BACKGROUND. Accurate diagnosis and attribution of the aetiology of pneumonia are important for measuring the burden of disease,
implementing appropriate treatment strategies and developing more effective interventions.
OBJECTIVES. To produce revised guidelines for the diagnosis of pneumonia in South African (SA) children, encompassing clinical,
radiological and aetiological methods.
METHODS. An expert group was established to review diagnostic evidence and make recommendations for a revised SA guideline. Published
evidence was reviewed and graded using the British Thoracic Society grading system.
RESULTS. Diagnosis of pneumonia should be considered in a child with acute cough, fast breathing or difficulty breathing. Revised World
Health Organization guidelines classify such children into: (i) severe pneumonia; (ii) pneumonia (tachypoea or lower chest indrawing);
or (iii) no pneumonia. Malnourished or immunocompromised children with lower chest indrawing should be managed as cases of
severe pneumonia. Pulse oximetry should be done, with hospital referral for oxygen saturation <92%. A chest X-ray is indicated in severe
pneumonia or when tuberculosis (TB) is suspected. Microbiological investigations are recommended in hospitalised patients or in outbreak
settings. Improved aetiological methods show the importance of co-infections. Blood cultures have a low sensitivity (<5%), for diagnosing
bacterial pneumonia. Highly sensitive, multiplex tests on upper respiratory samples or sputum detect multiple potential pathogens in
most children. However, even in symptomatic children, it may be impossible to distinguish colonising from causative organisms, unless
identification of the organism is strongly associated with attribution to causality, e.g. respiratory syncytial virus, Mycobacterium tuberculosis,
Bordetella pertussis, influenza, para-influenza or severe acute respiratory syndrome coronavirus 2 (SARS-CoV-2). Investigations for TB
should be considered in children with severe pneumonia who have been hospitalised, in a case of a known TB contact, if the tuberculin skin
test is positive, if a child is malnourished or has lost weight, and in children living with HIV. Induced sputum may provide a higher yield
than upper respiratory sampling for B. pertussis, M. tuberculosis and Pneumocystis jirovecii.
CONCLUSIONS. Advances in clinical, radiological and aetiological methods have improved the diagnosis of childhood pneumonia.