Poor blood pressure control and its associated factors among peoples living with Diabetes Mellitus in sub-Saharan Africa: A Systematic Review and Meta-analysis

Poor blood pressure control among people living with diabetes mellitus (DM) is one of the primary causes of cardiovascular complications and death in sub-Saharan Africa (SSA). However, there is paucity of evidence on the prevalence and associated factors of poor blood pressure control in SSA. Therefore, this review aimed to estimate the prevalence and its associated factors among people living with DM in SSA. access studies was using Cochrane Q test statistic and and using statistical at level. A random-effects model was used to estimates the pooled prevalence of poor blood pressure control with a 95% condence interval (CI).

Hypertension is the worldwide leading cause of cardiovascular diseases (CVD) and deaths (1), and accounts for around 7.5 million yearly deaths (2). Globally, an estimated 1.13 billion people are hypertensive, most (two-thirds) living in low-and middle-income countries (3). The highest prevalence of hypertension in the world is observed in SSA (4,5). It tends to occur more commonly with diabetes and as many as 70-80% of diabetic patients suffer from hypertension, which worsens and accelerate the progression of both micro-and macro vascular complications of diabetes and results in 7.2-fold increase in the risk of mortality (6)(7)(8)(9). Moreover, type 2 diabetes mellitus (T2DM) patients with hypertension have a two to four-fold greater risk of developing cardiovascular sequelae than age-matched normotensive type 2 diabetic controls (10).
Therefore, it is imperative to control cardiovascular disease (CVD) risk and mortality in diabetes patients and the most effective and powerful intervention to reduce it is controlling blood pressure by integrated use of life style modi cation and appropriate regimen and dose of anti-hypertensive medications (11,12).
The bene ts of tight BP control in patients with diabetes exceed the bene ts of tight glycemic control and extend to the prevention of macrovascular and microvascular complications (13,14). Many randomized controlled trials and the United Kingdom Prospective Diabetes Study (UKPDS) showed that strict BP control in patients with hypertension and diabetes reduces the risk of stroke, coronary heart disease, congestive heart failure and other CVD, macrovascular and microvascular complications, and death (13,15,16).
A meta-analysis has revealed that a 10 mm Hg reduction in systolic blood pressure reduced the risk of major cardiovascular disease events by 20%, coronary heart disease by 17%, stroke by 27%, heart failure by 28%, and all-cause mortality by 13% (17). Lowering of blood pressure to treatment targets is therefore, a priority in individuals with diabetes to prevent complications (18)(19)(20). Even though different guidelines differ in their recommendations on BP targets in diabetic patients (21), many guideline committees had recommended that in patients with DM and hypertension, the target systolic and diastolic BP should be below 130 and 80 mm Hg, respectively (6,21,22). However, most hypertensive diabetic patients fail to meet the recommended BP target. A study in USA by Andros et al. reported that poor blood pressure control among individuals with diabetes was still high and remains a major public health concern causing economic burden (23). In a longitudinal cohort study of 30,228 diabetic patients, only 43 and 30% of European American and African American diabetic hypertensive patients, respectively, demonstrated a target blood pressure of 130/80 mmHg (24). In SSA rates of BP control range between 11 and 35% (25,26) and the cardiovascular complications in this region diabetic individuals are attributed to poor blood pressure control (27). A study carried out in six specialized diabetes care centers of six SSA showed an overall poor BP control in T2DM individuals despite adherence to guidelines (26). This implies the presence of other factors attributed to this poor BP control, including demographic, health literacy, and socio-economic characteristics (28). Moreover, overweight and noncompliance with antihypertensive drugs were strongly associated with uncontrolled hypertension (29,30).
Even though extensive efforts to develop interventional BP control strategies to decrease risk of complications have been made in the past several decades, there is still signi cant rise in the risk of complications in diabetic patients with hypertension, and the control of BP is sub-optimal. To solve this problem and meet the blood pressure target, understanding of risk factors of blood pressure control particularly in SSA is important. Therefore this systematic review and meta-analysis aimed to determine the prevalence and associated factors of poor blood pressure control among people living with DM in SSA.

Search strategies
This systematic review and meta-analysis followed the Preferred Reporting Items for Systematic Reviews and Meta-Analysis guidelines (S1 Table, PRISMA Checklist) (31). PubMed, African Journals online, Hinari, Google scholar, and direct Google search were used to access relevant studies for this review. Moreover, references list of eligible studies were retrieved to account for the missed studies in the database searching. All studies reporting proportion or prevalence of poor blood pressure control among people living with DM (either T1DM or T2DM) in SSA countries were the target of this review.

