Prevalence and determinants of gestational diabetes mellitus in Africa based on the updated international diagnostic criteria: A systematic review and meta-analysis

Background: Gestational diabetes mellitus (GDM) is a major public health problem and threat to maternal and child health in Africa. No prior review has been conducted in Africa using the updated GDM diagnostic criteria. Therefore, this review aimed to estimate the pooled prevalence and determinants of GDM in Africa by using current international diagnostic criteria. Methods: A systematic review and meta-analysis was conducted by comprehensive search of the published studies in Africa. Electronic databases (PubMed, Scopus, Cochrane Library, EMBASE, Google Scholar, CINAHL, Web of Science, Science direct and African Journals Online) were searched using relevant search terms. Data were extracted on an excel sheet and Stata/ SE 14.0 software was used to perform the meta-analysis. Heterogeneity of included studies were assessed using I 2 and Q test statistics. I2 > 50% and Q test with its respective p-value < 0.05 were suggestive for the presence of a significant heterogeneity. Publication bias was assessed using the Egger's regression test and funnel plot. Subgroup and sensitivity analyses were done. A random effects model was used to estimate the pooled prevalence of GDM and odds ratio (OR) with 95% confidence interval (CI). Result: A total of 23 studies were included in the final analysis. The pooled prevalence of GDM in Africa was 13.61% (95% CI: 10.99, 16.23; I2 = 96.1%), and 14.28% (95% CI, 11.39, 17.16; I2 = 96.4%) in the sub-Saharan African region. The prevalence was highest in Central Africa 20.4% (95% CI, 1.55, 38.54), and lowest in Northern Africa 7.57% (95% CI, 5.89, 9.25) sub- regions. Overweight and obesity, macrosomia, family history of diabetes, history of stillbirth, history of abortion, chronic hypertension and history of previous GDM had positively associated with GDM. Conclusions: The prevalence of GDM is high in Africa. Being overweighed and/or obese, ever had macrocosmic baby, family history of diabetes, history of stillbirth, history of abortion or miscarriage, chronic hypertension and history of previous GDM were factors associated with GDM. Preventing overweighed and obese, giving due attention to women having high-risk cases for GDM in pregnancy are strongly recommended to mitigate the burden. The present review was registered with PROSPERO(2018:CRD42018116843) [27] and conducted according to the Preferred Reporting Items for Systematic Reviews and Meta Analyses (PRISMA) [28]. A protocol was developed during the planning process. A systematic review and meta-analysis was conducted using published articles on the prevalence and associated factors of GDM in Africa with the updated international diagnostic criteria. The databases used to search for studies were PubMed, Scopus, Cochrane Library, EMBASE, Google Scholar, CINAHL, Web of Science, Science direct and African Journals Online (AJOL). All potentially eligible studies were accessed through this searching strategy for “gestational diabetes mellitus OR hyperglycemia in pregnancy OR impaired glucose tolerance OR gestational hyperglycemia” AND “name of Africa countries” were used separately and in combination of the Boolean operators terms “OR” and “AND” as necessary. Likewise, terms like “determinant factors OR determinant variables OR associated factors” were used in combination with the above search terms. The search was also made by combining the above search terms with the names of all countries included in Africa and sub-region of Africa. A combination of expanded search term and free-text searches were used as shown in Table 4 in Appendix 1. Then the reference lists of the retrieved studies were also followed to access for additional articles and screened for its suitability to be recruited into this review. Any studies in Africa that reported prevalence and risk factors for GDM and fulfilled the following criteria were entered into the analysis, including the following factors: (1) being conducted in African countries classified by the United Nations Statistics Division [29]; (2) Epidemiological studies had reported prevalence and risk factors of GDM as primary results; (3) provided the prevalence and OR with 95% confidence interval (CI) or total of participants and number of GDM events (3) Being published in English language journals from January 1, 2013 to November 26, 2018; (4) studies conducted on pregnant women regardless of gestational age, sample size and study setting; and (5) studies used the updated international diagnostic criteria for GDM diagnosis was made by using the new 2013 WHO [1] or ADA [25] or modified IADPSG [30] diagnostic criteria. If an article failed to mention any of the above inclusion criteria it was excluded. In addition, studies were excluded if they were: (1) studies with poor definition of the outcome of interest; (2) qualitative studies, review articles, case reports, and case series regardless