Systematic Review and Meta-Analysis on the Aetiology of Bacterial Pneumonia in Children in Sub-Saharan Africa

Background Before the introduction of vaccination to protect children from pneumonia, Streptococcus pneumoniae and Haemophilus in�uenzae type B (HiB) were the most frequent aetiological agents causing bacterial pneumonia in children under �ve years old. However, under the in�uence of these vaccines, the aetiology of childhood pneumonia appears to be changing and non-vaccine-type S. pneumoniae, non-typeable H. in�uenzae, and Staphylococcus aureus are becoming more relevant. Methods We conducted a systematic review aimed at identifying the common causes of bacterial pneumonia in children in sub-Saharan Africa. We searched PubMed, Web of Science and African Index Medicus and included primary studies conducted since January 2010 that reported on the bacterial causes of pneumonia in children under �ve from sub-Saharan Africa. We extracted data items (about the study setting, pneumonia diagnosis, sampling, microbiological methods, and etiological agents) as well as study quality indicators.


Background
Pneumonia is an acute infection of the lungs and tissues of the lower respiratory tract. This condition disproportionately affects young children, in whom both incidence and mortality are high (1). Childhood pneumonia is one of the leading causes of illness and death among children less than ve years of age. According to the Global Burden of Disease (GBD) estimates, more than 100 million children under the age of ve years suffered from pneumonia in 2015, and approximately 700,000 died (1). The burden of childhood pneumonia is particularly high in sub-Saharan Africa (SSA) where the incidence and mortality are much higher than in any other region of the world (1,2). The region accounts for about half of all childhood pneumonia deaths, while only 20% of the global under-ve population lives in SSA (1,3).
Pneumonia in children can be caused by multiple organisms, of which bacteria and viruses are the most important. Although viral agents are responsible for most of the pneumonia cases, it is bacterial agents that are responsible for most of the severe cases resulting in hospitalization and death (4). According to the 2015 GBD estimates, 64% of pneumonia-related deaths in children below ve years were due to bacterial causes (1). Before the introduction of vaccination to protect children from pneumonia, Streptococcus pneumoniae and Haemophilus in uenzae type B (HiB) accounted for most of bacterial pneumonia in children under ve (3,4). With the use of highly effective pneumococcal conjugate vaccines (PCVs) and conjugate vaccines against H. in uenzae type B (HiB), the incidence of pneumonia due to vaccine-type (VT) S. pneumoniae and HiB have declined (2,3,5,6). However, under the in uence of these vaccines, the aetiology of childhood pneumonia appears to be changing (2). Serotypes of S. pneumoniae not included in PCV, nontypeable or non-type B H. in uenzae, and Staphylococcus aureus are becoming more relevant etiological agents of childhood pneumonia (2,7).
Knowledge of the common causative agents of pneumonia guides the choice of antibiotics to treat pneumonia. This is especially important in SSA where many cases of childhood pneumonia are treated empirically, without microbiological guidance. It is therefore essential that frontline health workers and policy-makers have up-to-date knowledge of organisms causing pneumonia in children. In this systematic review, we aim to summarize the current evidence on the causes of bacterial pneumonia in children under ve years of age in SSA after the introduction of conjugate vaccines.

Protocol and registration
The review was developed in line with the PRISMA guidelines (8). The original review protocol, as well as its amendments, were registered in PROSPERO (CRD42020203924).

Eligibility criteria
We sought records published after 2010 which report on pneumonia in children between four weeks and ve years of age in SSA. Studies reporting exclusively on children below four weeks of age were excluded because the etiological pattern of pneumonia is different in this age group (9). Studies were included if they reported on the frequency of bacterial causes of pneumonia in the relevant population, with or without comparison. Studies using culture or molecular methods on blood or any type of respiratory sample (nasopharyngeal swabs, induced sputum, lung aspirate) were eligible. We only included primary research and considered the following study designs: case series, surveillance, cross-sectional, case-control, cohort, and interventional studies. Modelling studies and reviews were not eligible.

Information sources
We searched the following electronic databases without language restrictions: MEDLINE using the PubMed interface (last search 10 October 2020), Web of Science database (last search conducted 16 October 2020), and African Index Medicus (last search 2 October 2020). The MEDLINE search was restricted to articles published after 1 January 2010; no restrictions were applied to other searches. We manually searched the reference lists of included records for other potentially relevant records.

Search
Our search strategy combined the key themes of the review question: (a) bacterial pneumonia (b) children and (c) sub-Saharan Africa. For each of the themes, we applied alternate terms and spelling combinations, including truncations and wildcards to improve sensitivity. This search strategy was applied to MEDLINE and Web of Science; in the search of African index Medicus, we omitted the theme of SSA. Full details of the search strategies and syntaxes are available as supplementary material (supplement 1).

