A cross-sectional study of the prevalence and risk factors for hypertension in rural Nepali women.

0.21 MB·English

by  Khan, Rumana J

Khanetal.BMCPublicHealth2013,13:55
http://www.biomedcentral.com/1471-2458/13/55
RESEARCH ARTICLE Open Access
A cross-sectional study of the prevalence and risk
factors for hypertension in rural Nepali women
Rumana J Khan1, Christine P Stewart1, Parul Christian2, Kerry J Schulze2, Lee Wu2, Steven C LeClerq2,3,
Subarna K Khatry3 and Keith P West Jr2*
Abstract
Background: The prevalence ofhypertension is increasing in much of theSouth Asian region, including Nepal. This
paper reports theprevalenceand risk factors of hypertension and pre-hypertension among adult women in a rural
community ofNepal.
Methods: Cross-sectional data on socioeconomic status (SES), lifestyle factors and blood pressure (BP)were
collected from a cohort of15,934 women in rural Nepal in 2006–08.Among a subsample (n=1679), anthropometry
and biomarkers of cardiovascular risk were measured.
Results: Themeanageofwomenwas34.2years(range16.4-71.2years).Morethanthreepercent(3.3%)had
hypertensionand14.4%hadpre-hypertension.Inanadjustedanalysis,lowerSES,especiallylowerhouseholdfarm
assetsandstorageoffoodforlongtermconsumption,wasassociatedwithincreasedoddsofhypertension(OR=1.14
formid-levelSESandOR=1.40forlowSES;pfortrend<0.01).Smoking,alcoholuseandnotworkingoutsidethe
homewerealsoassociatedwithhigherrisk.Inasubsample,bothsystolicBP(SBP)anddiastolicBP(DBP)were
positivelyassociatedwithhightriglycerides(SBPβ=4.1mmHg;DBPβ=3.6mmHg),highHbA1c(SBPβ=14.0;DBP
β=9.2),raisedfastingglucose(SBPβ=10.0;DBPβ=6.9),highBMI(SBPβ=6.7;DBPβ=5.1)andhighwaist
circumference(SBPβ=6.2;DBPβ=5.3)afteradjustingforpotentialconfounders(pforall<0.01).
Conclusions:Althoughtheprevalenceofhypertensionwaslowinthiscohort,itwasmoreprevalentamongthe
poorerwomenandwasstronglyassociatedwithothercardiovascularrisks.Theseassociationsatarelativelyyoungage
mayconfergreaterriskforcardiovasculardiseaseamongwomeninlaterlife,indicatingtheneedforinterventionsto
reducetheprogressionfrompre-hypertensiontohypertension.
Keywords:Bloodpressure,Hypertension,Cardiovascularrisk,Nepal,Rural
Background considered a disease of the aged, where only 23% of
Low and middle-income countries bear a large burden deaths occur below the age of 70 years; however, in
of cardiovascular disease (CVD), accounting for 80% of South Asia, 52% of CVD deaths occur among people
the global CVD-related deaths and 87% of disability under 70 years [3,6]. Additionally, it has been estimated
adjusted-life years lost [1]. CVD rapidly has become a that the age of onset of acute myocardial infarction was
major cause of mortality and morbidity in low income an average of 6 years of age earlier in South Asia com-
South Asian countries as well [2-4]. In developed coun- pared tootherregions [7].This trend hasled tosubstan-
tries, age adjusted death rates from CVD are declining tial loss of potential human productive years. Early
due to preventive interventions and improved treat- identification and rigorous control of intermediate risk
ments [5]. Thus, CVD in developed countries is factors are needed to prevent and control CVD in this
partoftheworld.
Hypertension is one of the leading risk factors for
*Correspondence:[email protected] CVD and the prevalence of hypertension has been in-
2PrograminHumanNutrition,DepartmentofInternationalHealth
creasing in the South Asian region including Nepal [8].
Bloomberg,SchoolofPublicHealthJohnsHopkinsUniversity,615North
WolfeStreet,Baltimore21205,Maryland Despite rapid urbanization, about 83% of Nepal’s
Fulllistofauthorinformationisavailableattheendofthearticle
©2013Khanetal.;licenseeBioMedCentralLtd.ThisisanOpenAccessarticledistributedunderthetermsoftheCreative
CommonsAttributionLicense(http://creativecommons.org/licenses/by/2.0),whichpermitsunrestricteduse,distribution,and
reproductioninanymedium,providedtheoriginalworkisproperlycited.
Khanetal.BMCPublicHealth2013,13:55 Page2of10
http://www.biomedcentral.com/1471-2458/13/55
inhabitants live in rural areas [9]. Few studies have repeated. Mid-upper arm circumference (MUAC) was
attempted to describe the burden and determinants for measured on the upper left arm at the mid-point of the
hypertension in rural Nepal and such data are limited in acromion process and the tip of the olecranon using a
the South Asian context. Exploration of such data in standardinsertiontape.
rural Nepal will help to understand the etiology of CVD During a second visit among a subsample of 1679
in a population at the cusp of the epidemiologic and nu- women, anthropometric measurements were recorded,
trition transition, with findings thatmay begeneralizable including height (Harpenden stadiometer, Crosswell,
tootherpartsofruralSouth Asia. UK), weight (Seca 881, Hamburg, Germany), and waist
Theprimaryaimofthispaperistoexploretheprevalence circumference (WC) (Seca 200, Hamburg, Germany).
and risk factors for hypertension and pre-hypertension Women were then asked to fast overnight and were
among adult women, previous participants of a nutrition visited the following morning by a team of phleboto-
interventiontrial[10] ina rural community of Nepal. We mists to collect venous blood. Total cholesterol (TC),
alsoreporttheprevalenceofotherCVDriskfactors,such HDL cholesterol, triglycerides (TG), and fasting glucose
asobesity,cholesterol,triglycerides,andHbA1c,andtheir were measured in plasma specimens (LDX Analyzer,
associationwithbloodpressure. Cholestech, Hayward, CA). Of the women who con-
sented totheblood draw(n=1447),251(17.3%) hadnot
Methods fasted, defined as no food or drink other than water
This study utilizes cross-sectional data from a follow-up within 8 hours of the blood draw. Glucose data was only
of women who had been enrolled in a double blind, pla- analyzedamongwomen who were fasting(n=1196).
