Cardiovascular risk assessment among rural population: findings from a cohort study in a peripheral region of Bangladesh

p u b l i c h e a l t h x x x ( 2 0 1 6 ) 1 e8

Available online at www.sciencedirect.com

Public Health journal homepage: www.elsevier.com/puhe

Original Research

Cardiovascular risk assessment among rural population: findings from a cohort study in a peripheral region of Bangladesh* K. Fatema a,b,*,f, N.A. Zwar b, A.H. Milton c, B. Rahman b, A.S.M.N. Awal d, L. Ali e a

Department of Epidemiology, Bangladesh University of Health Sciences (BUHS), 125/1 Darus Salam, Mirpur, Dhaka 1216, Bangladesh b The School of Public Health and Community Medicine, Faculty of Medicine, The University of New South Wales, Sydney, NSW 2052, Australia c Centre for Clinical Epidemiology and Biostatistics (CCEB), The School of Medicine and Public Health, Faculty of Health and Medicine, The University of Newcastle, Newcastle, NSW 2008, Australia d ORBIS International, Apartment No, B4 & C4, Plot #1, Road #137, Gulshan, Dhaka 1212, Bangladesh e Department of Biochemistry and Cell Biology, BUHS, 125/1 Darus Salam, Mirpur, Dhaka 1216, Bangladesh

article info

abstract

Article history:

Objectives: The incidences of non-communicable diseases including cardiovascular dis-

Received 19 March 2015

eases (CVDs) is increasing in Bangladesh. The reasons for this increasing trend need to be

Received in revised form

explored. The aim of this study is to assess the risk of CVDs among a peripheral rural

20 January 2016

Bangladeshi population and to explore the sociodemographic, anthropometric and clinical

Accepted 15 February 2016

variables associated with increased risk.

Available online xxx

Study design: Cohort study. Methods: From a cohort of 190,471 individuals of all ages, originally included in a diabetic

Keywords:

eye disease program initiated in 2008e2009, a purposive sub-cohort of 66,710 individuals,

Cardiovascular diseases

aged 31e74 years was recruited. During 2011e2012 these participants were assessed for

Diabetes

CVDs using the WHO's risk assessment tool designed for primary care settings in low

Hypertension

resource societies. Participant characteristics associated with higher risk were explored

Anthropometric indices

using univariable and multivariable regression analysis.

Bangladesh

Results: Out of all (95.5% participation rate) participants 1170 (1.84%) were found to be at high risk for CVD. The prevalence of hypertension (HTN), pre-HTN, obesity, underweight and self-reported DM were 8.9%, 15.2%, 9.6%, 7.8% and 0.5% respectively, among the study population. In multivariable regression analysis female sex, older age, temporary housing structure (i.e., tin shed), extremes of BMI (both underweight and obese) and central obesity were associated with higher risk for CVDs.

*

Project Place: Pirgonj of Thakurgaon district e one of the north-western districts in Bangladesh. * Corresponding author. The School of Public Health & Community Medicine, Faculty of Medicine, University of New South Wales (UNSW), Sydney, NSW 2052, Australia. Tel.: þ61 470505515. E-mail addresses: [email protected], [email protected], [email protected] (K. Fatema), N.Zwar@unsw. edu.au (N.A. Zwar), [email protected] (A.H. Milton), [email protected] (B. Rahman), [email protected] (A.S.M.N. Awal), [email protected] (L. Ali). f Permanent address. Department of Epidemiology, Bangladesh University of Health Sciences (BUHS), 125/1, Darus Salam, Mirpur, Dhaka 1216 Bangladesh. Tel.: þ880 2 8055352, þ880 1716502930 (mobile); fax: þ880 2 8611138. http://dx.doi.org/10.1016/j.puhe.2016.02.016 0033-3506/© 2016 The Royal Society for Public Health. Published by Elsevier Ltd. All rights reserved. Please cite this article in press as: Fatema K, et al., Cardiovascular risk assessment among rural population: findings from a cohort study in a peripheral region of Bangladesh, Public Health (2016), http://dx.doi.org/10.1016/j.puhe.2016.02.016

2

p u b l i c h e a l t h x x x ( 2 0 1 6 ) 1 e8

Conclusions: The prevalence of CVD risk factors and high CVD risk individuals in this cohort was found to be lower than previous studies. It may be the effects of urbanization are yet to reach this relatively traditional rural population. This study adds to the literature on use of the WHO risk assessment tool. © 2016 The Royal Society for Public Health. Published by Elsevier Ltd. All rights reserved.

