Metabolic, Anthropometric, and Type 2 Diabetes Mellitus Related Risk Factors in Normal and Pre-Diabetic Adults

Type 2 diabetes mellitus (T2DM) is a major global health problem. The present study examines the relationship between the metabolic, anthropometric and Finnish risk score (FINDRISC) among normal and pre-diabetic adults. Subjects (n = 1319, aged above 18 years) from the Qatari population were classified into two groups based on their hemoglobin A1c (HbA1c) measurements (non-diabetic A1c<5.6% and pre-diabetic 5.6% ≤ A1c ≤ 6.4%) were examined for their anthropometric (height, weight and waist circumference), metabolic [fat, fat free mass (FFM), muscle mass (MM), total body water (TBW), bone mass, degree of obesity, basal metabolic rate (BMR), body mass index (BMI), metabolic age, visceral fat rating, systolic blood pressure (SBP), diastolic blood pressure (DBP), total cholesterol (Total-C), low-density lipoprotein (LDL), high-density lipoprotein (HDL), triglycerides (TG), fasting / random plasma glucose (FPG/RPG), HbA1c and vitamin D (VitD)] and FINDRISC. Means and frequencies were determined in aggregate and by subgroups for all variables and correlations between categorical variables were tested to estimate the association between the anthropometric and metabolic risk factors with the FINDRISC. A percentage of 74.8% (n = 987) of the study population aged below 45 years old and their overall BMI was 28.8±5.2 kg/m (overweight). Pre-diabetic subgroup have shown a statistically higher FINDRISC compared to their non-diabetic counterparts (11.2±4.1 vs. 9.8±4, p<0.001). The FINDRISC was significantly and directly correlated with the BMI, HbA1c and FPG. However, HbA1c was correlated directly with BMI, SBP, DBP, FPG/RPG and indirectly with the levels of HDL. This study demonstrates an apparent relationship between the HbA1c and FINDRISC score. Pursuing further research on this association may permit using HbA1c with the FINDRISC in predicting the risk of T2DM to be a better tool rather than using the current FPG/RPG, OGTT methods.


Introduction
Type 2 diabetes mellitus (T2DM) is now a global health concern and is one of the main chronic diseases influencing people despite their geographic location and socio-economic status. T2DM has been rising dramatically over the last few decades and is currently identified as a leading cause of mortality worldwide (Wild, Rolic, & Green, 2004). The association between T2DM and increased mortality and morbidity is due to its presence as a key predisposing factor for cardiovascular disease, stroke and other life-threatening complications. Factors contributing to the progression of T2DM are mainly related to genes and the environment. However, T2DM prevalence is found to increase with age and other behaviors including and related to obesity, sedentary lifestyle, physical inactivity, and unhealthy dietary habits (Mokdad et al., 2001;Venkataraman et al., 2004). As incidence of T2DM continue to rise globally, it is critical to maintain early prediction for the risk of the disease in order to avoid or control it.
T2DM is mainly caused by impaired glucose intolerance (IGT) which is a result of insulin resistance and islet β-cell exhaustion (Stumvoll, Goldstein, & Haeften, 2005). It is a metabolic disorder with heterogeneous etiologies mostly characterized by chronic hyperglycemia and disturbances of carbohydrate, fat and protein metabolism. Other factors leading to T2DM include hypertension, hyperglycemia, obesity and dyslipidemia (Dandona, Aljada, & Bandyopadhyay, 2004). Dyslipidemia triggers increased levels of triglyceride (TG) and fatty acids either from diet or from accelerated lipolysis, the exposure of muscle cells to fatty acids causes an impairment of insulin-mediated glucose uptake and consequently contributes to insulin resistance (Dimopoulos et al., 2006;Bilam et al., 2009). In other words, insulin resistance, accompanied by impaired fasting plasma glucose (IFG) levels, lead to an increase in the levels of low density lipoprotein (LDL) and TG, and a decrease in levels of high density lipoprotein (HDL) cholesterol in the body (Sorrentino, 2005).
The leading risk factor for T2DM is the "Pre-diabetes" stage which is identified by hemoglobin A 1c (HbA 1c ) levels of a range between 5.6% to 6.4% (American Diabetes Association, 2010). This stage is characterized by the elevated blood glucose level (accumulated over the previous 2-3 months) which is above the normal range but below that of clinical diabetes. This "pre-diabetic" stage is just an indication of the risk of getting the disease and could be altered to the normal stage (<5.6%) if a healthier lifestyle and better control of the sugar intake is adopted.
An International Expert Committee was convoked in 2008 by the European Association for the Study of Diabetes, the American Diabetes Association (ADA), and the International Diabetes Federation to consider the means for diagnosing diabetes in non-pregnant individuals using HbA 1c as an alternative, if not superior better, tool (International Expert Committee, 2009). Diabetes has been diagnosed for decades with fasting/random plasma glucose (FPG/RPG) assessment or, much less frequently, with an oral glucose tolerance test (OGTT). The measurement of HbA 1c equals the assessment of hundreds (virtually thousands) of FPG and also captures postprandial glucose peaks; therefore, it is a more robust measurement than FPG and/or 2hr OGTT plasma glucose (Bonora & Tuomilehto, 2011).
A well-known screening tool to predict the risk of developing T2DM is the Finnish risk score (FINDRISC) questionnaire (www.diabetes.fi/files/502/eRiskitestilomake.pdf) which gives an estimate number (0-26 points) that identifies a 72% (Saaristo et al., 2005) risk of developing T2DM within the next 10 years (Lindstorm & Tuomilehto, 2003). It includes eight questions about age, body mass index (BMI), waist circumference, physical activity, dietary consumption of fruits, vegetables and berries, as well as history of antihypertensive medications, history of high blood glucose, and family history of diabetes. The total test score provides a measure of the probable development of T2DM among participants, where a score of 15 and above indicates a high risk. Therefore, both HbA 1c and FINDRISC can identify the biomarker features that may assist in the early prediction of T2DM and assist in future control of the disease. Eventually, interweaving HbA 1c in the FINDRISC will improve the power of the conventional tool and provide a more accurate predictor of developing the disease. This cross-sectional study was conducted consequently to prove the same.