cebo controlled, cluster randomized trial of vitamin A or Continuous variables such as age, SBP, DBP, MUAC,
β carotene supplementation provided to women for height, weight, BMI and WC were checked for normal-
three years (1994–1997). Details of the trial were pub- ity, outliers and missing values. Blood pressure and lipid
lished previously [10]. Briefly, the study was conducted parameters were categorized using standard cutoffs
in the rural, low-lying Sarlahi District of Nepal. The [12,13]. Hypertension was present if SBP and/or DBP
study area comprises 30 Village Development Commu- were ≥140/90 mm Hg [12]. Pre-hypertension was
nities (VDCs), each divided into 9 administrative wards. defined by SBP ≥120 mmHg but <140 mmHg and/or
Over a study period from July 1994 to June 1997, 17,531 DBP≥80 mmHg but <90 mmHg [12]. The cutoff for
infants were born to women enrolled in the trial. high TC was ≥5.17 mmol/L (200 mg/dL); high TG was
Women were supplemented before, during and after ≥1.7 mmol/L (150 mg/dL); low HDL cholesterol was
pregnancy throughout the study period. A subsample <1.03 mmol/L (40 mg/dL), raised fasting glucose
area was selected for more intensive monitoring during was ≥5.6 mmol/L (100 mg/dL) and high HbA c was
1
the trial. The area contained roughly 10% of the study ≥6.5% [14]. Overweight was defined as BMI ≥23 kg/m2
population and selection was based on ease of access to and abdominal obesity defined as WC ≥80 cm [15-17].
clinics and a paved road in order to facilitate biochem- Women were also classified according to the Inter-
ical sample collection and to visit women at birth. The national Diabetes Federation’s worldwide definition of
cohort of women who became pregnant during the trial the metabolic syndrome which includes a WC ≥80 cm
period have been followed over time along with their plus≥2 of the followings: 1) raised triglycerides:
children to monitor long-term health outcomes [11]. ≥1.7 mmol/L; 2) reduced HDL cholesterol: <1.03 mmol/
From 2006–2008, women were followed-up (n=16,469) L; 3) raised blood pressure: systolic BP ≥130 or diastolic
during a series of up to three household visits. The BP ≥85 mm Hg; and 4) raised fasting plasma glucose:
current analysis uses the data from this follow-up period ≥5.6mmol/L[16].
(n=15,934), excluding women who were currently SES was analyzed using Principal Component Analysis
pregnant (n=470) or who did not know their pregnancy (PCA) by examining 17 questions on SES, extracting
status (n=65). three components with eigen values >1 which explained
Duringthefirstvisit,fieldworkersconductedinterviews 41%ofthe variance.One componentrepresentedasum-
tocollectinformationonhouseholdsocioeconomicstatus, mary measure of household quality and status, with the
literacy, ethnicity and occupation. Information on their most heavily weighted variables including type of roof
smoking status and alcohol consumption were also col- and walls, type of latrines, number of servants in the
lected. Blood pressure was measured four times on the house, number of rooms in the house, having electricity,
right arm using an automated measurement device (BPM and owning a motorcycle. The secondcomponent repre-
300, BPTrue, Canada). The first measure was dropped sented household farming assets and food storage, with
andthemeanofthelastthreewasusedforanalysis.Ifthe the most heavily weighted variables being kilograms of
mean systolic or diastolic blood pressure (SBP and DBP, rice stored per household member, duration of time rice
respectively) were≥140/90, the measurements were stores would last, ownership of cultivable land, and
Khanetal.BMCPublicHealth2013,13:55 Page3of10
http://www.biomedcentral.com/1471-2458/13/55
Table1SelectedcharacteristicsofadultNepaliwomeninthecohortin2006–2008bybloodpressurecategories
Missingvalues Hypertension(n=530) Pre-hypertension Normalblood P-valuea
(n=2296) pressure(n=13,108)
n n(%)
Age(years) 11 <0.0001
16-30 109(1.9) 657(11.7) 4853(86.4)
31-45 366(3.8) 1503(15.6) 7742(80.6)
46andabove 55(7.9) 134(19.3) 504(72.7)
Smoking 11 <0.0001
No 353(2.9) 1745(14.1) 10246(83.0)
Yes 177(4.9) 550(15.4) 2852(79.7)
Alcohol 12 <0.0001
No 426(3.0) 1922(13.6) 11789(83.4)
Yes 104(5.8) 372(20.8) 1309(73.3)
MUAC(cm)b 0 23.7(3.0) 23.7(2.9) 23.0(2.8) <0.0001c
Pahadi/Madheshi 38 <0.0001
Pahadi 200(3.9) 877(17.0) 4068(79.1)
Madheshi 330(3.1) 1410(13.1) 9011(83.8)
Occupation 16 <0.0001
Doesnotwork 126(4.3) 468(16.1) 2310(79.5)
Worksoutsidethehome 404(3.1) 1826(14.0) 10784(82.9)
Literacy 8 0.991
No 451(3.4) 1949(14.4) 11120(82.2)
Yes 79(3.3) 347(14.4) 1980(82.3)
Landownership 63 <0.0001
No 254(3.8) 1027(15.3) 5414(80.9)
Yes 275(3.0) 1253(13.7) 7648(83.3)
Livestock 23 <0.0001
No 116(5.1) 368(16.1) 1799(78.8)
Yes 414(3.0) 1921(14.1) 11293(82.9)
TV 31 0.104
No 364(3.3) 1549(14.0) 9143(82.7)
Yes 166(3.4) 739(15.2) 3942(81.3)
Electricity 23 0.357
No 270(3.4) 1107(14.0) 6529(82.6)
Yes 260(3.2) 1181(14.8) 6564(82.0)
CompositeindicatorsofSES 577
Householdquality
High 167(3.2) 778(15.1) 4218(81.7) 0.332
Medium 178(3.5) 695(13.7) 4218(82.9)
Low 168(3.3) 731(14.3) 4204(82.4)
HouseholdAssets 0.326
High 177(3.4) 768(14.8) 4227(81.7)
Middle 177(3.5) 746(14.6) 4198(82.0)
Low 159(3.1) 690(13.6) 4215(83.2)
Khanetal.BMCPublicHealth2013,13:55 Page4of10
http://www.biomedcentral.com/1471-2458/13/55
Table1SelectedcharacteristicsofadultNepaliwomeninthecohortin2006–2008bybloodpressurecategories
(Continued)
Householdfarmingandfoodstoraged
High 142(2.8) 713(13.8) 4305(83.4) <0.0001
Medium 166(3.3) 685(13.4) 4253(83.3)
Low 205(4.0) 806(15.6) 4082(80.1)
Abbreviations:SES=SocioeconomicStatus;MUAC=MidUpperArmCircumference.