Introduction Non-communicable diseases (NCDs) are now major public health problems in both developed as well as low- and middleincome countries (LMICs).1 While death rates from cardiovascular diseases (CVDs) have been declining in developed countries the prevalence of CVDs has been rising greatly in LMICs.2 At the beginning of the present millennium 80% of the burden of CVDs was estimated to occur in these countries.3 South Asians exhibit greater susceptibility to these diseases4 and compared to the western population CVD-related death among South Asian populations has been found to occur 5e10 years earlier.5 This increased susceptibility to CVD among South Asian populations is not fully explained by traditional risk factors.2 Although some useful information has been generated on the present and future burden of CVDs in Bangladesh, organized epidemiological studies exploring the prevalence of CVDs as well as their risk factors among various groups and subgroups of the population are relatively scarce. Evidence suggests that CVD risk prediction based on the combined effect of multiple risk factors is more accurate than that based on individual risk factors.6 There are few studies on Bangladeshi population using multiple risk assessment approach for coronary heart disease (CHD) or CVDs as a whole. Moreover, the majority of studies to date were conducted either on Bangladeshi migrants (along with some other Asian ones) or on subjects recruited from tertiary hospital settings.7e10 Only two studies, so far, have explored multiple CVD risk factors among the rural Bangladeshi population,11,12 but the subjects were recruited from areas within 50 km of the capital (Dhaka) and thus, may not reflect the prevalence in peripherally located rural population. Given the scarcity of population data the present study was undertaken to assess CVD risk factors among members of a large population-based cohort in a more traditional rural setting in a peripheral region of the country (at Pirganj Upazilla of Thakurgaon district, about 467 km north-west of Dhaka) using the WHO Risk Management Package tool for CVDs.13 In this paper, we present the assessment of CVDs risk (using baseline data) in this population. The association between CVD risk and sociodemographic, anthropometric and clinical characteristics has also been explored.

started in 2008 under the ‘BADAS-ORBIS Eye Care Project’. The original cohort was initiated to generate epidemiological data on the burden of diabetic retinopathy and associated risk factors from a rural population and study area was selected as per that project design. The study was conducted in an upazilla (sub-district) named ‘Pirgonj’ of Thakurgaon district - one of the north-western districts in Bangladesh. ‘Pirgonj’ has a total of 10 unions, 13 wards and 165 villages. It is situated at a distance of ~467 km from Dhaka, the capital of Bangladesh. In 2011 we assessed a subset of this cohort for the presence of pre-existing CVDs symptoms (using the multiple risk assessment questionnaire of the WHO Risk Management Package) under the North Bengal Non-Communicable Disease Programme (NB-NCDP) of Bangladesh University of Health Sciences (BUHS). Eligible participants were those meeting the aged criteria (as recorded in the baseline data collection during 2008e09), usual residence of the study area, willing to participate and being able to communicate in a face to face interview and also do not have any physical disability that hinders the anthropometric measurements (Fig. 1). To collect this data, we recruited 20 interviewers and a field supervisor from the local community. The interviewers had completed at least 12 years of education and had experience in conducting interviews, surveys and using the census method. All project staff received two weeks intensive training on data collection techniques by the investigators of this study. The principal investigator (KF) supervised the fieldwork. A randomly selected sample of two percent of the completed questionnaires was cross checked by the investigators.

Description of the questionnaire to assess CVD risk The WHO questionnaire that was used to determine probable angina, heart attack, stroke and transient ischemic attack (TIA) consists of eight questions. These questionnaires were developed as a part of the ‘WHO CVD-RISK Management Package for low- and medium e resource settings’13 to identify pre-existing disease or likely disease participants on the basis of symptoms of a history of pre-existing CVD and it has been used in several studies.14,15 In this study we used this tool for a group of cohort participants based on age alone and did not select subgroups with known risk factors. Those who had probable angina, heart attack, stroke or TIA symptoms were considered at high risk participants for CVD.16

Methods Anthropometric and blood pressure measurements Study population and data collection procedures This study was conducted as the baseline survey of a prospective cohort among the rural Bangladeshi population,

Anthropometric measurements, such as height (ht) [using a portable, locally manufactured, stadiometer, standing upright on a flat surface machine], weight (wt) [modern electronic

Please cite this article in press as: Fatema K, et al., Cardiovascular risk assessment among rural population: findings from a cohort study in a peripheral region of Bangladesh, Public Health (2016), http://dx.doi.org/10.1016/j.puhe.2016.02.016

3

p u b l i c h e a l t h x x x ( 2 0 1 6 ) 1 e8

Location: "Pirgonj" upazilla,

BADAS-ORBIS Eye Care Project

Period:

September 2008—March 2009

Type:

Population census

Sample:

Included all ages (n=190471)

Baseline

data:

Information

on

sociodemographic,

anthropometry, blood pressure (BP), known diabetes mellitus (DM), active smoking habit and eye problem

Location: "Pirgonj" upazilla, Period:

September 2011—July 2012

Type:

Population census

North Bengal non

Sample: All BADAS-ORBIS participants aged between

communicable

31—74 years in 2008 (n=66701) were asked to

disease program

participate. Of these n=63708 (participation rate 95.5%)

(NB-NCDP)

agreed and included in this study. Objective: Screening for CVDs. WHO Screening questionnaire for probable angina, heart attack, stroke, transient ischemic attack (TIA) were used to identify current or past symptoms and considered as high risk. High risk n=1170 (participation rate 100%) and total not high risk n= 62538.

