Background
In sub-Saharan Africa (SSA), growth rates of diabetes mellitus (DM) and hypertension are among the highest worldwide. While today an overall DM prevalence of 4% is assumed, the number of affected patients is projected to double from 12 to 24 million within the next 20 years [
1‐
4].
DM and other chronic diseases hit Africa in particular: The health system does not reach a considerable portion of the population, has a focus on emergencies and infectious diseases, and is frequently limited in staff and infrastructure. Not rarely, health workers are insufficiently trained in chronic disease management [
2]. Severe complications and a reduced life expectancy for both diabetic and hypertensive patients are among the consequences [
4‐
6].
In urban Ghana, type 2 DM (DM2) affects at least 6% of adults and is associated with age and obesity. Some 23% of adults are overweight, and this has been related to advanced age, female gender, urban environment, high income and tertiary education [
7,
8]. Epidemiological data suggest interactions between acculturation, urbanisation, and genetic disposition to be involved in DM2 among Ghanaians [
5,
9,
10].
Contrasting increasing prevalence, severe complications and public health significance, studies on DM2 in SSA are remarkably scarce. Understanding manifestation and associated factors, however, is essential to guide diagnosis, management, and prevention of DM2 in this region. Here, we examined clinical, anthropometric, socio-economic, nutritional and behavioural parameters among 1466 urban Ghanaian adults with and without DM2 and hypertension, and present these data and an explorative analysis of associated factors in this population.
Methods
Study site and design
The study was conducted from August 2007 through June 2008 at Komfo Anokye Teaching Hospital (KATH) in Kumasi, Ghana. In this region, 6% and 29% of adults are affected by DM2 and hypertension, respectively [
5,
11,
12]. At KATH, the diabetes and hypertension clinics are frequented each by > 100 patients/week. The study aimed at examining factors associated with DM2 and hypertension among hospital attendants with DM2 and/or hypertension and controls. Secondary objective was to describe the patients' clinical and biochemical characteristics. The study protocol was reviewed and approved by the Ethics Committee, School of Medicine, Kwame Nkrumah University of Science and Technology, Kumasi, and informed written consent was obtained from all participants.
Recruitment procedures and examinations
Following study-related information, patients attending the diabetes center (n = 495) or the hypertension clinic (n = 451) were recruited. Patients encouraged members of their community to participate in the study as preliminary controls. After exclusion of DM2 and hypertension (see below), the latter were included into the study as controls (n = 222). Likewise, further controls were recruited among outpatients (n = 150) and hospital staff (n = 148).
From 10:00 p.m. prior to the examination day, the participants were instructed on fasting, alcohol and tobacco abstinence, and avoiding excessive physical activity. On the examination day, patients were physically examined and interviewed. Parameters assessed included: age, gender, residence, ethnic group, previous and current diseases and complaints, own and family history of DM and hypertension, medications, smoking behaviour, literacy, occupation, household size, wealth indicators, characteristics of work and recreational sports, fitness indicators as well as axillary temperature, blood pressure (0', 5', 10'; measured after resting for ten minutes; M8 Comfort, Omron, Japan), tuning fork test, and peripheral pulses and ulcers.
Fasting venous blood and urine samples were collected. Concentrations of fasting plasma glucose (FPG, fluoride plasma, +4°C) and of urinary albumin were measured photometrically (Glucose 201
+ Analyzer, Albumin Systems; HemoCue, Sweden). Serum triglycerides, total cholesterol and high-density lipoprotein (HDL)-cholesterol were measured by colorimetric tests (ABX Pentra400, Horiba Medical, Germany). Low-density lipoprotein (LDL) cholesterol was calculated according to the Friedewald formula [
13]. If triglycerides were > 3.0 mmol/L, LDL-cholesterol was quantified directly.
Anthropometrical examinations and nutritional interviews
Weight was measured in kg using a person scale, and height in cm by a statometer. Waist and hip circumferences were assessed in cm using a measuring tape (all devices, SECA, Germany). Body mass index (BMI) and waist-to-hip ratio (WHR) were calculated as: BMI (kg/m
2) = weight (kg)/[height (m)]
2 and WHR = waist (cm)/hip (cm). Triplicates of triceps, biceps, subscapular and suprailiac skin fold thickness were measured in mm on the right-hand side of the body (Harpenden calliper, Baty International, UK). Body fat (%) was calculated according to Durnin & Womersley [
14]. Also, body composition was determined under fasting conditions by bioelectric impedance analysis (BIA) and appropriate software (50 kHz, Nutrigard-S, NutriPlus 1.0; Data Input, Germany).
