THE NATIONAL MEDICAL JOURNAL OF INDIA VOL. 15, NO.2, 2002

Original Article63

Hypertension and its risk factors in tea garden workers ofAssam

N. C. HAZARIKA, D. BISWAS, K. NARAIN, H. C. KALITA, J. MAHANTA

ABSTRACTBackground. Hypertension is emerging as a major public

health problem in India. The diversity and heterogeneity ofdistribution of the population makes it difficult to arrive at theprecise prevalence. In Assam, reports from hospitals in teagardens reveal a high prevalence of hypertension among workersin tea gardens. However, no systematic study has been carriedout in this population.

Hethods. We selected, by systematic sampling, 1015individuals (512 men and 503 women) 30 years or more in age,who were interviewed and clinically examined for hypertension.Blood pressure of all the study participants was measured usinga standardized technique. Crude and adjusted odds ratios wereestimated by unconditional simple and multiple logistic regres-sion analyses. Mantel-Haenszel X2 analysis was also used to testfor the association of potential risk factors with hypertension aftercontrolling for co-variables in a stratified analysis. Risk factorsconsidered for hypertension included age, gender, marital status,occupation, alcohol consumption (locally prepared), extra saltintake, smoking history, khaini (a form of tobacco quid contain-ing a mixture oftobacco and lime) intake, body mass index andwaist-hip ratio.

Results. The overall prevalence of hypertension was 60.8%.Increasing age, consumption of locally prepared alcohol, intakeof extra salt in food and beverages and the habit of taking khainiwere found to increase the risk of hypertension. Multivariatelogistic regression models showed that the independent determi-nants of hypertension were age, gender, consumption of locallyprepared alcohol and intake of extra salt. Gender-specific andage-stratified analyses showed the association of increased riskwith intake of khaini in women only, while consumption of locallyprepared alcohol was an important risk factor for hypertensionin both men and women.

Conclusion. The disease burden of hypertension amongworkers in tea gardens is large, despite the community not beingobese. Interventions directed at these workers as well as studies

Regional Medical Research Centre, N.E. Region (lCMR), Post Box 105,Dibrugarh 786001, Assam, India

N.C.HAZARIKA, D.BISWAS, K.NARAIN, J.MAHANTA

Assam Medical College, Dibrugarh 786001, Assam, IndiaH. C. KALITA Department of Cardiology

Correspondence to J. MAHANT A; rmrcdib@rediffmail.com

© The National Medical Journal of India 2002

to determine the reasons for the high prevalence of hypertensionare required.Natl Med J India 2002; 15:63-8INTRODUCTIONDiseases of the circulatory system are responsible for more than 5out of 12 million deaths in developed countries and are rapidlyemerging as a major public health problem in most developingcountries.'> The most important circulatory diseases are highblood pressure (hypertension), coronary heart disease and cere-brovascular disease. Studies in developed countries have reportedthat 60%-80% of the elderly population have high blood pressure(BP).3.4A study on the general population, 25 years and older inMarondera, Zimbabwe, which defined hypertension as systolicBP (SBP) ~140 mmHg and or diastolic BP (DBP) ~90 mrnHgreported a 41% prevalence in women and 26% in men.'

In India, hypertension is a major public health problem. Adiverse and heterogeneous population distribution makes it diffi-cult to arrive at a precise prevalence." A recent study using theWorld Health Organization criteria has shown a 23.7% preva-lence of hypertension in north India in the age group of 25-64years.' A study on women in the age group of 25-64 years,defining hypertension as SBP >140 mmHg or DBP >90 mmHg,reported a prevalence of 30.7% in Thiruvananthapuram (southIndia), 28% in Mumbai (west India), 22.6% in Moradabad (northIndia), 24.2% in Nagpur (central India) and 19.1% in Kolkata(east India)."

The north-eastern states of India are ethno-culturally diverseand each ethnic group has a distinct lifestyle. There is no system-atic community-based study available on the prevalence of hyper-tension from this region. Data available from hospital sourcesshow a high prevalence of hypertension among workers of teagardens of Assam (Dr B. Deka, personal communication). There-fore, we aimed to assess the prevalence of hypertension and its riskfactors in workers of tea gardens of Assam, a major populationgroup linked to an important industry of Assam, India.