aPearsonChisquareexceptwherenoted.
bMean(SD).
cOnewayANOVA.
dpforchisquarelineartrend<0.001forhypertensionandpre-hypertension.
ownership of cattle, goats and bullock carts. The third adjusted for other predictors found to be significantly
component represented other household assets, with the associated with hypertension in the full cohort. In
most heavily weighted variables including bicycle, TV, addition, models examining the lipid profile, glucose and
radio and wristwatch ownership. Each of these three HbA cwerefurtheradjustedfor BMI.
1
components of SES were categorized into tertiles based The follow-up study was approved by the Institutional
ontheircomponent scores. Review Boards at the Johns Hopkins Bloomberg School
The primary outcome measure for this analysis was of Public Health and the Institute of Medicine in
blood pressure, evaluated both as a dichotomous and a Kathmandu, Nepal.
continuous variable. The mean differences of BP among
supplementation groups (vitamin A, β carotene or pla- Results
cebo), were within 0.5 mm Hg and the prevalence of The study sample included 15,934 women with a mean
hypertension and pre-hypertension did not differ be- age of 34.2 years (Standard deviation or SD=5.9), 60% of
tween groups. Thus, all intervention groups were com- whomwerebetweentheageof31–45years.Roughly,one
bined for this analysis. For the full sample, simple and third of them were Pahadi and the rest were Madheshi.
multiple logistic regression analyses were done to study Most women could not read or write in any language
the relationship between each independent variable with (84.9%) and most worked outside the home (81.7%).
the outcomes of hypertension or pre-hypertension. The Approximately42.2%didnotownanyland;49.7%hadno
independent variables included the three SES factors, age, electricity; and 14.5% owned no livestock. Around 22.5%
smoking status (yes/no), alcohol consumption (yes/no), ofwomenweresmokersand11.2%werealcoholdrinkers.
MUAC (in cm),ethnicity (womenofPahadi ethnicityhis- TheirmeanMUACwas23.1cm(SD=2.7).
torically originated from the hill areas of Nepal while The mean SBP was 104.3 mm Hg (SD=12.3) and ran-
those of Madheshi ethnicity originated from the plains) ged from 68 to 216 mm Hg and the mean DBP was
and occupation, which was grouped into two categories: 70.9 mm Hg (SD=9.3) and ranged from 48 to 122 mm
1) no reported work or study or 2) work or study outside Hg. A total of 530 (3.3%) women had hypertension
home,includingfarmers,unskilledorcontractedlaborers, while 2296 (14.4%) had pre-hypertension. Participant
business,privateorgovernmentservice,andstudents.For characteristics and selected SES variables by BP
both the regression models, effect modification by ethni- category are presented in Table 1. Notably,therewasa
city and age were evaluated with likelihood ratio tests by trend of increasing prevalence of hypertension and pre-
comparingtwonestedmultivariatemodelswithandwith- hypertension with increasing age and higher MUAC.
out the interaction term. Theresultswerestratified ifany Women with hypertension or pre-hypertension were
significantinteractionwasfound. more likely to be smokers and twice as likely to consume
For the subsample, blood pressure was analyzed con- alcohol.Pahadiwomenhadagreaterprobabilityofhyper-
tinuously, as the sample size was smaller and the num- tension (4% versus 3%) or pre-hypertension (17% versus
ber of women with hypertension was relatively low. In 13%) than Madheshi women. Women with hypertension
order to examine the associations of SBP and DBP with or pre-hypertension were also more likely to come from
other cardiovascular risk factors (highTC, highTG, low households that did not own any land or livestock.
HDL cholesterol, raised fasting glucose, high HbA c, Womenworkingoutsidethehomehadalowerprevalence
1
high BMI, high WC), simple linear regression was per- of both hypertension and pre-hypertension than women
formed with SBP or DBP as the outcome variable and whostayedathome.
each of these risk factors as predictors. Those with nor- Table 2 and Table 3 show the unadjusted and adjusted
mal values of the predictor variables were considered as relationships of BP categories with socioeconomic and
a reference group. Multivariable regression models were other predictors. In the adjusted analysis (Table 2) two
Khanetal.BMCPublicHealth2013,13:55 Page5of10
http://www.biomedcentral.com/1471-2458/13/55
Table2Associationbetweenselectedriskfactorsand stratified the findings for pre-hypertension by ethnicity
prevalenceofhypertensionamongadultNepaliwomen (Table 3). In the case of pre-hypertension, Madheshi
(n=13,638) women with low farming assets and food storage had a
Hypertensiona significant31%increasedriskcomparedtowomenofhigh
OddsRatio(95%CI) status (p for trend<0.001). This association was not sig-
Unadjusted Adjustedb nificant for Pahadi women. However, for Pahadi women
those with low household quality and low household
SES:Householdquality
assetshadincreasedoddsofhavingpre-hypertensionthan
High Reference -
those with high household quality (OR, 1.23 and 95% CI,
Middle 1.06(0.86to1.32) 1.09(0.87to1.37)
1.03-1.52) and high household assets (OR, 1.27 and 95%
Low 1.01(0.81to1.26) 1.00(0.78to1.26) CI, 1.04-1.55). In addition to different indicators of
SES:Farmingandfood SES, age, alcohol drinking, MUAC and occupation
High Reference - remained significantly associated both with hyperten-
sion and pre-hypertension in multivariable models.
Middle 1.18(0.94to1.49) 1.14(0.90to1.45)
Smoking remained significant only for hypertension and
Low 1.52(1.22to1.89) 1.40(1.12to1.76)
notforpre-hypertension.
