key: cord-0001559-b0tlco4t authors: Howie, Stephen R. C.; Morris, Gerard A. J.; Tokarz, Rafal; Ebruke, Bernard E.; Machuka, Eunice M.; Ideh, Readon C.; Chimah, Osaretin; Secka, Ousman; Townend, John; Dione, Michel; Oluwalana, Claire; Njie, Malick; Jallow, Mariatou; Hill, Philip C.; Antonio, Martin; Greenwood, Brian; Briese, Thomas; Mulholland, Kim; Corrah, Tumani; Lipkin, W. Ian; Adegbola, Richard A. title: Etiology of Severe Childhood Pneumonia in The Gambia, West Africa, Determined by Conventional and Molecular Microbiological Analyses of Lung and Pleural Aspirate Samples date: 2014-09-01 journal: Clin Infect Dis DOI: 10.1093/cid/ciu384 sha: 4ceca734c7185a61184a7e464e454b7f7d958dba doc_id: 1559 cord_uid: b0tlco4t Molecular analyses of lung aspirates from Gambian children with severe pneumonia detected pathogens more frequently than did culture and showed a predominance of bacteria, principally Streptococcus pneumoniae, >75% being of serotypes covered by current pneumococcal conjugate vaccines. Multiple pathogens were detected frequently, notably Haemophilus influenzae (mostly nontypeable) together with S. pneumoniae. Pneumonia remains the leading cause of death in children worldwide. A better understanding of the range of pneumonia pathogens is needed to reduce child mortality further [1] , but this is hampered by limitations at the bedside and in the laboratory. Pneumonia etiologic studies have depended largely on blood culture, which is insensitive [2] . Direct percutaneous aspiration from the site of lung infection, rarely done despite its good safety record, yields the highest-quality clinical specimen, but even then bacterial culture identifies a pathogen in no more than half of cases [3] . Molecular methods have advantages over conventional methods for the detection and characterization of pathogens in clinical samples, and hold promise for improving diagnostic sensitivity [4] . We describe the application of both approaches to the detection of pathogens in lung and pleural aspirates obtained from children with severe pneumonia in a West African setting. The objective of the study was to elucidate the etiology of severe pneumonia in this group more comprehensively than has been possible previously. The Gambia is a West African country of 1.8 million people with a human immunodeficiency virus (HIV) infection prevalence of <2%. A study of the etiology of severe childhood pneumonia was undertaken in the coastal area of The Gambia (Supplementary Appendix 1). At the time of the study (2007-2009), there was high coverage with conjugate Haemophilus influenzae type b (Hib) vaccine but no routine usage of pneumococcal conjugate vaccine (PCV), which is the case currently in many African countries. Cases were children aged 2-59 months with severe pneumonia defined clinically by modified World Health Organization (WHO) criteria (cough or difficulty in breathing, plus any of the following: lower chest wall indrawing, nasal flaring, or an oxygen saturation of <90% on pulse oximetry). Participants were recruited from the Medical Research Council (MRC) hospital in Fajara; the Royal Victoria Teaching Hospital in Banjul; and the major health centers at Fajikunda, Serekunda, and Brikama. Children with a cough of ≥2 weeks, or severe anemia (hemoglobin level <6 g/dL) or confirmed wheeze were excluded. Radiological pneumonia was defined using WHO criteria (ie, endpoint consolidation or pleural effusion). HIV testing was done if informed consent was given after standard counseling. A standard WHO guideline-based antibiotic regimen was used. Written informed consent was obtained for participation in the study from parents or guardians. The study was approved by the Gambian government/MRC Joint Ethics Committee (SCC/EC1062). Details of lung aspiration in our setting over a period of 25 years, during which there have been no associated serious adverse events, have been described previously [5] . Participants underwent lung aspiration if they had accessible consolidation adjacent to the chest wall, had no contraindications, and written informed consent had been obtained. Pleural aspiration was undertaken in those with pleural effusions. Culture, nonmolecular serotyping, HIV testing, singleplex polymerase chain reaction (PCR) for lytA and cpsA (for Streptococcus pneumoniae) and glpQ (for H. influenzae), 16S rRNA PCR, multilocus sequence typing (MLST), molecular serotyping, and multiplex fast-track 33 PCR, all using standard methods (Supplementary Appendix 2) [2, [6] [7] [8] [9] [10] [11] [12] [13] , were performed at the MRC Unit in The Gambia, and multiplex MassTag PCR [14] was performed at Columbia University. Multiplex pathogen targets are listed in Supplementary Appendix 3. Laboratory analyses were done blinded and independent of each other. Summary results of organisms identified to genus or species level using 1 or more detection methods were compiled. Demographic and clinical data were double-entered into an SQL database (Microsoft Corporation) and verified, and laboratory data were entered into an Access (Microsoft Corporation) database, cross-checked, and