Study selection
Screening of titles and abstracts and full-text screening for eligibility was conducted by blinded double-voting, with a third vote to resolve disagreements. CO, BE and OI screened the titles and abstracts while VW resolved disagreements.
Potentially eligible records from the title and abstract screening were considered for full-text assessment. The assessment of the full-texts was conducted by BE, OI and VW, with CO acting as a tiebreaker to resolve disagreements. MJ and CO subsequently searched the reference list of records included in the review for potentially relevant records. We used the Covidence platform (https://www.covidence.org/about-us-covidence/) to organize the screening and selection of records.

Data collection process
We developed a data extraction form, implemented it in Covidence, and re ned it after a pilot phase using ve included records. Next, one member of the review team extracted the relevant data items from all the included papers and a second member checked the extracted data. Disagreement between the primary extraction and data check was resolved by consensus between voting members in consultation with a third member of the team. No additional information was sought from investigators or authors.

Data items
The following categories of information were extracted: (a) study characteristics (study aim, design, and start and end date), (b) characteristics of the study population (description of cases, pneumonia case de nition, method of recruitment, the severity of pneumonia and number of children screened for pneumonia if applicable); (c) type of outcome measure (sample type, method of sample collection, method of bacterial identi cation, total number of samples collected, number of samples with positive test results for bacteria, and number of speci c bacterial isolates).
For case-control studies, we extracted similar data items for the control subjects.

Risk of bias in individual studies
Two members of the review team assessed the risk of bias, with disagreement resolved by consensus. We used the Joanna Briggs Institute (JBI) quality assessment tools for assessing the quality of included studies (10). As this review focuses on the cases with pneumonia we used the JBI tool for case-series for both case series and case-control studies.

Summary measure and analyses
Since this review aimed to summarize the prevalence of speci c bacterial agents among cases with pneumonia, our main summary measure was the proportion of pneumonia cases with speci c isolates. We rst conducted metaanalyses of these proportions per sample type and per pathogen, using a random-effects model and after a variancestabilizing transformation (double arcsine transformation). Second, for case-control studies of nasopharyngeal isolates, we also conducted a meta-analysis of the crude odds ratios of bacterial isolation comparing children with and without pneumonia. We assessed heterogeneity by computing Cochrane's Q and I 2 statistics which measure the proportion of the variation between studies that is due to heterogeneity and not by chance (11,12). R (package metafor) and STATA version 16 were used to conduct the analyses and to produce forest and funnel plots (13).

Risk of bias across studies
Assessing the risk of publication bias in meta-analyses of prevalence studies is not straightforward, as the prevalence is expected to vary across studies and funnel plots may not be relevant (14). We, therefore, discussed the possible presence of bias across studies and the implications it may have had on our ndings without making a quantitative evaluation. For the second type of meta-analysis in this review, i.e. the association between pneumonia and nasopharyngeal isolation of S. pneumoniae (expressed as odds ratio (OR)), we did construct a funnel plot. To assess the funnel plot symmetry, we relied mostly on visual inspection of the plot, with support from formal statistical tests (formal tests for asymmetry are underpowered when the funnel plot has fewer than 10 studies) (14).

Study selection
Eleven studies (reported in 12 records) were eligible for inclusion in the review (15,16,25,26,(17)(18)(19)(20)(21)(22)(23)(24). The search of PubMed, Web of Science and African Index Medicus retrieved 2279 records, 229 of which were duplicates. After title and abstract screening of 2050 records, we excluded 1954 because they were irrelevant. We assessed the remaining 96 fulltext records and excluded 84 (Fig. 1) because they did not report on bacterial causes of pneumonia (n = 3), were conducted before 2010 (n = 22), were not primary research (n = 13), and included persons outside the eligible age range (n = 7). Three additional reports were identi ed from manual searching of references, but all three were excluded after full-text assessment (not shown in PRISMA chart).
Study characteristics Table 1. Description of studies included in the review

Population
Two of the 11 studies in this review were multicenter (PERCH and GABRIEL networks) and nine were single-centre studies ( Table 1). Three studies were conducted before the introduction of PCV in the corresponding countries (17,25,26). Concerning study design, there were nine case-control studies and two case series. All studies recruited children in hospital and all were conducted prospectively. The diagnosis of pneumonia was mostly based on standard World Health Organization (WHO) de nitions of clinical (n = 6) or radiological (n = 6) pneumonia; one study (27) used a physician-based diagnosis. Taken together, the 11 studies contained information about 5362 pneumonia cases.

Outcomes
Three types of samples were used to determine the aetiological agents: nasopharyngeal samples (n = 9), blood (n = 5), and induced sputum (n = 1). The laboratory methods used were PCR (n = 10) and culture (n = 7), with 4 studies using more than one laboratory method

Quality appraisal of included studies
The quality appraisal of the included studies is summarized in Fig. 2. An important dimension of quality concern in the review was in case inclusion, some included studies did not provide enough information to make a judgement on completeness of case inclusion and consecutive case inclusion. Incomplete or non-consecutive case inclusion is a potential source of selection bias in case-control and case series studies.

