The Prevalence of Metabolic Syndrome in Ethiopian Population: A Systematic Review and Meta-analysis

Background The metabolic syndrome is a clustering of hyperglycemia/insulin resistance, hypertension, dyslipidemia, and obesity which are risk factors for cardiovascular disease, type 2 diabetes and stroke, and all-cause mortality. The burden of metabolic syndrome is emerging alarmingly in low- and middle-income countries such as Ethiopia; however, there is lack of comprehensive estimation. This study aimed to determine the pooled prevalence of metabolic syndrome in Ethiopia. Methods This systematic review and meta-analysis included original articles of observational studies published in the English language. Searches were carried out in PubMed, Google Scholar, and Africa Journals from conception to August 2020. A random-effects model was used to estimate the pooled prevalence of metabolic syndrome in Ethiopia. Heterogeneity was assessed using the I2 statistic. Subgroup analysis was also conducted based on sex/gender and study subjects. Egger's test was used to assess publication bias. Results Electronic and gray literature search retrieved 942 potentially relevant papers. After removing duplicates and screening with eligibility criteria, twenty-eight cross-sectional studies were included in this meta-analysis. The pooled prevalence of metabolic syndrome in Ethiopia was found to be 34.89% (95% CI: 26.77, 43.01) and 27.92% (95% CI: 21.32, 34.51) by using NCEP/ATP III and IDF criteria, respectively. The weighted pooled prevalence of metabolic syndrome was higher in females 36.74% (95% CI: 20.72, 52.75) and 34.09% (95% CI: 26.68, 41.50) compared to males 22.22% (95% CI: 14.89, 29.56) and 24.82% (95% CI: 18.34, 31.31) by using IDF and NCEP/ATP III criteria, respectively. Subgroup analysis based on the study subjects using NCEP/ATP III showed that the weighted pooled prevalence was 63.78%(95% CI: 56.17, 71.40), 44.55% (95% CI: 30.71, 52.38), 23.09% (95% CI: 19.74, 26.45), 20.83% (95% CI: 18.64, 23.01), and 18.45% (95% CI: 13.89, 23.01) among type 2 diabetes patients, hypertensive patients, psychiatric patients, HIV patients on HAART, and working adults, respectively. The most frequent metabolic syndrome components were low HDL-C 51.0% (95% CI: 42.4, 59.7) and hypertriglyceridemia 39.7% (95% CI: 32.8, 46.6). Conclusions The findings revealed an emerging high prevalence of metabolic syndrome in Ethiopia. Therefore, early intervention is required for the primary prevention of the occurrence of metabolic syndrome and the further reduction of the morbidity and mortality related to it.