Eligibility criteria
We used CoCoPop (Condition, Context, and Population) approach for prevalence studies to declare inclusion and exclusion criteria.

Inclusion criteria and exclusion criteria
We included all studies conducted in SSA countries, written in English language, published as of August 01, 2020 and reported prevalence and/or associated factors of poor blood pressure control among people living with type 1 or type 2 diabetes. All included studies were published. Studies which did not report prevalence of poor blood pressure control, case-reports, case-series, letters to the editors, and studies conducted on speci c population were also excluded.

Study selection
All studies retrieved using different electronic databases were exported into endnote version X7. After excluding duplicated articles, titles of all articles were screened and abstracts and their full-texts were independently reviewed by two authors (YA and SAT). Disagreement between reviewers was resolved by further discussion and other reviewers (DAA and MG) Quality assessment The quality of included studies were assessed using the validated modi ed version of a quality assessment tool for prevalence studies (32). Two reviewers (BD and GAT) independently assessed the quality of the included studies and discrepancy in quality appraisal between the two authors was synchronized by the third reviewer (SAT). The quality assessment tool has 9 risk of bias items which has a maximum score of "9" and minimum score of "0". Ranking of risk of bias is labeled as low risk (0-3), moderate risk (4-6), and high risk (7-9) (32).

Data extraction and management
Findings on the prevalence and associated factors of poor blood pressure control among diabetic patients from each study were summarized by two authors (YA & YY) using the data extraction format which was prepared with the assistance of the Joanna Briggs Institute (JBI) data extraction tool for prevalence studies and the extracted data were compared between the two authors (YA and YY). Discrepancies were resolved by consensus after discussion. For each study; name of the rst author, year of publication, study design, sample size, blood pressure cut point used (to de ne poor blood pressure control), the prevalence of poor blood pressure control or number of cases with poor blood pressure control, and associated factor (sex, BMI, adherence) estimates (Odds ratio or the cases in each cell) with their standard error were extracted.

Outcome measurement
Outcome variable of this study was poor blood pressure control among peoples living with diabetes mellitus. In the primary studies, the outcome variable was de ned using different blood pressure cut points. Included studies de ned poor blood pressure control in diabetes as; 1: systolic blood pressure (SBP) ≥ 140mmHg and /or a diastolic blood pressure (DBP) ≥ 90 mmHg; 2: SBP ≥ 140mmHg and /or a DBP ≥ 80 mmHg, and 3: SBP ≥ 130mmHg and /or DBP ≥80. The latest BP cut point to de ne poor blood pressure control is SBP ≥ 130mmHg and /or DBP ≥80 in DM patients (33)(34)(35).

Statistical analysis
The extracted data in the Microsoft excel were exported into STATA version 16.0 software for further analysis. The pooled estimate of poor blood pressure control prevalence and its associated factors were determined by the random-effects model using DerSimonian-Laird weight (36). Statistical heterogeneity was checked by Cochrane Q-test and I 2 statistics (37). To minimize the variance of point estimates between primary studies subgroup analysis was carried out by BP cut point used to de ne poor BP control, year of publication, income level, type of DM, and sample size. Besides, sensitivity analysis was also conducted to determine the effect of single studies on the pooled estimate. Univariable metaregression was conducted by publication year, mean age of the respondent from primary studies, sample size and income level using random-effects model.
Publication bias (small study effect) was checked using funnel plot and statistically by Egger's test (38).
Odds ratio(OR) and 95% con dence interval was used to identify factors associated with poor blood pressure control in people living with DM.

Systematic Review registration
The review protocol has been registered in PROSPERO with protocol number of CRD42020187901, URL: https://www.crd.york.ac.uk/PROSPERO/#myprospero

Description of included studies
We retrieved 1,043 records of journal articles in the electronic database search and 456 duplicates were removed. After scrupulous review of the titles and abstracts, we excluded 561 articles. Five articles were excluded in the full-text review because of difference in population under study (28), full text not found (39,40), and reviews (21,41). We used full-text copies of 21 records for further review with an overall sample size of 6, 308 ( Figure 1).