verified. Statistical analyses were done using Stata software, version 11 (StataCorp). Fifty-five children, representing 74% of the radiological severe pneumonias and 26% of all clinical severe pneumonia cases identified in the study period, underwent lung aspiration (n = 47) or pleural fluid aspiration (n = 9); 1 participant underwent 2 aspirates at different time points (Supplementary Appendix 4). HIV testing was done in 33 of 55 (60%) participants, 2 of whom were positive for HIV type 1. The characteristics of those who underwent lung or pleural aspiration were similar to the radiological pneumonia group from which they were drawn (Supplementary Appendix 5). Pathogens were identified to genus or species level in 53 of 56 (95%) samples by 1 or more laboratory methods (Table 1 ). An organism was cultured from 21 of 56 (38%) specimens: S. pneumoniae in 14 (25%), H. influenzae (all non-type b on latex agglutination) in 3 (5%), and Staphylococcus aureus in 3 (5%). Ziehl-Neelsen staining was done in 37 of 56 (66%) lung aspirate samples (all negative); 35 of 37 (95%) underwent culture for Mycobacterium tuberculosis, and all were negative. Streptococcus pneumoniae was detected by 1 or more molecular assays (Supplementary Appendix 6) in 48 of 53 (91%) samples and in 2 or more assays in 36 of 53 (68%) samples. Molecular pneumococcal serotyping (40 serotypes; Supplementary Appendix 7) showed the following prevalences: serotype 1, 22%; serotype 4, 18%; serotype 14, 18%; serotype 5, 16%; serotype 6A/B/C, 4%; and serotype 9V/A, 2%. Ten-and 13-valent PCVs both include serotypes that account for 76%-80% of serotypes identified in the samples in this study. Haemophilus influenzae was detected in 12 of 53 (23%) samples. Four of the 12 had sufficient deoxyribonucleic acid (DNA) for full MLST typing, 3 being nontypeable H. influenzae (NTHi), whereas 1 in a child with HIV infection had a serotype b infection. Two of the 12 had sufficient loads for 5 or 6 of the standard 7 alleles of MLST to be defined, and these were suggestive of NTHi, 1 being in a culture-positive case confirmed as being non-type b by latex agglutination assay. Staphylococcus aureus was detected in 3 of 53 (6%) samples, all of which were pleural fluid specimens. One or more other bacteria were identified to genus or species level in 8 of 53 (15%) samples. Viruses, led by respiratory syncytial virus, adenovirus, and bocavirus, were detected by 1 or more methods in 10 of 53 (19%) samples. Half (28/53 [53%]) of samples had >1 organism detected to genus or species level, predominantly 2 or more bacterial species. Codetection of S. pneumoniae and H. influenzae occurred in 11 of 53 (21%) samples, H. influenzae having the higher bacterial load in 4 of 11 samples. In the 3 specimens positive for S. aureus, low loads of S. pneumoniae were detected in 2 and of H. influenzae in the other. This study showed that S. pneumoniae was the predominant pathogen in this group of children with radiologic pneumonias, followed by H. influenzae, S. aureus, a range of gram-negative bacteria, and a few viruses. Potentially causative pathogens were found in all but 1 sample (98%), in contrast to the use of culture alone, which yielded an organism in 38% of specimens. Streptococcus pneumoniae and H. influenzae, predominantly NTHi, were detected together in around 1 in 5 cases in this Hib-vaccinated population. Current PCVs include >75% of serotypes identified in these samples. Gambian studies of pneumonia etiology in the pre-Hib vaccine era showed a predominance of S. pneumoniae followed by Hib, as have other low-income country studies, including a recent lung aspirate study from Malawi using PCR [15, 16] . The role of NTHi in pneumonia has been raised before [17] and has been controversial [15] . It is likely that this study's findings are relevant to similar patient groups in other developing-country settings, especially those where conjugate Hib vaccine is routinely used and where PCVs have not yet been introduced. The multicountry Pneumonia Etiology Research for Child Health (PERCH) study will provide important data in this respect [4] . The prominence of S. pneumoniae but low frequency of Hib in our study appears to contrast with recent Global Burden of Disease 2010 estimates [18] . The strength of this study is its detailed analysis of the best possible specimens for diagnosing the cause of pneumonia using sensitive methods from a well-defined patient group that is representative of radiologic pneumonias. Multiple assays have provided several lines of supporting evidence, and 2 laboratories analyzed the raw clinical specimens, strengthening the findings. The predominance of S. pneumoniae along with its serotype distribution is supported by a high level of concordance between independent laboratories [19] . The study's weaknesses are that the best possible specimens are still subject to sampling error, that the study is relatively small, and that the identification of potential pathogens does not in itself confirm that these are the causative agents, a fundamental challenge for the field. Nevertheless, if there