Background
Metabolic syndrome (MetS) is a cluster of interrelated risk factors that have been associated with cardiovascular disease (CVD), stroke, diabetes mellitus, and other comorbidities [1,2]. Insulin resistance, obesity, dyslipidemia, and hypertension are considered to be the primary components of MetS [3,4]. e worldwide prevalence of MetS and noncommunicable chronic diseases in the adult population is on the rise [5]. It has been estimated that the prevalence of MetS ranges 20-25% of the adult population globally [6]. e epidemiologic nature of MetS is also emerging alarmingly and being common in Africa in contrast to the earlier trend of being considered rare. e increase in the prevalence of MetS in Africa is thought to be due to divergence from traditional African to Western lifestyles [7]. High prevalence of MetS has been reported in some sub-Saharan Africa countries such as in Morocco (16.3%) and South Africa (33.5%) [8].
e individual with MetS has 2-3 times higher chance of developing stroke and CVD than without MetS [9,10]. It has also a six-fold greater risk of developing type 2 diabetes (T2DM) [11]. T2DM has become one of the major causes of premature illness and death, mainly through the increased risk of CVD disease [12]. MetS is also associated with other comorbidities such as cancer, nonalcoholic fatty liver disease, and other reproductive disorder [13][14][15]. It is also suggested that mortality due to MetS is more than twice that without the syndrome [16]. e prevalence of MetS varies across different populations. MetS appear to be more common in the presence of comorbidities such as diabetes mellitus, hypertension, and HIV infection than their counterparts. Around 85% of those with diabetes have MetS in the United States in contrast to 25% of working adults in Europe and Latin America [17,18]. Likewise, the high prevalence of HIV infection in developing countries and concomitant antiretroviral therapy is also associated with the rise of MetS [19]. Todowede O. O. et al. have reported in a meta-analysis that MetS prevalence among people living with HIV was 21.5% in contrast to uninfected 12.0% [20]. MetS rates are rising in developing countries in an epidemic situation [7]. For instance, in Ethiopia, the prevalence of MetS among T2DM, hypertensive, and HIV patients is estimated to be high [21][22][23]. e pathogenesis of the MetS is multidirectional, but obesity-induced insulin resistance is the major pathway for the occurrence of MetS [24]. Similarly, increased risk of MetS and its components is attributed to antiretroviral therapy and other related factors [25]. Sociodemographic factors, behavioral factors, stress, family history of disease, and some anthropometric measurements have been reported as associated factors of MetS in different literatures [26][27][28]. For instance, a study conducted among Malaysian population has showed that increasing age and low physical activity have been associated with increased odds of MetS [29].
Due to the dramatic increase in urbanization, smoking, severe stress-related problems of poverty, nutrition transition, reduced physical activity, and over nutrition increasing on top of the already high prevalence of undernutrition, sub-Saharan Africa is facing a rapid escalation of MetS and noncommunicable chronic diseases and associated mortality [30][31][32].
In Ethiopia, several studies were conducted to assess the prevalence of MetS among different study subjects having great disparity and inconsistent findings. However, there is lack of comprehensive estimation of MetS in Ethiopia. Hence, this systematic review aimed to determine the pooled prevalence of MetS among the Ethiopian population of various study subjects. is will provide the necessary information for policymakers, clinicians, and concerned stakeholders in the country to provide an appropriate strategy and intervene in the control, prevention, and management of MetS.

Protocol and Search Strategy.
is systematic review and meta-analysis was reported according to the Preferred Reporting Items for Systematic Reviews and Meta-Analyses statement guideline. e study protocol was registered in the PROSPERO International Prospective Register of Systematic Reviews (CRD42018090944) [33]. An inclusive literature search was conducted to identify studies about the prevalence of MetS reported among the Ethiopian population of various study subjects. Both electronic and gray literature searches were carried out systematically. PubMed, African journal of Medline, and Google Scholar were used to retrieve data. e search terms were used separately and in combination using Boolean operators like "OR" or "AND." An example of keywords used in PubMed to select relevant studies was as follows:

Design.
is systematic review and meta-analysis was conducted to estimate the pooled prevalence of MetS based on the International Diabetes Federation [6] and National Cholesterol Education Program-Adult Treatment Panel III (NCEP/ATP III) criteria [34] (Table 1) among different study subjects in Ethiopia.

Exclusion and Inclusion Criteria.
Observational studies (all of them were cross-sectional studies) that described the prevalence of MetS among the indigenous Ethiopian population were included. All included studies were original research articles published in English and contained the minimum information (study participants and number of MetS cases). Moreover, studies in which MetS has been reported using (i) IDF criteria and/or (ii) NCEP/ATPIII were included. e full text of studies meeting these criteria was retrieved and screened to determine its eligibility. Whereas studies in which MetS was described on other than the Ethiopian population, nonoriginal research (such as review, editorial, and a letter or commentary), and unknown/unclear methods of how MetS was diagnosed were not included.

Study Selection and Quality Assessment.
Two reviewers (S.A. and M.M.) independently screened the titles and abstracts in the abovementioned databases to consider the articles in the full-text review. e quality of the studies was appraised using Joanna Brigg's Institute quality appraisal criteria (JBI) [35]. All selected articles were evaluated by using the JBI appraisal checklist. Studies that got 50% and above on the quality scale were considered to have good quality ( Table 2).

