A

f

Ds

fi(Adw

i©

K

R

f

Léb

(Pl(cp

1d

Available online at

www.sciencedirect.com

Diabetes & Metabolism 38 (2012) 332–336

Original article

Prevalence of diabetes and associated risk factors in a Senegalese urban(Dakar) population

P. Duboz a,∗, N. Chapuis-Lucciani b, G. Boëtsch b, L. Gueye b

a UMR 6578 laboratoire d’anthropologie bioculturelle, CNRS, établissement francais du sang, faculté de médecine secteur Nord, université de la Méditerranée,boulevard Pierre-Dramard, 13955 Marseille cedex 15, France

b UMI 3189 environnement, santé, sociétés CNRS, université Cheikh Anta Diop, université de Bamako, CNRST laboratoire de physiologie exploratoire etfonctionnelle, faculté de médecine de Dakar, Dakar, Senegal

Received 14 December 2011; received in revised form 21 February 2012; accepted 26 February 2012

bstract

Aim. – The aim of this study was to estimate the prevalence of diabetes in the urban population living in Dakar, Senegal, and to investigate theactors associated with diabetes.

Methods. – Data from a 2009 survey of 600 individuals, aged 20 years or above and considered representative of the population of the city ofakar, were evaluated. Socioeconomic characteristics, hypertension, capillary whole blood glucose, and weight and height measurements of these

ubjects were collected during face-to-face interviews. The statistical analyses used chi-square (chi2) tests and binary logistic regressions.Results. – The percentage of participants with fasting blood glucose levels greater than or equal to 1.10 g/L and/or currently being treated

or diabetes was 17.9% (n = 107, 95% CI: 14.7–20.8). Observed rates of diabetes were significantly higher among women (chi2 = 6.3; P < 0.05),n subjects aged > 40 years (chi2 = 33.6; P < 0.001), in those with low educational levels (chi2 = 11.9; P < 0.05) and in those with hypertensionchi2 = 13.9; P < 0.001), and in those who were overweight (BMI ≥ 25 kg/m2 and < 30 kg/m2) or obese (BMI ≥ 30 kg/m2; chi2 = 40.3; P < 0.001).fter adjusting for gender, age, educational level, BMI and blood pressure, the results showed that gender, age and BMI were associated withiabetes: women, older people and those with a higher BMI had significantly greater chances of being diabetic than the rest of the population,hatever their blood pressure and educational level.Conclusion. – Diabetes is becoming a pressing public-health problem in Senegal, and the major risk factors for the increasing diabetes prevalence

n the city of Dakar are gender, age and body mass index. 2012 Elsevier Masson SAS. All rights reserved.

eywords: Biological anthropology; Epidemiology; Type 2 diabetes; Urban population; Prevalence; Obesity; Risk factors; Senegal

ésumé

Prévalence du diabète et des facteurs de risque associés dans une population urbaine du Sénégal (Dakar).Objectif. – L’objectif de cette étude était d’estimer la prévalence du diabète dans la population urbaine de Dakar (Sénégal) et de déterminer les

acteurs de risque associés au diabète dans cette population.Méthodes. – Un échantillon de 600 individus, représentatif de la population du département de Dakar âgée de 20 ans et plus, a été constitué.

es caractéristiques socioéconomiques des individus, la pression artérielle, la glycémie capillaire ainsi que le poids et la taille des individus ont2

té mesurés durant des interviews menées en face-à-face. Les analyses statistiques utilisées étaient des tests du Chi et des régressions logistiques

inaires.Résultats. – Le pourcentage d’individus dont la glycémie capillaire était supérieure ou égale à 1,10 g/L, et/ou sous traitement est de 17,9 %

ues étaient significativement plus importantes chez les femmes (Chi2 = 6,3 ;

n = 107) (IC à 95 % : 14,7–20,8). Les proportions observées de diabétiq < 0,05), les sujets âgés de plus de 40 ans (Chi2 = 33,6 ; P < 0,001), les sujets qui avaient un faible niveau d’éducation (Chi2 = 11,9 ; P < 0,05),