SUBJECTS AND METHODSStudy areaThe selected tea garden had an area of 969.63 hectares. Thisgarden was selected at random keeping in view the operationalfeasibility.

64

Study populationThe community of tea garden workers in Assam have theirancestral origins largely in central and south India (MadhyaPradesh, Bihar, Orissa and Andhra Pradesh). Their ancestorscame to Assam as tea garden labourers in the late nineteenth andearly twentieth century. Racially, they are Dravidians and belongto many tribes, e.g. Bhil, Saotal, Kol and Munda.? They are atpresent scattered over more than 800 tea gardens of Assam, butkeep their sociocultural identities intact. The selected tea gardenhad a population of 6138 individuals from 996 families.

The sample size was calculated from the findings of a previouspilot study in a tea garden, where the prevalence of hypertensionrecorded was 30.2% (hypertension defined as SBP >140 mmHg orDBP >90 mmHg). The systematic sampling method was adoptedand every alternate family from the family list was selected. Whenindividuals from selected families were not present during thesurvey period, the families were revisited. In spite of this, therewas a 2% drop-out rate from the selected population. Finally, atotal of 1015 (512 men and 503 women, M:F 1.02)) individualswere interviewed and examined. Both men and women subjects30 years of age and older (age range 30-80 years) were includedin the study. The mean (SD) age of men was 45.1 (11) years andthat of women 42.96 (11.2) years.

Survey instrumentThe subjects were interviewed using a specially designed pre-tested questionnaire. The interviewers were trained before thesurvey started. The questionnaire sought information on thedemographic characteristics, income status and medical historyof the participants. Information on sociodemographic variablessuch as age, marital and co-habitation status, religion, educa-tion, occupation and type of work engaged in was recorded fromthe participants. Dietary habits, including smoking pattern andconsumption of khaini (a form of tobacco quid containing amixture of tobacco and lime), drinking of alcohol and intake ofextra salt (used asa side dish and also in salted tea) were alsorecorded. 10

Anthropometric examinationAccording to the standard protocol," the height and body weightof the study subjects were measured by two trained techniciansthroughout the study period. The participants were barefoot andhad light clothing on when the measurements were taken. Weightand height measurements were recorded using a digital balance(SECA) and a portable stadiometer, respectively. Waist and hipcircumferences were measured to the nearest 0.5 em over lightclothing with the subjects standing erect with their feet together.The waist circumference was measured at the level of the umbili-cus and the hip circumference at the level of the greater trochanterusing a non-stretchable, metallic measuring tape. Body massindex (BMI; weight in kg/height in metre") was used as an indexof obesity. Waist-hip ratio (WHR) was used for measuring truncalobesity. A WHR of 0.88, based on the 75th percentile in the studypopulation, was used as the cut-off point to determine truncalobesity.

Measurement of BPA temporary clinic was set up daily for every 10-15 houses toavoid a long walk for the study subjects. The BP was measured bya mercury column sphygmomanometer using a standardized tech-nique in the sitting posture after the subjects had rested for at leasthalf an hour. Two readings were taken at an interval of 10-15

THE NATIONAL MEDICAL JOURNAL OF INDIA VOL. 15, NO. 2,2002

minutes for each subject who had not consumed alcohol and hadnot eaten for at least one hour before the measurement. The BP wasrecorded in the right arm using a cuff of appropriate size with theinstrument at the level ofthe subject's heart. The cuff pressure wasinflated 30 mmHg above the level at which the radial pulsedisappeared then deflated slowly at the rate of 2 mm/second andthe readings recorded to the nearest 2 mm. II The first and fifthKorotkoff sounds were taken as indicative of the SBP and theDBP, respectively. The average of the two readings of SBP andDBP was recorded. Whenever a high BP was encountered, thereading was rechecked on the next day in the same manner. Theaverage of the rechecked record was accepted for analysis. Thesame team of two doctors (NCH and DB) recorded the BPthroughout the study. For standardization, a pilot study wasundertaken in three different communities of Assam having high,low and medium prevalence of hypertension. The inter- and intra-observer variability was within agreeable limits.