Fig. 1 e Flow chart for the study design and selection of participants.

digital LCD weighing machines] and waist circumferences (WC) and hip circumferences (HC), were taken from the participants wearing light clothes and without shoes and following standard procedures. To minimize error, all instruments were calibrated regularly before data collection. Body mass index (BMI) was calculated using the formula [wt (kg)/ht (m2)]. Waist hip ratio (WHR) and waist height ratio (WHtR) were calculated by dividing WC (cm) by HC (cm) and ht (cm) respectively. Following auscultatory method BP was measured on the day of interview. Prior to the measurement, 10 min rest was assured and standard cuffs for adults fitted with mercury sphygmomanometer were used to minimize variation in the measurements. However, in the first cohort, only a single reading was recorded for wt, ht, WC and BP.

Diagnosis criteria for variables Asian BMI criteria was used to categorise and define underweight (less than 18$5 kg/m2) normal (18.5e23.0 kg/m2), overweight (23e27.5 kg/m2) and obesity (higher than 27.5 kg/ m2) for both sexes;17 cut-off values for central obesity including WC for men and women were 90 and 80 cm, respectively, WHR for men and women were <0.95 and <0.80, respectively,18 WHtR for both sexes were 0.50.19 Known DM was defined by the use of insulin or oral anti-diabetic medication(s) and self-reported information. Individuals were considered to have hypertension (HTN) if their average SBP was 140 mmHg or DBP was 90 mmHg, or if they were receiving treatment for HTN; pre-HTN if SBP was 120e139 mmHg or DBP was 80e89 mmHg.20 Income was

Please cite this article in press as: Fatema K, et al., Cardiovascular risk assessment among rural population: findings from a cohort study in a peripheral region of Bangladesh, Public Health (2016), http://dx.doi.org/10.1016/j.puhe.2016.02.016

4

p u b l i c h e a l t h x x x ( 2 0 1 6 ) 1 e8

classified according to the 2006 per capita Gross National Income (GNI) and according to World Bank (WB) calculations.21 The groups were: low-income, US$ 905 or Bangladeshi Taka BDT 5360; lower-middle-income, US$ (906e3595) or BDT (5361e21270).

Ethical consideration The present study was carried out according to the guidelines laid down in the Declaration of Helsinki on medical ethics. Study protocol and all procedures involving human participants were approved by the Human Research Ethical Committee (HREC) of the University of New South Wales (HREC ref: sHC12621), Sydney, Australia and the Ethics Review Committee of the Diabetic Association of Bangladesh for Medical Research. In Bangladesh the present illiteracy rate is 40.18%;22 so, given the low literacy level of the population community based research in Bangladesh commonly uses verbal informed consent in presence of a witness.11 All the participants were verbally informed regarding their right to withdrawal from the study at any stage or to restrict their data from further analysis or disclose for publication. A printed copy of the participants' rights regarding participation in the study was also provided after obtaining the verbal information.

Statistical analysis The prevalence of high CVD risk and its associated risk factors were determined as percentages. The sociodemographic, anthropometric and clinical characteristics were summarised by means and standard deviations (SD) for continuous variable and by proportions for categorical variables. Independent-sample t-tests (for continuous variables) and Pearson's chi-square test (for categorical variable) were carried out to examine the difference in the covariates between high risk and not high risk populations. A backward elimination approach of model building was used to decide the final model including all predictors of high risk group that were significant at 5% level. Under this approach, variables significant at 15% level in the univariable logistic regression model were included in a multivariable base model. From the base model, variables were excluded one by one based on their Pvalues that were greater than 0.05 to reach the final model. After forming the final model we checked for multicollinearity by estimating a variance inflation factor (VIF) by fitting a multiple linear regression model with the binary outcome variable.23 Data were analysed using Stata version 12 (StataCorp LP, College Station, TX).

Results Overall characteristics of the study participants

Sociodemographic characteristic of all the participants are shown in Table 1. High risk individuals were different from rest of the study participants with respect to gender, age group, education and gross national income (P < 0.001). Among the high risk group a smaller proportion was aged under 40 years or below (P < 0.001) compared to the other group.

Prevalence of risk factors and its correlates The overall prevalence of anthropometric, clinical and other risk factors in high risk and not high risk participants is reported in Table 2. The prevalence of HTN was 4.4 and 9.0%; pre-HTN was 15.4% and 15.2% among high risk and not high risk group, respectively. The prevalence of known DM among study participants was very low. Distribution of the WHtR showed that 28.5% in the high risk group and 30.5% in low risk

Table 1 e Sociodemographic characteristics of the study participants. Variablesa

Gender Male 31,786 (50.8) Female 30,752 (49.2) Age (M ± SD; years) 46 ± 11 31e40 26,231 (41.9) 41e50 18,090 (28.9) 51e60 12,187 (19.5) 61e70 4503 (7.2) 71eup to 74 1527 (2.4) Education Illiterate/Signature/ 26,793 (42.8) Gonoshikha Primary level 23,139 (37.0) Secondary level 11,022 (17.6) and above Graduation/ 1584 (2.5) postgraduation Employment status Unemployed/sacked 1076 (1.7) from the present job/Retired Office work/Business/ 6563 (10.5) Skilled labour House maker/farmer 26,689 (42.7) Rickshaw puller/ 28,210 (45.1) day labour/Others Gross national income (monthly, US$) Low income (905) 52,674 (84.2) Lower-middle 9864 (15.8) income (906e3595) Type of house Brick 5220 (8.3) Semi-brick 8967 (14.3) Tin-shed 18,210 (29.1) Bamboo and others 30,141 (48.2) a

According to the WHO's CVD risk assessment tool, only 1170 (1.84%) of the participants were at high risk for a further CVD event on the basis of pre-existing CVDs (angina, heart attack, stroke and TIA).

