Neonatal Septicaemia in Sub-Saharan Africa: A Protocol for Systematic Review and Meta-analysis.

Background: The morbidity and mortality from neonatal septicaemia (NNS) in low-middle income country remain high at the background of strained health care delivery system.The burden, pooled risks and outcomes of NNS are largely unknown. We aimed to produce a protocol for synthesizing evidence from available data for neonatal septicaemia in sub-Saharan Africa. Methods: We developed a search strategy using MeSH, text words and entry terms. Nine databases will be searched: PubMed, Embase, CINAHL, AJOL, Google Scholar, Web of Science, Cochrane Library, Research gate and Scopus. Only Observational studies retrievable in the English Language will be included. The primary measurable outcome is the proportion of neonatal with septicaemia while secondary outcomes include proportion of bacterial isolates and their antibiogram, risk factors for NNS, in hospital mortality, length of hospital stay, frequency of necrotizing enterocolitis and other sequel . All identied studies will be screened based on the inclusion criteria. Data will be deduplicated in Endnote version 9, before exporting to Rayyan QCRI for screening. Extractable data will include rst author’s name and year of publication, the country and regions in sub-Saharan Africa, total neonatal admissions, number with sepsis, the sample size, bacterial isolates, antibiogram, in-hospital mortality, length of hospital stay and frequency of necrotizing enterocolitis. All studies will be assessed for methodological, clinical and statistical heterogeneity. The NIH Quality assessment tool for observational studies and the Cochrane tool of risk of bias will be used to assess for the strength of evidence. Publication bias will be assessed using the funnel plot. Discussion: Results will be presented as the prevalence, standard error and condence interval of newborns with neonatal septicaemia in sub-Saharan Africa. Subgroup analysis using categorical data such as risk factors, bacterial isolates, antibiogram and outcomes of neonatal septicaemia will also be reported. A cumulative meta-analysis will be done to assess the time trend of the risk factors, pathogens and antibiogram.The CMA version 3 will be used for statistical analysis. Results will be presented in forest plots. This


Background
The presence of bacteria in the bloodstream of a newborn with associated clinical evidence of systemic acute in ammatory response and multiorgan dysfunction sydrome describes the state of a newborn with septicaemia. [1][2][3][4] Globally, neonatal septicaemia (NNS) remains one of the leading causes of morbidity and mortality in the rst four weeks of life accounting for 336,300 total deaths per annum. 2,5,6 The greatest burden disproportionately impact the low middle-income countries (LMICs) in contrast with high income countries. 7,8 Neonates compared with older children and adults are at increased risk for septicaemia. 3,6,9 They are immunolagic, lacking quantitatively and qualitatively in the cellular response to infection. 3,6,9 The neonates have defective humoral defence mechanism. 3 The responses in the classic and alternate complement system pathways are suboptimal. 3,9 The newborns have impaired chemotaxis, decreased neutrophil deformability, reduced complement receptor expression besides the poor oxidative metabolic response. 3,6,9 The inherent limitations in the immunological systems, suboptimal infection control and prevention, limited infection screening of high-risk mothers including the proportion of mothers with unknown HIV status, home deliveries practices, pervading cultural practices and myth that discourages safe deliveries in sub Saharan Africa; all these put the newborns at a higher risk for infection compared with neonates in the high-income countries. 1 There are several attempts to upscale the newborn survival strategies in sub Saharan African through essential newborn care, survival and thrive strategies. [10][11][12] But adequate measures and protocols for handling neonatal sepsis are still lacking.
However, there is increasing literature on NNS in the sub Saharan, eventhough the pooled burden in this setting is still lacking. 12 The variable de nition of terms, small sample sizes with con icting reports also makes the published data less generalizable, hence, its limited clinical application. Besides, recent literature has also shown a mismatch between empiric antibiotic choices and sensitivity pattern of isolates for neonatal sepsis in sub-Saharan Africa. This has implications for resource allocation, mobilization and programme implementation. The gap requires urgent attention if the sustainable development goal 3.2 of reducing neonatal mortality rate from 27 per 1,000 in 2019 to 12 per 1,000 will be achievable in all the member countries. 13,14 We, therefore, aimed at developing a protocol to enable transparent, accurate and reproducible systematic review and meta-analysis, which will provide robust evidence for an informed decision in the care and future policy regarding newborn care in sub Saharan Africa.

