Prevalence of Diabetes Mellitus Among the Adult Population in Rural Blocks of Haryana,India: A Community-Based Study

Abstract

Objective:

The aim of this was to study the prevalence of diabetes in the rural adult population of Haryana, India.

Research Design and Methods:

A total of 2606 adults aged ≥18 years were randomly selected from two rural blocks of Haryana State. Those without diabetes were subjected to a 75-gram oral glucose tolerance test. Demographic, socioeconomic, and anthropometric details along with blood pressure and physical activity were recorded, and their association with the prevalence of diabetes was studied.

Results:

The prevalence of newly detected diabetes was 7.3%, whereas the overall prevalence of known and newly detected diabetes was 13.3%. Multiple logistic analysis showed a statistically significant association between the prevalence of diabetes and increasing age, waist-to-hip ratio (WHR), positive family history of diabetes, low level of physical activity, and systolic blood pressure. No significant association was observed with education level and socioeconomic status.

Conclusions:

The prevalence of diabetes is rising, even in the rural population of Haryana. A positive family history of diabetes, low physical activity, and high WHR are strong predictors of diabetes in tested adult rural population of Haryana.

Introduction

T he prevalence of type 2 diabetes mellitus (DM) is increasing globally and reaching epidemic proportions. This impact is more marked in developing countries like India where various epidemiological studies have reported a 10-fold increase in prevalence of diabetes in urban areas from 1971 to 2000 (from 0.12% in 1971 to 12.1% in 2000).^1

–6 But there are wide variations in the prevalence of diabetes in different parts of India, including urban as well as rural areas. It has been observed that although very high levels of diabetes have been reported in urban areas of India, where a number of studies have been carried out, few data are available for rural areas of India, where more than 70% of the population lives. Recent rural studies have reported the prevalence of DM ranging from 2.7% to 13.2%.^5,6

India is a multiethnic, multireligious, and multicultural country with wide differences in living conditions, socioeconomic levels, and dietary habits, thus it is difficult to predict a uniform level of DM. Therefore, multiple regional studies in different subtypes of populations are required for quantifying the prevalence data. Recent data for the prevalence of DM are not available in rural Haryana; moreover, available data do not adequately represent the rural areas of northern India. Sadikot et al.⁵ reported the standardized prevalence for diabetes in total Indian, urban, and rural populations as 4.3%, 5.9%, and 2.7%, respectively. The corresponding impaired glucose tolerance (IGT) rates were 5.2%, 6.3%, and 3.7%, respectively. Of 817 subjects found to have DM, 71.1% were newly diagnosed cases and 28.9% were previously known cases of diabetes. But even this multicentric study done by Sadikot et al. had no representation from Haryana. Haryana has a long agricultural background with population of 2.1 crore (21 million). Because of the high per capita income of the state, the lifestyle of the rural population has also changed. The present study was planned keeping this perspective in mind so that the diabetic load in the rural population of District Jhajjar and District Rohtak in Haryana could be determined and the health system better prepared to better manage it.

Material and Methods

The study was community based and cross-sectional and carried out in rural blocks of the Districts of Jhajjar and Rohtak. A list of 268 anganwaris was prepared from the Jhajjar and Rohtak Districts. Out of 268 anganwaris, 87 were chosen randomly by a lottery method. A total of 2606 participants (20–75 years old) who were permanent residents of the area for more than a year were selected randomly from each anganwari. Before starting the study, the help of local leaders [members of panchayat, teachers, nongovernmental organization (NGOs), etc.] was taken to sensitize the people regarding the need for the study. The procedure of the study was explained to the participants and informed consent was taken. The day before the study, subjects were advised to observe overnight fasting (at least 8 h) and were called at the nearest health center/Anganwari Centre (AWC) in the morning.