Not high risk (n ¼ 62,538)

b

High risk (n ¼ 1170)

P-valueb

463 (39.6) 707 (60.4) 51 ± 11 261 (22.3) 378 (32.3) 313 (26.8) 188 (16.1) 30 (2.6)

<0.001 <0.001

<0.001

583 (49.8) 404 (34.5) 169 (14.4)

<0.001

14 (1.2)

13 (1.1)

95 (8.1)

0.186

608 (52.0) 454 (38.8)

1035 (88.5) 135 (11.5)

68 (5.8) 121 (10.3) 470 (40.2) 511 (43.7)

<0.001

0.032

Values expressed as numbers and percentages in parentheses or M ± SD, as appropriate; M, mean; SD, standard deviation; yrs, years; US$1 ¼ 80 Bangladeshi Taka (BDT). P-value, significance between not high risk and high risk participants.

Please cite this article in press as: Fatema K, et al., Cardiovascular risk assessment among rural population: findings from a cohort study in a peripheral region of Bangladesh, Public Health (2016), http://dx.doi.org/10.1016/j.puhe.2016.02.016

p u b l i c h e a l t h x x x ( 2 0 1 6 ) 1 e8

Table 2 e Overall baseline risk factors among not high risk and high risk study participants. Variablesa Systolic blood pressure (mmHg) Diastolic blood pressure (mmHg) Hypertension Normotensive Pre-hypertensive Hypertensive Known diabetes Yes No Smoking pattern Non-smoker Smoker Body mass index Normal (18.51e23.0) Underweight (<18.5) Overweight (23.01e27.5) Obese (>27.51) Waist circumference Normal (<0.90 male, <0.80 female) High risk (>0.90 male, >0.80 female) Waist hip ratio Normal (<0.95 male, <0.80 female) Moderate (0.96e1.0 male, 0.81e0.85 female) High risk (>1.0 male, >0.85 female) Waist height ratio 0.5 (non-central fat distribution e pears) >0.5 (central fat distribution e apples) a b

Not high risk High risk P-valueb (n ¼ 62,538) (n ¼ 1170) 117.6 ± 10.2

116.9 ± 10.4

0.022

77.7 ± 8.4

77.1 ± 8.3

0.010

47,391 (75.8) 9517 (15.2) 5630 (9.0)

939 (80.3) 180 (15.4) 51 (4.4)

<0.001

301 (0.5) 62,237 (99.5)

8 (0.7) 1116 (99.3)

0.482

56,380 (90.2) 6158 (9.8) 21.3 ± 2.3 41,541 (66.4) 4917 (7.9) 10,053 (16.1) 6027 (9.6) 74.5 ± 6.5 53,807 (86.0)

1040 (88.9) 130 (11.1) 21.6 ± 2.5 821 (70.2) 75 (6.4) 202 (17.3) 72 (6.2) 74.3 ± 7.1 1031 (88.1)

0.151

8731 (14.0)

139 (11.9)

0.87 ± 0.04 30,452 (48.9)

0.86 ± 0.04 472 (40.6)

9419 (15.1)

243 (20.9)

22,459 (36.0)

448 (38.5)

0.48 ± 0.04 43,445 (69.5)

0.48 ± 0.04 836 (71.5)

19,093 (30.5)

334 (28.5)

5

Correlate of high risk for CVDs is reported in Table 3. On univariable analyses, gender, age, education, employment status, gross national lower-middle income (monthly), type of house, HTN, BMI, WC and WHR were found to be significantly associated with higher risk of CVD. On multivariable analysis all these factors, except education and WHR, also showed independent significant association with high risk for CVD.

Discussion

0.004 <0.001

0.346 0.042

0.056

<0.001

0.219 0.150

Values expressed as numbers and percentages in parentheses. P-value, significance between not high risk and high risk participants.

participants had the apple shaped central fat distribution tendency which has been found to increase the risk of CVDs in western populations.24 Prevalence of the combination of one or more risk factors (hypertension, diabetes, smoking/tobacco use and obesity) in both high and not high risk groups was also observed. Only 28.2% of the total participants were free from any risk factors. Most (34.2%) of the participants in the present study had at least one of these risk factors, about 37.6% had at least 2 coexisting risk factors, and 9.9% had all types of risk factors. Moreover females showed significantly higher prevalence of risk factors than males (40% vs 29% has one and 57% vs 19% has more than two risk factors, in women and men respectively, P < 0.001) (data not shown in a table).