of the number of cases, narrative reviews, conference abstracts with no full information or if authors have not responded to our inquiry on the full text, editorials, commentaries, letters to the editor, author replies, and other publications that do not include quantitative data on the prevalence and/or associated factors of GDM; (3) studies presenting contradictory/unclear quantitative measures that could not be verified with authors; (4) duplicated studies on GDM ascertainment in the same population. In the case of duplicated publications, only the study containing the most important information in the context of prevalence and ascertainment methodologies or most recent results was included; and (5) studies including GDM patients with other metabolic disorders or other non-communicable diseases (NCDs) in the same category. Gestational diabetes mellitus was diagnosis, if one or more of the following abnormality are met, fasting plasma glucose 5.1–6.9 mmol/l (92–126 mg/dl), one-hour plasma glucose ≥10.0 mmol/l (180 mg/dl), 2-h glucose 8.5–11 mmol/l (153–199 mg/dl) after overnight fasting with 75 g glucose load [1, 25, 30]. WHO endorsed the modified IADPSG criteria by 2013 [31]. Relevant papers identified from the aforementioned databases and websites were imported into an EndNote X7, and duplicates were removed. Retrieved articles were assessed by two review authors with extensive experience in systematic reviews. Screening of titles abstracts and full text quality was conducted independently by two review authors (A & Y) on the bases of these inclusion criteria. Disagreement between the two reviewers was resolved by consensus or the third reviewer (O) made the decision regarding inclusion of the article in the final review. The selected papers were fully reviewed and the required information for the systematic review was extracted and summarized using an extraction table in Microsoft Office Excel software. The Preferred Reporting Items for Systematic Reviews and Meta-Analyses (PRISMA) guideline were followed throughout the review and analysis processes [32]. The study quality or risk of bias was assessed using the adopted a risk of bias tool developed by Hoy et al. [33] and modified it to suit to our study. The tool consists of ten items that assess sampling, attrition, measurement and reporting bias. The validity of methodology, appropriateness and reporting of results were also assessed (Table 5 in Appendix 2). When the information provided was not adequate to assist in making judgment for a certain item, we agreed to grade that item with a ‘NO’ meaning high risk of bias. Each study was graded depending on the number of items judged ‘YES’ as low (≥ 8), moderate (6 to 7) or high risk of bias (≤ 5) (Table 6 Appendix 2). The main findings regarding the prevalence and risk factors for GDM were summarized by two authors, and excel sheet was prepared under subheadings agreed upon by all authors. Data were extracted from each study regarding name of author (s), country and sub-region, Study design, setting, year of publication, year of study conducted (year of survey), sample size, response rate, gestational age when GDM screen, participant selection, age of pregnant women, test approach (one step vs two step), screening criteria (Universal vs selective), Blood glucose levels measured by (Glucometer vs Laboratory method), prevalence of GDM (including percentage and 95% CI), odds ratio, relative risk of certain risk factors. The outcome measures extracted were prevalence of GDM and risk factors in terms of differences of proportion/percent of GDM in the total pregnant women were participated. The data were entered into Microsoft Excel, exported into STATA/SE version 14 software for analysis. The heterogeneity test of the included studies was assessed by using the I2 statistics and Q test with its respective p-value. The presence of heterogeneity was considered to I2 test statistics results > 50% [34, 35] and Q test and its respective P-value < 0.05. Furthermore, the heterogeneity was presumed in the protocol based on an estimate of a potential variation across studies and depicted in the analyses, we used a random effects model as a method of analysis [34]. The publication bias was assessed using the Egger‘s regression test objectively and funnel plot subjectively [36, 37]. Any asymmetry of a funnel plot and statistical significance of Egger’s regression test (P-value < 0.05) was suggestive of publication bias. Therefore, the Duval and Tweedie nonparametric trim and fill analysis using the random effect analysis was performed [38]. Forest plots used to present the combined prevalence and 95% confidence interval (CI). Subgroup analyses for prevalence were performed by sub regions of Africa, publication year of studies, quality of the study and study design. In addition, a sensitivity analysis was done to point out the study (s) that caused variation. The different factors associated with GDM were presented using odds ratios (ORs) with 95% confidence interval (CI).