SES:HouseholdAssets
High Reference -
Cardiovascularriskfactorsinthesubsample
Middle 1.01(0.81to1.24) 1.05(0.84to1.31)
Among the subsample of 1679 women, there was a low
Low 0.90(0.72to1.12) 0.92(0.73to1.15)
prevalence of high TC and raised fasting glucose, yet a
Age high prevalenceof low HDL cholesterol (Table 4). About
16to30 Reference - 10.2% women had high TG, about 12.0% were over-
31to45 2.10(1.69to2.61) 2.00(1.60to2.52) weight and 14.0% of women had abdominal obesity.
Only 5% of women had metabolic syndrome. The mean
46andabove 4.86(3.47to6.80) 4.30(3.09to6.26)
SBP and DBP were higher among women with highTC,
Smoking
high TG, high HbA c, high BMI and high WC and
1
No Reference -
raised fasting glucose compared to women with normal
Yes 1.80(1.50to2.17) 1.31(1.10to1.58) values (Table4).
Alcohol These differences in SBP and DBP were statistically
No Reference - significant (p≤0.01) for TG, HbA1c, fasting glucose,
BMI and WC. The greatest differences were between
Yes 2.20(1.76to2.75) 1.53(1.23to1.90)
women with either high HbA c (SBP β=14.5 mm Hg;
MUAC(cm) 1.09(1.06to1.13) 1.08(1.05to1.12) 1
DBP β=9.5 mm Hg) or raised fasting glucose (SBP
Ethnicity β=11.8 mm Hg; DBP β=8.0 mm Hg) compared to
Pahadi Reference - women with normal values. Though the beta coefficients
Madheshi 0.75(0.62to0.89) 1.03(0.83to1.27) were slightly attenuated in multivariate models control-
Occupation ling for factors found to be associated with hypertension
in the full cohort, the association withTG, HbA c, fast-
Doesnotwork Reference - 1
ing glucose, BMI and WC remained highly significant.
Worksoutsidehome 0.69(0.56to0.84) 0.65(0.52to0.81)
TG, fasting glucose and HbA c remained significantly
1
Abbreviations:SES=SocioeconomicStatus;MUAC=MidUpperArm
associated with both SBP and DBP even after further ad-
Circumference.
aHypertensiondefinedassystolicbloodpressure≥140mmHgand/ordiastolic justmentforBMI(datanot shown).
bloodpressure≥90mmHg.
bAdjustedforallothervariablesinthetable.
Discussion
measuresofSES(householdqualityandhouseholdassets) In this sample of relatively young, rural Nepalese women,
were not associated with hypertension, but household the prevalence of hypertension and pre-hypertension was
farming assets and food storage was. Women with low 3.3% and 14.4%, respectively. Household SES was
status in this latter SES indicator had a significant 40% associated with hypertension, notably indicators of
increased odds of hypertension compared to women of household farming assets and food storage. Hyperten-
highstatus. sion was more prevalent among older women, smo-
We found a significant interaction between ethnicity kers and alcohol drinkers, but significantly less
and socioeconomic status while assessing the association prevalent among women who worked outside the
ofpre-hypertensionanditspredictors.Wethereforehave home. Although a low percentage of women had high
Khanetal.BMCPublicHealth2013,13:55 Page6of10
http://www.biomedcentral.com/1471-2458/13/55
Table3Associationbetweenselectedriskfactorsandprevalenceofpre-hypertensionamongadultNepaliwomen
(n=15,404)
Pre-Hypertensiona
OddsRatio(95%CI)
Pahadi Madheshi
(n=4945) (n=10,421)
Unadjusted Adjustedb Unadjusted Adjustedb
SES:Household quality
High Reference - - -
Middle 1.08(0.90to1.29) 1.04(0.86to1.27) 0.84(0.73 to0.98) 0.92(0.80 to1.07)
Low 1.23(1.03to1.46) 1.23(1.03to1.52) 0.89(0.75 to1.01) 0.93(0.79 to1.08)
SES:Farmingandfood
High Reference - - -
Middle 0.87(0.71to1.03) 0.78(0.69to1.12) 1.04(0.90to1.20) 1.09(0.95to1.27)
Low 1.03(0.86to1.24) 0.95(0.78to1.15) 1.28(1.12to1.47) 1.31(1.14to1.55)
SES:HouseholdAssets
High Reference - - -
Middle 1.11(0.93to1.32) 1.11(0.9to1.33) 0.94(0.81to1.08) 0.99(0.85to1.14)
Low 1.25(1.04to1.51) 1.27(1.04to1.55) 0.81(0.67to0.96) 0.86(0.74to1.00)
Age
16to30 Reference - - -
31to45 1.74(1.48to2.05) 1.67(1.40to1.98) 1.30(1.14to1.47) 1.34(1.17to1.52)
46andabove 1.58(1.11to2.25) 1.63(1.12to2.34) 2.23(1.72to2.88) 2.47(1.88to3.24)
Smoking
No Reference - - -
Yes 1.15(0.99to1.34) 0.96(0.80to1.16) 1.05(0.87to1.16) 0.93(0.79to1.10)
Alcohol
No Reference - - -
Yes 1.55(1.33to1.81) 1.57(1.30to1.88) 1.51(1.07to2.12) 1.46(1.02to2.10)
MUAC(cm) 1.07(1.04to1.10) 1.11(1.08to1.14) 1.08(1.06to1.11) 1.08(1.05to1.10)
Occupation
Doesnotwork Reference - - -
Worksoutsidehome 0.89(0.72to1.10) 0.82(0.65to1.03) 0.77(0.67to0.88) 0.80(0.70to0.92)
Abbreviations:SES=SocioeconomicStatus;MUAC=MidUpperArmCircumference.
aPre-hypertensiondefinedbysystolicbloodpressure≥120mmHgbut≤140mmHgand/ordiastolicbloodpressurewas≥80mmHgbut≤90mmHg.
bAdjustedforallothervariablesinthetable.