Daily energy expenditure was calculated as metabolic equivalents (MET) × body weight × duration of activity [
15]. Trained nurses recorded weekly aliment intake with a locally-adapted food frequency questionnaire. Quantity and type of ingested foods of one day were documented by participants using a 24-hours dietary recall and handy-measured food units. Based on local and international food composition tables [
16], the individual intake of macronutrients (total energy, protein, fat, carbohydrates), sodium, and fibre was calculated.
Definitions
DM2 was defined as FPG ≥ 7 mmol/L and/or documented anti-diabetic medication [
17]. Likewise, hypertension denoted a mean BP ≥ 140/90 mmHg and/or documented anti-hypertensive treatment [
18]. Controls were negative for both conditions. Increased serum triglycerides were defined as ≥ 1.695 mmol/L, increased total cholesterol as ≥ 5.17 mmol/L and decreased HDL-cholesterol as ≤ 0.9 mmol/L (male) or ≤ 1.0 mmol/L (female) [
17,
19]. Overweight, obesity and central adiposity were classified as BMI ≥ 25.0 kg/m
2, BMI ≥ 30.0 kg/m
2 and WHR > 0.90 (male) or > 0.85 (female), respectively [
17]. Body fat percentage was increased at ≥ 20% (male) or ≥ 33% (female) [
20].
Statistical analysis
Assessed parameters were compared between controls and patients with diabetes, and following stratification of diabetic patients by hypertension. Between-group-comparisons of continuous parameters were done by Mann-Whitney-U-test and of proportions by χ2-test. We applied an explorative analysis, i.e., not hypothesis-driven, of factors associated with DM2 and/or hypertension. For that, all factors found to be univariately associated with e.g., DM2, were entered into a logistic regression model, a priori including age and gender, and odds ratios (ORs) and 95% confidence intervals (95% CI) were calculated. Stepwise backward removal of parameters loosing significant association in multivariate analysis (P > 0.05) identified factors associated with DM2 independently of each other. These models were applied separately for DM2, DM2 only, DM2 with hypertension, and hypertension only. Further, an alternative, fully adjusted model was established including age, gender and identified confounders.
Discussion
DM2 is emerging literally epidemically in SSA [
3]. This hospital-based study from urban Ghana shows that DM2 patients are predominantly middle-aged and of low socio-economic status, and characterized by high proportions of central adiposity, a respective family history, hypertension, albuminuria and hyperlipidaemia.
The study population - although not representative for the community as a whole - displays many features of urban life in Africa: life-style is mainly sedentary, socio-economic conditions are severely restricted, overweight is frequent, particularly among women, alcohol and tobacco use are generally low, and recreational sports not very popular [
21‐
25]. In comparison with previous studies on DM2 in urban Africa, our findings on obesity and SES are similar. However, our diabetic patients exhibited lower FPG and blood pressure, worse lipid profiles, more frequently a family history of DM2, and less use of tobacco and alcohol [
10,
23,
26‐
28]. In comparison to African American DM2 patients, blood pressure and lipid profiles were similar in the present study but obesity, family history, and smoking and alcohol intake less frequent [
21,
29]. In comparison to African and African American DM2 patients, Caucasian DM2 patients show the worst lipid profiles, the highest rates of tobacco and alcohol consumption, a male predominance in abdominal adiposity, and more physical activity [
21,
30,
31].
Rates of hypertension and albuminuria were high among our DM2 patients, confirming previous findings from Kumasi [
32]. This suboptimal management of patients may reflect both institutional and individual factors including drug cost and availability, health policy disparities, culturally inappropriate lifestyle recommendations, and diluting effects of traditional medicine [
33]. Still, complications occurred at only half the figure reported elsewhere in SSA [
28,
34]. Diagnostic restrictions and a majority of medicated patients may be involved. Typically, in African diabetic patients, late-onset microvascular complications predominate over macrovascular events [
6,
34]. In fact, reported rates of retinopathy in African DM2 patients (25%) exceed those in African Americans (10%) and Caucasians (9%), whereas cardiovascular diseases are estimated at 8%, 33% and 48%, respectively [
34,
35].
