Abstract
Objectives. This study documented the prevalence of and cardiovascular risk factors associated with obesity and undernutrition in the Gambia.
Methods. Adults (≥15 years; N = 5373) from rural and urban areas completed a questionnaire; their height, weight, and waist and hip circumferences were measured, and their cardiovascular risk factors were assessed.
Results. Prevalence of undernutrition (body mass index < 18 kg/m2) was 18.0%; all strata of society were affected. Prevalence of obesity (body mass index ≥30 kg/m2) was 4.0% but was higher (32.6%) among urban women 35 years or older. Cardiovascular risk factors were more prevalent among obese participants.
Conclusions. Undernutrition coexists with obesity, demonstrating a “double burden of disease.” Differential interventions should focus on high-risk groups; prevention needs a multisectorial approach.
Undernutrition remains a serious health problem in sub-Saharan Africa; adult undernutrition is much more widespread than commonly recognized.1 Adults are responsible economically for children and the elderly; their well-being is crucial. The detrimental effect of undernutrition on physical, economic, and mental well-being establishes a vicious circle of undernutrition and disease.2
Overnutrition is an emerging problem in segments of sub-Saharan African society, particularly where lifestyles become urbanized and westernized,3 and data have accumulated on the adverse health effects of obesity in sub-Saharan Africa.4 Central obesity (obesity in the waist and hips) is particularly associated with cardiovascular disease.5 Obese people have higher blood pressure measurements, blood glucose concentrations, and blood lipid concentrations,6 although not all studies have confirmed these findings.7
In a nationwide rapid assessment among Gambians older than 15 years, 18.0% were undernourished (body mass index [BMI] <18 kg/m2), and 2.3% were obese (BMI ≥30 kg/m2).8 The coexistence of undernutrition and obesity is a “double burden of disease.” To facilitate planning of health interventions, we assessed BMI and associated cardiovascular risks among Gambian adults.
We conducted a community-based survey, from 1996 to 1997, to examine prevalence of and risk factors for major noncommunicable diseases in urban and rural Gambians, described in detail elsewhere.9
Participants (15 years or older) answered a questionnaire. We conducted anthropometry with standard protocols10; electronic Seca scales (Seca Ltd, Birmingham, UK) were used to measure weight, and stadiometers were used for height measurements.
Blood pressure and pulse rate were recorded with a validated automated blood pressure machine (HEM-705CP; Omron, Tokyo, Japan).11 Among participants 35 years or older (n = 2301), blood glucose was measured with a glucose meter (Haemocue-B; AB Haemocue, Ängelholm, Sweden) 2 hours after a 75-g glucose load. In a subsample (n = 1075), serum was analyzed for total cholesterol, triglycerides, uric acid, and creatinine with a centrifugal analyzer (Cobas Fara; Roche, Kent, UK).
Obesity was defined as BMI ≥30 kg/m2 and undernutrition as BMI <18 kg/m2. Participants with a BMI of 18 to 30 kg/m2 were considered, for the purposes of this study, to have adequate BMI. Central fat deposition was estimated by waist circumference and waist-to-hip ratio.
Data were analyzed with Stata, Version 6.0 (Stata Corp, College Station, Tex). Means of continuous variables were compared with Student t tests; proportions were compared with c2 tests. Linear and logistic regression models were used to analyze associations between anthropometric and other measurements. Significance was assigned by a 2-sided a level of .05.
The number of eligible participants was 5373 of 6901 (77.9%). Mean BMI was higher in the urban (22.4) than in the rural (20.2) area and higher among women (21.7) than among men (20.3) (both P < .001). Differences remained highly significant after adjustment for age. Undernutrition was found among 966 of the 5373 (18.0%) participants—young and old, men and women, and urban and rural. Of the study population, 216 of 5373 (4.0%) were obese; however, the prevalence of obesity was much higher (32.6%) among urban women 35 years or older (Figure 1).
Undernourished participants had lower mean blood pressure and lower cholesterol levels but higher mean glucose levels than did persons with a BMI of 18 to 30. Prevalence of impaired glucose tolerance was higher among undernourished persons, but no significant difference emerged in the prevalence of diabetes. On adjustment, mean glucose levels did not differ between undernourished and adequately nourished participants (Table 1).
Compared with persons with adequate BMI, obese persons had higher mean blood pressure and mean glucose, blood lipid, and uric acid concentrations and higher prevalence of hypertension and diabetes but reported less physical activity. These differences remained significant after adjustments. Smoking was less prevalent among obese participants. However, most obese persons were women, and most smokers were men. Of the obese women, 8.2% smoked, compared with 4.9% of the women of adequate weight (P = .04) (Table 1).
Of the obese persons, 110 of 216 (50.9%) had a waist-to-hip ratio in the upper quartile of their sex-specific population distribution. In each age group, urban women had the largest mean waist circumference. Mean waist circumference increased with age, especially in the urban population. After adjustments, waist circumference, but not waist-to-hip ratio, was significantly associated with blood pressure and cholesterol concentration. Increased waist circumference and waist-to-hip ratio were both significantly associated with elevated levels of glucose, triglycerides, and uric acid.
This study confirmed a “double burden of disease”: adult undernutrition is prevalent overall, whereas obesity is mainly confined to urban women 35 years or older. Both conditions are chronic, are detrimental to health and well-being, and lack an effective medical intervention. Although related to nutrition, both undernutrition and obesity are multifactorial in etiology.