Bacterial pathogens isolated from nasopharyngeal swabs
The most frequently identi ed bacteria from NPS were S. pneumoniae and M. catarrhalis. As shown in Table 2 and Fig. 3   The forest pot of bacterial agents isolated from blood among cases showed a relatively high proportion of S. pneumoniae and H. in uenzae (Fig. 4 below). S. pneumoniae was isolated from blood in an estimated 8% of cases (95% CI: 4% − 14%); while H. in uenzae was isolated in an estimated 3% of cases (95% CI: 1% -17%). S. aureus was less frequently isolated from blood samples, except for one study, where S. aureus was found in 5% of cases (15). M. catarrhalis was identi ed in the blood sample of 1 of 2189 children included in this analysis.

Association between bacterial nasopharyngeal carriage and pneumonia in children in SSA
We conducted separate analyses of the association between nasopharyngeal isolates of each bacterial agent and pneumonia by computing pooled odds ratio (OR) for case-control studies. As shown in Fig. 5, we found no evidence in favour of an association between nasopharyngeal carriage and pneumonia: the pooled OR was very close to 1.0. Also, in these meta-analyses, there was considerable heterogeneity in 3 of our analyses (I²= 86-93%; P < 0.01).

Risk of bias across studies
During this review, we found no indications of publication bias. Visual inspection of the funnel plot suggested no obvious asymmetry and the Peters test was not statistically signi cant (p = 0.62).

Summary of evidence
The main ndings of this systematic review are: (a) bacterial pathogens remain a relevant cause of pneumonia in children in SSA, and (b) the usual bacterial culprits persist. S. pneumoniae was the most commonly detected organism in blood samples from children with pneumonia. S. pneumoniae was also the most common organism identi ed from nasopharyngeal swabs in both cases and non-cases. When we compared nasopharyngeal isolates from children with pneumonia and those without pneumonia, we found no obvious difference in the proportion of children in whom S. pneumoniae, H. in uenzae, S. aureus and M. catarrhalis were isolated from the nasopharynx. We were unable to describe isolation patterns in severe versus non-severe pneumonia cases as nearly all the studies in the review were on children with severe pneumonia.
Before the introduction of PCV, nasopharyngeal colonization by S. pneumoniae among children in SSA was high, even in healthy children (29,30). The ubiquitous nature of pneumococcal carriage implies that inferring aetiology based on nasopharyngeal samples is problematic. Our review shows that in SSA, nasopharyngeal carriage is largely similar in children with and without pneumonia. One previous study found that in some instances, bacteria may be more easily isolated by culture from healthy children than from those with pneumonia (28). A possible explanation for this is the use of antibiotics before sample collection among children with pneumonia. However, comparing the frequency of carriage reported in our review with those reported before conjugate vaccine introduction suggests that overall carriage has not changed much (29)(30)(31)(32).
We also report on the continued importance of S. pneumoniae and H. in uenzae as bacterial agents causing pneumonia in children in the region. Studies conducted before conjugate vaccine introduction showed that these pathogens were the most commonly identi ed among children with pneumonia (4,33). We also found that vaccine-type serogroups of S. pneumoniae are still important colonizers of the nasopharynx.

Limitations
The small number of studies included in the meta-analysis combined with the random-effect model applied greatly increases the level of uncertainty around our meta-analysis estimates. Another limitation of our review was that the samples collected and the method of bacterial identi cation differed between studies. The different methods of bacterial identi cation have different levels of accuracy and this may account for some of the heterogeneity in observed results. The study designs and case de nition applied across studies also varied. On the quality assessment of individual studies, nearly all had at least one area of concern. The most commonly observed quality concerns were in the areas of consecutive case recruitment and complete case inclusion. Therefore, there is the potential for selection bias within these studies (see Table 2).
There was no evidence of publication bias amongst the assessed studies. However, there persists the possibility of bias due to language bias. Indeed, we observed that only one record screened for full-text inclusion was published in French.
This could result in an under-representation of studies from parts of SSA and a bias toward English predominant areas.
Furthermore, with the small number of countries presented in this review, it is unclear how representative they are of the wider SSA region.

Conclusions
Despite the widespread implementation of vaccination against S. pneumoniae and H. in uenzae in the past decade, these bacteria continue to colonize the nasopharynx of children and cause pneumonia. This, therefore, suggest that in resource-limited settings without microbiological support, the current empirical approach to the treatment of childhood

Competing interests
The authors declare that they have no competing interests Funding No speci c funding was available for this review. The COVIDENCE license was provided by Dr Kristien Verdonck from core research funding. All named authors receive a salary from their primary a liated institutions. The a liated institutions had no roles in the conceptualization, conduct, interpretation nor decision to publish this review.
Authors' contributions CO conceptualized the review and developed the review methodology, wrote the review protocol, searched the online databases, screened records, extracted data items, conducted risk of bias assessment, conducted the meta-analysis and prepared the draft manuscript. BE, IO, and VW reviewed the review protocol, screened records, extracted data items, conducted risk of bias assessment, and reviewed the manuscript. BS reviewed the protocol, and reviewed the manuscript. MJ conducted the manual literature search, screened manual records, and reviewed the manuscript. KV conceptualized the review, developed the review methodology, conducted the meta-analysis, guided interpretation of results, and reviewed the manuscript.