es hypertendus (Chi2 = 13,9 ; P < 0,001), ainsi que chez les individus en surpoids (25 kg/m2 ≤ BMI < 30 kg/m2) ou obèses (BMI > 30 kg/m2)Chi2 = 40,3 ; P < 0,001). Les régressions logistiques binaires ont montré que, toutes choses égales par ailleurs, le sexe, l’âge et l’indice de masseorporelle (IMC) étaient associés au diabète : les femmes, les sujets les plus âgés et les sujets qui présentaient un IMC élevé ont significativementlus de risque d’être diabétique que les autres.

∗ Corresponding author. Tel.: +221 77 310 77 77.E-mail address: prisciduboz@yahoo.fr (P. Duboz).

262-3636/$ – see front matter © 2012 Elsevier Masson SAS. All rights reserved.oi:10.1016/j.diabet.2012.02.011

à©

M

1

pgeita2ril2pratAd

i1lCedf[cas

irotg

cibbFsdm2ud

P. Duboz et al. / Diabetes & Metabolism 38 (2012) 332–336 333

Conclusion. – Le diabète devient actuellement un problème de santé publique majeur au Sénégal. Les principaux facteurs de risque de diabète Dakar (Sénégal) sont le sexe, l’âge et l’IMC.

2012 Elsevier Masson SAS. Tous droits réservés.

isque

lttbCtdf

2

2

ciSptpT(pFgT3wmGAsgsMu1t

adhhyay

ots clés : Anthropologie biologique ; Épidémiologie ; Obésité ; Facteurs de r

. Introduction

Diabetes is affecting people worldwide, and poses majorublic-health and socioeconomic challenges. Several reports oflobal estimates and projections have confirmed the diabetespidemic, and indicate that the numbers of people with diabetess destined to rise [1–4]. Estimates from 2011 by the Interna-ional Diabetes Federation (IDF) [1] suggest that the number ofdults with diabetes around the world will expand by 50% by030. The largest increases are expected to be in the developingegions of the world because of population ageing and urban-zation [2–4]. In 2010, 12.1 million people were estimated to beiving with diabetes in Africa, and this is projected to increase to3.9 million by 2030 [4]. Diabetes is thus becoming a pressingublic-health problem for sub-Saharan Africa (SSA). While theegion faces the double burden of communicable diseases (suchs HIV and tuberculosis) and non-communicable diseases andheir risk factors (such as diabetes) [5], the paucity of data fromfrica on diabetes is significant, as it considerably limits theevelopment of potential preventative strategies.

Type 2 diabetes accounts for more than 90% of diabetes casesn SSA [6,7]. Studies conducted from the 1960s to the early980s showed that the prevalence of type 2 diabetes was mostlyess than 1% except in South Africa (0.6–3.6%) and the Ivoryoast (5.7%) [8,9]. In fact, diabetes was regarded as a rare dis-ase in SSA prior to the 1990s [10]. However, unlike the past,iabetes is no longer rare in SSA: reported prevalences rangerom 3% in Benin to 14.5% in the Democratic Republic of Congo11]. Considerable regional variations persist within the Africanontinent [12] due, in part, to differences in diagnostic methodsnd criteria [13], and variation within countries is commonlyeen between urban and rural environments [7].

Previous suggestions that diabetes prevalence is increasingn Africa have been confirmed [7,13–19]. In SSA, the identifiedisk factors are not markedly different from those reported inther populations [13,14], including increasing age, urbaniza-ion, physical inactivity, hypertension and obesity [5], whereasender has little effect on diabetes.