Limitations of the study and of associated factor analysis in particular need to be considered. Our DM2 definition by FPG is based on the IDF consensus valid during study conduct and follows general practice. Validated nutritional questionnaires possibly could have improved respective assessments. Importantly, the present study was not matched for e.g., age and sex, and used a convenience sample. Women predominated. Controls were younger than patients and, related to that, roughly a quarter originated from hospital staff. This basically was due to limited project funds and reluctance of community members to participate. Although the multivariate analyses are adjusted for these and other differences, the selection bias has implications for interpreting the results: for instance, the increased odds of DM2 among unemployed individuals are partially an artefact due to the proportion of staff members among controls. This, however, does not invalidate the identification of major risk factors such as, e.g., family history or abdominal adiposity. In analysis, we followed an exploratory approach, i.e., lacking pre-formulated hypotheses, and identified independently associated factors. Thus, associations of e.g. unemployment or crowded living conditions with DM2 are statistically independent which does not mean that they are unrelated in real life. We stratified analysis by the presence of hypertension to illustrate differences in associated factors. For comprehensibility, we abstained from (further) stratification by e.g. age groups but adjusted for age, sex and obesity. Lastly and as a matter of fact, association does not necessarily mean causality, and the direction of an association is subject to interpretation.
Notwithstanding the above mentioned limitations, several and partly inter-related parameters reflecting low SES strongly associated with DM2. The propagation of DM2 among low social classes worldwide may be due to low health care utilization, reduced uptake of prevention messages and SES-dependent differences in risk factors including nutrition and physical inactivity [
36]. Associations of DM2 with outskirts residence and illiteracy point to the possibility of inadequate access to health information. Unemployment, expanded working hours, hard work, and overcrowded households were all associated with DM2 and may reflect stressful living conditions. Psycho-social stressors are known to be capable of adversely influencing the metabolic constitution [
37]. Stress may lead to overeating and poor exercise. Also, increased sympathetic activity may affect adipose and pancreatic tissue regulation and contribute to insulin resistance [
38]. Detailed investigations into the association of psycho-social stress and DM2 in SSA are thus warranted.
The strong association in the present study of DM2 with a respective family history underlines the pronounced predisposition in Africans towards DM2 [
39]. However, replication of risk alleles established in Caucasians not rarely has failed in African populations [
40], possibly as a result of their higher genetic diversity [
41]. Because of this, validated genetic markers of an increased risk of DM2 in Africans are rare. Large-scale studies accounting for environmental variation and, possibly, epigenetic priming, will thus be needed to disentangle predisposition in, e.g. the Ghanaian population.
Obesity, a prominent DM2 risk factor worldwide [
42] and also in the present study, shows an outstanding prevalence in SSA, particularly in urban women [
22]. In many areas of SSA, obesity constitutes an obvious social marker of affluence, and poor knowledge and misconceptions about lifestyle risk factors conflict with appropriate prevention and control of obesity and DM2. Clearly, more research into the traditional cognitive imagery as well as into DM2-related knowledge, attitudes and behaviour is needed to be able to implement socioculturally appropriate health promotion campaigns [
1,
43]. Such is of particular importance considering the specifically increased risk for adulthood obesity (and DM2) as a result of frequent undernutrition in African infants [
44].
Serum lipid profiles are constantly associated with DM2 [
30], and so did hyperlipidaemia in the present study, particularly when DM2 was complicated by hypertension. Hyperlipidaemia may result from a combination of low SES and comparatively high urban food prices which, in turn, favours the intake of inexpensive and highly refined foods, i.e. poor in fibre and protein, but rich in simple carbohydrates, fats and sodium [
10,
44]. In fact, such corresponds to the diet assessed for most study participants. Contrariwise, popular meals based on peanut and fermented maize may improve lipid profiles and underlie the observed weak association of DM2 with total and HDL-cholesterol [
45,
46].
Competing interests
The authors declare that they have no competing interests.
Authors' contributions
ID, GBA, MvdG, JS and FPM conceived and designed the study. ID, GBA, KJT, FM, YAAm, and YAAw were responsible for recruitment, interviews and examinations of study participants. JS did the lipid analysis. ID, ED, and FPM performed the statistical analysis. ID and FPM wrote the manuscript with contributions of all authors. All authors read and approved the final manuscript.