total cholesterol, 73% of women had low HDL choles- reported prevalence of hypertension in rural areas of
terol and about 10% had high triglycerides. BP was India, Bangladesh and Pakistan is in the range of 4.5 -
not associated with cholesterol levels, but was strongly 22% [21-25]. These studies varied in the included age
associated with triglycerides, BMI, WC, fasting glucose groups, study settings, and criteria for classifying hyper-
andHbA c. tension. Nevertheless, the prevalence of hypertension
1
A number of reports suggest that cardiovascular dis- among our cohort based on a cut-off of 140/90 mm Hg
ease and hypertension are rapidly increasing both in was 3.3%, one of the lowest reported estimates among
urban and rural areas of South Asia [3,4,18,19], yet there rural communities in South Asia. One likely explanation
have been few population based studies and prevalence is the young age of our participants, in contrast to many
estimates vary widely. The only evidence from rural of the aforementioned studies which were among older
Nepal comes from a 1981 study in which the prevalence participants. None of the previous studies reported the
of hypertension among people aged 20 or more was prevalence of pre-hypertension. Individuals with pre-
about 6% in the hill areas and 8% in the plains [20]. The hypertension have a high likelihood of progression to
Khanetal.BMCPublicHealth2013,13:55 Page7of10
http://www.biomedcentral.com/1471-2458/13/55
Table4AssociationbetweencardiovascularriskfactorsandmeanbloodpressureamongadultNepaliwomen(N=1679)
Riskfactorsa Prevalence SystolicBP DiastolicBP
N(%) Betac(95%CI) Pvalue Betac(95%CI) Pvalue
Hightotalcholesterol 23(1.6)
Unadjusted 3.9(-0.5to8.4) 0.085 1.8(-1.8to5.5) 0.320
Multivariateb 4.4(-0.5to9.3) 0.080 1.8(-2.2to5.8) 0.880
Hightriglycerides 147(10.2)
Unadjusted 4.3(2.4to6.1) <0.001 3.7(2.2to5.2) <0.0001
Multivariateb 4.1(2.1to6.0) <0.001 3.6(2.1to5.2) <0.0001
LowHDL-cholesterol 1046(72.3)
Unadjusted -0.6(-1.8to0.7) 0.374 -0.7(-1.6to0.3) 0.190
Multivariateb -0.5(-1.8to0.8) 0.464 -0.5(-1.6to0.6) 0.351
Raisedfastingglucose 15(1.3)
Unadjusted 11.8(6.2to17.4) <0.001 8.0(3.5to12.6) <0.0001
Multivariateb 10.0(4.6to15.4) <0.001 6.9(2.5to11.3) <0.0001
HighHbA c 9(0.6)
1
Unadjusted 14.5(7.4to21.6) <0.001 9.5(3.7to15.3) 0.003
Multivariateb 14.0(6.9to21.0) <0.001 9.3(3.5to14.9) 0.004
Highbodymassindex 214(12.7)
Unadjusted 6.9(5.3to8.4) <0.001 5.1(3.9to6.3) <0.0001
Multivariateb 6.7(5.0to8.4) <0.001 4.8(3.5to6.2) <0.0001
Highwaistcircumference 234(14.0)
Unadjusted 6.6(5.0to8.1) <0.001 5.3(4.1to6.5) <0.0001
Multivariateb 6.2(4.6to7.8) <0.001 5.2(3.9to6.4) <0.0001
Missingfortotalcholesterol=232,triglycerides=232,HDLcholesterol=232,fastingbloodglucose=483,HbA1c=234,BMI=10,waistcircumference=2.
aCut-offsforhightotalcholesterol=5.17mmol/L,hightriglycerides=1.7mmol/L,lowHDLcholesterol=1.04mmol/L,highBMI=23,abdominalobesity=waist
circumference≥80cm,Raisedfastingglucose=5.6mmol/LandhighHbA1c=7.Thosewithnormalvalueswereconsideredasthereferencegroup.
bAnalysisdoneusingsimpleandmultivariatelinearregression.Themultivariatemodelsadjustedforage,cigarettesmoking,alcoholconsumption,household
farmingandfoodassets,ethnicity,andoccupation.
cBetacoefficientsrepresentthemeandifferenceinsystolicordiastolicbloodpressurecomparingthosewithabnormalvaluesofeachpredictorvariabletothose
withnormalvalues.
hypertension over the subsequent 5 years [26]. We have variability we created SES indices on three different
also seen a significant positive association between age aspects. The SES component representing household
and prevalent hypertension and it is likely that many of farming assets and food storage (such as land, goats, cat-
the women who currently have pre-hypertension will tle, bullock carts, and per capita amount of rice stored
progress to hypertension as they age. Moreover, pre- in the household) was negatively associated with hyper-
hypertension itself is also a risk factor for cardiovascular tension. This SES component was also negatively asso-
diseaseandcardiovascular mortality [27]. ciated with pre-hypertension for Madheshi women. This
To explore the determinants of hypertension, several finding might suggest that in this rural setting in Nepal,
researchers have investigated the association with SES. those who are more food insecure are more vulnerable
In rural India and Bangladesh, most studies reported a to non-communicable diseases, a finding that has been
positive relationship between SES and hypertension reported in developed countries [32,33]. The other two
[22,25,28,29], while only one found the opposite [30]. aspects of SES, household quality and household
Defining SES is complex in the developing world. Many assets, were not associated with hypertension but
studies combined data on education, occupation, hous- showed a negative and significant relationship with pre-
ing quality, land ownership, durable goods, income and hypertension for Pahadi women. While it is unclear why
livestocktocreateacompositeindex ofSES.Creatingan two of the factors were more strongly associated within
overall index in this manner might not reflect the true one group and the third factor more strongly associated
relationship between SES and hypertension consistently within the other ethnic group, the general pattern
as different components in the index could have varying appears to be that low SES is associated with high blood
influences on hypertension [31]. To capture this pressure for women of both the groups. Our findings do
Khanetal.BMCPublicHealth2013,13:55 Page8of10
http://www.biomedcentral.com/1471-2458/13/55
not support the finding from other studies in rural Our evidence suggests that modifiable risk factors like
South Asia where higher SES were associated with obesity, lipid abnormalities and glucose intolerance tend
increased prevalence of hypertension. Rather, these find- to coexist with raised BP and may begin in early life.
ings indicate that in the context of rural women in a low This finding implies that screening for and managing
income country, having a higher SES is a protective fac- multiple risk factors should be considered instead of
tor for health. We examined literacy and occupation as only a single one. In contexts such as these, early life
separate variables from the SES scores in our multiva- malnutritionmayalsobeassociatedwithanelevatedrisk
riate models. We observed that those working outside of CVD, particularly if followed by later life overweight
home had a lower risk of hypertension than those who or obesity [51]. It is important, therefore, to consider
did not work or worked in the home. While it is un- health andnutrition riskfactors across thelife course.