The present study shows a small number of high risk subjects for CVDs (1.84%) on the basis of a positive finding on the WHO tool in this rural population which is located in a peripheral area of Bangladesh. Older age, female sex, temporary housing structure, underweight and obesity define by BMI and central obesity were associated with being a higher risk for CVD in this population. A few recent studies have also used the WHO tool. In a study among Chinese and Nigerian population, only 2% of the study participants were labelled with CVD high risk.25 When analysed with the individual risk factor approach, 34.2% of the participants in the present study were found to have at least one and 37.6% had more than two CVD risk factors with females showing higher prevalence than men (40% has one and 57% has more than two risk factors, P < 0.001). These overall proportions are consistent with the data from another study in Bangladesh.26 However, the prevalence of hypertension, self-reported DM and smoking are lower in our study than other studies conducted among rural Bangladeshi population.27e29 It is well established that self-reporting of health events is more influenced by overstatement or underestimate, stemming from recall, telescoping and social acquiescence biases.30e32 In this study population we consider underestimation of DM is likely to have occurred due to lack of health awareness, inadequate access to health facilities and other social factors such as religious barrier.33 We only collected data on participants' current smoking status and we did not collect any information on smokeless tobacco consumption. Age, female sex, type of temporary housing structure, and both obesity by BMI and central obesity are important correlates of CVD risk positivity in the study population. With the notable exception of female sex, similar determinants of CVD risk have been reported earlier in other studies.34,35 Education status, occupation, type of temporary housing structure and social status were associated with cardiometabolic risk factors in Indian population.36,37 However, the association of female sex with CVD risk positivity was an unexpected finding that needs further investigation. Although it is very clear that Bangladesh has double burden of nutritional disorders specially in the rural population38 the association between being underweight and CVDs is still unclear. In this study, underweight defined by BMI showed significant inverse association with CVD risk. However, obesity defined by both BMI and central obesity showed direct associated with high risk of CVD in multivariable model. A recent study in Denmark reported that underweight is one of the risk factors for CVDs which, in association with increased age, can aggravate this risk even further.39 Based on

Please cite this article in press as: Fatema K, et al., Cardiovascular risk assessment among rural population: findings from a cohort study in a peripheral region of Bangladesh, Public Health (2016), http://dx.doi.org/10.1016/j.puhe.2016.02.016

6

p u b l i c h e a l t h x x x ( 2 0 1 6 ) 1 e8

Table 3 e Univariable and multivariable analyses of factor associated with CVD risk (n ¼ 63,708). Variables

Univariable a

ORb (95% CI)

Sociodemographic Gender Age (yrs)

Education

Employment status

Gross national income (monthly, US$) Type of house

Risk factors: Hypertension

Body mass indexd

Waist circumferenced Waist hip ratiod

a b c d

Male Female 31e40 41e50 51e60 61e70 71eup to 74 Illiterate/Signature/Gonoshikha Primary level Secondary level and above Graduation/postgraduation Office work/Business/Skilled labour Unemployed/sacked from the present job/Retired House maker/farmer Rickshaw puller/day labour/Others Low income (905) Lower-middle income (906e3595) Brick Semi-brick Tin-shed Bamboo and others

1.00 1.58 (1.41e1.78) 1.00 2.10 (1.79e2.46) 2.58 (2.19e3.05) 4.19 (3.47e5.07) 1.97 (1.35e2.89) 1.00 0.80 (0.70e0.91) 0.70 (0.59e0.83) 0.41 (0.24e0.69) 1.00 0.83 (0.46e1.49)

Normotensive Pre-hypertensive Hypertensive Normal (18.51e23.0) Underweight (<18.5) Overweight (23.01e27.5) Obese (>27.51) Normal (<0.90 male, <0.80 female) High risk (>0.90 male, >0.80 female) Normal (<0.95 male, <0.80 female) Moderate (0.96e1.0 male, 0.81e0.85 female) High risk (>1.0 male, >0.85 female)

1.00 0.95 (0.81e1.12) 0.46 (0.34e0.61) 1.00 0.77 (0.61e0.98) 1.02 (0.87e1.19) 0.60 (0.47e0.77) 1.00 0.83 (0.70e0.99) 1.00 1.66 (1.42e1.95) 1.29 (1.13e1.47)

1.57 (1.26e1.96) 1.11 (0.89e1.39) 1.00 0.70 (0.58e0.83) 1.00 1.04 (0.77e1.40) 1.98 (1.53e2.56) 1.30 (1.01e1.68)

Multivariable P-value

<0.001 <0.001 <0.001 <0.001 <0.001 0.001 <0.001 0.001 0.544 <0.001 0.351 <0.001 0.817 <0.001 0.043

