Application of country-specific Globorisk score to estimate next 10 years risk of cardiovascular diseases and its associated predictors among postmenopausal rural women of Bangladesh: A cross-sectional study in a primary care setting

Introduction: Risk of cardiovascular disease (CVD) among postmenopausal Bangladeshi women has not yet been evaluated using a country-specific tool. Hence, we prompted to estimate the risk and identify the predictors that were not typically included in any CVD risk assessment tool. Methods: This cross-sectional study used a web version of country-specific lab-based Globorisk calculator to estimate the risk of CVD among 265 postmenopausal women who visited a primary healthcare centre in a rural area of Bangladesh. The centre was selected purposively and the participants were recruited using a convenient sampling technique. Data were collected using a modified STEP-wise approach to surveillance of non-communicable disease risk factors questionnaire of the World Health Organization. The risk levels were presented using descriptive statistics and the associated predictors were identified using adjusted multiple linear regression analysis. Results: Overall, 56.7% of the subjects were identified as ‘at risk’ of future CVD events. After adjusting the confounders, CVD risk factors including age of onset of menopause ( β = 0.441, p < 0.001),durationofmenopause( β = 0.603, p < 0.001),smokelesstobacco use ( β = –1.047, p = 0.003), added salt intake ( β = 1.081, p = 0.002), waist–hip ratio ( β = 0.094, p = 0.03) and diastolic blood pressure ( β = 0.145, p = 0.001) were identified as significant predictors of CVD risk. Conclusion: This finding suggests screening program among postmenopausal women for early detection of CVD risk and efforts to control the associated predictors.


INTRODUCTION
Cardiovascular diseases (CVDs) are the major cause of death globally and the burden is also higher among the Bangladeshi population. 1 Regarding gender specificity, mortality from CVD is again rising among women and it has already higher than their counter-part men. 2 In women, the clean-cut difference in CVD mortality has been observed between pre-menopausal and postmenopausal age groups. 3 The global incidence rate of CVD among postmenopausal women is equal in men who are 10 years younger. 4 So all of these have suggested that men to women and pre-to postmenopausal differences in CVD mortality are possibly due to menopause, the permanent cessation of menstruation in women's lives. However, this issue is still inconclusive and exact causes need to be fully elucidated through systematic research, especially in developing countries such as Bangladesh.
A recent study showed that Bangladeshi women in rural areas had higher mortality (47-fold) from CVD than men (30-fold). 5  ). In Bangladesh, no country-specific CVD risk prediction tool has applied, and predictors not included in a tool have examined CVD risk. Hence, this study had two objectives: (1) to estimate next 10 years risk of CVD using country-specific risk prediction tool and (2) to find out the predictors of CVD risk that are not typically included in the conventional risk prediction tool.

METHODS
This was a cross-sectional study conducted in 2016 among 265 postmenopausal women, aged 40-70 years, who visited a rural primary healthcare centre of Bangladesh. The centre was selected purposively and data were collected conveniently through a face-to-face interview.
We used a pre-tested questionnaire adapted from STEP-wise approach to Surveillance (STEPS) of non-communicable diseases risk factors of World Health Organization (WHO) with appropriate modifications. 8 Details of study design, data collection procedures, physical measure-ment, tools and technique and quality assurance issues were described elaborately in previous papers. 7,9 In brief, we included those partici-

CVD risk estimation using country-specific Globorisk score
Globorisk is the first CVD risk score that predicts the risk of heart attack or stroke in healthy individuals (those who have not yet had a heart attack or stroke) for all countries globally. Previously formulated CVD risk tools were based on the data of developed or high-income countries; however, the burden of CVD mortality is higher among lowincome countries. The Globorisk uses information on a person's country of residence, age, gender, smoking, diabetes, blood pressure and cholesterol to predict the chance that they would have a heart attack or stroke in the next 10 years. It has two versions: lab-based version and office-based version. If the person does not have recent blood glucose or total cholesterol (TC) test, they can use the office-based version of Globorisk which is based on body weight and height instead. 15 This unique feature makes it possible to apply the tool in a low-resource setting where blood cholesterol measurement and/or blood glucose measurement is not possible. The CVD risk can be calculated using the web version of risk calculator or risk charts. To achieve our objective, we used the web version of lab-based Globorisk calculator that expresses CVD risk as a percentage. 16 The Globorisk tool has no categorical risk classification as like as other existing CVD risk classification system. Hence, we categorized the risk as low (<10%), moderate (10%-19.9%), high (20%-29.9%) and very high (≥30%). For intervention purpose, we further re-categorized the CVD risk as dichotomized variable such as 'at risk' and 'no risk' . Here, 'at risk' population comprised those who were at either moderate or high or very high risk.