Handling Heterogeneity
The extent of heterogeneity among the included studies was high in the random-effects model pooled estimate. To handle this, sensitivity, subgroup and, meta-regression analyses were performed. In the sensitivity analysis, no in uential study was found ( Figure 4). We further did and report estimates from a sub group analysis considering other possible sources of variations including the cut point used to diagnosis poor BP control, type of DM, sample size, income, and year of publication. However, heterogeneity was not handled. A pooled prevalence from sub-group analysis showed, studies which used BP cut point of 130/80 mmHg systolic and diastolic blood pressure, respectively, had the highest prevalence of poor BP control (77.7%) (Supplementary gure (S. Figure 1)), and the least pooled prevalence (55.3%) was observed in the subgroup of studies those used 140/80 mmHg as a BP cut point. Pooled prevalence of poor BP control was higher (71%) in studies whose study population was both (type 1 and/or 2) types of DM than studies with study population of type 2 DM only (68.8%) (S. Figure ). Higher pooled prevalence of poor blood pressure control (70.0%) was found in studies with sample greater than or equal to 296 than the counter parts (S. Figure 3). Studies published from 2005-2012 had the highest pooled prevalence of poor blood pressure control (82.0%) (S. Figure 4). Sub group analysis by income was also done and the highest pooled prevalence was found in lower middle income SSA countries (Ghana, Nigeria, Cameron) (77.3%) followed by low income SSA countries (Ethiopia, Kenya, Uganda, Tanzania) (71.1%) (S. Figure 5) ( Table 2).

Meta-regression
To handle heterogeneity, we further tted meta-regression on the aggregated study level variables using the random effects model. The univariable meta-regression analysis revealed that mean age, publication year, sample size and income level were not signi cantly associated with poor blood pressure control (Table 3). Factors associated with poor blood pressure control among people living with diabetes mellitus in sub-Sahara Africa In the random effect model of meta-analysis of identi ed associated factors, the pooled effect of four studies (11,50,55,65) showed, poor adherence to antihypertensive treatments (POR =1.7; 95% CI: 1.03, 2.80, I 2 =0.0%, p=0.531) was signi cantly associated with poor blood pressure control among peoples living with diabetes mellitus ( Figure 5). The pooled effect of other three studies (52,53,55) also showed that body mass index (BMI) of ≥25kg/m 2 (POR =2.4, 95% CI: 1.57, 3.68), I 2 =0.00%, p=0.47) was signi cantly associated with poor blood pressure control (Figure 06). On the other hand the pooled effect of eight studies (11,50,52,53,55,(64)(65)(66) showed that sex (POR=0.7, 95% CI, 0.49 -1.12, I 2 = 74.1%, p=0.001) was not signi cantly associated with poor blood pressure control (Figure 07). The odds of poor blood pressure control among diabetic peoples with poor adherence to antihypertensive treatment was 1.7 times higher than the counter parts. Body mass index of ≥25 kg/m 2 was associated with a 2.4 times more likelihood of having poor blood pressure control (Table 4).