Methods And Design
Objectives: The main objective of this study is to determine the pooled prevalence of neonatal septicaemia in the sub Saharan Africa.

Study Design
This is a protocl for systmatic review and meta-analysis of observational studies conducted in the Sub Saharan Africa. This protocol is design to enable a reliable, transparent and accurate systematic review and meta-analysis. Using this protocol will enable determination of pooled prevalence of neonatal sepsis, pooled prevalence of risk factors, sub-group analysis of pathogens, antibiogram and the pooled prevalence of outcomes of neonatal septicaemia and cumulative trends over the years in the sub Saharan Africa. There is no timeframe or restriction in selecting eligible studies using this protocol.

INCLUSION CRITERIA:
A. Observational studies: Cohort studies, case controls, cross-sectional studies, historic cohort studies.
B. Studies must report the primary outcome: proportion of neonatal septicaemia.
C. Studies must be retrievable in the English Language.
D. Studies that report any of the following secondary outcomes, e.g risk factors, outcomes, bacterial isolates and antibiogram, in addition to primary outcome, will be included. PICOs: The details for participants, intervention, comparison and outcomes (PICOs) are shown in Table 1.
This review will be reported in line with Preferred Reporting Items for Systematic Reviews and Meta-Analyses (PRISMA 2015 Statement). 15 The Protocol has the PRISMA-P Checklist attached as a supplementary material.

Information sources
The search will employ sensitive topic-based strategies designed for each database. The search will be carried out in the following databases: PUBMED, EMBASE, CINAHL, RESEARCHGATE, AJOL, GOOGLE SCHOLAR, WEB OF SCIENCE, SCOPUS and COCHRANE LIBRARY. Only observational studies will be included. There is no time frame for selection of studies.

Search strategy
The search strategy includes MESH terms, text words and entry terms. Table 2 shows the search strategies as used in the Pubmed. The same search strategy will be used in other databases with slight modi cations.

Data Extraction and Management a. Data Extraction
Four main tools will be used for data extraction and management: a) EndNote version 9, b) Rayyan QCRI systematic review screening tool and c) Microsoft Excel and d) CMA Version 3 Software Five levels of data screening will be used for searched studies: i. Level 1 would involve screening of identi ed studies for the study design. Only observational studies, retrievable in the English Language would be selected.
ii. Level 2 will involve screening of selected studies in the titles and abstracts using entry terms, keywords and meSH terms.
iii. At Level 3, selected studies will be further screened for content by reading the full article using the same search strategy.
iv. Level 4 will involve snowballing of literature on references from included studies.
v. Level 5 will involve grey literature that report primary outcome and or secondary outcomes.
Seven reviewers are involved in this study. A pair of reviewers will independently screen the selected article for eligibility using Rayyan QCRI systematic review screening tool. 16 The two reviewers will be blinded from each other using the screening tool. Con ict between the paired reviewer shall be resolved by a third reviewer who would will server as a tie -breaker. The data of all screened studies will be exported into MICROSOFT EXCEL. Snowballing search of relevant studies through the review of the references of selected studies and grey literature will be performed manually.
b. Selection Process: Agreement between two independent and blinded reviewers who screened titles, abstracts, full texts of eligible observational studies, snowballed articles and grey literature will form the basis for selecting studies for inclusion for systematic review and meta-analysis. Where there are con icts in decision, this will be resolved by a third reviewer. Authors of eligible studies with any missing data will be contacted via email and telephone.
c. Data Collection Process: Deduplication of studies will be done using End note version 9. 17 Eligible studies will be screened using Rayyan QCRI and exported into Microsoft Excel. All templates for housing extracted data items will be created in Excel.
The following data items will be extracted from eligible studies and entered into MICROSOFT EXCEL: 1

Effect Sizes
The primary effect size is prevalence.
Different primary indexes in individual studies of same design and report outcome will be converted to prevalence in the CMA Software.
Categorical outcomes will be used for subgroup analysis.
Numerical outcomes will be used for meta-regression.
Data synthesis a. Studies that passed the methodological quality assessment using the NIH quality assessment tool will be extracted. The results will be presented in tabular format.
b. in addition to a narrative synthesis, the following will be included in the meta-analysis: 1. The proportion of babies with neonatal septicaemia.