Urbanization associated with the westernization of lifestyles,haracterized by decreased amounts of physical activity andncreased consumption of energy-dense or high-fat diets, cane considered a major determinant of the rising burden of dia-etes and other cardiovascular diseases in Africa [4,10,20,21].rom reported studies, the prevalence of diabetes varies con-iderably between rural and urban areas and, where examined,iabetes prevalence has been higher in urban than in rural com-unities in the same country [9,22]. Projections show that, by

050, 60% of Africans will live in cities, with a regional annualrban growth rate of 3.6% [23]. This suggests that the rates ofiabetes will continue to rise in SSA.

oIt

; Diabète de type 2 ; Sénégal ; Population urbaine ; Prévalence

As for Senegal, in 1960, Payet et al. [24] estimated the preva-ence of diabetes in Dakar to be 1.1%. In 2010, the IDF estimatedhe prevalence to be 4.7%. Today, no epidemiological informa-ion is available from Senegal on diabetes, but it is thought toe increasingly widespread, based on clinical observations [25].onsequently, and considering the previous findings on diabetes,

he aim of the present study was to estimate the prevalence ofiabetes in the urban population of Dakar and to investigate theactors associated with diabetes.

. Subjects, materials and methods

.1. Population sample

To carry out the present study, a comprehensive survey wasonducted from January to May 2009 in Dakar. In 2009, accord-ng to the National Agency of Statistics and Demographics ofenegal (Agence Nationale de la Statistique et de la Démogra-hie du Sénégal), a total of 990,019 individuals were living inhe environs of Dakar. The population sample selected for theresent study comprised 600 individuals age 20 years or above.he sample was constructed using the combined quota method

cross-section by age, gender and area of residence) to obtain aopulation aged 20 years or above and representative of Dakar.or this, data from the National Agency of Statistics and Demo-raphics of Senegal dating from the last census (2002) were used.he quota variables were gender (male/female), age (20–29,0–39, 40–49 and ≥ 50 years) and area of residence. These areasere grouped according to the four main divisions (arrondisse-ents) of the greater city of Dakar: Plateau-Gorée (five areas);rand Dakar (six areas); Parcelles Assainies (four areas); andlmadies (four areas). In practice, this method required con-

tructing a sample that reflected the proportions found in theeneral city population: for example, according to the last cen-us, 2% of men aged 20–29 years were living in the area of

edina (Plateau-Gorée arrondissement). Thus, our sample pop-lation was constructed to match this proportion by including2 men aged 20–29 years living in this area. The method washe same for each quota by gender, age and area.

In each area, four investigators (PhD students in medicinend pharmacy) started out from different starting points eachay to interview individuals in Wolof or French in every thirdome. Investigators had a certain number of individuals theyad to interview (women aged 20–29 years/men aged 20–29ears/women aged 30–39 years/men aged 30–39 years/womenged 40–49 years/men aged 40–49 years/women aged ≥ 50ears/men aged ≥ 50) in each area to meet the quotas. Only

ne person was selected as the respondent for each home.nvestigators entered these homes, asked to see the inhabi-ants and then chose the first person they saw who met the

3 Meta

cDt3t

2

2

l(es

2

tsdgpgew1vg

2

Ir

rTioIu

et

2

1Ft(Boa

2

n(ivtr

TD

V

G

A

E

H

B

34 P. Duboz et al. / Diabetes &

haracteristics needed for the quotas. The participation rate inakar was extremely high: 99.2% of the individuals agreed

o participate to the study. In-person interviews, lasting from0–45 min depending on the respondent’s availability and desireo talk, were then conducted.

.2. Variables studied

.2.1. Socioeconomic variablesThe socioeconomic and demographic variables col-

ected were: age (20–29/30–39/40–49/≥ 50 years); gendermale/female); educational level, defined in accordance with theducational system in Senegal (0/1–5/6–9/10–12/> 12 years ofchooling).

.2.2. DiabetesSubjects were examined in the morning after fasting since

he previous evening meal. The day before the investigation,ubjects were informed of the necessity to have nothing torink and eat, to properly measure capillary whole bloodlucose. Capillary whole blood was obtained from a fingerrick, and was immediately analyzed using a HemoCue® bloodlucose analyzer. Participants were then divided into two cat-gories according to international standards [26]: those whoere not diabetic, with capillary whole blood glucose less than10 mg/dL; and those with diabetes who had been either pre-iously diagnosed with diabetes or had capillary whole bloodlucose greater than or equal to 110 mg/dL.