Initial evaluation included detailed history and clinical examination to exclude any systemic diseases. Anthropometric measurements, including body weight (to the nearest 0.1 kg) and height (to the nearest 0.001 meter), were recorded in subjects without shoes and socks. The waist circumference was measured between the iliac crest and the lower costal margin, and the hip circumference was measured at the maximum circumference of the buttocks while the subjects were standing with feet placed together. These readings were taken three times, and the mean was taken for calculation of the waist-to-hip ratio (WHR). Percentage body fat (BF) was calculated by impedance plethysmography (bioelectrical impedance meter Omron BF 302). Percentage body fat >25% in men and >30% in women was used as a criterion to define overweight in healthy subjects.^7,8 Blood pressure was recorded twice at an interval of 5 min in a seated position in the nondominant arm to nearest 2 mmHg using a standard adult mercury sphygmomanometer. The mean of the two readings was taken as the blood pressure.

After estimating fasting capillary glucose, those without diabetes were subjected to an oral glucose tolerance test (OGTT) with 75 grams of anhydrous glucose powder dissolved in 250–300 mL of water that was to be consumed over 5 min. Time was counted from the start of the drink. Fasting and 2-h postglucose load and plasma glucose were estimated by glucometer (Ultra 2; Johnson and Johnson, New Brunswick, NJ), which was validated. In every tenth case, venous plasma glucose was estimated by using glucose oxidase method. The correlation coefficients for fasting plasma glucose (FPG) and 2 h PG by glucometer and laboratory methods were 0.96 and 0.87. While waiting after the intake of 75 grams of glucose, the subjects were asked to avoid physical activity during the next 2 h.

A pretested semistructured interview schedule was used for collecting the details of demographic and socioeconomic status, family history, and physical activity. The sample size was calculated as 1800 subjects by assuming prevalence of DM in rural areas as 9.9% and allowable errors of 10% at level of significance of 95%, using the following formula for sample size, (n)=Z ² _1−α/2 p(1−p)/d ²; however, 2606 subjects participated in the study. Patients suffering from chronic renal, pancreatic, or other severe illness, pregnant women and women who delivered 2 months or less preceding the study, and patients on steroids/nicotinic acid or other medication likely to cause dysglycemia were excluded from the study. For the physical activity of individuals, pro forma adopted (with modifications) from Global Physical Activity Questionnaire (GPAQ) of the World Health Organization (WHO) was used.⁹ It took into account activity at the workplace, travel to and from places, recreational activity, and sedentary behavior. The work-related and recreational activities were divided into vigorous and moderate physical activities. Metabolic equivalent (METs), defined as the ratio of working metabolic rate to resting metabolic rate, was commonly used in the analysis of physical activity. The total MET minutes per week were calculated from the sum of MET minutes of each category. Finally they were classified as high, moderate, or low physical activity as described under the GPAQ. The coding column was used as a guide for analysis of the physical activity data. The American Diabetes Association (ADA) criteria were used to categorize the subjects into prediabetes and diabetes.

According to the objectives of the study, the collected data were compiled, tabulated, and analyzed using appropriate statistical tests. The nonparametric tests, chi-squared test, and t-test were applied wherever applicable to see the association with difference among the various factors of the study.

Observations

The prevalence of newly diagnosed diabetes was 7.3% whereas overall prevalence of known and newly diagnosed diabetes was 13.3%. Prediabetes, i.e., impaired fasting glucose (IFG), impaired glucose tolerance (IGT), and IFG+IGT, was diagnosed in 26.85%, 3.25%, and 8.35% of study subjects, respectively. The mean age of study population was 49.39±15.45 years. The anthropometric variables of total 2606 subjects, 1212 males with a mean age of 51.89±16.34 years and 1394 females with a mean age of 48.32±14.92 years, are given in Table 1. Prevalence of diabetes was higher in males (15.8%) as compared with females (11.1%), and the sex-wise difference in prevalence was statistically significant (p<0.001). Body mass index (BMI), waist circumference, WHR, waist-to-height ratio (WHtR), percentage body fat, mean systolic (SBP) and diastolic blood pressure (DBP), fasting and 2-h plasma glucose, and family history of diabetes were statistically different in diabetic males and females as compared to nondiabetic males and females (Table 1). In males, the association of smoking and alcoholism with prevalence of diabetes was found to be statistically significant (p<0.001). Multiple logistic regression analysis using diabetes as the dependent variable and each independent variable was performed (Table 2). The variables of age, positive family history, low physical activity, WHR, and SBP showed a significant association with diabetes. No definite association of prevalence was observed with increasing or decreasing educational levels (p=0.775) and socioeconomic status (p=0.078) in the study subjects.

Table 1.

Baseline Characteristics of Study Subjects Stratified by Gender (N=2606)

	Female (n=1394)		Male (n=1212)
	Without diabetes n=1239	With diabetes n=155	Without diabetes n=1020	With diabetes n=192
Age (years)	47.37±14.91	55.88±12.87^a	50.95±16.80	56.99±12.45^a
BMI (kg/m²)	22.43±4.60	23.50±4.90^b	21.38±3.87	22.63±4.57^a
Waist (cm)	80.61±12.02	84.64±12.72^b	81.78±10.87	85.78±13.29^a
Waist-to-hip ratio	0.88±0.08	0.91±0.08^b	0.90±0.07	0.93±0.08^b
Waist-to-height ratio	0.52±0.08	0.54±0.08^b	0.49±0.06	0.51±0.08^a
Systolic blood pressure (mmHg)	123.3±19.8	131.8±22.5^b	125.2±17.9	131.6±21.6^a
Diastolic blood pressure (mmHg)	80.0±11.2	82.0±11.9^b	81.8±11.6	84.2±11.8^a
Fasting plasma glucose (mg/dL)	96.0±11.38	206.0±93.27^b	98.1±10.71	199.7±76.27^b
2-h plasma glucose (mg/dL)	121.0±18.08	251.0±37.01^b	125.0±16.60	226.0±22.06^a
Body fat (%)	25.9±6.63	28.3±8.43^a	24.7±6.96	28.8±6.66^a
Family history of diabetes	26 (2.1%)	8 (5.2%)^a	11 (2.2%)	11 (11.5%)^b
Smokers	123 (9.9%)	19 (12.3%)	298 (58.4%)	48 (50.0%)
Alcohol	4 (0.3%)	1 (0.6%)	112 (22.0%)	33 (34.4%)^b
Sedentary workers (low physical activity)	725 (58.5%)	129 (83.8%)^b	241 (47.3%)	72 (75.0%)^b
Socioeconomic status (upper middle/high)	155 (13.1%)	21 (14.9%)	67 (13.1%)	20 (20.8 %)^a

Compared with those do not have diabetes.

p<0.05.

p<0.001.

Table 2.

Odds Ratio for Risk Factors Found to Be Associated with Diabetes

	Female		Male
Variable	Odds ratio	P value	Odds ratio	P value
Age	1.03	0.001^*	1.04	0.005^*
Family history	3.70	0.003^*	7.46	0.000^*
Low physical activity	2.25	0.002^*	2.52	0.002^*
Waist-to-hip ratio	2.89	0.002^*	8.10	0.001^*
Systolic blood pressure	1.04	0.036^*	1.01	0.057

P<0.05 is statistically significant.

, significant P value.

Discussion

Studies from different parts of India have suggested a rising trend in the prevalence of type 2 DM in almost all populations.^2,10,11 Earlier studies on the prevalence of diabetes had reported a higher prevalence in urban areas and a quite low prevalence in rural areas. But recent studies from rural areas have been showing quite high rates of prevalence.^12
–14 The Elluru survey by Rao et al.¹² in 1989 reported the prevalence of known diabetes in four villages of Andhra Pradesh as 6.1% in individuals aged 40 years and above that was unexpectedly high at that time for a rural area with low socioeconomic status and decreased health awareness. Sudha et al.¹³ in 2006 reported a 9.3% prevalence of diabetes as per WHO criteria and 8.9% as per ADA criteria from rural areas of Maharashtra. A large-scale survey by Chow et al.⁶ in 2005 sampled 4535 subjects >30 years of age from 20 villages of Andhra Pradesh and on the basis of finger prick measurements. These investigators reported a prevalence of 13.2%, out of which 6.4% were known cases of diabetes and 6.8% were newly detected cases of diabetes; furthermore, 15.5% had impaired fasting glucose levels. This study reported the highest prevalence of diabetes until now from a rural area.