Study Design and Population
This cross-sectional study was conducted to explore the relationship between metabolic, anthropometric, and T2DM-related risk factor profiles in normal and pre-diabetic stages in a Qatari population and their link to the early prediction of the disease as well as the profile of vitamin D serum levels in study subjects. Screening campaigns were held in different regions in Qatar for a period of six months; three researchers were trained and certified to work on participant data. Consequently, we finalized the statistical evaluation of the metabolic, anthropometric and T2DM related risk factors from 1,319 participants excluding diabetic patients. No definite sample size was planned as it was a cross-sectional study. The inclusion criteria were Qatari residents (both males and females), above 18 years of age, and with no history of diabetes. Participants excluded were those already diagnosed with diabetes, gestational diabetes, or those whose HbA 1c levels were higher than 6.4%, in addition to pregnant and nursing women (due to hormone imbalance). Those classified as diabetics had their HbA 1c measured twice to confirm. The primary selection criteria were the HbA 1c levels and those were limited to individuals with less than 6.5%. Accordingly, participants were classified based on these HbA 1c measurements into two groups: the non-diabetics (A 1c <5.6%) and pre-diabetics (5.6%≤A 1c ≤ 6.4%). HbA 1c was measured by Quo-lab analyzer. This study was approved by Hamad Medical Corporation/Weill Cornell Medical College in Qatar Joint Institutional Review Board and registered with ClinicalTrials.gov, number: NCT02098980.

Anthropometric and Metabolic Measures
Anthropometric measurements, including height, weight and waist circumference, were measured. The waist circumference was measured by using a measuring tape at the midpoint between the last floating rib and the top of the iliac crest in the midaxillary line. Other weight and metabolic rate-related measurements were obtained using a Tanita body composition analyzer model BC-420MA, (TANITA BC-420MA, Tanita Corporation, Japan). These were mainly BMI (kg/m 2 ), fat percentage (%), fat mass (kg), fat free mass (FFM; kg), muscle mass (MM; kg), total body water (TBW; %), bone mass (kg), degree of obesity, basal metabolic rate (BMR; Kcal), metabolic age (years) and visceral fat rating (%). These factors were assessed according to the manufacturer's protocol (Tanita Corporation). Blood was taken from the finger tip from each subject after a 12hr overnight fast to measure their lipid profile including total cholesterol (Total-C; mmol/L), HDL (mmol/L), LDL (mmol/L), and TG (mmol/L) using the CardioCheck P-A machine and the Lipid panel strips, while the serum glucose (FPG / RPG; mmol/L) was measured by the same machine but using the Glucose panel strips. Systolic and Diastolic blood pressure (SBP, DBP; mmhg) were also measured. Vitamin D (VitD) serum levels were assessed using ARCHITECT machine.
FINDRISC sheet was filled for each participant by the researcher; from the data obtained on height and weight the BMI was calculated (weight in kg ÷ height in m square). The data on waist circumference was obtained from earlier tests. The rest of the FINDRISC questions were self-reported from the participants' knowledge including information on: age, fruits and vegetables intake, 30 minutes daily exercise, blood pressure medication, previous high blood glucose and family history of diabetes.