.2.3. Blood pressureBlood pressure (BP) was measured, using an OMRON M5-

automatic BP monitor, in the sitting position, using theight upper arm and an appropriately sized cuff after a 5-min

tvui

able 1istribution of sociodemographic variables based on fasting blood glucose levels (n =

ariables Categories Fasting blood glucose

n %

ender Male 264 53.6

Female 229 46.4

ge range (years) 20–29 219 44.4

30–39 139 28.2

40–49 70 14.2

≥ 50 65 13.2

ducational level (years) 0 108 21.9

1–5 126 25.6

6–9 82 16.6

10–12 78 15.8

> 12 99 20.1

ypertension No 373 75.7

Yes 120 24.3

ody mass index(Kg/m2) < 18.5 67 13.6

≥ 18.5 and < 25 302 61.3

≥ 25 and < 30 92 18.7

≥ 30 32 6.4

Total 493 100.00

bolism 38 (2012) 332–336

est period. Three readings were taken during the interview.he first was discarded, and the mean of the last two read-

ngs used in the analysis. The first measurement was takenn both arms to detect any difference in BP between arms.f this was the case, the higher measure of arterial BP wassed.

Hypertension was defined as a systolic BP greater than orqual to 140 mmHg and/or a diastolic BP greater than or equalo 90 mmHg or reported treatment for hypertension.

.2.4. Body mass indexWeight was measured using a digital scale (accurate to within

00 g), with subjects wearing minimal clothing and barefoot.ollowing World Health Organization (WHO) recommenda-

ions, body mass index (BMI) was calculated by dividing weightkg) by the square of height (m2). Overweight was defined asMI greater than or equal to 25 kg/m2 and less than 30 kg/m2,besity corresponded to a BMI greater than or equal to 30 kg/m2

nd underweight was a BMI less than 18.5 kg/m2.

.3. Statistical analysis

Files compiled on the basis of the 600 participants’ question-aires were processed and coded in Excel (2007). Chi-squarechi2) tests were used to measure the presence, strength andndependence of the statistical association of sociodemographicariables, BMI and hypertension with diabetes. Binary logis-ic regression analysis was also carried out to estimate theisk factors for diabetes. Binary logistic regression sets out

o produce a model that permits the prediction of binaryariable values from a set of illustrative variables (contin-ous and/or binary). Each illustrative variable was adjustedn relation to the others. The association measure computed

600).

< 1.10 g/L) Fasting blood glucose ≥ 1.10 g/L Total

n %

43 40.2 30764 59.8 293

31 28.9 25016 14.9 15529 27.1 9931 29.1 96

34 31.8 14231 29.1 15721 19.6 10311 10.3 8910 9.2 109

62 57.9 43545 42.1 165

7 6.5 7440 37.4 34242 39.3 13418 16.8 50

107 100.00 600

P. Duboz et al. / Diabetes & Meta

Table 2Odds ratios (OR) adjusted for diabetes (n = 600).

Variables Categories P OR 95% CIfor OR

Gender (male) Female 0.047* 1.59 1.01–2.52Age (continuous) 0.003** 1.03 1.01–1.04Educational level (< 6years)

≥ 6 years 0.301 0.79 0.50–1.24

Systolic bloodpressure (continuous)

0.800 1.02 0.90–1.15

Body mass index(continuous)

0.001** 1.08 1.03–1.13

itrtA1s

3

lt(kh

sPstwotll(

BwBtl

4

evti1

tmtoBIrdho0pd

ewws(Bb(pbtoap

aiotcaiictAwpt[

cDohwt

5

* P < 0.05; **P < 0.005.

n this model was the odds ratio (OR), which quantifieshe strength of the association between an event’s occur-ence (represented by a dichotomous variable) and the factorshat could influence it (represented by illustrative variables).ll analyses were performed using SPSS software, version6. A two-sided P value < 0.05 was considered statisticallyignificant.

