Pneumocystis pneumonia in South African children diagnosed by molecular methods

Morrow, Brenda M. ; Samuel, Catherine M. ; Zampoli, Marco ; Whitelaw, Andrew ; Zar, Heather J. (2014-01)

Please cite as follows:

Morrow, B. M. 2014. Pneumocystis pneumonia in South African children diagnosed by molecular methods. BMC Research Notes, 7(1):26, doi:10.1186/1756-0500-7-26.

The original publication is available at


Abstract Background Pneumocystis pneumonia (PCP) is an important cause of hospitalization and mortality in HIV-infected children. However, the incidence of PCP has been underestimated due to poor sensitivity of diagnostic tests. The use of polymerase chain reaction (PCR) for pneumocystis has enabled more reliable diagnosis. This study describes the incidence, clinical features and outcome of PCP in South African children diagnosed using PCR. Methods A prospective study of children hospitalised in South Africa with suspected PCP was done from November 2006 to August 2008. Clinical, laboratory and radiological information were collected. Lower respiratory tract specimens were obtained for PCP immunofluorescence (IF), real- time PCR for pneumocystis, bacterial and mycobacterial culture. Nasopharyngeal aspirates were taken for immunofluorescence (IF), real-time PCR for pneumocystis and PCR for respiratory viruses. A blood specimen for bacterial culture and for cytomegalovirus PCR was taken. Children were followed for the duration of their hospitalisation and the outcome was recorded. Results 202 children [median (interquartile range, IQR) age 3.2 (2.1– 4.6) months] were enrolled; 124 (61.4%) were HIV infected. PCP was identified in 109 (54%) children using PCR, compared to 43 (21%) using IF and Grocott staining (p < 0.0001). Most PCP cases (88, 81%) occurred in HIV-infected children. All 21 cases (19%) occurring in HIV- negative children had another risk factor for PCP. On logistic regression, predictive factors for PCP were HIV infection, lack of fever, high respiratory rate and low oxygen saturation whilst cotrimoxazole prophylaxis was protective (OR 0.24; 95% CI 0.1 to 0.5; p < 0.002). The case fatality of children with PCP was higher than those without PCP (32.1% versus 17.2%; relative risk 1.87; 95% confidence interval (CI) 1.11 – 3.15). Amongst HIV-infected children, a CD4 less than 15% was the only independent predictor of mortality. Conclusions The diagnostic yield for PCP is more than 2.5 times higher on PCR than other detection methods. PCP is a very common cause of severe hypoxic pneumonia and is associated with high mortality in HIV-infected African infants.

Please refer to this item in SUNScholar by using the following persistent URL:
This item appears in the following collections: