Glycaemic control and its associated factors in patients with type 2 diabetes in the Middle East and North Africa: An updated systematic review and meta‐analysis

Abstract Aims To examine the patient‐related factors that have been linked to glycaemic control in people living with type 2 diabetes mellitus in Middle Eastern countries. Design A systematic review and meta‐analysis. Data Sources A computerized search was conducted using the databases MEDLINE (via PubMed and Ovid), EMBASE, Scopus and CINAHL to identify peer‐reviewed articles published in English between 1 January 2010 and 21 May 2020. On 28 June 2021, the search was updated with the same keywords and databases; however, no further relevant studies were identified. Review Methods Extracted data were analysed using Review Manager 5.4. Results The final sample consisted of 54 articles with a total of 41,079 participants. Pooled data showed an increased risk of inadequate glycaemic control in smokers [OR = 1.26, 95% confidence interval (CI): 1.05, 1.52; p = .010], obese patients (OR = 1.30, 95% CI: 1.10, 1.54; p = .002), patients with elevated waist to hip ratio (OR = 1.62, 95% CI: 1.16, 2.26; p = .004) and longer disease duration (OR = 2.01, 95% CI: 1.64, 2.48; p < .001). A lower risk of inadequate control was associated with physical activity (OR = 0.40, 95% CI: 0.24, 0.67; p < .001) and self‐management (OR = 0.49, 95% CI: 0.29, 0.82; p = .006). Conclusion These findings highlight the opportunity to address factors to improve glycaemic control. Further longitudinal studies are required to better understand these variations, to assess all predictors of glycaemic control in participants with type 2 diabetes, and to provide a strong basis for future measures to optimize glycaemic control.

of patients with diabetes in the MENA region will reach 76 million by 2030 (Williams et al., 2019). Effective management of diabetes is, therefore, a priority for the countries of this region.
If not properly controlled, diabetes and its complications can result in frequent hospital admissions and premature death. Glycaemic control is a fundamental determinant of the prognosis of diabetes (Mamo et al., 2019). As a result, it is targeted by physicians to prevent diabetesrelated complications and mortality (Khattab et al., 2010). A wide variety of factors are associated with glycaemic control among participants with type 2 diabetes mellitus (T2DM); these include demographic factors such as age, gender, marital status and education (Abdullah et al., 2020;Aghili et al., 2016). Lifestyle parameters have also been linked with T2DM, such as smoking, obesity and physical inactivity (Al Slamah et al., 2020). In addition, correlations have also been noted between psychological status (e.g. depression and anxiety) and glycaemic status (Ahmadian et al., 2018;Al Hayek et al., 2014). Many of these factors have the potential to be influenced by the local context and cultural considerations. However, little information is available to inform priorities for the MENA region and clinical guidelines to direct healthcare professionals on how this information may be used to improve health and prevent or delay the onset of the complications of diabetes.

| Aims
Given the high and rising prevalence of T2DM in the MENA region, and to minimize cultural and geographic heterogeneity, this systematic review and meta-analysis was conducted to examine the patient-related factors that have been linked to glycaemic control in people living with T2DM in the MENA region, to make recommendations for locally appropriate diabetes care.

| Design
This systematic review and meta-analysis followed the Preferred Reporting Items of Systematic Reviews and Meta-analysis (Moher et al., 2016). Review methods and analyses were conducted in accordance with international recommendations (Agency for Healthcare Research and Quality, 2014;Higgins et al., 2019). The protocol of this study has been registered in PROSPERO (https:// www.crd.york.ac.uk/prosp ero/; ID:CRD42021227250). lists of included and any review articles were checked for potentially relevant studies. MESH terms were applied, and synonyms identified and used as keywords, linked with Boolian operands, as follows:

Impact
What problem did the study address?
Rising rates of diabetes, particularly in Middle Eastern regions, make it imperative to identify factors linked to glycaemic control among patients with Type 2 Diabetes in these areas.

| Eligibility criteria
To be included in the review, studies had to meet the criteria set out in the PEO framework. Studies were also required to be observational in design (case-control, cohort or cross-sectional studies).
Studies that were not available in English; animal studies; casereports or not full reports of primary research (e.g. conference abstracts and review articles) were excluded.

| Search outcomes
Citations were downloaded into EndNote x9, and duplicate studies were removed using the "remove duplication" function. Two independent reviewers (O. H. A. and S. J.) screened the retrieved articles in two steps. First, only the titles and abstracts were independently screened to remove papers not meeting eligibility criteria. Then, the full-texts of the remaining papers were retrieved for full-text screening of their eligibility for inclusion. Full-text screening focused predominantly on the methods and results sections to confirm the eligibility of the study and provision of relevant data. Any discrepancies between reviewers' decisions were addressed in discussion of the eligibility criteria with the third reviewer (L. P.).

| Data abstraction
A standardized extraction form was prepared using MS excel (Microsoft), including the domains shown in Table 1. Data extraction was done independently by two investigators (O. H. A. and S. J.), with any discrepancies resolved in discussion with a third reviewer (L. P.). Missing odds ratios were calculated with review manager 5.3 using the required data (event and total of experiment and control groups). Missing standard deviations (SD) were calculated from the standard error (SE) with the following formula: and from the confidence interval (CI) using this formula: The Newcastle-Ottawa Scale (NOS) for observational studies (Case-control, Cohort, and Cross-sectional studies) was used (Wells et al., 2014) to critically appraise all included studies (see Supplementary File 1).

| Synthesis
Data were analysed in three forms based on our objectives: 1. We performed a generic inverse variance analysis using the logarithmic odds ratio (LogOR) and SE to identify the association between glycaemic control and characteristics reported as associated with glycaemic control in bivariate analyses, such as age, gender, marital status, education, mental status, physical activity, smoking, obesity and self-monitoring.
2. We used the Inverse Variance (I-V) model to calculate the mean difference (MD) between adequate (HbA1c <7) and inadequate (HbA1c equal to or more than 7) levels of glycaemic control according to participants' characteristics such as age, Body mass index (BMI) and duration of diagnosis (ADA, 2021).

| Sensitivity analysis and publication bias
According to Egger and Smith (1998), publication bias assessment is relicable where ≥10 studies are pooled. Therefore, we assessed the publication bias using funnel plots and the Egger test. We performed a sensitivity analysis to ensure that none of the included studies affected the results and whether the overall effect size was statistically robust. We excluded one study in each scenario to determine the effect of each study on the overall effect size (Agency for Healthcare Research and Quality, 2014).

| Study selection
The literature search of databases yielded 3345 records. Following the title/abstract and full-text screening, 54 articles with a total of 41,079 participants were retained for inclusion in the systematic review. Of these, 32 articles reported adequate and relevant data for inclusion in the meta-analysis. The flow of the study selection process is shown in Figure 1.

| Study characteristics
Of the included studies, 17 articles were conducted in Saudi Arabia, seven each in Turkey and Iran, six in Jordan, three each in SD = SE × √ N, SD = √ N × (upper limit − lower limit) ∕ 3.92.

| Quality of included studies
The quality of included studies, assessed using the NOS tool, ranged from good to very good for cross-sectional studies, and was deemed very good for all included cohort studies participants ( In studies conducted in Palestine, older participants were associ-

| Physical activity
Overall random-effect estimates of eight studies showed a significant association between the lower risk of inadequate glycaemic control and physical activity (OR = 0.40, 95% CI: 0.24, 0.67; p = .0005). Overall pooled data were heterogenous (I 2 = 80%, p < .0001), see Figure 9. Heterogeneity was best resolved by ex-

| Diabetes knowledge
The pooled analysis of four studies showed no significant association between diabetes knowledge and glycaemic control (OR = 0.67, TA B L E 2 The Newcastle-Ottawa Scale (NOS) of Cross-sectional studies and Cohort studies Cohort studies

Representativeness of the exposed cohort
Selection of the nonexposed cohort

Ascertainment of exposure
Demonstration that outcome of interest was not present at start of study

Subgroup analysis showed no significant association in Saudi
Arabia (  One suggested explanation for these observations is that older age is associated with a reduction in the lifespan of the red blood corpuscles, reducing their exposure to circulating glucose, thereby lowering blood levels of HbA1c (Dubowitz et al., 2014;Shperling & Danon, 1990). However, the overall odds ratio in our study

F I G U R E 3 Forest plot of the odds ratio of inadequate glycaemic control by gender
showed no association between these variables. Similarly, Wiener & Roberts (1999) found no significant correlation in 399 participants between age and HbA1c levels. The lack of significance was ascribed to the exclusion of participants who had impaired glucose tolerance yet were eligible for inclusion in the normoglycaemic group based on a fasting plasma glucose value <6.4 mmol/L, a criterion that does Subgroup analysis surprisingly showed a significant inverse relationship between age and HbA1c among Palestinian participants.
This relationship can perhaps be attributed to differences in attitudes among young participants compared with their elders. Younger individuals are often regarded or observed to be more negligent of the hazards of diabetes and the necessary self-management activities to control their glucose levels; conversely, older participants tend to be more careful and adherent to their treatment regimen (Toh, 2011). Radwan et al. (2018) found a significant association between old age and adequate glycaemic control (OR = 0.97; 95% CI: 0.945, 0.995).
Their results contradict previous findings because of differences in the demographics of the studied populations, especially in terms of the age distribution (Radwan et al., 2018). Higher percentages of participants were of older age in the Palestinian samples than in other study groups, with 34% of participants >61 years (Radwan et al., 2018) and more than 50% ≥58 years (Mosleh et al., 2017). The second demographic aspect examined was gender. No significant correlation was seen between gender and the risk of inadequate glycaemic control. This finding is inconsistent with a previous patient-level pooled analysis of six RCTs, totalling 2600 participants with diabetes, which found women more prone to hypoglycaemia and high HbA1c levels (Kautzky-Willer et al., 2015).
Another systematic review and meta-analysis from 2015 found that the risk of inadequate glycaemic control was higher in women; however, three of their included studies were deemed unreliable due to the high risk of bias (Sobers-Grannum et al., 2015).
Various explanations have been put forward for these observations. It has been proposed that women show lower levels of adherence to medications, to which they more frequently report side effects (Thunander Sundbom & Bingefors, 2012). Differences in lifestyles (e.g. exercise) as well as in psychosocial perceptions might also play a role in gender-based differences in glycaemic control (Walker et al., 2006). Socio-economic differences might also play a part; however, these were not considered in any study.
Differences in physiology have also been suggested to underlie any sex-based differences in HbA1c levels; however, further investigation is required of any sex-based differences in efficacy/

F I G U R E 6
Forest plot of the odds ratio of inadequate glycaemic control by smoking status treatment response, as current evidence for factors such as dissimilarities in body composition is unlikely to provide sufficient explanation for such observations (Radwan et al., 2018). In Lebanon, Noureddine et al. (2014) found no association between gender and the risk of T2DM within the Lebanese population (Noureddine et al., 2014). Similarly, in Saudi Arabia, Al-Rasheedi (2014) failed to demonstrate a significant connection between the two factors (Al-Rasheedi, 2014). However, though insignificant, both studies showed that the percentage of males within the inadequate glycaemic control group was higher than in the adequate glycaemic control group.
With regard to the anthropometric measures, this review found a significant association between increased WHR and high HbA1c levels (p < .05). Such results are congruent with the findings from (Vazquez et al., 2007), where 32 different populations from 29 studies were analysed. In that study, WHR was found to correlate well with the incidence of diabetes. WHR is considered by some a reliable measure of visceral fat (Gadekar et al., 2020). An elevated WHR indicates greater accumulation of visceral fat stores, a feature that negatively influences the metabolism of hormones such as insulin through free fatty acid secretion, which hinders insulin uptake by the liver leading to hyperinsulinemia and subsequent insulin resistance (Despres et al., 1995;Kahn & Flier, 2000).
Nevertheless, reports have questioned the reliability of WHR for accurately reflecting changes in visceral fat (Van Der Kooy et al., 1993). This suggests that more sensitive measures should F I G U R E 7 Forest plot of the odds ratio of inadequate glycaemic control in relation to extended duration of diabetes (>10 years) be used to properly explore the relationship between fat mass and glycaemic control.
Obesity, signified by BMI >30 kg/m 2 , was significantly associated with an increased risk of inadequate glycaemic control. This observation supports the findings of Abdullah et al. (2020). Moreover, a significant association was detected between BMI and HbA1c levels.
As mentioned above, obesity (represented by both high WHR and BMI) is usually accompanied by insulin resistance and insensitivity due to the effect of visceral fat on insulin secretion and clearance.
Accordingly, this sheds some light on a possible association between insulin resistance itself and HbA1c. Saha and Schwarz (2017) found HbA1c a reliable tool for predicting insulin resistance and inadequate F I G U R E 8 Forest plot of the odds ratio of inadequate glycaemic management in relation to self-management F I G U R E 9 Forest plot of the odds ratio of inadequate glycaemic control in relation to physical activity glycaemic control among 3578 normoglycaemic and hyperglycaemic individuals (Saha & Schwarz, 2017).
The third category of factors related to glycaemic control is that of lifestyle. This includes physical activity, smoking, and psychological and social patterns. This review indicated a significant association between smoking and inadequate glycaemic control. Other meta-analyses have found significant associations between smoking (whether active or passive) and the risk of developing T2DM (Pan et al., 2015;Soulimane et al., 2014;Wang et al., 2013). With the exception of Alzaheb and Altemani (2018), all the included studies that analysed smoking used HbA1c as an index for glycaemic control, and reported significant positive correlations between smoking and HbA1c. Numerous explanations have been put forward to explain the relationship between smoking and glycaemic control.
One theory proposed that smoking raises HbA1c levels by altering the erythrocytic membrane permeability to glucose, consequently increasing haemoglobin glycation (Higgins et al., 2009 (Beutler, 1975;Smith et al., 1982). The strong association between smoking and T2DM may offer a new approach for avoiding a rise in HbA1c levels, and therefore reducing the incidence of the disease. For example, awareness campaigns could educate people living with diabetes regarding the particular seriousness of smoking and its harmful influence on their glycaemic status.
Regarding physical activity, this review found that increased physical effort led to a significant reduction in the risk of inadequate glycaemic control (p < .001). Similarly, Aune et al. (2015) presented a meta-analysis of 87 studies (including three RCTs and 84 cohorts) demonstrating an inverse relationship between total physical activity (i.e. leisure-time, occupational and transport) and the risk of inadequate glycaemic control (Aune et al., 2015). The consequential reduction in body adiposity, a major contributing factor to the development of diabetes, can explain this response to sustained, regular physical activity (Mozaffarian et al., 2011;Rana et al., 2007).
Alternatively, Aune et al. (2015) pointed out that physical exercise involves concentric contractions of the body's skeletal muscles, which take up glucose from the circulation (Aune et al., 2015). This process is mediated through the GLUT4 glucose transporter located on the membranes of striated muscle cells. Studies have reported that physical activity enhances GLUT4 translocation on the myocytic membrane, thereby increasing the rate of glucose uptake by muscles and improving glucose homeostasis (Colberg et al., 2010;Perseghin et al., 1996;Röckl et al., 2008). On sensitivity analysis, three records Lastly, a highly significant link was demonstrated between longer disease duration and inadequate glycaemic control. By contrast, a UK-based cross-sectional study found an inverse relation between the two variables. This finding may be attributed to the different healthcare systems of developing and developed countries. Abubakari et al. (2016) hypothesized that longer disease duration allows participants to better understand their condition's nature, consequences and responsibilities, which assists in the management of negative emotions that might affect their response, allowing them to better control their disease (Abubakari et al., 2016).