known whythis relationshipmay exist,wespeculate that
this variable might be a proxy for regular physical acti- Conclusion
vity or it may suggest that women going outside of the Although the prevalence of hypertension was low in this
home regularly have more chance to interact with differ- relatively young cohort, nearly 18% were classified as ei-
ent people and are more aware of health irrespective of ther pre-hypertensive or hypertensive. We found a nega-
theirliteracyorSES. tive association between SES and hypertension, which
We have found that a total of 72.3% of women had could be an indication of the maturation of CVD epi-
low HDL cholesterol, a pattern of dyslipidemia that was demic in that region. This epidemic has to be prevented
also observed in studies mainly from India, where the because across the South Asian region, it has been pre-
prevalenceoflowHDLcholesterolhasrangedfrom41- dicted that if rigorous and early preventive measures are
68% [34-37]. These studies also reported that the not taken, treatment costs of CVD will become a sub-
prevalenceofhighTGwasbetween10-48%.Inaddition, stantial burden on national economies [19]. It is also
arecent studyofurbanNepalesewomenreporteda simi- predicted that a reduction in the population distribution
lar pattern-- 40%ofthemhad lowerHDLcholesteroland of systolic blood pressure of 2 mm Hg results in 6%, 4%
36% had high TG [38]. Such lipid patterns observed in and 3% reductions in 1-year stroke related mortality,
South Asia are different from patterns in Western coun- coronaryheartdiseaserelatedmortalityandoverallmor-
tries. By comparison, a nationally representative sample tality, respectively [52]. It also has been projected that
from the United States reported that 30% of females had reducing chronic disease mortality even by only 2% per
low HDL cholesterol and 21% had highTG [39]. Some of year by 2015 could save 10% of the expected loss of in-
the apparent difference between regions could be come and around $8 billion collectively in South Asian
explainedbythewideusageandavailabilityoflipidlower- andothermiddle orlowincome countries[19].Wehave
ing drugs in developed world or could be reflective of a shownthatdifferentcardiovascular riskfactorsincluding
distinct South Asian diet pattern largely dependent on blood pressure tend to be strongly associated with each
carbohydrates,whichmaycontributetohypertriglyceride- other. Identification ofthese risk factors at an early stage
miaandlowerHDLcholesterol[40]. of life is an important opportunity for primary preven-
The prevalence of raised fasting glucose was somewhat tion of hypertension through lifestyle modification to
lower in our study (1.3%) than other reports from rural prevent disease progression. We hope that these data
Nepal and India (~3.5%) [41-43]. We found that 14.0% will contribute to the evidence of CVD risk in a poor,
women had abdominal obesity, lower than the 32% preva- rural population, helping to devise policies to prevent
lencereportedbyanotherstudyofruralIndianwomen[44] disease in this rural population of Nepal as well as in
and50%prevalenceofurbanNepalimen[45].Yet,mostof othersimilarsettingsinSouth Asia.
the measured metabolic factors (TG, BMI, WC, fasting
plasma glucose and HbA c) were positively and strongly Competinginterests
1 Theauthors’declarethattheyhavenocompetinginterest.
associated with blood pressure. The strongest relationship
wasforglucoseabnormalities(fastingglucoseandHbA c)– Authors’contributions
1
women with high glucose or high HbA c had a mean SBP RJK,CPS,PCandKPWconceptualizedtheanalysisplanforthispaper.RJK
that was 11–14 mm Hg higher than t1hose with normal andCPSwrotethepaperandperformedthestatisticalanalysistogether
withLSFW.CPS,SCLandSKKwereinvolvedinthedatacollectionand
values.Thesefindingsagreewithpreviousinvestigations projectmanagement.KJSoversawthelaboratoryanalysisandreviewedthe
on the tendency of hypertension to occur concurrently manuscript.KPWservedasthePIfortheoriginalstudyandsubsequent
follow-up.Allauthorshavereviewedandapprovedthefinalmanuscript.
with other metabolic disorders [46-48]. The combin-
ation of hypertension with lipid abnormalities, glucose Acknowledgements
intolerance and abdominal obesity comprise metabolic ThisstudywassupportedbyGrant(#614)fromtheBill&MelindaGates
Foundation,Seattle,WAandtheNationalInstitutesofHealth
syndrome [49], which increases the risk of cardiovascu-
(1R03HD062634).TheoriginalNepalNutritionInterventionProject-Sarlahi
lardiseasesanddiabetes[50]. (NNIPS)-2trialwassupportedthroughtheVitaminAforHealthCooperative
Khanetal.BMCPublicHealth2013,13:55 Page9of10
http://www.biomedcentral.com/1471-2458/13/55
Agreement(HRN-A-00-97-00015-00)betweenJohnsHopkinsUniversityand 19. AbegundeDO,MathersCD,AdamT,OrtegonM,StrongK:Theburdenand
theOfficeofHealth,InfectiousDiseasesandNutrition,UnitedStatesAgency costsofchronicdiseasesinlow-incomeandmiddle-incomecountries.
forInternationalDevelopment(USAID),WashingtonDC,withadditional Lancet2007,370:1929–1938.
supportfromtheSightandLifeResearchInstitute,Baltimore,MD,USAand 20. PandeyMR,UpadhyayaLR,DhungelS,PillaiKK,RegmiHN,NeupaneRP:
Basel,Switzerland. PrevalenceofhypertensioninaruralcommunityinNepal.IndianHeartJ
1981,33:284–289.
Authordetails 21. JafarTH,LeveyAS,JafaryFH,WhiteF,GulA,RahbarMH,KhanAQ,
1PrograminInternationalandCommunityNutrition,UniversityofCalifornia, HattersleyA,SchmidCH,ChaturvediN:Ethnicsubgroupdifferencesin
Davis,CA,USA.2PrograminHumanNutrition,DepartmentofInternational hypertensioninPakistan.JHypertens2003,21:905–912.
HealthBloomberg,SchoolofPublicHealthJohnsHopkinsUniversity,615 22. SayeedMA,BanuA,HaqJA,KhanamPA,MahtabH,AzadKhanAK:
NorthWolfeStreet,Baltimore21205,Maryland.3NepalNutritionIntervention PrevalenceofhypertensioninBangladesh:effectofsocioeconomicrisk
Project-Sarlahi,Kathmandu,Nepal. factorondifferencebetweenruralandurbancommunity.Bangladesh
MedResCouncBull2002:7–18.
