Using a purposive sampling method, we selected 2 isolated urban settlements (Agu-Abor and Ugbodogwu) in Enugu, the capital of Enugu State, south east Nigeria. The two settlements have an estimated adult population of 7000 - 9000 individuals (based on church and local records). The total area occupied by both settlements is approximately 2.5 - 5 km² and are located about 1 - 2.5 km from the nearest state-owned teaching hospital. The two settlements were selected purposively because of their relatively isolated location. The inhabitants of Agu-Abor were surveyed over a 4-week period (August 12-September 9, 2013), while Ugbodogwu inhabitants were surveyed between November 25 and December 21, 2013. This study was approved by the ethics committee of the University of Nigeria Teaching Hospital Ituku/Ozalla, P M B 01129 Enugu. No NHREC/05/01/2008B-FWA00002458-1RB00002323. Date of approval 28^th July 2013.

A cross-sectional descriptive study was done to survey the adult population living in both localities. In order to facilitate the participation of the populace, sensitization meetings in the community which included both religious and elected leaders. Community-wide awareness announcements were carried out in churches and by town criers followed by sensitization meeting(s) on selected days. Following community entry, all participants who came out for the survey were interviewed by teams of research assistants. Using the WHO STEPS instrument [20] data on selected socio-demographic characteristics and lifestyle behaviors including, physical activity was collected. Inclusion criteria were all consecutive consenting adults 20 years and above, while the exclusion criteria were refusal to participate. Pregnant women, patients on steroids and individuals who could not be weighed or have their heights were interviewed but their data were not included in the final analysis. Clinical assessment of the participants was carried out in a field clinic. Data on weight, height, blood pressure, fasting blood glucose as well as past medical history were collected at this phase of the study. Only data on behavioral and physical measurements were collected. Biochemical measurements (excluding fasting blood glucose) were not done.

Fasting blood glucose (FBG) was measured using a glucometer (Fine test premium, Infobia co Ltd., Dongan-gu. South Korea) after an overnight fast. Fine test coding strips contains glucose oxidase and potassium ferricyanide and displays results between 0.6 - 33.3 mmol/L with a coefficient of variation of <10%. Accuracy of the results obtained from the Fine test meter compared with Hitachi glucose auto analyzer 747 and showed a system accuracy within ±0.8 mmol/L of 100%. Fine test control solution was used to check and test strips to make sure they are properly working. Participants with fasting blood glucose within normal range on two consecutive measurements (measured 2 - 5 days apart) were considered as having normal blood glucose [14]. If blood glucose was abnormal in two occasions it was categorized as prediabetes or diabetes. Subjects with one abnormal glucose levels were referred for glucose tolerance test in teaching hospital and were not included in the study.

Blood pressure was measured after 5 - 10 minutes rest in a sitting position and was measured thrice by means of mercury sphygmomanometer according to the guidelines of the European Society of Hypertension [21]. The measurements were done by one of the investigators or a doctor not below the rank of a registrar who was trained for the study from the department of medicine, UNTH using mercury sphygmomanometers and stethoscopes. All blood pressure measurements were obtained from the non-dominant arm. Systolic blood pressure (SBP) and diastolic blood pressures (DBP) were taken at the first and fifth Korotkoff sounds, respectively. The mean of 3 consecutive blood pressure measurements was documented. Weight was measured using a standard bathroom scale in kilograms (nearest 0.5 kg). Scales were calibrated and recalibrated on daily basis by re-adjusting their pointers to zero. At the beginning of the study, scales were compared with standard scales used in the hospital. Bathroom scales were used because of cost and availability. Height was measured in centimetres using a straight centimetre ruler with the patient standing erect on a flat surface. Body mass index (BMI) was calculated as weight (Kg) divided by squared height (m²) and categorized as underweight < 18.5 kg/m², normal weight 18.5 - 24.9 kg/m², Overweight 25 - 29.9 kg/m², Obese ≥ 30 kg/m². Obesity was defined as BMI ≥ 30mg/m² based on WHO criteria [22]. Patients with impaired fasting blood glucose were subsequently referred to the state hospital for follow-up.

Classification of diabetes and prediabetes was based on the WHO criteria [23].

FBG < 6.1, mmol/L, n(%)-normal range

FBG 6.1 - <7, mmol/L, n(%)-Prediabetes range

FBG ≥ 7, mmol/L, n(%)-Diabetes range

Definition of diabetes included two categories:

1) Previous diabetes and/or use of hypoglycemic agents;

2) FBG within the diabetic range on two occasions during the survey period.

Awareness was assessed based on past medical history of diabetes diagnosed by a health professional. Normal blood glucose control was defined as normal FBG measurements in individuals taking hypoglycemic agents at the time of survey. Individuals with systolic blood pressure (SBP) of ≥140 mmHg and/or diastolic blood pressure (DBP) of ≥90 mmHg, past medical history of hypertension and/or use of anti-hypertensive drugs were considered as having hypertension. Stroke was defined as a clinical syndrome of rapidly progressive symptoms and signs of focal or global neurological deficit lasting more than one hour of which there is no apparent cause other than vascular origin, and/or past medical history stroke diagnosed by a qualified personnel (doctors). Current tobacco use was defined as the use of any form of tobacco in the past 4 weeks. Alcohol use and quantity was defined as (mean quantity) the consumption of any alcoholic beverage in a week. The safe limit of alcohol was defined based on WHO guidelines of 21 units for men and 14 units for women per week [24].

Artisans were defined as skilled manual laborers. Level of education was the individual’s highest educational (formal) attainment based on the Nigerian school system.

Sample size was calculated using the Taro Yamane formula [25] , N = N/1 + N(e)². Where: n = required sample size, e² = error limit and N = estimated adult population in both settlements.

N = estimated population of the community (9000), e = 0.05.

N = 9000/9000 × 0.0025 = 9000/22.5 = 400. With an expected 10% attrition rate, a minimum of 440 individuals will be screened.

For database management and statistical analyses, we used the SPSS version 20 (IBM Corporation, New York, USA). Data were presented in tables. Age standardization was done using the WHO standard population distribution and Enugu East (EE) 2006 population data [26] [27]. For continuous variables, mean values and 95% confidence intervals were calculated. Prevalence of diabetes and prediabetes were expressed as percentages and confidence interval calculated. Mean values were compared using the independent t-test and the Mann-Whitney U test where applicable. In all, p values of <0.05 was regarded as statistically significant. The confidence level was kept at 95%.

Out of the 811 individuals invited to the clinic, 605 (74.6%) participants (414 women (68.4%) and 191 (31.6%) men p < 0.01) had their fasting blood glucose measured based on the study protocol. The male to female ratio of those screened was 0.5:1. The distribution of the participants showed that 444 (54.7%) came from Ugbodogwu while 367 (45.3%) were from Aguabor. The participants’ age ranged from 20 to 90 years, averaging 44.5 (43.3 - 45.8) years. Males were older than females by about a decade (mean age: 51.1 vs 41.5 years, p < 0.01). The age distribution and other characteristics of the participants are shown in Table 1. Females were shorter but weighed the same as males. Most participants were artisans/business men or women and had at least a secondary school education (51.6%). Current alcohol (males 160 (83.8%), females 306 (73.9%). P < 0.01) and tobacco (males 73 (38.2%), females 61 (14.7%). P < 0.01) use were more prevalent in males than females.

Fasting blood glucose (p < 0.01; skewness, 3.5; kurtosis, 21.6) and body mass index (P < 0.01; skewness, 0.9; kurtosis, 0.9) were skewed to the right. Twenty-four (4%) subjects were underweight, 285 (47.1%) had normal BMI while 296 (48.9%) were either overweight or obese. The mean body mass index (BMI) was 25 kg/m² higher in females (26.7 kg/m²) than males (24 kg/m²) P < 0.01. The mean fasting glucose level was similar in males and females. P = 0.98 (Table 1).

The mean blood pressure of the participants was 133.7 mmHg (95% CI, 131.7 - 135.9) systolic and 83.4 mmHg (95% CI, 82.4 - 84.7) diastolic. Systolic blood pressure was significantly higher in males than in females. P = 0.01. Past medical history of hypertension, previous diabetes and stroke was documented in 23%, 5.3% and 2.8% of the population sample respectively. The overall prevalence of hypertension (newly diagnosed and previous cases) was (52.2%) 316/605 (Table 1).

The overall prevalence of diabetes and prediabetes in the population was 71 (11.7%), (95% CI; 9.2 - 14.3) and 7.6% (95% CI; 5.5 - 9.7); similar in males and females P = 0.21 and 0.25 respectively. More than half of all the cases of diabetes (39/71 (54.9%) were newly detected (65.9% in females and 37% in males, p = 0.02). Out of the remaining 32 (45.1%) who were aware of having diabetes 28.1% had normal fasting serum glucose at the time of the study. Control was almost similar in males and females P = 0.06. The highest prevalence of diabetes was in females 60 years and above (17.2%, 95% CI, 7.5 - 26.7) (Table 2). Prevalence of diabetes varied widely with BMI and was most prevalent in those with BMI category of 25 - 29.9 kg/m². The sex distribution of diabetes and BMI showed that it was most prevalent in males with BMI ≥ 30 kg/m². Individuals with previous history of stroke also showed a very high prevalence, however their number was small.

The highest prevalence of prediabetes was in males 40 - 49 years (95% CI, 16.7%, 3.4 - 30). The prevalence of prediabetes was similar in all sub-groups compared. It however increased with increasing BMI reaching 10.2% in those with a BMI of 30 kg/m² and above (Table 3). The Enugu East Age adjusted prevalence of diabetes was 9.8% and the WHO population adjusted prevalence was 11%. The age specific rates adjusted for the two populations are shown in Table 4. The odds ratio of diabetes in different groups is shown in Table 5. It was significantly higher in subjects ≥ 45 years, hypertensives, stroke patients and those with sedentary lifestyles.

P-values are for the sex differences. Peripheral systolic and diastolic blood pressure were the average of 3 consecutive measurements.

*percentage of overall prevalence of hypertension, β previous history of diabetes. **Fisher’s exact test, †Controlled percentage of Aware with fasting blood glucose within normal range during the survey, P-values for sex differences.

*Fisher’s exact test.

*significant P-values.