Aetiology of severe hospitalised pneumonia in HIV-uninfected children from Africa and Asia: the Pneumonia Aetiology Research for Child Health (PERCH) Case-Control Study

PERCH Study Group (2019). The Lancet. In Press

Abstract

Background: Pneumonia is the leading cause of death among children under 5 years. We estimated pneumonia aetiology in young African and Asian children, using novel analytics applied to clinical and microbiology findings.

Methods: We conducted a multi-site, case-control study of severe pneumonia among hospitalised children 1-to-59 months of age; controls were community, age-group matched children. Sites in Bangladesh, The Gambia, Kenya, Mali, South Africa, Thailand, and Zambia each enrolled for 24 months. Nasopharyngeal/oropharyngeal (NP/OP), urine, blood, induced sputum, lung aspirate, pleural fluid, and gastric aspirates were tested by culture and/or multiplex PCR. Primary analyses were restricted to HIV-uninfected cases with abnormal chest radiographs and HIV-uninfected controls. We applied a Bayesian, partial latent class analysis to estimate individual and population level probabilities of aetiological agents, incorporating case and control data.

Findings: Of the 4,232 cases and 5,325 controls enrolled, the 1,769 (41.8%) HIV-uninfected cases with CXR abnormalities and 5,102 (95·8%) HIV-uninfected controls comprised the primary analysis group. Wheezing was present in 31·7% of cases (555/1,752) (range by site: 10·6% to 97·3%). Thirty-day case fatality was 6·4% (114/1,769). Blood cultures were positive in 3·2% (56/1,749) of cases; S. pneumoniae was commonest (33·9%; 19/56). Virtually all cases (99·1%) and controls (98·1%) had at least one pathogen on NP/OP testing; detection of respiratory syncytial virus (RSV), parainfluenza, human metapneumovirus, influenza, S. pneumoniae, Haemophilus influenzae type b (Hib), H. influenzae non-type b, and Pneumocystis jirovecii from NP/OP were associated with case status. The aetiology analysis estimated that viruses accounted for 61·4% (95% Credible Interval (CI): 57·3, 65·6), bacteria for 27.3% (95% CI: 23·3, 31·6) and Mycobacterium tuberculosis for 5.9% (95% CI: 3.9, 8.3) of aetiology. Viruses were more common [68·0% (95% CI: 62·7, 72·7) vs. 54·5% (95% CI: 47·4, 61·5)] and bacteria less common [22·8% (18·3, 27·6) vs. 33·7% (95% CI: 27·2, 40·8)] in severe than very severe cases. RSV contributed the greatest aetiologic fraction (31·1%, 95% CI: 28·4, 34·2). Human rhinovirus, human metapneumovirus A/B, human parainfluenza virus, Streptococcus pneumoniae, M. tuberculosis, and Haemophilus influenzae each accounted for >5% of the aetiologic distribution. Differences in aetiologic fraction by age were observed for B. pertussis, parainfluenza types 1 and 3, parechovirus/enterovirus, P. jirovecii, RSV, rhinovirus, S. aureus, and S. pneumoniae, and by severity for RSV, S. aureus, S. pneumoniae, parainfluenza type 3, and all viral causes combined. Each site’s leading 10 pathogens accounted for > 79% of the site’s aetiologic fraction.

Interpretation:
A limited set of pathogens account for most hospitalised pneumonia cases. Preventing and treating a subset of pathogens could substantially impact childhood pneumonia outcomes.

MeSH Key Words: pneumonia, child health, case-control studies, aetiology, developing countries