Data Extraction.
To collect relevant data, a data extraction tool was developed, and one reviewer (SA) was responsible for the extraction of data from the included studies. Information regarding authors, year of publication, the study population, type of study/study design, number of participants, gender, diagnostic criteria for defining Mets, region, study area, sampling techniques, and the prevalence of MetS in each study were collected. e extracted data were checked by the second reviewer (AE) for its accuracy and consistency. A third reviewer (YT) was also engaged where necessary.

Data Analysis.
e extracted data were entered into Microsoft Excel and analyzed using STATA version 14 (StataCorp. 2009. Stata Statistical Software. College Station, TX: StataCorp LP). A random-effects model was used to obtain an overall summary estimate of the prevalence across studies. Point estimation with a confidence interval of 95% was used. Sensitivity analysis was conducted to assess the role of each study in the final result by excluding each study one by one. e presence of publication bias was assessed by using Egger's test. Trim and fill method analyses have been conducted to obtain a bias-adjusted effect estimate. Heterogeneity across studies was checked by Cochran's Q statistic and I 2 statistics. Moreover, meta-regression has been conducted that represent linear predictions for the MetS prevalence as a function of published year. Subgroup analysis was performed based on sex/gender, and study subjects since there have unexplained significant heterogeneity.

Characteristics of Included Studies.
Out of 942 potential articles, 233 articles were found to be relevant to the topic of interest screened by title and abstract. en, 115 were found to be eligible for full-text assessment. Of these full-text screened articles, 28 of them comprising 20652 study participants were found to be eligible for meta-analysis (Table 3). Figure 1 showed the results of the search and reasons for exclusion during the study selection process. e prevalence of MetS was estimated based on the IDF and NCEP/ATPIII criteria among the Ethiopian population of various study subjects. Fourteen studies reported the prevalence of MetS based on both IDF and NCEP/ATPIII criteria [36-44, 48-50, 54, 56], seven studies based on NCEP/ATPIII criteria only [57][58][59][60][61][62][63], and again seven studies by IDF criteria only [45-47, 51-53, 55]. Table 3 presents the characteristics and outcomes of the reviewed studies.

Prevalence of MetS Using IDF and NCEP ATP III Criteria.
e random-effects model was applied since the heterogeneity index of the studies was significant. e pooled prevalence of metabolic syndrome in Ethiopia was found to be 34 (Table 4).
Moreover, meta-regression has been conducted that characterize the linear predictions for the prevalence of MetS as a function of the increment of published year (Figures 4  and 5). An increasing trend in pooled prevalence of MetS is observed during the 2010-2020 time period in both IDF and NCEP/ATP III criteria.

Prevalence of the MetS Components.
e prevalence of the individual components of MetS among the Ethiopian population varied considerably between studies (Table 5). In   Table 5.

Sensitivity Analysis.
Sensitivity analysis was performed to assess the effect of each study on the pooled estimated prevalence of MetS by excluding each study step-by-step from the analysis process based on the two given diagnostic criteria (IDF and NCEP/ATP III). e result showed that excluded studies led to no significant changes in the shared estimation of the prevalence of MetS (Figures 6 and 7).  (Figures 8 and 9). However, there was somehow a difference compared with our previous results indicating minimally impacted by publication bias.