0.571 <0.001 0.033 0.835 <0.001 0.042 <0.001 <0.001

ORc (95% CI) 1.00 2.0 (1.59e2.50) 1.00 2.07 (1.76e2.44) 2.69 (2.27e3.20) 5.19 (4.24e6.34) 3.25 (2.17e4.86) 1.00 1.03 (0.90e1.18) 1.12 (0.92e1.35) 0.83 (0.47e1.48) 1.00 0.32 (0.17e0.60) 0.77 (0.57e1.05) 0.81 (0.64e1.03) 1.00 0.68 (0.56e0.82) 1.00 1.0 (0.74e1.35) 1.88 (1.45e2.44) 1.19 (0.92e1.55) 1.00 0.91 (0.77e1.07) 0.30 (0.19e0.49) 1.00 0.60 (0.47e0.77) 1.10 (0.94e1.29) 1.82 (1.20e2.74) 1.00 1.32 (1.06e1.66) 1.00 1.15 (0.88e1.49) 1.03 (0.79e1.34)

P-value

<0.001 <0.001 <0.001 <0.001 <0.001 0.679 0.251 0.540 <0.001 0.105 0.094 <0.001 0.982 <0.001 0.175

0.261 <0.001 <0.001 0.230 0.004 0.015 0.297 0.849

OR, Odds ratio; CI, Confidence interval. Crude odds ratio after univariable logistic regression. Adjusted odds ratio after multivariable logistic regression. Different obesity indicators (i.e., BMI, WC and WHR were added in the model separately).

this information, it would appear that the progression of atherosclerosis with age occurs irrespective of general nutritional status for individual with higher CVD risk. Another study showed that the prevalence of metabolic syndrome and central obesity increased with age, peaking at 50e59 years and decreasing in the older age group.40 These findings further compound the prevailing controversy surrounding the relationship between ages, underweight, central obesity and CVDs. Moreover, it is observed that in univariable analysis obesity (both BMI and high waist circumference) showed inverse association with high CVD risk. But in the multivariable analysis all these variables showed strong positive association which indicate Simpson's paradox, that is, reverses the direction of association after adjusting for other predictors.41 The direct association observed in the multivariable analysis is consistent with existing literature. These findings suggest that a single risk factor model for CVD can be misleading and multivariable models should be examine CVD risk factors.

While the WHO Tool used in the present study is primarily targeted for a risk factor assessment for CVDs, the clinical nature of the questionnaire permits some assumption on the prevalence of cardiovascular disease themselves. The finding of low proportion of individual components of the CVD risk in the present population is somewhat surprising given the fact that other studies have found high prevalence of CVDs in various groups of Bangladeshi population.29,36,42 The differences may not be unusual if the methodological differences as well as the socio-economic and lifestyle heterogeneity of the various Bangladeshi groups are taken into consideration. Due to regional inequity and for age-old deprivation mainly due to extreme difficulty in road communication in the NorthWestern region in Bangladesh (the present study area), the participants are from a rural background with a relatively preserved traditional lifestyle. The rural population included in this study is more racially homogeneous than many urban samples and provides insights into the susceptibility of South

Please cite this article in press as: Fatema K, et al., Cardiovascular risk assessment among rural population: findings from a cohort study in a peripheral region of Bangladesh, Public Health (2016), http://dx.doi.org/10.1016/j.puhe.2016.02.016

p u b l i c h e a l t h x x x ( 2 0 1 6 ) 1 e8

Asian to CVD. To the best of our knowledge no study on CVD risk factors has so far been published on this population and, accordingly, the truly rural and peripheral nature of this location needs to be taken into account when comparing the present data even with other Bangladeshi population groups. Similar low prevalence of CVD risk has been reported in other underdeveloped populations.15,43 There are a number of limitations of the study. The data are cross sectional and do not permit any conclusion on the trends of CVD risk factor prevalence and CVD risk over time. The associations reported cannot be assumed to be causal but do generate a hypothesis that needs to be investigated in a longitudinal studies. We used the WHO risk assessment tool to determine CVD risk in a modified manner (i.e., based on age cohort rather than selecting a known risk group like HTN or DM or Smoking). This may have contributed to an underestimation of high CVD risk participants than other studies.44,45 In this study we only identified high risk for CVD using the WHO risk assessment tool and did not have the resources to measure blood glucose and lipids to calculate absolute CVD risk. It is therefore possible that some high risk individuals have been categorised as lower risk and vice versa introducing some probable non-differential measurement error. Moreover, some of the variables (e.g., wt, ht, SBP and DBP) have been measured with single measurement and the diabetes status was determined by selfreporting. These measurement issues have also presumably introduced some non-differential error. All these nondifferential measurement errors are expected to dilute the association between CVD and its risk factors. Some risk determinants of CVD (including family history of CVDs) were not included (due to lack of reliable information or lack of logistics). Also it is possible that females may be more prone to over report symptoms and this could cause a response bias. Despite these limitations, this populationbased study with a substantial sample size provides valuable baseline information on CVD risk in a typical rural Bangladeshi population. The information provided will be of value to policy makers as well as researchers to design future strategies in this field. In conclusion, the present study findings indicate that the risk of CVDs in a relatively traditional rural population may still be substantially lower compared to that in the urbanizing rural population in Bangladesh. The difference between the two population groups seems to be present on both multiple or individual risk factor analysis. We suspect that the WHO risk assessment tool to predict CVDs has misclassified the true CVD cases to some extent, which in turn, has diluted the association with its risk factors. Thus, it is recommended to examine the validity of this tool in a longitudinal study.