. Results

The percentage of individuals having fasting blood glucoseevels greater than pr equal to 1.10 g/L and/or currently beingreated for diabetes in our representative sample was 17.9%n = 107, 95% CI: 14.7–20.8). Ten (9.3%) of the 107 subjectsnew of their diabetes before screening, and only one (0.9%)ad a fasting blood glucose level less than 1.10 g/L.

As shown in Table 1, the observed proportions of diabeticubjects were significantly higher among women (chi2 = 6.3;

< 0.05) in subjects aged over 40 (chi2 = 33.6; P < 0.001), inubjects with low levels of education (chi2 = 11.9; P < 0.05), inhose with hypertension (chi2 = 13.9; P < 0.001), and in thoseho were either overweight (BMI ≥ 25 kg/m2 and < 30 kg/m2)r obese (BMI ≥ 30 kg/m2; chi2 = 40.3; P < 0.001). The rela-ionships between diabetes and gender, age, educationalevel, hypertension and BMI were also verified using binaryogistic regression, taking into account all variables studiedTable 2).

After adjusting for gender, age, educational level, BMI andP, the results showed that gender, age and BMI were associatedith diabetes: women, older-age subjects and those with a higherMI had significantly greater chances of becoming diabetic than

he rest of the sample population, whatever their educationalevel and systolic BP [1].

. Discussion

This study provides the first representative population-basedstimates of the prevalence of diabetes in Dakar. These data pro-ide baseline figures for the planning of healthcare policy and

he establishment of medical priorities in Dakar, the most urban-zed region of Senegal. Based on our results, it is estimated that7.9% of the urban population has diabetes. Not surprisingly,

st

bolism 38 (2012) 332–336 335

his high urban prevalence was higher than the IDF 2010 esti-ate, which included data from rural areas. The prevalence of

ype 2 diabetes in Dakar was also slightly higher than the rangef prevalence in other SSA countries, which varies from 3% inenin to 14.5% in the Democratic Republic of the Congo [11].

t is also noteworthy that, in comparison to the data from Dakareported by Payet et al. [24], it appears that the prevalence ofiabetes in the Senegalese capital has increased by 16.7% overalf a century. Furthermore, it appears that, in Dakar, awarenessf the diabetic condition and levels of control are low (9.3% and.9%, respectively), indicating the need to develop a nationalrogramme for the prevention, early detection and control ofiabetes.

On bivariate analysis, those with diabetes were older, lessducated, had hypertension more often, and were more over-eight and obese [22]. The prevalence of diabetes was greater inomen compared with men. However, binary logistic regression

howed that, after taking into account all the variables studiedgender, age, level of education, BMI and BP), gender, age andMI proved to be the principal risk factors for developing dia-etes. Therefore, the hypothesis that the demographic transitionthe increase in life expectancy leading to an increasingly elderlyopulation) is at the root of the drastic increase of diabetes haseen confirmed by our findings [5,13,22]. In addition, the facthat BMI is associated with diabetes highlights the importancef the impact of the nutritional transition [27] – which is associ-ted with a greater prevalence of obesity – that is currently takinglace in Dakar.

Finally, the acceleration of urbanization in Africa – and thessociated rise in obesity – plays an important role in thencreased prevalence of diabetes [11]. Indeed, the prevalencef diabetes in African communities is clearly increasing withhe lifestyle changes associated with rapid urbanization: dietaryhanges (greater consumption of refined sugars and saturated fat,nd a reduction in fiber intake); less physical activity; and grow-ng rates of obesity [28]. The rural community is still engagedn more physical occupations such as farming, walking andooking, and is probably more active than the urban popula-ion in general [28]. Thus, the projections that, by 2025, 70% offricans will live in cities suggest that levels of type 2 diabetesill continue to rise in the region [14]. In Senegal, 45% of theopulation is concentrated in Dakar and, by 2030, it is projectedhat more than half of all Senegalese will be living in urban areas29].

For these reasons, preventing obesity is important for redu-ing the onset of type 2 diabetes. In 2010, more than 30% ofakar citizens aged 20 years or above were either overweight orbese [30], and 27.5% had hypertension [31]. These data showow urgent it is to combat the rise of obesity and hypertension ife are to prevent the burden of chronic disease from becoming

oo heavy for Senegal to bear.