Definition of hypertension for the studyThe hypertension status of the study participants was assessedusing standard criteria formulated by the US Sixth Joint NationalCommittee on detection, evaluation and treatment of hypertension.Hypertension was defined as SBP ~140 mmHg and/or DBP ~90mmHg and/or on treatment with antihypertensive medication. 12

Statistical analysisContingency tables with Chi-square tests were used to detect anassociation between potential risk factors and hypertension. Thecrude odds ratios and 95% confidence intervals (CI) were calcu-lated using univariate analysis. The difference of mean wascalculated by the t-test and the Z-test for equality of proportions oftwo independent samples was used when appropriate. The datawere also analysed by unconditional multiple logistic regressionanalysis in order to find risk factors independently associated withhypertension after adjusting for confounding variables. The maxi-mum likelihood method was used to estimate the parameters ofregression models and p~0.05 (two-tailed) was considered sig-nificant. Multi variate analysis was done by the backward elimina-tion technique (cut -off p=0.1). Chi-square test for linear trend wasused for studying the association between age and hypertension.Mantel- Haenszel X2 test was used in the stratified analysis to studythe association between hypertension and potential risk factorsafter controlling for the effect of co-variables. Analysis of covari-ance (ANCOV A) using the general linear model (GLM) was usedto compare the mean SBP and DBP according to smoking status,consumption of alcohol, chewing of khaini and extra salt con-sumption after adjusting for potential confounding variables inorder to identify factors influencing or modulating SBP and DBPlevels to corroborate our findings. ANCOV A analysis was alsoused to find gender differences in mean SBP and DBP after ageadjustment.

RESULTSDescriptive findingsThe majority of subjects (959; 94.5%) were not educated. Con-sumption of locally prepared alcohol was very common in thestudy subjects. Though subjects from both genders consumed thisalcoholic beverage, the practice was more common in men (91.2%)than in women (78.9%; p

HAZARIKA et al. : HYPERTENSION AND ITS RISK FACTORS IN TEA GARDEN WORKERS OF ASSAM 65

Most of the subjects were underweight (86.1 %), having a BMIof

66 THE NATIONAL MEDICAL JOURNAL OF INDIA VOL. 15, NO. 2,2002

TABLE III. Distribution of blood pressure (BP in mmHg) in the study sample by JNC-VI categorization*

Normotensive High normal Hypertensives(SBP

HAZARIKA et al. : HYPERTENSION AND ITS RISK FACTORS IN TEA GARDEN WORKERS OF ASSAM 67

TABLEV. Risk factors of hypertension in a tea garden community ofAssam. Reduced multivariate models derived by multiple logisticregression analysis with backward elimination of non-significantdeterminants (cut-off p=O.lO)

Determinants Overall model Gender-specific models

In men In women

Adjusted odds ratio Adjusted odds ratio Adjusted odds ratio~%~ ~%~ ~%~

Age (in years)30-39.940-49.9

1(Reference)1.75

(1.28-2.40)*2.88

(1.90-4.37)*8.80

(4.94-15.66)*

1 (Reference)1.41

(0.90-2.20)2.46

(1.37-4.42)*

7.66(3.52-16.69)*

1 (Reference)1.20

(1.26-3.18)*3.43

(1.85-6.38)*9.73

(4.02-23.55)*

50-59.9

260

GenderMenWomen

1 (Reference)1.36 (1.01-1.82)t

Extra salt intakeNo 1(Reference)'A to \12 tsfper day 1.76

(1.30-2.46)t

1(Reference)1.90

(1.19-3.02)*

1(Reference)1.77

(1.13-2.78)*

Consumption of local alcoholNil 1(Reference)200-400 ml 2.33

(1.50-3.60)*>400 ml 2.10

(1.40-3.15)*

Chewing of khainiNoYes

1(Reference)3.87

(1.65-9.06)*3.01

(1.46-6.23)*

1 (Reference)2.0 (1.19-3.36)*

Waist-hip ratio0.88

1 (Reference)1.68 (0.87-3.23)

2.83 (1.47-5.46)*2.24 (1.22-4.12)*

tp

68

effect in measuring BP. Necessary precautions were also taken toavoid any fallacy due to the 'use of improper cuff size' and'dangling arm' .35

There are some limitations of our study. The study is a cross-sectional one and BP measurement was obtained during one studyperiod to determine the hypertensive status of the subjects. Thisshortcoming may affect the prevalence of hypertension. Preva-lence may be also overestimated because of the phenomenon ofpseudohypertension."