Author statements Acknowledgements We are indebted to all participants in this study. Also acknowledge important contribution of the field workers, Thakurgaon Swasthoseba Hospital and the ORBIS International to make this work possible.

7

Ethics approval Human Research Ethical Committee (HREC) of the University of New South Wales (HREC ref: sHC12621), Sydney, Australia and the Ethics Review Committee of the Diabetic Association of Bangladesh (DAB).

Funding This work was supported by grants from the Bangladesh University of Health Sciences (BUHS).

Competing interests The authors declare that they have no competing interests.

Patient consent Obtained.

Data sharing statement No additional data are available.

references

1. Alwan A. Global status report on noncommunicable diseases 2010. World Health Organization; 2011. 2. Murray CJ, Lopez AD. Alternative projections of mortality and disability by cause 1990e2020: global burden of disease study. Lancet 1997;349:1498e504. ^ 3. Yusuf S, Hawken S, Ounpuu S, Dans T, Avezum A, Lanas F, et al. Effect of potentially modifiable risk factors associated with myocardial infarction in 52 countries (the INTERHEART study): case-control study. Lancet 2004;364:937e52. 4. Antman EM, Braunwald E. Acute myocardial infarction. Harrisons Princ Intern Med 2001;1:1386e98. 5. Murray CJ, Lopez AD. Global health statistics: global burden of disease and injury series, Vol. I and II. Boston: Harvard School of Public Health; 1996. ~ as A, de la Noval R, Armas N, L de la 6. Nordet P, Mendis S, Duen Noval I, et al. Total cardiovascular risk assessment and management using two prediction tools, with and without blood cholesterol. MEDICC Rev 2013;15:36e40. 7. McKeigue P, Marmot M, Cottier D, Rahman S, Riemersma R. Diabetes, hyperinsulinaemia, and coronary risk factors in Bangladeshis in east London. Br Heart J 1988;60:390e6. 8. Siddique Md Abu, MPS, Salman Mohammad, Sirajul Haque KMHS, Ahmed Md Khurshed, Sultan Md Ashraf Uddin, et al. Age-related differences of risk profile and angiographic findings in patients with coronary heart disease. BSMMU J 2010;3:5. 9. Rahman M, Zaman M. Smoking and smokeless tobacco consumption: possible risk factors for coronary heart disease among young patients attending a tertiary care cardiac hospital in Bangladesh. Public Health 2008;122:1331e8. 10. Silbiger JJ, Ashtiani R, Attari M, Spruill TM, Kamran M, Reynolds D, et al. Atheroscerlotic heart disease in Bangladeshi immigrants: risk factors and angiographic findings. Int J Cardiol 2011;146:e38e40. 11. Bhowmik B, Munir SB, Diep LM, Siddiquee T, Habib SH, Samad MA, et al. Anthropometric indicators of obesity for

Please cite this article in press as: Fatema K, et al., Cardiovascular risk assessment among rural population: findings from a cohort study in a peripheral region of Bangladesh, Public Health (2016), http://dx.doi.org/10.1016/j.puhe.2016.02.016

8

12.

13.

14.

15.

16.

17.

18.

19.

20.

21. 22. 23. 24.

25.

26.

27.

28.

p u b l i c h e a l t h x x x ( 2 0 1 6 ) 1 e8

identifying cardiometabolic risk factors in a rural Bangladeshi population. J Diabetes Investigation 2013;4:361e8. Sayeed MA, Mahtab H, Sayeed S, Begum T, Khanam PA, Banu A. Prevalence and risk factors of coronary heart disease in a rural population of Bangladesh. Ibrahim Med Coll J 2010;4:37e43. WHO. Cardiovascular Disease Programme. WHO CVD-risk management package for low-and medium-resource settings. World Health Organization; 2002. Ogedegbe G, Plange-Rhule J, Gyamfi J, Chaplin W, Ntim M, Apusiga K, et al. A cluster-randomized trial of task shifting and blood pressure control in Ghana: study protocol. Implement Sci 2014;9:73. Mendis S, Lindholm LH, Anderson SG, Alwan A, Koju R, Onwubere BJ, et al. Total cardiovascular risk approach to improve efficiency of cardiovascular prevention in resource constrain settings. J Clin Epidemiol 2011;64:1451e62. Anderson KM, Odell PM, Wilson PW, Kannel WB. Cardiovascular disease risk profiles. Am Heart J 1991;121:293e8. World Health Organization. Appropriate body-mass index for Asian populations and its implications for policy and intervention strategies. (WHO Expert Consultation). Lancet 2004;363:157. Alberti K, Zimmet P, Shaw J. Metabolic syndromeda new world-wide definition. A consensus statement from the international diabetes federation. Diabet Med 2006;23:469e80. Ashwell M, Gibson S. Waist to height ratio is a simple and effective obesity screening tool for cardiovascular risk factors: analysis of data from the British National Diet and Nutrition Survey of adults aged 19e64 years. Obes Facts 2009;2:97e103. Whitworth JA. 2003 World Health Organization (WHO)/ International Society of Hypertension (ISH) statement on management of hypertension. J Hypertens 2003;21:1983e92. Haque AN. By the numbers: the middle-income matrix; 2007. Ameen DMA. Country's literacy rate 59.82 pc; 2013 February 5. Allison P. Logistic regression using SAS: theory and application. 2nd ed. Cary, NC: SAS Institute; 2012. Second Edition ed: Inc. Ashwell M, Gunn P, Gibson S. Waist-to-height ratio is a better screening tool than waist circumference and BMI for adult cardiometabolic risk factors: systematic review and metaanalysis. Obes Rev 2012;13:275e86. Mendis S, Johnston SC, Fan W, Oladapo O, Cameron A, Faramawi MF. Cardiovascular risk management and its impact on hypertension control in primary care in lowresource settings: a cluster-randomized trial. Bull World Health Organ 2010;88:412e9. Fatema K, Natasha K, Ali L. Cardiovascular risk factors among Bangladeshi ready-made garment workers. J Public Health Afr 2014;5. Alwan A, MacLean DR, Riley LM, d'Espaignet ET, Mathers CD, Stevens GA, et al. Monitoring and surveillance of chronic noncommunicable diseases: progress and capacity in highburden countries. Lancet 2010;376:1861e8. Akhter A, Fatema K, Afroz A, Bhowmik B, Ali L, Hussain A. Prevalence of diabetes mellitus and its associated risk