| Strengths and limitations
There are several limitations to this meta-analysis. First, most included studies used self-report to collect outcome data, potentially leading to recall bias, which might either exaggerate or underestimate the true effect size. Second, in terms of the anthropometric markers, only two measures were observed, WHR and BMI. These measurements have been reported as less informative measures of obesity than other measurements such as waist circumference and the waist-to-height ratio (Rajput et al., 2014;Vazquez et al., 2007).
Third, the high heterogeneity observed between studies was not always resolvable, potentially indicating diversity in the methodology or baseline criteria of the included studies. Greater consistency in design among studies is recommended for the future. Fourth, we only included English language studies, and peer-reviewed published articles to ensure the quality of the included studies' methodologies.
The strengths of this review include the relatively large number of included studies, adding to the strength of association between the predictors of glycaemic control.
In conclusion, the current evidence suggests an increased risk of inadequate glycaemic control in patients with elevated WHR, longer disease duration, obesity and smokers, with lower risk of inadequate control associated with physical activity and self-management.
Further longitudinal studies are required to better understand these variations and to assess all predictors of glycaemic control in participants with T2DM, and to further provide a strong basis for future recommendations to optimize glycaemic control.

| RECOMMENDATIONS
Review findings illustrated some variability in variables which exert significant impact across the countries of the Middle East and North Africa. Irrespective of whether they are middle-income or highincome, countries should use national and regional data to inform the development of strategies sensitive to local and global influences. Broadly, however, policymakers and clinicians should pay attention to spreading awareness campaigns to educate people living with diabetes regarding the seriousness and implications of obesity and smoking for glycaemic status. Regular exercise is essential, especially in people with living with diabetes for long periods, who should be encouraged to follow a balanced physical program to improve their glycaemic status. Diabetes education and self-care management should be integral to the management of all people living with diabetes regardless of their medication and management pathways.

ACK N OWLED G EM ENT
None.

CO N FLI C T O F I NTE R E S T
None.

PE E R R E V I E W
The peer review history for this article is available at https://publo ns.