Our study has also supported this trend and has shown the prevalence of diabetes to be 13.3% (347/2606 participants), out of which 6.0% were already known cases of diabetes on treatment and 7.3% were newly diagnosed during this study. The higher prevalence of newly detected cases indicates the low level of health awareness in rural areas. The prevalence of prediabetes (IFG, IGT, and IFG+IGT) was 38.45%. Individually, IFG was seen in 26.85% and IGT was seen in 3.25%, whereas IFG+IGT together were found in 8.35%. Here the collective prevalence of IGT (IGT and IFG+IGT), which was 11.6%, is similar to the NUDS study, where it was 14%.³

This prevalence of diabetes is one of the highest reported from rural areas in India. One of the reasons is that there are no previous data from the rural areas of Haryana, and this might be due to regional differences in population characteristics. Second, most of the older studies used the old WHO criteria for diagnosing diabetes (fasting plasma glucose=140 mg%), and IFG (fasting plasma glucose=110 mg%), whereas we used the current ADA diagnostic criteria (diabetes=fasting plasma glucose ≥126 mg%, IFG=fasting plasma glucose ≥100 mg%). Obviously lowering the cutoff levels for diagnosing IFG and diabetes by 9%–10% from WHO to ADA criteria is going to increase the prevalence of disease. Similar observations were made by others investigators from various parts of the world and the Indian subcontinent while applying this revised ADA criteria on urban populations. Zarger et al.¹⁵ reported that 19.4% of Kashmir valley population had IFG whereas Johnsen et al.¹⁶ found that the prevalence of IFG in Denmark increased from 11.8% to 37.6%. Thus, lowering the diagnostic criteria for IFG caused a dramatic increase in prevalence of IFG, but the concordance rate between IFG and IGT remained poor, as has been shown in the present study also. Third, Haryana is a small state so villages are not far distanced from urban areas. This proximity of the study area to towns has also contributed to villagers in this area adopting more urban behaviors and seeking more sedentary jobs. This fact is also supported by the findings of present study, where low physical activity was found to be associated with odds ratios of 2.25 and 2.52, respectively, in females and males for developing diabetes.

Previous studies done in rural populations had shown that the prevalence of diabetes is related to age, BMI, WHR, physical activity, and family history of diabetes.^12
–14,17 In our study, the prevalence of diabetes shows an association with sex and alcohol consumption apart from the above-mentioned risk factors. More prevalence of diabetes in the case of males than females can be attributed to alcohol consumption and smoking, whereas females in rural areas are physically quite active and share many of the outdoor responsibilities of males, such as working in the agricultural fields and taking care of cattle. This trend of more prevalence in males was also seen in previous studies from South India.^3,18 Multiple regression analysis showed a higher odds ratio for developing diabetes with WHR and a family history of diabetes—8.10 and 2.89 in the case of WHR and 7.28 and 3.70 in the case of a positive family history in males and females, respectively.

To the best of our knowledge, no previous study had reported an association between percentage body fat and prevalence of diabetes in the rural population of India. In the present study, we observed that participants having diabetes had a higher level of body fat percentage in both males and females, and this was significant when compared to normal group. This association can be due to the fact that the body fat percentage of Indians is significantly higher than their western counterparts having similar BMI values and plasma glucose levels.²⁰ It has also been hypothesized that excess body fat and less lean muscle mass explain the high prevalence of hyperinsulinemia and higher risk of diabetes in Asian Indians.¹⁹ No association of the disease with socioeconomic status was established in our study (p=0.078), thus refuting the old notion that diabetes is a “disease of the affluent.” This is in contrast to the findings of the study done by Menon et al.,¹⁸ who showed a positive association with high income. No definite association was observed with increasing or decreasing levels of education (p=0.775). Similar findings were reported by other researchers from south India as well.^12,13,18,20

The three main factors associated with increasing prevalence of prediabetes and diabetes in the rural population are decreased physical activity, high WHR, and positive family history. This rising trend of diabetes and prediabetes in a rural population shows a disturbing phenomenon and should be an alert to the health-care authorities to formulate proper primary prevention health-care policies to decrease future burden of new diabetes on the health-care system.