Numerical variables for meta-regression c. Quantitative analysis
Pooled prevalence of neonatal septicaema, standard error and variance will be calculated with 95% CI. This is the primary measurable outcome.
Sub-group analysis will be performed using categorical data on i) regions in Sub Saharan Africa, ii) types of pathogens, iii) the antibiogram, iv) risk factors and v) the secondary outcomes.
Subgroup analysis will be performed using categorical data such as geographical location, types of bacterial isolates, antibiogram, components of risk factors and outcomes as moderators. Results of subgroup analysis will be presented in forest plots.
Meta-regression will be performed on prevalence of neonatal septicaemia as outcome variable using frequency of risk factors, length of hospital admission, gestational age at birth and mortality as explanatory variables. This will establish a regression model for predicting prevalence of neonatal septicaemia.
A cumulative meta-analysis will performed to check for trend in prevalence of neonatal septicaemia of the years.
Eligible studies will be quantitatively analysed using the Comprehensive Meta-Analysis (CMA) software version 3 (Biostat, USA). 18 .

Criteria for Quantitative data Synthesis
The pooled prevalence, standard error, variance and 95% con dence interval, and. Heterogeneity will be assessed using the Q statistics and its p-value, tau 2 and the Higgins I 2 . I² values of less than 40% will be considered low heterogeneity while values > 40 but < 75% will be considered moderate and values > 75% are high. A random-effect model will be used for computation in this study. A sensitivity analysis will be performed to check for outlying studies and their effects on pooled prevalence of neonatal septicaemia. Publication bias in the selection of studies will be visually assessed on the funnel plot (trim and ll method) and tested for asymmetry. Other statistical tests such as Egger's regression intercept, Begg and Mazumdar's rank correlation and Orwin's fail-safe N will be used where appropriate.

Risk of bias
The risk of bias in the included studies will be accessed for the individual article using the National Institute of Health (NIH) Quality assessment tool for observational cohort and cross-sectional studies.
The NIH Quality assessment tool has 10 questions. This will be cross-checked with the Cochrane tool of risk of bias assessment for the strength of the body of evidence; i.e. using speci c relevant items from this tool to assess the strength of the body of evidence.. Studies with extreme bias will be subjected to sensitivity testing using the include/exclude function in the CMA Software.

Assessment of Meta-bias
Meta-bias will be assessed as follows: i. Method of testing/reporting of neonatal septicaemia at the outcome level.
ii. Reporting of study: Studies that were reported in different units but similar in outcome and design will be converted based on individual case evaluation. This will be evaluated for individual studies by assessing the unit of reporting of studies, for example, whether prevalence with con dence intervals or incidence or proportion will be reported.
iii. Heterogeneity will be assessed at the study level using the Q statistics and its p-value, tau 2 and the Higgins I 2 iv. Publication bias will be assessed at the study level using the funnel plot (trim and ll method) and test for asymmetry v. Sensitivity analysis will assessed at the study level using include and exclude function in the CMA Software

Results
The study selection process will be summarized in a ow diagram according to the PRISMA 2015 Statement and PRISMA-P Checklist. A table of the search strategy in various databases showing text words, MeSH and entry terms will be included. List of included studies will be summarized in a table.
Quantitative data such as pooled prevalence, standard error and 95% CI, p values, relative weights assigned to studies and heterogeneity tests will be included in the forest plots. A table of quality scores and risk of bias of each eligible study will be included. Forest plots to show sub-group analysis will be included. Meta-regression and sensitivity analysis will be shown in Figures and Tables.

Discussion
The study will examine the type of pathogen that predominantly causes neonatal septicaemia and the most widely used antibiotics (from antibiogram report). It will report associations, if any, between risk factors and secondary outcomes to prevalence of neonatal septicaemia. It will discuss the implications of the pooled prevalence of septicaemia in the sub Saharan Africa. The nal study will be published in a peer-review journal and the ndings will be disseminated to relevant health authorities and clinical groups that manage pregnant women and newborns. Ethical Approval/ Dissemination: The study will use published data, thus, no ethical approval is required.
Consent for Publication: Table 1 The details for participants, intervention, comparison and outcomes (PICOs)

Population
Neonates with septicaemia Intervention Antibiotics

Comparison
Neonates without septicaemia Outcome Primary outcome: prevalence of neonatal septicaemia Secondary outcomes: Risk factors, antibiogram sensitivity, mortality, mortality and recovery Table 2 The search strategy

Search No
Search strategy No of articles