This protocol has been registered within the PROSPERO database (CRD42020202115). This review will follow the relevant domains of the PRISMA-P (Preferred Reporting Items for Systematic Reviews and Meta-Analyses-Protocols) statement for quantitative aspects. Figure 1 summarizes the flow of the systematic review process.
Eligibility criteria
Types of studies
We will include all study designs that have the potential to address one or more of the review questions outlined above. These include observational studies, notably prospective cohort studies, retrospective cohort studies, case–control studies, and case series. Data may be available from one or other arms of randomized and non-randomized clinical trials as well; hence, these will be included also. We will also include any existing systematic reviews of literature that address any of the review questions.
We will exclude narrative reviews that do not provide objective data and also case series with less than 5 participants. We will also exclude studies related to animal experiments, in vitro experiments, and ex vivo human studies.
Broadly, we expect to include the following types of studies:
-
1.
Those which describe micro-organisms in the upper airway respiratory tract and/or lungs of healthy children
-
2.
Those which report micro-organisms in the upper airway respiratory tract and simultaneously the lungs of children with respiratory diseases
-
3.
Those which compare micro-organisms in the upper airway respiratory tract of healthy children versus those with respiratory diseases
-
4.
Those which compare micro-organisms in the lungs of healthy children versus those with respiratory diseases
Types of participants
We will include studies conducted in children (age range birth to 18 years) who are normal or healthy or asymptomatic (as defined by authors of individual publications) as well as those with acute or chronic respiratory disease (as defined by the authors of publications).
We will exclude studies that report findings in participants older than 18 years of age, or where data of children and adults are presented together, without the possibility of extracting data of children separately. We will also exclude post-mortem studies. If potentially eligible studies report data from specimens such as tracheal aspirate, tracheostomy tube secretions, and endotracheal tube aspirates, these will be excluded (as it is difficult to clearly classify them as upper or lower respiratory specimens).
Types of exposure
We will include studies reporting the microbiota identified in children during health and respiratory disease, irrespective of the age of the children, timing of specimen collection, methods used for specimen collection, and microbiologic processing technique. Since this is not a systematic review of interventions, no specific interventions will be considered. However, if studies addressing one or more of the review questions include any intervention(s), these will be not be excluded.
Types of outcome measures
-
1.
Organisms (bacteria, viruses, fungi) identified in upper respiratory tract specimens of healthy children
-
2.
Organisms (bacteria, viruses, fungi) identified in the lungs of healthy children
-
3.
Organisms (bacteria, viruses, fungi) identified in upper respiratory tract specimens and simultaneous lung specimens of children with respiratory diseases
Information sources
Two authors will independently search the following electronic databases to identify relevant studies: Epistemonikos and Cochrane Library for systematic reviews and MEDLINE (through PubMed), EMBASE, Cochrane CENTRAL, LIVIVO, Web of Science, Scopus, and CINAHL databases for primary studies. All searches will be run from inception to 30 June 2021, or the actual date of publication of the protocol, whichever is later. There will be no restrictions based on language or geographies.
Handsearching
The authors will check reference lists of all primary studies and review articles for additional references. Articles identified through reference lists and bibliographic searches will also be considered for data extraction.
Grey literature
We will conduct a grey literature search (to identify studies not indexed in the databases listed above) using OpenGrey (www.opengrey.eu/), ProQuest, and Google Scholar.
Search strategy
MeSH terms and synonyms for the following keywords will be combined together to run the literature search: “child, micro-organism, respiratory.” A typical search string based on this is as follows: “(organism OR microorganism OR micro-organism OR micro organism OR microbe OR bacteria OR virus OR fungus OR microbiome OR microbiota) AND (child OR pediatr*) AND (respiratory OR airway OR lung OR nasopharyngeal OR oropharyngeal OR nasal OR throat).” Pilot testing of this search string yielded over 67,000 citations in PubMed, confirming that the search strategy is comprehensive and unlikely to miss any relevant citations.