The high prevalence of undernutrition, existing in all layers of society in the absence of war or famine, is disconcerting. Adult undernutrition reflects childhood stresses of undernutrition and disease and subsequent accommodation to a high-carbohydrate, low-protein diet, with hard physical labor.12 A study in Tanzania observed an association between low BMI and high glucose levels that was similar to the one we found; the authors believed it reflected a larger glucose load per kilogram rather than an etiologic role for undernutrition in the genesis of diabetes.13 Poverty, poor education, lack of nutritional knowledge, lack of availability of nutritious foods or poor quality of harvested foods, and coexisting (chronic) diseases all interact and lead to inadequate food intake. It is unlikely that health education alone has the potential to improve undernutrition.14
Obesity appears to be a problem of urban women. The association between urbanization and obesity in people of West African origin has been well documented,3 although it is less clear why this association is more marked among women. This sexual dimorphism points to behavioral factors, because most genetic and environmental factors are shared by men and women. Central obesity is the most common cause of insulin resistance, which is primary in the metabolic syndrome (obesity, glucose intolerance, hypertension, hyperlipidemia, hyperuricemia, and other metabolic abnormalities).15 Our data showed clear associations of obesity with hypertension, diabetes, and hyperlipidemia, suggesting emergence of the metabolic syndrome.
People in sub-Saharan Africa appear not always to perceive obesity as a health risk. The local perception of obesity and physical inactivity as signs of prosperity is common. To challenge such traditional beliefs is difficult when potential role models (including health workers) are obese.
Obesity is associated with lifestyle. Nothing can be done about age, sex, or genetic susceptibility, but lifestyle choices can be reconsidered. Also, lifestyle can be modified through centralized health policy. A population-wide intervention program promoting healthy lifestyle was implemented in Mauritius. Although the prevalence of obesity did not decrease, that of other cardiovascular risk factors (hypertension, smoking, inactivity, hyperlipidemia) did decrease.16
Individual counseling and intervention strategies enable early intervention and will be more effective than general health education messages. These strategies should be backed up by periodic screening for hypertension and glycosuria among obese people; such screening is affordable and practical in sub-Saharan African primary health care settings.17 The prevention of obesity would have major public health implications through the associated reduction in incidence of hypertension and diabetes.5
The major challenge is to develop effective preventive strategies for both undernutrition and obesity in a low-resource environment, which will require multisectorial commitment.
Note. BP = blood pressure; IGT = impaired glucose tolerance. aMean (SD) and percentages given. bAmong people 35 years or older, 2 hours after 75-g glucose load. cMeans and proportions weighted to reflect stratified subsample. *P < .05; **P < .01; ***P < .001, in comparison with the adequately nourished group. FIGURE 1— Distribution of body mass index (BMI) in the study population, by age group, sex, and area of residence: the Gambia, 1996–1997.
BMI < 18 18 < BMI < 30 BMI ≥ 30 (n = 966) (n = 4191) (n = 216) Weight, kg 46.0 (6.3) 58.9 (9.1) 91.8 (14.0) Height, cm 165.1 (9.3) 165.6 (8.7) 162.9 (7.5) Waist, cm 69.6 (5.1) 78.3 (7.5) 104.8 (11.1) Hip, cm 83.0 (5.2) 93.3 (7.2) 119.9 (10.6) Waist-to-hip ratio 0.84 (0.05) 0.84 (0.06) 0.88 (0.08) Smokers, % Current 19.5 19.0 4.2*** Ever 25.1 25.9 11.2*** Physically active, % 41.2 44.3 36.3* Systolic BP, mm Hg 116.3 (24.3)*** 121.3 (22.6) 130.9 (26.7)*** Diastolic BP, mm Hg 67.6 (12.7)*** 71.4 (12.0) 82.4 (12.3)*** Pulse, beats/min 82.4 (14.2)* 81.2 (14.3) 84.4 (11.7)** % BP ≥ 160/95 mm Hg 6.2 6.7 18.1*** % BP ≥ 140/90 mm Hg 14.4** 18.3 36.6*** (n = 164) (n = 750) (n = 160) Glucose,b mmol/L 6.4 (1.7)** 6.1 (1.8) 7.2 (2.7)*** IGT,b % 35.5*** 25.4 50.3*** Diabetes,b % 2.8 3.1 7.9*** (n = 164) (n = 750) (n = 160) Cholesterol,c mmol/L 3.7 (0.9) 4.0 (1.0) 5.1 (1.3)*** Triglycerides,c mmol/L 0.73 (0.33) 0.74 (0.35) 0.90 (0.49)* Uric acid,c mmol/L 0.24 (0.07)* 0.26 (0.08) 0.30 (0.09)** Creatinine,c μmol/L 64.8 (30.0) 71.7 (36.2) 71.6 (17.1) Elevated cholesterol,c % 4.7** 11.9 41.3*** Elevated triglyceride,c % 4.7 2.0 5.2** Elevated uric acid,c % 3.0 3.8 20.5*** Elevated creatinine,c % 4.6 8.9 10.7 
M. A. B. van der Sande coordinated the writing of the paper and contributed to the design, analysis, and supervision of the fieldwork. S. M. Ceesay and O. A. Nyan contributed to the design, analysis, and supervision of the fieldwork. P. J. M. Milligan, W. A. S. Banya, and A. Prentice contributed to the design and analysis of the fieldwork. K. P. W. J. McAdam contributed to the design of the fieldwork. G. E. L. Walraven coordinated the design and supervision of the fieldwork and contributed to the analysis. All authors contributed to the writing of the paper.
The study was funded by the Medical Research Council, United Kingdom.
We are grateful to the Department of State for Health for their support and cooperation during the study. We thank DUNN Nutrition Unit for their support in design and execution of the study and Prof Th. Thien and Dr W. Dolmans for their feedback on previous drafts of this paper.