. Limitations and perspectives

The main limitation of the present study is its cross-ectional design, which does not allow longitudinal evaluationo examine how diabetes has evolved over time. For this

3 Meta

rlspetpop

6

flaAul2m

D

c

A

tgDR

R

[

[

[

[

[

[

[

[

[

[

[[

[

[

[

[

[

[

[

[

[

36 P. Duboz et al. / Diabetes &

eason, caution should be used when interpreting any causalinks between different risk factors (excess weight, obe-ity) and diabetes. Also, given the limited resources andoor infrastructure during the study, it was not possible tostablish the procedures to be followed when administeringhe 75g oral fasting blood glucose test. Nevertheless, theresent study is the first to provide data on the prevalencef diabetes in Senegal using a representative sample of theopulation.

. Conclusion

Diabetes is fast becoming a pressing public-health problemor SSA. The major risk factors for the increasing diabetes preva-ence in the city of Dakar, Senegal, are hypertension, overweightnd obesity. These risk factors are associated with urbanization.s there continues to be a growing number of people moving torban areas from the rural parts of Senegal, these results high-ight the emergence of an increasingly high prevalence of type

diabetes, with implications for an increased disease-relatedortality.

isclosure of interest

The authors declare that they have no conflicts of interestoncerning this article.

cknowledgments

The authors wish to thank all the subjects of Dakar who tookhe time to answer their questionnaires. We also thank the Sene-alese interviewers who participated in this study: Mohamedia, Alioune Badara Badgi, Ousmane Dieye and Mohamedassoul.

eferences

[1] Whiting DR, Guariguata L, Weil C, Shaw J. Global estimates of theprevalence of diabetes for 2011 and 2030. Diabetes Res Clin Pract2011;94:311–21.

[2] Wild S, Roglic G, Green A, Sicree R, King H. Global prevalence of dia-betes: estimates for the year 2000 and projections for 2030. Diabetes Care2004;27:1047–53.

[3] IDF. Diabetes Atlas. 3rd ed. Brussels: International Diabetes Federation;2006.

[4] IDF. Diabetes Atlas. 4th ed. Brussels: International Diabetes Federation;2009.

[5] Mbanya JC, Motala AA, Sobngwi E, Assah FK, Enoru ST. Diabetes insub-Saharan Africa. Lancet 2010;375:2254–66.

[6] Gill GV, Mbanya JC, Ramaiya KL, Tesfaye SA. Sub-Saharan Africanperspective of diabetes. Diabetologia 2009;52:8–16.

[7] Hall V, Thomsen RW, Henriksen O, Lohse N. Diabetes in Sub-SaharanAfrica 1999–2011: epidemiology and public health implications. A sys-tematic review. BMC Public Health 2011;14:11–564.

[8] McLarty DG, Swai AB, Kitange HM, Masuki G, Mtinangi BL, Kilima

PM, et al. Prevalence of diabetes and impaired glucose tolerance in ruralTanzania. Lancet 1989;1:871–5.

[9] Motala AA, Omar MA, Pirie FJ. Diabetes in Africa. Epidemiology of type1 and type 2 diabetes in Africa. J Cardiovasc Risk 2003;10:77–83.

[

bolism 38 (2012) 332–336

10] BeLue R, Okoror TA, Iwelunmor J, Taylor KD, Degboe AN, Agye-mang C, et al. An overview of cardiovascular risk factor burden insub-Saharan African countries: a socio-cultural perspective. Global Health2009;22:5–10.

11] WHO. STEPwise approach to chronic disease risk factor surveillance.http://www.who.int/chp/steps/reports/en/index.html. (Accessed 20th Octo-ber 2011).

12] Christensen DL, Friis H, Mwaniki DL, Kilonzo B, Tetens I, Boit MK, et al.Prevalence of glucose intolerance and associated risk factors in rural andurban populations of different ethnic groups in Kenya. Diabetes Res ClinPract 2009;84:303–10.