We will use the Peer Review of Electronic Search Strategies (PRESS) checklist for systematic reviews, for structured reviews of our literature search strategies [19]. The checklist is designed to identify errors in the search strategy and enhance the search.
Study records
Selection process
Two review authors will independently screen titles followed by abstracts of all studies identified through the searches and then retrieve the full-text study reports/publications. Two authors will independently screen the full text and identify studies for inclusion. They will also identify and record reasons for exclusion of the ineligible studies. A table will be presented listing the studies excluded from our synthesis at the full-text stage and the reasons for exclusion.
Any disagreements will be discussed and resolved among review authors, with arbitration by an external expert if necessary. Where the same study, using the same sample and methods, is presented in different reports, such reports will be collated so that each study, rather than each report, is the unit of interest in the review. After removal of duplicate publications, the final list of included studies will be created. The study screening form as well as the data extraction form to be used in this systematic review will be pilot tested in advance to ensure there are no errors. A PRISMA flow diagram will be used to illustrate the search results and the process of screening and selecting studies for inclusion.
Data management
We will use Rayyan (https://www.rayyan.ai) for the management of the screening and data extraction stages of the systematic review.
Translation of publications in languages other than English
For papers that are published in a language other than English, the abstract will be subject to initial translation through open source software. If this indicates potential inclusion, or if the translation is inadequate to permit a decision, an attempt will be made to obtain a formal translation of the full text. If this cannot be done, the authors will categorize the study as “awaiting classification” to ensure transparency in the review process.
Data collection process (data extraction and management)
Dealing with missing data
We will contact the corresponding authors of studies where data is/are missing and try to obtain the missing data. If this fails, we will try and impute data where possible. If that is not feasible, we will state as such.
Outcomes and prioritization
-
1.
Organisms (bacteria, viruses, fungi) identified in upper respiratory tract specimens of healthy children, with the frequency of each organism identified and proportion of children with each organism type
-
2.
Organisms (bacteria, viruses, fungi) identified in the lungs of healthy children, with the frequency of each organism identified and proportion of children with each organism type
-
3.
Comparison of organisms (bacteria, viruses, fungi) identified in upper respiratory tract specimens of healthy children versus organisms identified in their lungs, with the frequency of each organism identified and proportion of children (in each group) with each organism type
-
4.
Organisms (bacteria, viruses, fungi) identified in upper respiratory tract specimens and simultaneous lung specimens of children with respiratory diseases, with the frequency of each organism identified and proportion of children with each organism type (in both sites)
-
5.
Comparison of organisms identified in upper respiratory tract specimens of healthy children versus those with respiratory diseases, with the frequency of each organism identified and proportion of children (in both groups) with each organism type
-
6.
Comparison of organisms identified in lungs of healthy children versus those with respiratory diseases, with the frequency of each organism identified and proportion of children (in both groups) with each organism type
Data synthesis
The data obtained will be described in detail. We will try to pool data and perform meta-analysis where feasible. In general, this will be feasible for studies that address similar questions with respect to the population, exposure, comparison (if any), and outcomes. For data that cannot be pooled by meta-analysis, we will use the Synthesis Without Meta-analysis (SWiM) guideline checklist [20] to ensure that the quantitative narrative synthesis of data remains as free of bias, as is feasible. The checklist has 9 items that encompass critical aspects of data synthesis including grouping of studies, methods for synthesizing data, presentation of the data, and limitations of the synthesis.
Statistical analysis
We will present the data with descriptive statistics and provide pooled estimates of outcome parameters, wherever it is feasible to perform a meta-analysis. Pooled estimates will be presented with 95% confidence intervals. Outcomes reported through dichotomous variables will be expressed as proportions and compared within and/or between groups (where applicable) using odds ratios. Outcomes reported through continuous variables will be expressed as mean (SD) and compared within and/or between groups (where applicable) using weighted mean difference. For continuous variables expressed as median (IQR), efforts will be made to convert the values to mean (SD). The default analysis will be a random effects model.