Data for risk factor analysis of dietary salt consumption wasestimated by the recall method. Though this is an accepted methodand the intake is a regular phenomenon, yet sodium/potassiumestimation in both urine and blood samples would have added tothe strength of recall.

In conclusion, there is a high prevalence of hypertensionamong tea garden workers which requires intervention. Highlevels of BP, more so in those ~60 years of age, is a major causefor hypertension-related complications. The underlying reasons,apart from the risk factors discussed here, need to be explored todesign a rational intervention strategy. Awareness, detection andcontrol of hypertension are poorin this population probably due tolow literacy levels and socio-economic status. To develop aneffective intervention strategy, the treatment -seeking behaviour ofindividuals would also need to be improved through awarenessprogrammes.

ACKNOWLEDGEMENTSWe are grateful to the management, Dr P. Khaund, members of the hospital staffand members of the 'Mother's Club' of the tea garden for their active coopera-tion. We also appreciate the assistance received from Mr D. Hazarika, Mr P.Doioi and Mr S. K. Goswami of the Centre.

REFERENCESHealth across the lifespan. The World Heallh Report. Geneva:World HealthOrganization, 1998.

2 Reddy KS. Hypertension control in developing countries: Generic issues. ] HumHypertens 1996;10 (suppl 1):S33-S38.

3 Burt VL, Whelton P, RoccellaEJ, Brown C, Cutler JA, Higgins M, etal. Prevalenceof hypertension in the US adult population. Results from the Third National Health andNutrition Examination Survey, 1988-1991. Hypertension 1995 ;5:305-13.

4 Kuramoto K. Treatment of elderly hypertension in Japan: National InterventionCooperative Study in Elderly Hypertensives. J Hypertens 1994;12 (suppl): S35-S40.

5 MufundaJ, ScottLJ, ChifambaJ, MatengaJ, Sparks B, Cooper R, etal. Correlates ofblood pressure in an urban Zimbabwean population and comparison to other populationsof African origin.J Hum Hypertens 2000;14:65-73.

6 Wasir HS. Primary prevention of hypertension in IndiaJ' Hum Hypertens 1996;10(suppll):S55-S56.

7 Singh RB, Beegom R, Ghosh S, Niaz MA, Rastogi V, Rastogi SS, et al. Epidemiologicalstudy of hypertension and its determinants in an urban population of north India.J HumHypertens 1997;11:679-85.Singh RB, Beegom R, Mehta AS, Niaz MA, De AK, Haque M, et al. Prevalence andrisk factors of hypertension and age-specific blood pressures in five cities: A study ofIndian women. N.K.P Salve Institute of Medical Sciences, Nagpur, India-Five citystudy group. Int] CardioI1998;63: 165-73.

9 Griffiths P. The history of the Indian tea industry. London, 1967.

THE NATIONAL MEDICAL JOURNAL OF INDIA VOL. 15, NO.2, 2002

10 Rose EA, Blackburn H, Gillum RF, Prineas RJ. Cardiovascular survey methods.Geneva: World Health Organization, 1982:68-70.

II Hypertension control. Reportofa WHO expert committee. WorldHeallh Organ TechRep Ser 1996;862:1-5.

12 Joint National Committee on Detection, Evaluation and Treatment of high bloodpressure. The sixth report of the Joint National Committee on prevention, detection,evaluation and treatment of high blood pressure. JNC- VI. Arch lntern Med1997;157:2413-46.

13 Olatunbosun ST, Kaufman JS, Cooper RS, Bella AF. Hypertension in a blackpopulation: Prevalence and biosocial determinants of high blood pressure in a group ofurban Nigerians.i/ Hum Hypertens 2000;14:249-57.