are any clinical samples for which pathogen detection alone is sufficient to assign causation, then these are lung and pleural aspirates. The findings of this study emphasize the importance of bacteria, prominently S. pneumoniae, as a cause of severe pneumonia, and the potential for current PCVs to reduce the burden of this disease. This study also confirms that molecular methods are able to detect potential pathogens far more readily than culture and have a role in defining the etiology of pneumonia. The challenge of accurately assigning causation to the pathogens detected remains, particularly when using more generally available specimen types, and this will require the additional development of robust biomarkers of pathogen-specific disease. Notes contributed to study oversight and interpretation of data. M. A., O. S., E. M. M., and M. D. contributed to laboratory design, conduct, or interpretation. J. T. contributed to data analysis and interpretation. W. I. L. and T. B. oversaw and contributed to analyses at Columbia University and contributed to interpretation. K. M., B. G., P. C. H., and T. C. contributed to design and interpretation. M. J. and M. N. contributed to the design and conduct of the study. All authors contributed to critical review of the manuscript. Disclaimer. The Bill & Melinda Gates Foundation had no role in drafting the manuscript or the decision to publish. Financial support. This work was supported by the MRC and the Bill & Melinda Gates Foundation. Potential conflicts of interest. R. A. A. is an employee of GlaxoSmith-Kline Vaccines in Belgium and received previous grant awards for studies of bacterial diseases while working as an employee of the MRC Unit, The Gambia. All other authors report no potential conflicts. Breathing new life into pneumonia diagnostics The etiology of pneumonia in malnourished and well-nourished Gambian children The value and complications of percutaneous thoracic lung aspiration for etiologic diagnosis of community-acquired pneumonia The Pneumonia Etiology Research for Child Health Project: a 21st century childhood pneumonia etiology study Use of percutaneous transthoracic lung aspiration for the etiologic diagnosis of pneumonia: a 25 year experience from The Gambia Serotyping Streptococcus pneumoniae by multiplex PCR A detailed analysis of 16S ribosomal RNA gene segments for the diagnosis of pathogenic bacteria A multilocus sequence typing scheme for Streptococcus pneumoniae: identification of clones associated with serious invasive disease Characterization of encapsulated and noncapsulated Haemophilus influenzae and determination of phylogenetic relationships by multilocus sequence typing A novel Campylobacter jejuni sequence type from a culture-negative patient in The Gambia Determination of bacterial load by real-time PCR using a broad-range (universal) probe and primers set Measuring nasal bacterial load and its association with otitis media Longitudinal molecular microbial analysis of influenzalike illness Diagnostic system for rapid and sensitive differential detection of pathogens Epidemiology and etiology of childhood pneumonia PCR improves diagnostic yield from lung aspiration in Malawian children with radiologically confirmed pneumonia Aetiology of pneumonia in children in Goroka Hospital Global and regional mortality from 235 causes of death for 20 age groups in 1990 and 2010: a systematic analysis for the Global Burden of Disease Study Serotype analysis of Streptococcus pneumoniae in lung and nasopharyngeal aspirates from children in The Gambia by MassTag PCR Acknowledgments. The authors thank the participants and their parents/guardians; the clinical, field, laboratory, data, and administrative teams at the Medical Research Council (MRC) Unit in The Gambia; the staff of the Centre for Infection and Immunity, Columbia University; Simon Donkor for his contribution to managing data; Nuru Adams and Essa Jarra, who assisted with the molecular work in The Gambia; Ramu Sarge-Njie and her team, who performed the HIV testing; Christian Bottomley and Yin Bun Cheung, who assisted with statistics advice; and Pamela Collier Njai, Charles Onyeama, Danlami Garba, Uduak Okomo, and Augustin Fombah, who contributed to the clinical aspects of the study.Author contributions. S. R. C. H. proposed and led the study; designed epidemiological and clinical aspects; contributed to interpretation; completed the first draft of the article; and finalized the article. G. A. J. M. led Gambian-based molecular analysis design and conduct; contributed to interpretation; and contributed to the first draft of the article. R. T. undertook laboratory analyses at Columbia University and contributed to interpretation. B. E. E., R. C. I., O. C., and C. O. collected lung and pleural aspirates and clinical data from participants. R. A. A. contributed to the conception of the study; led conventional microbiological aspects; and All authors have submitted the ICMJE Form for Disclosure of Potential Conflicts of Interest. Conflicts that the editors consider relevant to the content of the manuscript have been disclosed.