Classification | Description of classification | Advantages | Disadvantages |
---|---|---|---|
WHO [16] | Pneumonia: Age 2–59 months with cough or difficult breathing and fast breathing and/or chest in-drawing. Severe pneumonia: pneumonia with any danger sign | Clinical: simple programmatic implementation to guide treatment Research: easy to enrol patients and findings directly generalisable | Clinical: no definition of aetiology, high levels of empiric antibiotic therapy Research: highly heterogeneous including viral, bacterial and other aetiologies |
NIH [17] | Community/hospital-acquired, health care-associated, aspiration, and atypical (caused by Legionella, Mycoplasma, Chlamydia) | Clinical: simple, guides empiric therapy Research: easy to enrol patients, findings directly generalisable | Clinical: little definition of aetiology or pathology, empiric antibiotic therapy Research: heterogeneous phenotypes |
Pathology | Acute inflammation of lung parenchyma, inflammatory alveolar infiltrate | Clinical: resolve cases of difficult diagnosis Research: highly homogenous | Clinical: limited availability and relevance Research: difficult to enrol patients |
ICD-10 | Uses clinical and laboratory diagnoses with known or unknown aetiology and many potential classifications | Clinical: not used clinically, primarily used for audit and administration Research: Analyses of clinical databases | Clinical: limited relevance Research: little definition of aetiology, heterogeneous, not systematic |
Harrison’s textbook [11] | Infection of pulmonary parenchyma by various pathogens, not a single disease. Terms lobar or bronchopneumonia not recommended. Clinical categories: community-acquired, nosocomial, aspiration | Clinical: encourages aetiologic diagnosis and guides empiric therapy Research: aetiologic diagnosis provides homogeneity, findings are directly generalisable | Clinical: difficult to confirm aetiology, substantial empiric antibiotic therapy Research: difficult to enrol patients with a single aetiology, clinical categories give heterogeneous aetiology and phenotype |
Clinical | Features: Age, acute/chronic, bronchiolitis, nosocomial, recurrent, comorbidity, HIV-related, complications, severity, mortality | Clinical: multiple inputs to guide treatment Research: may be easy to enrol patients, flexible, may define ‘important’ subgroups | Clinical: no aetiology, empiric therapy Research: heterogeneity, not standardised, difficult to generalise |
Chest radiograph | Interstitial/alveolar/lobar/air bronchogram WHO: dense, fluffy consolidation of entire lung or portion of a lobe; often with air bronchograms and possibly pleural effusion [15] | Clinical: supports viral or bacterial aetiology, identifies complications Research: some homogeneity and alignment with aetiology, standardised | Clinical: availability, time, expense Research: some difficulty enrolling patients, heterogeneous aetiology, unable to detect co-infection |
Ultrasound | Subpleural consolidation, B-lines, pleural line abnormalities, pleural effusion, air bronchogram | Clinical: fast, no radiation, for complications Research: simpler than radiograph, some homogeneity | Clinical: availability, no aetiology Research: detection in non-peripheral lung, not standardised, heterogeneity |
Microbiology | Culture of blood, lung/pleural aspiration, BAL Bacterial – viral – co-infection | Clinical: directs specific therapy Research: homogenous | Clinical: slow, limited detection Research: difficult to enrol patients |
Serology/antigen | Blood, urine, NPS (Legionella, S. pneumoniae) | Rapid, pathogen-specific | Range/sensitivity of tests, misclassification |
CRP | High CRP correlates with bacterial aetiology | Increased sensitivity for bacterial disease | Optimal threshold unclear, no aetiology |