Abstract

Background

Metabolic syndrome has become a major public health problem worldwide and is attributable to the spread of different non-communicable diseases such as type 2 diabetes mellitus, coronary artery diseases, stroke, and permanent or temporary disabilities. It is not a single disease entity but encompasses different risk factors. However, there were inconsistencies among previously conducted primary studies, hence this systematic review and meta-analysis aimed to determine the pooled prevalence and determinants of metabolic syndrome among type2 diabetes patients in Ethiopia.

Method

First-hand studies about the metabolic syndrome among adult type 2 diabetic patients in Ethiopia were searched through known and international databases (PubMed, Scopus, Web of Science, and Cochran Library) and search engines (Google and Google Scholar). Data was extracted using a standard data extraction checklist developed according to Joanna Briggs Institute (JBI). The I² statistics are used to identify heterogeneity across studies. Funnel plot asymmetry and Egger鈥檚 tests were used to check for publication bias. A random effect model was used to estimate the pooled prevalence of the metabolic syndrome among non-insulin-dependent patients in Ethiopia. The STATA version 11 software employed for statistical analysis was conducted using STATA version 11 software.

Result

The overall pooled prevalence of metabolic syndrome among type2 diabetic patients in was 54.56% [95%CI (43.73, 65.38), I²鈥�=鈥�97.0%, P鈥�=鈥�0.001] using NCEP-ATP III, 48.32% [95%CI (42.1, 54.44), I²鈥�=鈥�97.0%, P鈥�=鈥�0.001] IDF diagnosis criteria, 47.0[95%CI(27.01鈥�66.99)], I²鈥�=鈥�97.5%, p鈥�=鈥�0.001 using WHO and 59.37%(95%CI(47.21鈥�71.52), I²鈥�=鈥�91.2%, p鈥�=鈥�0.001 using harmonized diagnosis criteria respectively. This meta-analysis identified several significant predictors of metabolic syndrome among type 2 diabetes patients in Ethiopia. The odds of having metabolic syndrome was reduced for females (Adjusted Odds Ratio [AOR]鈥�=鈥�0.55, 95% CI: 0.35鈥�0.87) compared to males. However, the odds of metabolic syndrome increased with alcohol intake (AOR鈥�=鈥�1.44, 95% CI: 1.03鈥�2.01), the odds of living in urban areas(AOR鈥�=鈥�2.12, 95% CI: 1.55鈥�2.88), and the odds of having a diabetes duration of six or more years since diagnosis (AOR鈥�=鈥�2.94, 95% CI: 1.17鈥�7.41) were significant predictors.

Conclusion

The pooled prevalence was considerably high among type 2 diabetic patients in Ethiopia. The pooled prevalence of metabolic syndrome varies as per the diagnosis criteria used with the highest observed in harmonized diagnosis criteria. Being female, being rural residency, alcohol intake, and duration of diabetes since diagnosis were significant predictors of metabolic syndrome among type 2 diabetic patients in Ethiopia.

Introduction

Metabolic syndrome (Mets) is non-communicable; a group of metabolic risk factors encompasses central obesity, insulin resistance, hypertension, and atherogenic dyslipidemia which strongly contribute to developing cardiovascular disease and non-insulin-dependent diabetes mellitus [1, 2]. It is not a single disease also known as syndrome X, with the constellation of cardiovascular disease risk factors and has been defined slightly differently by various organizations [3, 4]. The three most common organizational definitions used for the diagnosis of metabolic syndrome are the World Health Organization (WHO), the National Cholesterol Education Program (NCEP), and the International Diabetes Federation (IDF) [3, 5]. This non-communicable disease (NCD) has become a major public health problem worldwide; attributable to the spread of different non-communicable diseases such as type 2 diabetes mellitus, coronary artery diseases, stroke, and permanent or temporary disabilities [6, 7]. Human behaviour such as a sedentary lifestyle, high calorie and low fibre fast food consumption, and physical inactivity due to high access to modern transportation are determinants of an increased metabolic syndrome [3, 8, 9].

According to IDF the global prevalence of diabetes is expected to increase from 415听million in 2015 to 642听million by 2040 [10]. The pooled prevalence of Metabolic syndrome in Sub-Saharan Africa among the general population according to IDF, NCEP-ATP III, and WHO diagnosis criteria was 18.0%, 17.1%, and 11.1% respectively [5]. Furthermore, the pooled prevalence of metabolic syndrome in Africa was 32.4% and according to IDF, NCEP-ATP III, and WHO diagnosis criteria 29.3%, 33.1%, and 44.8%, respectively among the general population [11].

Moreover, the pooled prevalence of metabolic syndrome was 30.0% in Ethiopia among the general population [9]. Furthermore, Ethiopia is an especially relevant country to study metabolic syndrome among individuals with Type 2 diabetes (T2DM) due to several unique social, economic, and health-related factors that influence the prevalence and impact of metabolic disorders [12]. Ethiopia is undergoing an epidemiological transition characterized by a shift from predominantly communicable diseases to a growing prevalence of non-communicable diseases (NCDs), including T2DM [13]. As urbanization increases, so does exposure to sedentary lifestyles and high-calorie diets, both of which are major risk factors for Type 2 diabetes and metabolic syndrome. Besides, the healthcare infrastructure in Ethiopia faces challenges in screening, diagnosing, and managing chronic diseases [14]. This results in a large number of undiagnosed or inadequately managed cases of diabetes, increasing the risk of metabolic syndrome and its complications [15].

However, there is little study conducted on the pooled prevalence of metabolic syndrome and its determinants among type 2 diabetes patients in Ethiopia using different diagnosis criteria. Therefore, this systematic review and meta-analysis aimed to determine the pooled prevalence and associated factors of metabolic syndrome among type 2 diabetes patients in Ethiopia.

Materials and methods

Registration and report of the study protocol

This systematic review and meta-analysis have utilized the guidelines developed by the Joanna Briggs Institute (JBI) for Systematic Reviews [16] and the report is written consistent with the revised 2020 PRISMA guidelines [17]. This systematic review and meta-analysis title and its protocol have been registered in the PROSPERO online database (with registration number CRD42024512769). The result of this review presentation was consistent with the standard preferred Reporting Items [17] for the Systematic Review and Meta-analysis (PRISMA) checklist (Additional File S1).

Searching strategies

The known and international databases (PubMed, Scopus, Web of Science, and Cochran Library) and search engines (Google and Google Scholar) were used to locate research articles on the prevalence of metabolic syndrome and its determinants among type 2 diabetic patients in Ethiopia. The string for searching was developed using 鈥淎ND鈥� and 鈥淥R鈥� Boolean operators with the keywords extracted from the Medical Subject Headings (MeSH) database. The search strategy was based on the research question of this review and utilized the CoCoPop (颁辞鈥�=鈥塁辞苍诲颈迟颈辞苍, 颁辞鈥�=鈥塁辞苍迟别虫迟, 笔辞辫鈥�=鈥塒辞辫耻濒补迟颈辞苍) model. The article locating strategy was through Prevalence OR magnitude OR 鈥淢etabolic Syndrome prevalence鈥� OR 鈥淢etabolic Syndrome magnitude鈥� OR 鈥淢etabolic Syndrome X鈥� OR 鈥淚nsulin Resistance Syndrome X鈥� OR 鈥�Dysmetabolic Syndrome X鈥� OR 鈥淢etabolic Cardiovascular Syndrome鈥� OR 鈥淐ardio-metabolic Syndrome鈥� AND 鈥渄iabetes patients鈥� OR 鈥渁dult diabetic patients鈥� OR 鈥淚nsulin non-dependent diabetes鈥� OR 鈥淭ype2 diabetic patients 鈥淎ND determinant* OR factor* AND *Ethiopia OR Ethiopia. This search strategy aimed to trace all reviewed (published) and unpublished primary studies. The list of all retrieved primary articles and systematic review and meta-analysis references were also screened or cross-referenced to get extra studies鈥攕ources of information range from electronic databases to direct contact with the principal investigator if mandatory. The first search through Pub Med, Cochran Library, Scopus, Web of Science, Google, and Google Scholar was done in December 2024. The final search for updating was conducted from January 15 / 2024 to February 15/ 12/2024. The publication date was used as a filter mechanism in which articles published from January 2014 to January 2024 were included in the current systematic Review and Meta-analysis study to generate the most recent evidence for the scientific community.

Eligibility criteria

In this review the primary articles were eligible if and only if (1) they reported diabetic patients (2) aged鈥夆墺鈥�18 years old (3) those articles reported the prevalence or magnitude of metabolic syndrome (4) observational studies (both analytical and descriptive cross-sectional) (5) articles published in English and only from Ethiopia were included, and (6) published from 2014 to 2024 were included in this particular systematic review and Meta-analysis. Nonetheless, after a thorough examination of the titles and abstracts using eligibility criteria, irrelevant studies such as conference reports, qualitative papers, and published in languages other than English were excluded.

The study selection and outcome

After comprehensive searching, all located citations were selected and exported to Endnote citation manager software version X7. Following this, irrelevant and duplicated articles were removed. Then three independent researchers (HKA, CKM and AWA) screened each particular article for its title, abstract, and full text by far and cross-checked it against the inclusion criteria. The other research team (CKM, AFZ, and AWA) checked the screened articles with full text for details by already defined criteria to take it to the final review process. Any sort of disagreement between the research team while including and excluding articles on predefined criteria of this particular review was resolved by a thorough discussion of the team. The exclusion of the articles was presented with countable reasons which could be consistent with the pre-defined criteria. The result of searching further screening and inclusion process of articles in this review was done in agreement with the PRISMA guidelines for Systematic Review and Meta-analysis 2020.

Quality appraisal of included studies

The methodological quality of the included studies was critically appraised by three independent reviewers (AWA, CKM and HKA) by using the standardized JBI critical appraisal tool for the studies reporting prevalence [16]. This critical appraisal tool consists of 9 items designed to oversee the study population, sample size adequacy of the study, study subject and setting, reliability of condition or problem measurement, the appropriateness of the Statistical test used to analyze the data, and adequacy of the response rate of each selected article for the review process (Additional File Table S1). Each item has an answer of 鈥淣o鈥�, 鈥淵es鈥� or 鈥渦nclear鈥� to rate its risk of bias (ROB). After the critical appraisal, the reviewers decided to include or exclude screened articles based on the overall quality of the appraisal score in which those articles with the lowest score out of 9 were considered as poor quality or high risk of bias. The article was prone to exclude when the score was below average that is 4.5 of the three independent reviewers. In this regard, there had to be more than 3 鈥淣o鈥� or unclear鈥� quality categories for the article to be excluded from the review. This critical appraisal threshold was supported by a previously published systematic review and meta-analysis study [18]. Any sort of disagreement between the involved reviewers was solved through the discussion of the reviewers. Furthermore, if the disagreement unfolds, the fourth reviewer was indicated to oversee the source of the doubt and reach to consensus.

Data extraction

Data were independently extracted by five authors using a standardized data extraction format that was developed according to the 2014 Joanna Briggs Institute Reviewers鈥� Manual [19]. The tool includes Authors, Region, study year, study design, sample size, prevalence of metabolic syndrome among type 2 diabetic patients, data collection method, factors associated with metabolic syndrome, response rate, and risk of bias assessment score included in the extraction. The data were extracted by two independent reviewers and any inconsistent data was cross-checked (Additional File S2 File). The disagreement between the reviewers was solved by a thorough discussion.

Data synthesis and analysis

The outcome of included primary studies was narratively presented and expanded with supplementary materials in text, tables, and figures where necessary. All necessary and relevant information from every article was extracted through a Microsoft Excel spreadsheet, and exported to STATA Version 11 for further analysis. The random-effects model was employed to estimate the pooled effect size of health-related quality of life of epilepsy patients due to the presence of heterogeneity. Heterogeneity was identified by using the standard chi-square and I² statistical tests. The variation between different study characteristics such as the region where the primary article was conducted, outcome ascertainment tool, and year of publication were investigated through subgroup analyses. This subgroup analysis could demonstrate the sources of heterogeneity and let the researcher for another remedy such as the use of meta-regression to treat this heterogeneity. Moreover, the publication bias was assessed through visual inspection of the funnel plot, Begg- Mazumdar Rank correlation tests and Egger鈥檚 test to see the funnel plot鈥檚 asymmetry. The influence of individual articles on the overall pooled prevalence or effect size was assessed by using sensitivity analysis. The forest plot with 95% CI was used to present the overall pooled prevalence of metabolic syndrome as well as the subgroup pooled prevalence by metabolic syndrome diagnosis criteria among diabetic patients in Ethiopia. Moreover, the bivariate and multivariate meta-regression analysis was conducted to identify the potential covariates that brought about the variations.

Random effects model

A random-effects model was employed in this study due to the high heterogeneity observed across the included studies. High heterogeneity, indicated by metrics such as a high I² statistic, suggests substantial variability in effect sizes that cannot be attributed merely to chance [20]. Instead, this variability is likely due to genuine differences in study characteristics, methodologies, or underlying population factors. Moreover, the random-effects model accommodates variability and assumes that the true effect size varies between studies rather than being constant [21]. It also gives broader generalizability by accounting for between-study variation; the random-effects model provides a pooled estimate that reflects the average effect across a range of contexts, making it more generalizable to varied populations. Unlike a fixed-effects model, which weights studies solely based on within-study variance, the random-effects model incorporates both within-study and between-study variances. This prevents studies with large sample sizes from disproportionately influencing the pooled estimate when heterogeneity is high [20].

The article selection and outcome

In this systematic review and meta-analysis study we reviewed a total of (695) articles related to the prevalence and associated factors of metabolic syndrome among type 2 diabetic (DM2) patients in Ethiopia. We located these articles using electronic databases and search engine websites. Among overall articles, 386 were removed for being irrelevance and duplicated and the other pretty sizable articles were removed for not being ineligible (study design and Title difference) by automation tools and other reasons (192 vs. 64) respectively. The remaining 53 articles were eligible for screening. Of these screened 31 papers were excluded due to the country of study or not being conducted in Ethiopia and target population difference (those articles conducted among children). With further screening, 22 articles were sought for retrieval and 6 were not retrieved for one and the other reason. Moreover, 16 research articles were assessed for eligibility to be included in the review process, but with the outcome of interest and measurement tool ambiguity a total of 5 articles were excluded. Finally, 11 original research articles were incorporated with the systematic review and meta-analysis (Fig.听1).

Fig. 1

PRISMA flow diagram of the included studies

The methodological quality assessment of included studies

There are a total of 11 articles assessed for methodological quality using a 9-point score tool developed by JBI for observational prevalence studies. The outcome of the quality appraisal ranged from moderate to high methodological quality in which five studies [22,23,24,25,26] scored 9 points, four studies [6, 27,28,29] scored 8 points, and the remaining two studies [30, 31] scored 7 points (Additional File Table S1).

Description of included studies

In the current systematic review and meta-analysis study, the articles retrieved were from three regions and four regions of the country. The smallest sample size was 159 from the Amhara region of Ethiopia [32]. Whereas, the largest sample size was 419 from the study conducted in the Tigray region, Ethiopia [26].

In this systematic review and meta-analysis, we tried to pool the prevalence of metabolic syndrome using different diagnosis criteria such as IDF, NECP-ATP III, WHO, and harmonized criteria. We found seven studies [22,23,24,25,26, 31, 32] with a total sample of 1919 study participants who reported the prevalence of metabolic syndrome using IDF diagnosis criteria. Of these three were from Amhara region, Ethiopia, two from Oromia region, Ethiopia, and one each for Tigray and South region, Ethiopia. (Table听1)

Table 1 Characteristics of included studies for the prevalence of metabolic syndrome using IDF diagnosis criteria

Moreover, nine articles [22,23,24,25, 27,28,29, 31, 32] reported the prevalence of metabolic syndrome using NECP-ATP III, diagnosis criteria. Of the nine articles four were from the Southern region, Ethiopia, three from the Amhara region, Ethiopia, and the rest two from the Oromia region, Ethiopia (Table听2).

Table 2 Characteristics of included studies for the prevalence of metabolic syndrome using NCEP-ATP III diagnosis criteria

Three articles [22,23,24] were used by WHO with a total sample of 900 where two studies were located in the Amhara region, Ethiopia and one South region, Ethiopia. Two articles [23, 28] were located in South region Ethiopia which used the Harmonized diagnosis criteria with a sample of 708 study participants (Table听3).

Table 3 Characteristics of included studies for the prevalence of metabolic syndrome using WHO and harmonized diagnosis criteria

Prevalence of metabolic syndrome using NCEP-ATP III diagnosis criteria in Ethiopia

A total of 12 primary articles were appraised and retrieved to pool the overall prevalence of metabolic syndrome among type 2 diabetic patients in Ethiopia. However, we did not get the overall prevalence of metabolic syndrome in the reviewed articles, all the authors rather reported the outcome using IDF, NCEP-ATP III, WHO, and harmonized diagnosis criteria. Hence, we have pooled similar outcomes based on the diagnosis criteria, and therefore, based on NCEP-ATP III diagnosis criteria we had pooled nine articles. The prevalence of metabolic syndrome in individual articles ranged from 27.7 to 70.30% in the South region, and Amhara, region, Ethiopia respectively. The overall pooled prevalence of metabolic syndrome among type 2 diabetic patients was 54.56% [95%CI (43.73, 65.38), I²鈥�=鈥�97.0%, P鈥�=鈥�0.001]. The effect size of the overall pooled prevalence using NCEP-ATP III diagnosis criteria among type 2 diabetic patients in Ethiopia was presented using a forest plot (Fig.听2).

Fig. 2

Forest plot of the pooled prevalence of metabolic syndrome using NCEP-ATP III diagnosis criteria among type 2 diabetic patients in Ethiopia

Subgroup analysis by region

We conducted a sub-group analysis based on the region where the primary articles were located. Hence, the lowest pooled prevalence of metabolic syndrome using the random-effects model was 41.25% [95%CI (36.66, 45.85), I²鈥�=鈥�0.00%, p鈥�<鈥�0.983] in the Oromia region, Ethiopia followed by 53.04% [95%CI (32.64, 73.39), I²鈥�=鈥�98.4%, p鈥�<鈥�0.001] in South region, Ethiopia. Whereas, whereas, the highest sub-group pooled prevalence of metabolic syndrome was 65.36% [95%CI (58.44, 72.28), I²鈥�=鈥�74.8%%, p鈥�<鈥�0.019] in Amhara region, Ethiopia (Fig.听3).

Fig. 3

Forest plot of the subgroup analysis of the regional prevalence of metabolic syndrome using NCEP-ATP III diagnosis criteria among type 2 diabetic patients in Ethiopia

Prevalence of metabolic syndrome using IDF diagnosis criteria in Ethiopia

In a systematic review and meta-analysis, a total of seven primary articles with IDF diagnosis criteria were appraised and retrieved to pool the overall prevalence of metabolic syndrome among type 2 diabetic patients in Ethiopia. Hence, the prevalence of metabolic syndrome in individual articles ranged from 31.40 to 57.00% in Oromia and Amhara, regions, Ethiopia respectively. The overall pooled prevalence of metabolic syndrome using the IDF diagnosis criteria among type 2 diabetic patients was 48.32% [95%CI (42.1, 54.44), I²鈥�=鈥�97.0%, P鈥�=鈥�0.001]. The effect size of the overall pooled prevalence using IDF diagnosis criteria among type 2 diabetic patients in Ethiopia was presented using a forest plot (Fig.听4).

Fig. 4

Forest plot of the pooled prevalence of metabolic syndrome using IDF diagnosis criteria among type 2 diabetic patients in Ethiopia

Subgroup analysis by region using IDF diagnosis criteria

We conducted a sub-group analysis based on the region where the primary articles were located. Hence, the lowest pooled prevalence of metabolic syndrome using the random-effects model was 36.61% [95%CI (26.42, 46.81), I²鈥�=鈥�80.7%, p鈥�<鈥�0.001] in the Oromia region, Ethiopia followed by 51.10% [95%CI (46.31, 55.89), I²鈥�=鈥�0.0001%, p鈥�<鈥�0.001] in Tigray region, Ethiopia. Whereas, whereas, the highest sub-group pooled prevalence of metabolic syndrome was 53.40% [95%CI (49.25, 57.54), I²鈥�=鈥�23.7%%, p鈥�<鈥�0.270] in Amhara region, Ethiopia (Fig.听5).

Fig. 5

Forest plot of the subgroup analysis of the regional prevalence using IDF diagnosis criteria among type 2 diabetic patients in Ethiopia

Bias assessment

We conducted a thorough assessment of potential publication bias using multiple methods to ensure the validity of our meta-analysis results. Funnel plot analysis, which visually represents the distribution of study results relative to their precision, revealed symmetric plots (Figs.听6 and 7) for NCEP-ATP III diagnosis criteria and IDF diagnosis criteria respectively. Symmetry in the funnel plots suggests an even distribution of study results around the pooled effect size, indicating minimal evidence of publication bias. However, it is essential to acknowledge that the reliability of this visual inspection can be limited by the number of included studies. Small sample sizes, in particular, may reduce the statistical power to detect asymmetry, which is a potential indicator of bias. To reinforce these findings and assess robustness, we performed a trim-and-fill analysis. This method estimates the number of potentially missing studies due to publication bias and adjusts the pooled effect size accordingly. Our trim-and-fill analysis did not identify any missing studies or alter the pooled estimate, providing further confidence that publication bias did not significantly influence our results.

Fig. 6

Funnel plot analysis of the metabolic syndrome using NCEP-ATP III diagnosis criteria among type 2 diabetic patients in Ethiopia

Fig. 7

Sensitivity analysis of included studies for the metabolic syndrome using NCEP-ATP III diagnosis criteria among type 2 diabetic patients in Ethiopia

Additionally, we conducted Begg鈥檚 test, a rank correlation method that evaluates the relationship between study effect sizes and their variances. This statistical test did not reveal any significant correlation, as indicated by a p-value of 0.675, suggesting the absence of publication bias. Similarly, we applied Egger鈥檚 test, a regression-based approach that examines the relationship between the standardized effect sizes and their standard errors. With a p-value of 0.755, Egger鈥檚 test provided no evidence of significant asymmetry in the funnel plot, further supporting the minimal risk of publication bias in our meta-analysis.

Sensitivity analysis

We have conducted a sensitivity analysis to identify whether there is evidence of the influencing effect of one study on the other. The output of leave-one-out sensitivity analysis through the random-effects model revealed that there was no individual study that influenced the overall pooled prevalence of metabolic syndrome among type 2 patients in this particular review. For each single study, the effect size indicated relates to the overall pooled effect size generated from meta-analysis omitted that particular study (Additional Files: Figures听5 and 9).

Fig. 8

Funnel plot analysis of using IDF diagnosis criteria among type 2 diabetic patients in Ethiopia

Trim and fill analysis

We have conducted a Trime and fill analysis that is shown in (Figs.听8 and 9) for using NCEP-ATP III diagnosis criteria and metabolic syndrome using IDF diagnosis criteria respectively.

Fig. 9

Sensitivity analysis of included studies using IDF diagnosis criteria among type 2 diabetic patients in Ethiopia

Determinants of metabolic syndrome

We have found the IDF and NECP-PAT-III reported determinants of metabolic syndrome. Hence, we conducted pooled effect size (OR) for variables sex, age, residency, duration of diabetes since diagnosis, alcohol intake, body mass index (BMI/), and physical exercise as a determinant of Metabolic Syndrome among type 2 diabetic patients. We found sex, alcohol intake, and did not have physical exercise significant predictors of metabolic syndrome among type 2 diabetic patients in Ethiopia using IDF as diagnosis criteria. In this regard, seven [22,23,24,25,26, 30, 32] articles reported sex as a determinant were pooled and the pooled effect size of the analysis revealed that females had 45% lower odds (Adjusted Odds Ratio [AOR]鈥�=鈥�0.55, 95% CI: 0.35鈥�0.87) of metabolic syndrome compared to males, after adjusting for other factors in the model. Four studies [17, 19, 20, 26] identified alcohol intake as a significant predictor, showing that individuals who consumed alcohol had 1.44 times higher odds of developing metabolic syndrome compared to non-drinkers (AOR鈥�=鈥�1.44, 95% CI: 1.03鈥�2.01). Additionally, six studies [16,17,18,19,20, 26] reported physical exercise as a predictor, and the pooled effect size indicated that individuals who did not engage in physical exercise were 1.86 times more likely to develop metabolic syndrome than those who exercised (AOR鈥�=鈥�1.86, 95% CI: 1.43鈥�3.55).

Predictor variables associated with metabolic syndrome using the NECP-PAT-III diagnostic criteria in Ethiopia were also analyzed. Five studies [6, 16,17,18, 21] were pooled and the effect size revealed that females had 0.44 times (AOR鈥�=鈥�0.44, 95% CI: 0.29鈥�0.68) the odds of metabolic syndrome compared to males. The residency was examined in three studies [16, 21, 22], which found that individuals living in urban areas had 2.12 times higher odds of developing metabolic syndrome compared to those in rural areas (AOR鈥�=鈥�2.12, 95% CI: 1.55鈥�2.88). Lastly, three studies [16, 17, 21] examined diabetes duration as a predictor, with a pooled effect size showing that patients with a diabetes duration of six or more years since diagnosis had 2.94 times higher odds of developing metabolic syndrome compared to those with a shorter duration (AOR鈥�=鈥�2.94, 95% CI: 1.17鈥�7.41). (Table听4).

Table 4 Determinants of metabolic syndrome among Type2 diabetic patients using IDF and NECP-ATP III diagnosis criteria in Ethiopia

Discussion

The prevalence of metabolic syndrome among type2 diabetic patients in Ethiopia found 54.56% [95%CI (43.73, 65.38), I²鈥�=鈥�97.0%, P鈥�=鈥�0.001], 48.32% [95%CI (42.1, 54.44), I²鈥�=鈥�97.0%, P鈥�=鈥�0.001], 47.0% [95%C (27.00, 66.95), I²鈥�=鈥�97.0%, P鈥�=鈥�0.001], and 59.37%[95%CI(47.21, 71.52), I²鈥�=鈥�97.0%, P鈥�=鈥�0.001] using NCEP-ATP III, IDF, WHO, and harmonized diagnosis criteria respectively. The subgroup analysis by region revealed significant variation in the prevalence of metabolic syndrome among diabetic patients. This might be due to population socio-cultural and health access differences across the regions of Ethiopia. For instance, the Amhara region has the highest prevalence of unmet health needs of the community as a result of human-made problems such as war. The finding was consistence with the study results of 50.2% and 53.9% in India [33] and 53.1% in Mexico [34] of the study conducted in India using NCEP-ATP III and IDF diagnosis criteria respectively. It was also consistent with study results of 45.8% in India [35] and 43.83 in Ghana [36] using NCEP-ATP III diagnosis criteria. However, the finding was higher than the study results of 35.1% and 29.5% in Brazil [37] using IDF and NCEP-ATP III diagnosis criteria respectively. It was also higher than the study results of 36% and 31% in Mexico [38] using NCEP-ATP III and WHO diagnosis criteria respectively. The possible explanation might be due to socio-demographic and health-related literacy variations among Ethiopia, Brazil and Mexico, which had a lion鈥檚 share in self-care and risk minimization among diabetic patients. However, the current study result is lower than the study results of 68.6% in Ghana [39] and 58.00% in Ghana [40]. It was also lower than the study results of 73.4%, and 64.9% in Iran [41] using NCEP-ATP III and IDF diagnosis criteria respectively. Moreover, it was lower than the study result 57.7% in India [35] using IDF diagnosis criteria. The possible explanation might be the study in Ghana was only among both type 1 and type 2 diabetic patients living in two big districts, but the present study is the pooled prevalence results of four different regions. The current Systematic review and meta-analysis result is also lower than the study result of 70.4 in Iran [41] and 72.7% in India [33] using harmonized diagnosis criteria. Moreover, it was lower than the study results of 80% and 85% in Saudi [42] using WHO and NCEP-ATP III diagnosis criteria respectively. Furthermore, the finding of the current result is lower than the study results of 60.4% and 71.7% in South Africa [43] using NCEP-ATP III and IDF diagnosis criteria. Besides, the findings of a current systematic review and meta-analysis are lower than the results of 63.58% and 69.1% in Ghana [36] using WHO and NCEP-ATP III respectively. The variation might be socio-demographic, socio-economic, healthcare access and lifestyle factors; for instance, growing urbanization in South Africa and Ghana may have higher levels of urbanization compared to Ethiopia. Urban lifestyles are often associated with physical inactivity, unhealthy diets, and increased stress, all of which contribute to higher Mets prevalence. The study results of 66.2% and 58.4% in Nepal [44] using IDF and NCEP-ATP III diagnosis criteria were higher than the current. The possible explanation might be socio-demographic and lifestyle-related variations between patients in Nepal and Ethiopia. For example; Nepal may have higher rates of urbanization and associated sedentary lifestyles, compared to Ethiopia鈥檚 largely rural population. Urban lifestyles often lead to reduced physical activity and higher rates of obesity and Mets. Besides, the Nepalese diets may include more processed foods, high-calorie diets, or sugary foods, contributing to Mets risk. In contrast, Ethiopia鈥檚 traditional diet, which includes high consumption of whole grains and less processed food, may lower Mets prevalence. This is evidenced by studies conducted in the USA [45], Brazil [46] and China [47].

It is also supported by the study conducted in Ghana [48] in which a sedentary lifestyle such as smoking, weight gain, and physical inactivity were significant predictors of metabolic syndrome among type 2 diabetic patients.

Being female was identified as a significant predictor of metabolic syndrome (Mets) among T2DM patients in Ethiopia. Females had lower odds of developing metabolic syndrome compared to males, with odds of having metabolic syndrome decreased by 45% and 56% based on the IDF and NCEP-ATP III criteria, respectively. This finding aligns with studies conducted in Nepal and Ghana [44, 48, 49]. Furthermore, type 2 diabetic patients living in urban areas in Ethiopia were 2.12 times more likely [AOR鈥�=鈥�2.12, 95% CI (1.55, 2.88)] to develop metabolic syndrome compared to those residing in rural areas. The possible explanation might be because of majority of urban residents are always overscheduled and physically inactive as compared to rural. This has been supported by the studies conducted in Cameroon in which the population in urban had a lion鈥檚 share in the prevalence of metabolic syndrome [50]. Besides, the odds of having metabolic syndrome were 1.44 times higher [AOR鈥�=鈥�1.44, 95% CI: 1.03鈥�2.01] among those who consumed alcohol compared to those who did not. The possible explanation might be that alcohol is a serious risk factor for an unhealthy liver and impairs metabolism [51, 52]. The odds of having metabolic syndrome were significantly influenced by the duration of diabetes since diagnosis, with the odds of being diagnosed for 鈮モ��6 years being 2.94 times higher [AOR鈥�=鈥�2.94, 95% CI: 1.17鈥�7.41] compared to those with a shorter duration of diabetes. The possible explanation might be that as age increases the risk of getting metabolic disorders increases.

Among type 2 diabetic patients in Ethiopia, the odds of developing metabolic syndrome were 1.86 times higher [AOR鈥�=鈥�1.86, 95% CI: 1.43鈥�3.55] for those who did not engage in physical exercise compared to those who did engage in physical exercise. A possible explanation could be the effect of exercise on healthy metabolism and not exercise related to unhealthy metabolism related to cardiovascular and liver problems. This was supported by the studies conducted in Nepal and Spain [44, 53].

Strengths and limitations of the study

The strengths of this systematic review and meta-analysis lie in its comprehensive analysis, its identification of key determinants of Mets (Metabolic Syndrome), and its practical implications for improving public health awareness and long-term health outcomes. The study highlights the need for targeted interventions and increased awareness to reduce the prevalence of Mets and its related comorbidities. Through its robust evidence, the study contributes significantly to advancing knowledge in this field and offering actionable insights for healthcare providers and policymakers. However, the study has limitations such as exclusively including articles from Ethiopia, limiting its findings to the Ethiopian population. This narrow geographical focus may restrict the generalizability of the results to other regions or populations with different genetic, lifestyle, or environmental factors. As Mets prevalence and associated factors are influenced by global variations in healthcare access, socioeconomic conditions, and cultural practices, broader inclusion of studies from diverse regions could have enriched the analysis and provided more universal insights.

The inclusion of only English-language articles introduces the possibility of language bias. Research published in other languages, potentially containing relevant data, may have been excluded, which could limit the comprehensiveness of the findings. This restriction may inadvertently overlook valuable studies conducted in Ethiopia but published in local languages or non-English journals. The study did not present a pooled effect size for factors associated with health-related quality of life (HRQoL) due to limited data availability. This limitation prevents a comprehensive understanding of the broader determinants of HRQoL in individuals with Mets. Identifying these factors could have provided actionable insights into improving patient outcomes and tailoring interventions. The study did not adequately address why one diagnostic criterion (NCEP-ATP III) might result in a higher prevalence estimate, while another (IDF) might yield a lower prevalence. This gap limits the discussion on how differences in the thresholds, definitions, or components of the diagnostic criteria influence prevalence estimates. For instance: The NCEP-ATP III focuses on uniform cutoff points, which may overestimate Mets in populations with smaller body frames or lower baseline waist circumferences. The IDF includes ethnicity-specific waist circumference thresholds, potentially leading to more conservative prevalence estimates in certain populations. An exploration of these variations could have provided critical insights into the applicability and appropriateness of specific criteria in the Ethiopian context.

Implications of the study

This systematic review and meta-analysis study has crucial benefits for those who are involved in metabolic syndrome care directly and indirectly. For instance, it helps healthcare providers, policymakers, and program planners by providing current and up-to-date information regarding patients suffering from metabolic-related metabolic syndrome. It also helps to set off possible strategies to improve the quality of life of patients as well as families by providing insight information how to minimize other comorbid metabolic-related disorders. Hence, minimizes disease burden and a cost demanded from the patient, family, and the country at large.

Conclusion

This systematic review and meta-analysis reveals that more than half of individuals with type 2 diabetes are affected by metabolic syndrome, highlighting a substantial burden of coexisting metabolic abnormalities in this population. There was a significant variation in the prevalence of metabolic syndrome based on the diagnosis criteria utilized across the studies. Besides, being female, being rural residency, alcohol intake, and duration of diabetes since diagnosis were significant predictors of metabolic syndrome among type 2 diabetic patients in Ethiopia.