13] Sobngwi E, Mauvais-Jarvis F, Vexiau P, Mbanya JC, Gautier JF. Diabetesin Africans. Part 1: epidemiology and clinical specificities. Diabetes Metab2001;27:628–34.

14] Levitt NS. Diabetes in Africa: epidemiology, management and healthcarechallenges. Heart 2008;94:1376–82.

15] Mbanya JC, Ngogang J, Salah JN, Minkoulou E, Balkau B. Prevalence ofNIDDM and impaired glucose tolerance in a rural and an urban populationin Cameroon. Diabetologia 1997;40:824–9.

16] Olatunbosun ST, Ojo PO, Fineberg NS, Bella AF. Prevalence of diabetesmellitus and impaired glucose tolerance in a group of urban adults inNigeria. J Natl Med Assoc 1998;90:293–301.

17] Mbanya JC, Cruickshank JK, Forrester T, Balkau B, Ngogang JY, RisteL, et al. Standardized comparison of glucose intolerance in west African-origin populations of rural and urban Cameroon, Jamaica, and Caribbeanmigrants to Britain. Diabetes Care 1999;22:434–40.

18] Levitt NS, Unwin NC, Bradshaw D, Kitange HM, Mbanya JC, MollentzeWF, et al. Application of the new ADA criteria for the diagnosis of diabetesto population studies in sub-Saharan Africa. Diabet Med 2000;17:381–5.

19] Motala AA, Esterhuizen T, Gouws E, Pirie FJ, Omar MA. Diabetes andother disorders of glycemia in a rural South African community: prevalenceand associated risk factors. Diabetes Care 2008;31:1783–8.

20] Godfrey R, Julien M. Urbanisation and health. Clin Med 2005;5:137–41.21] Mayosi BM, Flisher AJ, Lalloo UG, Sitas F, Tollman SM, Bradshaw

D. The burden of non-communicable diseases in South Africa. Lancet2009;374:934–47.

22] Baldé NM, Diallo I, Baldé MD, Barry IS, Kaba L, Diallo MM, et al. Dia-betes and impaired fasting glucose in rural and urban populations in FutaJallon (Guinea): prevalence and associated risk factors. Diabetes Metab2007;33:114–20.

23] Population Division of the Department of Economic and Social Affairsof the United Nations Secretariat, World Population Prospects: The2008 Revision and World Urbanization Prospects: The 2009 Revision,http://esa.un.org/wup2009/unup/. (Accessed 20th October 2011).

24] Payet M, Sankalé M, Pène P, Bao O, Trellu H. The chief aspects of diabetesmellitus in an african environment at Dakar. Bull Soc Pathol Exot Filiales1960;53:901–10.

25] Gning SB, Thiam M, Fall F, Ba-Fall K, Mbaye PS, Fourcade L. Le diabètesucré en Afrique Sub-Saharienne. Aspects épidémiologiques, difficultés deprise en charge. Med Trop 2007;67:607–11.

26] World Health Organization. In: Definition, diagnosis and classification ofdiabetes mellitus and intermediate hyperglycemia. Geneva: World HealthOrganization; 2006.

27] Popkin BM. The nutrition transition and obesity in the developing world.J Nutr 2001;131:871S–3S.

28] Sobngwi E, Mbanya JC, Unwin NC, Kengne AP, Fezeu L, MinkoulouEM, et al. Physical activity and its relationship with obesity, hypertensionand diabetes in urban and rural Cameroon. Int J Obes Relat Metab Disord2002;26:1009–16.

29] United Nations, Department of Economic and Social Affairs, PopulationDivision. World Urbanization Prospects: The 2009 Revision; 2009.

30] Macia E, Duboz P, Gueye L. Prevalence of obesity in Dakar. Obes Rev

2010;11:691–4.

31] Duboz P, Macia E, Dia M, Ba A, Chapuis-Lucciani N, Gueye L. Prévalencede l’hypertension et facteurs associés dans le département de Dakar. DakarMed 2011;56:232–42.