Discussion
In peoples living with diabetes, the risk of developing CVD and life-threatening complications is determined by the degree of control of hypertension. However, the rate of control of hypertension is often inadequate and the cardiovascular complications in SSA diabetic individuals are attributed to poor blood pressure control (14,27,68). This systematic review and meta-analysis estimated the pooled prevalence and associated factors of poor blood pressure control among peoples living with diabetes mellitus in sub-Saharan Africa.
The pooled prevalence of poor blood control in diabetes patients among 21 studies in SSA was 69 (70), and a systematic review and meta-analysis of 44 studies (88%) that use the same BP cut point to diagnose poor blood pressure control (71). The high prevalence of poor BP control in people with diabetes might be due to different factors, like the likelihood that a more emphasis on glucose control and under emphasis of treatment for associated risk factors such as hypertension (71). In addition, inadequate access to follow-up care and prescription medications, inappropriate or ineffective treatments, poor adherence to prescription medication and lifestyle modi cations or a combination of these factors may be responsible (72,73). Low compliance is the main reason for poor control of blood pressure in SSA (74). This high prevalence of poor blood pressure control among diabetes suggests urgent need for initiatives to improve and control of hypertension.
However, the prevalence of poor BP control in this review is higher than the nding of a review of 16 studies (64.8%) (75), a nationwide study in Korea (60.1%) (76), a study in Malaysia (52.8%) (77), and other large scale cohort studies (59.7%) (78) and (60%) (79). This inconsistency could be explained by the fact that patients in SSA countries have low socioeconomic status and poor wealth index that are attributed to the weak health systems resulting in poor access to medications and healthier life style (80,81). They have low access to quality healthcare services, are more likely to have low healthier lifestyle and to be non-adherent to their medications due to barriers in accessing medical care, unaffordable health care costs, lack of transport money to visit the hospital and other reasons, leading to poor blood pressure control (82,83). A great proportion of populations in SSA countries have no access to more than one blood pressure lowering drugs and, when they are available, they are not affordable (84). Poor dietary quality and practice such as, high saturated-and trans-fatty acid intake, and low fruit and vegetable consumption and physical inactivity are increasingly becoming prevalent in low and middle income countries (LMICs) (81,(85)(86)(87). The mean salt intake in most of the LMICs is also beyond the recommended maximum intake (88). These all factors might contribute to the higher prevalence of poor BP control in SSA.
On the contrary, the prevalence of poor BP control among peoples living with diabetes in this review is lower than the prevalence of poor blood pressure control in Spain (90.2%) (89). This variation may possibly due to difference in; proportion of overweight and obesity, magnitude of co-morbidity, age of study participants, and patient, physician and sociocultural factors, which are hypothesized to have impact on BP control (90,91).
On the other nine studies of this review which used BP cut point of 140/90, the pooled prevalence of poor blood pressure control was 70.0% (95% CI: 61.9%, 78.08). A similar study that uses the same BP cut point to diagnose poor blood pressure control in Saudi Arabia among 1178 diabetic reported a consistent nding (71.8%) (92). In this systematic review and meta-analysis, the lowest pooled prevalence of poor blood pressure control (55.3%) was found in three studies that uses BP cut point of 140/80. This nding is lower than any of the above prevalence reports, which is clearly due to the BP cut point used to diagnose poor BP control.
In this review, diabetes patients with poor adherence to antihypertensive medications were more likely to have poor blood pressure control. This nding is in agreement with existing literature showing higher prevalence of poor blood pressure control among patients with poor adherence (70,76,93,94). This is due to the fact that adherence to antihypertensive therapies is a primary determinant of treatment success and poor medication adherence is the primary cause for poor control of BP (95). Poor adherence attenuates effectiveness of antihypertensive drugs. Moreover, it is noticeable that poorly adherent patients are less likely to undertake a healthier life style, contributing to poor BP control (94). Therefore, this review indicates the need of counseling and encouraging patients to adhere to antihypertensive medications, as adherence to antihypertensive medications is a key to achieving an optimal BP.
In the current review, BMI of 25 kg/m 2 or greater was associated with poor blood pressure control. Consistently, other studies reported that being obese is associated with poor blood pressure control (70,76,(96)(97)(98)(99)(100). In obesity there is an increased production of leptin, a polypeptide produced from adipocyte which stimulate sympathetic activity leading to renal water retention, increased heart rate and peripheral vascular resistance, and nally worsening of hypertension (101). Besides, obesity decreases sensitivity of peripheral tissues to insulin (causing insulin resistance) resulting compensatory hyperinsulinemia. This increased insulin has a primary effect on activation of sympathetic nervous system, endothelial dysfunction, and failure of peripheral vasodilation (102). Insulin or leptin induced activation of sympathetic nervous system is also associated with tubular sodium reabsorption and volume expansion, and consequent increased blood pressure (103). Moreover, an increase in BMI results in increased renin and plasma angiotensin-converting enzyme activity, angiotensin level, and plasma aldosterone (104), thereby causing poor blood pressure control.
Findings from the current review might be helpful for clinicians, programmers, and policymakers to design a strategy and take prompt interventions which would prevent complications and death due to uncontrolled hypertension in peoples living with diabetes and hypertension comorbidity.

Limitations
Though sensitivity, subgroup, and meta-regression analyses were conducted to minimize the effect of heterogeneity, the extent of heterogeneity among the included studies was high, which might be due to the difference in the study area, methodology, study period, blood pressure cut point used, and other unexplained variations. Use of different BP cut points by primary studies to de ne poor blood pressure control in DM patients was the other possible limitation of this review. Hence, clinicians and policymakers should consider these during the interpretations of results.

Conclusion
The pooled prevalence of poor blood pressure control among peoples with diabetes was high. Only one third of diabetic patients had achieved target blood pressure. This high prevalence of poor BP in hypertensive diabetic peoples highlights the urgency for primary health care re-engineering and an urgent need for initiatives to improve and control of hypertension. BMI of 25 kg/m 2 or greater and poor adherence to antihypertensive treatments were signi cantly associated with poor blood pressure control. Preventive measures should concentrate on overweight patients and patients with poor adherence to antihypertensive medications. Not applicable.

Availability of data and materials
All the materials and data are presented within the manuscript.
PRISMA ow diagram for a systematic review and meta-analysis of poor blood pressure control among people living with diabetes in sub-Saharan African countries    Sensitivity analysis between studies included in a meta-analysis Figure 5 Forest plot for the association between poor adherence to antihypertensive drugs and poor blood pressure control among diabetic peoples in sub-Saharan countries Figure 6 Forest plot for the association between BMI and poor blood pressure control among diabetic peoples in sub-Saharan countries.

Figure 7
Forest plot for the association between sex (being female) and poor blood pressure control among diabetic peoples in sub-Saharan countries.