Ethical approval
The study's purpose, the necessity of invasive procedures and data safety issues were explained to the participants. Data were collected after written informed consent was obtained.

Data processing and analysis
All questionnaires were reviewed thoroughly for consistency and completeness. For this, data were cleaned, edited and verified to exclude any error or inconsistency before coding and entering them into the

Reproductive and risk factors profile of the study population
The mean age of the study population was 53.51 ± 7.5 years and three fourth of them were at high-risk age groups (≥50 years) of CVD.
Their mean age of onset of menopause and duration of menopause was 44.83 ± 5.22 and 8.79 ± 6.45 years, respectively. Among the behavioural risk factors, more than half (58.1%) were physically inactive, 44.9% used smokeless tobacco and 44.5% used to take added salt with a meal. Majority of the study population were centrally obese (73.2%), hypertensive (41.9%) and had an abnormal lipid profile (59.6%) ( Table 1).

3.2
Risk of CVD among postmenopausal women using Globorisk score   (Table 3).

DISCUSSION
Application of country-specific tool, the Globorisk score, detected more than half of the postmenopausal women (56.7%) resided in a rural area of Bangladesh were at risk (moderate and high) of a CVD event in next 10 years. This finding was much higher than the previous study that used World Health Organization/International Society of Hypertension (WHO/ISH) 'with' and 'without' cholesterol risk charts and FRS in the same population. 7 Other than postmenopausal women, there are few studies conducted among the rural population of Bangladesh to estimate the next 10 years risk of CVD and their finding was much lower than the current finding. [17][18][19] In the current study, around 28% participants were at high risk (≥20%) which was 5.5% and 2.2% for Fatema et al. 17 and Cravedi et al. 19 among rural Bangladeshi population, respectively. Evidence on CVD risk assessment among postmenopausal women in SEA region is lacking.
Moreover, we did not find any study that applied Globorisk score to predict risk among the Asian postmenopausal women. In consideration of the women participants, the current study detected more participants as at high CVD risk compared to countries of Indian subcontinent (India, Pakistan, Nepal and Sri Lanka) and other parts of Asia (Cambodia, Malaysia and Mongolia). [20][21][22] The difference in CVD risk is possibly due to applying different tools, variation in sociodemographic factors, cultural background, environmental factors and genetics of the participant. In this regard, country-specific CVD risk prediction tool has broader applicability than the universally applied CVD risk tool due to less chance of overestimation or underestimation. From our study perspective, gender is another concern as some tools underestimate TA B L E 1 Reproductive and CVD risk factors profile of the study population (n = 265) Abbreviation: CVD, cardiovascular diseases.

