Cluster Identification of Diabetic Risk Factors among Saudi Population

1 Department of Pharmacology and Toxicology, College of Pharmacy, University of Hail, Hail, 81442, Saudi Arabia. 2 Department of Pharmaceutical Chemistry, Pharmacology and Toxicology Unit, Faculty of Clinical Pharmacy, Albaha University, Albaha, 65431, Saudi Arabia. Department of Pharmacy Practice, East Point College of Pharmacy, Rajiv Gandhi University of Health Sciences, Bangalore–560049, India. A. O. University of Naples, Federico II –Centralized Pharmacy, Naples, 80131, Italy. 5 Department of Clinical Pharmacy, Faculty of Clinical Pharmacy, Albaha University, Albaha, 65431, Saudi Arabia. Department of Clinical Pharmacy, College of Pharmacy, University of Hail, Hail, 81442, Saudi Arabia. 7 Department of Clinical Laboratory Sciences, College of Applied Medical Science, Taif University, P. O. Box 11099, Taif, 21944, Saudi Arabia.


INTRODUCTION
There is an exponential rise in the prevalence of diabetes in the Middle East and North Africa (MENA) region, including Saudi Arabia. According to the International Diabetic Federation (IDF) report in 2019, around 463 million adults (20-79 years) were diagnosed with diabetes, and this will grow up to 700 million by 2045 globally. While in the MENA region, 55 million adults  live with diabetes in 2019 and this number is expected to grow to 108 million by 2045 [1]. Estimates of the prevalence of diabetes mellitus in Saudi Arabia indicate that the current prevalence is around 17 percent, with predictions that by 2030 it will plateau at more than 20 percent [2,3]. Linking diabetes with COVID-19 shows a greater chance of death associated with COVID-19 than people without diabetes. However, the probability of death for individuals younger than 40 years of either type 1 or type 2 diabetes is very low [4].
Lifestyle variables, including obesity and being overweight, lack of physical exercise, smoking, inadequate diet, stress and urbanization are critical for the development of type 2 diabetes (T2DM). The risk can be minimized to a greater extent through diet and lifestyle interventions [5]. Furthermore, diet is considered as an adjustable risk factor for T2DM. Studies have shown that a low-fiber diet with high glycemic index is favorably correlated with a greater risk of T2DM [6], and particular dietary fatty acids can influence insulin resistance and the risk of varying degrees of diabetes [7]. Total and saturated fat intake, independent of body mass index (BMI), is associated with an increased risk of T2DM, but higher linoleic acid intake has the opposite effect, especially among leaner and younger males [8]. Following BMI modification, previous weight shift, and alcohol and energy intake, repeated consumption of processed meat but no other meats may increase the risk of T2DM [8]. Soft drinks are also related to an elevated risk of T2DMM [9] and metabolic syndrome [10], since they are correlated specifically with BMI [11]. As a wide difference in information and attitude has been found, a good education and awareness program will improve patients ' knowledge and attitude.

Data collection
Data was collected by trained pharmacists. The data collectors used a questionnaire that included demographic data, family history, smoking and Qat habits, assessment of blood pressure and BMI, physical exercise. On an average, each of the interviews took about 20-30 minutes. Prior to beginning the survey, data collectors were trained for data collection. Height and weight were estimated using standardized methods, and the National Heart, Lung, and Blood Institute (NIH) USA table was used to measure the BMI. Random blood sugar level and blood pressure assessment were carried out according to the World Health Organization (WHO) standard guidelines by qualified practitioners [12].

Agglomerate Hierarchical Cluster Analysis for Identifying National Diabetic Risk Factors
Secondly, we collected all the articles published until December 2020 that reported risk factors for diabetes like gender, age, smoking, family history of diabetes, physical inactivity, BMI, hypertension, metabolic syndrome in normal volunteers from thirteen region of Saudi Arabia. Thirteen regions were clustered into five regions: Central region -Riyadh [13,14], Eastern region -Dammam and related cities [15], Northern region-Qassim [16], Tabuk [17], Hail [18], Arar [19], Aljouf [20], Southern region -Jazan [3], Albaha and Western region -Makkah [21], Jeddah [22], Madinah [23]. Studies that only considering physical inactivity and smoking risk factors, and risk factors in diabetic patients, were excluded. Some cities in the Asir region, such as Abha and Najran, did not report any publications relating to risk factors screening in normal volunteers among thirteen regions. An agglomerative hierarchical cluster analysis with Euclidean distance Group average (Unweighted pair group) centroid method was used to determine the groupings between the different variables. In order to determine the association between risk factors for developing diabetes in healthy volunteers, first hierarchical clustering was performed and secondly to determine the region that is most at risk of developing diabetes using the cluster heat map. Third comparative risk factors clusters among the various regions were identified using K means analysis.