Sensitivity analysis
A sensitivity analysis will be done by excluding the studies with high risk of bias.
Subgroup analysis
Subgroup analysis will be conducted (where possible) based on the following characteristics:
-
1.
Age groups: <1 year, 1–5 years, 6–12 years, and 13–18 years. The cut-off of 12 years has been chosen for one of the subgroups, because in some developing countries (including India), children till 12 years of age are managed by pediatricians, and thereafter by physicians caring for adults. Therefore, there may be literature that included children till 12 years of age only
-
2.
Type of upper respiratory tract specimen respiratory specimen: nasopharyngeal swab/aspirate, oropharyngeal swab/aspirate, others
-
3.
Methods for identification and comparison of microbial flora: culture versus molecular methods
-
4.
Type of respiratory disease: infectious condition versus non-infectious condition, and acute versus chronic disease
Assessment of heterogeneity
We will test for heterogeneity using the I2 statistic. We will interpret heterogeneity as outlined in the Cochrane Handbook for Systematic Reviews of Interventions [21]. An I2 statistic < 50% will be considered to be a low level of heterogeneity, 50 to 75% a moderate level, and >75% a high level. Where substantial heterogeneity is identified, we will explore possible causes for it.
Assessment of reporting biases
Wherever possible, we will obtain the original trial protocols for comparison with the published papers to ensure that all outcomes were reported. If it is not possible to obtain the trial protocols, we will scrutinize the “Methods” section of the published paper(s) to ensure full reporting of all measured variables. We will use the Outcome Reporting Bias in Trials (ORBIT) classification system to highlight missing or incomplete outcome reporting of the outcomes [22].
If negative data were not fully reported, we will contact the primary investigators for these data. We will explore reporting bias using a funnel plot. We will also assess publication bias by looking for evidence of conference presentations not followed by subsequent journal publications.
Risk of bias in individual studies and assessment of methodological quality
Two authors will independently assess the risk of bias and methodological limitations of each included observational study using the Newcastle–Ottawa Scale (NOS) [23]. NOS is used to assess the quality of non-randomized studies including case–control and cohort studies. The NOS contains eight items, categorized into three broad domains viz. (i) selection of the study groups, (ii) comparability of the groups, and (iii) ascertainment of either the exposure or outcome of interest for case–control or cohort studies, respectively. We will rate the quality of the studies (good, fair, and poor) by awarding stars in each domain following the guidelines of the Newcastle–Ottawa Scale. A “good” quality score will require 3 or 4 stars in selection, 1 or 2 stars in comparability, and 2 or 3 stars in outcomes. A “fair” quality score will require 2 stars in selection, 1 or 2 stars in comparability, and 2 or 3 stars in outcomes. A “poor” quality score will reflect 0 or 1 star(s) in selection, or 0 stars in comparability, or 0 or 1 star(s) in outcomes [24].
For case series and case studies, we will use the Joanna Briggs Institute (JBI) Critical Appraisal Checklist for case series [25].
We will assess the risk of bias across included studies in two ways, as per the Cochrane Handbook guidelines [21]. First, we will assess the risk of bias for an individual outcome, by making judgments about evidence quality. Second, we will try to assess the overall risk of bias across included studies by making judgments on empirical evidence of bias, likely direction of bias, and likely magnitude of bias.
Confidence in cumulative evidence
We will present a summary of finding tables, displaying a structured summary of each review question, findings, and references to the studies contributing data to each review question.
Two reviewers will independently assess the quality of evidence of each outcome based on five GRADE considerations, i.e., study limitations, consistency of effect, imprecision, indirectness, and publication bias. We will use the methods described in the Cochrane Handbook for Systematic Reviews of Interventions [21], employing GRADEpro GDT software. We will justify decisions to downgrade or upgrade the quality using footnotes with comments. We will also consider the overall quality of evidence across outcomes. The quality of evidence will be rated as high, moderate, low, or very low.