Received:8May2012Accepted:22December2012 23. GuptaR:TrendsinhypertensionepidemiologyinIndia.JHumHypertens
Published:21January2013 2004,18:73–78.
24. MalhotraP,KumariS,KumarR,JainS,SharmaBK:Prevalenceand
determinantsofhypertensioninanun-industrialisedruralpopulationof
References NorthIndia.JHumHypertens1999,13:467–472.
1. LeederSR,RaymondSU,GreenbergH,LuiH,EssonK:Raceagainsttime:The
25. KinraS,BowenLJ,LyngdohT,PrabhakaranD,ReddyKS,RamakrishnanL,
challengeofcardiovasculardiseaseindevelopingeconomies.NewYork:
GuptaR,BharathiAV,VazM,KurpadAV,SmithGD,Ben-ShlomoY,Ebrahim
TrusteesofColumbiaUniversity;2004.
S:Sociodemographicpatterningofnon-communicablediseaserisk
2. MurrayCJ,LopezAD:Theglobalburdenofdisease:Acomprehensive
factorsinruralIndia:acrosssectionalstudy.BMJ2010,341:c4974.
assessmentofmortalityanddisabilityfromdiseases,injuries,andriskfactorsin
26. VasanRS,LarsonMG,LeipEP,KannelWB,LevyD:Assessmentoffrequency
1990andprojectedto2020.Cambridge,MA:HarvardUniversityPress;1996.
ofprogressiontohypertensioninnon-hypertensiveparticipantsinthe
3. GhaffarA,ReddyKS,SinghiM:Burdenofnon-communicablediseasesin
SouthAsia.BMJ2004,328:807–810. FraminghamHeartStudy:acohortstudy.Lancet2001,358:1682–1686.
27. HsiaJ,MargolisKL,EatonCB,WengerNK,AllisonM,WuL,LaCroixAZ,Black
4. GoyalA,YusufS:TheburdenofcardiovasculardiseaseintheIndian
subcontinent.IndianJMedRes2006,124:235–244. HR:PrehypertensionandcardiovasculardiseaseriskintheWomen’s
HealthInitiative.Circulation2007,115:855–860.
5. GazianoTA:Reducingthegrowingburdenofcardiovasculardiseasein
thedevelopingworld.HealthAff(Millwood)2007,26:13–24. 28. GilbertsEC,ArnoldMJ,GrobbeeDE:Hypertensionanddeterminantsof
bloodpressurewithspecialreferencetosocioeconomicstatusinarural
6. GuptaR:BurdenofcoronaryheartdiseaseinIndia.IndianHeartJ2005,
57:632–638. southIndiancommunity.JEpidemiolCommunityHealth1994,48:258–261.
29. TodkarSS,GujarathiVV,TapareVS:Periodprevalenceand
7. JoshiP,IslamS,PaisP,ReddyS,DorairajP,KazmiK,PandeyMR,HaqueS,
sociodemographicfactorsofhypertensioninruralMaharashtra:across-
MendisS,RangarajanS,YusufS:Riskfactorsforearlymyocardial
sectionalstudy.IndianJCommunityMed2009,34:183–187.
infarctioninSouthAsianscomparedwithindividualsinothercountries.
JAMA2007,297:286–294. 30. SinghRB,SharmaJP,RastogiV,NiazMA,SinghNK:Prevalenceand
determinantsofhypertensionintheIndiansocialclassandheartsurvey.
8. KearneyPM,WheltonM,ReynoldsK,MuntnerP,WheltonPK,HeJ:Global
JHumHypertens1997,11:51–56.
burdenofhypertension:analysisofworldwidedata.Lancet2005,
365:217–223. 31. FernaldLC,AdlerNE:Bloodpressureandsocioeconomicstatusinlow-
9. NepalDemographicsProfile2008.https://www.cia.gov/library/publications/ incomewomeninMexico:areversegradient?JEpidemiolCommunity
the-world-factbook/geos/np.html. Health2008,62:e8.
10. WestKPJr,KatzJ,KhatrySK,LeClerqSC,PradhanEK,ShresthaSR,ConnorPB, 32. StuffJE,CaseyPH,SzetoKL,GossettJM,RobbinsJM,SimpsonPM,Connell
DaliSM,ChristianP,PokhrelRP,SommerA:Doubleblind,clusterrandomised C,BogleML:Householdfoodinsecurityisassociatedwithadulthealth
trialoflowdosesupplementationwithvitaminAorbetacaroteneon status.JNutr2004,134:2330–2335.
mortalityrelatedtopregnancyinNepal.TheNNIPS-2StudyGroup. 33. SeligmanHK,LaraiaBA,KushelMB:Foodinsecurityisassociatedwith
BMJ1999,318:570–575. chronicdiseaseamonglow-incomeNHANESparticipants.JNutr2010,
11. StewartCP,ChristianP,KatzJ,SchulzeKJ,WuLSF,LeClerqSC,ShakyaTR, 140:304–310.
KhatrySK,WestKP:MaternalsupplementationwithvitaminAorβ-carotene 34. DasM,PalS,GhoshA:Ruralurbandifferencesofcardiovasculardisease
andcardiovascularriskfactorsamongpre-adolescentchildreninrural riskfactorsinadultAsianIndians.AmJHumBiol2008,20:440–445.
Nepal.JournalofDevelopmentalOriginsofHealthandDisease2010,1:262–270. 35. RamachandranA,SnehalathaC,SatyavaniK,SivasankariS,VijayV:Metabolic
12. ChobanianAV,BakrisGL,BlackHR,CushmanWC,GreenLA,IzzoJLJr,Jones syndromeinurbanAsianIndianadults–apopulationstudyusing
DW,MatersonBJ,OparilS,WrightJTJr,RoccellaEJ:TheSeventhReportof modifiedATPIIIcriteria.DiabetesResClinPract2003,60:199–204.
theJointNationalCommitteeonPrevention,Detection,Evaluation,and 36. GuptaR,GuptaVP,SarnaM,BhatnagarS,ThanviJ,SharmaV,SinghAK,
TreatmentofHighBloodPressure:theJNC7report.JAMA2003, GuptaJB,KaulV:Prevalenceofcoronaryheartdiseaseandriskfactorsin
289:2560–2572. anurbanIndianpopulation:JaipurHeartWatch-2.IndianHeartJ2002,
13. ExecutiveSummaryofTheThirdReportofTheNationalCholesterol 54:59–66.
EducationProgram(NCEP)ExpertPanelonDetection,Evaluation,And 37. SinghRB,RastogiV,NiazMA,GhoshS,SyRG,JanusED:Serumcholesterol
TreatmentofHighBloodCholesterolInAdults(AdultTreatmentPanel andcoronaryarterydiseaseinpopulationswithlowcholesterollevels:
III).JAMA2001,285:2486–2497. theIndianparadox.IntJCardiol1998,65:81–90.