Statistical Analysis
The overall prevalence of DM among volunteers and parameters for risk for diabetes, was calculated as percent and 95 percent confidence interval (CI). Chi-square test was used to identify the association between risk for diabetes and the independent variables. Gender, age group, marital status, family income, educational attainment, occupation, family history, BMI, Khat chewing, smoking, hypertension, daily exercise and work involved physical activity included in the model as independent variables using GraphPad Prism 5.01, Software Inc., San Deigo, CA, USA. All statistical tests were two-sided; and a level of P < 0.05 was used to indicate statistical significance.
Whereas an agglomerative hierarchical cluster analyses with Euclidean distance Group average (Unweighted pair group) centroid method, clustered heat map and K means cluster analyses were used to determine the groupings between the different variables by using software NCSS 2020 Statistical Software (2020). NCSS, LLC. Kaysville, Utah, USA, ncss.com/software/ncss.

Risk Factors Screening for Diabetes at Albaha Region, Saudi Arabia
Five hundred and eleven individuals including 96.28% (n=492) males and 3.71% (n=19) females were screened for risk for diabetes in this study (  Table 1). The prevalence rates of diabetic and non-diabetic among the screened population were 11.54 percent and 88.45 percent respectively. The study characteristics of the participants in the non-diabetic and diabetic studies are shown in Table 1.
Male volunteers were more among non-diabetic and diabetic, 441 and 51 respectively. The agespecific prevalence rate indicated a general rise in middle-aged DM, with a substantial variation between the multiple age groups. In the age group, the highest prevalence was observed in (15- (Fig. 1). The dendrogram identifies the tests that trend together; the most distal branch points reveal those tests that are most closely associated. The cluster identifies relation between risk factors and diabetes in total volunteers and are assigned to clusters based solely on risk factors, a total of three clusters of subjects emerges. Cluster 1 (male, physical inactivity) with largest Euclidean distance indicating highest risk, followed by cluster 2 (female and BMI risk), cluster 3 (age >60, 50-59, smoking, metabolic syndrome, hypertension and 40-49 age) The Euclidean distance was in following order cluster 1 > cluster 2 > cluster 3. Cluster 1 and cluster 2 being higher risk factor and linked to all other risk factors. In cluster 1 and 2 apart from gender, BMI risk and physical inactivity being the most higher risk factors for diabetes and its linked other risk factors. In Fig. 2 represents cluster heat map defining diabetic risk factors in five major region that identified region with most of the risk factors. Two clusters, cluster 1 (western region and eastern region) and cluster 2 (southern region, northern region and central region), and sub-cluster (northern and southern region) were identified. Following risk factors such as male sex, physical inactivity in all regions, family history in the western region, BMI risk factors in the Eastern region are identified as prominent risk factors for developing diabetes. In Fig. 3, K mean clusters are shown, clusters of risk factors are compared between various regions, in Fig. 3 A, B, C, D, E, F, G, H and I the cluster of physical inactivity and BMI risk being more prominent is all regions. Again, demonstrating in all three-cluster analysis technique, physical inactivity and BMI risk being the highest risk factors in Saudi Arabia and it is linked to other diabetic risk factor clusters.