Footnotes

Author Disclosure Statement

No competing financial interests exist.

References

Mohan

, Sandeep

. Epidemiology of type 2 diabetes: Indian scenario. Indian J Med Res, 2007; 125:217–230.

Sandeep

, Mohan

. The changing trends in epidemiology of diabetes. Med Update, 2008; 18:133–141.

Ramachandran

, Snehlatha

, Kapur

et al. Diabetes epidemiology study group in India (DESI). High prevalence of diabetes and impaired glucose tolerance in India: National Urban diabetes Survey. Diabetologia, 2001; 44:1094–1101.

Ramachandran

. Epidemiology of diabetes in India—three decades of research. JAPI, 2005; 53:34–38.

Sadikot

, Nigam

, Das

. The burden of diabetes and impaired glucose tolerance in India using the WHO 1999 criteria: Prevalence of diabetes in India study (PODIS) Diab Res Clin Prac, 2004; 66:301–307.

Chow

, Raju

et al. The prevalence and management of diabetes in rural India. Diabetes Care, 2006; 29:1717–1718.

Pollock

, Wilmore

. Exercise in Health and Diseases. Philadelphia, PA: WB Saunders, 1990.

Hortobagyi

, Israel

, O'Brien

. Sensitivity and specificity of the Quetelet index to assess obesity in men and women. Eur J Clin Nutr, 1994; 48:769–775.

Global physical activity questionnaire: Department of chronic diseases and Health promotion surveillance and population based prevention. World Health Organization, Geneva, Switzerland. www.who.int/chp/steps/.

10.

Ahuja

MMS

. Recent contributions to the epidemiology of diabetes in India. Int J Diab Dev Countries, 1991; 11:5–9.

11.

Mishra

, Pandey

, Rama

Devi

et al. High prevalence of diabetes, obesity and dyslipidaemia in urban slum population in northern India. Int J Obesity, 2001; 25:1–8.

12.

Rao

, Ushabala

, Seshaiah

et al. The Elluru survey: prevalence of known diabetes in rural Indian population. Diabetes Res Clin Pract, 1989; 7:29–31.

13.

Sudha

, Avinash

, Rajashree

. To identify the risk factors for high prevalence of diabetes and impaired glucose tolerance in Indian rural population. Int J Diab Dev Countries, 2006; 26:19–23.

14.

Mohan

, Mathur

, Deep

. Urban rural differences in prevalence of self-reported diabetes in India—The WHO–ICMR Indian NCD risk factor surveillance. Diab Res Clin Prac, 2008; 80:159–168.

15.

Zarger

, Bashir

, Masoodi

et al. Prevalence of prediabetes in Kashmir valley using revised ADA criteria for impaired fasting glycaemia. Diab Res Clin Pract, 2006; 72:329–330.

16.

Johnsen

, Colagiuri

, Balkau

et al. Creating a pandemic of prediabetes: The proposed new diagnostic criteria for impaired fasting glycaemia. Diabetologia, 2004; 47:1396–1402.

17.

Ramachandran

, Snehlatha

, Baskar

et al. Temporal change in prevalence of diabetes and impaired glucose tolerance associated with lifestyle transition occurring in the rural population in India. Diabetologia, 2004; 47:860–865.

18.

Menon

, Kumar

, Gilchrist

. Prevalence of known and undetected diabetes and associated risk factors in central Kerala—ADEPS. Diab Res Clin Prac, 2006; 74:289–294.

19.

Wang

, Thornton

, Russel

et al. Asians have lower body mass index (BMI) but higher percentage body fat than do white: Comparisons of anthropometric measurements. Am J Clin Nutr, 1994; 60:23–28.

20.

Ramachandran

, Snehlatha

. Rising prevalence of NIDDM in an urban population in India. Diabetologia, 1997; 40:232–237.