14. InternationalExpertCommitteereportontheroleoftheA1Cassayin 38. LimbuYR,RaiSK,OnoK,KurokawaM,YanagidaJI,RaiG,GurungN,RaiCK:
thediagnosisofdiabetes.DiabetesCare2009,32:1327–1334. LipidprofileofadultNepalesepopulation.NepalMedCollJ2008,10:4–7.
15. Appropriatebody-massindexforAsianpopulationsanditsimplications 39. GhandehariH,Kamal-BahlS,WongND:Prevalenceandextentof
forpolicyandinterventionstrategies.Lancet2004,63:157–163. dyslipidemiaandrecommendedlipidlevelsinUSadultswithand
16. AlbertiKG,ZimmetP,ShawJ:Themetabolicsyndrome–anewworldwide withoutcardiovascularcomorbidities:theNationalHealthandNutrition
definition.Lancet2005,366:1059–1062. ExaminationSurvey2003–2004.AmHeartJ2008,156:112–119.
17. MisraA,ChowbeyP,MakkarBM,VikramNK,WasirJS,ChadhaD,JoshiSR, 40. RadhikaG,VanDamRM,SudhaV,GanesanA,MohanV:Refinedgrain
SadikotS,GuptaR,GulatiS,MunjalYP:Consensusstatementfordiagnosis consumptionandthemetabolicsyndromeinurbanAsianIndians(Chennai
ofobesity,abdominalobesityandthemetabolicsyndromeforAsian UrbanRuralEpidemiologyStudy57).Metabolism2009,58:675–681.
Indiansandrecommendationsforphysicalactivity,medicalandsurgical 41. SrinathReddyK,ShahB,VargheseC,RamadossA:Respondingtothe
management.JAssocPhysiciansIndia2009,57:163–170. threatofchronicdiseasesinIndia.Lancet2005,366:1744–1749.
18. ReddyKS:Cardiovasculardiseaseinnon-Westerncountries.NEnglJMed 42. SasakiH,KawasakiT,OgakiT,KobayashiS,ItohK,YoshimizuY,SharmaS,
2004,350:2438–2440. AcharyaGP:Theprevalenceofdiabetesmellitusandimpairedfasting
Khanetal.BMCPublicHealth2013,13:55 Page10of10
http://www.biomedcentral.com/1471-2458/13/55
glucose/glycaemia(IFG)insuburbanandruralNepal-thecommunities–
basedcross-sectionalstudyduringthedemocraticmovementsin1990.
DiabetesResClinPract2005,67:167–174.
43. SinghDL,BhattaraiMD:Highprevalenceofdiabetesandimpairedfasting
glycaemiainurbanNepal.DiabetMed2003,20:170–171.
44. ChowCK,NaiduS,RajuK,RajuR,JoshiR,SullivanD,CelermajerDS,NealBC:
Significantlipid,adiposityandmetabolicabnormalitiesamongst4535
IndiansfromadevelopingregionofruralAndhraPradesh.Atherosclerosis
2008,196:943–952.
45. VaidyaA,ShakyaS,KrettekA:ObesityprevalenceinNepal:publichealth
challengesinalow-incomenationduringanalarmingworldwidetrend.
IntJEnvironResPublicHealth2010,7:2726–2744.
46. BarretoSM,PassosVM,FirmoJO,GuerraHL,VidigalPG,Lima-CostaMF:
Hypertensionandclusteringofcardiovascularriskfactorsina
communityinSoutheastBrazil–TheBambuiHealthandAgeingStudy.
ArqBrasCardiol2001,77:576–581.
47. NguyenNT,MagnoCP,LaneKT,HinojosaMW,LaneJS:Associationof
hypertension,diabetes,dyslipidemia,andmetabolicsyndromewith
obesity:findingsfromtheNationalHealthandNutritionExamination
Survey,1999to2004.JAmCollSurg2008,207:928–934.
48. GoelR,MisraA,AgarwalSK,VikramN:Correlatesofhypertensionamong
urbanAsianIndianadolescents.ArchDisChild2010,95:992–997.
49. GrundySM:Hypertriglyceridemia,atherogenicdyslipidemia,andthe
metabolicsyndrome.AmJCardiol1998,81:18B–25B.
50. KahnR,BuseJ,FerranniniE,SternM:Themetabolicsyndrome:timefora
criticalappraisal:jointstatementfromtheAmericanDiabetes
AssociationandtheEuropeanAssociationfortheStudyofDiabetes.
DiabetesCare2005,28:2289–2304.
51. GluckmanPD,HansonMA,SpencerHG:Predictiveadaptiveresponsesand
humanevolution.TrendsEcolEvol2005,20:527–533.
52. WheltonPK,HeJ,AppelLJ,CutlerJA,HavasS,KotchenTA,RoccellaEJ,
StoutR,VallbonaC,WinstonMC,KarimbakasJ:Primarypreventionof
hypertension:clinicalandpublichealthadvisoryfromTheNationalHigh
BloodPressureEducationProgram.JAMA2002,288:1882–1888.
doi:10.1186/1471-2458-13-55
Citethisarticleas:Khanetal.:Across-sectionalstudyoftheprevalence
andriskfactorsforhypertensioninruralNepaliwomen.BMCPublic
Health201313:55.
Submit your next manuscript to BioMed Central
and take full advantage of:
• Convenient online submission
• Thorough peer review
• No space constraints or color figure charges
• Immediate publication on acceptance
• Inclusion in PubMed, CAS, Scopus and Google Scholar
• Research which is freely available for redistribution
Submit your manuscript at
www.biomedcentral.com/submit