical ventilation, and in whom the bedside health care providers suspect bacterial pneumonia. Bacteria will be studied by sampling lung fluid through the breathing tube less than 12 hours after starting antibiotics, using a procedure known as "non-bronchoscopic-bronchoalveolar lavage (NB-BAL)." Participants may be involved in study related procedures for up to 31 days.

The primary objective of this prospective cohort study is to derive a clinical score for predicting the absence of bacterial pulmonary infection in children requiring mechanical ventilation. This study is in preparation for designing an interventional trial whose aim will be to reduce antibiotic use in the pediatric intensive care unit (ICU) and thereby reduce the incidence of acquisition of antibiotic resistant bacteria. Secondary study objectives are to: define key characteristics of bacterial pulmonary infection in children requiring mechanical ventilation: demographics, microbiology, and predisposing factors; and assess the duration and number of antibiotics used for the treatment of pneumonia in this study population and to further assess the amount of antimicrobial therapy that could potentially be avoided in patients with a low probability of bacterial pneumonia as determined by the scoring system. The study hypothesis is that a non-invasive clinical scoring system can be devised for bacterial pulmonary infection in children requiring mechanical ventilation, to identify a threshold score below which the probability of bacterial pulmonary infection is sufficiently low that the clinician can safely withhold or stop antibiotics. The primary outcome (bacterial pneumonia) will be defined as meeting the following 2 criteria: the presence of greater than 10,000 organisms/mL of one or more pathogenic bacterial species obtained by blind non-bronchoscopic bronchoalveolar lavage (NB-BAL) on Day 0 or the isolation of a pathogenic organism from pleural fluid culture obtained between Day -1 through Day 3, in a patient with a pulmonary infiltrate; and persistence of Day 0 pulmonary infiltrate on Day 3 based on a comparison of results of chest x-rays from Days 0 and 3. The secondary endpoint is the duration and number of antibiotics used, for the treatment of pneumonia in the study population and separately in subgroups defined by the threshold score for bacterial pneumonia. Potential predictive parameters for bacterial pulmonary infection will be calculated at the initiation of antibiotics and 3 days later. These parameters will include factors in the Clinical Pulmonary Infection Score (CPIS) developed for adult patients. Additional potential predictors of bacterial pulmonary infection that may be relevant in children also will be recorded and added to the model in an attempt to improve its predictive accuracy. Study outcome will be microbiologically defined bacterial infection, using quantitative lower respiratory tract specimens collected by blind NB-BAL at entry and persistence of infiltrate on chest x-ray at Day 3. The study population will include children, age greater than 48 weeks post-conception and less than 18 years, requiring mechanical ventilation in the pediatric ICU, presence of an abnormal chest x-ray, and initiation of antibiotics less than or equal to 12 hours prior to NB-BAL by the assigned health care providers for suspected bacterial pulmonary infection.