Accuracy of Clinical and Diagnostic Studies for Pneumonia in Children

Background Pneumonia is the leading cause of death in children under five years of age worldwide.1 These deaths may be prevented by early detection and targeted antibiotic therapy.2 However, the diagnosis is not always clear on presentation to health care facilities. Missed diagnosis may lead to increased morbidity and mortality, while over-diagnosis may lead to unnecessary antibiotic use, which may further lead to increased antibiotic resistance, cause allergic reactions, and create unnecessary costs for patients. Therefore, using clinical tools and diagnostic capabilities to better improve diagnosis is critical.

Many studies have evaluated clinical signs and symptoms that may predict pneumonia.3-6 A recent meta-analysis found that no singular physical exam finding predicted pneumonia.5 The World Health Organization (WHO) Criteria diagnose pneumonia based on fast breathing; however, tachypnea has not been shown to strongly predict pneumonia.7 Additionally, most children with fever have compensatory tachypnea as a result, making the criteria of fast breathing difficult to diagnose pneumonia alone.8 In Nepal, one study evaluating the WHO criteria for pneumonia found a sensitivity of only 69.6% and specificity of 59.6%.9 Despite this, many providers rely on clinical exam findings and the WHO criteria for diagnosing pneumonia.

When available, diagnostic imaging is used regularly to confirm suspected pneumonia. Chest x-ray has been the standard for diagnosis in most facilities worldwide. However, remote facilities in resource-limited settings often lack radiographic imaging capabilities. Many facilities have bedside ultrasound available as it is easily portable, repeatable, and not associated with radiation. Ultrasound has been shown to be sensitive and specific for the diagnosis of pneumonia, yet few studies have evaluated the accuracy of lung ultrasound for pneumonia in pediatric patients in a resource-limited setting.10-14

The objective of this study is to evaluate the diagnostic accuracy of clinical history, physical exam, laboratory findings, and lung ultrasound compared to chest x-ray for the diagnosis of pneumonia in pediatric patients in a resource-limited setting. Determining diagnostic accuracy of these findings may help derive a clinical decision rule that may more accurately predict which children have pneumonia than current WHO guidelines.

Study Design

A prospective observational cross-sectional study of pediatric patients presenting for fever or respiratory complaints to the emergency department and outpatient department at Patan Hospital in Lalitpur, Nepal will be done over one year. Ethical approval will be obtained from the Nepal Health Research Council Ethical Review Board.

Study Setting and Population

Located in the Kathmandu valley, Patan Hospital is a large urban hospital with a 35-bed emergency department. The emergency department has an annual volume of approximately 48,000 patients, including approximately 8,000 pediatric visits. The admission rate is 20%.

Inclusion Criteria: Patients presenting under age 5 years of age with fever, respiratory complaints, or concern for pneumonia and receiving chest x-ray imaging.

Study Protocol

Parents will be consented for inclusion of child in the study (See consent form). Data will be collected on pediatric patients meeting the above inclusion criteria. Data will include demographics (age, gender), duration of symptoms, symptoms (presence or absence of fever, cough, chest pain, difficulty breathing, or vomiting), vital signs (temperature, respiratory rate, oxygen saturation), other physical exam findings (grunting, nasal flaring, retractions or indrawing of chest, crepitations, wheezing, or diminished breath sounds). These will be collected on the data collection form (see Appendix 1). Additionally, white cell counts with neutrophil counts and c-reactive protein will be collected if ordered by the clinician.

As a part of the evaluation, a bedside lung ultrasound will be performed by a clinician trained to perform lung ultrasounds. The bedside lung ultrasound is provided free for patients. Sonographers will be blinded to clinical information and results of any chest imaging. A Sonosite M Turbo (Fujifilm Sonosite, Inc.) ultrasound machine with a curvilinear probe will be used. In accordance with previous literature, the ultrasound examination will include ten views: two anterior views and two lateral views (one including the costophrenic angle), and one posterior view on each hemithorax. The physician will record ultrasound findings and interpretation directly after the ultrasound is complete. An ultrasound diagnosis of pneumonia is defined as the presence of unilateral B lines or subpleural lung consolidation. All ultrasounds will be reviewed for accuracy by a medical sonographer.

All patients will have a single posterioranterior (PA) chest x-ray as a part of the standard evaluation. The chest x-ray will be read by a board-certified radiologist, who is blinded to the clinical presentation and the results of any other imaging. Chest x-ray readings will be recorded on the standardized data form.