DISCUSSION
Although other survey-based studies are reported on diabetes, screening for prevalence and risk factors is not reported. The focus of this study was to create awareness among obese, hypertensive and physically inactive participants, and additionally to report prevalence rate. As per the 2015 census, the province of Albaha had a population of 466115, but according to new World Population data estimates, Albaha accounted for 1.5 percent of the total population of Saudi Arabia. The prevalence of diabetes from our sample was 11.54 percent, which is lower than the national average, this number can vary because screening was conducted at malls and banks with 511 volunteers. If the mapping cluster survey design of the World Health Organization (EPI) is applied, then the prevalence rate in Albaha might be higher. According to International Diabetes Federation, the prevalence rate of diabetes in Saudi Arabia in 2019 was 18.3 percent for the age group 20-79 years [1]. In 2016, in Jeddah province, Bahijri et al studied for diabetes and prediabetes by mapping cluster design based on World Population as standard and found that the prevalence was 18.3 percent for DM and 11.9 percent for prediabetes [25]. In our study, the higher percentage of the age group for nondiabetics was 15-39, whereas 15-39 and 40-49 for diabetics, since more middle-aged people visit malls and banks. The distribution of screening volunteers was comparable to other studies performed in other regions of Saudi Arabia for both diabetic and non-diabetic volunteers, married with average income with university education with private jobs as reported at Al-Kharj region by Aldossari et al, Eastern province by Al Bagli et al, Riyadh region by Hadlaq et al [14,15,26]. The prevalence of obesity among adults in KSA in 2016 was 35.4 percent, also one of the highest in the MENA region. The prevalence of hypertension among adults in KSA in 2015 was also one of the highest in the GCC region, at 23.3 percent [1]. The link between diabetes mellitus and obesity is well known and has been reported, nationally and internationally, in many other surveys. In a global study involving forty-nine developing countries indicated that overweight (BMI) and obesity (BMI > 30 kg⁄m2) were substantially correlated for developing diabetes than in normal weight individual [27]. The prevalence of DM is higher among individuals with a history of diabetes in their families and is associated independently. A correlation between diabetes and physical inactivity was further recorded in the study. The level of physical inactivity can affect prevalence by its association with other factors such as obesity and hypertension [28]. When interpreting the outcomes of the current research, some limitations should be taken into account. The current study is focused on crosssectional results, so with caution it should be interpreted as associating diabetes with other independent variables. The gender composition of research participants was prejudice against males, and this may influence the approximate prevalence of DM in this study, which may explain why the prevalence of DM among women was much lower than at national level.
The major cluster identified was physical inactivity and BMI risk that was linked all other risk factors. In several trials, the association between physical inactivity as a risk factor for type 2 diabetes has been tested. Unfortunately, there is a lack of physical activity among the Saudi population [29]. For instance, AlQuaiz and Tayel (2009) conducted a cross-sectional analysis at King Khalid University Hospital (KKUH) in Riyadh city on 450 Saudi participants and documented a prevalence of physical inactivity of 82 percent among participants. They also stressed that 88 percent of women were more physically inactive than 72 percent of males [30]. The third cluster with age, smoking, male, hypertension and metabolic syndrome. Family history is an important risk factor which is not clustered in our findings. Poor lifestyle habits have been dramatically related to metabolic syndrome. Physical inactivity is directly and inversely related to metabolic syndrome, which lowers with weight loss and daily physical activity [24]. In some trials, metabolic syndrome has been strongly linked with smoking. Tobacco use has been involved in insulin resistance pathogenesis, as smoking acutely impairs insulin action and causes insulin resistance [31,32].
This study was carried out at selected locations only, e.g., malls and banks whereas, healthcare centers could not be covered which would have noticeable implications. This was the main limitation of the study.

CONCLUSION
The prevalence of DM is 11.54 percent in the Albaha area. In Saudi population, the physical inactivity and high BMI risk factors are identified, and they are linked to other factors which are likely to develop diabetes. The increased incidence of diabetes calls for immediate efforts to encourage prevention and health promotion, which are interventions intended to minimize the burden of diabetes. Modification of the life style and awareness on obesity and physical inactivity are encouraged to be made available to the public to mitigate the risk of developing diabetes.

CONSENT AND ETHICAL APPROVAL
The study was reviewed and approved by the Ethics Committee of Faculty of Clinical Pharmacy, Albaha University. Participation was voluntary and verbal consent was acquired from all of the participants. Confidentiality of all the participants was maintained as no names were mentioned in the questionnaires, participants were told that they have the complete freedom to quit the study at any time.