Risk factors Mean ± SD Median (IQR) n (%)
CVD risk among women. 23 The application of country-specific risk tool has public health importance as most of the CVD risk prediction tools were based on the data of European descent populations in developed countries, whose background CVD risk may be significantly different from that in developing countries. An assessment of a CVD risk tool in terms of country specificity and gender specificity, FRS is the exam-ple that overestimated risk among the external population with a lower background of CVD risk 24 and underestimated risk among women. 25 Considering all of these critiques, it has been suggested recalibrating the CVD risk tools using local or country-specific data as it improved predictability among the external population. 26 Hence, we may assume that our estimated CVD risk among postmenopausal women using country-specific Globorisk score is significant and provided better estimation than the previous one. 7 Over 300 CVD risk factors have been discovered and only a few of them used to predict CVD risk. 27 Most CVD risk prediction tools Note: Multiple linear regression analysis (adjusted) was computed considering Globorisk score percentage as dependent variable and all the precisions are estimated at 95% confidence interval. Abbreviations: CVD, cardiovascular diseases; DBP, diastolic blood pressure; LDL-C, low-density lipoprotein cholesterol; TG, triglyceride; WHR, waisthip ratio. *The standardized coefficient β was statistically significant at a threshold of p < 0.05.
significant association with the risk of CVD. In this regard, a systematic review and meta-analysis of 32 studies explored the association between age of menopause and risk of CVD. The study reported a higher risk of coronary heart diseases (CHD), CVD mortality and overall mortality in women who experience premature or early-onset menopause (<45 years). 28 However, the association between duration of menopause and risk of CVD is controversial. 28 A Chinese population-based study showed that women with 2-6 years duration of menopause were at a greater risk of CVD than women less than 1 year duration, with evidence of increased risk after 6 years. 29 In contrast to this finding, two other studies reported no association between duration of menopause and CVD risk. 30,31 Among the behavioural risk factors, smokeless tobacco use (β = - with non-users. 32 However, country to country variation has been observed and contradictory outcomes were reported. 32,33 Although in our study a significant association was detected, the direction of that association contradicts with the aforementioned study. 32 Similarly, the findings of another Bangladeshi study contradict the current study as they did not found any association between smokeless tobacco use and coronary heart disease. 34 Regarding salt intake and CVD risk, data are inconsistent due to variation in measuring sodium intake and population characteristics variation. 35 Among the available studies, two large international cohort studies sought to evaluate the association of sodium intake with risk of CVD events. It has been reported that sodium intakes between 3.0 and 6.0 g/day were associated with the lowest CVD risk with higher or lower levels. 36,37 Extensive statistical analysis was conducted by the studies to control the potential confounders, reverse causality and potential biases. Still, the J-shaped association between sodium intake and CVD outcomes remained the same. Current study demonstrated significant association of CVD risk with WHR (β = 0.094, p = 0.03) as anthropometric parameter and DBP (β = 0.145, p = 0.001) as blood pressure parameter. Predictive value of WHR on CVD risk events was previously evaluated and found highly significantly associated with CHD (p < 0.01) and death (p < 0.01), myocardial infarction (p < 0.01), stroke (p < 0.01), total CVD (p < 0.001) and death (p < 0.01). In the same study, it was reported that the number of subjects exceeding a 15% risk of developing a CHD event over the next 10 years increased with increasing quintile of WHR and showed higher odds (2.60, CI: 1.09-6.54) than the other anthropometric parameters. 38 An Australian study has also reported that obesity, as measured by WHR, is a dominant, independent and predictive variable for CVD and CHD deaths. 39 Regarding DBP and CVD risk, most recent article reported that DBP could independently predict CVD risk. 40 This study reported J-curved association between DBP and CVD outcome which has both clinical and public health importance.
This is because the observed relation of DBP with CVD risk was not altered by choice of the threshold as per new 2017 USA hypertension guideline. 41 The present study had several limitations including cross-sectional design, small sample size and convenient sampling technique. As this was a self-funded Master's thesis in a specialized group of population with the limited time frame in a low-resource setting, we had to design the study as cross sectional and follow convenient sampling technique that enabled us to achieve the sample size we wanted in a relatively fast and inexpensive way. Moreover, it was difficult for us to discuss the most of the predictors as data related to the predictive value of associated risk factors on CVD risk among postmenopausal women lack global research, more specifically in Bangladesh.
Despite these limitations, this study has several positive aspects.
First, this is the first study of Bangladesh that used a country-specific CVD risk prediction tool to estimate the next 10 years risk of CVD. Second, an association of CVD risk with different risk factors not included in a CVD risk tool was examined for the first time to identify potential predictors in a low-resource setting. Third, from a gender perspective, Globorisk tool was applied for the first time among postmenopausal women residing in a rural area of Bangladesh.
A significant proportion of Bangladeshi postmenopausal rural women were classified as 'at risk' of next 10 years CVD event that demands screening program and prevention effort for this neglected population to reduce the future burden of CVD. Highly significant association with the predictors not included in a risk tool indicated the necessity of recalibration with these factors to better predict CVD risk in postmenopausal women of Bangladesh.
centre. The authors would also like to thank the participants for their kind cooperation for conducting the study.

CONFLICT OF INTEREST
The authors declare no conflict of interest.

FUNDING INFORMATION
This research received no specific grant from any funding agency in the public, commercial or not-for-profit sectors.

DATA AVAILABILITY STATEMENT
Data are available on request from the authors.