Discussion
e current systematic review provides evidence of an estimated pooled prevalence of MetS among various study subjects of the Ethiopian population. According to this review, the combined pooled prevalence of MetS was 34.89% (95% CI: 26.77, 43.01) and 27.92% (95% CI: 21.32, 34.51) by using NCEP/ATP III and IDF criteria, respectively. ese results were slightly in line with the reports in Brazil (29.6%) [64], Bangladesh (30%) [65], Iran 31% [66], and the United States (33%) [67]. But the current pooled estimate was also higher than the prevalence estimated in Ghana (12.4%) [68], Philippine (11.9%) [69], and the global estimate (20-25%) [6]. e findings were also lower as compared to Palestine (37.0%) [   Journal of Obesity (43.4%) [71], and the population of Nepal (52.7%) [72]. e variation in reports might be due to the difference in the underling important risk factors such as urbanization, westernization of lifestyle including unhealthy diet and physical inactivity [73], mental stress due to economic, social, and cultural factors [74], genetics, proinflammatory condition [75], intrauterine growth retardation, and over nutrition increasing on top of the already high prevalence of under nutrition [76]. e prevalence of MetS was high (34.89%) when NCEP/ATPIII criteria was used compared to IDF criteria (27.92%) in the current meta-analysis. ese might be due to the relative flexibility of the NCEP/ATPIII in which central obesity is not the absolutely required criterion, unlike IDF [77]. However, it was different from a systematic review conducted among the Ghanaian population using these two criteria, NCEP/ATPIII (12.4%) and IDF criteria (21.2%) [68]. e discrepancy between the results of different studies could be attributed to the difference in abdominal obesity and waist circumference in different populations.
We observed high between-study heterogeneity across studies. In the subgroup analysis, the pooled prevalence of MetS using NCEP ATP III criteria was 63.78% among T2DM patients which was higher than the result of the rest of the study subjects, 18.45%, 20.83%, 23.09%, and 44.55% among working adults, HIV patients on HAART, psychiatric patients, and hypertensive patients, respectively. is is also the same using IDF criteria. ough hyperglycemia is one of the components of Mets that increased its prevalence, insulin resistance-linked obesity might be the aggravated reason for the high prevalence of MetS among T2DM patients. In T2DM patients, particularly when glycemic control is poor, lipoprotein lipase activity is reduced [78]. e reduced lipoprotein lipase activity added with poor dieting and lack of regular exercise might be the reason for the higher prevalence of MetS among T2DM patients [79].
In line with other reviews [66,80], the current metaanalysis demonstrated that the prevalence rate of MetS was higher in females compared to that in males. is has been shown in all the Middle Eastern countries, and the prevalence was much higher among women than men [81]. e       Journal of Obesity higher prevalence of MetS among the women is attributed to abdominal obesity, which is mainly due to low physical activity, higher birth rate, presence of estrogen receptors, and menopause [82]. Besides genetic variation, lifestyle might be the reason.
Low HDL-C was the most frequent individual component of MetS in the current meta-analysis which was similar to reports in Bangladesh (89%) [65], Colombia (62.9%) [18], and Venezuela (58.6%) [83]. However, hypertriglyceridemia was shown the second most prominent MetS component to the contrary high blood pressure report in Bangladesh (30%) [65]. e low concentration of HDL-C favors the accumulation of low-density lipoprotein in the blood vessels greatly due to its poor scavenging capacity of low-density        lipoprotein from the body. e coexistence of hypertriglyceridemia and low HDL-C are risk factors for the development of CVD. It is well-known that a low concentration of the HDL-C level is a strong independent risk factor for CVD [84]. As a limitation to this review, a point needs to be noted that there is no uniformity of MetS definitions and waist circumference cutoffs points. Besides, high heterogeneity has been observed. Furthermore, there has been publication bias.

Conclusion
In conclusion, our review summarizes that a large proportion of the Ethiopian population have MetS regardless of the study subjects, gender, and the definitions used. One fact is clear here.
MetS is growing at an alarming rate and is high in Ethiopia. erefore, policymakers, clinicians, and concerned stakeholders shall urge effective strategies in the control, prevention, and management of MetS.
Lower CI limit Estimate Upper CI limit

Data Availability
e data used to support the findings of this study are available from the corresponding author upon request.

Conflicts of Interest
e authors declare that they have no conflicts of interest.

Authors' Contributions
S.A. was involved in conception of the research protocol, design of the study, review of the literature, data extraction, and statistical analysis. S.A. and A.E. were involved in data analysis and interpretation and drafting of the manuscript. M.M. and A.W. were involved in data interpretation and review of the manuscript. S.A., Y.T., and B.B. were involved in data extraction and quality assessment. All authors critically revised the paper, and they agreed to be accountable for all aspects of the work.