Statistical Analyses
All statistical analyses were conducted using the IBM Statistical Package for Social Sciences (SPSS version 21.0, IBM Corp., Armonk, NY, USA). Means and frequencies were determined in aggregate and by subgroup analysis of the classified HbA 1c categories (normal:<5.6% and pre-diabetes: 5.6% -6.4%). Continuous variables are given as mean±SDs and an independent sample t-test was used to examine the association between the aforementioned groups. The α error was set at 0.05 and 2-sided P values of ≤ 0.05 were reported as significant.
Associations between categorical variables were tested through chi-square χ² test which aimed to analyze the difference between these two groups. Pearson correlation coefficients (r) were calculated to evaluate the cross-correlation between the cardio-metabolic risk factors and T2DM risk scores.

Results
As shown in Table 1, means and frequencies of the different characteristics where identified in the entire study subjects and stratified by their HbA 1c status (non-diabetic:<5.6% and pre-diabetic: 5.6%-6.4%). The study results showed that there was a predominance in males (male to female ratio 2.3:1). The majority of the study population was below the age of 45 years old (74.8%), overweight and with waist circumference between 94 -102cm in males and over 88cm in females. The majority of the study participants reported the absence of physical activities of over 30 minutes/day, and daily consumption of fruits and vegetables, and the presence of first degree relatives with T2DM. Only 7-10% of the study population had a previous history of high-blood pressure and high-blood glucose. There was a statistically significant difference (p=0.001) between the non-diabetic and pre-diabetic subgroups for their FINDRISC. The metabolic-and weight-related factors are represented in Figure 1. Fat mass, FFM, MM and TBW were non-significantly higher in the pre-diabetic stage compared to the non-diabetic stage, while the bone mass and degree of obesity were statistically significantly higher. The mean value of the metabolic age for the entire study (44.8±13.1 years) was 18% higher than the actual age of the group (37.9±10.2 years) which describes the declined relationship between the body composition and BMR measurements in the total subjects (data not shown). The metabolic age was also significantly higher in the pre-diabetic subjects compared to their non-diabetic counterparts (49.4±13.2 vs. 43.3±12.8 years, p<0.001; Figure 1-i). The visceral fat rating in the entire population is 9.8±5.7 which is below the ideal range of 13%, indicating an overall healthy population. However, there was a significantly higher visceral fat rating in the pre-diabetic subgroup compared to the non-diabetic one.   The cardiometabolic risk factors were also assessed in the entire study and in the non-diabetic and pre-diabetic subgroups (see Table 2). In general, the pre-diabetic subjects have shown a statistically higher FINDRISC compared to their non-diabetic counterparts (11.2±4.1 vs. 9.8±4, p<0.001). The mean levels of cardiometabolic factors in the study population fell within the normal clinical range. Age, BMI, SBP and Total-C were all significantly higher in the pre-diabetic subgroups compared to the non-diabetic one. There was no significant difference between the non-diabetic and pre-diabetic subgroups for the DBP and TG. However, HDL was significantly lower in pre-diabetic subjects and LDL was significantly higher compared to the non-diabetic subjects. Both fasting and random plasma glucose where higher; as expected; in the pre-diabetic subjects compared to the non-diabetic subgroup. VitD serum levels, measured as 25(OH)D in the entire group was on the deficient range (44.2±17.8 nmol/l), there was a slight but non-significant further drop in the pre-diabetic subjects compared to the non-diabetic subgroup.
A cross-correlation analysis among the cardiometabolic risk factors as well as the FINDRISC is shown in Table  3. The risk of T2DM identified by FINDRISC was significantly and directly correlated with the BMI, levels of HbA 1c and TG. However, HbA 1c was correlated directly with BMI, SBP, DBP, Total-C, LDL, FPG/RPG and indirectly with the levels of HDL. HbA 1c tended to increase dramatically with increasing FINDRISC scores (Costa et al., 2013). The above findings indicate that introducing HbA 1c into the FINDRISC can be a more precise prediction of T2DM than utilizing the FINDRISC alone.
VitD serum levels were indirectly associated with the BMI and cholesterol levels (Total-C, TG, LDL), whereas directly correlated with the risk of T2DM (FINDRISC), blood pressure, HbA 1c and HDL. Although these associations were identified to be insignificant (perhaps due to the small sample size and the large inter-individual variation in vitamin D serum level), they substantiate a possible role of VitD deficiency in the emergence of pre-diabetic stage.

