| Abstract |
Introduction
Respiratory failure (RF) is one of the most common causes of hospitalization of children in pediatric intensive care units (PICU) and is associated with a large number of heterogeneous causes and diseases. Children with RF present significant differences in the manifestation of the entity among themselves and compared to adults. Assessment of respiratory function at presentation and during hospitalization includes evaluation of clinical findings, respiratory function tests, lung mechanics, pulse oximetry and blood gases, imaging findings and occasionally bronchoscopy. Imaging of the respiratory system can be achieved by radiography, ultrasonography (US), computed tomography (CT), and Magnetic Resonance Imaging (MRI). These tools along with the clinical and laboratory results allow a holistic approach of the child with RF and effective planning of treatment.
Each imaging method displays different sensitivities and specificities depending on the wide range of RF causes. Imaging severity scoring systems have been employed mainly in adults and refer to single respiratory diseases like asthma and acute respiratory distress syndrome (ARDS). To our knowledge, such studies are limited in pediatric patients, mainly focusing in grading severe respiratory infection. There are sparse data on the contribution of imaging of the chest in critically ill children with respiratory illness with regard to the treatment, prognosis and outcome of RF, regardless of its cause.
Objectives
To explore the degree of contribution of imaging methods (radiography, US, CT, MRI) and bronchoscopy to the diagnosis, monitoring, treatment and outcome of children with acute respiratory failure who are admitted to PICU.
To propose imaging-based severity scores separately for radiographs, CT and MRI for imaging assessment of children with RF. To investigate the correlation of imaging findings and imaging severity scores with indicators of severity based on clinical and laboratory findings and evaluate these as tools for diagnosis, treatment and prediction of outcome.
Material and Methods
This retrospective study included 104 patients, aged 30 days to 18-year-old who were hospitalized in PICU of the University Hospital of Heraklion from January 2014 to June 2018 due to RF.
Demographics, length of stay, cause of RF, length of stay, cause for admission, clinical severity scores (PRISM and PELOD), and outcome (survival or not) were recorded. Respiratory support, duration of mechanical ventilation, chest tube placement, fraction of inspired oxygen (FiO2), and blood gas values were all recorded. The degree of ventilatory support, the respiratory failure severity index PaO2 / FiO2 and indices of oxygenation and ventilation were studied on the first day of the hospitalization and the day of most deteriorated function (worse pediatric ICU day) as defined by maximum needs for respiratory support and worse PaO2 and PaCO2 levels.
Imaging data regarding number of chest x-rays, US, CT and MT per patient were recorded where available. Radiographic findings of radiographs on the day of admission and on worse pediatric ICU day were blindly reviewed by four independent observers, were graded based on a radiographic severity score and scores were compared between readers and between admission and worse pediatric ICU day.
Additionally, data on available US, CT and MRI findings were recorded. CT imaging findings of any PICU day were reviewed and scored by two readers in consensus. The contribution of each imaging modality to the diagnosis and/or the modification of applied treatment was assessed.
Results
We studied 104 patients with an average age of 5,44 ± 0,6 years, 68 (65,4%) were males, average duration of hospitalization was 15,1 ± 2,1 days and mortality rate was 5%. 62 (59,6%) patients were mechanically ventilated with an average duration of mechanical ventilation 10,8 ± 2,04 days. Diagnoses on admission mostly included deterioration of chronic lung disease (21,15%), bronchiolitis or asthmatic attack (18,27%) and pneumonia (18,27%).
Overall 283 chest radiographs were performed. 162 radiographs, 85 upon admission and 77 at worse PICU day were studied and scored. The radiographic severity scores correlated positively with the severity of the disease, the duration of mechanical ventilation and the duration of hospitalization (p <0,001) while there was a negative correlation with the ratio PaO2 / FiO2 (p<0,001), both on the first day of hospitalization and on the worse PICU day. The change in the radiographic scores during hospitalization was associated with improvement or worsening of clinical respiratory disease severity indices, with RF indicators (p = 0.04) and the duration of mechanical ventilation (p <0,001).
Twenty chest CTs were performed. The CT imaging respiratory failure scores correlated positively with the clinical respiratory failure severity indices, the duration of hospitalization and the duration of mechanical ventilation (p &llt;0,001). Chest-US contributed mainly to the recognition and monitoring of complex parapneumonic effusions. Bronchoscopy was performed in patients with either a deterioration in radiographic scores and difficulty to interpret radiographic findings or an increased predicted mortality index (p <0,001). There were no recorded MR imaging tests.
Repeat radiographs showed no significant change in radiographic scores in 73.3% of children. The performance of a repeat chest CT revealed additional information that led to modification of treatment in 5.56% of children, while the corresponding percentage for chest US was 12%. Bronchoscopy showed additional findings in 17.65% of children.
Conclusion
The proposed radiographic and CT imaging severity scores can be a reliable diagnostic and prognostic tool for the evaluation of RF in children, as they are related to the severity of the disease, the length of hospitalization and mechanical ventilation and the final outcome in children hospitalized in the pediatric ICU.
Repeated imaging of the respiratory system with chest radiographs, CT, and US in children with respiratory failure during PICU hospitalization contributes to diagnosis by adding new information, and may result in modification of treatment.
|
Search