29.

30. 31.

32.

33.

34.

35.

36.

37. 38.

39.

40.

41. 42.

43.

44. 45.

indicators in a rural Bangladeshi population. Open Diabetes J 2011;4. Joshi P, Islam S, Pais P, et al. RIsk factors for early myocardial infarction in south asians compared with individuals in other countries. JAMA 2007;297:286e94. Bourne PA. Paradoxes in self-evaluated health data in a developing country. North Am J Med Sci 2010;2:18. Short ME, Goetzel RZ, Pei X, Tabrizi MJ, Ozminkowski RJ, Gibson TB, et al. How accurate are self-reports? an analysis of self-reported healthcare utilization and absence when compared to administrative data. J Occup Environ Medicine/ American Coll Occup Environ Med 2009;51:786. Dodd-McCue D, Tartaglia A. Self-report response bias: learning how to live with its diagnosis in chaplaincy research. Chaplain Today 2010;26:2e8. Bajaj S, Jawad F, Islam N, Mahtab H, Bhattarai J, Shrestha D, et al. South Asian women with diabetes: psychosocial challenges and management: consensus statement. Indian J Endocrinol Metabolism 2013;17:548. Williams ED, Steptoe A, Chambers JC, Kooner JS. Psychosocial risk factors for coronary heart disease in UK South Asian men and women. J Epidemiol Community Health 2009;63:986e91. Suastika K, Dwipayana P, Saraswati M, Gotera W, Gde Budhiarta A. Coronary heart disease in a remote area. J Clin Exp Cardiol S 2012;6:2. Gupta R, Deedwania PC, Sharma K, Gupta A, Guptha S, Achari V, et al. Association of educational, occupational and socioeconomic status with cardiovascular risk factors in Asian Indians: a cross-sectional study. PloS One 2012;7:e44098. Po JY, Subramanian S. Mortality burden and socioeconomic status in India. PloS One 2011;6:e16844. Shafique S, Akhter N, Stallkamp G, de Pee S, Panagides D, Bloem MW. Trends of under-and overweight among rural and urban poor women indicate the double burden of malnutrition in Bangladesh. Int J Epidemiol 2007;36:449e57. Azimi A, Charlot MG, Torp-Pedersen C, Gislason GH, Køber L, Jensen LO, et al. Moderate overweight is beneficial and severe obesity detrimental for patients with documented atherosclerotic heart disease. Heart 2013;99:655e60. Suastika K, Dwipayana P, Ratna Saraswati I, Kuswardhani T, Astika N, Putrawan IB, et al. Relationship between age and metabolic disorders in the population of Bali. J Clin Gerontology Geriatrics 2011;2:47e52. Pearl J. Comment: understanding Simpson's Paradox. Am Statistician 2014;68:8e13. Hussain SM, Oldenburg B, Wang Y, Zoungas S, Tonkin AM. Assessment of cardiovascular disease risk in South Asian populations. Int J Vasc Med 2013;2013:786801e10. Otgontuya D, Oum S, Buckley BS, Bonita R. Assessment of total cardiovascular risk using WHO/ISH risk prediction charts in three low and middle income countries in Asia. BMC Public Health 2013;13:539. Reddy KS, Yusuf S. Emerging epidemic of cardiovascular disease in developing countries. Circulation 1998;97:596e601. Kelly BB, Fuster V. Promoting cardiovascular health in the developing world:: a critical challenge to achieve global health. National Academies Press; 2010.

Please cite this article in press as: Fatema K, et al., Cardiovascular risk assessment among rural population: findings from a cohort study in a peripheral region of Bangladesh, Public Health (2016), http://dx.doi.org/10.1016/j.puhe.2016.02.016