Discussion
The present study examined the association between the anthropometric, metabolic and FINDRISC factors in 1319 Qatari -subjects (local and expat, excluding diabetics, nursing and pregnant women) which represent 2.169 million Qataris. There was an overall strong relation between the aforementioned factors and the prevalence of T2DM risk factors where 24% (n = 323) of the population sample were identified as pre-diabetic subjects in their middle age (42±9.8 years). This is considered to be a high percentage in accordance to their age due to their sedentary lifestyle and food habits, where 61% (n = 806) were not doing 30 minutes per day of physical activity and 42% (n = 559) were occasionally consuming fruits and vegetables (as shown in Table 1). There are even high genetic factors involved as 52% (n = 680) were reported to have first-degree relatives with T2DM. Eventually, Qatar is among the countries with the highest percentage prevalence of diabetes (Christos et al., 2014). Our findings are similar to previous studies that showed the relation between the anthropometric factors along with FINDRISC diabetes scores in obese populations (Stijn et al., 2015).
There was a statistical relation between the cardiometabolic risk factors and FINDRISC. BMI was the oddest figure with 43% of the sample population has their BMI value laying between 25-30 kg/m 2 with the pre-diabetic stage being the highest (30.3±6 kg/m 2 ). Statistics also reflected a strong correlation between the BMI and FINDRISC. As known obesity plays a vital role in the etiology of T2DM from different prospective; (a) activation of innate-immune system, (b) release of cytokines (tumor necrosis factor-α, interleukin-1 β and interleukin-6) blocking major anabolic cascades of insulin, (c) triggering of acute-phase proteins (C-reactive protein, plasminogen activator, inhibitor-1, serum amyloid-A, haptoglobin) which characterize the pre-diabetic stages of T2DM (Badawi et al., 2010). Eventually, the weight-related factors were slightly higher in the pre-diabetic stage compared to their non-diabetic counterparts.
Results have clearly shown that vitamin D values were in the deficient range (44.2±17.8 nmol/l) for the entire study population, however, was not significant. A potential explanation of such insignificancy would be due to the large inter-individual variation between the subjects in each subgroup and the relative small sample size. In the present study, BMI was inversely associated with plasma 25(OH)D which suggests that vitamin D plays an active role in obesity, with low levels stimulating synthesis and release of parathyroid hormone, increasing calcium in adipocytes, and promoting weight gain (McCarty & Thomas, 2003). As studied earlier, vitamin D deficiency has been implicated to the development of metabolic syndrome, obesity and T2DM (Brenner et al., 2011). Recently it has been reported that about 90% of the Qatari population has insufficient levels of vitamin D (Badawi et al., 2012), which contributes to the prevalence of T2DM.
The current cross-sectional study impacted the utility of HbA 1c as a useful tool in T2DM prediction. The strong direct correlation of HbA 1c with the FINDRISC, BMI, SBP, DBP, Total-C, LDL and FPG/RPG and the indirect correlation with HDL statistically shows that HbA 1c can be predictive to all aforementioned. Altogether are markers of T2DM, so HbA 1c can be a representative tool and even a better one. Therefore, FINDRISC along with HbA 1c factors are considered to be the best prognostication of the T2DM disease.
There were several limitations in the collection of data such as: (a) the self-reporting nature of the data (e.g. physical activity), (b) the absence of 5% of data from the 1,319 participants, (c) the sample size was not representative of the total population -a cross sectional study, (c) the high cost of HbA 1c testing compared to the inexpensive FPG, and (d) the inability to rule out the possibility of residual confounding factors (e.g. diet, transportation-related physical activity and sun exposure).
In conclusion, the growing prevalence of T2DM requires the development of a better preventative method to reduce the incidence of the disease, which in turn will help in reducing health costs. It is currently recommended that screening for diabetes and pre-diabetes should be carried out using the FINDRISC, followed by a glucose test (FPG, 2hr OGTT and HbA 1c ). This cross-sectional study showed a direct association between HbA 1c along with BMI and FINDRISC, which indicates that interweaving HbA 1c into the FINDRISC can be a strong predictive tool. The early prediction of T2DM will help in applying strategies for the future prevention of the disease and reducing the burden of its chronic complications.