14 Wilson PW. An epidemiologic perspective of systemic hypertension, ischaemic heartdisease and heart failure. Am] Cardia I 1997;80 (9B):3J-8J.

15 GuptaR, al-OdatNA, Gupta VP. Hypertension epidemiology in India: Meta-analysisof 50-year prevalence rates and blood pressure trends. ] Hum Hypertens 1996;10:465-72.

16 PrimatestaP, BostL, Poulter NR. Blood pressure levels and hypertension status amongethnic groups in England.Z Hum Hypertens 2000;14: 143-8.

17 Rodriguez-Saldana J, Chavez-Dominguez R, Marmolejo-Henderson R, Reynosco-Marenco MT. [Di versity in studies ofthe prevalence of systemic hypertension.] ArchInst Cardiol Max 1998;68:130-9.

18 Kalavathy MC, Thankappan KR, Sarma PS, Vasan RS. Prevalence, awareness,treatment and control of hypertension in an elder! y community-based sample in Kerala,India. Natl Med J India 2000;13:9-15.

19 Tanji JL, Arevalo JA, Paliescheskey M, Lee L, Alcalde O. Prevalence rate ofhypertension among recent Southeast Asian refugees to northern California. JAmBoard Fam Pract 1994;7:105-9.

20 Woo J, Ho SC, Yuen YK, Yu LM, Lau J. Cardiovascular risk factors and IS-monthmortality and morbidity in an elderly Chinese population aged 70 years and over.Gerontology 1998;44:51-5.

2!1. Yunis C, Krob HA. Status of health and prevalence of hypertension in Brazil. EthnDis1998;8:406-12.

22 Carbajal HA. Optimal blood pressure and high normal blood pressure in La Plata,Argentina. Can] CardioI1994;10:749-52.

23 Dwivedi S, Singh G, Agarwal MP. Profile of hypertension in elderly subjects.J AssocPhysicians India 2000;48:1047-9.

24 Agarwal AK, Yunus M, Khan A, Ahmed 1. A clinico-epidemiological study ofhypertension in rural population of Jawan Block, Distt Aligarh (UP), India.U R SacHealth 1994;114:17-19.

25 Sonmez HM, Basak 0, Camci C, Baltaci R, Karazeybek HS, Yazgan F, et al. Theepidemiology of elevated blood pressure as an estimate for hypertension in Aydin,Turkcy.i/ Hum Hypertens 1999;13:399-404.

26 Joshi SV, Patel JC, Dhar HL. Prevalence of hypertension in Mumbai. Indian] MedSci 2000;54:380-3.

27 Weinberger MH. Salt sensitivity. Does it play an important role in the pathogenesis andtreatment of hypertension? Curr Opin Nephrol Hypertens 1996;5:205-8.

28 KolevaM. Nutrition, nutritional habits, obesity and prevalence of chronic diseases inworkers. Rev Environ Health 1999;14:21-9.

29 Beilin LJ, Puddey lB. Alcohol, hypertension andcardiovasculeardisease-implicationsfor management. Clin Exp Hypertens 1993;15: 1157-70.

30 Britton A, Mckee M. The relation between alcohol and cardiovascular disease inEastern Europe: Explaining the paradox. ] Epidemiol Community Health 2000;54:328-32.

31 JoshiPP, Kate SK, Shegokar V. Blood pressure trends and life styleriskfactors in ruralIndia.U Assoc Physicians India 1993;41:579-81.

32 Seedat YK, Seedat MA, Hackland DB. Biosocial factors and hypertension in urban andrural Zulus. S Afr Med J 1982;61 :999-1002.

33 Scrimshaw NS. The relation between fetal malnutrition and chronic diseases in laterlife. BM] 1997;315:825-6.

34 The Silent Emergency. United Nations Children Fund. The state of the world'schildren, Published for UNICEF by Oxford University Press, 1998.

35 Hall WD. Pitfalls in the diagnosis and management of systolic hypertension. SouthMed] 2000;93:256-9.

36 Messerli FH, Ventura HO, AmodeoC. Osler's maneuver and pseudohypertension. NEngl] Med 1985;312:1548-51.